Alkaline Igneous Rocks
Geological Society Special Publications Series Editor
K. C 0 E
GEOLOGICAL
SOCIETY SPECIAL PUBLICATION
Alkaline Igneous Rocks E D I T E D BY
J. G. F I T T O N & B. G. J. U P T O N Grant Institute of Geology University of Edinburgh Edinburgh EH9 3JW
1987 Published for The Geological Society by Blackwell Scientific Publications OXFORD LONDON
EDINBURGH
BOSTON PALO ALTO MELBOURNE
N O 30
Published by Blackwell Scientific Publications Editorial offices: Osney Mead, Oxford OX2 0EL 8 John Street, London WC1N 2ES 23 Ainslie Place, Edinburgh EH3 6AJ 52 Beacon Street, Boston Massachusetts 02108, USA 667 Lytton Avenue, Palo Alto California 94301, USA 107 Barry Street, Carlton Victoria 3053, Australia
DISTRIBUTORS USA and Canada Blackwell Scientific Publications Inc PO Box 50009, Palo Alto California 94303 Australia Blackwell Scientific Publications (Australia) Pty Ltd 107 Barry Street, Carlton, Victoria 3053
9 1987 The Geological Society. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by The Geological Society for libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that a base fee of $02.00 per copy is paid directly to CCC, 27 Congress Street, Salem, MA 01970, USA. 0305-8719/87 $02.00
British Library Cataloguing in Publication Data
First published 1987
ISBN0-632-01616-7
Typeset, printed and bound in Great Britain by William Clowes Limited, Beccles and London
Library of Congress Cataloging-in-Publication Data
Alkaline igneous rocks.--(Geological Society special publications, ISSN 0305-8719) 1. Alkalic igneous rocks I. Fitton, J.G. II. Upton, B. G. J. III. Series 552'. 1 QE462.A4
Alkaline igneous rocks. (Geological Society special publication; no. 30) Bibliography: p. Includes index. 1. Alkalic igneous rocks. I. Fitton, J. G. II. Upton, B. G.J. III. Geological Society of London. IV. Series. QE462.A4A43 1987 552'. 1 86-26364 ISBN0-632-01616-7
Contents Preface Introduction: FITTON, J. G. & UPTON, B. G . J . BAILEY, D. K. Mantle metasomatism--perspective and prospect
vii ix 1
MENZIES, M. Alkaline rocks and their inclusions: a window on the Earth's interior
15
EDGAR,A. D. The genesis of alkaline magmas with emphasis on their source regions: inferences from experimental studies
29
LE Bhs, M. J. Nephelinites and carbonatites
53
TWYMAN, J. D. & GITTINS, J. Alkalic carbonatite magmas: parental or derivative ?
85
DAWSON, J. B. The kimberlite clan: relationship with olivine and leucite lamproites, and inferences for upper-mantle metasomatism
95
BERGMAN, S. C. Lamproites and other potassium-rich igneous rocks: a review of their occurrence, mineralogy and geochemistry
103
ROCK, N. M. S. The nature and origin of lamprophyres: an overview
191
CLAGUE, D. A. Hawaiian alkaline volcanism
227
WEAVER, B. L., WOOD, D. A., TARNEY, J. & JORON, J. L. Geochemistry of ocean island basalts from the South Atlantic: Ascension, Bouvet, St. Helena, Gough and Tristan da Cunha
253
HARRIS, C. & SHEPPARD, S. M. F. Magma and fluid evolution in the lavas and associated granite xenoliths of Ascension Island
269
FITTON, J. G. The Cameroon line, West Africa: a comparison between oceanic and continental alkaline volcanism
273
BAKER, B. H. Outline of the petroIogy of the Kenya rift alkaline province
293
MACDONALD,R. Quaternary peralkaline silicic rocks and caldera volcanoes of Kenya
313
WOOLLEY, A. R. & JONES, G. C. The petrochemistry of the northern part of the Chilwa alkaline province, Malawi
335
BOWDEN, P., BLACK, R., MARTIN, R. F., IKEI E. C. KINNAIRD, J. A. & BATCHELOR, R . A . Niger-Nigerian alkaline ring complexes: a classic example of African Phanerozoic anorogenic mid-plate magmatism
357
LI/~GEOIS, J. P. & BLACK, R. Alkaline magmatism subsequent to collision in the Pan-African belt of the Adrar des Iforas (Mali)
381
FLETCHER, C. J. N. & BEDDOE-STEPHENS, B. The petrology, chemistry and crystallization history of the Velasco alkaline province, eastern Bolivia
403
BARKER, D. S. Tertiary alkaline magmatism in Trans-Pecos Texas
415
EBY, G. N. The Monteregian Hills and White Mountain alkaline igneous provinces, eastern North America
433
vi
Contents
UPTON, B. G. J. & EMELEUS, C. H. Mid-Proterozoic alkaline magmatism in southern Greenland: the Gardar province
449
LARSEN, L. M. & SORENSEN, H. The Ilimaussaq intrusion--progressive crystallization and formation of layering in an agpaitic magma
473
NIELSEN,T. F. D. Tertiary alkaline magmatism in East Greenland: a review
489
DOWNES, H. Tertiary and Quaternary volcanism in the Massif Central, France
517
KOGARKO, L. N. Alkaline rocks of the eastern part of the Baltic Shield (Kola Peninsula)
531
Index
545
Preface The papers contained in this volume were presented at a symposium held in Edinburgh in September 1984, which marked the passage of ten years since the publication of The Alkaline Rocks edited by Henning Sorensen. In organizing the symposium and compiling this volume we aimed to review recent developments in the petrology and geochemistry of alkaline igneous rocks. We have, for example, paid particular attention to work on lamprophyres and carbonatites which are rock associations of current interest not covered in Sorensen's book. Reviews of recent work on some of the classic alkaline provinces, such as East Africa, southern Greenland and the Kola Peninsula, are included together with reviews of less wellknown areas. Other papers discuss the impact of experimental, geochemical and isotopic studies on our understanding of the generation and evolution of alkaline magmas. We are indebted to the contributors for their collaboration in producing this volume and it is with sadness that we note the death, on 14 February 1986, of Brian Baker, whose pioneering field studies formed the basis for much of our knowledge of the tectonic and volcanic evolution of the East African Rift. An obituary and appreciation of his work is published in the Journal of Volcanology and Geothermal Research (28, v-vii). We are grateful to The Geological Society and The Royal Society of Edinburgh for their generous assistance with the symposium costs, to Lucian Begg and Dodie James for their help with organizing the symposium and producing this volume, and to our colleagues for the care and enthusiasm with which they reviewed the manuscripts. The efforts of Edward Wates and his staff at Blackwell Scientific Publications are also gratefully acknowledged. To all these we offer our sincere thanks. J.G.F. B.G.J.U.
vii
Introduction J. G. Fitton & B. G. J. Upton Alkaline igneous rocks may be defined as those which have higher concentrations of alkalis than can be accommodated in feldspars alone, the excess appearing as feldspathoids, sodic pyroxenes, sodic amphiboles and other alkali-rich phases. These rocks are, therefore, deficient in silica and/or alumina with respect to alkalis and will have nepheline and/or acmite in their norms. In practice the term 'alkaline' is used to encompass a wide range of igneous rocks, not all of which conform to this rigid definition. Carbonatites, for example, are certainly silica-deficient but are rarely alkali-rich. True (nepheline-normative) alkali basalts grade into hypersthene-normative transitional basalts without any obvious change in mineralogy. Since transitional basalts are often closely associated with alkali basalts in the field, they are traditionally regarded as alkaline. It is now usual practice to define alkaline igneous rocks simply in terms of their alkali ( N a 2 0 + K20) and silica contents (see, for example, Le Bas et al. 1986). We have not attempted to review the classification of alkaline igneous rocks in this volume as this has been dealt with elsewhere (e.g. Sorensen 1974; Streckeisen 1967, 1980). The only alkaline rocks not covered in previous reviews are those hydrous mafic to ultramafic hypabyssal rocks known as the lamprophyres. The present volume includes three papers on this group. A comprehensive overview of lamprophyres is given by Rock and of the sub-group of lamproites by Bergman. The relationship between lamproites and kimberlites (which arguably belong to the lamprophyres) is discussed by Dawson. Volumetrically, alkaline rocks account for less than one per cent of all igneous rocks. Despite this, their remarkable mineralogical diversity has brought them repeatedly to the attention of petrologists and mineralogists, with the result that alkaline rocks account for about half of all igneous rock names. Sorensen (1974) lists no fewer than 400 alkaline rock types. This diversity springs largely from an abundance of alkalis and deficiency in silica which together generate a large number of mineral species not stable in more silica-rich, alkali-poor magmas. However, a large part of the attention given to alkaline rocks is due to their characteristic high concentrations of incompatible or large-ion lithophile elements (LILE). These are often of more than academic interest as most of the world's resources of niobium, tantalum and the rare-earth elements are found in or around alkaline igneous rock bodies. The economic importance of alkaline igneous rocks is further enhanced by their association with economic deposits of apatite (Kogarko) and with diamonds (Dawson;
Bergman). Evidence for continental alkaline magmatism can be found as far back as the late Archaean. For example, biotites from the Poohbah Lake syenite in north-western Ontario have been dated at around 2.7 Ga (Mitchell 1976) and a similar age has been reported by Larsen et al. (1983) for the Tupertalik carbonatite in western Greenland. At the present time, alkaline magmas are erupted in all tectonic environments with the possible exception of mid-ocean ridges. Even here, though, mildly alkaline lavas are sometimes erupted from off-axis volcanoes, as in the Vestmann Islands of Iceland. Alkaline igneous rocks are found on all the continents and on islands in all the ocean basins. Their occurrences may be classified on the basis of tectonic setting into three categories; continental rift valley From: FITTON, J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. ix-xiv.
ix
x
Introduction
magmatism, oceanic and continental intraplate magmatism without clear tectonic control, and alkaline magmatism related to subduction processes. In practice, however, this classification is not always easy to apply. Continental rift valleys provide, volumetrically, the most important occurrences of alkaline igneous rocks although continental rifting is not always accompanied by magmatism. The best known example (arguably the type example) is the East African Rift which, in the course of its long history, has yielded almost the entire spectrum of alkaline magmas. Three papers in this volume are devoted to various aspects of magmatism in the East African Rift. Baker reviews its magmatic associations with respect to tectonic development and discusses the origin of the magmas, particularly in relation to the problems presented by the strongly bimodal distribution of basic and salic lava compositions. He concludes that the salic magmas evolved from basic parental magmas by processes of crystal fractionation (cf. Bailey). Macdonald focusses attention on the peralkaline silicic central volcanoes of Kenya and also favours an origin by crystal fractionation for most of the evolved magmas. There is little evidence for contamination of the evolving magmas with ancient continental crust except in the case of the Naivasha comendites. It is not always possible to demonstrate a genetic link between basic and evolved magmas in the East African Rift, however. The Chilwa alkaline province in Malawi, at the southern end of the rift, is an essentially intrusive province in which salic rocks predominate. The scarcity of basic rocks in this part of the rift has led Woolley & Jones to suggest that the evolved magmas were produced directly by melting of metasomatised mantle and lower crust. Insight into the processes occurring at depth beneath rift valleys may be gained by studying ancient and deeply eroded examples. The Proterozoic Gardar province in SouthWest Greenland is probably the best studied of these and is reviewed by Upton & Emeleus. One of the most striking features of the province is the presence of giant dykes, up to 800 metres wide. These are dominantly basic but in places show in situ differentiation into more salic rocks. Salic magma generated in the wider portions of these dykes migrated upwards and may ultimately have accumulated to produce central complexes in which basic magma was subordinate or absent (e.g. the Ilimaussaq intrusion, Larsen & Serensen). The giant dykes, therefore, play a crucial role in understanding the relationship between basic and salic magmas in this and possibly other rift systems. The separation of continents to form ocean basins must always be preceded by a phase of continental rifting, leaving volcanic and intrusive complexes stranded along passive continental margins. The vigorous magmatism which accompanies continental separation is generally tholeiitic in character as, for example, in the Karoo and Deccan flood basalt provinces. Alkaline magmas, however, may be emplaced along the trailing continental margin during the waning phase of magmatism, long after the spreading centre is established off-shore. The Tertiary volcanic rocks exposed along the east coast of Greenland (described by Nielsen) provide an excellent example of such an alkaline province. A second major occurrence of alkaline igneous rocks is provided by intraplate magmatic provinces whose activity and siting are not subject to any obvious tectonic control. In the ocean basins such magmatism manifests itself as ocean islands which are sometimes aligned in chains with ages increasing away from the active centres, as in the Hawaiian islands. In these cases it is possible to relate the magmatism to convective plumes within the asthenosphere. The Hawaiian islands (reviewed by Clague) show clearly defined
Introduction
xi
magmatic cycles starting with alkaline magmas (represented by Loihi seamount), passing through a voluminous tholeiitic shield-building stage and returning to alkaline magmatism during the waning phases of individual volcanic centres. These cycles seem to result from movement of the oceanic plate over a rising plume of partially molten asthenosphere in which the degree of melting increases towards the centre. They may be typical of ocean islands in general and are broadly analogous to similar cycles seen in flood basalt provinces preserved on passive continental margins. Continental intraplate magmatism may also show age progressions as in the NigerNigeria province (Bowden et al.). These progressions are, however, very rare and not so clearly defined as in ocean island chains, probably because the continental lithosphere is thicker and less easily penetrated than oceanic lithosphere. The Monteregian Hills and White Mountain provinces of eastern North America (Eby), for example, show no obvious progression but their seaward extension, the New England seamounts, show a regular decrease in age eastwards. Reviews of other continental intraplate alkaline provinces are given by Fletcher & Beddoe-Stevens (Velasco province, Bolivia) and Kogarko (Kola Peninsula). The Cameroon line in West Africa (Fitton) includes both continental and oceanic alkaline volcanic centres. None of these examples shows any clear progression of ages. Some continental provinces undergo repeated alkaline magmatism in one place over long periods. For example, the Kola Peninsula (Kogarko) was the site of alkaline magmatism in the mid Proterozoic and again in the Devonian. Such examples could be the result of coincidence but are more likely the result of the repeated exploitation of zones of weakness in the lithosphere. Destructive plate boundaries provide the third tectonic setting in which alkaline igneous rocks may occur. During the life of a subduction zone the characteristic calc-alkaline magmas tend to become more potassic with time and may give way to volcanic rocks of the shoshonitic association, some members of which may contain leucite. A discussion of subduction zone processes is beyond the scope of this volume and the reader is referred to the reviews of Gill (1981) and Ewart (1982). There are, however, two circumstances under which subduction processes can lead to the generation of more 'normal' alkaline magmas. Once the descending slab has become dehydrated at depth it loses its capacity to stimulate the generation of calc-alkaline magmas but it can still cause melting in the overlying asthenosphere. This can lead to the production of alkaline magmas from the mantle above the deepest parts of subduction zones. One such example of alkaline magmas erupted under a compressive regime is provided by the Trans-Pecos province of west Texas (Barker). The alkaline rocks in this area grade south-westwards into the calc-alkaline rocks of the Sierra Madre Occidental in Mexico and Barker relates both suites to subduction of the Farallon Plate. After the cessation of subduction, relaxation of the former compressive regime often results in extension and the generation of alkaline magmas. The resulting switch from subduction-related calc-alkaline to extensional alkaline magmatism appears to be a common phenomenon. It occurred, for example, in the western U.S.A. about 17 Ma ago and in parts of Africa and Arabia at the end of the Pan-African metamorphic episode during the late Precambrian. An example from the Pan-African belt in Mali is discussed by Liegeois & Black. The origin of alkaline magmas has attracted a great deal of interest among igneous petrologists over the last ten years or so. This interest has been stimulated by two important
xii
Introduction
and characteristic features of alkaline rocks. Firstly, they contain high concentrations of LILE and yet isotopic evidence suggests that their parental magmas had a mantle source which had been depleted in these elements for a long time. Thus alkaline igneous rocks may provide useful information about enrichment and/or melting processes in the mantle. Secondly, many mafic alkaline volcanic rocks contain xenoliths inferred to have originated within the mantle (Menzies). These are often enriched in LILE when compared with concentrations expected of chondritic mantle material and sometimes contain amphiboles and micas of metasomatic origin (Bailey). Such clear evidence for the existence of metasomatically enriched mantle, coupled with the problem of extracting LILE-rich magmas from LILE-poor mantle, has led to hypotheses involving mantle metasomatism as a precursor to alkaline magmatism. These hypotheses, reviewed by Bailey, have gained popularity over the past fifteen years and are invoked by several contributors to this volume. Mantle metasomatism neatly explains many of the features of alkaline magmatism. For example, the frequent association of alkaline magmatism with areas of large-scale regional uplift is consistent with the relatively low density of metasomatized mantle. An essential feature of all models involving a metasomatized mantle source for alkaline magmas is that this source must lie in the lithosphere. This is the only part of the mantle where enriched material can remain in one place for long periods without being swept away by convection. The lithospheric mantle beneath the continents is likely to be chemically and isotopically different from that beneath the oceans. Continental lithospheric mantle is old and will have had as complex a metamorphic and magmatic history as the overlying crust. Oceanic lithosphere, on the other hand, is relatively young and probably depleted in LILE. These differences should be reflected in the compositions of continental and oceanic alkaline rocks. However, alkali basalts erupted in continental and oceanic settings are generally identical both chemically (Fitton) and isotopically (Menzies). Since enriched mantle xenoliths are only commonly found in continental regions it follows that the enriched lithospheric mantle represented by these xenoliths is not the source of most continental and oceanic alkaline magmas. An asthenospheric source is therefore implied. This is not to say that enriched continental lithospheric mantle is never involved in the generation of alkaline magmas. There is good evidence (e.g. Edgar) that pockets of ancient enriched mantle beneath cratonic regions provide the source for LILE-rich mafic and ultramafic alkaline rocks such as micaceous kimberlite (Dawson), and lamproite and other potassic igneous rocks (Bergman). It is significant that these rock types are exclusively continental. More extensive melting may involve the continental lithosphere mantle in the production of less exotic rock types such as flood tholeiite and mildly alkaline basalt. Upton & Emeleus, for example, argue for a lithospheric mantle source for the Gardar alkaline magmas. If most alkaline magmas have an ultimate source in the asthenosphere then they must share this source with unequivocally asthenosphere-derived rocks such as mid-ocean ridge basalt (MORB). The consistent isotopic differences between MORB and alkali basalts (and indeed all intraplate basalts) requires that the asthenosphere be heterogeneous. This heterogeneity may result from the entrainment of lower mantle material in deep mantle plumes as suggested for the Hawaiian island chain (Clague). The entire convecting upper mantle may also be heterogeneous on a small scale and alkaline magmas may be generated by the selective melting of LILE-enriched streaks while more extensive melting produces MORB (Fitton). Geochemical studies on ocean island basalts from the South Atlantic
Introduction
xiii
(Weaver et a/.) suggest that one component of these enriched streaks is provided by subducted ocean-floor sediment. Derivation of LILE-rich magmas from an asthenospheric source depleted in these elements requires either very small degress of partial melting or extensive crystal fractionation. There can be no doubt that many alkaline rocks are the products of extensive low pressure crystal fractionation but this cannot be true of those alkaline rocks which host mantle xenoliths. Even the most magnesian alkaline rocks, which must represent nearprimary magmas, are rich in LILE. Small-degree partial melting ( < l ~ ) is therefore required to produce such magmas. McKenzie (1985) has recently shown that the extraction of melt fractions as small as 0.2~ is not only physically possible but inevitable where the melt viscosity is low, as it probably is in the case of alkaline magmas. Experimental studies on alkaline rocks and synthetic analogues (reviewed by Edgar) provide useful constraints on the feasibility of fractional crystallization and partial melting models and on the temperatures and pressures involved. Many lines of evidence suggest that volatile components form a significant part of alkaline magmas. The development of extensive zones of metasomatised country rock (fenite) around alkaline plutons, the abundance of chlorine and fluorine in some alkaline igneous rocks, and the frequently explosive eruption of alkaline magma all point to high concentrations of volatiles. These volatile components play an important role in the evolution of alkaline magmas and yet relatively little is known about them. Constraints on their composition have been provided by fluid inclusion studies (Harris & Sheppard) and by thermodynamic considerations (Kogarko). The effects of volatile components on the evolution of alkaline magmas can be seen clearly in both intrusive and extrusive rocks. Larsen & Serensen, for example, discuss the crystallization history of the Ilimaussaq intrusion in South-West Greenland and show how the upward migration of low-density, low-viscosity volatile-rich magma delayed crystallization under the roof of the intrusion. Silicic alkaline pyroclastic deposits around central volcanoes in Kenya often show striking variations in the abundance of some incompatible elements within a single vertical section, implying compositional zonation in the magma chamber before eruption. Macdonald shows that these variations are too large to be accounted for by crystal fractionation alone and suggests that some elements have been transported to the magma chamber roof zones as complex ions in a volatile phase. Carbonatites provide perhaps the best illustration of all of the influence of volatile components on the origin and evolution of alkaline rocks. There is now a consensus that their parental magmas originate by the separation of an immiscible carbonate liquid phase from a CO2-saturated nephelinite or phonolite magma. There is, however, some disagreement over the nature and subsequent evolution of this parental carbonate magma. Le Bas argues that the parental magma is rich in alkalis and similar in composition to the natrocarbonatite lavas erupted from Oldoinyo Lengai. This magma evolves at low pressure towards the more common calcite carbonatite (s6vite) by loss of alkalis to the surrounding country rocks which are metasomatized (fenitized) as a result. Twyman & Gittins offer an alternative scheme in which s6vite magmas are parental and natrocarbonatite magmas are derived from them by crystal fractionation. Most petrologists now believe that evolved alkaline magmas are produced by the fractional crystallization of basic magma. The more highly undersaturated parental magmas represented by basanite and nephelinite will produce undersaturated derivatives
xiv
Introduction
such as phonolite and foyaite. Mildly alkaline and transitional basalt m a g m a is likely to produce trachyte and, with extreme fractionation, alkali rhyolite. The production of peralkaline acid rocks by crystal fractionation alone is likely to be an inefficient process. The operation of this process, however, is clearly demonstrated by their association with transitional basalt on some ocean islands, such as Ascension (Harris & Sheppard). Peralkaline acid rocks only occur in a b u n d a n c e in continental environments, however, and here there is often good evidence that crustal contamination has a c c o m p a n i e d crystal fractionation. Well-documented examples of the operation of crustal c o n t a m i n a t i o n in the evolution of alkaline m a g m a s are presented by several of the contributors to this volume. Downes, for example, shows that the assimilation of lower crustal granulite has affected the evolution of alkaline m a g m a s in the F r e n c h Massif Central and uses isotope data to estimate the extent of this contamination. Other examples are presented by Bowden et al. ( N i g e r - N i g e r i a granite ring complexes), Eby (Monteregian Hills and White Mountain provinces, N o r t h America), Fitton (Cameroon line, West Africa) and Fletcher & BeddoeStevens (Velasco province, Bolivia). Other authors propose the derivation of evolved alkaline m a g m a s directly from m e t a s o m a t i z e d mantle or lower crust (Bailey; Woolley &
Jones). Our understanding of the origin and evolution of alkaline m a g m a s has come a long way since the publication of Sorensen's book in 1974, largely through the acquisition of a far larger geochemical and isotopic data base. The contributions to this volume review the current state of this understanding. Emphasis has shifted from crustal to mantle processes with the recognition of mantle m e t a s o m a t i s m and its possible role as a precursor to alkaline magmatism. More recently, though, there has been a swing towards the opposite view, that mantle m e t a s o m a t i s m is c a u s e d b y alkaline m a g m a t i s m . Theoretical and experimental studies on the migration and segregation of small-degree melts seem destined to accelerate this swing. Despite these advances, however, m a n y mysteries remain unsolved and alkaline rocks will still provide a fruitful field of research for m a n y years to come, yielding further insights into the nature of mantle processes and the evolution of magmas.
References EWART, A. 1982. The mineralogy and petrology of Tertiary-Recent orogenic volcanic rocks: with special reference to the andesitic-basaltic compositional range. In Thorpe, R. S. (ed.) Andesites. Pp. 25-87. John Wiley & Sons, London. GILL,J. B. 1981. Orogenic Andesites and Plate Tectonics, 390 pp. Springer-Verlag, Berlin. LARSEN,L. M., REX,D. C. & SECHER,K. 1983. The age of carbonatites, kimberlites and lamprophyres from southern west Greenland: recurrent alkaline magmatism during 2500 million years. Lithos 16, 215-21. LE BAS,M. J., LE MAITRE, R. W., STRECKEISEN,A. & ZANETTIN, B. 1986. A chemical classification of
volcanic rocks based on the total alkali--silica diagram. J. Petrol. 27, 745-50. MCKENZIE, D. 1985. The extraction of magma from the crust and the mantle. Earth planet. Sci. Lett. 74, 81-91. MITCHELL,R. H. 1976. Potassium-argon geochronology of the Poohbah Lake alkaline complex, northwestern Ontario. Can. J. Earth Sci. 13, 1456-9. SORENSEN,H. (ed.) 1974. The Alkaline Rocks. 622 pp. John Wiley & Sons, London. STRECKEISEN,A. 1967. Classification and nomenclature of igneous rocks. N. Jb. Miner. Abh. 107, 144-240. - - , 1980. Classificationand nomenclature of volcanic rocks, lamprophyres, carbonatites and melilitic rocks. Geol. Rundschau, 69, 194-207.
J. G. FITTON& B. G. J. UPTON,Grant Institute of Geology, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, U.K.
Introduction J. G. Fitton & B. G. J. Upton Alkaline igneous rocks may be defined as those which have higher concentrations of alkalis than can be accommodated in feldspars alone, the excess appearing as feldspathoids, sodic pyroxenes, sodic amphiboles and other alkali-rich phases. These rocks are, therefore, deficient in silica and/or alumina with respect to alkalis and will have nepheline and/or acmite in their norms. In practice the term 'alkaline' is used to encompass a wide range of igneous rocks, not all of which conform to this rigid definition. Carbonatites, for example, are certainly silica-deficient but are rarely alkali-rich. True (nepheline-normative) alkali basalts grade into hypersthene-normative transitional basalts without any obvious change in mineralogy. Since transitional basalts are often closely associated with alkali basalts in the field, they are traditionally regarded as alkaline. It is now usual practice to define alkaline igneous rocks simply in terms of their alkali ( N a 2 0 + K20) and silica contents (see, for example, Le Bas et al. 1986). We have not attempted to review the classification of alkaline igneous rocks in this volume as this has been dealt with elsewhere (e.g. Sorensen 1974; Streckeisen 1967, 1980). The only alkaline rocks not covered in previous reviews are those hydrous mafic to ultramafic hypabyssal rocks known as the lamprophyres. The present volume includes three papers on this group. A comprehensive overview of lamprophyres is given by Rock and of the sub-group of lamproites by Bergman. The relationship between lamproites and kimberlites (which arguably belong to the lamprophyres) is discussed by Dawson. Volumetrically, alkaline rocks account for less than one per cent of all igneous rocks. Despite this, their remarkable mineralogical diversity has brought them repeatedly to the attention of petrologists and mineralogists, with the result that alkaline rocks account for about half of all igneous rock names. Sorensen (1974) lists no fewer than 400 alkaline rock types. This diversity springs largely from an abundance of alkalis and deficiency in silica which together generate a large number of mineral species not stable in more silica-rich, alkali-poor magmas. However, a large part of the attention given to alkaline rocks is due to their characteristic high concentrations of incompatible or large-ion lithophile elements (LILE). These are often of more than academic interest as most of the world's resources of niobium, tantalum and the rare-earth elements are found in or around alkaline igneous rock bodies. The economic importance of alkaline igneous rocks is further enhanced by their association with economic deposits of apatite (Kogarko) and with diamonds (Dawson;
Bergman). Evidence for continental alkaline magmatism can be found as far back as the late Archaean. For example, biotites from the Poohbah Lake syenite in north-western Ontario have been dated at around 2.7 Ga (Mitchell 1976) and a similar age has been reported by Larsen et al. (1983) for the Tupertalik carbonatite in western Greenland. At the present time, alkaline magmas are erupted in all tectonic environments with the possible exception of mid-ocean ridges. Even here, though, mildly alkaline lavas are sometimes erupted from off-axis volcanoes, as in the Vestmann Islands of Iceland. Alkaline igneous rocks are found on all the continents and on islands in all the ocean basins. Their occurrences may be classified on the basis of tectonic setting into three categories; continental rift valley From: FITTON, J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. ix-xiv.
ix
x
Introduction
magmatism, oceanic and continental intraplate magmatism without clear tectonic control, and alkaline magmatism related to subduction processes. In practice, however, this classification is not always easy to apply. Continental rift valleys provide, volumetrically, the most important occurrences of alkaline igneous rocks although continental rifting is not always accompanied by magmatism. The best known example (arguably the type example) is the East African Rift which, in the course of its long history, has yielded almost the entire spectrum of alkaline magmas. Three papers in this volume are devoted to various aspects of magmatism in the East African Rift. Baker reviews its magmatic associations with respect to tectonic development and discusses the origin of the magmas, particularly in relation to the problems presented by the strongly bimodal distribution of basic and salic lava compositions. He concludes that the salic magmas evolved from basic parental magmas by processes of crystal fractionation (cf. Bailey). Macdonald focusses attention on the peralkaline silicic central volcanoes of Kenya and also favours an origin by crystal fractionation for most of the evolved magmas. There is little evidence for contamination of the evolving magmas with ancient continental crust except in the case of the Naivasha comendites. It is not always possible to demonstrate a genetic link between basic and evolved magmas in the East African Rift, however. The Chilwa alkaline province in Malawi, at the southern end of the rift, is an essentially intrusive province in which salic rocks predominate. The scarcity of basic rocks in this part of the rift has led Woolley & Jones to suggest that the evolved magmas were produced directly by melting of metasomatised mantle and lower crust. Insight into the processes occurring at depth beneath rift valleys may be gained by studying ancient and deeply eroded examples. The Proterozoic Gardar province in SouthWest Greenland is probably the best studied of these and is reviewed by Upton & Emeleus. One of the most striking features of the province is the presence of giant dykes, up to 800 metres wide. These are dominantly basic but in places show in situ differentiation into more salic rocks. Salic magma generated in the wider portions of these dykes migrated upwards and may ultimately have accumulated to produce central complexes in which basic magma was subordinate or absent (e.g. the Ilimaussaq intrusion, Larsen & Serensen). The giant dykes, therefore, play a crucial role in understanding the relationship between basic and salic magmas in this and possibly other rift systems. The separation of continents to form ocean basins must always be preceded by a phase of continental rifting, leaving volcanic and intrusive complexes stranded along passive continental margins. The vigorous magmatism which accompanies continental separation is generally tholeiitic in character as, for example, in the Karoo and Deccan flood basalt provinces. Alkaline magmas, however, may be emplaced along the trailing continental margin during the waning phase of magmatism, long after the spreading centre is established off-shore. The Tertiary volcanic rocks exposed along the east coast of Greenland (described by Nielsen) provide an excellent example of such an alkaline province. A second major occurrence of alkaline igneous rocks is provided by intraplate magmatic provinces whose activity and siting are not subject to any obvious tectonic control. In the ocean basins such magmatism manifests itself as ocean islands which are sometimes aligned in chains with ages increasing away from the active centres, as in the Hawaiian islands. In these cases it is possible to relate the magmatism to convective plumes within the asthenosphere. The Hawaiian islands (reviewed by Clague) show clearly defined
Introduction
xi
magmatic cycles starting with alkaline magmas (represented by Loihi seamount), passing through a voluminous tholeiitic shield-building stage and returning to alkaline magmatism during the waning phases of individual volcanic centres. These cycles seem to result from movement of the oceanic plate over a rising plume of partially molten asthenosphere in which the degree of melting increases towards the centre. They may be typical of ocean islands in general and are broadly analogous to similar cycles seen in flood basalt provinces preserved on passive continental margins. Continental intraplate magmatism may also show age progressions as in the NigerNigeria province (Bowden et al.). These progressions are, however, very rare and not so clearly defined as in ocean island chains, probably because the continental lithosphere is thicker and less easily penetrated than oceanic lithosphere. The Monteregian Hills and White Mountain provinces of eastern North America (Eby), for example, show no obvious progression but their seaward extension, the New England seamounts, show a regular decrease in age eastwards. Reviews of other continental intraplate alkaline provinces are given by Fletcher & Beddoe-Stevens (Velasco province, Bolivia) and Kogarko (Kola Peninsula). The Cameroon line in West Africa (Fitton) includes both continental and oceanic alkaline volcanic centres. None of these examples shows any clear progression of ages. Some continental provinces undergo repeated alkaline magmatism in one place over long periods. For example, the Kola Peninsula (Kogarko) was the site of alkaline magmatism in the mid Proterozoic and again in the Devonian. Such examples could be the result of coincidence but are more likely the result of the repeated exploitation of zones of weakness in the lithosphere. Destructive plate boundaries provide the third tectonic setting in which alkaline igneous rocks may occur. During the life of a subduction zone the characteristic calc-alkaline magmas tend to become more potassic with time and may give way to volcanic rocks of the shoshonitic association, some members of which may contain leucite. A discussion of subduction zone processes is beyond the scope of this volume and the reader is referred to the reviews of Gill (1981) and Ewart (1982). There are, however, two circumstances under which subduction processes can lead to the generation of more 'normal' alkaline magmas. Once the descending slab has become dehydrated at depth it loses its capacity to stimulate the generation of calc-alkaline magmas but it can still cause melting in the overlying asthenosphere. This can lead to the production of alkaline magmas from the mantle above the deepest parts of subduction zones. One such example of alkaline magmas erupted under a compressive regime is provided by the Trans-Pecos province of west Texas (Barker). The alkaline rocks in this area grade south-westwards into the calc-alkaline rocks of the Sierra Madre Occidental in Mexico and Barker relates both suites to subduction of the Farallon Plate. After the cessation of subduction, relaxation of the former compressive regime often results in extension and the generation of alkaline magmas. The resulting switch from subduction-related calc-alkaline to extensional alkaline magmatism appears to be a common phenomenon. It occurred, for example, in the western U.S.A. about 17 Ma ago and in parts of Africa and Arabia at the end of the Pan-African metamorphic episode during the late Precambrian. An example from the Pan-African belt in Mali is discussed by Liegeois & Black. The origin of alkaline magmas has attracted a great deal of interest among igneous petrologists over the last ten years or so. This interest has been stimulated by two important
xii
Introduction
and characteristic features of alkaline rocks. Firstly, they contain high concentrations of LILE and yet isotopic evidence suggests that their parental magmas had a mantle source which had been depleted in these elements for a long time. Thus alkaline igneous rocks may provide useful information about enrichment and/or melting processes in the mantle. Secondly, many mafic alkaline volcanic rocks contain xenoliths inferred to have originated within the mantle (Menzies). These are often enriched in LILE when compared with concentrations expected of chondritic mantle material and sometimes contain amphiboles and micas of metasomatic origin (Bailey). Such clear evidence for the existence of metasomatically enriched mantle, coupled with the problem of extracting LILE-rich magmas from LILE-poor mantle, has led to hypotheses involving mantle metasomatism as a precursor to alkaline magmatism. These hypotheses, reviewed by Bailey, have gained popularity over the past fifteen years and are invoked by several contributors to this volume. Mantle metasomatism neatly explains many of the features of alkaline magmatism. For example, the frequent association of alkaline magmatism with areas of large-scale regional uplift is consistent with the relatively low density of metasomatized mantle. An essential feature of all models involving a metasomatized mantle source for alkaline magmas is that this source must lie in the lithosphere. This is the only part of the mantle where enriched material can remain in one place for long periods without being swept away by convection. The lithospheric mantle beneath the continents is likely to be chemically and isotopically different from that beneath the oceans. Continental lithospheric mantle is old and will have had as complex a metamorphic and magmatic history as the overlying crust. Oceanic lithosphere, on the other hand, is relatively young and probably depleted in LILE. These differences should be reflected in the compositions of continental and oceanic alkaline rocks. However, alkali basalts erupted in continental and oceanic settings are generally identical both chemically (Fitton) and isotopically (Menzies). Since enriched mantle xenoliths are only commonly found in continental regions it follows that the enriched lithospheric mantle represented by these xenoliths is not the source of most continental and oceanic alkaline magmas. An asthenospheric source is therefore implied. This is not to say that enriched continental lithospheric mantle is never involved in the generation of alkaline magmas. There is good evidence (e.g. Edgar) that pockets of ancient enriched mantle beneath cratonic regions provide the source for LILE-rich mafic and ultramafic alkaline rocks such as micaceous kimberlite (Dawson), and lamproite and other potassic igneous rocks (Bergman). It is significant that these rock types are exclusively continental. More extensive melting may involve the continental lithosphere mantle in the production of less exotic rock types such as flood tholeiite and mildly alkaline basalt. Upton & Emeleus, for example, argue for a lithospheric mantle source for the Gardar alkaline magmas. If most alkaline magmas have an ultimate source in the asthenosphere then they must share this source with unequivocally asthenosphere-derived rocks such as mid-ocean ridge basalt (MORB). The consistent isotopic differences between MORB and alkali basalts (and indeed all intraplate basalts) requires that the asthenosphere be heterogeneous. This heterogeneity may result from the entrainment of lower mantle material in deep mantle plumes as suggested for the Hawaiian island chain (Clague). The entire convecting upper mantle may also be heterogeneous on a small scale and alkaline magmas may be generated by the selective melting of LILE-enriched streaks while more extensive melting produces MORB (Fitton). Geochemical studies on ocean island basalts from the South Atlantic
Introduction
xiii
(Weaver et a/.) suggest that one component of these enriched streaks is provided by subducted ocean-floor sediment. Derivation of LILE-rich magmas from an asthenospheric source depleted in these elements requires either very small degress of partial melting or extensive crystal fractionation. There can be no doubt that many alkaline rocks are the products of extensive low pressure crystal fractionation but this cannot be true of those alkaline rocks which host mantle xenoliths. Even the most magnesian alkaline rocks, which must represent nearprimary magmas, are rich in LILE. Small-degree partial melting ( < l ~ ) is therefore required to produce such magmas. McKenzie (1985) has recently shown that the extraction of melt fractions as small as 0.2~ is not only physically possible but inevitable where the melt viscosity is low, as it probably is in the case of alkaline magmas. Experimental studies on alkaline rocks and synthetic analogues (reviewed by Edgar) provide useful constraints on the feasibility of fractional crystallization and partial melting models and on the temperatures and pressures involved. Many lines of evidence suggest that volatile components form a significant part of alkaline magmas. The development of extensive zones of metasomatised country rock (fenite) around alkaline plutons, the abundance of chlorine and fluorine in some alkaline igneous rocks, and the frequently explosive eruption of alkaline magma all point to high concentrations of volatiles. These volatile components play an important role in the evolution of alkaline magmas and yet relatively little is known about them. Constraints on their composition have been provided by fluid inclusion studies (Harris & Sheppard) and by thermodynamic considerations (Kogarko). The effects of volatile components on the evolution of alkaline magmas can be seen clearly in both intrusive and extrusive rocks. Larsen & Serensen, for example, discuss the crystallization history of the Ilimaussaq intrusion in South-West Greenland and show how the upward migration of low-density, low-viscosity volatile-rich magma delayed crystallization under the roof of the intrusion. Silicic alkaline pyroclastic deposits around central volcanoes in Kenya often show striking variations in the abundance of some incompatible elements within a single vertical section, implying compositional zonation in the magma chamber before eruption. Macdonald shows that these variations are too large to be accounted for by crystal fractionation alone and suggests that some elements have been transported to the magma chamber roof zones as complex ions in a volatile phase. Carbonatites provide perhaps the best illustration of all of the influence of volatile components on the origin and evolution of alkaline rocks. There is now a consensus that their parental magmas originate by the separation of an immiscible carbonate liquid phase from a CO2-saturated nephelinite or phonolite magma. There is, however, some disagreement over the nature and subsequent evolution of this parental carbonate magma. Le Bas argues that the parental magma is rich in alkalis and similar in composition to the natrocarbonatite lavas erupted from Oldoinyo Lengai. This magma evolves at low pressure towards the more common calcite carbonatite (s6vite) by loss of alkalis to the surrounding country rocks which are metasomatized (fenitized) as a result. Twyman & Gittins offer an alternative scheme in which s6vite magmas are parental and natrocarbonatite magmas are derived from them by crystal fractionation. Most petrologists now believe that evolved alkaline magmas are produced by the fractional crystallization of basic magma. The more highly undersaturated parental magmas represented by basanite and nephelinite will produce undersaturated derivatives
xiv
Introduction
such as phonolite and foyaite. Mildly alkaline and transitional basalt m a g m a is likely to produce trachyte and, with extreme fractionation, alkali rhyolite. The production of peralkaline acid rocks by crystal fractionation alone is likely to be an inefficient process. The operation of this process, however, is clearly demonstrated by their association with transitional basalt on some ocean islands, such as Ascension (Harris & Sheppard). Peralkaline acid rocks only occur in a b u n d a n c e in continental environments, however, and here there is often good evidence that crustal contamination has a c c o m p a n i e d crystal fractionation. Well-documented examples of the operation of crustal c o n t a m i n a t i o n in the evolution of alkaline m a g m a s are presented by several of the contributors to this volume. Downes, for example, shows that the assimilation of lower crustal granulite has affected the evolution of alkaline m a g m a s in the F r e n c h Massif Central and uses isotope data to estimate the extent of this contamination. Other examples are presented by Bowden et al. ( N i g e r - N i g e r i a granite ring complexes), Eby (Monteregian Hills and White Mountain provinces, N o r t h America), Fitton (Cameroon line, West Africa) and Fletcher & BeddoeStevens (Velasco province, Bolivia). Other authors propose the derivation of evolved alkaline m a g m a s directly from m e t a s o m a t i z e d mantle or lower crust (Bailey; Woolley &
Jones). Our understanding of the origin and evolution of alkaline m a g m a s has come a long way since the publication of Sorensen's book in 1974, largely through the acquisition of a far larger geochemical and isotopic data base. The contributions to this volume review the current state of this understanding. Emphasis has shifted from crustal to mantle processes with the recognition of mantle m e t a s o m a t i s m and its possible role as a precursor to alkaline magmatism. More recently, though, there has been a swing towards the opposite view, that mantle m e t a s o m a t i s m is c a u s e d b y alkaline m a g m a t i s m . Theoretical and experimental studies on the migration and segregation of small-degree melts seem destined to accelerate this swing. Despite these advances, however, m a n y mysteries remain unsolved and alkaline rocks will still provide a fruitful field of research for m a n y years to come, yielding further insights into the nature of mantle processes and the evolution of magmas.
References EWART, A. 1982. The mineralogy and petrology of Tertiary-Recent orogenic volcanic rocks: with special reference to the andesitic-basaltic compositional range. In Thorpe, R. S. (ed.) Andesites. Pp. 25-87. John Wiley & Sons, London. GILL,J. B. 1981. Orogenic Andesites and Plate Tectonics, 390 pp. Springer-Verlag, Berlin. LARSEN,L. M., REX,D. C. & SECHER,K. 1983. The age of carbonatites, kimberlites and lamprophyres from southern west Greenland: recurrent alkaline magmatism during 2500 million years. Lithos 16, 215-21. LE BAS,M. J., LE MAITRE, R. W., STRECKEISEN,A. & ZANETTIN, B. 1986. A chemical classification of
volcanic rocks based on the total alkali--silica diagram. J. Petrol. 27, 745-50. MCKENZIE, D. 1985. The extraction of magma from the crust and the mantle. Earth planet. Sci. Lett. 74, 81-91. MITCHELL,R. H. 1976. Potassium-argon geochronology of the Poohbah Lake alkaline complex, northwestern Ontario. Can. J. Earth Sci. 13, 1456-9. SORENSEN,H. (ed.) 1974. The Alkaline Rocks. 622 pp. John Wiley & Sons, London. STRECKEISEN,A. 1967. Classification and nomenclature of igneous rocks. N. Jb. Miner. Abh. 107, 144-240. - - , 1980. Classificationand nomenclature of volcanic rocks, lamprophyres, carbonatites and melilitic rocks. Geol. Rundschau, 69, 194-207.
J. G. FITTON& B. G. J. UPTON,Grant Institute of Geology, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, U.K.
Mantle metasomatism
perspective and prospect
D. K. Bailey S U M M A R Y : Mantle replenishment in lithophile elements has been discerned in the patterns of trace elements and isotopes in lavas. One replenishment process is identified as metasomatic replacement, seen in ultramafic xenoliths brought up in high-velocity alkaline eruptions. Thus alkaline magmatism provides the best primafacie evidence of metasomatism and open-system conditions in the upper mantle. The list of added lithophile elements includes the following: H, C, F, Na, A1, P, S, C1, K, Ca, Ti, Fe, Rb, Y, Zr, Nb, Ba and rare earths. Some metasomatism may be due to wall-rock alteration near magma bodies, but the evidence for metasomatism prior to melting opens the possibility that the process is a precursor to alkaline magmatism, giving the necessary source enrichment in lithophile elements. In some igneous provinces the metasomafism is widespread, intensive and pervasive; in others it appears as veining of variable intensity. Metasomatism as a largescale process is best indicated by the widespread distribution of alkaline magmatism in space and time: volatile flux through the lithosphere would then be the necessary precursor of metasomatism and magmatism. Volatile activity, metasomatism and melt enrichment clearly widen the scope for mafic magma generation in the mantle, but some long-standing problems of the alkaline associations (and indeed the calc-alkaline) also call for re-examination in terms of volatile activity in the lithosphere mantle. These include the diversity of magmas, the generation of large felsic volumes, and composition gaps in magma series. Experiments show that felsic minerals are stable to 30 kb, indicating the possibility of felsic-melt generation in the upper part of the mantle. A combination of volatile flux and melt percolation along geotherms that intersect the solidus at depths of less than 80 km would lead to enrichmentand metasomatism, providing distinct mantle sources for felsic magmas. Initial (or residual) melts from such a region, as distinct from those from greater depths, would be constrained by equilibria involving felsic minerals. Thus an igneous cycle could generate a bimodal association, with felsic melts forming in the upper regime and the mafic melts originating at depths below the range of felsic-mineral stabilities. Such a magma system is consistent with the observed eruptive characteristics, explains the typical ultramafic nodule and megacryst suites in the alkali olivine-basalt association and is free of difficulties with relative volumes of melts, with eruption timing and with rapid changes in erupted compositions.
Perspective From its Greek origins the term metasomatism should imply a 'change of body' or a change of substance. It must indicate a chemical change in a pre-existing rock or mineral, and traditionally has been used to make the distinction from metamorphism ('change of form') indicating reconstitution without chemical introduction. Unfortunately, early definitions of metasomatism are not rigorous, but common usage applies to cases where material has been transferred through a vapour or fluid without melting. More recently, in reference to the mantle, the usage has widened considerably, sometimes referring to simple melt infiltration, and even unspecified enrichment processes of a pre-existing mantle composition. Obviously the latter usages are not justified but there is a boundary problem. W h e n melt is introduced into a rock it may react with its surroundings and metasomatism may correctly describe the alteration process in the wall-rocks;
this borderline condition is perhaps part of the reason for some of the current ambiguity. There is at present a need for general agreement about terminology and it would certainly help to retain scientific precision if 'metasomatism' could be restricted to those cases where there is petrographic evidence of replacement of a pre-existing rock. Certainly the term should not be used to describe simple melt injection, producing a hybrid mixed rock, nor when there is no petrographic evidence that there has been replacement of an earlier mineralogy. In other words, the term should be applied only when there is unequivocal evidence of the previous substance--otherwise, to say that a rock has changed its chemistry is supposition. If there is a case for chemical introduction but the process is unclear or unknown, it would be better, and more straightforward, to use the term enrichment. Evidence for metasomatism in the mantle has
From: FITTON, J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 1-13.
2
D.K.
been steadily accumulating during the past 10 years (Harte et al. 1975; Lloyd & Bailey 1975) and the topic has provoked such interest that some surveys and reviews have already appeared (e.g. Boettcher & O'Neil 1980, Introduction and tabulation; Bailey 1982). The cited examples are relatively succinct and in some ways complementary, and may be taken as constituting a comprehensive starting point for the following discussion which aims to sketch an overall perspective to the subject. Occurrence and distribution Mantle samples showing evidence of metasomatism are provided in ultramafic nodules carried to the surface by high-speed volcanic eruptions. Some peridotite massifs (e.g. Lherz) contain veins of less refractory minerals, while others also contain evenly distributed minerals such as phlogopite and amphibole (e.g. Finero). These are sometimes cited as examples of 'enriched' and even metasomatized mantle but the evidence that such minerals have been introduced from an external source while the peridotite was part of the mantle has still to be established. Even in volcanically derived mantle fragments there may still be problems in distinguishing reaction products produced after incorporation in the magma; various criteria can be applied, however, and it is clear that in many cases minerals were introduced into the peridotite prior to eruption (Bailey 1982). Mantle nodules showing signs of metasomatism are brought to the surface exclusively by alkaline activity, especially in ultramafic and mafic eruptions. These range from kimberlites, which are characteristically fragmental, through to magmas in the alkali olivine-basalt association (basanites through to mugearites) which may occasionally carry ultramafic nodules. Hence, in terms of distribution, samples may be found in any parts of the stable plates, both oceanic and continental, where there has been alkaline igneous activity; such activity is usually associated with uplift and dislocation, in other words with some kind of lesion in the lithosphere. Compositions Metasomatism is recognized when there is replacement of a pre-existing peridotite mineralogy, new minerals being characterized by their content of mobile and volatile elements, as listed in Table 1. Introduction of the listed minerals into garnet or spinel peridotite (generally taken as representing average mantle composition) signifies mantle enrichment in lithophile ele-
Bailey TABLE 1. Lithophile-element-bearing m&erals in peridotite xenoliths which either may show metasomatic replacement of pre-existing peridotite mineralogy or are associated with metasomatism a (Major) b (Minor)
biotite, amphibole, clinopyroxene, carbonate phosphate, titanates, oxides, sulphides
Possible candidates for higher-level metasomatic minerals c (< 25 kb) feldspathoids, alkali feldspars ments. Most descriptions of mantle metasomatism deduce the introduction of some or all of the following elements: H, C, F, Na, A1, P, S, C1, K, Ca, Ti, Fe, Rb, Y, Zr, Nb, Ba, rare earths. If metasomatism were wholly a replacement process then there should be subtraction of equivalent material, balancing the introduction of the above elements. As yet this aspect has received little attention and this may be because the fabric of the rocks showing metasomatism is characteristically veined, with metasomatism proceeding along grain boundaries and fissures: hence the introduced material could be largely accommodated by an increase in rock volume, with redistribution rather than removal of the displaced elements. A parallel case may be seen in alkaline metasomatism in the crust (fenitization) which is characteristically marked by the introduction of new minerals along a close network of cracks. Timing In many cases the indications are that the metasomatic introduction of new minerals, although it must have preceded eruption, is part of the cycle of igneous activity, and there is isotopic harmony between the nodules and the magmas. In some instances there is evidence of an earlier metasomatic event (Erlank & Shimizu 1977; Menzies & Murthy 1980a) and there are cases of possible complex metasomatism and/or enrichment dating as far back as 3 Ga (Erlank et al. 1980; Menzies & Murthy 1980b). No doubt the picture will become still more complex as new data become available; this may be expected because alkaline activity has often been repeated through the same segment of lithosphere (Bailey 1977) and even if metasomatism were nothing more than a minor part of igneous activity there should be samples showing complex histories. Conditions of metasomatism Using a combination of experimentally determined mineral stabilities and solidi it is possible
Mantle metasomatism
3
to put limits on the pressures and temperatures of metasomatism, and hence on the conditions in the source mantle. Figs 1 and 2 summarize the conditions for stability of phlogopite, amphibole, clinopyroxene and felsic minerals in the mantle, and by reference to Fig. 4 the relative stability of carbonates can readily be envisaged. Obviously the diagrams must be generalizations because the individual mineral stabilities will depend on bulk composition (mineral and rock), coexistence with other metasomatic minerals, and the presence or absence of vapour, fluid or melt. In spite of these reservations the diagrams offer useful limits and indicate the following relationships. 1 Amphibole is stable near the peridotite solidus only in the upper part of the mantle, its lower boundary corresponding approximately to that of spinel and showing broad equivalence to the lower stability boundaries of feldspars and feldspathoids and the upper stability boundary of carbonate (see Figs 1, 2 and 4). 2 Phlogopite stability extends to higher temperatures and greater depths than amphibole, its
FIG. 2. P-T relationships of mineral stabilities and solidi, and geotherms. Solidi: as in Fig. 1 with the addition of WE, the solidus in the system KA1SiO4MgO-SiO2-H20-CO2 (Wendlandt & Eggler 1980). WE is used to provide a measure of phlogopite stability in peridotite mantle; its extrapolation to intersect KS at point P indicates a depth limit for phlogopite in the presence of melt (this is similar to the boundary given by Wyllie (1979)). Geotherms: as in Fig. 1. Mineral stabilities: C, coesite: Ks, kalsilite; AB, albite + nepheline; (AB), albite (jadeite + quartz); OR, K-feldspar; SC, solvus crest for alkali feldspars. stability limit at the solidus approximately coinciding with the inflexion region in some kimberlite geotherms (about 180km). Phlogopitebearing nodules or lavas from greater depths would therefore seem to be ruled out (Bailey 1986). 3 Clinopyroxene stability can extend up to the vapour-absent solidus and this mineral is arguably the most important but least appreciated metasomatic mineral.
FIG. 1. P-Trelationships of solidi, and geotherms. Solidi : PSD, peridotite vapour-absent solidus; KS, kimberlite solidus (Eggler & Wendlandt 1978); OE, peridotite solidus in the presence of H20 and CO2 (limited) (Olafsson & Eggler 1983). For convenience the stability region of spinel peridotite, with or without amphibole, is shown here rather than in Fig. 2. D is the diamond stability boundary (Kennedy & Kennedy 1976). Geotherms: S and O, shield and ocean (Clark & Ringwood 1964); 180, oceanic lithosphere 180 Ma old (Sclater et al. 1980); 30, oceanic lithosphere 30 Ma old (Oldenburg 1981). If geotherms should converge more rapidly in the depth range around 200 km as proposed by Tozer (1967) the geometry of the melt systems would change but the principles would remain the same.
From the above it may be deduced that carbonate and clinopyroxene may appear in any metasomatic regime, but the geological and experimental evidence suggests that high-pressure, low-temperature conditions (e.g. kimberlitic) favour carbonates while low-pressure hightemperature conditions (basanitic) favour clinopyroxene (see Fig. 5). Consideration of the magmas and their nodule suites also points to the fact that ultramafic alkaline magmas such as kimberlites, melilitites and nephelinites are characterized by phlogopite (in melts and xenoliths) while other magmas such as some nephelinites, but especially basanites, are characterized by xenoliths containing amphibole. Basanites in particular are likely to carry ultramafic nodules containing both spinel and amphibole, which
4
D.K. Bailey
may be an indicator of a generally shallower source depth for the nodules compared with ultramafic melts.
Outstanding issues While the reality of mantle metasomatism seems to have found wide acceptance, some differences may be perceived concerning such questions as the following: 1 Which comes first, the magmatism or the metasomatism? 2 Is metasomatism always a local phenomenon (in the vicinity of alkaline intrusions) or can it develop regionally? 3 What is the source of materials erupted at the surface, e.g. are they from the lithosphere or the asthenosphere ? The last question seems irresolvable until the nature of the asthenosphere can be unequivocally defined. At the moment, the ascription of a rock to an asthenosphere source (on the basis of selected aspects of its trace-element or isotope chemistry) is a convoluted way of saying that it has some attributes of mid-ocean ridge basalt; it would be wholly inappropriate to digress into this minefield. The first two questions are related, and controversy on these issues seems likely to be fruitless. Obviously, samples showing recognizable metasomatism can do so only on a local scale, and because volcanically transported fragments of metasomatized mantle must be small the evidence for metasomatism on a large scale must be sought in other ways. Appropriate lines of evidence, indicating large-scale metasomatism, have been described elsewhere (Bailey 1982) but may be summarized as follows: (a) an ultramafic nodule population dominated by metasomatized fragments throughout an igneous province; (b) a range of samples showing various degrees of metasomatism through to completely transformed nodules (alkali clinopyroxenites); and (c) the chemical equivalence of extensively metasomatized nodules and erupted melt compositions. Much of the cogency of the case for large-scale metasomatism is lost if discussion is allowed to focus just on laboratory data: it is then easy to lose sight of the wider geological perspective, where alkali- and volatile-rich magmatism has to be seen in the context of the geothermal and tectonic conditions of stable plates (Bailey 1983). This larger-scale (geological) evidence must be taken together with that of metasomatism on a small scale if we are to retain a balanced view. Both kinds of evidence exist (O'Reilly & Griffin 1984; Wilshire 1984) and they are not mutually exclusive. No amount of evidence or
argument in favour of one can, of itself, falsify the case for the other. Good petrographic evidence of local metasomatism is obviously immune to rebuttal, and no evidence has yet emerged to reject the case for regional development of the process; until such evidence emerges it is important that the argument should not become polarized (as in some older geological disputes) lest we end up with a controversy without roots in nature. The dilemma of the relationships between melting and metasomatism has been highlighted by Hawkesworth et al. (1984), who distinguish two enrichment processes, one typified by subsolidus metasomatism in kimberlite nodules and the other typified by injection of small-volume melts (typically associated with basanitic magmatism). These are essentially expressions of the effects of pressure and temperature, and it will be shown that they may be explained by the interplay of geothermal gradient, mantle solidus and mineral stabilities. It will be seen there that subsolidus metasomatism can also play a vital role in basanitic magmatism. Although it is valuable to identify the chemical characteristics of the two types of enrichment (Hawkesworth et al. 1984) it should be remembered that there is a spectrum of alkaline magmatism between kimberlites and basanites, and a spectrum of enrichment processes may be expected (see Fig. 5 and Bailey 1986).
Geological factors If all the observed metasomatism were a localized consequence of alkaline igneous activity, then the quantities of metasomatized mantle and the petrological significance of the process would be essentially trivial. It is when the geological picture is seen in total that such a conclusion becomes most suspect. Association with alkaline magmatism signifies the high activities of alkalis, mobile elements and volatiles, and the high-velocity eruptions that bring the samples to the surface emphasize the role of gases in this type of magmatism. When the tectonic framework, especially the connection between alkaline magmatism and lesions in stable lithosphere (and the repetition through time), is taken into account, it is difficult to escape the conclusion that we are observing a process in which volatiles must play a vital role. Alkali rich means volatile rich and, although there may be room for debate about whether the metasomatism is a precursor to magmatism in any particular instance, there can be little doubt that volatile migration may have far reaching effects without the necessary intervention of magma. One of the most graphic examples is in kimberlite peridotite
Mantle metasomatism nodules containing phlogopite: these were sampled from points on the geotherm well below the kimberlite solidus (Bailey 1982, Fig. 3) and hence well outside conditions where any melt could exist. Certainly there are strong grounds for supposing that volatile influx is a natural precursor to both metasomatism and alkaline magmatism.
Prospect Given an outline of present concepts of mantle metasomatism it becomes fruitful to look beyond and to consider other consequences of volatile movement through the mantle. One aspect, of immediate relevance to a volume on Alkaline Rocks, has been largely neglected in discussion of mantle metasomatism: this concerns the broader spectrum embracing not just mafic and ultramafic compositions but felsic alkaline magmas. At what has been described as the limiting case of cratonic magmatism (Bailey 1980a) there is kimberlite with no felsic associates, and closely related must be the lamproites and melilitites with very limited, if not uncertain, felsic connections. These and the ultramafic lamprophyres share other common features, such as style of eruption, that require special consideration in terms of petrogenesis; they have been discussed elsewhere (Bailey 1986). The more voluminous expressions of alkaline mafic magmatism are essentially basanitic (having some connections with nephelinites) grading through to transitional basalts, and these typically form major volcano complexes and have associated felsic magmas. Such alkali basalt magmas show evidence of mantle enrichment and characteristically may carry mantle nodules, which themselves show signs of metasomatism and/or enrichment. Their mantle source is hardly in doubt, and the possibility that this source may have undergone metasomatism raises the question whether the felsic magmas could be part of the same process. Some of the problems of alkaline felsic magmatism, such as regional magma development and comparative volumes, could be resolved in the context of regional metasomatic processes (Bailey 1972, 1974) and there have been no data on relevant mineral stabilities to exclude the possibility of sources in the upper mantle (Bailey 1976); more recent data lend credence to this possibility and allow a more penetrating look at the question. Before embarking on an examination of the more recent experimental information it is useful to look over the case in geological terms. The popular view for a long time has been that
5
felsic alkaline rocks are products of differentiation from mafic parents at relatively shallow (usually crustal) depths. Processes such as fractional crystallization are still currently in vogue in spite of many difficulties (see for instance Yoder 1979; Bailey 1981) and discussions of felsic magmas in particular seem impervious to contrary evidence. When the various pieces of evidence are assembled together, however, the case for low-pressure differentiation as a universal mechanism for producing felsic magmas can be seen to be without foundation. Essentially, evidence contrary to continuous differentiation at low pressure takes two forms: (1) felsic and intermediate volcanics carry ultramafic nodules; and (2) there are substantial composition and volume discrepancies in the igneous products of a given complex or province. As will be seen later, the first could be considered as the evidence from the high-energy ranges of the magmatic system and the second from the lower-energy ranges. Nodules
In addition to the dramatic examples of felsic lavas carrying ultramafic nodules (Wright 1966, 1969), mugearites and hawaiites are commonly reported as the hosts in ultramafic-nodule localities (see Boettcher & O'Neil (1980) for an informative listing). High-speed eruption is essential, and these are unequivocal examples of the existence of felsic and intermediate melts in the mantle: it is clear that low pressure differentiation is not essential for the formation of magmas in the range hawaiite-trachyte-phonolite. In fact, the composition of these melts may be a direct indication of physical and chemical conditions in the mantle source. Composition gaps
Daly (1925, 1927) first drew attention to the basalt-trachyte bimodality of oceanic lavas, a finding later supported by Barth (Barth et al. 1939) and statistically verified by Chayes (1963) who suggested that this raised doubts about the evolution of trachyte from basalt by continuous differentiation. Quite naturally, this suggestion provoked spirited opposition from advocates of fractional crystallization; the case was reviewed by Yoder (1973) who concluded that the gap was real and that this was not a realistic outcome of continuous fractional crystallization. More recent studies (e.g. Zielinski 1975) still invoke this mechanism, however, without apparent question, and so it is permissible to cite some additional geological facts.
6
D.K. Bailey
1 In deeply exposed continental sections revealing syenite plutons, comparable alkaline mafic and intermediate plutons are rare or absent. 2 Zoned plutons are rare, and where known are relatively small and generally considered to be composite (e.g. Monteregian Hills). 3 In keeping with 1 and 2, the common nodules in felsic volcanics are typically cognate, e.g. syenite in trachyte, peralkaline granite in comendite. To the above it may be added that the typical nodules in basanites are peridotite and gabbro. Hence it is hardly reasonable to argue that the basalt-trachyte bimodality in the ocean basins is an artefact of exposure level, eruption characteristics or sampling bias: the continental plutonic bodies and the nodule suites confirm the scarcity of intermediate magmas. Perhaps the most vivid evidence of magmatic bimodality lies on Graciosa in the Azores. Here there have been alternating eruptions of felsic and mafic pyroclastics, without detectable time breaks, requiring the simultaneous coexistence of contrasting magmas within the one volcanic system (Maund & Bailey 1982; Maund, 1985). This is simply a graphic example of the contemporaneous eruption of trachyte and basalt on oceanic islands generally, to which Daly called attention so long ago. Other composition gaps are present in other magma associations, such as basanite-phonolite and nephelinite-nephelinitic phonolite, but it may be relevant that even among the felsic rocks themselves there seem to be marked distribution maxima, as between phonolites, trachytes and rhyolites (and the plutonic equivalents). These gaps, too, militate against continuous differentiation, tending to favour partial melting or at least multiprocess origins (see Bailey (1976) for a full discussion).
Volumediscrepancies Another major problem for continuous differentiation at low pressure lies in the volume relations of the magmas in many provinces. The usual image of a large alkali basalt centre with small spines and flows of trachyte may be reasonable for some oceanic volcanoes, but it is a dangerous generalization, even for the oceans, e.g. Azores and Canaries. On the continents it is probably the exception rather than the rule: here again we have deep sections providing plutonic evidence. In major syenite provinces, such as Kola and Malawi, the basic and intermediate rocks required by continuous differentiation are not merely insufficient--they are lacking altogether.
Large syenite plutons (and large monotonic trachyte and phonolite volcanoes, as in the Kenya rift) are further evidence against low-pressure evolution of felsic magmas from a basaltic parent. Indeed, to suppose such differentiation has occurred when there is no sign of the earlier stages is to assume that syenite can form in only one way. In the volcanic regime the volume discrepancy can be seen in the Miocene-Recent activity of the East African rift zone where the calculated volumes of felsic and mafic volcanics are approximately equal (Williams 1972) and there have been regional floods of phonolite and trachyte. The latter are difficult to provide for in any process of magma generation but pose an acute problem for high-level differentiation (Bailey 1974, 1978). It should be said, too, that detailed studies of particular trachyte and pantellerite suites have not only failed to find any links with basalt (Bailey 1978) but have even revealed distinctive differences between felsic volcanoes erupting contemporaneously in the same province. In the area around Lake Naivasha in Kenya, there are Holocene phonolite, trachyte, pantellerite, comendite and basalt eruptions from overlapping centres that provide not only a series of major composition gaps but also distinctive major and trace-element patterns that are inexplicable in terms of continuous differentiation (Bailey & Macdonald, 1987). Clearly, there is commonality in that the magmas are part of a cycle of igneous activity, but the need is for a multisource/ multiprocess system of magma genesis. Yoder's solution (1973) to the dilemma of the Daly gap was the production of two contrasting magmas by fractional melting, along lines indicated by Presnall (1969). This envisages a source composition with two invariant points which by continuous melt extraction during an igneous cycle yields two contrasting magmas. Melt generation from two distinct invariant points would clearly alleviate some of the problems of alkaline magmatism and would explain composition gaps, but the generation of large volumes of felsic magma from a peridotite mantle source remains a difficulty, and a new problem is introduced of irreversibility and timing. Either the source volume would have to keep changing or the liquid from the lower-temperature invariant point could be erupted only at the start of a melting cycle. These difficulties could be solved through replenishment of felsic components by metasomatism or other forms of enrichment. At the very least therefore Yoder's solution requires an open system with melting pulses acting on periodically enriched sources (cycles of melting and enrichment). Even so the volume problem remains, and
Mantle metasomatism an additional factor is introduced by alternating mafic-felsic volcanism, exemplified by Graciosa, where there is effectively simultaneous eruption of felsic-mafic magmas poor in phenocrysts. Such activity implies the need for two (or more) distinctly different source compositions to be melted and tapped during the same igneous cycle. Geological evidence indicates that felsic sources richer than normal peridotite are feasible in the mantle and recent experimental evidence confirms that appropriate mineral stabilities extend into the upper part of the mantle. When these are taken together with eruption characteristics, depth indicators in the nodule suites and evidence of metasomatism, it becomes possible to propose a scheme of magma genesis that can reconcile all the seemingly conflicting observations and give an integrated pattern to the activity.
Metasomatism and the development of separate felsic sources Peridotites containing minor amounts of plagioclase, trachytes containing peridotite nodules, and sanidine-coesite eclogite nodules are some of the evidence indicating that felsic melts and sub-solidus felsic mineralogies are possible in the mantle, and it has been clear for some years that there is no a priori reason to suppose otherwise (Bailey 1976). In view of the problems with felsic magmatism, it is appropriate to update the evidence on mineral stabilities and then to enquire whether and how two sources could develop, and how magmas could be generated and erupted. In Figs 1 and 2 some of the relevant stability fields are plotted. Obviously the bulk mantle composition must influence some of these, but broadly speaking felsic minerals could exist to depths of about 100 kin, the exact limit depending also on the prevailing temperature. The special importance of the stability ranges shown in Figs 1 and 2 lies in the fact that the upper mantle within the felsic stability zone has the potential to yield melts of felsic character at a relatively low-temperature invariant point. No such invariant melting point is possible at greater depths, so that the first melts (from a phase assemblage containing no felsic minerals) must of necessity be different, and in all probability distinctly mafic. Thus the possibility of two distinct sources with completely different first melts is evident. Could their potential for producing contrasting magmas be established and enhanced by processes of enrichment and/or metasomatism ? Starting from the critical observation that high
7
volatile activity is a hall-mark of alkaline magnatism, it has been possible to develop a hypothesis relating magmatic variations to melting and metasomatism by volatile flux along different geothermal gradients (Bailey 1970, 1980a). Further consideration of the effects of the processes along low geothermal gradients has been given elsewhere (Bailey 1984, 1986). For the generation of felsic melts, it is necessary to look to activity along steeper geotherms more appropriate to oceanic conditions (or those away from continental craton nuclei). On steeper geotherms, such as those that would cross the solidus at high levels in the mantle (in the felsic stability region), volatiles could migrate along the geotherm only if the channelways were lined (Olafsson & Eggler 1983); melting is then more likely. When the possibility of flux melting along steeper geotherms was discussed previously (Bailey 1980a, 1983) attention was focussed on the eruption of melt through the overlying lithosphere with increasing departure of melt temperature from that of the wall-rocks, i.e. an intrusion, or melt diapir, rising as a detached thermal anomaly. In recent years, however, the concept of migration of very small percentage melts has found increasing favour (Walker et al. 1978; Waft 1980; Stolper et al. 1982). Essentially the melt is envisaged as percolating through the mantle, in which case it should maintain thermal equilibrium with its path, and in the simplest case could be envisaged as migrating along the geothermal gradient. Where the melt is generated by volatile flux, it would effectively extend the path of flux migration. Gradual percolation over a long distance would mean that the melt must also maintain phase equilibrium with the mineralogy through which it moves: most alkaline ultramafic melts are of alkali clinopyroxenite composition (Lloyd & Bailey 1975; Lloyd 1981) and could percolate only along channelways akin to the source mineralogy. Either there must be pervasive distribution of all the required phases or the movement must be restricted to channelways lined with these phases. As the initial melting is seen here as resulting from volatile influx, the pathways are more likely to be enriched by interaction with the melt, which may be contrasted with the scavenging process of melt percolation envisaged by Fitton & Dunlop (1985). In the higher section of the path the percolating melt will, in any case, start to reapproach the solidus and will then progressively precipitate solids such as clinopyroxene and, closer to the solidus, minerals such as amphibole, phlogopite, carbonate, feldspar and feldspathoids. Consequently the melt channels will be further lined or plated with minerals containing lithophile ele-
8
D.K. Bailey
ments. This may lead to occlusion of the percolation holes, requiring changes of course for subsequent percolating melt and thus extending and intensifying an enriched zone in the mantle. At the same time the temperature will be raised by the release of latent heat of fusion so that continuing percolation will act as an effective means of heat transfer and gradually steepen the geothermal gradient. Under conditions of 'equilibrium' percolation, melt will be used up where the geotherm passes back through the solidus, leaving only residual volatiles to continue along the sub-solidus path. These may be expected to produce an intense zone of metasomatism in the immediately overlying lithosphere segment indicated in Fig. 3. Melts or volatiles migrating along steep geotherms will not encounter the stability regions of either phlogopite or amphibole (see Fig. 2) so that most of the trapping of lithophile elements would be in clinopyroxene and/or carbonate until the solidus is approached. Thus, most of the mobile constituents such as Na, K, OH, F and C1 will be concentrated into any surviving lowtemperature melts and fluids, leading to enrichment in felsic minerals and amphibole at nearsolidus conditions. During the early phase of a new igneous cycle along steeper geotherms, therefore, there would be enrichment and meta-
800
IO00
1200
T ~
-,o
5o-
~.:/..-..: ~ u s
I 0 0 --
. ..........
I
"~" ~.~ : ~ ACCUMULATION "<':;: ~].. :.i ::.. ? AND ,,,~,-~1::. i .) ENRICHMENT "20
I ~~.
h
l;. : S.::" ~::!~i:, ~ i
I ' IY'~ ._./ LUX
I
,
I\
"50
t Depth km
FIG. 3. The consequences of flux migration and melt percolation along an initial geotherm G1 (similar to O in Fig. 1) so that continued flux would develop alkaline ultramafic and/or mafic melts. As melt (arrows) percolates back towards the solidus (OE, as in Fig. 1) it precipitates lithophile-element-bearing minerals and releases heat, 'instantaneously' steepening the channel geotherm (G2). Vapour released, on solidification of melt, heats and metasomatizes the sub-solidus sector. With time, G2 will be smoothed by exchange with the greater mass of the surrounding mantle, and ultimately will be restored to a steady-state geotherm when activity ceases. (See text for further discussion.)
somatism in the amphibole-felsic zone of the upper mantle. Furthermore, the conditions would locally favour pegmatitic formation, with scope for the development of large megacrysts of amphibole, alkali felspars and soda-rich plagioclase; near-solidus volatile accumulation could then lead to small, explosive high-velocity eruptions, bringing such megacrysts to the surface. Melt percolating along grain boundaries obviously cannot penetrate below the solidus boundary, but continued percolation towards the solidus may lead to melt accumulation in this sector of the path, through a combination of steepening geothermal gradient and channel occlusion. The solidus boundary may thus act as a melt dam allowing the development of near-solidus melt pockets, capable of diatremic eruption, of which a typical manifestation would be lamprophyric breccias (see Bailey (1986) for further details). Accumulation of felsic melts would also be possible at temperatures below the peridotite solidus shown in Fig. 3 whenever the mantle becomes extensively modified so that the peridotite solidus no longer applies and other lowertemperature solidi become appropriate. In the strict sense this should take effect as soon as felsic minerals are present, but in the initial stages any melt would be volumetrically insignificant; these interstitial melts may themselves percolate, however, and accumulate into eruptible volumes at temperatures below the peridotite solidus. The preceding discussion provides the framework for generation of felsic melts in the alkaline magma association; it remains now to examine the conditions under which the mafic melts are generated, and in particular how some of these melts achieve the conditions necessary for highvelocity eruption.
Mafic melt diatresis Under the peridotite near-solidus conditions described above, the distinctive products are felsic magmas, high-velocity diatremic eruptions (felsic to mafic) carrying megacrysts and nodules, and lamprophyric breccias. In contrast, it is implicit in the discussion that many associated mafic volcanics originate below the felsic stability zone and in general will be at higher temperatures (the latter is also an essential element of Yoder's solution). Alkaline ultramafic and mafic rocks are indeed sometimes erupted as liquid, examples being available from lamproite through melilitite and nephelinite and extending to basalts; such melt eruptions present a fundamentally different facet of alkaline igneous activity. In addition to nearsolidus melts, the magma system has to permit
Mantle metasomatism melt accumulation at a temperature well above the vapour-saturated peridotite solidus, such that the eruption trajectory in P - T space intersects the Earth's surface above the solidus, as seen in Fig. 4. Some ultramafic eruptions may take the form of lapillae sprays in which the lapillae are lava droplets (Lloyd 1985); the associated pyroclastic deposits may also contain ultramafic nodules testifying to high-velocity eruption from a great depth. This must represent the case of a fluidized system originating in the mantle with a high liquid content and presumably driven by rapidly exsolved gas. Elsewhere ultramafic nodules are also to be found in lava flows, which must indicate more coherent high-velocity eruptions of melt plus fragments, and this style of eruption seems to apply most commonly in the case of nodule-bearing basanitic eruptions. The last factor, along with other basanite-eruption characteristics, is consistent with generally lower gas contents compared with nephelinite-melilitite magmas. High-velocity eruption requires a mechanism or agent of acceleration, and the only plausible 60(
800 I
I000 I
Deplh
T ~ IA
1200 I
OE/
km
G-_
""
CS
~ ~
T~\
\
1400 I |
',
\
L
5O
. I00
~
\
Ao;TN;X2E kb ] .... - 20 -30
~,B~\ \~
- 40 ,,
150 \1/ ~ 0 ~ -- MELT DIATRESIS
~ ~
'~'~ ,,
-- 5 0
FIG. 4. P-Trelations of detached melt bodies rising from a region of accumulation on a geotherm (G) along an adiabatic path (A, an adiabat line with its origin at 1100~ from the Nyiragongo lava lake). Carbonate peridotite solidus is CSQR, with TQ the decarbonation boundary of Wyllie (1979); the equivalent boundary of Olafsson & Eggler (1983) is shown by the parallel (broken) line originating in OE (as in Fig. 1). Basanitic melt B accumulates and starts buoyant ascent from a point on the geotherm around 1200~ (approximating the asthenosphere boundary). High-velocity eruption ('melt diatresis') is triggered by massive gas exsolution as the magma passes through the melt structure transition where CO32- becomes unstable (taken as approximating the decarbonation boundary TQ). The pressure limit for amphibole and spinel peridotite stability (at near-solidus temperatures) is indicated at the right margin.
9
agent seems to be gas, either injected into an existing melt chamber or produced by rapid exsolution from the melt itself; at the moment the latter mechanism seems the logical candidate. If vapour-saturated melt is rising through the lithosphere it is to be expected that depressurization will cause progressive gas release; this would tend to quicken the rate of ascent but may be partly offset by increasing viscosity due to melt structure changes resulting from falling temperature and loss of volatiles. The effectiveness of viscosity changes must depend on initial melt composition and temperature. Speeding u p by gas expansion will be a function of release rate and gas density. High-speed liquid eruptions carrying mantle nodules are not the norm, however, and it would be useful if some additional 'firing' or 'lift-off mechanism could be identified in addition to simple decompression vesiculation. One potential trigger for the rapid and massive emission of gas would be when a rising body of melt passes through a phase boundary or melt structure transition. In the case ofbasanitic melts, the most common nodule mineralogies are spinel lherzolite, frequently including amphibole, to which experimental evidence can be applied to give a clear indication of the typical sampling depths of nodule-bearing basanitic eruptions (Fig. 4). This characteristic association with spinel lherzolite strongly suggests that nodulebearing basanitic eruptions get a sudden acceleration (or lift-off) within or close to the spinel and amphibole stability fields, i.e. massive gas exsolution occurs at depths around 80 km. As shown in Fig. 4, there is a ready explanation for this particular case, because around this level carbonate becomes unstable, and the passage of basanitic melt containing dissolved CO2 (in the form of CO32 -) through the carbonate stability boundary could provide the necessary trigger for highvelocity eruption. For liquid diatresis, it is necessary to have melt accumulation to the point where it starts to rise independently through cooler lithosphere, i.e. detached from the ambient mantle geotherm. It is important to realize, however, that although the melt and wall-rocks may be thermally separated they may still be chemically similar. This is because melt piercement of the lithosphere is the probable culmination of a period of volatile flux and melt percolation. Thus, a rising mass of melt is likely to encounter earlier-enriched wall-rocks in its upward passage. Olafsson & Eggler (1983) and Brey et al. (1983) have pointed out that amphibole, phlogopite and carbonate disappear within a small temperature interval above the solidus, so an intrusive melt may be expected to consume more of these phases in any interaction
I0
D.K. Bailey
it may have with the wall-rocks. An ascending melt body would thus tend to develop and preserve similar characteristics to its enriched path zone, in keeping with the observation that the magmas in a given province are chemically similar to the metasomatized nodules that they bring to the surface (Lloyd & Bailey 1975). The conclusion drawn then, that the magmas were formed by direct melting of the nodule compositions, may be too simple. The eruption characteristics require that at the level from which the nodules were picked up the carrier melt was at a temperature well above the vapour-saturated peridotite solidus. Hence the initial melt must have a deeper source than the bulk of the ultramafic nodule population. Depending on the source depth, its initial chemistry might even have been alkaline ultramafic, but it attained its erupted chemistry by interactions with metasomatized wall-rocks up to the point of lift-off. Most of the nodules were derived from the wallrocks around and in front of the point of rapid gas exsolution, which provided the impetus for their high-velocity transport to the surface. In areas where the ambient geothermal gradient is steep, an alkaline magma system can now be envisaged as comprising two distinct source regions, distinguished on the basis of felsic mineral stability. One region, below the maximum depths of felsic mineral stability (approximately coincident with spinel and amphibole stability) is the potential source for mafic hypersolidus melts whenever there is volatile flux through the lithosphere. The overlying zone of felsic stability is the locus of intense melt enrichment and metasomatism at near- and subsolidus conditions during volatile fluxing: this forms the source region for the felsic and nearsolidus portions of the activity. Mafic and felsic magmas may thus both be primary products of the same igneous cycle: the first forms by volatile influx leading to melt diapirism (mafic) well above the peridotite solidus (Fig. 4), the second by melt percolation leading to melt diapirism (felsic) near or below the peridotite solidus (Fig. 3).
Conclusion By combining the geological and experimental evidence it is possible to produce a unified scheme for the generation and eruption of mafic and felsic alkaline magmas, constituting the range sometimes described as the alkali olivine basalttrachyte association. A combination of volatile and melt percolation along geothermal gradients appropriate to thicker oceanic and younger
continental lithosphere leads to enrichment and metasomatism in the P-Tregion of the amphibole peridotite solidus, providing the major source region for felsic magmas: the main source of mafic magmas is at depths below the stability range of felsic minerals, eruption resulting from melt accumulation and buoyant uprise. These conditions are summarized in sections 0 and 30 in Figs 5 and 6. Melt generation from two distinct sources can account for the following characteristics of this magmatic association: 1 Composition gaps. Intermediate compositions may be formed but are not essential to petrogenesis. 2 Volume discrepancies. The relative volume of mafic-to-felsic magma results from the interplay of melt percolation and mafic melt diapirism. 3 Timing. Characteristic magmatic diversity results simply from the fluctuating interplay of volatile flux and melt generation (irreversibility is not an issue). 4 Synchroneity. Both felsic and mafic magmas can be available for eruption at all times. 5 Megacrysts. These are typically erupted in explosive diatremic activity and show significant differences from megacrysts associated with ultramafic magmatism. They are commonly amphibole and feldspar in the alkali basalt association and are explicable in terms of near-solidus formation. 6 Basanites may be erupted in a highly liquid state, carrying nodules of amphibole and spinel peridotite which are indicative of higher-temperature generation from a deeper source than felsic melts. High-velocity eruption is 'fired' by gas exsolution when CO32- becomes unstable in the melt. 7 Characteristic abundance of lithophile elements. Volatile fluxing, metasomatism and enrichment provide the required concentration mechanism. All the above features are also consistent with the observation that alkaline magmatism is controlled by lithosphere structures acting as focussing zones for volatile escape to the Earth's surface (Bailey 1977, 1980b, 1983). In earlier discussions attention has been concentrated on continental examples where, in addition to the advantages of many different erosion levels, there is frequently evidence of repetition of the activity through the same lesion in the lithosphere. Since a crucial observation for the present paper has been the alternating mafic-felsic volcanism in the Azores, it is appropriate to conclude by pointing out that the Azores constitute an excellent oceanic case for lithosphere control.
Mantle metasomatism GEOTHERM : NEAR S (SHIELD)
S
BETWEEN
AND O
O
II (OCEAN)
Depth km + Amph(?)
! ( TOO COOL 9 ) II
50
740 ~
I
Phi + Cpx
I00
Phi +
150
--
Cpx + Carb + v
p ~180 o
200
+
880"11\1\11\\/1\\1/\//\\//\\//k\/
- -
-MANTLE
Corb 1030,~ + V
I 10750 (DIAMOND)
-T- V i 175*----
J" I I P OL-MELIL ? I K I M B T I - LAMPROITE I I K I M B T CARBONATITE I I MELTS
INCREASINGDECREAStNG "CPX"CARBONATE"
MAGMA
EVOLUTION
I DEVELOPS 402Ocpx + KS I +Ab(+Phl) I . . . . . . . + Melt UAbAr~,.r.. - ~ I NEPHELINITE 1200o* "
I I
,,
I
Me
Amph+Ab/Of I FELSICS (+Cpx) + v I AS CYCLE I GROWS
,o8o'c~
{-i-Phl)_+Melt }TRANSITIONAL-12OO=-T ', V ~ BASALT
IO 20 5O P kb 4O
Phi + Cpx + Carb +- Mell
I
I
Amph+Ab/Or IFELSICMELTS (+Cpx) +V II AS CYCLE
50
I NEPHELINITE 9 I I LEUCITITE II MELILITITE CARBONATITE - - --t (DIAMOND)
5O 6O
MANTLE }
IN M O D I F I E D
WITH
[
MELTS
MANTLE
MANTLE
1
MELTS
MANTLE
I MELTS
AND MELTS
CONTINUOUS
FLUX
THROUGH
LESION
FIG. 5. Sample lithosphere sections along four selected geothermal gradients, derived directly from Figs 1 and 2. In each case it is assumed that the initial mantle composition is peridotite. On the left of each column are indicated the mantle changes induced by volatile flux along the geotherm, and on the right the kinds of melt developed in this regime. 'Melt' is not mentioned on the left in column 1 to emphasize the limited melting expected when the geotherm S is in grazing incidence with the kimberlite solidus (KS in Fig. 1). The base of columns 3 and 4 has been set at 1200~ on the geotherms, as an approximation to the asthenosphere boundary. As the geotherm becomes steeper (columns 1-4) the melts become more siliceous, with increasing separation from the solidus, in the manner prescribed by Olafsson & Eggler (1983). Volatile fluxing along geotherm O could not produce strong phlogopite enrichment but instead gives rise to assemblages dominated by clinopyroxenes, felsics and (at higher levels) amphibole, i.e. directly related to basanitic melts.
Felsic volcanoes, especially trachytes, form a major component of the activity on some of the islands, notably those lying along the Terceira rift: Sao Miguel, Terceira and Graciosa. On these three islands not only are the composition gaps pronounced but the large volumes of felsic
_
JL
Z
Z THOSP
ERE
,!:i;i
MOBILE ELEMENTS FLOW TO LESION
Fio. 6. Sketch to illustrate the operation of volatile focussing through lesions in the lithosphere of different thicknesses and geothermal gradients. Columns correspond to those depicted in Fig. 5. Discussion of the relevant applications of channelling volatiles through a narrow zone in the lithosphere ('pie-funnel effect') can be found in Bailey (1980b; 1983).
magmas constitute major embarrassments for continuous differentiation. Explanations for individual volcanoes have been attempted, but the case for continuous differentiation is lost when the activity is seen in regional context. It is hardly conceivable that the volcanoes strung along the Terceira rift should all be spectacularly successful products of high-level differentiation while the volcanoes on the other islands outside the rift have effectively failed. It is more reasonable to assume that the Terceira rift volcanoes are the result of structural control that has permitted the formation, below the rift, of mantle source regions capable of providing felsic melts. Such a conclusion is consistent with the scenario whereby the rift zone acts as a focus for the release of volatiles through the Azores lithosphere. ACKNOWLEDGMENTS: My appreciation of magmatic composition gaps has been sharpened by discussions with Felix Chayes and Ray Macdonald, and by Julian Maund's systematic observations on the 'zebra-striped' tephra deposits of Graciosa. Constructive reviews of the manuscript were provided by Godfrey Fitton and Martin Menzies.
I2
D. K. Bailey
References BAILEY, D. K. 1970. Volatile flux, heat focussing and the generation of magma. Geol. J. Spec. Issue No. 2, 177-86. - 1972. Uplift, rifting and magmatism in continental plates. J. Earth Sciences (Leeds), 8, 225-39. - - 1 9 7 4 . Continental rifting and alkaline magmatism. In: SORENSON, H. (ed.) The Alkaline Rocks, pp. 148-59. Wiley, New York. - - 1 9 7 6 . Applications of experiments to alkaline rocks. In: BAILEY, D. K. & MACDONALD, R. (eds) The Evolution of the Crystalline Rocks, pp. 416-69. Academic Press, London. --1977. Lithosphere control of continental rift magmatism. J. geol. Soc. London, 133, 103-6. - 1978. Continental rifting and mantle degassing. In: NEUMANN, E.-R. & RAMBERG, I. B. (eds) Petrology and Geochemistry of Continental Rifts, pp. 1-13. Reidel, Dordrecht. - 1980a. Volatile flux, geotherms, and the generation of the kimberlite-carbonatite-alkaline magma spectrum. Mineralog. Mag. 43, 695-9. - 1980b. Volcanism, Earth degassing and replenished lithosphere mantle. Phil. Trans. R. Soc. A 297, 309-22. - - 1 9 8 1 . Bowen and the Post-Impressionists. Geol. Mag. 118, 311-8. - 1982. Mantle metasomatism--continuing chemical change within the Earth. Nature, Lond. 296, 525-30. 1983. The chemical and thermal evolution of rifts. Tectonophysics, 94, 585-97. - 1984. Kimberlite: 'the mantle sample' formed by ultrametasomatism. In: KORNPROBST, J. (ed.) Kimberlites. 1 : Kimberlites and Related Rocks, pp. 323-33. Developments in Petrology 11 A, Elsevier, Amsterdam. - 1986. Fluids, melts, flowage and styles of eruption in alkaline ultramafic magmatism. Trans. geol. Soc. S.Afr. 88 (2), 449-57. -& MACDONALD, R. 1987. Dry peralkaline felsic liquids and carbon dioxide flux through the Kenya rift zone. In: MYSEN, B. O. (ed.) Magmatic Processes: Physiochemical Principles. Geochem. Soc., Spec. Pub. No. 1, in press. BARTH, T. F. W., CORRENS, C. W. & ESKOLA, P. 1939. Die Entstehung der Gesteine, 422 pp. Springer, Berlin. BOETTCHER,A. L. & O'NEILL, J. R. 1980. Stable isotope, chemical, and petrographic studies of high-pressure amphiboles and micas: evidence for metasomatism in the mantle source regions of alkali basalts and kimberlites. Am. J. Sci. 280A, 594621. BREY, G., BRICE, W. R., ELLIS, D. J., GREEN, D. H., HARRIS, K. L. & RYABCHIKOV, I. D. 1983. Pyroxene-carbonate reactions in the upper mantle. Earth planet. Sci. Lett. 62, 63-74. CHAYES, F.1963. Relative abundance of intermediate members of the oceanic basalt-trachyte association. J. geophys. Res. 68 (5), 1519-34. CLARK, S. P. & RINGWOOD, A. E. 1964. Density
distribution and constitution of the mantle. Rev. geophys. 2, 35. DALY, R. A. 1925. The geology of Ascension Island. Proc. Am. Acad. Arts Sci 60, 1-80. -1927. The geology of St. Helena Island. Proc. Am. Acad. Arts Sci. 62, 31-92. EGGLER, D. H. & WENDLANDT, R. F. 1978. Phase relations of a kimberlite composition. Cam. Inst. Wash. Yr Bk, 77, 751-6. ERLANK, A. J. & SHIMIZU, N. 1977. Strontium and strontium isotope distributions in some kimberlite nodules and minerals. Extended Abstracts, 2nd Int. Kimberlite Conference. - - A L L S O P P , H. L., DUNCAN, A. R. & BRISTOW, J. W. 1980. Mantle heterogeneity beneath southern Africa: evidence from the volcanic record. Phil. Trans. R. Soc. A297, 295-307. FITTON, J. G. & DUNLOP, H M. 1985. The Cameroon line, West Africa, and its bearing on the origin of oceanic and continental alkali basalt. Earth planet. Sci. Lett. 72, 23-38. HARTE, B., COX, K. G. & GURNEY, J. J. 1975. Petrography and geological history of upper mantle xenoliths from Matsoku kimberlite pipe. Phys. Chem. Earth, 9, 389-416. HAWKESWORTH, C. J., ROGERS, N. W., VAN GALASTEREN, P. W. C. & MENZIES, M. A. 1984. Mantle enrichment processes. Nature, Lond. 311 (5984) 331-5. KENNEDY, C. S. & KENNEDY, G. C. 1976. The equilibrium boundary between graphite and diamond. J. geophys. Res. 81, 2467-70. LLOYD, F. E. 1981. Upper-mantle metasomatism beneath a continental rift: clinopyroxenes in alkali mafic lavas and nodules from S. W. Uganda. Mineralog. Mag. 44, 315-23. - - 1 9 8 5 . Melting experiments on glassy olivine melilitites from strongly potassic mafic volcanics ofS. W. Uganda. Contrib. Mineral. Petrol. 90, 23643. - & BAILEY, D. K. 1975. Light element metasomatism of the continental mantle: the evidence and the consequences. Phys. Chem. Earth, 9, 389-416. MAUND, J. G. 1985. The volcanic geology, petrology and geochemistry of Caldeira volcano, Graciosa, Azores, and its bearing on contemporaneous felsicmafic oceanic island volcanism. Ph.D. Thesis, University of Reading (unpublished). - - , J. G. & BAILEY, D. K. 1982. Caldeira volcano, Graciosa, Azores: Progress Report. J. geol. Soc. Lond, 139, 658. MENZIES, M. & MURTHY, V. R. 1980a. Mantle metasomatism as a precursor to the genesis of alkaline magmas--isotopic evidence. Am. J. Sci. 280A, 622-38. - & 1980b. Nd and Sr isotope geochemistry of hydrous mantle nodules and their host alkali basalts: implications for local heterogeneities in metasomatically veined mantle. Earth planet. Sci. Lett. 46, 323-34. OLAFSSON, M. & EGGLER, D. H. 1983. Phase relations of amphibole, amphibole-carbonate, and phlogo-
Mantle metasomatism pite-carbonate peridotite: petrologic constraints on the asthenosphere. Earth planet. Sci. Lett. 64, 305-15. OLDENBURG, D. W. 1981. Conductivity structure of oceanic upper mantle beneath the Pacific plate. Geophys. J. R. astron. Soc. 65, 359-94. O'REILLY, S. Y. & GRIFFIN, W. L. 1984. Sr isotopic heterogeneity in primitive basaltic rocks, southeastern Australia: correlations with mantle metasomatism. Contrib. Mineral. Petrol. 87, 220-30. PRESNALL, D. C. 1969. The geometrical analysis of partial fusion. Am. J. Sci. 267, 1178-94. SCLATER, J. G., JAUPART, C. & GALSON, D. 1980. The heat flow through oceanic and continental crust and the heat loss of the Earth. Rev. Geophys. Space Phys. 18, 269-311. STOLPER, E., WALKER,D., HAGER, B. H. & JAYS, J. F. 1982. Melt segregation from partially molten source regions: the importance of melt density and source region size. J. geophys. Res. 86, 6261-71. TOZER, D. C. 1967. Towards a theory of mantle convection. In: GASKELL, T. F. (ed.) The Earth's Mantle, pp. 325-53. Academic Press, New York. WAFt, H. S. 1980. Effects of the gravitational field on liquid distribution in partial melts within the upper mantle. J. geophys. Res. 85, 1815-25. WALKER, D., STOLPER, E. & HAYS, J. F. 1978. A numerical treatment of melt/solid segregation: size of the eucrite parent body and stability of the
13
terrestrial low-velocity zone. J. geophys. Res. 83, 6005-13. WENDLANDT,R. F. & EGGLER, O. H. 1980. The origins of potassic magmas: 2. Stability of phlogopite in natural spinel lherzolite and in the system KA1SiO4-MgO-SiOE-H20-CO 2 at high pressures and high temperatures. Am. J. Sci. 280, 421-58. WILLIAMS,L. A. J. 1972. The Kenya Rift Volcanics: A note on volumes and chemical composition. Tectonophysics, 15, 83-96. WILSHIRE, H. G. 1984. Mantle metasomatism: The REE story. Geology, 12, 395-8. WRIGrtT, J. B. 1966. Olivine nodules in phonolite of the East Otago Alkaline Province, New Zealand. Nature, Lond. 210, 519-20. --1969. Olivine nodules and related inclusions in trachyte from the Jos Plateau, Nigeria. Mineral. Mag. 37, 370-4. WYLLIE, P. J. 1979. Petrogenesis and the physics of the Earth. In: YODER, n . S., Jr (ed.) The Evolution of the Igneous Rocks, pp. 481-520. Princeton University Press, Princeton, NJ. YODER, H. S. Jr. 1973. Contemporaneous basaltic and rhyolitic magmas. Am. Mineral. 58, 153-71. --(Ed) 1979. The Evolution of the Igneous Rocks. Fiftieth Anniversary Perspectives, 588 pp. Princeton University Press, Princeton, N.J. ZIELINSKI, R. A. 1975. Trace element evaluation of a suite of rocks from Reunion Island, Indian Ocean. Geochim cosmochim. Acta, 39, 713-34.
D. K. BAILEY, Department of Geology, University of Reading, Whiteknights, Reading, RG6 2AB, U.K.
Alkaline rocks and their inclusions: a window on the Earth's interior Martin Menzies S U M M A R Y : The majority of inclusion-bearing alkaline rocks have a source region identical with that of ocean island basalts regardless of whether they were erupted through oceanic or continental crust. Models have been proposed where oceanic alkaline rocks are derived from the asthenosphere and during passage to the surface there have occurred variable amounts of interaction with lithosphere of the mid-ocean ridge basalt type. A similar scenario would account for most inclusion-bearing continental alkaline rocks. Genesis of micaceous kimberlites and lamproites, however, may require that asthenospheric plumes have impinged on 'nuggets' of aged heterogeneous enriched sub-continental lithosphere. This would lead to mixing of isotopicallydistinct asthenospheric and lithospheric melts.
Introduction Inclusion-bearing alkaline rocks are erupted at the Earth's surface in the ocean basins and on the continental crust. They include basanites, nephelinites, hawaiites, alkali olivine basalts, phonolites, Group I and Group II kimberlites and lamproites. All these magmas or melts are known to transport spinel- or garnet-bearing peridotites and/or high-pressure megacrysts to the surface (e.g. diamond, sapphire, garnet etc.). The fact that they have disrupted and entrained mafic and ultramafic fragments stable at mantle pressures and temperatures argues very strongly for an ultimate mantle source. Important information about the geochemistry of the upper mantle is retained in the host magmas. Moreover, the inclusions provide added insight into the nature of the mantle occurring at shallower depths than the source of the host magma. Early research into the possible origins of alkaline rocks stressed the 'mantle-type' 87Sr/ S6Sr < 0.706) and 'crustal-type' (s 7Sr/S6Sr > 0.710) isotope ratios as a means of identification of possible source regions (Powell & Bell 1974; Rock 1976). Any similarity between isotope ratios measured in alkaline rocks and mid-ocean ridge basalts (MORBs) or ocean island basalts (OIBs) was taken as unequivocal proof that the particular rock in question was mantle derived (e.g. East African rift). Similarly, any alkaline rock with an isotopic composition similar to crustal values was assumed to be a crustal derivative or to have experienced crustal contamination (e.g. Kimberley, Australia). For example, Rock (1976) commented that in certain alkaline rocks the Sr isotopic ratios were 'so high (>0.710) that a significant contribution from crustal material was inescapable'. It is implicit in these models that interaction with a 'crustal' component occurred after magma
formation. More recently, alkaline rocks with radiogenic Sr (and non-radiogenic Nd) isotopic ratios have been interpreted as extracts from mantle source regions where addition of a crustal component to the mantle occurred prior to magma production (Chase 1981; Hofmann & White 1982; Zindler & Jagoutz 1987) by recycling processes involving subduction and/or delamination of aged sub-continental mantle (McKenzie & O'Nions 1983). Alkaline rocks may include components from the following:
1 Lower mantle." below the boundary layer (about 670 kin). 2 Upper mantle." oceanic lithosphere; continental lithosphere; convecting asthenosphere. 3 Recycled components." oceanic and continental crust and mantle. In this review isotopic data pertinent to inclusion-bearing alkaline rocks and their cargo of mantle inclusions erupted in ocean basins, continental rift valleys and stable cratonic regions will be used to construct an Earth model. Earth models will be considered where the structural elements that form the upper mantle below both oceans and continents are lithosphere (crust and mantle) and asthenosphere. These in turn are underlain by the lower mantle below the 670 km boundary layer.
Alkaline rocks and their mantle inclusions Random fragments of mantle are transported to the surface in alkaline rocks. The inclusions entrained in alkaline rocks range from garnet
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 15-27.
15
M. Menzies
[6
and spinel peridotites to mica clinopyroxenites, glimmerites, apatite-amphibole pyroxenites and mica-rutile-ilmenite-diopside (MARID) rocks. It is useful to think of the variety of inclusions as disrupted mantle pegmatites viz. pieces of mantle wall-rock (i.e. peridotites) and fragmented conduits of silicate melts or aqueous fluids (i.e. pyroxenites, glimmerites etc.). Analogues can be found in inclusion suites for the various melts or fluids that have migrated through the mantle. A m p h i b o l e - a p a t i t e - p y r o x e n ites found as veins in spinel lherzolite inclusions are believed to be a product of silicate-melt migration (Hawkesworth et al. 1984; Menzies et al. 1985c). A by-product of this magmatism is the Fe-Ti metasomatism observed in wall-rock peridotites. In other words, upwelling basanitic or nephelinitic melts may have crystallized in the mantle as these vein networks. It can be shown that their isotopic and trace-element characteristics are consistent with such an interpretation (Irving and Frey 1984; Roden et al. 1984). M A R I D rocks (Kramers et al. 1983; Smith 1983) found in kimberlite inclusion suites are believed to be portions of the conduit lining left after passage of a hydrous fluid enriched with incompatible elements. Glimmerites and micaceous pyroxenites (Erlank et al. 1987; Menzies et al. 1987) may also represent disrupted vein systems formed by crystallization of LIL- (LIL stands for large
05132 ~ N I V A K ~ " 1 0
o.51~L__~ 0.702
~
Ocean basins Before any discussion of the isotopic variability of alkaline basaltic and kimberlitic rocks erupted in continental regions, it would seem logical to review the data on alkaline rocks from ocean basins (Fig. 1). Although not strictly oceanic, the Nunivak data are included in Fig. 1. Alkaline rocks associated with disruption and entrainment of fragments of oceanic mantle have a very narrow range of Sr and Nd isotopic ratios (i.e. STSr/86Sr < 0.706 and 143Nd/144Nd > 0.51270 (Fig. 1, Table 1)). These magmas are clearly derived from mantle depleted in the light REE for a considerable period of time. A concomitant depletion in Rb is also apparent, but in the case of Tahiti and Malaita the magmas have slightly higher SVSr/86Sr ratios. This could result from (a) mixing of MORB mantle with a recycled radiogenic component or (b) influx of a high Rb/Sr fluid.
0.513211
j-5
|
I
TAHITII|
|A / ~10
|
I
07.02
i
I
0.704
I
0.706
Inclusions
Host Magmas
0.513211_
'
0.51~4t-,
0.706
0.704
(~
ion lithophile elements) and LREE-enriched hydrous fluids in the mantle (LREE stands for light rare-earth elements). Mantle inclusions provide a very important record, not only of ancient depletion events (more than 103 Ma ago) related to crust formation, but more importantly of the passage of silicate melts and hydrous fluids from the mantle to the surface.
3 0.5132~,inlcl "-
~
i
HAWAllli
+10
-5 "O 0.5128 z
Z-O 9-
0.5128
_
9
0.5124 I 0.702 87Sr/86S
I
I 0.704
I
I 0.706
o
0.5124
-I 0.702
1
I 0704
I
I 0.706
-5
-o Z
uo
r
FIG. 1. Nd versus Sr isotopic variation (present day) in mantle ultramafic and mafic inclusions and their host basaltic rocks from oceanic basins. Note the following: (i) the host magmas are derived from mantle depleted in Rb and light REE; (ii) the inclusions exhibit a greater range of Nd and Sr isotopes than the host magmas: (iii) the high 8VSr/a6Srratios in the Malaita and Tahiti inclusions. The vertical and horizontal reference lines in this and other diagrams define the generally accepted values for undifferentiated or chondritic mantle or bulk earth (87Sr/S6Sr= 0.7045 and 143Nd/144Nd=0.51264). Data are taken from several sources. Nunivak : Menzies & Murthy (1980a); Roden et al. (1984a). Malaita: Bielski-Zyskind et al. (1984). Tahiti: Vidal et al. (1984); Menzies & Hawkesworth (1987). Hawaii: Chert & Frey (1983).
Alkaline rocks and their inclusions
I7
TABLE 1.87Sr/S6Sr, 143Nd/144 Nd, R b / S r a n d S m / N d ratios f o r inclusion-bearing alkaline, kimberlitic a n d lamproitic m a g m a s
Locality
87Rb/86Sr
esra
/;Ndb
147Sm/
Ref
144Nd Afar (Assab) - 17 to - 10.0 -Africa (Tanzania) - 14.2 -Alaska (Nunivak) - 2 8 . 4 to - 17 0.049-0.121 Antarctica Foster Crater - 13.8 to - 11.4 - Ross Island - 19.2 0.089-0.165 Arizona Geronimo - 23.4 to - 17 0.118-0.240 San Carlos - 22.7 0.091 Australia Kiama - 7.1 0.051-0.164 Kimberley + 85 to + 218.6 0.593-9.855 Victoria - 28.4 to + 4.3 -California (Sierran Province) c + 4.3 to + 34.1 0.056-1.28 Fiji - 13.4 -France (Massif Central) - 14.2 to 0 0.153-1.225 French Polynesia (Tahiti) d + 7.1 to + 34.1 -Hawaii (Honolulu Series) - 18.5 to - 17.03 0.015-0.103 Scotland (Midland Valley and N W High- - 4.3 to + 25.6 0.077-0.203 lands) Solomon Islands (Malaita) + 1.4 to + 2.8 0.059-0.183 South Africa (various localities) Group I kimberlites - 11.4 to + 14.2 0.038-1.52 Group II kimberlites + 41.2 to + 90.8 0.140-0.786 South Yemen (Ataq) - 17.0 to - 2.8 0.028-0.692 Mantle xenoliths Keel e - 3 5 . 5 t o +1867.0-Asthenosphere -35.5to+21.3 --
+4.7 to +7.2 + 1.4 + 4 . 9 t o +7.8
-0.096 0.119-0.133
+3.3
--
to
+4.7
1 2 3
--
--
4
+ 4 . 9 t o +8.0 +7.6
0.101-0.126 0.130
5 6
- 0 . 6 t o +0.4 - 16.2 to - 8 . 0 - 4 . 7 to +5.0 - 9 . 6 to +2.9
0.116-0.120 0.065-0.082 0.112-0.119 0.103-0.140
+3.7
--
+ 7 . 0 t o +7.6 - 6 . 6 t o +3.1
0.120-0.168 0.068-0.108
7 8 9 10 11 12 13 14 15
+ l . 9 t o +2.9
0.102-0.121
16
- 4 . 1 to +1.0 - 12.9 to - 8.8 - 2 . 3 to +8.0
0.085-0.098 0.062-0.087 0.100-0.130
17 18
- 2 4 . 2 t o +14.8 -4.7to+14.8
---
19
+2.1 to +4.1
0.087-0.119
-
--
1.9
to
+2.5
"es~ represents the present-day 87Sr/86Sr ratio assuming a bulk earth s 7Sr/86Sr of 0.7045. beyo represents the present-day 143Nd/144Nd ratio assuming a bulk earth 143Nd/144Nd of 0.51264. cRange in Rb/Sr and Sm/Nd given for the Sierran Province lavas are for basaltic rocks that are not necessarily host to the xenoliths. The isotope data and trace-element data are for different samples. aTahitian samples are not the host magmas to the xenoliths. eThe range given for the 'keel' inclusions excludes inclusions found in diamonds (Richardson et al. 1984). 1 2 3 4 5 6 7 8 9 10
Betton & Civetta 1984; Cohen et al. 1984; Menzies & Murthy 1980a; Roden et al. 1984a; Stuckless & Ericksen 1976; Menzies et al 1985b; Menzies et al. 1985a; Evans & Nash unpublished data; Zindler & Jagoutz 1987; Wass & Rogers 1980; Menzies & Wass 1983; McCulloch et al. 1983 ; Frey & Green 1974; McDonough & McCulloch 1985; Van Kooten etal 1985;
T h e spinel a n d g a r n e t p e r i d o t i t e s a n d p y r o x e n i t e s e n t r a i n e d b y t h e s e m a g m a s (Fig. 1) h a v e a s i m i l a r r a n g e o f Sr a n d N d isotopes e x c e p t t h a t t h e i n c l u s i o n s do r e c o r d s o m e h i g h e r 87Sr/S6Sr ratios (e.g. M a l a i t a a n d Tahiti). It is v e r y difficult to state u n e q u i v o c a l l y t h a t t h e c h e m i c a l c h a r a c t e r istics o f t h e inclusions are n o t a n a d j u n c t o f
11 12 13 14 15 16 17 18 19
Gill 1984; Chauvel & Jahn 1984; Downes 1984; White & Hofmann 1982; Clague & Frey 1982; Chen & Frey 1983; Stille et al. 1983; Thirlwall 1982; Menzies & Halliday 1984; Bielski-Zyskind et al. 1984; Smith 1983; Menzies & Murthy 1980a; Menzies et al. 1987, and references cited therein; Menzies & Hawkesworth 1987.
m a g m a t i c processes, i.e. r e p e a t e d cycles o f m a g m a t i s m in t h e m a n t l e . R e g a r d l e s s o f this d i l e m m a , the basalt and inclusion data reveal mantle r e g i o n s d e p l e t e d in t h e light R E E a n d v a r i a b l y d e p l e t e d or e n r i c h e d in Rb. T h e latter e n r i c h m e n t m u s t be a r e l a t i v e l y r e c e n t p h e n o m e n o n (less t h a n 0.2 x 103 M a ago) w i t h i n t h e l i f e t i m e o f t h e
18
M. Menzies I
I
0.5132
I
I
I
t
/
MIDDLE EASTt!10 (Jd~ ~L'T-----":~ ~?-al ,.? --ETHIOPIA
I..~--~ I iFRANCE i I&4+1o /
0.5132~.~ , ~ I~-
~
GERMANY|
"
0.5128 0.5124 1 0.702
I
0.704
,'
l
o.51241-,
I
,
0.702
0,706 I I I -]40
I ~
F,
,
0.704 ~
'
I '
, 4-5
0.706 ' |
sw u
~ 0,5124[1
0.702
l
0704
0.5130
l
; 1-s
0706
TANZANIA
1
L I
I
-
0.5124 [-1 I 0.702 5
,o
-25 t
0.750
t
0.800
I ~ I 0704
2 I ~ -5 0.706
Magmas
-15 -20
~
II I
0.700
0 . 5 1 2 8 ~
[ 1_0
. . . .
F 05110t
o
L
0.850
-30
(j
Inclusions z
%9
87Sr/86Sr
FIG. 2. Nd versus Sr isotopic variation (present day) in mantle ultramafic and mafic inclusions and their host basalts from rifted continental regions. Note the following : (i) the continental host magmas are derived from source regions not unlike the oceanic magmas in Fig. 1 ; (ii) in the majority of cases the host magmas and inclusions are derived from different mantle source regions; (iii) in the case of the African rift valley, fragments of heterogeneous mantle have been entrained. Data are taken from several sources. Middle East-Ethiopia: Menzies & Murthy (1980a); Betton & Civetta (1984); Menzies & Hawkesworth (1987). France and Germany: Stosch et al. (1980); Chauvel & Jahn (1983); Downes (1984). Eastern Australia: Menzies & Wass (1983). SW U.S.A. : Jagoutz et al. (1980); Stosch et al. (1980); Menzies & Hawkesworth (1987); Menzies et al. (1985a); Van Kooten et al. (1985); Zindler & Jagoutz (1987). ocean basin. It can be deduced from these data that the source regions of alkaline magmas and of the overlying mantle as represented by the inclusions are isotopically quite similar. Exposures of peridotite in the ocean basins (Roden et al. 1984b) exhibit isotopic variability that falls within the limits of the data in Fig. 1. Roden et al. (1984b) believe that the St Paul's rocks contain evidence of a recent (less than 0.2 • 103 Ma ago) trace-element enrichment related to alkaline magmatism.
Continental rift valleys Inclusion-bearing alkaline rocks erupted in the major rift valleys of the world have present-day Sr and Nd isotopic ratios not unlike those of the inclusion-bearing alkaline rocks from ocean basins. 87Sr/S6Sr ranges from 0.7027 to 0.7055, and 143Nd/144Nd ranges from 0.51305 to 0.51245 (Table 1). Interestingly, however, the marked disposition of the oceanic data towards high 87Sr/ S6Sr ratios is not evident in the continental
alkaline rocks. In fact, these data are more inclined to plot towards lower 143Nd/144Nd ratios (Fig. 2). Again, the alkaline host magmas are derived from a mantle source depleted in Rb and the light REE for a considerable period of time. Modifications to this mantle source have clearly occurred and account for the scatter of the data. For example, alkaline rocks from the Middle East (Betton & Civetta 1984; Menzies & Hawkesworth 1987) and SW U.S.A. (Van Kooten et al. 1985) plot towards more radiogenic 87Sr/86Sr ratios, a fact that indicates input of Rb into the source region or mixing of end-members with diverse 8VSr/86Sr ratios. The variety of mafic and ultramafic inclusions entrained in alkaline magmas erupted in rift valleys exposes a complex mantle structure that is at first glance extremely heterogenous for both Sr and Nd isotopes and is thus distinct from oceanic mantle. It is apparent, however, that the majority of inclusions have 87Sr/86Sr ratios ranging from 0.702 to 0.7065 and t43Nd/t44Nd ratios ranging from 0.51330 to 0.51240. Considering the paucity of data from ocean basins and the prolific number of isotopic
Alkaline
rocks and their inclusions
analyses from continental rift valleys, these data are only marginally different from oceanic-basin data. However, isotopic data from Tanzania appears to complicate matters (Fig. 2) since the inclusions display an extreme range in Sr and Nd isotopic composition. Nothing of this type has been seen in the ocean basins or in most of the continental rift valleys (compare Figs 1 and 2). These extremely radiogenic Sr and non-radiogenic Nd isotopic ratios record ancient (more than 103 Ma ago) trace-element enrichments involving addition of the light R E E and/or Rb to the mantle. With the passage of time, the isotopic ratios have adjusted so as to record the modified
0.5120
~
/
19
trace-element ratios. It is unlikely that such old trace-element enrichment events would be preserved in actively convecting asthenosphere, A mantle layer with fundamentally different isotopic characteristics clearly exists beneath the continents and not beneath the ocean basins. Furthermore, the similarity between oceanic and continental rift valley alkaline basalts points to a similar mantle source at depth for the host magmas. While most of the inclusions record an old depletion in both Rb and the light REE not unlike the oceanic mantle, certain fragments of 'anomalous' enriched mantle are occasionally entrained within continental rift valley magmas.
~-10 0'5125I
0,110
'
1o
0.702 Q704 0.706 0.708 0.710
0.700 0.710 0.720
( •Host Magmas 0.5132 0.5130 0.5128 0.5126
!
~
I
~
0.5124
0.740
( ~ Inclusions
I
~ t
0.730
I
...
-,10
I
SOUTH
AFRICA
-+~
o ~
' ,NOLUS,O,S
/
o.5122 o.5120 t Z ,r,-~., 0.5118 ~- I "O z
OY02
-
GROUP II~ K,MBERL,TES garnet inclusions in diamonds I
I
I
I
0.706
0.710
0.714
0.718
-lO
-15 "t~ Z
87Sr / 86 Sr
FIG. 3. Nd and Sr isotopic variation (present day) in mantle ultramafic and mafic inclusions and their host basalts and kimberlites from other continental regions. Note the following: (i) the host alkaline magmas from Scotland and Antarctica and the basaltic kimberlites (Group 1) from S Africa have Sr and Nd isotopic compositions not unlike OIBs; (ii) the mantle inclusions display, in most instances, a greater range of Sr and Nd isotopic ratios than the host magmas; (iii) a greater proportion of the inclusions than were entrained in oceanic basalts are from regions of the mantle enriched in Rb and light REE for a considerable time period; (iv) micaceous kimberlites (Group II) have Sr and Nd isotopic ratios distinctly different from those of OIB and similar to some of the inclusions. Data are taken from several sources. Scotland: Menzies & Halliday (1984); Halliday et al. (1985). Antarctica: Stuckless & Ericksen (1976); Kyle et al. (1985); Menzies et al. (1985b). South Africa: Menzies & Murthy (1980b); Kramers et al. (1981 ; 1983); Erlank et al. (1982, 1987); Smith (1983); Menzies et al. (1987).
M. Menzies
20
Although the source regions of the host magmas may be similar, clearly the potential exists in rift valleys for magmas to entrain fragments of mantle that does not exist under ocean basins.
istries. The range in isotopic composition is extreme, particularly with respect to Sr isotopic ratios (Fig. 3). These isotopic data record a timeintegrated response to (a) major partial melting or depletion events (similar to that observed in the alkaline volcanic rocks erupted in the oceans and rift valleys) and (b) ancient enrichment or metasomatic events caused by infiltration of silicate melts and aqueous fluids that have not as yet been reported from the ocean basins.
Other continental regions Alkaline basalts, lamproites, lamprophyres and kimberlites intrude regions of old crust. The alkaline basalts have Sr and Nd isotopic ratios consistent with derivation from either enriched (e.g. Scotland) or depleted (e.g. Antarctica) portions of the Earth's mantle (Fig. 3). Moreover, most of these continental alkaline basalts and basaltic kimberlites (Group I) are isotopically indistinguishable from ocean island alkaline basalts or rift valley basalts (Table 1). In contrast, micaceous kimberlites (Group II) and lamproites have more radiogenic Sr and less radiogenic Nd isotopic ratios than anything erupted in the ocean basins or rift valleys. In the case of the Group II kimberlites and lamproites we must invoke a source region distinctly different from that of alkaline rocks erupted in the ocean basins or continental rift valleys. It is perhaps to be expected that the petrographically and chemically diverse mafic and ultramafic inclusions entrained from below regions of old crust would display heterogeneous isotopic chemI
I
Asthenosphere-lithosphere The available host magma and inclusion data can now be integrated into some kind of Earth model. To assist us, the Sr and Nd isotopic data for continental and oceanic alkaline basaltic rocks (excluding Group II kimberlites) are compared with available data from several ocean islands and MORB (Fig. 4). The first point to be made is that the source is identical (Fig. 4) regardless of whether the rocks are erupted in continents or ocean basins. Their source must therefore lie not in the lithosphere but in or below the convecting asthenosphere since (a) the oceanic lithosphere is relatively sterile and not capable of producing large volumes of alkaline magma and (b) the I
I
I
+15
0.5134 MORB 0.5132
+1o
~Nd
0.5130 4"
A
+5
V
0.5128
<>
erguelen It
0.5126 9
A9 ~9
9
"-~
,O-
0.5124 I
0.702
0.703
-5
I
0.704
0.705
0.706
87Sr/86Sr
FIG. 4. Nd versus Sr isotopic variation in alkaline volcanic rocks and basaltic kimberlites. Comparative basalt data are taken from Dosso & Murthy (1980), Cohen & O'Nions (1982), Chen & Frey (1983) and Stille et al. (1983). Other sources of data are: o, San Quintin, Mexico (De Paolo 1978); , San Carlos, Arizona (Zindler & Jagoutz 1987); o, Geronimo, Arizona (Menzies et al. 1985a); O, Pisgah Crater, California (De Paolo 1978); ~ Nunivak, Alaska (Menzies & Murthy 1980a); ,t, Ataq, South Yemen (Menzies & Murthy 1980a); 4, Kiama, New South Wales (Menzies & Wass 1983); ~, Afar, Ethiopia (Betton & Civetta 1984); in, Massif Central, France (Chauvel & J ahn 1984); ~ , Malaita, Solomon Islands (Bielski-Zyskind et al. 1984); 4,, Victoria, Australia (McDonough & McCulloch 1985); ~, Sierran Province, California (Van Kooten et al. 1985); ~, Ross Island and vicinity, Antarctica (Menzies et al. 1985b); V, Tanzania (Cohen et al. 1984); v, Tahiti (White & Hofmann 1982); v, Hawaii (Stille et al. 1983); , , Scotland (Menzies & Halliday 1984; Halliday et al. 1985); *, South Africa (Smith 1983).
Alkaline rocks and their inclusions continental lithosphere is heterogeneous and differs markedly from the oceanic lithosphere. The second point to note concerns the similarity between mantle sampled by volcanic rocks erupted in ocean basins and continental rift valleys (summarized in Fig. 5). Essentially this fragmented and entrained lithosphere has a prehistory not unlike the source regions of alkaline magmas ultimately responsible for the transport of inclusions to the surface. This can be interpreted to mean that lithosphere has been rehomogenized by impingement of asthenospheric melts. Since rift valleys are relatively young (less than 0.2 x 10a Ma) any trace-element enrichment locked into the lithosphere would have had insufficient time to evolve isotopically or would have been obliterated by recent rehomogenization of the isotope systems. The third point has relevance to the presence
+1!
+10 I
+5 I E~
r ANTARCTICA
<~ISCOTLAND
0 I
-5 I
21
of aged isotopically heterogeneous lithosphere below the stable continental cratons. This is sampled during eruption of magmas through stable cratons or in some rift valleys (e.g. Tanzania) (Fig. 5). As a group, ultramafic and mafic inclusions from alkaline and kimberlite host rocks erupted through cratonic regions have a considerable range in Nd and Sr isotopic composition (Fig. 3). These values encompass mantle both enriched and depleted in light REE for a prolonged period of time (Menzies & Murthy 1980b; Erlank et al. 1982, 1987; Kramers et al. 1983; Cohen et al. 1984). Inclusions with the least radiogenic Nd isotopic composition were entrained by kimberlitic and alkaline magmas erupted through old stable cratonic regions (Menzies & Murthy 1980b; Cohen et al. 1984; Menzies & Halliday 1984; Richardson et al. 1984). Diopsides sepa-
-10 I
-15 I
-20 I
-25 I
s
-30 I
Nd
] ] ULTRAMAFIC
O~LSOUTH AFRICA
XENOLITHS
(Lithosphere)
F TANZANIA
1
CALIFORNIA-
i
] COL::A:~N-A NEW MEXICO
I
......
I
El AUSTRALIA ~:
[ EUROPE k S.YEMEN RED SEA GUADALUPE HAWAII
~
~ ~
HOST
MAGMAS
(Asthenosphere)
TAHITI MALAITA NEW ZEALAND
t
i
' ULTRAMAFIC HOST
ALASKA ALPINE
PERIDOTITE 0.5135
I 0.5130
I 0.5125
I 0.5120
1 0.5115
XENOLITHS
MAGMA
I 0.5110
84
I 0.5105
0.5100
143Nd / 144Nd
FIG. 5. Nd isotopic variation in host magmas and mantle inclusions erupted through stable cratons, in rift valleys and ocean basins. Note that, regardless of the tectonic setting, the alkaline host magmas are derived from a mantle source region with a narrow end range of + 8 to -- 7, similar to that observed in oceanic basalts. The inclusions exhibit a considerable range of eNdrelative to the host magmas. Mantle entrained from below stable cratons (e.g. Antarctica, Scotland and South Africa) is heterogeneous, whereas mantle sampled in rift valleys and ocean basins is not unlike MORB-OIB. Rift valleys formed through regions of old crust (e.g. Tanzania), however, appear to sample both types of mantle. Arguments can be put forward for the presence of old heterogeneous lithospheric mantle below stable cratons, which tends to be disrupted and replaced by asthenospheric processes during rift formation. Well established rift valleys would eventually rupture this carapace of aged lithosphere by upwelling of asthenospheric melts. Data are taken from the sources given in the captions to Figs 1-3 and data for Guadalupe and New Zealand (Menzies, unpublished data) and alpine peridotites (Menzies 1984, and references cited therein).
M. Menzies
22
formed by asthenospheric processes. Similarly, in the ocean basins the relatively young age of the crust makes the widespread survival of aged cold stable lithosphere somewhat unlikely.
rated from garnet peridotites have model Nd ages of less than 3.0 x 103 Ma (Menzies et al. 1985b; Erlank et al. 1987). Such ages indicate stabilization and isolation of the continental lithosphere for a considerable portion of the Earth's history. Furthermore, the isotopic variability of the lithospheric mantle overlaps that of the crust. Inclusions with an isotopic signature equivalent to M O R B or OIB have been found in oceanic regions and in continental regions that have experienced recent extension and presumably upwelling (i.e. rift valleys). Inclusions with isotopic ratios distinct from MORB and OIB are sampled by kimberlite or alkaline magmas erupted through old cratonic regions. In the case of stable cratons, the host magma is isotopically equivalent to OIB while the inclusions display a distinct isotopic signature unlike anything recovered from regions of young crust or the ocean basins (Fig. 5). It seems logical to infer that the difference must be that an aged heterogeneous component exists below regions of old crust. During rift valley formation this lithospheric carapace was presumably ruptured and trans-
Models for genesis of alkaline rocks The identification and characterization of the continental lithosphere, tectosphere or keel (Jordan 1975; Brooks et al 1976; Oxburgh & Parmentier 1978) has added a new dimension to our interpretation of the isotopic and traceelement characteristics of volcanic rocks erupted through continental crust. The continental lithosphere is defined as that portion of the upper mantle that moves as a coherent mechanical unit with a velocity similar to the surface crustal plate. Thermal and/or diapiric accretion may be responsible for stabilization of this region of the mantle (Jordan 1978; Oxburgh & Parmentier 1978). The thickness of the lithosphere below continental regions remains the subject of some debate (Jordan 1978; Oxburgh 1981). It is generally
TABLE 2. Source regions of inclusion-bearing alkaline and kimberlitic magmas Locality Afar (Assab) Africa (Tanzania) Alaska (Nunivak) Antarctica Foster Crater Ross Island Arizona Geronimo San Carlos Australia Kiama Kimberley Victoria California (Sierran Province)
Ref.*
Source region
1 Mantle mixing sources MORB and subducted lithosphere 2 Depleted mantle source 3 Mantle sourceOIB 4 Asthenospheric processes=OIB source 5 OIB source 6 OIB source 7 8 9 10
Mantle source modified by CO, MORB-enrichedmantle mixing Sub-continental heterogeneous mantle Subduction component in mantle-micaceous source
Fiji France (Massif Central) French Polynesia (Tahiti) Hawaii (Honolulu Series) Nigeria (Bokkos) Scotland (N Highlands) Solomon Islands (Malaita) South Africa (Various localities) South Yemen (Ataq)
11 OIB pods or dykes in MORB 12 Metasomatizedmantle source 13 Mantle source with recycled component 14 Multistage history of lithosphere-asthenosphere mixing 19 Fractionation in mantle, crustal contamination of evolved rocks 15 Old enriched sources: lithosphere-asthenosphere mix 16 Young depleted mantle plus recycled component 17 Lithosphere (enriched)--asthenosphere (depleted) sources 18 Mantle heterogeneity
* References 1-18 as in Table 1 ; 19, Irving & Price 1981.
Alkaline rocks and their inclusions accepted, however, that the sub-continental lithospheric mantle is considerably thicker than the crust and is of the order of 100 km or more compared with tens of kilometres beneath ocean basins. While the lithosphere below oceans is relatively young (less than 0.2x 103 Ma) and geochemically uncomplicated, the continental lithosphere retains a complex history that has been preserved in the stable non-convecting environment below the cratons. It has been suggested (Oxburgh & Parmentier 1978) that the lithosphere records the last major thermal event in the overlying crust. If so, aged heterogeneous mantle should be restricted to regions of Proterozoic and Archean crust. The prolonged stability of this region throughout geological time has important implications for the geochemical evolution of the sub-continental mantle. Brooks e t al. (1976) proposed that the keel soaked up fluids and melts from the asthenosphere and as a consequence the subcontinental lithosphere may contain a significant fraction of the incompatible-element inventory of the Earth. In theory, upwelling of silicate melts and aqueous fluids from the asthenosphere will modify the trace-element inventory of the lithospheric mantle. These enhanced trace-element
23
ratios will eventually register as modified isotopic ratios in regions of the mantle left undisturbed for considerable periods of time (more than 103 Ma). Extremely heterogeneous mantle will be produced by such processes. Any study of the evolution of continental alkaline (and tholeiitic) magmas must acknowledge the presence of aged heterogeneous lithosphere that may attain thicknesses in excess of 200 km below continental regions. Any magma believed to have been generated in the asthenosphere or lower mantle must first rupture this carapace of lithospheric mantle and then the crust, and establish by magma fracture a network of conduits to the surface. Since lithospheric mantle will be considerably thicker than crust, it presents a potential source of 'mantle contamination' for the asthenospheric melts. Mixing of asthenospheric and lithospheric melts may be a more significant process than crustal interaction in determining the isotopic and chemical characteristics of alkaline magmas (Table 2). Group II kimberlites and lamproites have Sr and Nd isotopic ratios distinctly different from those of OIB and more akin to those of crustal rocks or ultramafic xenoliths disrupted from the mantle below regions of old crust (Fig. 6).
I
~Nd
'
....MORB ~---~ASTHENOSPHERE 0.5130
~
R
-+10
BASALTS
m
0
OUP i K]-MBERLITES GROUP II KIMBERLITES
"O
z
"-~. 0 . 5 1 2 0
-
-10
-
-20
-
-30
LAMPROITES ~ "" ~
"O
z ,r-
'
[LITHOSPHERE I MANTLE INCLUSIONS
\MANTLE INCLUSIONS 0.5110 I,
I
I
I
0700
0750
0~00
0.850
87Sr/86Sr FIG. 6. t43Nd/~44Ndversus 878r/S6Sr variation in mantle inclusions and inclusion-bearing alkaline rocks. The extreme range in Nd and Sr isotopic composition defined by minerals separated from ultramafic inclusions encompasses the variation in Nd and Sr isotopic composition observed in alkaline rocks. Alkaline basalts and Group I kimberlites have isotopic characteristics identical with those of OIB derived from asthenosphere. In contrast, lamproites and Group II kimberlites may have a lithospheric component in their source regions. Data are taken from Menzies & Murthy (1980a, b), Erlank et al. (1987), Menzies (1983, and references cited therein), Menzies & Wass (1983), Cohen et al. (1984) and Menzies & Halliday (1984). Other data are taken from the references given in the caption to Fig. 4.
z4
M. Menzies
Contamination with significant amounts of crust is unlikely because of the rapidity with which these melts must reach the surface in order to maintain their load of high-density inclusions and diamonds. Furthermore, the extremely high concentrations of Sr (678-1845 ppm) and Nd (48-336 ppm) would require mixing with a highly radiogenic crustal component or stoping and assimilation of considerable amounts of crust. This is believed to be very unlikely. Now the problem arises of exactly which structural elements in the sub-continental mantle contribute to the genesis of these rocks. The isotopically heterogeneous nature of lamproites compares well with the heterogeneities recorded in continental lithospheric mantle inclusions (Fig. 6). Arguments could therefore be presented for a source region solely within the lithospheric mantle. This is rather unlikely as an asthenospheric component or influence may be needed to provide heat and volatiles and to trigger anatexis. Impingement of an asthenospheric plume on the continental lithosphere is believed to elevate the isotherms and induce melting in the lithospheric mantle immediately above the plume. The column of mantle above the plume may display a considerable range of Sr, Nd and Pb isotopic composition. To illustrate this it is
worth noting that ultramafic inclusions from the Bultfontein pipes, S Africa (Fig. 3), have an extreme range in Sr and Nd isotopes. If we accept that these inclusions represent a random vertical profile through the lithosphere, melts extracted from such a heterogeneous mantle would tend to be distinctly different from OIBs and the bulk of continental basalts particularly with regard to their isotopic ratios. If lamproites and micaceous kimberlites contain a major contribution from the lithospheric mantle it follows that their isotopic variability could theoretically be as great as that observed in lithospheric inclusions (Fig.
6). As more data become available, the range of Sr, Nd and Pb isotope ratios observed in mantle rocks is continually revised. Therefore a unique solution to the origin of lamproites and Group II kimberlites is impossible. It is perhaps prudent to limit comments on their origin to interaction between relatively homogeneous asthenosphere and extremely heterogeneous aged lithospheric mantle. The observed isotopic ratio in erupted lamproite or micaceous kimberlite will depend on (1) the mixing proportions of asthenosphere and lithosphere, (2) the Sr/Nd ratio in the asthenospheric and lithospheric end-members and (3) the isotopic ratio of these end-members.
FIG. 7. Earth models pertinent to the origin of inclusion-bearing alkaline rocks (IBAR). The primordial hotmantle plume (PHMP) and depleted low velocity layer (DLVL) model of Schilling (1973) and the model of Tatsumoto (1978) are based on REE and Pb isotopes. The third model is modified after Wasserburg & De Paolo (1979) and incorporates ideas from McCulloch et al. (1983), Smith (1983) and Chen& Frey (1983). The most significant aspect of this model is the involvement of the continental lithospheric mantle (CM) in the genesis of lamproites and Group II kimberlites. Wasserburg & De Paolo (1979) believe the CM to be 'sterile', but this is not compatible with geochemical data on inclusions. Continental flood basalts may have a contribution from CM early in their evolution while upwelling of asthenosphere triggers melting in the lithosphere and encourages mixing of the melts. Eventually, once the volcanic conduits are well established, continental flood basalts will essentially have the isotopic signature of OIBs. (CC, continental crust; OC, oceanic crust; IBAR, inclusion bearing alkaline rocks.)
Alkaline rocks and their inclusions McCuUoch et aL (1983) pursued a somewhat similar line of reasoning in their interpretation of the isotopic and trace-element variation observed in Australian lamproites. Mixing of a depleted end-member, not unlike OIB in isotopic composition, and an enriched mantle end-member, not unlike average crust in isotopic composition, produced the required range of St and Nd isotopes observed in the diamond-bearing lamproites. Similarly, in a detailed study of Group I (basaltic) and Group II (micaceous) kimberlites, Smith (1983) outlined a model of asthenosphere-lithosphere interaction. Since Group I kimberlites are isotopically identical with OIBs, their evolutionary history was believed to involve derivation from asthenospheric sources with little or no lithospheric mantle input. In contrast, Group II kimberlites require a more dominant lithospheric mantle component because of their radiogenic Sr and non-radiogenic Nd isotopic composition. McCulloch et al. (1983) defined a unique composition for the enriched mantle end-member. As more and more data become available, it is obvious that the sub-continental lithospheric mantle is extremely heterogeneous and could provide a variety of end-members for mixing models. To use the terminology of Powell and Bell (1974), several suites of garnet peridotites and pyroxenites stable at mantle pressures and temperatures have 'crustal-type' St, Nd and Pb isotopic compositions. As a consequence of this, the isotopic variability in the continental mantle is almost as diverse as in the overlying crust,
25
perhaps owing to incorporation of subducted melts or fluids over billions of years. Continental lithospheric mantle is in part transitional between M O R B - O I B type mantle and the crust. In its trace-element geochemistry and ultimately in its isotope geochemistry certain portions of the lithospheric mantle have more similarity to the crust than to M O R B - O I B mantle. It remains to be seen whether this is due to diapiric upwelling related to subduction of continental material or to other recycling processes.
Summarizing comments Impingement of asthenospheric or lower mantle plumes on oceanic or continental lithosphere is believed to initiate anatexis and promote mixing of isotopically distinct melts. Interaction with MORB-like lithosphere in ocean basins and continental regions gives rise to inclusion-bearing alkaline magmas and basaltic kimberlites (Group I) with isotopic signatures not unlike M O R B and OIB ( e N d = 8 to -- 9 ; eSr = -- 28 to + 34) (Table 2). Interaction with aged enriched continental lithospheric mantle may produce a more diverse suite of melts with isotopic characteristics unlike M O R B - O I B and more like that of ultramafic inclusions entrained in alkaline and kimberlitic melts (eNd= -- 24 to + 15 ;/3Sr= - - 35.5. to + 1867). Magmas believed to be produced by this process are micaceous kimberlites (Group II) and certain lamproites (Table 2).
References BETTON,P. J. & CIVETTA,L. 1984. Sr and Nd isotopic evidence for the heterogeneous nature and development of the mantle beneath Afar (Ethiopia). Earth planet. Sci. Lett. 71, 59-70. BIELSKI-ZYSKIND,M., WASSERBURG,G. J. & NIXON,P. H. 1984. Sm-Nd and Rb-Sr systematics in volcanics and ultramafic xenoliths from Malaita and Solomon Islands and the nature of the Ontong Java Plateau. J. geophys. Res. 89, 2415-24. BROOKS,C., JAMES,D. E. & HART,S. R. 1976. Ancient lithosphere: its role in young continental volcanism. Science, 193, 1086-994. CHASE, C. G. 1981 Oceanic island Pb: two-stage histories and mantle evolution. Earth planet Sci. Lett. 52, 277-84. CHAUVEL,C. & JAHN, B.-M. 1984. Nd-Sr isotope and REE geochemistry of alkali basalts from the Massif Central, France. Geochim. cosmochim. Acta, 48, 269-78. CHEN, C.-Y. & Frey, F. A. 1983. Origin of Hawaiian tholeiite and alkalic basalt. Nature, Lond. 302, 785-9.
CLAGUE, D. & FREY, F. A. 1982. Petrology and trace element geochemistry of the Honolulu volcanics, Oahu: implications for the oceanic mantle below Hawaii. J. Petrol. 23, 447-504. COHEN, R. S. & O'NIONS, R. K. 1982. The lead, neodymium and strontium isotopic structure of ocean ridge basalts. J. Petrol. 23, 299-324. , & DAWSON,J. B. 1984. Isotope geochemistry of xenoliths from East Africa: implications for development of mantle reservoirs and their interaction. Earth planet. Sci. Lett. 68, 209-20. DE PAOLO,D. 1978. Nd and Sr isotopic systematics of young continental igneous rocks. Short Papers of the 4th ICGCIG. Geological Survey Open File Report 78-701, pp. 91-3. United States Department of the Interior, Geological Survey, U.S. Printing Office. Dosso, L. & MURTHY, V. R. 1980. An Nd isotopic study of the Kerguelen islands: inferences on enriched oceanic mantle sources. Earth planet Sci. Lett. 48 268-76. DOWNES, H. 1984. Sr and Nd isotopic geochemistry of co-existing alkaline magma series, Cantal massif
26
M. Menzies
Central, France. Earth planet Sci. Lett. 69, 221334. ERLANK, A. E., ALLSOPP, H., HAWKESWORTH, C. J. & MENZIES, M. A. 1982. Chemical and isotopic characterization of upper mantle metasomatism in peridotite nodules from the Bultfontein kimberlite. Terra Cognita 2, 261-3. --, WATERS, F. G., HAWKESWORTH, C. J., HAGGERTY, S. E., ALLSOPP, H. L., RICKARD, R. S. & MENZIES, M. A. 1987. Evidence for mantle metasomatism in peridotite nodules from the Bultfontein floors, Kimberley, S. Africa. In : MENZIES, M. A. & HAWKESWORTH, C. J. (eds) Mantle Metasomatism, Academic Press, New York, 221311. FREY, F. A. & GREEN, D. H. 1974. The mineralogy, geochemistry and origin of lherzolite inclusions in Victorian basanites. Geochim. cosmochim. Acta, 38, 1023-59. GILL, J. B. 1984. Sr-Pb-Nd isotopic evidence that both MORB and OIB sources contribute to ocean island arc magmas in Fiji. Earth planet. Sci. Lett. 68, 44358. HALLIDAY,A. N., MENZIES,M. A. PALACZ,Z., HUNTER, R. H. & HAWKESWORTH, C. J. 1985. Hebridean mantle geochemistry. Trans. Am. Geophys. Union, 66, 414. HAWKESWORTH, C. J., ROGERS, N., VAN CALSTEREN, P. W. C. & MENZIES, M. A. 1984. Mantle enrichment processes. Nature, Lond. 311, 331-55. HOFMANN, A. & WrotE, W. M. 1982. Mantle plumes from ancient oceanic crust. Earth planet Sci. Lett. 57, 421-36. IRVING, A. J. & FREY, F. A. 1984. Trace element abundances in megacrysts and their host basalts: constraints on partition coefficients and megacryst genesis. Geochim. cosmochim. Acta, 45, 1201-21. & PRICE, G. 1981. Geochemistry and evolution of Iherzolite-bearing phonolitic lavas from Nigeria, Australia, East Germany and New Zealand. Geochim. cosmochim. Acta, 45, 1309-20. JAGOUTZ, E., CARLSON, R. W. & LUGMAIR, G. 1980. Equilibrated Nd, unequilibrated Sr isotopes in mantle xenoliths. Nature, Lond. 286, 708-10. JORDAN, T. H. 1975. The continental tectosphere. Rev. Geophys. space Phys. 13, 1-12. - 1978. Composition and development of the continental tectosphere. Nature, Lond. 274, 544-8. KRAMERS, J., RODDICK, J. & DAWSON, J. 1983. Trace element and isotope studies on veined, metasomatic and 'MARID' xenoliths from Bultfontein, South Africa. Earth planet. Sci. Lett. 65, 90-106. - - , SMITH, C. B., LOCK, N. P., HARMON, R. S. & BOYD, F. R. 1981. Can kimberlites be generated from ordinary mantle? Nature, Lond. 291, 53-6. KYLE, P., KIRSCH, I., GAMBLE, J. & MENZIES, M. 1985. Metasomatised ultramafic xenoliths from Foster Crater, Antarctica. Trans. Amer. geophys. Union 66, 409. MCCULLOCH, M. T., JAQUES, A. L., NELSON, D. R. & Lewis, J. D. 1983. Nd and Sr isotopes in kimberlites and lamproites from Western Australia: an enriched mantle origin. Nature, Lond. 302, 400-3. MCDONOUGH, W. F. & MCCULLOCI-I, M. T. 1985.
Isotopic and geochemical systematics in TertiaryRecent basalts from southeastern Australia and implications for the evolution of the sub-continental lithosphere. Geochim. cosmochim. Acta 49, 205168. MCKENZlE, D. P. & O'Nions, R. K. 1983. Mantle reservoirs and ocean island basalts. Nature, Lond. 301, 229-31. MENZlES, M. A. 1983. Mantle ultramafic xenoliths in alkaline magmas: evidence for mantle heterogeneity modified by magmatic activity. In: HAWKESWORTH, C. J. & NORRY, M. J. (eds) Continental Basalts and Mantle Xenoliths, pp. 111-38. Shiva, U.K. - - 1 9 8 4 . Chemical and isotopic heterogeneities in orogenic lherzolite massifs. In: GASS, I. G., LIPPARD, S. J. & SrIELTON, A. W. (eds), Ophiolites and Oceanic Lithosphere, Geol. Soc. Spec. Pubi. 13, pp. 231-40. - & HALLIDAY,A. N. 1984. Enriched mantle below the Scottish Highlands: Sr and Nd isotope and rare earth elements in peridotite and pyroxenite xenoliths from Loch Roag, Streap Comlaidh and Kilchatten, Scotland. Abstracts of 27th Int. Geological Cong., Moscow, p. 348. - - & HAWKESWORTH, C. J. 1987. Mantle xenoliths and upper mantle structure. In: NIXON, P. (ed.) Mantle Xenoliths, Wiley, New York, in press. --& MURTHY, V. R. 1980a. Nd and Sr isotope geochemistry of hydrous mantle nodules and their host alkali basalts: implications for local heterogeneities in metasomatically veined mantle. Earth planet Sci. Lett. 46, 323-34. - - & 1980b. Enriched mantle: Nd and Sr isotopes in diopsides from kimberlite nodules. Nature, Lond. 283, 634-6. --& WASS, S. Y. 1983. COz rich mantle below eastern Australia: REE, Sr and Nd isotopic study of Cenozoic alkaline magmas and apatite-rich xenoliths, Southern Highlands Province, New South Wales, Australia. Earth planet. Sci. Lett. 65, 287-302. - - , KEMPTON, P. & DUNGAN, M. 1985a. Interaction of continental lithosphere and asthenospheric melt below the Geronimo Volcanic Field, Arizona, U.S.A.J. Petrol., 26, 663-93. - - , KYLE, P. & GAMBLE,J. 1985b. Fragments of aged enriched sub-continental lithosphere entrained in asthenospheric magmas erupted at Foster Crater, Antarctica. Trans. amer. geophys. Union, 66, 409. - - - , ROGERS, N., TINDLE, m. & HAWKESWORTH,C. J. 1987. Metasomatic and enrichment processes in lithospheric peridotites, an effect of asthenosphere-lithosphere interaction. In: MENZlES, M. A. & HAWKESWORTI.I,C. J. (eds) Mantle Metasomatism, Academic Press, New York, 313-61. OXBURGH, E. R. 1981. Heat flow and differences in lithospheric thickness. Phil. Trans. R. Soc. Lond., Ser. A, 310, 337-46. - & PARMENTIER,E. M. 1978. Thermal processes in the formation of continental lithosphere. Phil. Trans. R. Soc. Lond., Set. A, 288, 415-29. POWELL, J. L. & BELL, K. 1974. Isotopic composition of strontium in alkalic rocks. In: SORENSEN, H.
Alkaline rocks and their inclusions (ed.) The Alkaline Rocks, pp. 412-420. Wiley, New York. RICHARDSON, S. H., GURNEY, J. J., ERLANK, A. J. & HARRIS, J. W. 1984. Origins of diamonds in old enriched mantle. Nature, Lond. 310, 198-202. ROCK, N. M. S. 1976. The comparative strontium isotopic composition of alkaline rocks: new data from southern Portugal and East Africa. Contrib. Mineral. Petrol. 56, 205-28. RODEN, M. F., FREY, F. A. & FRANCIS, D. i . 1984a. An example of consequent mantle metasomatism in peridotite inclusions from Nunivak Island, Alaska. J. Petrol. 25, 546-77. - - , HART, S. R., FREY, F. A. & MELSON, W. G. 1984b. Sr, Nd and Pb isotopic and REE geochemistry of St. Paul's Rocks: the metamorphic and metasomatic development of an alkali basalt mantle source. Contrib. Mineral. Petrol. 85, 379400. SCHILLING,J-G. 1973. Iceland mantle plume: geochemical evidence along the Reykjanes Ridge. Nature, Lond. 242, 565-571. SMITH, C. B. 1983. Pb, Sr and Nd isotopic evidence for sources of southern African Cretaceous kimberlites. Nature, Lond. 304, 51-4. STILLE, P., UNRUH, D. & TATSUMOTO,M. 1983. Pb, Sr, Nd and Hf isotopic evidence of multiple sources for Oahu, Hawaii basalts. Nature, Lond. 304, 259. STOSCH, H. G., CARLSON, R. W. & LUGMAIR, G. W. 1980. Episodic mantle differentiation: Nd and Sr isotopic evidence. Earth planet. Sci Lett. 47, 26371.
27
STUCKLESS, J. S. & ERICKSEN, R. L. 1976. Strontium isotope geochemistry of the volcanic rocks and associated megacrysts and inclusions from Ross Island and vicinity, Antarctica. Contrib. Mineral. Petrol. 58, 111-26. TATSUMOTO, M. 1978. Isotopic composition of lead in oceanic basalts and its implication to mantle evolution. Earth planet. Sci. Lett. 38, 63-87. THIRLWALL,i . 1982. Systematic variation in chemistry and Nd-Sr isotopes across a Caledonian calcalkaline volcanic arc: implications for source materials. Earth planet. Sci. Lett. 58, 27-50. VAN KOOTEN, G., LEEMAN, W. P. & MENZIES, M. A. 1985. Lithospheric peridotites and pyroxenites in alkaline basaltic rock from the Sierra Nevada, California, U.S.A. Trans. Amer. Geophys. Union, 66, 414. VIDAL, P., CHAUVEL, C. & BROUSSE, R. 1984. Large Mantle Heterogeneity beneath French Polynesia. Nature, Lond. 307, 536-8. WASS, S. Y. & ROGERS, N. K. 1980. Mantle metasomatism--precursor to continental alkaline volcanism. Geochim. cosmochim. Acta, 44, 1811-23. WASSERBURG, G. J. & DE PAOLO,D. J. 1979. Models of earth structure inferred from neodymium and strontium isotopic abundances. Proc. Natl Acad Sci. 76, 3594-8. WHITE, W. M. & HOFMANN, A. W. 1982. Sr and Nd isotope geochemistry of oceanic basalts and mantle evolution. Nature, Lond. 296, 821-5. ZINDLER, A. & JAGOUTZ, E. 1987. Trace element and Nd and Sr isotope systematics ofperidotite nodules from Peridot Mesa San Carlos, Arizona. Geochim. cosmochim. Acta, in press.
MARTIN MENZIES, Department of Geology, University of London, Royal Holloway and Bedford New College, Egham, Surrey, TW20 0EX, U.K.
The genesis of alkaline magmas with emphasis on their source regions: inferences from experimental studies A. D. Edgar SUMMARY: Experimental studies bearing on the genesis of alkaline magmas are reviewed, with emphasis on alkaline mafic-ultramafic magmas and the nature of their mantle sources. In contrast with previous experiments on simplified systems mainly at low pressures, more recent experiments under mantle pressures have resolved some of the problems of the earlier work. Inadequacies of petrogenetic schemes based on the model pyrolite mantle source for the generation of alkaline magma are also partly resolved by recent studies which suggest that with increasing alkalinity the possibility of deriving mafic-ultramafic magmas from a dry lherzolitic mantle source, or one with H20 as the only volatile, becomes increasingly difficult. The addition of CO2 and H20 to the source regions allows generation of magmas such as olivine melilitites, but few magmas of the olivine leucitite kindred can be generated from lherzolitic sources. Such magmas appear to be partial melts of clinopyroxenite or orthopyroxenite sources containing both volatiles. Preliminary experimental modelling of mantle metasomatism suggest that this process may readily occur. Experiments on felsic alkaline rocks with emphasis on liquid immiscibility and halogens and experiments aimed at solving specific problems are briefly considered.
Introduction Experiments on systems pertinent to alkaline rocks have undoubtedly made an immense contribution to our understanding of the genesis of these rocks, particularly in problems such as the attainment of peralkalinity, the depth of derivation of alkaline magmas and, most recently, the variation in the source materials from which alkaline magmas are derived relative to nonalkaline varieties. Some of the major experimental studies carried out since the publication of The Alkaline Rocks by Sorensen (1974) are reviewed in this paper. The term 'alkaline rocks' will be used to denote rocks containing feldspathoidal minerals (modal or normative) or rocks not necessarily containing a feldspathoid but with low SiOz and high alkali contents. Such a definition clearly precludes alkali granites but could include certain kimberlites and carbonatites which are not discussed in any detail. Within the past decade a number of reviews of the role of experimental petrology in our understanding of the genesis of alkaline rocks have been published. Edgar (1974) included data on alkaline systems published up to the early 1970s, while Bailey (1974) reviewed the application of experimental petrology to oversaturated peralkaline volcanic rocks and later (Bailey 1976) discussed experiments applicable to a wide spectrum of alkaline rocks ranging from strongly SiO2-undersaturated mafic varieties to undersaturated and oversaturated felsic types. More recently, the Basaltic Volcanism Study Project
(1981) (see particularly pp. 529-30, 561-2 and 564--7) has reviewed experiments on alkaline and highly undersaturated potassic basalts. Zyryanov (1981) discussed phase relations in feldspars and feldspathoids with some emphasis on petrological aspects, and Kogarko & Romanchev (1982) discussed phase equilbria in alkaline melts. Both of these papers include important references to the Russian literature. Writing from a petrologist's viewpoint, Edgar (1984) discussed the chemistry, occurrence and paragenesis of feldspathoids, including experiments bearing on the origin of pseudoleucites, the formation of melilite in nephelinites, melilitites and alkaline lavas of comparable compositions, geothermometry and the question of the origins of primary analcite. In the past 10-15 years the emphasis of experimental studies bearing on alkaline rock genesis has changed from low-pressure experiments in relatively simple synthetic systems to studies using rock compositions under mantle conditions in excess of 100 km depth and controlled partial pressures of volatiles. As a result of these high-pressure experiments the importance of H20, CO 2 and other volatiles on the genesis and evolution of alkaline magmas is much better known than it was a decade ago. The hypothesis that alkaline magmas as well as kimberlites, carbonatites and lamproites are partial melts of slightly to vastly modified lherzolite mantle sources has been strongly supported by experimental studies and has led to experiments in which processes of mantle metasomatism are being simulated. Investigations
From : FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 29-52.
29
30
A. D. Edgar
have been undertaken to determine the compositions of liquids produced from partial melting of metasomatized mantle nodules. Results of such experiments, combined with major-element, trace-element and isotope geochemistry, have changed many of the models based on a pyrolitetype mantle source proposed in the late 1960s and early 1970s (e.g. O'Hara 1968; Green 1971) for basalt genesis, and particularly for the origin of alkali basalt magmas. Many experiments have also been carried out on the structures of silicate melts, many of alkaline affinities, which have important implications on the physical properties of magmas. Although these studies are beyond the scope of this paper, they are now being incorporated into petrology texts (e.g. Barker 1983). Other important advances in experimental petrology pertinent to alkaline rocks include studies on liquid immiscibility, the role of halogens and traceelement partitioning between crystals and melts. This review concentrates on experimental studies pertinent to the genesis of alkaline basalts and mafic rocks of extreme alkaline affinities. The implications of these studies on the mantle source regions of alkaline magmas are emphasized. Experiments at lower pressures pertinent to felsic alkaline magmatism are briefly considered.
Alkali mafic and ultramafic magmas--background studies Synthetic systems at low pressures Alkali basalts, in which potassium is a minor element, can be represented in the expanded basalt system (CaSiO3 (Wo)-NaA1SiO4 (Ne)Mg2SiO4 (Fo)-SiO2) by recalculating the rock composition into the CIPW normative minerals representing this system. In this system (Fig. 1) mildly alkaline basalts and associated rocks plot within the D i - N e - F o - A b volume and are separated from the more alkali-rich nephelinites and melilitites, plotting within the A k - D i - F o - N e volume, by the F o - D i - N e pseudoternary join. Experiments at low pressure within these two sub-systems of the expanded basalt tetrahedron and in systems with additional components such as CaA12Si20 8 (An) resulted in discrepancies in the paragenesis of the minerals, based on the interpretations of the experimental results and on textural observations on mildly alkali-enriched to highly alkali-enriched rocks. Part of this problem was due to treating the expanded basalt system on flow sheets as a quaternary system
Lo
Ne~
Fo
Mol per cent
FIG. 1. The expanded basalt system. Abbreviations for the CIPW normative minerals are as follows: La, larnite; Ne, nepheline; Fo, forsterite; Qz, quartz; Ab, Albite; En, enstatite; Di, diopside ; Ak, akermanite; Wo, wollastonite; Sm, soda-melilite; Mo, monticellite; Mer, merwinite; Ra, rankinite. Shaded joins represent planes of critical SiO2 undersaturation. when it is at least quinary (Schairer et al. 1969; Schairer & Yoder 1970). For similar reasons, the CaO-MgO-AI203-SiO 2 system (O'Hara & Biggar 1969; Schairer et al. 1969; Schairer & Yoder 1970) cannot be used to model alkaline mafic magmas. The obvious chemical deficiencies of the CaO-MgO-A1203-SiO2 system relative to the compositions of nephelinites and melilitites, and the now accepted hypothesis that nephelinites and melilitites form from the mantle-derived magmas at considerable depth make experiments at 1 atm in either the expanded basalt tetrahedron or the CaO-MgO-AI203-SiO 2 system quite inadequate for investigating the genesis of these rocks ( c f Thompson 1972). Although low-pressure experiments are largely inappropriate in explaining the genesis of alkaline magmas, some features such as the incompatibility of melilite and plagioclase in the same lava can be accounted for by using the schematic representation of liquidus volumes in the WoD i - A b - N e sub-system of the expanded basalt tetrahedron (Fig. 2). Data for this diagram are given by Yoder (1979). Along the Di-Ne join in this system diopside and nepheline do not coexist as liquidus phases but are separated by a field of
Genesis o f alkaline magmas olivine (Bowen 1922: Schairer et al. 1962; Yoder & Kushiro 1972). At sub-liquidus temperatures melilites with compositions comparable with those of melilite in melilitites and melilite nephelinites (E1-Goresy & Yoder 1974) occur. Extension of the primary olivine field into the D i - N e - W o - A b volume (Fig. 2) separates the primary fields of melilite and plagioclase and accounts for the incompatibilityof these minerals in nature. The stability of melilite molecules in various sub-systems of the expanded basalt tetrahedron and in other systems in which Ne and Di are components has been used by Yoder (1979) to explain the genesis of melilite-bearing rocks and associated lamprophyres. The limitations of the expanded basalt and CaO-MgO--A1203-SiO2 systems to explain the genesis of Na-rich mafic magmas become more pronounced for K-rich varieties of these magmas. Although not as common, K-rich mafic and ultramafic lavas may be more important than Na-rich varieties in holding the key to heterogeneities in the mantle source regions. Yoder (1975) has modelled these K-rich magmas using the system Ca2SiO 4 (La)-KA1SiO4 (Ks)Mg2SiO4 (Fo)-SiO2_H20. Figure 3(a) shows the relationship of leucite and akermanite to other major constituents of these rocks, whereas in Fig. 3(b) the system with H20 as an additional
31 Ks
o
LO
Met
Mo
Fo
Mol per cent
Ks
(b) Ne
Sm
Lo-
, Pl ~'"
wo wo
Ab
D. FIG. 2. The CaSiO3 (Wo)-CaMgSi206 (Di)NaA1Si308 (Ab)-NaA1SiO4(Ne)system. Abbreviations are as in Fig. 1 and as follows: Mel, melilite; O1, olivine; Pl, plagioclase; SS, solid solution. (After Yoder 1979.)
Mb
Mo Mol per cent
Fo
FIG. 3. The system Ca2SiO4 (La)-KA1SiO4 (Ks)MgzSiO4(Fo)-SiO2(Qz) (a) without H20 and (b) with H20. Abbreviations are as in Fig. 1 and as follows: Lc, leucite; Sa, sanidine; Phlog, phlogopite. (After Yoder 1979.) component has phlogopite, a common constituent of K-rich mafic lavas, as an additional phase. Many joins in this system have been investigated up to high pressures and, despite the obvious lack of Na20, the results provide some explanations for the genesis of K-rich alkaline mafic rocks. Using this system, Yoder (1975) related ugandite (olivine melilite leucitite) to mafurite (olivine kalsilite pyroxenite) by the reaction
A. D. E d g a r
32
2[Ca2MgSi207] (Ak) 4- 3[KA1Si206] (Lc) 4- (x 4-1)[MgzSiO4] (Fo) ugandite
4[CaMgSi206] (Di) + 3[KA1SiO4] (Ks) + x[MgzSiO4] (Fo) mafurite
As shown in Fig. 3(a)), the leucite + akermanite assemblage (bold lines) reacts to give the kalsilite + diopside assemblage (broken line). Both ugandite and mafurite contain olivine. If the reaction based on relationship in Fig. 3(a) is correct the simplified ugandite assemblage must be richer in olivine than the mafurite assemblage as indicated by the unknown value x. Highpressure experiments on these rocks (Edgar et al. 1980; Arima & Edgar 1983a) suggest that these magmas are not interrelated but may be derived independently from different metasomatized mantle sources. Using the relationship in Fig. 3(b) and adding COz, Yoder (1975) indicated a possible relationship between mafurite magmas and kimberlite magmas by the reaction CaMgSi206 (Di) + 3[KAISiO4] (Ks) 4- (x + 4) [Mg2SiO4] (Fo) 4- CO2 4- 3H20 rnafurite
3[KMg3A1Si3Olo(OH)2] (Phlog) + CaCO3(Ct) + x[Mg2SiO4] (Fo) kimberhte
In this reaction x represents the amount of olivine required to balance the equation and indicates that a kimberlitic groundmass will have less olivine than a simplified mafurite. On the basis of these equations, Yoder (1975) proposed that melilite-bearing magmas should be transformed to kimberlitic magmas. Melilite-bearing magma might be produced at low pressures under dry conditions (Fig. 3(a)) caused by volatile loss during ascent or by low-pressure crystallization. With the addition of H20 and CO2, such a melilite magma could be transformed to kimberlite. Although the kimberlite-melilitite association is not uncommon, kimberlites are only rarely associated with highly K-rich mafic volcanics such as mafurite. However, kimberlites are associated with lamproites sensu stricto (Bergman 1987) but probably not by the mechanism proposed by Yoder (1975). Many mafurites contain both kalsilite and primary phlogopite (Holmes 1950; Edgar 1979) which cannot be explained by this mechanism.
Rock systems--the pyrolite model As an alternative to the chemically simplified synthetic systems, experiments on rocks of alkali
olivine basalt, basanite, nephelinite and melilitite compositions (el Green & Ringwood 1967; Bultitude & Green 1968, 1971; Green 1969, 1970a, b, 1973a; Green & Hibberson 1970; Brey & Green 1977) under conditions ranging from crustal to upper-mantle depths clearly indicated the importance of mantle processes in the genesis of alkaline mafic magmas. Based on these studies and assuming a model pyrolite mantle source for these magmas (Ringwood 1975), Green (1970a, b) developed a petrogenetic grid for basaltic magmas (Fig. 4). For a model pyrolite with 0.1 H20 and no other volatile component, Fig. 4 shows that progressively more alkali-rich mafic to ultramafic magmas, ranging from alkali olivine basalts to olivine melilitites, might be produced at progressively greater depths with decreasing degrees of partial melting and increasing amounts of H20 in the melt. The grid shown in Fig. 4 can be used to divide fractionation trends, based on the liquidus and near-liquidus minerals, according to depth of generation of the magma. Green & Ringwood (1976), Green & Hibberson (1970) and Merrill & Irving (1977) proposed that mildly alkali basalts could be formed at approximately 45-60 km by fractionation of aluminous enstatite 4- aluminous subcalcic clinopyroxene from an olivine basalt magma to yield alkali olivine basalt magma with about 25~ normative olivine and 1 ~ - 2 ~ normative nepheline (Fig. 4). Green & Ringwood (1976) extended this fractionation scheme by suggesting that further fractionation of aluminous clinopyroxene_ aluminous enstatite from this magma can produce an olivine basanite with about 5~ normative nepheline. The absence of orthopyroxene as a liquidus phase at any pressure under dry conditions in alkali olivine basalt compositions (Ito & Kennedy 1968; Bultitude & Green 1971; Arculus 1975) makes this scheme highly unlikely. At greater depths (60-120 km) no fractionation trends linking the more extreme alkali magmas of nephelinitic and basanitic affinities could be found on the basis of results of experiments in the absence of H20. However, hydrous experiments on nephelinitic compositions with an estimated 5-7~ H20 in the melt (Bultitude & Green 1968, 1971; Green 1969, 1970a, b, 1973b; Green & Hibberson 1970) indicated that between 18 and 27 kb the primary field of orthopyroxene crystallization was extended, in contrast with anhydrous experiments in which liquidus phases were olivine, clinopyroxene and garnet. These experiments suggested a fractionation scheme whereby less extreme alkali basaltic magmas could be linked to more alkali-rich compositions, provided that H20 was available in their source regions. This scheme has
Genes& of alkaline magmas
33
% MELTING 5
I0
15
20
25
Qz.' Tholeiite Tholeiite/ ~ -----.__ 0-5 / i '
olivine
/,
,
~ ~
~
C\~e'/Oliv in 9 Th ole iit 9
\
5-10
?\~High AI203
.
20
Or tho'~vr ~
9OXene~
High AI2I~3
Hawc(iite"/"Alkaline Olivine
35
Tholeiite
5-15
High-AI203 e~ ~
30
Olivine
Olivine
Olivine Tholeiite
Tholeiite
~" 40
Olivine-rich ~-/~<~Y
.13
So.sanite
/
/
Alkaline Olivine Bo.satt 20 - 30
~s-25 / / ~O~ine Olivine-rich NL:phelinite Bosonite 2o-3o
@ L
D20
//2o-2s
U~ @
Otiv/ne Basalt 2~-25
5O
Otivine Thoteiite 20-25
Alkc{t'ine Picrite ~ 3o-35
Thoteiitic Picrite -3o
L
n
-~344v].Qe ~ M e t i t i t e ~ -~ Olivine Nepheti~te i . . o,~ ~30-35 / 3O Nephetln~te %, 25-35
I O,iv,ne
J~lMel.ilitit e ~ /
I~
.~35-~0
10 4
2
-O'[ivine N~.|i rite Nephelin'i, te ~35-40\
i
Picritic Basonite 30-35
Atkotine Picrite 30-35
E
60 v,
Thoteiitic Picrite -35
t70 .~ n 80 ~J
90 100
110 120
0%
0',3
%H20 in melt for simple m e l t i n g of p y r o l i t e c o n t a i n i n g
0.1~ H20
FIG 4. Green's (1970a, b) petrogenetic grid based on a model pyrolite mantle containing 0.1% H20 (the numbers below each rock represent normative olivine): - - - , stability limits of minerals indicated.
now been abandoned for highly alkaline magmas in favour of their being primary melts. The petrogenetic scheme for alkali basalt genesis outlined above differs considerably from that of O'Hara (1968) who suggested that alkaline magmas were generated at a greater depth (at least 100 km) by higher degrees of partial melting (up to 30%) and by a more complex fractionation scheme in which as much as 40% garnet+ clinopyroxene and 40% olivine are fractionated from a tholeiitic picrite at various stages of ascent. Ringwood (1975, pp. 166-71) gave a (biased?) review of the differences between O'Hara's scheme and that of Ringwood, Green and coworkers. His principal criticisms of O'Hara's hypotheses are the oversimplification of O'Hara's four-component system relative to natural conditions, the inconsistency in the Mg/ (Mg+Fe), Ni, Cr and incompatible elements required by the large amount of fractionation involved in O'Hara's hypothesis and those observed in alkaline olivine basalts, and O'Hara's lack of quantitative assessment of his petrogenetic scheme. The genesis of alkaline mafic-ultramafic magmas based on Green's petrogenetic grid (Fig. 4)
presents a number of difficulties. None of the compositions represents rocks with K20 > Na20 and those with high Na20 contents, e.g. olivine melilite nephelinites, require very small degrees of partial melting. The genesis of these highly alkaline SiO2-undersaturated magmas is much more likely to be the result of direct partial melting than of fractionation processes as these magmas have a primitive chemistry which precludes even moderate fractionation and they are in equilibrium with a pyrolitic mineral assemblage (cf Ringwood 1975). Because these magmas have much greater abundances of incompatible, major and trace elements than are present in less alkaline types, very low degrees of partial melting seem to be required (cf Kay & Gast 1973). Nevertheless the problem of segregation of these very small amounts of partial melt and their subsequent ascent from 100 km or greater is formidable (see Yoder 1976). This difficulty is even greater if, as Ringwood (1975, p. 153) suggests, less than 1% partial melting of a hydrous source may be required for the generation of olivine nephelinite, olivine melilitite and kimberlite magmas. More recently McKenzie (1984) has developed models based on equations involving
34
A. D. Edgar
the movement of melt and matrix in partialmelts. Based on his calculations, very low degrees of partial melting (less than 3~) at depths commonly accepted for alkali basalts can readily allow melts to ascend to the surface. For highly alkaline magmas derived by direct partial melting of a mantle source the grid shown in Fig. 4 may not be as valuable as for less alkaline types. The grid is based on a pyrolite (lherzolite) source in which H20 is the only volatile component. Recently the concept that alkaline magmas are partial melts of mantle sources metasomatized to produce regions of quite different compositions from that of pyrolite and the importance of volatiles other than H20 in such source regions has been widely discussed ( c f Bailey 1982). Much of the remainder of this review is devoted to experiments pertinent to these concepts.
Recent experimental studies on alkaline mafic-ultramafic magmas Alkaline olivine basalts
The Basaltic Volcanism Study Project (1981, chapter 3) defines alkali olivine basalts as those with not more than 5~ normative nepheline and lists (Appendix 3.1, pp. 614-15) 11 experimental studies carried out on natural or synthetic samples up to 1978. The pressures varied from 1 atm to 33 kb under dry to H20-excess conditions, many with controlled or reasonably estimated oxygen fugacity (fO2) and temperatures from nearliquidus to solidus. Experiments performed at low pressures are not pertinent to the problems of the generation of alkali magmas under mantle conditions, but are important in establishing the sequence of mineral crystallization particularly near the liquidus and possible parental magma compositions (see Thompson 1973). These experiments have also been important in determining the effects offO2 on the stabilities of minerals in alkali olivine basalts (see Helz 1973, 1976). Unfortunately, fewer data are available for the effects offO2 on alkali magmas than are available for other basalts. Thompson (1974) studied a primitive alkali olivine basalt composition from Skye, Scotland (Mg number, 65; ne, 2.84~), between 8 and 31 kb at fOz>iron-wustite (IW) buffer under dry conditions. Because of its high Mg value it is unlikely that this alkali olivine basalt could be derived from a tholeiitic parental magma by fractionation; it is more likely to represent a pristine melt as suggested by Green (1970a) for similar alkali olivine basalts. Near-liquidus rela-
tionships in these experiments are very complicated at about 17kb (see Thompson 1974) suggesting that liquids must be saturated in olivine + clinopyroxene_ pigeonite. On the basis of these results Thompson proposed that this alkali olivine basalt was either a direct partial melt of an Fe-rich lherzolite at about 50km depth or a partial melt of a more Mg-rich lherzolite from deeper sources which underwent some olivine fractionation. Frey et al. (1978) estimated that 11~-15~ partial melting of garnet-free lherzolite can produce an alkali olivine basalt magma. Experiments without volatiles give no information on the roles of amphiboles and micas on the genesis of alkali olivine basalt. Allen et al. (1975) studied a Hualalai, Hawaii, alkali olivine basalt under pressure conditions similar to those used by Thompson (1974) but with excess H20 and f O 2 approximating those of the hematitemagnetite (HM), nickel-nickel oxide (NNO) and magnetite-wustite (MW) buffers. This composition has a comparable Mg number (67) but a lower ne (0.20~) than the sample used by Thompson (1974). Their experiments are shown in Fig. 5 which indicates that under HzO-excess conditions amphibole + olivine + clinopyroxene are near-liquidus phases around 13 kb, olivine is a liquidus phase up to 25 kb and garnet is present between 18 and 23 kb (Fig. 5(a)). The main differences between the results of the dry experiments (Thompson 1974) and the H20-excess conditions used by Allen et al. (1975) are the presence of amphibole and absence of pigeonite in the HzO-excess experiments (Fig. 5(a)). Allen et al. (1975) found that a n f O 2 corresponding to the N NO buffer produced amphibole at higher temperatures relative to the HM and MW buffer conditions (Fig. 5(b)). The absence oforthopyroxene in the experiments reported by Allen et al. places some constraints on the potential of orthopyroxene fractionation in the genesis of these basalts under H20-saturated conditions. The experiments support the concept that andesitic magmas are derived from basaltic magmas by amphibole-liquid fractionation processes as amphibole is stable at temperatures greater than the H20-saturated liquidus of andesite. Thus fractionation of amphibole from basaltic liquids during partial melting or fractional crystallization processes would lead to SiO 2 enrichment. Melting at depths corresponding to pressures greater than 23 kb, where amphibole is no longer stable (Fig. 5), would enrich liquids in Na20 and H20. Figure 6 shows the results of dry experiments performed by Takahashi (1980) on a primitive alkali olivine basalt from Oki-Dogo Island, Japan (Mg number, 63; ne, 3.40~), under low-fO2
Genesis o f a l k a l i n e m a g m a s
[--,
| I"
O I + C p x + G a + Ru + O p
25 ~-rA'mphibo'0"~ "
~
9 f, . / LOlivine Out] "4="
35
, ~
,
,
Y+
x 20
13_
10
,, ,,'
800
900
,
+ Cpx
o
__ J
1000 1100 800
900
1000
1100
Temperature, ~ (b)
(o)
FIG. 5. Phase relations of an alkali olivine basalt from Hawaii : (a) under H20-saturated conditions; (b) effects of variousfO2 on amphibole stability. Abbreviations are as in Fig. 2 and as follows : Cpx, clinopyroxene; Am, amphibole; Ga, garnet; Ru, rutile; Op, opaque; N-NO, nickel-nickeloxide buffer assemblage; M-H, magnetite-hematite buffer assemblage ; M-W, magnetite-wustite buffer assemblage. Liquid(s) coexist with all assemblages. (Modified from Allen et al. 1975.) conditions within the wustite stability field. This basalt, which contains spinel lherzolite nodules, gives results similar to those obtained by Thompson (1974) with orthopyroxene+clinopyroxene on the liquidus at 14-15kb and olivine+orthopyroxene (pigeonite) + clinopyroxene + liquid between 13.5 and 14 kb. This suggests that the
i
I
I
magma could have coexisted with the latter assemblage at depths corresponding to this pressure. This assemblage, along with spinel, also occurs at an isobaric invariant point within the D i - A n - F o tetrahedron between 7 and 20 kb (Presnall et al. 1978, 1979). Kushiro (1973), Presnall et al. (1978, 1979) and Mysen & Kushiro
i
i
I
i
1500
ou 1 4 0 0
Ga§
.
OL+Opx
~
+ Cpx+L t
d
OL+Cpx+L
\Opx+Cpx+L
/
_ /
r 1300 n
E
r
1200
1100
j
~
OL§
0
n 5
n 10
n 15 Pressure,
n 20 kb
~ 25
I 30
35
FIG. 6. Phase relations in an alkali olivine basalt from Japan (abbreviations as in Figs 2 and 5). (Modified from Takahashi 1980.)
36
A. D. E d g a r
(1977) suggest that liquids in more complex natural systems become nepheline normative between 10 and 20 kb. All these results suggest that the alkali olivine basalt lay on the SiOz-poor side of the A b - F o - D i plane (Fig. 1) at 14 kb. Based on these experiments Takahashi (1980) proposed that alkali olivine basalt magmas can be partial melts oflherzolite at depths corresponding to 14 kb provided that H20 and CO2 are unimportant in the melting relations, or that they were generated at greater depths but remained in equilibrium with lherzolite in the upper mantle. Alkali olivine basalts and systems of corresponding composition have also been investigated experimentally at crustal pressures under a variety of conditions. Duke (1974) studied an alkali olivine basalt (Mg number, 57.3; he, 1.63~) from the Azores at 1 atm and withfO2 determined using CO2 and C O + C O 2 mixtures. He found that oxidation increased the stability of spinel and clinopyroxene but decreased that of olivine and plagioclase, promoted crystallization of magnesioferrite-rich spinel and acmitic pyroxene, and caused nephetine-normative liquids to become SiO2 saturated or oversaturated. Using the system CaO-MgO-AlzO3-SiO2-Na20-H2 O at 5 kb to model various basalt types, Cawthorn (1976) found that nepheline-normative liquids could crystallize amphibole___ olivine + Ca-rich clinopyroxene to produce quartz-normative residual liquids of andesitic composition.
Green (1973a) investigated a lherzolite-bearing olivine basanite composition from Mount Laura, Victoria, with 10% added olivine (Mg number, 70; ne, 10.5%) between 5 and 36 kb and over a range of H20-undersaturated liquid conditions. As shown in Fig. 7, liquids from this composition are saturated with olivine + orthopyroxene + clinopyroxene+garnet only under very limited conditions corresponding to 25-30kb, 12001300~ and 2 - 7 ~ H20. Based on these experiments, Green (1973a) concluded that basanites could be generated from a lherzolite mantle with 0.2-0;4% H20 by 6 ~ partial melting at a depth of around 100 km. He suggested that the excess olivine fractionated from the liquid prior to the incorporation of lherzolitic mantle xenoliths, presumably during ascent. Using a kaersutite eclogite nodule from Kakanui, Hawaii, of composition corresponding to olivine basanite (Mg number, 63; ne, 6.4~), Merrill & Wyllie (1975) showed that orthopyroxene+ clinopyroxene + garnet coexisted at 20 kb under HzO-excess conditions, and at greater than 30 kb with progressively lower H=O contents. With the addition of normative olivine to their bulk composition, they found a more extensive field of olivine + orthopyroxene + clinopyroxene+ garnet in equilibrium with liquid than that reported by Green (1973a) and they also found that kaersutite was stable on the H20-undersaturated liquidus between 12 and 21 kb for liquids 40
Basanites and nephelinites Basic rocks more alkaline than alkali olivine basalts are generally defined as those with more than 5% normative nepheline. They may also contain normative leucite and a few have normative kalsilite. These magmas have primitive Mg numbers and other chemical features indicative of an origin by direct partial melting from a mantle source. Some are extremely rich in K and other incompatible and large-ion-lithophile (LIL) elements. Because of the many problems involved in the genesis of these rocks, including the fact that many cannot apparently be derived by partial melting of a (normal) lherzolitic mantle source, a large number of experiments on these magma compositions have been carried out in recent years. Those experiments particularly applicable to the nature of the source regions are concentrated on in this section. Over a decade ago it was recognized that highly alkaline magmas could not be derived by partial melting of volatilefree mantle peridotite but that some of these magmas could represent partial melts of an HzOand CO2-bearing peridotitic mantle (see Basaltic Volcanism Study Project 1981, pp. 529-30).
Ol
/
,/ go+'~px,/ I
I
~1 t 'IX
l +wx/ l + opx /
I go
~',,~ + ol l
I
go
/o,/// /op,+.l
:~ 20
!
,/ I g ~
/
/
/
/
/
/
J
/
/ /
l,,
II
I/ II
I jq.~
/
/
~'~
~ ~
Olivine/
/
,~
\1
!9
/ O........ livine
\//
o
L 1100
,
J
"-.I 1300
/ ,
. 1500
TEMPERATURE, ~ FIG. 7. Phase relations in an olivine basanite + 10~ olivine under dry to HzO-saturated conditions: - - -, percentage HzO added. Abbreviations are as given in Figs 2 and 5. (After Green 1973b).
Genesis of alkaline magmas containing at least 2-3% H20. From this Merrill & Wyllie (1975) inferred that basanites with kaersutite megacrysts may contain 2-3% H20 at depths corresponding to these pressures. Arculus (1975) investigated basanites from Mount Shadwell, Victoria (Mg number 68; ne, 11.92%), and from Grenada, Lesser Antilles (Mg number, 74; ne, 7.84%), under dry conditions between 10 and 35 kb. Owing to the absence of liquidus assemblages compatible with a dry peridotitic source (notably orthopyroxene), he concluded that basanites could not be derived from such a source unless H20 and CO2 were present; the former stabilizes olivine relative to pyroxene, and the latter stabilizes orthopyroxene relative to olivine and clinopyroxene (Eggler 1974). Arculus (1975) also concluded that oxides of polyvalent cations such as TiO2 and P205, commonly abundant in basanites, might lower the olivine stability field by 2-3 kb. Thus the absence of volatiles and such polyvalent cations in previously proposed models (e.g. O'Hara & Yoder 1967; Kushiro & Yoder 1974) suggests that such models might be inapplicable for natural systems of basanitic affinities. Evolved alkali basalts Evolved alkali basalts with lower Mg numbers and less normative nepheline than the primitive basanites and nephelinites may be produced from these more primitive magmas by various fractionation mechanisms. Bultitude & Green (1971) suggested that many basanites and nephelinites might be products of olivine fractionation of lowSiO2 picritic magmas. Irving & Green (1972) demonstrated that kaersutite was a near-liquidus mineral at 14 kb from a nepheline mugearite composition with about 3.5% H20 which contained lherzolite nodules. These studies indicate that fractionation of olivine+kaersutite from basanitic magma could produce nepheline hawaiites and nepheline mugearites at depths around 40-45 km. Although experimental data are lacking, it seems likely that further fractionation of clinopyroxene+biotite could produce nepheline benmoreites and phonolites (cf Irving & Price 1981). Olivine-rich nephelinites and melilitites In experiments with olivine nephelinites and comparable compositions up to 36 kb and under dry to H20-excess conditions (Bultitude & Green 1968, 1971; Allen et al. 1975; Merrill & Wyllie 1975) it was shown that such magmas could not be in equilibrium with a lherzolitic mantle source. In all cases clinopyroxene rather than orthopyrox-
37
ene appeared on the liquidus. Bultitude & Green (1971) indicated the improbability that tholeiitic picrite or alkaline picritic magmas are related to olivine and olivine melilite nephelinites by fractionation processes, thus implying that these magmas are primary. Brey & Green (1975, 1976, 1977) and Brey (1976, 1978) conducted a very detailed study of olivine melilitite compositions to investigate the effects of H20:CO2 ratio a n d f O 2 up to 40 kb on the genesis of these magmas. These authors used an olivine melilitite from Laughing Jack Marsh, Tasmania (Mg number, 74; ne, 17.2%; lc, 6.6%), and a synthetic Ca- and Mg-rich olivine melilitite (Mg number, 75; ks, 4.4%). In addition to analyses of the solid phases, the H20 and CO2 contents of the liquids were also determined. Using the Laughing Jack Marsh sample Brey & Green (1975) found no combination of pressure, temperature and percentage H20 conditions under which this olivine melilitite could be in equilibrium with a pyrolite mantle assemblage. In the presence of H20 the solubility of CO2 is high in the olivine melilitite liquid, and it suppresses olivine and clinopyroxene and promotes orthopyroxene and garnet as liquidus or near-liquidus phases. These results, using an external HM buffer assemblage, are summarized in the 30 kb section shown in Fig. 8. With increasing COz (decreasing HzO/(H20 + CO2)) the liquidus temperature rises almost 200~ between 0.5 H20/ !
!
! 1400
? e
\
13oo
0
ox
§
0
9
~ 9
0.'25 0.~5
- 1200
0
0175
C02
38wt7.
T*C
H20 H20 * C02 mole fraction in c h a r g e
1100 H20 20wt7.
Fla. 8. Phase relations in an olivine melilitite showing variations in the liquidus phases at 30 kb and at f O 2 = HM buffer with varying HzO/(HzO + C02). Liquidus and near-liquidus phases are indicated above the liquidus: O ; garnet; +, orthopyroxene; 0, clinopyroxene; o, all liquid. (Modified from Brey & Green 1975.)
38
A.D.
(H20 + CO2) and CO2-only experiments. Based on these results, Brey & Green (1975) inferred that olivine melilitite magmas could be derived at 30kb and 1150-1200~ by less than 5~ equilibrium partial melting of pyrolite containing 0.5-1.0 H20/(H:O+CO2). They suggested low fO2 values (less than the N N O buffer) as most appropriate. Brey & Green (1977) refined the results of their earlier study by systematic investigation of the liquidus and near-liquidus relations with varying H 2 0 / ( H 2 0 + C O 2 ) and f O 2 conditions. The resuits shown in Fig. 9 represent near-liquidus phases at 30 kb andfO2 = HM buffer. Of particular importance is the narrow field of clinopyroxene separating the garnet+orthopyroxene+ clinopyroxene field from the olivine + clinopyroxene field for varying CO2 and H20 contents. Experiments at 27 kb with a composition containing 0.92 H20/(H20 + CO2) closely approached a near-liquidus olivine + orthopyroxene + clinopyroxene+garnet assemblage, compatible with equilibration of the olivine melilitite magma with a pyrolite source. Brey & Green (1977) proposed that the additon of 5% olivine (implying fractionation of a similar amount of olivine from the primary magma) would produce this four-phase assemblage. They also determined that liquidus and near-liquidus assemblages were not appreciably different at more realistic mantlefO2 values corresponding to those less than the N N O buffer assemblage. Based on these results, Brey & Green (1977) proposed that olivine melilitite magmas could be derived at depths corresponding to 27 kb at 1160~ by partial melting of a pvrolite mantle source containing both H20 and CO2 equivalent to 6 - 7 ~ CO2 and 7-8~o H20 dissolved in the melt. Brey (1978) proposed that olivine melilitites which were more SiO2-undersaturated than the ,
/ O~
4Q
O~. /"'f I"j1~
20
Laughing Jack Marsh sample might be derived at greater depths and with higher CO2 in the source region. Experiments at 30-35 kb with higher CO2 contents using a very SiO2-undersaturated synthetic olivine melilitite composition produced olivine+clinopyroxene as liquidus phases. At 30 kb this assemblage is clearly separated from the garnet + orthopyroxene field by an orthopyroxene field. He found that garnet stability increased at lower H20/(H20+CO2) and suggested that highly SiO2-undersaturated olivine melilitites could be derived at deeper levels (corresponding to 35 kb) by partial melting of a more CO2-rich pyrolite source. Although Brey & Green (1975, 1977) and Brey (1978) demonstrated the role of CO2 in addition to H20 for the genesis of olivine melilitite magmas, Eggler (1974, 1978), Huang & Wyllie (1974), Wyllie & Huang (1975, 1976) and Wendlandt & Eggler (1980a, b) proposed that CO2 was an essential component in the source regions for mafic SiO2-undersaturated magmas. Reviews of the experimental evidence for the role of CO2 are given by Eggler & Holloway (1977) and Wyllie (1977, 1979). Eggler (1978) studied joins in the system Na20-CaO-A1203-MgO-SiO2CO2 up to 30 kb and showed that orthopyroxene stability increased relative to that of olivine, thus producing liquids containing less SiO 2 than those formed in the absence of CO2. He suggested that nephelinites and melilite nephelinites could be partial melts of a peridotitic source at 50-90 km provided that CO2 was the predominant volatile component. If the total volatile contents (H20 + CO2) of the mantle source region are greater than 0.5 wt.% all variations between H20-saturated and CO2-saturated melts can occur, whereas with low total vapour concentrations in the source regions the compositions of the vapour a r e restricted and hence solidus temperatures and near-solidus melt compositions are also restricted. As an example of the role the amount of H20 and CO2 plays in the composition of melts. Eggler (1978) proposed that partial melts of an amphibole-peridotite in the presence of both H20 and CO2 (not exceeding 0.37 wt.%) will produce a nephelinitic melt composition at 15 kb.
/"
~i~ 20
Olivine Melilitite
,,
Edgar
Kimberlites "
ga + opx
i 40
ga + opx + cpx C pX ol § c p x ol 60 H20
FIG. 9. Phase relations in the system olivine melilititeH20-CO 2 at 30 kb andfO 2= HM buffer. Abbreviations are as in Fig. 5. (Modified from Brey & Green 1977).
The role of CO2 in the genesis of magmas is particularly relevant to kimberlites and carbonatites. Eggler & Wendlandt (1979) determined phase relations in a kimberlite with variable H20 and CO2 contents at 30 and 55 kb. At the higher pressure they found that kimberlite-like liquids coexisted with a lherzolite assemblage in experiments with 5% H20 and 5% CO2, but kimberlites
Genesis of alkaline magmas
could not be derived from a lherzolite source at the lower pressure. At 30 kb phlogopite and dolomite were sub-solidus phases in experiments with HzO-rich buffered vapour, whereas magnesite solid solution was a sub-solidus mineral at 55 kb. On the basis of phase relationships in the peridotite-COz-H20 system, particularly those determined by Eggler (1978), Wyllie (1977, 1978, 1979) and Ellis & Wyllie (1980), Wyllie (1980) suggested that kimberlite originated by partial melting of peridotitic mantle near 260 km in which solidus temperatures were reduced by upward migration of volatiles. This melting causes a density inversion and results in a diapiric uprise under adiabatic conditions with subsequent crystallization of the partially melted diapiric material beginning at depths of 10080 kin. Leucitites and olivine leucitites With the exception of kimberlites, all the alkali magmas so far described have N a z O > K 2 0 . Mafic-ultramafic rocks with K z O > N a 2 0 are essentially leucitites and olivine leucitites, but with a plethora of varietal names. Most of these rocks fall within the broad category of lamproites (cf. Bergman 1987). They commonly have normative leucite but may be SiO2-saturated to oversaturated and contain normative quartz. Problems involved in the genesis of these often highly LIL-element-rich rocks have been summarized by Bell & Powell (1969). Edgar et al. (1976) have pointed out some of the improbabilities of the older theories given by Bell & Powell. These theories fall into three categories. Firstly, differentiation of basaltic or peridotitic magma with fractionation of olivine (Wade & Prider 1940), eclogite and olivine (Holmes 1932), or eclogite (O'Hara & Yoder 1967). Secondly, assimilation of magmas of such diverse compositions as granite, carbonatite, nephelinite and monchiquite by a wide variety of crustal materials (cf Daly 1910; Bowen 1928; Williams 1936; Holmes & Harwood 1937; Larsen 1940; Holmes 1950, 1965; Turner & Verhoogen 1960). And thirdly, partial melting of biotites and hornblendes (Waters 1955) and kimberlite (Harris, quoted by Bell & Powell 1969). Any hypothesis involving basaltic magma seems unlikely, as in most occurrences of rocks of the leucititic kindred basalts are sparse or absent. Even if basaltic material was involved, very extreme degrees of crustal contamination or assimilation would be required to account for their high incompatibleelement content and to maintain their high MgO and CaO values. The theories involving partial melting of some types of mantle material are now widely accepted by most workers. Recent wide-
39
spread interest in these rocks has been prompted by the possibility of greater heterogeneity in the mantle source regions than previously proposed and by the current popularity of the process of mantle metasomatism (cf Bailey et al. 1980; Bailey 1982). In general, the similarities in high Mg number, low SiO2 content and high volatile contents of olivine leucitites, olivine basanites, olivine melilitites, olivine nephelinites and kimberlites suggest that these magmas are primary rather than derivative and hence can only be evolved from a mantle source containing H20 and CO2. Cundari & O'Hara (1976) conducted dry experiments on a mafic leucitite from New South Wales, Australia (Mg number, 79; lc, 0H; ne, 9.08%). They found clinopyroxene as the only liquidus phase between 20 and 40 kb and olivine at lower pressures. Absence of the assemblage olivine + orthopyroxene+ clinopyroxene _+garnet_+ spinel as near-liquidus assemblages precludes this magma's being derived by partial melting of dry lherzolite. Cundari & O'Hara (1976) concluded that this rock may have formed by partial melting of an olivine pyroxenite source containing essential phlogopite. Absence of normative leucite or quartz in this composition sets it apart from other compositions. Thompson (1977) conducted experiments under similar conditions on an apparently less-primitive leucitite from the Alban Hills, Italy (Mg number, 63; lc, 32.5%; ne 13.6%), with the same results. High-pressure experiments using H20 and H20 + CO2 have been carried out for compositions ranging from potassic intermediate to ultramafic from SW Uganda (Edgar et al. 1976, 1980; Ryabchikov & Green 1978; Arima & Edgar 1983a), from the Leucite Hills, Wyoming (Barton & Hamilton 1979, 1982), and from the W Kimberley area, W Australia (Arima & Edgar 1983b). All these compositions have primitive Mg numbers (70-78) but widely varying degrees of K enrichment (K/(K + Na) mol = 0.41-0.86). In Fig. 10 these rocks are plotted on the Mg2SiO4 (Fo)-KA1SiO~ (Ks)-SiO2 system which indicates the variable degrees of SiO2 saturation. The three compositions from SW Uganda (olivine ugandite, katungite and biotite mafurite) and the madupite from the Leucite Hills are all SiO2 undersaturated, whereas the orendite from the same locality and the wolgidite from W Kimberley are SiO2 saturated and slightly oversaturated respectively. The SiO2-undersaturated group shows distinctly different liquidus to near-liquidus phases from the group which is SiO 2 saturated to oversaturated. The primary phase fields at 3 kb PH2o (Luth 1967) and 28 kb Pmo (Green et al. 1984) in this system are also included in Fig. 10. With
40
A . D. E d g a r Oi
Sil wt.% FIG. 10. K-rich mafic-ultramafic rocks (O), for which experimental data are available, plotted as CIPW normative minerals in the system Mg2SiO4-KA1SiO4-SiO2 together with phase relations at PHzo of 3 and 28 kb (Luth 1967; Green et al. 1984)" HI, phlogopite composition. Abbreviations are as in Figs t, 2 and 3" Hy, hypersthene. Ks
Lc
So
increasing pressure the fields of phlogopite and enstatite expand relative to olivine. Thus partial melting of a mantle source with H20 as the only volatile phase results in liquids which may crystallize these minerals. With H20 as the only volatile phase neither orthopyroxene nor garnet crystallize from any of the SiO2-undersaturated compositions. Thus these magmas cannot have equilibrated with a normal lherzolitic mantle source. Figure 11 is a schematic diagram of near-liquidus phase assemblages for these compositions with H20 contents between 3 and 15 wt.~, representing vapourabsent conditions above about 10 kb. Clinopyroxene___phlogopite are the commonest near-liquidus minerals at high pressures and olivine+ clinopyroxene are the commonest at lower pressures. Based on these near-liquidus phase assemblages, possible source regions for these magmas range from a slightly K-enriched wehrlite for the least K-rich olivine ugandite magma to a phlogopite-rich clinopyroxenite for the highly K-rich mafurite magma (Fig. 11). The appearance of
phlogopite as a near-liquidus assemblage is directly correlated with the degree of K enrichment (Edgar & Arima 1983). Figure 12 shows near-liquidus assemblages for the SiO2-saturated (orendite) and SiO2-oversaturated (wolgidite) compositions in which only H20 has been added. Orthopyroxene is a common near-liquidus mineral for these compositions. In the orendite the near-liquidus assemblage of olivine + orthopyroxene + clinopyroxene + garnet + phlogopite (trace) at 30 kb suggests that this magma could be derived from a phlogopitebearing garnet lherzolite containing small amounts of H20. A phlogopite orthopyroxenite and a phlogopite harzburgite seem the likeliest sources for the wolgidite composition at high and low pressures respectively. The nature of the source materials and the depth of generation depend on the amount of H20 in the source. Ryabchikov & Green (1978) studied the effects of CO2 and H20 at 30kb on the mafurite composition used by Edgar et al. (1976). Figure 13 shows that, with total volatiles equivalent to 5
Genesis of alkaline magmas
40-
Madupite OI-ugandite Katungite Mafurite (Barton 8~ (Edgar e/o/. (Arima 8~ (Edgar elo/. Hamilton 1980) 1980) Edgar 1983a) 1976)
41
OI-ugandite (Edgareta~. 1980)
Katungite Mafurite (Arima & (Edgar et o/. Edgar 1983a) 1976) -140
T
T
35
t
30-
s_HlightlyK-enriched wehrlite (ol+cpx)
K-bearing clinopyroxenite (cpx) a_ 2 5 -
I
3%
-
RI---#-~ h g~ phlog_o_pite clinopyroxenite clinopyroxenite (cpx+phl) (cpx+ phi)
p_hloggRitewehrlite (cpx + ol + phi)
phlogopitewehrlite (cpx +ol+phl)
20-
, ]
P_h I._9_og_opit e___: clin.~_opyroxenite (cpx+phl)
t
120
slightly K-enriched wehrlite (ol+ cpx)
(ol only) p_hl og_op_it_eewehrlite (ol+cpx+phl)
I
~
(ol only)
l
5%
4
H20
p_hlogo!it.__eewehrlite (ol.cpxephl)
I
added
l
p_hl og_opit_.ee wehrlile (cpx + ol + phi) 15%
I00
80
I
60 I
FIG. 11. Schematic diagram showing near-liquidus phases for SiO2-undersaturated K-rich magmas in which H20 is the only added volatile. The underlined rock names are approximately equivalent to the near-liquidus assemblages in brackets. wt.% H20 , olivine+clinopyroxene are liquidus phases at CO2/(CO 2 Jr H 2 0 ) mol < 0.50, whereas orthopyroxene stability increases above this value. For CO2/(CO2 + H20) mol between 0 and 0.50 phlogopite is stable but clinopyroxene rapidly becomes unstable at higher ratios. For CO2/(CO2+H20 ) equivalent to 15 wt.% H20 , phlogopite and clinopyroxene stability decrease at low ratios, with orthopyroxene as the liquidus phase over most of the range of CO2/(C0 2 Jr H20) values. These relations and the absence of garnet suggest that a lherzolitic source may be possible for this highly K-rich composition. Figure 14 is a schematic diagram of three SiO 2undersaturated SW Ugandan lavas at CO2/ (CO2+H20) mo1=0.75 (equivalent to 15 wt.% H20 added). Only the low-K ugandite with orthopyroxene + clinopyroxene_+ garnet 4- spinel as near-liquidus phases from 22 to 40 kb can be derived from a lherzolitic source under these conditions. The katungite liquidus relationships suggest a garnet clinopyroxenite or a clinopyroxenite as the likeliest source compositions. Although phlogopite does not occur in these highCO2 experiments, the source regions must be K enriched in order to produce the K enrichment in the magmas provided that these are direct mantle melts. No experiments with high CO2 have been carried out on the SiO2-saturated to SiO 2oversaturated orendite or wolgidite.
Wendlandt & Eggler (1980a, b) modelled the genesis of potassic magmas based on melting relations in the systems KAISiO4-Mg2SiO4SiO2, KA1SiO~-MgO-SiO2-CO2 and KA1SiO4MgO-SiO2-H20-CO2 as simple mantle sources at pressures up to 50 kb. In these systems they were able to determine a large number of reactions producing a wide variety of possible liquids ranging from simplified tholeiites to carbonatites. On the basis of previous studies they suggested that the absence of Ca (as CaMgSi206) and Na (as NaA!SiO4 and NaA1Si3Os) would not appreciably alter their conclusions. The expansion of the primary phase fields of enstatite and sanidine in the KA1SiO~-Mg2SiO,~SiO2 system with increasing pressure is illustrated by the schematic diagrams given by Wendlandt & Eggler (1980a, Fig. 6). This results in liquids becoming progressively enriched in K20 and depleted in SiO2 at pressures ranging from 5 kb to greater than 30 kb. In the KA1SiO4-MgOSIO2-CO2 system, phase relations are similar but occur at lower pressures than in the CO2-free system. The trends of liquids with increasing pressure are schematically shown in Fig. 15 which is based on the variation in the extent of the primary solid phase fields with pressure in the systems KA1SiO4-MgO-SiO2-CO2 and KA1SiO4-Mg2SiO4-SiO 2 (Wendlandt & Eggler 1980a). The approximate pressures at which the
A. D. Edgar
42 Orendite {Barton 8 Hamilton 1982)
Wolgidite (Arima 8 Edgar !983b)
40 -f
-140
35
120 ph Iogopite
orthopyroxenite (opx+phl)
p_h I__qgo_pite ?garnet- Iherzolite {ol+opx+cpx+gt+ph)
30
Io0 T s s
25
ph I__0_ogo_pit3 orthgpy_roxenite (opx+phl)
p hl_#_ogop~t__e harzbuEgit__e (oi+ opx + phl)
K3
H. p_ ogoplte harzburqit._ee (opx +ol + phi)
80
20
(ol Only) I te p.hi ogopj harzburglte (o1+ opx+ phi) 5%
1'2%
60 15%
H20 added FIG. 12. Schematic diagram showing near-liquidus phases for SiO2-saturated and SiO2-oversaturated K-rich magmas in which H20 is the only added volatile. The underlined rock names are approximately equivalent to the near-liquidus assemblages in brackets. 1400
m
L
Iq
o
1400 1 30 kb (equiv to 15 wt% H20)
.+v/
J+FTa 1300
,,'
(bq~//
1300 -
,,
r
OI
,p
/
-....--
/ /
L O
/
~ ~ +
Xtols
/ /
1200
_
/
1200
-
,., C~4\\
-
L+Xtals
-% \\ \ \ \,o. \ \ \-o \%
30 kb {equiv. to 5 wf% H20) I100 H20 5 wt~
I
0.25
I
0-5
I
0.75
C02/(C%+ H20} mole fraction in charge
CO2 12.2 wt~
I100 H20 15 w1%
I
0.25
\
I
0"5
I
0.75
C02/(C02-1- H20) mole fraction in charge
C% 36.7 wt%
F]G. 13. Phase relations in the biotite mafurite system with varying C02/(H20 --kC O 2 ) contents equivalent to 5 and 15 wt.% H20 at 30 kb. Abbreviations are as in Figs 2-4. (Modified from Ryabchikov & Green 1978.)
Genesis of alkaline magmas Katungite (Arima a Edgar 1983a)
OI-ugandite (Edgar et 0/ 1980
40
t
:55 -
Mafurite (Ryabchikova Green 1978)
I
garnetIherzolite ? (cpx + opx + ga)
43
i
' ' garnet' clinopyroxenite (cpx + ga)
-
140
'
120
,r
E
v (D
50
Iherzolite (cpx + opx + ol)
clinopyroxenite (cpx)
r
ioo
d3
13._
25
_
garnet or spinel Iherzolite (opx + cpx + ol + ga + sp) 80
wehrlite (ol +cpx)
20 COz H201
27.4 3-5
27.5 3.8 % CO2 and H20 added
27.5 5"8
FIG. 14. Schematic diagram showing near-liquidus phases for SiO2-undersaturated K-rich magmas with CO2/(CO2 + H20) mol = 0.75 (equivalent to 15 wt.% HzO). The underlined rock names are approximately equivalent to the near-liquidus assemblages in brackets.
Mag Pressure (kb) of appearance of liquids Vol absent 19 34-5 --
CO2 satd. 14 27.5 > 29
Liquids Lc normative Ks normative Mag normative Carbonatitic
/
/
/
/
~.// 7
/
7
// "
~
/
/
/ 9
/
/
/~hl
/
Olivine tholeiitic
~/
/
/
/ //
/ /
/
/
Quartz.. tholeiitic
\
/
7
/.. Ks
/ Lc [ Sa (Ks normative] Alkali basaltic / Lc normative}
Qz
FIG. 15. Trends of liquids with increasing pressure in the systems KAISiO4-MgzSiO4-SiO2 and KA1SiO,-MgOSIO2-H20-CO2 (modified from Wendlandt & Eggler 1980a). Abbreviations are as in Figs 1 and 3 ; Mag, normative magnesite. The inset shows the normative nature of liquids at increasing pressures (as indicated by the line on the diagram) under volatile-free and CO2-saturated conditions.
44
A. D. Edgar
compositions of the residual liquids on partial melting become leucite, kalsilite and magnesite normative are shown in the inset to Fig. 15. At approximately 14 kb for CO2-saturated conditions and 19 kb for volatile-free conditions, liquids in equilibrium with sanidine, enstatite and forsterite are leucite normative and change from tholeiitic to alkali basaltic with increasing pressure. At greater than 27.5 kb under CO2saturated conditions and greater than 34.5 kb in the volatile-free system, sanidine is no longer stable and melts become extremely K20 rich and SiOz poor, yielding kalsilite-normative liquids. At pressures above 29 kb in the KA1SiO4-MgOSIO2-CO2 system melts are actually magnesite normative. Based on these results Wendlandt & Eggler (1980a) concluded that quartz tholeiitic liquids cannot fractionate to alkali basaltic ones. Liquids generated at high pressures with some fractionation could result in increasingly SiO2poor residual liquids. In contrast, fractionation at lower pressures could result in SiO2 enrichment. Wendlandt & Eggler's (1980a) experiments have been used by Kuehner et al. (1981) to model the Leucite Hills, Wyoming, magmas which show this trend of SiO2 enrichment. Although it is theoretically possible that leucite-normative rocks could be products of fractionation rather than direct partial melting, Wendlandt & Eggler (1980a) consider this to be unlikely. Using the KA1SiO4-MgO-SiO2-CO2-H20 system in which phlogopite is an additional phase, Wendlandt & Eggler (1980b) modelled the melting behaviour of phlogopite peridotite and its role in the genesis of K-rich magmas, kimberlites and carbonatites. Peridotite with small amounts of H20 + CO2 undergoes univariant melting in which H20 and CO2 vapour compositions are buffered by phlogopite and magnesite respectively. Such buffering is pressure dependent and hence the compositions of liquids produced by various reactions depend on the pressure of melting. With increasing pressure up to 30 kb, melting of a phlogopite peridotite with insufficient CO2 and H20 to hydrate all the potential phlogopite involves four reactions between combinations of phlogopite, enstatite, sanidine (or kalsilite), forsterite, liquid and vapour. As the pressure increases up to 30 kb the liquids change in composition progressively from quartz to enstatite to leucite to kalsilite normative. These relations are comparable with those for the KA1SiO4-Mg2SiO4-SiO2 system (Wendlandt & Eggler 1980a). Above 30 kb magnesite becomes stable and melting occurs by two reactions involving combinations of phlogopite, enstatite, magnesite, forsterite, kalsilite, liquid and vapour. The liquids produced are carbonatitic. With
increasing pressure phlogopite has decreased hypersolidus stability and at 50 kb ceases to be a solidus phase. Thus between 30 and 50kb, melting of phlogopite peridotite will produce increasingly K20-rich SiO2-poor liquids.
Mantle metasomatism Experimental studies and other data indicate that as the alkalinity of mafic to ultramafic magmas increases there is a decreasing possibility that these magmas are partial melts of a 'normal' pyrolitic model mantle composition. Although there are a number of mechanisms whereby local heterogeneities can occur in the mantle source, the concept of heterogeneities being caused by mantle metasomatism (cf. Lloyd & Bailey 1975; Bailey 1982) is now widely accepted. Experimental modelling of metasomatic processes is extremely difficult, particularly at mantle depths where very little is known of either the composition or the speciation of the fluids involved. Among the many elements considered to be responsible for metasomatism (Bailey 1982), K and Na are very important. Ryabchikov & Boettcher (1980) showed that the solubility of potassium (as K/O) in aqueous fluids increased about sixfold between 11 and 30 kb in the range 1050-1100~ Their values, ranging from 4 to 25 g K20 per 100 g H20, indicated that the solubility of K:O is sufficiently high to produce radical changes in the mantle. Ryabchikov et al. (1982) determined that the amount of sodium silicate leached by hydrous fluids at constant temperature from omphacitic pyroxenes increases with decreasing pressure. On the basis of both sets of experiments, Ryabchikov et al. (1982) proposed that K/Na fractionation in the upper mantle might produce K-rich metasomatism at deeper levels and Na-rich metasomatism at shallower levels which, on partial melting, would result in K-rich magmas generated at greater depths than Na-rich magmas. Edgar & Arima (1984; unpublished data) have carried out preliminary metasomatism-modelling experiments involving the system pyroliteK2Oaqueou s and pyrolite-(K20+Na20)aq ..... at 2 0 - 3 0 kb and between 850 and 950~ In these experiments K20 , as dried K2CO3, was dissolved in deionized H20. Na20 was added as Na2CO 3. This resulted in an aqueous solution containing negligible amounts of dissolved CO2. Figure 16(a) shows that at 30 kb an increased concentration of K20 in solution produces an initial increase in the proportions of phlogopite relative to orthopyroxene, clinopyroxene and olivine. Although
Genesis of alkaline magmas not in sufficient quantity to be observed in the Xray diffraction patterns (Fig. 16) garnet (at 30 kb) and spinel (at 20 kb) decrease with increasing K20 as observed optically. At about 3.8 g K20 per 10 g of solution, liquid appears at both 850 and 950~ indicating a decrease in the amount of phlogopite with K partitioning preferentially into the liquid. These results are in accordance with those of Takahashi & Kushiro (1983). Edgar & Arima (1984) were unable to analyse their run products using the electron microprobe because of the fine grain size, but they speculated that at 30 kb the formation of phlogopite involved a reaction between olivine, orthopyroxene and garnet in the approximate proportions ofpyrolitetype mantle at 30 kb (Wyllie 1975) with the K20 in solution to produce a product with less garnet and olivine and more orthopyroxene. Clinopyroxene is probably not involved in the formation of phlogopite. These results are in accordance with the observed decrease in the olivine/orthopyroxene and olivine/clinopyroxene intensities (Fig. 16(a)). In contrast with the results at 30 kb, at 20 kb an amphibole of pargasitic composition as analysed by microprobe shows an inverse relationship to phlogopite which remains the dominant 'metasomatic' product (Fig. 16(b)). On the basis of the relative proportions of phlogopite to the other mineral phases (Fig. 16(b)) and the ratios of olivine to orthopyroxene, olivine to clinopyroxene and orthopyroxene to clinopyroxene (Fig. 16(c)), the formation of phlogopite at 20 kb and at K20 concentrations where amphibole is no longer present (more than 2.0 g K20 per 10 g of solution) may involve a complex reaction between olivine, orthopyroxene, clinopyroxene, amphibole and spinel, together with K20 in solution, to form a product containing a higher amount of olivine and clinopyroxene and lower amounts of orthopyroxene relative to the reactants. A few preliminary experiments at 20 kb and 950~ in which both Na and K have been added in various proportions equivalent to 3.8 g total alkali per 10 g of solution show that, with increasing Na relative to K, the field of amphibole stability increases as phlogopite decreases. Although vastly simplified, these experiments indicate that alkali metasomatism can readily occur. Nodules with mantle mineralogy found in alkali basalts, kimberlites and related rocks often contain phlogopite and amphibole and range in overall composition from lherzolite to pyroxenite. Lloyd & Bailey (1975) and Lloyd (1981) studied nodules mainly of phlogopite clinopyroxenite composition from the K-rich volcanic province of SW Uganda and proposed that they might represent metasomatized mantle sources for these
45
lavas. Alternatively these nodules could be depleted residual source materials. Lloyd et al. (1985) determined the compositions of liquids produced from varying degrees of partial melting at 20 and 30 kb of an average nodule with no added volatiles (based on analyses of 84 nodules from this area). Figure 17 shows the compositions of liquids produced by various degrees of partial melting at 30 kb together with the coexisting minerals. At approximately 25% partial melting of this nodule, the liquid corresponds closely with that of an average katungite composition found in SW Uganda. Although this is a higher degree of partial melting than generally considered likely for the production of such K-rich magmas, it seems feasible in the light of the upwarped mantle isotherms, postulated in rifted areas such as SW Uganda (Bailey 1970), caused by mantle degassing (Bailey 1980), and in view of the high proportion of radioactive elements associated with these magmas.
Felsic alkaline rocks Although experiments pertinent to felsic alkaline rocks have been less numerous than those carried out on mafic varieties in the past decade, studies have been undertaken on liquid immiscibility, on the role of C1, F and P20 5 in alkaline magmas and on the problems of the genesis of pseudoleucites and primary analcites.
Liquid immiscibility Roedder (1979) reviewed the role of experimental work on the concept of liquid immiscibility as a viable process in the genesis of igneous rocks. Only studies applicable to alkaline rocks are considered here, most of which concern silicatecarbonatitic magmas. Freestone (1978) showed that the liquid immiscibility field in the fayalite-leucite-silica system was extended to more K-rich compositions by the addition of 1 mol ~ P205 and 3 mol ~ TiO2. He suggested that this expansion supports the liquidunmixing origin suggested by ocelli textures in syenites and sheets within alkali gabbro intrusions. The role of liquid immiscibility on the genesis of carbonatites was considered by Hamilton et al. (1979) and Freestone & Hamilton (1980). These workers used lavas ranging in composition from felsic nephelinites to phonolites comparable in composition with those in the Oldoinyo Lengai volcano, Tanzania. These were mixed with synthetic Na- and Ca-rich carbonatites. Experiments at 0.7-7.6 kb and 900-1250~ showed that
46
A. D. Edgar 8
_- q~p
/
6
"~ ~
o
o=~=I=ua 2 0
-
@
o
k
,~ ~
~ H
o
1.5~ I'0
o
u H
4
\ Iamph 2
o
oI
0
1"5~~I'0
9
-~---I
H
0.5 8
4
q,t-fl~p
4
6 n
"~
3 2 I
0
v
I'0
2.0
3-0
4-0
I
~
6.0
5.0
1
i0
2. 0
50
l
4-0
g K20/IO g sOIn.
g K 2 0 / I O g soln.
(a)
(b)
I 5.0
I 6,0
3-0 -
2o' I.O
o
u
-6
1.5
0"5
2"0 I-0 I ~ - -
0
a-
1
I-O
o
@ I
2.0
I
3.0
I
4-0
g K20/IO g soln. (C)
I
5-0
I
6-0
FIG. 16. Results of experiments in the pyrolite-K20 aqueous solution system plotted as ratios of X-ray diffraction intensities (I) of minerals against K 20 concentrations: (a) runs at 30 kb, 850~ ( 0 ) , and at 30 kb, 950~ (O); (b) runs at 20 kb, 950~ showing variations in the ratios of phlogopite and amphibole relative to other minerals; (c) runs at 20 kb, 950~ showing variations in the ratios of olivine, orthopyroxene and clinopyroxene. Abbreviations are as in Figs 2 and 3 ; q, quench phlogopite; p, primary phlogopite. (After Edgar & Arima 1984.)
Genesis of alkaline magmas
~_ApI
(not used in Freestone & Hamilton's experiments) liquid immiscibility could account for the carbonate-rich groundmass in kimberlites and the latestage associated carbonatite dykes and diapirs.
Phi C p x ~
Im. . . .
37
Halogen-bearing
I0
~
8-
u
/
-
e
-
systems
Hards (1976) determined the distribution of certain elements between the melt and vapour phases in granite and nepheline syenite compositions at 1 kb between 680 and 850~ in the presence of C1 and F. He found that C1 was more strongly concentrated in the vapour phase in granitic melts than it was in nepheline syenite melts, whereas F was concentrated preferentially in the liquid phase for both compositions and only entered the vapour phase after crystallization
O
~
6-
NazO
0 ~- ~ 9
e
O~
xcr..-"
if
0
47
J
w,o,o
II
L
I
i
I
~
I
20 40 60 80 Degree of melting (wt.%)
i
I
I00
FIG. 17. Compositions of liquids with varying degrees of partial melting of an average clinopyroxenite nodule from SW Uganda at 30 kb : - - , determined compositions; - - - , calculated compositions based on analyses of minerals. The phases present are given at the top of the figure. Cpx, clinopyroxene: Phi, phlogopite: Ilm, ilmenite: Ap, apatite. (After Lloyd et al. 1985)
Si% + AiaO3
CaO (a)
Na20
the miscibility gap expands with increasing Pco2 and decreasing temperature (Fig. 18). Results at 0.7 kb and l l00~ are plotted on the N a 2 0 (SiO2+AlzO3)-CaO system (Fig. 18(a)). Under the miscibility gap inincreasing P t o t a l = P c o 2 , creases and there is a rotation of tie-lines resulting in more CaO-rich liquids (Fig. 18(b)). From the compositions of the melts, Freestone & Hamilton (1980) concluded that the natrocarbonatites from Oldoinyo Lengai separated from a phonolitic rather than a nephelinitic magma. Because the miscibility gap closes away from the Na20-rich compositions (Fig. 18), exsolution of a carbonatitic melt is more likely in salic than in mafic silicate magmas. Freestone & Hamilton's (1980) results do not support the idea that kimberlite sills and dykes form by an immiscibility process at crustal pressures. Bedson & Hamilton (1981), however, showed that in the presence of H 2 0
SiOa + Ala03
CaO
(b) FIG. 18. Liquid immiscibility in the system NazOCaO-(A1203 + SiO2) at (a) 0.7 kb and 1100~ and (b) 1100~ and varying pressures Pc% = Ptotal. (After Freestone & Hamilton 1980.)
A. D. E d g a r
48
was completed. Sodium was the most common element in the vapour phase of both compositions. These experiments may explain the presence of Cl-bearing sodalite in SiO2-undersaturated rocks but the absence of Cl-bearing minerals from granitic rocks. On the basis of Hard's experiments the presence of F-bearing minerals in both types of rocks is to be expected. On the basis of experiments at 1 kb in the system NaA1SiO4 (Ne)-SiO2-NaC1-H20, Barker (1976) found that in the SiO2-undersaturated portion of the system NaC1 stabilized sodalite whereas in the SiO2-oversaturated part it has no effect on phase relations because of immiscibility between NaCl-rich liquid and SiO2 melt. These results suggest that the high C1 content of feldspathoidal rocks is not a cause of their SiO 2 undersaturation and that such rocks are unlikely to be produced by anatexis or assimilation of halite-bearing sediments (Appleyard 1974; Ayrton 1974). For the SiO2-undersaturated part of the same system, Binsted (1981) showed that sodalite might exist as a primary mineral between about 25% and 60% Ne in the pseudobinary ( A b + Ne)gs(NaC1)5 join (Binsted 1981, Fig. 13). He found that melts crystallizing sodalite follow a non-linear trend (Binsted 1981, Fig. 14) owing to partial replenishment of NaCI in the melt from the fluid reservoir. With increasing NaC1, nepheline will crystallize as the second phase following albite. These relationships may be important in the genesis of agpaitic rocks such as the lujavrites in Ilimaussaq, Greenland, where sodalite is a primary mineral.
sial for many years. Taylor & Mackenzie (1975) studied differences in leucite morphologies with different cooling methods at 2 kb Pnz o in the system NaA1SiO4-KAISiO4-SiOz-H20. As shown in Fig. 19 the compositions of the leucites produced appeared to depend on their morphologies. From the common occurrence of natural pseudoleucites along the leucite-analcite join, Taylor & MacKenzie (1975) suggested that metastable Na-rich leucites are produced by extremely rapid alkali-ion exchange rather than by the breakdown of high-temperature leucites or by leucite breaking down to nepheline+ feldspar at the reaction point in this system.
Primary versus secondary origin of analcites Experimental studies bearing on the problem of the origin of analcite phenocrysts in a few rare volcanic rocks have been carried out by Roux & Hamilton (1976) in the system NaA1SiO4KA1SiO4-SiO2. They found that analcite coexisted with liquid above about 5 kb PH2o and 600~ supporting the concept of primary analcites. In contrast, Gupta & Fyfe (1975) found a very rapid reaction between leucite and saturated NaCI solutions at 1 kb Pn2 o at low temperatures. This supported the idea that these natural analcite phenocrysts could be transformed from leucites in a low-temperature hydrothermal environment. The plausibility of both of these mechanisms occurring in nature has been discussed by Edgar (1984).
Other experiments on felsic alkaline rocks Pseudoleucites The origin of pseudoleucites has been controver-
Dolfi et al. (1978) and Ruddock & Hamilton (1978) investigated the KA1Si206 (Lc)-
Zoned
c Ne
"x,~~Ks Greatly exsolved
leucite
Slightly ( I11r-='_lib IIIII
exsolved I--'l
leucite
~
FIG. 19. Recalculated microprobe analyses and morphologies of leucites crystallized from a bulk composition (+) crystallized in part of the system NaA1SiO4-KA1SiO4-SiO2-H20. Lines join spots probed on leucite crystals to compositions in the system. (After Taylor & Mackenzie 1975.)
Genesis of alkaline magmas CaMgSi20 6 (Di) and KA1Si206-CaMgSi20 6S i O 2 - H 2 0 - C O 2 systems respectively. In the former system, Dolfi et al. showed that increasing pressure up to 12 kb increased the primary diopside field at the expense of leucite. In experiments at 4 kb in the system KA1Si20 6CaMgSizO6-H20, Ruddock & Hamilton (1978, Fig. 19) indicated that with increasing pressure the fields of leucite and quartz contract while those of diopside and sanidine expand. Phlogopite is restricted to compositions with low diopside contents. On the basis of these results they explain the common phenocryst assemblage of diopside and phlogopite along with groundmass quartz and potash feldspar in many lamprophyres. High-pressure experiments up to 30 kb on the composition Lc60DisQ35 with H 2 0 and CO2 (Ruddock & Hamilton 1978, Fig. 19) confirmed these conclusions. These experiments also suggested that minettes are derived by partial melting of the mantle and that primary carbonate minerals in these rocks could have formed at mantle pressures. In order to determine the effect of CaA12Si2Os
49
(An) on phase relations in the system NaA1SiO4KA1SiO4-SiOz-H20, Norris & MacKenzie (1976) and Whiteley (1981) have determined phase relations at 1 kb PH20 with 3, 5 and 10 wt.% CaA12SizOs. With increasing An contents, the leucite and alkali feldspar fields are reduced as the plagioclase field expands. For the 3% An plane, Norris & MacKenzie (1976, Fig. 30) located a eutectic between nepheline, K-rich feldspar and Na-rich feldspar. For the 5% and 10% An planes (Norris & MacKenzie 1976, Figs 31 and 32) the projected phase volumes of Kfeldspar and Na-feldspar cut the leucite phase volume at progressively more K-rich compositions as the K-rich feldspar field contracts. In the absence of a Ca-bearing mafic mineral, such as diopside, these systems more closely approximate felsic magmas.
ACKNOWLEDGMENTS: I am grateful to M. Arima and A. Kolisnik for their assistance in the preparation of the diagrams for this paper. The Natural Science and Engineering Research Council of Canada provided financial support.
References ALLEN,J. C., BOETTCHER,A. L. & MARLAND,G. 1975. Amphiboles in andesite and basalt: I. Stability as a function of P - T - f 0 2 . Am. Mineral. 60, 1069-85. APPLEYARD,E. C. 1974. Syn-orogenic igneous alkaline rocks of eastern Ontario and northern Norway. Lithos, 7, 147-60. ARCULUS, R. J. 1975. Melting behaviour of two basanites in the range 10-35 kbar and the effect of TiO 2 on the olivine-diopside reactions at high pressures. Yearb. Carnegie Inst. Washington, 512-15. ARIMA, M. & EDGAR, A. D. 1983a. High-pressure experimental studies on a katungite and their bearing on the genesis of some potassium-rich magmas of the west branch of the African rift. J. Petrol. 24, 166-87. & - - 1983b. A high pressure experimental study of a magnesian-rich leucite-lamproite from the West Kimberley area, Australia: petrogenetic implications. Contrib. Mineral. Petrol. 84, 228-34. AYRTON, R. 1974. Rifts, evaporites and the origin of certain alkaline rocks. Geol. Rundsch. 63, 430-50. BAILEY, D. K. 1970. Volatile flux, heat focussing and the generation of magma. Geol. J. Spec. Issue 2, 177-86. 1974. Experimental petrology relating to oversaturated peralkaline volcanics, a review. Bull. Volcanol. 38, 637-52. Applications of experiments to alkaline rocks. In: BAILEY,D. K. & MACDONALD,R. (eds) The Evolution of the Crystalline Rocks, pp. 419-69. Academic Press, New York. 1980. Volatile flux, geotherms, and the generation 74,
-
-
-
-
- - 1 9 7 6 .
-
-
of the kimberlite-carbonatite-alkaline magma spectrum. Mineral. Mag. 43, 695-9. - 1982. Mantle metasomatism--continuing chemical change within the earth. Nature, Lond. 296, 525-30. --,TARNEY, J. & DUNHAM, K. C. (eds) 1980. Evidence for chemical heterogeneity in the earth's mantle. Phil Trans. R. Soc. Lond. Ser. A, 297, 134493. BARKER, D. S. 1976. Phase relations in the system NaA1SiO4-SiO2-NaC1-H20 at 400-800~ and 1 kilobar, and petrologic implications. J. Geol. 84, 77-106. - - 1983. Igneous Rocks, 417 pp. Prentice-Hall, Englewood Cliffs, NJ. BARTON, M. & HAMILTON,D. L. 1979. The melting relationships of a madupite from the Leucite Hills, Wyoming, to 30 kb. Contrib. Mineral. Petrol. 69, 133-42. & - - 1 9 8 2 . Water undersaturated melting experiments bearing upon the origin of potassiumrich magmas. Mineral. Mag. 45, 267-78. BASALTIC VOLCANISMSTUDY PROJECT, 1981. Basaltic Volcanism on the Terrestrial Planets, 1286 pp. Pergamon Press, Oxford. BEDSON, P & HAMILTON, n. L. 1981. Kimberlites, carbonatites and liquid immiscibility. Prog. Exp. Petrol., NERC Publ. Set. D, 29-32. BELL, K. & POWELL, J. L. 1969. Strontium isotopic studies of the Burunga and Toro-Ankole regions, east and central equatorial Africa. J. Petrol. 10, 536-72. BERGMAN,S. C. 1987. Lamproites and other potassium-
5o
A. D. Edgar
rich igneous rocks: a review of their occurrence, mineralogy and geochemistry. In: FITTON, G. J. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 103-90. BINSTED, N. 1981. The system Ab-Ne-NaC1-HzO. Prog. Exp. Petrol., NERC Publ. Ser. D, 34-35. BOWEN, N. L. 1922. The reaction principle in petrogenesis. J. Geol. 30, 177-98. --1928. Evolution of the Igneous Rocks, 251 pp. Dover, New York. BREY, G. 1976. CO2 solubility and solubility mechanisms in silicate melts at high pressures. Contrib. Mineral. Petrol. 57, 217-21. 1978. Origin of olivine melilitites--chemical and experimental constraints. J. Volcanol. geotherm. Res. 3, 61-88. & GREEN, D. H. 1975. The role of CO2 in the genesis of olivine melilitite. Contrib. Mineral. Petrol. 49, 93-103. & -1976. Solubility of COz in olivine melilitite at high pressures and the role of CO2 in the earth's upper mantle. Contrib. Mineral. Petrol. 55, 217-30. & -1977. Systematic study of liquidus phase relations in olivine melilitite + H20 + CO 2 at high pressures and petrogenesis of an olivine melilitite magma. Contrib. Mineral. Petrol. 61,141-62. BULTITUDE, R. J. & GREEN, D. H. 1968. Experimental study at high pressures on the origin of olivine nephelinite and olivine melilite nephelinite magmas. Earth planet. Sci. Lett. 3, 325-37. & -1971. Experimental study of crystal-liquid relationships at high pressures in olivine nephelinite and basanite compositions. J. Petrol. 12, 12147. CAWTI-IORN, R. G. 1976. Melting relations in part of the system CaO-MgO-A1203-SiO2-Na20-H20 under 5 kb pressure. J. Petrol. 17, 44-72. CUNDARI, A. & O'HARA, M. J. 1976. Experimental study at atmospheric and high pressures of a mafic leucitite from New South Wales, Australia. Prog. Exp. Petrol., N.E.R.C. Publ. Ser. D, 260-2. DALY, R. A. 1910. Origin of the alkaline rocks. Geol. Soc. Am. Bull. 21, 87-115. DOLFI, O., HAMILTON,D. H. & TREGILA, R. 1978. The system KA1Si206-CaMgSi206 at 4 and 12 kilobars. Prog. Exp. Petrol., NERC Publ. Ser. D, 21-2. DUKE, J. M. 1974. The effect of oxidation on the crystallization of an alkali basalt from the Azores. J. Geol. 82, 524-8. EDGAR, A. D. 1974. Experimental studies. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp. 355-89. Wiley, New York. 1979. Mineral chemistry and petrogenesi: of an ultrapotassic-ultramafic volcanic rock. Contrib. Mineral. Petrol. 71, 171-5. 1984. Chemistry, occurrence and paragenesis of feldspathoids: a review. In: BROWN, W. L. (ed.) Feldspars and Feldspathoids, pp. 501-32. Reidel, Dordrecht. & ARIMA, M. 1983. Conditions of phlogopite crystallization in ultrapotassic volcanic rocks. Mineral. Mag. 47, 11-19. -& -1984. Experimental studies on K-metasomatism of a model pyrolite mantle and their -
-
-
-
-
-
-
-
-
-
bearing on the genesis of ultrapotassic magmas. Proc. 27th Int. Geological Cong. on Petrology (Igneous and Metamorphic Rocks), Vol. 9, 509-41. V N U Science Press, Utrecht. CONDLIFFE,E., BARNETT,R. L. & SHIRRAN, R. J. 1980. An experimental study of an olivine ugandite magma and mechanisms for the formation of its K-enriched derivatives. J. Petrol. 21,475-97. - - , GREEN, D. H. & HIBBERSON, W. O. 1976. Experimental petrology of a highly potassic magma. J. Petrol. 17, 339-56. EGGLER, D. H. 1974. Effects of CO2 on the melting of peridotite. Yearb. Carnegie Inst. Washington, 73, 215-24. 1978. The effect of CO2 upon partial melting of peridotite in the system Na20-CaO-A1203-MgOSIO2-CO2 to 35 kb, with an analysis of melting in a peridotite-H20-CO2 system. Am. J. Sci. 278, 305-43. --& HOLLOWAY, J. R. 1977. Partial melting of peridotite in the presence of H20 and CO2: principles and review. Oreg. Dep. Geol. mineral Industries Bull. 96, 15-36. • WENDLANDT, R. F. 1979. Experimental studies of the relationship between kimberlite magmas and partial melting of peridotite. In: BOYD, F. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds: Their Geology, Petrology and Geochemistry, pp. 330-8. American Geophysical Union, Washington, DC. EL-GORESY, A. & YODER, H. S. 1974. Natural and synthetic melilite compositions. Yearb. Carnegie Inst. Washington, 73, 359-71. ELLIS, D. E. & WYLLIE, P. J. 1980. Phase relations and their petrological implications in the system MgOSIO2-H20-CO2 at pressures up to 100 kbar. Am. Mineral. 65, 540-56. FREESTONE, I. C. 1978. Liquid immiscibility in alkalirich magmas. Chem. Geol. 23, 115-23. --& HAMILTON, D. L. 1980. The role of liquid immisciblity in the genesis of carbonatites--an experimental study. Contrib. Mineral. Petrol. 73, 105-17. FREY, F. A., GREEN, D. H. & ROY, S. D. 1978. Integrated models of basalt petrogenesis: a study of quartz tholeiites to olivine melilitites from southeastern Australia utilizing geochemical and experimental petrological data. J. Petrol. 19, 463-513. GREEN, D. H. 1969, The origin of basaltic and nephelinitic magmas in the earth's mantle. Tectonophysics, 7, 409-22. 1970a. A review of experimental evidence on the origin of basaltic and nephelinitic magmas. Phys. Earth planet. Inter. 3, 221-35. The origin of basaltic and nephelinitic magmas. Trans. Leicester Lit. Phil. Soc. 44, 26-54. 1971. Composition of basaltic magmas as indicators of conditions of origin: applications to oceanic volcanism. Phil. Trans. R. Soc. Lond. Ser. A, 268, 707-25. 1973a. Conditions of melting of basanite magma from garnet peridotite. Earth planet. Sci. Lett. 17, 456-65. Experimental melting studies on model -
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1
9
7
0
b
.
1
9
7
3
b
.
Genesis of alkaline magmas upper mantle compositions at high pressure under both water-saturated and water-undersaturated conditions. Earth planet. Sci. Lett. 19, 37-53. - - , GUPTA, A. K. & TAYLOR, W. R. 1984. Experimental studies on the effects of H20 and CO2 on liquidus relations in Fo-Ks-SiO2 and F o - N e SiO2 at 28 kb. Abstracts, Workshop on Experimental Geochemistry, Monash University, Melbourne. HmBERSON, W. O. 1970. Experimental duplication of conditions of precipitation of high pressure phenocrysts in a basaltic magma. Phys. Earth planet. Inter. 3, 247-54. & RINGWOOD, A. E. 1967. The genesis of basaltic magmas. Contrib. Mineral. Petrol. 15, 103-90. GUPTA, A. K. & FYFE, W. S. 1975. Leucite survival: the alteration to analcime. Can. Mineral. 13, 3613. HAMILTON, D. L., FREESTONE, I. C., DAWSON, J. B. & DONALDSON, O. H. 1979. Origin of carbonatites by liquid immiscibility. Nature, Lond. 279, 52-4. HARDS, N. 1976. Distribution of elements between the fluid phase and silicate melt phase of granites and nepheline syenites. Prog. Exp. Petrol., NERC Publ. Ser. D, 88-90. HELZ, R. T. 1973. Phase relations of basalts in their melting range at PH2o=5 kb as a function of oxygen fugacity. Part I. Mafic phases. J. Petrol. 14, 249-302. 1976. Phase relations of basalts in their melting ranges at PH2o = 5 kb. Part II. Melt compositions. J. Petrol. 17, 139-93. HOLMES, A. 1932. The origin of igneous rocks. Geol. Mag. 69, 543-58. 1950. Petrogenesis of katungite and its associates. Am. Mineral. 35, 772-92. -1965. Principles of Physical Geology (2nd edn), 623 pp. Ronald Press, New York. --& HARWOOD, F. 1937. The petrology of the volcanic rocks of Bufumbira. Mere. Geol. Surv. Uganda, 3 (2), 1-300. HUANG, W. L. & WYLLIE, P. J. 1974. Eutectic between wollastonite I1 and calcite contrasted with thermal barrier in i g O - S i O 2 - C O 2 at 30 kilobars with emphasis to kimberlite-carbonatite petrogenesis. Earth planet. Sci. Lett. 24, 305-10. IRVING, A. J. & GREEN, D. H. 1972. Experimental study of phase relationships in a high pressure mugearitic basalt as a function of water content. Geol. Soc. Am. Meet. Abstracts with Programs 4, 550-1. - - - & PRICE, R. C. 1981. Geochemistry and evolution of high pressure phonolitic rocks from Nigeria, Australia, Eastern Germany and New Zealand. Geochim. cosmochim. Acta, 45, 1309-20. ITO, K. & KENNEDY, G. C. 1968. Melting and phase relations in the plane tholeiite-lherzolite-nepheline basanite to 40 kilobars with geological implications. Contrib. Mineral. Petrol. 19, 177-211. KAY, R. W. & GAST,P. W. 1973. The rare earth content and origin of alkali rich basalts. J. Geol. 81, 65382. KOGARKO, L. N. & ROMANCHEV, B. P. 1982. Phase equilibria in alkaline melts. Zap. vses. mineral. Ova. 8, 167-182 (in Russian). -
-
-
-
-
-
-
-
&
51
KUEHNER, S. M., EDGAR, A. D. & ARIMA, M. 1981. Petrogenesis of the ultrapotassic rocks from the Leucite Hills, Wyoming. Am. Mineral. 66, 663-77. KUSHIRO, I. 1973. Origin of some magmas in oceanic and circum-oceanic regions. Tectonophysics, 17, 211-22. -& YODER, n . S. 1974. Formation of eclogite from garnet lherzolite: liquidus relations in a portion of the system MgSiO3-CaSiO3-A1203 at high pressures. Yearb. Carnegie Inst. Washington, 73, 266-9. LARSEN, E. S. 1940. Petrographic province of central Montana. Geol. Soc. Am. Bull. 51,887-948. LLOYD, F. E. 1981. Upper mantle metasomatism beneath a continental rift: Clinopyroxenes in alkalic mafic lava and nodules from South West Uganda. Mineral. Mag. 44, 315-24. & BAILEY, D. K. 1975. Light element metasomatism of the continental mantle: the evidence and the consequences. Phys. Chem. Earth, 9, 389-416. , ARIMA,M. & EDGAR, A. D. 1985. Partial melting of a phlogopite-clinopyroxenite nodule from south-west Uganda: an experimental study bearing on the origin of highly potassic continental rift volcanics. Contrib. Mineral. Petrol., 91, 321-9. LUTH, W. C. 1967. Studies in the system of KA1SiO4Mg2SiO4-SiO2-H20. I. Inferred phase relations and petrologic applications. J. Petrol. 8, 372-416. MCKENZlE, D. P. 1984. The generation and compaction of partially molten rock. J. Petrol. 25, 713-65. MERRILL, R. B. & IRVING, A. J. 1977. Chemistry and phase relations of an orthopyroxene-bearing transitional alkalic basalt (abstract). Los, 58, 526. --& WYLLIE, P. J. 1975. Kaersutite and kaersutite eclogite from Kakanui, New Zealand--water excess and water deficient melting relations to 30 kilobars. Geol. Soc. Am. Bull. 86, 555-70. MYSEN, B. O. & KUSHIRO, I. 1977. Compositional variations of coexisting phases with degree of melting of peridotite in the upper mantle. Am. Mineral. 62, 843-65. NORRIS, G. & MACKENZIE,W. S. 1976. Phase relations in the system NaA1SiO4-KAISiO4-CaA12Si2OsSiO2 at PH2o = 1 kb. Prog. Exp. Petrol., NERC Publ. Set. D, 79-81. O'HARA, M. J. 1968. The bearing of phase equilibria studies on synthetic and natural systems on the origin and evolution of basic and ultrabasic rocks. Earth Sci. Rev. 4, 69-133. & BIGGAR, G. M. 1969. Diopside-spinel equilibria, anorthite and forsterite reaction relationships in silica-poor liquids in the system C a - M g O A1203-SiO 2 at atmospheric pressure and their bearing on the genesis of melilitites and nephelinites. Am. J. Sci. 267A, 364-90. YODER, H. S. 1967. Formation and fractionation of basic magmas at high pressures. Scott. J. Geol. 3, 67-117. PRESNALL, D. C., DIXON, J. R., O'DONNELL, T. H. & DIXON, S. A. 1979. Generation of mid-ocean ridge tholeiites. J. Petrol. 20, 3-35. , DIXON, S. A., DIXON, J. R., O'DONNELL, T. H., BRENNER, N. J., SCHROCK, R. L. & DYCUS, D. W. 1978. Liquidus phase relations on the join diopside-forsterite-anorthite from 1 atm to 20 kbar; -
-
-
-
-
-
&
52
A. D. Edgar
their bearing on the generation and crystallization of basaltic magma. Contrib. Mineral. Petrol. 66, 203-220. RINGWOOD, A. E. 1975. Composition and Petrology of the Earth's Mantle, 618 pp. McGraw-Hill, New York. ROEDDER, E. 1979. Silicate liquid immiscibility in magmas. In: YODER, H. S. (ed.) The Evolution of the Igneous Rocks: Fiftieth Anniversary Perspectives, pp. 15-58. Princeton University Press, Princeton, NJ. Roux, J. & HAMILTON, D. L. 1976. Primary igneous analcite--an experimental study. J. Petrol. 17, 244-57. RUDDOCK, D. I. & HAMILTON, D. L. 1978. The system KA1SizO6-CaMgSi206-SiO2-H20 at 4 kilobars. Prog. Exp. Petrol., NERC Publ. Ser. D, 25-31. RYABCHIKOV, I. D. & BOETTCHER,A. L. 1980. Experimental evidence at high pressure for potassic metasomatism in the mantle of the earth. Am. Mineral. 65, 915-19. & GREEN, D. H. 1978. The role of carbon dioxide in the petrogenesis of highly potassic magmas. In: Problems of Earth's Crust and Upper Mantle, Trudy Instituta Geologii GeofizikL So An SSR 403, Nauka, Novosibirsk (in Russian). , SCHREYER, W. & ABRAHAM, K. 1982. Compositions of aqueous fluids in equilibrium with pyroxenes and olivines at mantle pressures and temperatures. Contrib. Mineral. Petrol. 79, 80-4. SCHAIRER, J. F. & YODER, H. S. 1970. Critical planes and flow sheet for a portion of the system CaOA1203-SiO 2. Yearb. Carnegie lnst. Washington, 68, 202-14. , TILLEY, C. E. & BROWN, G. M. 1969. The join nepheline-diopside-anorthite and its relation to alkali basalt fractionation. Yearb. Carnegie Inst. Washington, 66, 467-71. , YAGI, K. & YODER, H. S. 1962. The system nepheline-diopside. Yearb. Carnegie lnst. Washington, 61, 96-8. SORENSON, H. (ed.). 1974. The Alkaline Rocks, 622 pp. Wiley, New York. TAKAHASHI, E. 1980. Melting relations of an alkali olivine basalt to 30 kbar, and their bearing on the origin of alkali basalt magmas. Yearb. Carnegie Inst. Washington, 79, 271-6. - & KUSHIRO, I. 1983. Melting of a dry peridotite at high pressures and basalt magma genesis. Am. Mineral. 68, 859-79. TAYLOR, D. & MACKENZIE,W. S. 1975. A contribution to the pseudoleucite problem. Contrib. Mineral. Petrol. 49, 321-33. THOMPSON, R. N. 1972. The one atmosphere melting patterns of some basaltic volcanic series. Am. J. Sci. 272, 901-32. - - 1 9 7 3 . One atmosphere melting behaviour and nomenclature of terrestrial lavas. Contrib. Mineral. Petrol. 41, 197-204. - 1974. Primary basalts and magma genesis. I. Skye, North-West Scotland. Contrib. Mineral. Petrol. 45, 317-41.
Primary basalts and magma genesis. III. Alban Hills, Roman comagmatic province, central Italy. Contrib. Mineral. Petrol. 60, 91-108. TURNER, F. J. & VERHOOGEN, J. 1960. lgneous and Metamorphic Petrology, 694 pp. McGraw-Hill, New York. WADE, A. & PRIDER, R. T. 1940. The leucite-bearing rocks of the West Kimberley area, Western Australia. Q.J. geol. Soc., Lond. 96, 39-98. WATERS, A. C. 1955. Volcanic rocks and the tectonic cycle. Geol. Soc. Am. Spec. Pap. 63, 703-722. WENDLANDT, R. F. & EGGLER, D. H. 1980a. The origins of potassic magmas. 1. Melting relations in the systems KA1SiO~-Mg2SiO4-SiO2 and KA1SiO4-MgO-SiO2-CO2 to 30 kilobars. Am. J. Sci. 280, 385-420. - & 1980b. The origins of potassic magmas. 2. Stability of phlogopite in natural spinel lherzolite and in the system KA1SiO4-MgO-SiOz-H20CO2 at high pressures and temperatures. Am. J. Sci. 280, 421-58. WIalTELEY, C. 1981. Phase relations in the system NaA1SiO4-KA1SiO4-CaA12Si2Os-SiO2-HzO at 1 kb. Prog. Exp. Petrol., NERC Publ. Ser. D, 33-4. WILLIAMS, H. 1936. Pliocene volcanoes of the NavajoHopi county. Geol. Soc. Am. Bull. 47, 111-72. WYLLIE, P. J. 1975. The earth's mantle. Sci. Am. March 1975. - - 1 9 7 7 . Mantle fluid compositions buffered by carbonates in peridotite-CO2-H20. J. Geol. 8 5 , 187-207. - - 1 9 7 8 . Mantle fluid compositions buffered in peridotite-H20--CO2 by carbonates, amphibole and phlogopite. J. Geol. 86, 687-713. - - 1 9 7 9 . Magmas and volatile components. Am. Mineral. 64, 469-500. - 1980. The origin of kimberlite. J. geophys. Res. 8 5 , 6902-10. - & HUANG, W. L. 1975. Peridotite, kimberlite, and carbonatite relations in the system CaO-MgOSiO2-COz. Geology, 3, 621-4. - - & 1976. Carbonation and melting reactions in the system CaO-MgO-SiO2-CO2 at mantle pressures with geophysical and petrological applications. Contrib. Mineral. Petrol. 54, 79-107. YODER, H. S. 1975. Relationship of melilite-bearing rocks to kimberlite: a preliminary report on the system akermanite-CO2. Phys. Chem. Earth, 9, 883-94. -1976. Generation of Basaltic Magmas, 265 pp. National Academy of Sciences. Washington, DC. - 1979. Melilite-bearing rocks and related lamprophyres. In." YODER, H. S. (ed.) The Evolution of Igneous Rocks. Fiftieth Anniversary Perspectives, pp. 391-411. Princeton University Press, Princeton, NJ. - & KUSrtIRO, I. 1972. Composition of residual liquids in the nepheline-diopside system. Yearb. Carnegie Inst. Washington, 71,413-16. ZYRYANOV, V. N. 1981. Phase Correspondence in the Systems of Feldspars and Feldspathoids, 217 pp. Nauka, Moscow (in Russian). - - 1 9 7 7 .
A. D. EDGAR, Department of Geology, University of Western Ontario, London, Ontario, N6A 5B7, Canada.
Nephelinites and carbonatites M. J. Le Bas S U M M A R Y : Carbonatites found in strongly alkaline intra-plate petrographic volcanic provinces are associated with olivine-poor nephelinites and with phonolites. Olivine-rich nephelinites occur in basanitic and alkali basalt provinces, normally without carbonatites. These igneous provinces are marked by epeirogenic crustal uplift. Nephelinites, ijolites and carbonatites form discrete magmatic events within individual complexes and correspond to the silicate-carbonate conjugate immiscible liquids observed in the laboratory. Carbonate liquids of variable alkali content can separate from both nephelinitic and phonolitic liquids. The silicate liquids give rise to pyroxenites, ijolites, nepheline syenites and nephelinitic pyroclast-rich strato-volcanoes. The carbonate liquids lose alkalis and fractionate to s6vite, alvikite and ferrocarbonatite, each with increasing incompatible-element content. Further fractionation of carbonatite magma can produce mineralizing fluids rich in rare-earth elements, F, Ba, U and Th. Dolomite carbonatite forms only in the deeper parts of carbonatitic complexes, perhaps at depths greater than 2 km. Explosive carbonatite volcanism can occur giving widespread carbonate tufts, rarely with lavas. Fenitization characterizes ijolite-carbonatite intrusive complexes. Syenitic fenites containing alkali feldspars, sometimes perthitic, are formed as aureoles 500 m wide around ijolites in granitic terranes, with nepheline syenite commonly formed as a contact-reaction rock. The feldspar-rich syenitic fenite aureoles which develop around the early carbonatites usually contain pure orthoclase or pure albite.
Carbonatites achieved geological respectability only about 30 years ago, largely through the impact of work in Scandinavia and Africa. Br6gger's work (1921), at Fen in S Norway, which was strongly disputed by Bowen (1924) but later re-established by Saether (1957), demonstrated that calcite-rich rocks could have all the characteristics of intrusive magmatic bodies and that they alkali metasomatize (fenitize) their host rocks. In Sweden, von Eckermann (1948) conducted a meticulous study of the Aln6 complex. He maintained that alkali carbonatite magma was the prime intrusive material at Aln6, that the associated alkaline silicate rocks were all metasomatic products of fenitization and that the calcite-rich carbonatite was the residue of the intrusive carbonate magma. Dixey et al. (1935) were the first to identify carbonatites in Africa. Garson re-investigated the Malawi complexes of Chilwa, Tundulu, Kangankunde, Songwe and others, and wrote a series of memoirs culminating in a valuable summary account (Garson 1965). The association of carbonatites with nephelinitic volcanism was first recognized by King (1949) at the Napak complex in Uganda. Tuttle & Gittins (1966) and Heinrich (1966) wrote comprehensive accounts of intrusive carbonatites, but extrusive carbonatites have only recently become widely appreciated although they were first identified in 1960 by Dawson (1962).
Strongly alkaline igneous rocks are notorious for their diversity of rock type and abundance of names. Some names, such as ijolite and urtite, are necessary, but many can be discarded. The names used here are based on the work of the IUGS Subcommission on the Systematics of Igneous Rocks (Streckeisen 1976, 1978). Carbonatites are defined as igneous carbonate rocks with more than 50% modal carbonate minerals. If the carbonate is calcite, the rock is a calcite carbonatite which is called a s6vite if it is coarse grained with adcumulate texture or a micro-s6vite if it is medium or fine grained. Beforsite is equivalent to dolomite carbonatite. A less common variety of carbonatite is alvikite, which typically has a texture of small calcite rhombs packed together and is enriched in incompatible elements such as rare-earth elements (REE), Ba, Mn, Zn and commonly Nb. The boundary between s6vite and alvikite is still uncertain but may be set at 0.4% MnO, 1500 ppm Ba and 2000 ppm REE. The calcite of alvikite is slightly ferroan. Ferrocarbonatite is a ferroan calcite to ankeritic calcite carbonatite even more strongly enriched in some or all of REE, Ba, Mn, Fe, Zn, F and U, with possible lower limits of 1.0% MnO, 5000 ppm REE and 5000 ppm Ba. This rock type is very rare, but it is the main source for mineralization of these elements in carbonatites. Natrocarbonatite is carbonatite lava. It is fine grained, extrusive and composed of nyerereite ((Na,K)2Ca(CO3)2) and gregoryite
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 53-83.
53
54
M. J. Le Bas
((Na,K)2CO3). Some natrocarbonatites are richer in Ca and have calcite in place of gregoryite. The purpose of this contribution is to give a petrological synthesis of the nephelinite-carbonatite complexes, at both plutonic and volcanic levels. It begins with the rise of the primary magma through the uppermost mantle, then considers the processes suffered by the magma when it stops in reservoirs at or near the base of the crust and finally traces the passage of the different magmas, carbonate and silicate, through the crust to the ultimate products at the surface. Kimberlitic carbonatites are not included because they have recently been reviewed by Dawson (1980) and because they are different from nephelinitic carbonatites in their primary lack of alkalis, particularly Na. This lack is expressed by the absence of fenitization around the kimberlitic carbonatites. However, the zonation of the E African and E Siberian ultraalkaline provinces, recently compared by Le Bas (1986), indicates that these two types of carbonatite are genetically related in their mantle sources.
Distribution Although nephelinite-carbonatite complexes are rare, being only a fraction of 1~ of all igneous rocks, their distribution is worldwide (Fig. 1) with ages back to the early Proterozoic. Nephelinite-carbonatite complexes occur in intra-plate environments, both continental and, rarely, oceanic. For instance they occur in the Canary 150 r
I
120 I
I
90 ~
89
60 ~f",./
I
30 T
I
Islands and the Cape Verde Islands, where there is a clear regional association with ocean island basalts. Petrographically there is no distinction between the complexes that occur singly, for example as at Kaiserstuhl in S Germany, and those in groups such as the complexes composing the 20 Cenozoic carbonatitic volcanic centres of the inner portion of the E African province (Le Bas 1980). The E African province (Fig. 2) is zoned, with kimberlites in a central core about 300 km across. Carbonatites, olivine-poor nephelinites and ijolites, but not basalts, comprise the next inner concentric zone 100-300 km wide, and at the edge is a carbonatite-free outer zone ranging in width from 100 to 300 km which is composed of basanites, alkali basalts and some olivine-rich nephelinites (Fig. 3). The large Palaeozoic-Mesozoic E Siberian province is similarly zoned (Dawson 1980; Le Bas 1986). In the Angola province, or half-province if S Brazil is rejoined to it, the distribution is different. There the alkali basaltic centres cluster near the coast line, with carbonatitic complexes further inland and kimberlite diatremes furthest inland. Namibia is similar (Prins 1981). The scale of these distributions suggests deep-seated mantle processes. They are not related to plate motion or plume traces. These alkaline and ultra-alkaline provinces are also marked by epeirogenic uplift (Le Bas 1971, 1980; Baker 1987) which produces broad swells and plateau uplift both continental (e.g. the Rhine upwarp and the East African plateau) and oceanic (e.g. Cape Verde Rise). These uplifts
0 f
I
I
30 C~-/I
60 I
I
90 I
r~.....~,r~J
120 ]
150 [
I
180 I
9
b
~
9
"
';~.
~
FIG. ]. World distribution of nephelinite-carbonatRe complexes: I I , r215
or province of comple•
[
Nephelinites and carbonatites "J,.
SUDAN
..""
".... ...... .'" ~ ,~v
"9 ~ X \"
/
fl
''..
"-
55
"''..
A
ETHIOPI/~
A'"
'/ U "~'.
.....
'^~ "~/
-.""~/ "~ANDA u~
/ U ' ~ IIAI / ." U~ I " 1 ~ As~" ,I i
. /.. /."~
0*
-
/
K
i
;1:/\. - u3-'~."\ -r
~
\\
-/:/
",
KENYA
u
__-/~.'\- " :.
/
I 2 ~ i
TANZANIA
9
~
'
"q'/~'L I I I 3;"~I
"".
IIi1
A " ""
~
~ I~])
Ill
,
\\ \ I' 3\~ j
\
-/
~ ~'~ ~
/
. . . . ,' ,,I~176
,J
' 3 ""---
\
K
;-
-
'u\
,,
\\
L~"~
u)
I
\
x
" . . .A
K'
i
"'
(
\" L~ 250 km "
]'~
\
40~' t
FIG. 2. Map of the E African Alkaline Province (or Lake Victoria Province) showing the zonation ( - - - ) : 1, central zone of kimberlite diatremes (K); 2, inner zone of ultra-alkaline complexes (U); 3, outer zone of alkali basaltic volcanism (A). Radial lines indicate the edge of the E African uplift. Bold lines indicate rift faults. The faults show no apparent relation to the zonation. The inset square marks the area shown in Fig. 3. (RW, Rwanda; BU, Burundi.)
evidently relate to the same deep-mantle processes as those giving rise to the ultra-alkaline magmatism (Bailey 1987). Such up-arching produces tension in the upper crust which is quite independent of the motion of the whole tectonic plate. The two motions, the first vertical, covering an equidimensionat area and emanating from the mantle associated with ultra-alkaline magmatism, and the second essentially horizontal and related to plate motion, are interpreted to be distinct and not necessarily related (Le Bas 1986). It is argued that ultra-alkaline magmatism and regional uplift are not necessary initial stages of the Wilson cycle which describes the break-up of continents followed by sea-floor spreading.
Nephelinite magma There are nephelinites and nephelinites, the two varieties being distinguished by the properties given in Tables 1-3. Those defined as Group I
nephelinites are olivine rich, and are typically associated with alkali basalt and basanite volcanic provinces. Group II nephelinites are rarer and comprise olivine-poor nephelinites which occur with carbonatitic and ijolitic complexes (Fig. 3). The olivine-poor nephelinites commonly have abundant euhedral clinopyroxene phenocrysts, whereas the few olivine phenocrysts often show resorbed crystal margins, sometimes even reaction rims, indicating that olivine was no longer in equilibrium. The abundance of pyroxene usually means that these rocks are metanephelinites rather than nephelinites, and many could be better described as clinopyroxenephyric melanephelinites. Aphanitic varieties are rare. Olivine-poor nephelinite can fractionate to olivine-free nephelinite and to phonolite (Table 1). These phonolites are often glassy, commonly making abundant dykes and plugs, and Lippard (1973) distinguished them (his Gwasi-type phonolites) from the phonolites in and around the
56
M. J. Le Bas 134o +
--
+
r
t 3 8~
4-
ETHIOPIA
+ + +
,~
+
~1 40 ~
4- •
LODWAR
~t~
MOROTO
z
--2 0
NAPAK
~
KENYA
C e n o z o i c l a v a s and pyroclastic deposits
x +
NYAMBENI
I&UISHU
~ Kavirondo ~ M T .
~ A
KENYA
Olivine-poor volcanic A Olivine-rich volcanic
ABERDARES
nephelinitic centres nephelinitic centres
AL
Mixed nephelinitic volcanic centres
SJNGIR I
~
~
NAIROBI Z o n e of o v e r l a p of p r o v i n c e s
Rift
faults
SHOMBOLE
( L.N~tro~ OLDOINYO LENGA :::::K?IMASI
KILIMANJARO
SADIMAN 9 100 L
1
i !
I
ESSIMINGOR 4o
TANZANIA 134~
~
HANANG
~
1380
FIG. 3. Map of Kenya and adjacent areas showing the present distribution of the Group II olivine-poor nephelinitic (and carbonatitic) volcanic centres corresponding to a portion of the inner zone shown in Fig. 2, and the Group I olivine-rich nephelinitic (non-carbonatitic) volcanic centres corresponding to the Kenyan region of the outer zone shown in Fig. 2. Between these two zones lies a zone of mixed nephelinitic types resulting from an overlap or mixing of the inner- and outer-zone magmas.
Kenya rift valley (his Plateau and Kenya types). The latter are derived from basanitic and basaltic sources, and form extensive lavas evidently of low viscosity. The phonolites derived from the nephelinites are rich in Sr (more than 1000 ppm) and Ba (more than 100 ppm, and some more than 1000 ppm), and are richer in total alkalis (about 14%) relative to silica (about 53%) than those derived from plagioclase-bearing m a g m a s w h e n both Sr and Ba contents are less than 100 ppm.
The Miocene lavas of the Kisingiri volcano in W Kenya provide a good example of the normal fractionation trend from olivine-poor nephelinites and they are not dissimilar from those at other centres nearby except in detail. The early stages of fractionation are dominated by the precipitation of m u c h clinopyroxene plus a few oxides and some olivine at first (Fig. 4), producing only a m i n i m a l rise of Si in the m a g m a with fractionation. On some occasions Si falls, partic-
Nephelinites and carbonatites
-
~ e ~6
~ --'lk
57
\
,," /
~.
~09o'~0
9 9
10 : /
I
~
2.. 3
\ 1 phonolites
/
0
e ~.Nx :
l nephelinites
t-~
I
t
I
10
I
t
t
i
wt
I
20 %
t
I
t
t
I
30
I
I
[
40
MgO
FIG. 4. CaO-MgO plot of fields for olivine-rich nephelinites (Group I) and olivine-poor nephelinites or melanephelinites (Group II), and their mutual fractionation product of ordinary nephelinites and phonolites. • is the composition of olivine fractionated; + - - + are successive compositions of clinopyroxenes fractionated from melanephelinites, nephelinites and phonolites. Dotted lines are paths of fractionation, and to the right of the fields are crystal extracts with the pyroxene-to-olivine ratio indicated. (Data from Le Bas (1978), Varne (1968), Saggerson (1970), Spencer (1969), Strong (1972) and EAGRU data file (unpublished.)
ularly if no oxides are precipitated, and melilite nephelinite magma can form. Some data are given in Table 1, and further data can be obtained from the U K - I G B A data base which can be accessed via BGS (London) or the World Data Center-A (Colorado). They show that the production of phonolite from olivine-poor nephelinite is accompanied by a fall in Ca, Mg, total Fe, P and sometimes Ti, with a rise in Si, A1, Na, K and the ratios Fe/Mg and K/Na. The fractionation of trace elements from olivine-poor nephelinites can be shown on twoelement plots, but is best seen on primordial mantle-normalized plots (Figs 5 and 6). Figure 5 is a plot of four samples from Kisingiri (data in Table 1) and shows a marked relative increase in the incompatible elements Th, Nb and Zr, with depletion in Sr, P, Ti and V; the latter is pulled out by the precipitation of pyroxene, apatite and any titanomagnetite. The increase in the ratio Zr/Ti appears to be a sensitive index of fractionation, and a similar rotation of the Zr-Ti tie-line (i.e. Zr/Ti increases) is also seen from the trace data from other complexes. Figure 6 compares melanephelinites of carbon-
atitic complexes of different ages in E Africa and reveals that, while the high-field-strength elements Nd, P, Zr, Ti and Y remain closely similar, the others vary slightly, particularly the largeion-lithophile (LIL) elements. The younger complexes of Homa Mountain and Elgon both show a slight enrichment in Ba and Th, and Ba mineralization is seen at Homa (Le Bas 1977, p. 237). If data ranges were marked on Fig. 6(a), the values of Rb, Sr, Nb and REE would show overlap, except in the case of Homa Mountain where all the data consistently show higher REE and lower K values. This is not instrument bias since all the data were obtained on the X-ray fluorescence (XRF) spectrometers at both Edinburgh and Leicester University. If the Kisingiri and N a p a k averages plotted in Fig. 6(a) are taken as normal for nephelinitic rocks associated with carbonatites (but whether these are nephelinites from before or after possible liquid-immiscible separation from carbonate melt is not yet certain), they should be chemically comparable with other carbonatitic nephelinites. Appropriate data are sparse, but they do compare closely with the Triassic nephel-
58
M. J. Le Bas
TABLE 1. Selected nephelinitic lavas from carbonatitic and other provinces Elgon, WKenya
Kisingiri, W Kenya
Napak, EUganda
RR659 RR585 RR457 U1041
SUN6 SUN89 SUN21
ElK8
ElK6 ElK7
54.73 0.47 21.37 5.13 0.23 0.21 2.70 9.74 5.72 0.16
43.50 2.03 10.30 15.03 0.18 9.18 16.94 1.26 1.68 0.58
42.84 2.77 11.47 15.46 0.29 7.16 15.33 2.76 1.27 0.99
42.20 44.12 2.48 2.61 12.04 13.89 1 3 . 9 7 14.11 0.27 0.25 6.22 5.03 1 5 . 1 7 11.42 3.93 5.52 2.54 2.78 0.88 0.95
101.27 100.46
100.68
97.76
97.23
100.35
99.68
100.68
82 92 333 31 6 749 1098 46 49 88 38 135 18
19 121 291 54 7 1251 1167 67 69 132 50 184 25
12 133 260 55 8 1212 1267 73 74 133 56 206 26
29 117 286 60 15 1857 1946 163 96 154 57 271 36
29 114 211 45 12 1211 910 111 92 158 61 193 25
30 121 218 58 12 1128 684 123 94 157 57 214 28
Major elements (wt.%) SiO 2 TiO 2 AI20 3 Fe203 T MnO MgO CaO Na20 K20 P205
Total
41.99 3.08 12.94 15.22 0.24 7.43 13.65 3.46 1.58 0.89
45.42 49.73 2.56 1.95 1 4 . 8 8 16.72 13.46 10.62 0.23 0.22 4.39 2.36 9.17 6.75 7.26 8.35 2.80 3.82 0.93 0.75
100.48 101.10
40.38 42.56 2.52 2.48 1 1 . 6 8 13.62 1 6 . 2 7 14.48 0.25 0.23 6.27 5.52 1 2 . 5 3 10.76 4.42 4.58 2.33 2.50 0.81 0.79
Trace elements (ppm) Ni Zn V Rb Th Ba Sr Nb La Ce Nd Zr Y
47 114 323 40 12 1862 3135 119 105 194 79 284 32
25 119 238 63 11 1161 1342 111 92 174 70 317 29
9 129 143 81 15 1401 1448 132 95 161 132 362 30
4 168 52 141 56 1139 808 336 74 135 40 816 24
RR659, aphanitic melanephelinite lava, top of Kaniamwia, Kisingiri (Leicester XRF); RR585, nephelinite lava, Rangwe, Kisingiri (Leicester XRF); RR457, nephelinite dyke, with katophorite, Kiawindu, Kisingiri (Leicester XRF); U 1041, phonolite lava, trachytic texture, Nyamaji, Homa Bay (Leicester XRF); SUN6, pyroxene-phyric melanephelinite lava, Napak (wet analysis by R. C. Tyler 1966; traces Edinburgh XRF); SUN89, melanephelinite lava, Napak (Leicester XRF); SUN21, aphanitic melanephelinite lava, Okathim (Edinburgh XRF); ELK8, melanephelinite lava above Kitum Cave, E flanks Mount Elgon (Leicester XRF); ELK6, melanephelinite lava, Kitum Cave, E flanks Mount Elgon (Leicester XRF); ELK7, nephelinite lava, Kitum Cave, E flanks Mount Elgon TABLE 2. Worm average nephelinites
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO Na20 K20 P205 H20
Group I (olivine rich) (wt.% (SD))
Group II (olivine poor) (wt.~(SD))
39.8 (1.5) 3.0 (0.3) 11.3 (0.9) 5.1 (1.9) 7.9 (1.2) 0.2 (0.01) 12.3 (1.8) 12.8 (1.3) 3.4 (0.3) 1.1 (0.3) 0.9 (0.6) 2.1 (0.6)
41.5 (2.0) 2.8 (0.3) 11.4 (1.6) 7.4 (1.5) 6.0 (1.0) 0.2 (0.01) 8.1 (2.0) 13.2 (3.2) 3.7 (1.9) 1.8 (1.5) 0.8 (0.6) 3.0 (1.1)
99.9 n=81
99.9 n--97
After Le Bas 1978.
inites of the Karoo province with w h i c h carbonatites are associated (Bristow & Saggerson 1983). The Miocene nephelinites of the southern portion (south of the equator) of the K e n y a rift valley, particularly those in the Kishalduga region, are almost identical (cf. Figs 6(a) and 6(b), and have olivine contents m i d w a y between the rich and poor types defined above (Kishalduga data from R. Tarzey, personal communication). Miocene nephelinites of N K e n y a are m a r k e d l y poorer in K and the more incompatible trace elements. N o olivine-free nephelinites have yet been observed in the northern rift. The few data that are available from the olivine-rich nephelinites of the Massif Central in France and of Greenland, Hawaii and the Comores Islands are all closely similar to the carbonatitic nephelinites, apart from variable K / T h ratios and Rb and K contents.
Nephelinites and carbonatites
Homa Mt., W. Kenya
N Kenya rift
S Kenya rift
59
HF566
HC42
11.108
8.428
15.591
17.548
Massif central 42466
Hawaii
Comores
65KAPAAI 1
Moh24
45.17 2.94 11.16 12.06 0.13 6.48 11.81 2.84 1.75 0.79
53.28 0.59 21.14 4.23 0.16 0.48 3.33 9.97 4.82 0.18
41.71 2.23 11.52 12.22 0.18 12.72 12.64 3.06 0.77 0.41
42.84 2.13 12.53 11.74 0.18 10.64 12.01 3.99 1.27 0.49
42.78 4.05 10.09 15.13 0.20 9.13 15.79 1.62 1.16 0.73
43.46 3.73 9.82 15.49 0.21 8.54 14.67 2.81 1.36 0.68
43.61 3.62 8.80 11.08 0.15 13.44 11.45 2.38 3.66 0.82
40.15 2.55 11.45 14.40 0.22 12.77 12.10 3.51 0.77 0.80
41.37 2.71 11.49 13.89 0.24 11.03 12.10 3.39 1.56 1.00
95.12
98.18
97.45
97.81
100.69
100.76
99.01
98.67
98.78
37 113 350 46 10 4668 1213 105 92 170 71 249 25
4 120 62 104 33 1224 1153 143 103 151 37 490 26
214 78 317 9 7 551 604 69 56 108 39 147 23
180 83 297 17 10 697 685 80 73 135 49 166 24
135 109 357 45 13 623 915 128 105 202 85 345 33
121 114 344 47 13 651 1075 116 96 183 74 335 29
309 80 -78 9 824 764 93 79 174 73 409 23
340 107 290 20 6 780 1050 -53 110 56 171 --
168 127 -47 -1250 880 ----257 --
(Leicester XRF); HF566, pyroxene-phyric melanephelinite plug, N flanks Homa Mountain (Edinburgh XRF); HC42, feldspar-phyric phonolite plug, S flanks Homa Mountain (Edinburgh and Durham XRF); 11.108, olivine nephelinite lava, Kasorogol, S Turkana (Edinburgh XRF); 8.428, olivine nephelinite, Nasaken, S. Turkana (Edinburgh XRF); 15.591, olivine nephelinite, S. Nakuru (Edinburgh XRF); 17.548, olivine nephelinite, Kishalduga, Narok (Leicester XRF); 42466, olivine nephelinite, Cantal, France (Downes 1984); 65APAAll, olivine nephelinite, Honolulu (Clague & Frey 1982); MOH24, olivine nephelinite, Moheli Island, Comores (Strong 1972). T h e similarities of the trace-element signatures in Fig. 6 and Table 1 indicate broadly similar chemical sources for the m a g m a s in the mantle, w h e t h e r or not carbonatites are associated. A c h a r a c t e r c o m m o n to all nephelinites is the e n r i c h m e n t in K , Th, N b and the light R E E w h i c h points to a m a n t l e source rich in the heatproducing elements, although the K / T h ratio does vary, indicating a variable m a n t l e source for these components. A further study of the detailed variation in nephelinitic and other primitive m a g m a s across the zoned alkaline igneous province of E Africa, c o m p a r i n g early Cenozoic with later Cenozoic m a n t l e m a g m a t i c products, is being carried out at present by R. Tarzey. A similar investigation of the Cape V e r d e m a g m a tism is being u n d e r t a k e n by N. Hodgson. The olivine-rich and olivine-poor nephelinites, in terms of their m a j o r - e l e m e n t chemistry, could
be derived from e a c h other, but their geographical separation by some h u n d r e d s of kilometres into the c o n t e m p o r a n e o u s outer and inner zones of the E A f r i c a n province shows that distinct processes occurred in the two cases.
Magma plumbing These two types of nephelinites are distinct chemically and physically (Table 3). The olivinerich nephelinites have high Mg and N i contents a n d few or no fractionation products. T h e y build small volcanic cones with an occasional abund a n c e of lherzolite nodules, w h i c h m a y be t a k e n to indicate that these m a g m a s c a m e directly from the m a n t l e with few or no intervening crustal reservoirs. In strong contrast, the olivine-poor nephelin-
M. J. Le Bas
6o
~
U
1041
RR 4 5 7
-----+
10 3
"
..
."
~
"...
102
o
i ...,.
101
KISINGIRI
LAVAS
W. K E N Y A
100
I
Rb
I
I
I
1
I
I
I
I
I
K Th Ba Sr Nb La Ce Nd
I
P
1
Zr Ti
I
Y
TI
V
I -
Zn
FIG. 5. Primordial mantle (PM) normalized plot of four selected lavas from the Kisingiri strato-volcano (Table 1). Note the decreasing coherence of distribution, particularly of the high-field-strength elements during fractionation. The PM values of Wood et al. (1979) were used; V was taken as 84 ppm.
ites make large central strato-volcanoes and almost never carry mantle-type xenoliths. They exhibit strong fractionation to phonolite and ijolite, which suggests that magma produced in the mantle must have stopped in a magma reservoir during its upward passage to the surface. This is corroborated by the low Mg number and Ni content of magmas that reach the surface. The question arises: why should one type of nephelinite consistently stop on its rise through the continental lithosphere, and the other not? Trace elements do not support the argument of mantle chemical heterogeneity as the cause. Therefore, either the physical conditions of passage of the two differ, i.e. one portion of the lithosphere is easier to penetrate than another, or the physical properties of the two types of magma must differ. No lateral variation of the lithosphere is apparent in any of the provinces and therefore the second alternative must be examined. The density of primitive nephelinite magma would be about 2.9-3.0 g cm- 3, whilst the density contrast across Moho from mantle to continental crust is about 3.4-2.8 g cm-3. Thus a nephelinite magma would be expected to stop rising in the lower continental crust, but perhaps not in the lower oceanic crust (density about 3.0 g cm-3). Why then do nodule-bearing nephelinites pass rapidly up through continental crust without stopping ? The answer may lie in the volatile content. Nephelinites associated with carbonatites by liquid immiscibility must be carbonated nephelinite magmas and hence have a high C O 2 / ( H 2 0 -kCO2) volatile fraction. If, however, the COz/ ( H 2 0 + C O 2 ) ratio is lower, the nephelinite magma would be more hydrous and less viscous
500
I.
/ I "t,,, /
~o
200
\\
/_
,R,
,?-~
i~
,
-
I00 o >o
q
d
50
1:;7 ;;2~ 20
-~0 x)\\_
i
]
I
L
i
I
i
I
b K Th Ba Sr Nb La Ce Nd P Zr Ti
Y
9 S Kenya rift
L-~'XX
[] N Kenya rift
\
-
'~
o Kisingiri (Miocene) I
-
s
9 Napak (Miocene)
I
[]
I0
Rb K Th Ba Sr Nb La Ce Nd P Zr Ti
Y
FIG. 6. (a) Normalized plot of average trace-element abundances for melanephelinites from four Tertiary carbonatitic volcanic complexes in E Africa (Fig. 3); (b) normalized plot of averaged values of trace elements in olivine nephelinites from the northern and southern sectors of the Kenya rift valley.
Nephelinites and carbonatites
6i
TABLE 3. Contrasts between nephelinites of Groups I and H Character
Principal phenocrysts Clinopyroxene MgO content of rock Mg no. (100Mg/(Mg+Fe2+)) at 100 K/(Na + K) at Associated basalts Chemical fractionation Volcanic structure Percentage pyroclastic Plutonic equivalents Mantle-type xenoliths Examples
Group I (olivine-rich nephelinites) Olivine and clinopyroxene Lilac and Ti rich > 10 wt.% 74 18 Abundant Very slight Small parasitic Moderate Rare Common Kenya Rift Comores Islands Hawaiian Islands Texas Massif Central
and would perhaps be able to rise through the lithosphere without stopping. Some dynamic fractionation during ascent may also have taken place. Now that the style of nephelinitic magmatism associated with carbonatites has been distinguished from other nephelinitic developments, the term nephelinitic volcanism in the remainder of this contribution will refer to the olivine-poor nephelinites and their derivatives. During or before retention in a magma reservoir, olivine was precipitated and hence largely lost to the remaining nephelinitic magma. Resorption and reaction rims on remaining olivine phenocrysts indicate that they were no longer stable in the magma. Augite was also being precipitated in abundance. It seems that the initial precipitation of olivine increased the already high CO2 content in the magma, which eventually destabilized any remaining olivine and widened the pyroxene stability field. Cumulate bodies of dunite within carbonatite complexes are rare (e.g. Kovdor, Kola, and Shawa, Zimbabwe), whilst alkali pyroxenites occur in most carbonatitic complexes.
Pyroxene fractionation Well over half the extrusive products of nephelinite volcanism are pyroclastic, indicating a volatile-rich source. Among the products in W Kenya are numerous megacrysts of clinopyroxene up to 3 cm long. The chemistry of these and other clinopyroxenes in nephelinitic rocks is revealing
Group II (olivine-poornephelinites (melanephelinites)) Clinopyroxene and rarely olivine Pale brown or greenish and Ti poor < 10 wt.% 67 24 None Strong Large central High Frequent Very rare W Kenya E Uganda Malawi Siberia Amba Dongar, India
in the deciphering of the upward passage of the magma. Table 4 gives the analyses of a range of pyroxenes as xenocrysts, phenocrysts and groundmass phases from melanephelinites, nephelinites and phonolites. Most of them have low Ti and A1 contents which, if they came from groundmass phases, would be at variance with the higher Ti and A1 contents normally encountered in strongly alkaline rocks (Thompson 1974). Instead, as pointed out by Thompson, the low Ti and AI appear to indicate a high temperature of formation. The high Ca content of these magmas would also favour partition of the Ti into perovskite or sphene, which are both quite abundant as microphenocrysts in these rocks. Until analysed by microprobe, the presence of Cr-diopside among the xenocrysts and phenocrysts in the melanephelinites was not appreciated (Table 4, column 1). Cr-diopside is typical of pyroxenes from mantle-type lherzolites, and sometimes forms whole xenocrysts. These xenocrysts can also be present as colourless to palegreen cores to more normal pyroxene compositions (Table 4, columns 4a and 4b) and fall in the mantle pressure field for chromiferous pyroxenes described by Onuma & Tohara (1984) where Na > A1vI. The presence of Cr-diopside is thought to indicate that the nephelinitic magma does record a history of passage through upper-mantle lherzolite. Olivine, with more than 90% Fo, has not been recognized among the xenocrysts or phenocrysts in the lavas, and the few olivines present range from 86% to 81% Fo. Olivine with Fo > 90% does occur in the cumulates, however.
62 TABLE 4.
M. J. Le Bas
Analyses of clinopyroxenes from melanephelinites, nephelinites and phonolites 2a
2b
53.08 0.62 0.58 . 3.19 0.11 16.16 24.31 0.39 0.00 0.14 98.58
51.67 0.87 1.61 . 6.73 0.13 13.58 23.80 0.66 0.00 0.15 99.20
51.25(0.55) 1.06 (0.19) 1.84 (0.26) . . 4.63 (0.28) 0.05 (0.02) 14.83 (0.34) 24.08 (0.09) 0.41 (0.03) 0.02 (0.00) 0.28 (0.07) 98.45
51.94 49.95 0.71 1.95 1.53 3.09
Total
53.20(0.45) 0.45 (0.07) 1.00 (0.14) . . 2.93 (0.13) 0.06 (0.01) 16.85 (0.18) 23.43 (0.37) 0.48 (0.07) 0.00 0.86 (0.27) 99.33
48.51 1.56 4.49 3.19 4.34 5.44 4.81 0.05 0.05 0.14 15.12 14.02 13.58 23.81 24.55 22.40 0.47 0.40 0.96 0.01 0.02 0.00 0.44 0.01 . 98.42 99.47 99.64
46.55 47.80 50.45 50.41 2.19 1.37 0.80 0.82 5.04 4.82 2.30 2.27 4.01 -12.80 13.51 7.19 17.00 8.45 8.27 0.30 0.65 0.55 0.52 10.29 6.05 4.87 4.53 22.07 19.20 13.98 13.43 1.22 2.79 5.08 5.89 0.18 -0.00 0.34 . . . . 99.04 99.68 99.28 99.99
Si Ti A1 Cr Fe 3+ Fe 2+ Mn Mg Ca Na K Ca Mg Fe* Na Mg Fe z+ + M n Mg/(Mg + Fe*)
1.956 0.013 0.046 0.025 . 0.090 0.002 0.924 0.923 0.034 . 47.6 47.7 4.7 3.4 91.2 5.4 91
1.971 0.017 0.025 0.004 . 0.099 0.003 0.894 0.968 0.028 . 49.3 45.5 5.2 2.8 90.0 7.2 90
1.940 0.025 0.071 0.005 . 0.211 0.004 0.760 0.958 0.048 . 49.6 39.3 11.1 5.0 79.1 15.9 80
1.923 0.030 0.081 0.008 . . 0.145 0.002 0.830 0.968 0.030 . . 49.8 42.7 7.5 3.1 86.9 10.0 85
1.943 1.867 1.817 0.020 0.055 0.044 0.068 0.136 0.198 0.013 . . . 0.090 0.136 0.170 0.151 0.002 0.002 0.004 0.843 0.781 0.758 0.954 0.983 0.899 0.034 0.029 0.070 . 49.3 50.8 47.3 43.6 40.3 39.9 7.1 8.9 12.9 3.5 3.2 7.8 86.5 85.9 84.9 10.0 10.9 7.3 86 82 76
1.786 0.063 0.228 . 0.116 0.231 0.010 0.589 0.907 0.091 0.009 48.9 31.8 19.3 10.9 70.5 18.6 62
1
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO Na20 K20
Cr203
.
.
3
4a
4b
5
6
7
8
9
1.871 1.949 1.942 0.040 0.023 0.024 0.222 0.105 0.103 . . -0.372 0.392 0.557 0.273 0.266 0.022 0.018 0.017 0.353 0.280 0.260 0.805 0.579 0.554 0.212 0.381 0.440 --0.017 46.4 38.0 37.2 20.3 18.4 17.5 33.3 43.6 45.3 23.6 40.0 46.8 39.4 29.4 27.7 37.0 30.6 25.5 38 30 28
1, Cr-diopside xenocryst from strongly porphyritic olivine melanephelinite lava, Kisingiri volcano, W Kenya (RR710) (mean of six wavelength-dispersive microprobe analyses (WDMAs); standard deviation to 2tr given in parentheses); 2a, core of pale-green zoned pyroxene in block of pyroxenite ejected from Oldoinyo Lengai volcano, N Tanzania (24407) (energy-dispersive microprobe analysis (EDMA)); 2b rim of 2a; 3, euhedral megacryst (2 cm) of black diopside in agglomerate between melanephelinite lavas of Kisingiri volcano, Kenya (K79) (mean of four WDMAs; standard deviation given); 4a, core of colourless zoned Cr-diopside in phenocryst in normaltype melanephelinite lava, lower N flanks of Kisingiri volcano, Kenya (U16) (WDMA); 4b rim of 4a; 5, bulk pyroxene from fine-grained melanephelinite lava, S Ruri, W Kenya (S197) (wet analysis by H. Lloyd); 6, bulk pyroxene from nephelinite lava, S. Ruri ($33) (wet analysis by H. Lloyd); 7, Aegirine-augite phenocryst in feldspar-phyric phonolite lava of Kisingiri volcano, Wasaki Peninsula, W Kenya (U410) (EDMA); 8, Aegirineaugite microphenocrysts in glassy phonolite, volcanic plug, S Ruri (S 17) (wet analysis by H. Lloyd); 9, Aegirineaugite microphenocrysts in porphyritic phonolite, volcanic plug, S Ruri ($29/1) (wet analysis by H. Lloyd). Fe*, total F e + Mn. Titaniferous lilac-coloured Al-augite phenocrysts are exceedingly rare in these nephelinites, but are a b u n d a n t in the olivine-rich nephelinites and have been interpreted (Fodor et al. 1982) as being precipitated at pressures in the vicinity of 1520 kb during passage t h r o u g h the u p p e r mantle. Since Cr-diopside occurs but Ti-Al-augite is rare in the olivine-poor nephelinitic rocks, it is inferred that the latter were not precipitated during the passage of the p a r e n t m a g m a t h r o u g h the mantle. Similar chromiferous diopsides in pyroxenite (e.g. Table 4, c o l u m n 2a) are inter-
preted as cumulates, w h i c h further suggests that a c c u m u l a t i o n in a deep reservoir did occur. T h e more n o r m a l pyroxene phenocrysts in the m e l a n e p h e l i n i t e are a b u n d a n t low-Ti low-A1 augites (Table 4, columns 2b and 3) w h i c h zone out to more aluminous augites (Table 4, columns 4b and 5). They also show m o d e r a t e Fe enrichment, indicative of fractionation in a m a g m a chamber. Large euhedral black glistening megacrysts showing signs of resorption on the crystal faces also occur. T h e y have the same composition as
Nephelinites and carbonatites the phenocrysts (Table 4, column 3) and hence presumably have the same source. Evidently they were transported directly to the surface without stopping. The rare presence of pargasitic amphibole in some E African pyroclastic rocks (Varne 1968; Le Bas 1977) may be related to fractionation at mantle pressures as proposed by Varne, but they may also be the product of strongly hydrous conditions obtaining at only slightly lower pressures in a deep crustal magma chamber and the consequent explosive eruption of this magma to the surface. If this interpretation of the magmatic plumbing is correct, then the subvolcanic coarse-grained bodies of ijolite exposed by erosion at carbonatitic centres are products of nephelinitic magma rising from the lower crustal reservoir which stopped and crystallized. Therefore, the subvolcanic bodies would not mark the position of the main magma chambers feeding the volcanic structures above. Crystal fractionation of the melanephelinitic magma in the lower crustal reservoir produced nephelinitic and finally phonolitic liquids with increasingly Fe-rich sodic pyroxenes (Table 4, columns 6-9) as plotted in Fig. 7. Compositions richer in acmite than shown in
63
Fig. 7 do occur, but only rarely, and are confined to needles growing in late-stage fractionates. Also significant, with a bearing on the plumbing, is that some pyroxenes have green cores and show reverse zoning. They are quite common in melanephelinites and nephelinites. Table 5 lists analyses of three such zoned pyroxenes which have cores richer in Fe and Na than their colourless rims. The green cores correspond to pyroxenes from phonolites, while the rims correspond to the nephelinitic and melanephelinitic hosts. These give evidence for magma mixing, whereby pyroxene phenocrysts from phonolite are mixed into incoming magma to produce the reverse-zoned pyroxenes which can then be carried to the surface. Such a process necessitates a magma reservoir as proposed above. Analysis 3 in Table 5 further shows that mixing did not always lead to eruption, but that cumulative pyroxenites could be produced. Similar magma mixing revealed by green cores with reverse zoning is well known in many parts of the world. Brooks & Prinzlau (1978) give an excellent summary covering such mixed products from Antarctica, Africa, Europe, America and Greenland.
Crustal intrusive processes
\
~/
\
ee/
...:,.'.. Fe 2+ + Mn
FIG. 7. Na-(Fe 2+ + Mn)-Mg at.% plot of pyroxenes taken from Table 4 together with unpublished energy dispersive microprobe data on further E African samples: II, Cr-diopside from RR710 (Table 4, column 1); A, megacryst from K79 (Table 4, column 3); + - - +, zoned core-to-rim pale-green pyroxene in pyroxenite (Table 4, columns 2a and 2b); O, phenocryst and groundmass pyroxenes (the tie-line with the arrow indicates zoning); - - - , pyroxene trend for the sequence pyroxenite to ijolite to nepheline syenite in the sub-volcanic complexes of E Uganda (Tyler & King 1967) and W Kenya (Le Bas 1977, and unpublished data). The path from melanephelinite to nephelinite to phonolite corresponds to fractionation beginning with lowfO2.
Recent geochemical modelling by the author using major-element, trace-element and isotopic data shows that mixing of phonolite with melanephelinite magma can also produce lavas of mugearitic and trachyandesitic compositions. Such lavas are rare, but have been recognized in the nephelinitic volcanoes of E Africa (King 1949; Le Bas 1977). In the absence of any basalts in these volcanoes, they had been thought to be products of crustal contamination of melanephelinite magma. Another feature indicative of magma residing in a continental crustal chamber for long periods is that, the more extreme is the phonolitic fractionate from a nephelinite parent, the greater is the STSr/86Sr ratio. In W Kenya the 8781"/8651" initial ratios of melanephelinites are usually between 0.7032 and 0.7038 (Norry et al. 1980), whilst phonolites and trachyphonolites have values of 0.7046 and 0.7053 respectively (Le Bas 1977, p. 140). These higher values in the fractionates are interpreted as the product of crustal contamination. The density-contrast conditions cited above for arresting the upward passage of nephelinitic magmas near the mantle-crust boundary hold true only for continental crust. They cannot be applied to oceanic islands such as the Canary and
M . J. Le Bas
64
TABLE 5. Analyses of clinopyroxenes with green cores and reverse zoning 1
Core
2
Rim
SiOz TiOz A1203 FeO T MnO MgO CaO NazO CrzO 3 Total
51.12 0.52 1.96 15.50 0.36 8.47 19.44 2.21 . 99.58
50.80 1.50 2.19 6.50 0.12 14.36 23.42 0.45 . . 99.34
Si Ti A1 Cr Fe r Mn Mg Ca Na Na Mg Fe 2+ +Mn Mg/(Mg + Fe*)
1.972 1.903 0.015 0.042 0.089 0.098 . . . 0.500 0.204 0.012 0.004 0.487 0.802 0.804 0.940 0.165 0.033 16.2 3.4 47.8 82.3 36.0 14.3 57 85
3
Core
Rim
Core
Rim
47.56 1.53 5.12 17.90 0.39 5.32 19.98 1.88 . 99.68
48.32 2.15 4.67 7.58 0.12 12.53 22.58 1.02
51.26 0.37 1.07 15.69 0.35 8.31 20.36 2.09 0.16 99.66
50.43 0.87 2.19 10.11 0.22 11.45 22.88 0.76 0.19 99.10
1.987 0.011 0.049 0.005 0.509 0.011 0.480 0.845 0.157 15.8 48.2 36.0 57
1.924 0.025 0.098 0.006 0.323 0.007 0.651 0.935 0.056 5.9 67.9 26.2 72
98.97
1.866 0.045 0.237
1.830 0.061 0.208
0.587 0.013 0.311 0.840 0.143 15.4 33.5 51.1 40
0.240 0.004 0.707 0.916 0.075 8.3 78.5 13.2 86
.
1, Fine-grained nephelinite lava with occasional phenocrysts of pyroxene, mostly with green cores and colourless rims, showing reversed zoning, Kisingiri volcano, Kenya (K61) (EDMA); 2, hybrid fine-grained lava with scattered large phenocrysts of phonolitic pyroxene rimmed by groundmass composition pyroxene of melanephelinite host, NE flanks of Kisingiri volcano (U926) (EDMA); 3, apatite pyroxenite block from Oldoinyo Lengai volcano, N Tanzania (24420) containing subhedral pyroxenes with green cores and pale-green margins (EDMA). Cape Verde Islands, where nephelinites and carbonatites occur and where the distinction between the olivine-rich and olivine-poor nephelinites is not so clear. Green pyroxene cores exist, and so do mantle nodules. If a deep crustal reservoir is lacking, then fractionation must take place during ascent or in high-level chambers such as those now occupied by the subvolcanic plutonic complexes where fractionation and segregation can be seen to have occurred. A final question may be asked concerning the stage and depth at which magma chambers are formed in continental or oceanic crust. What relation do these feeders have to the onset of liquid immiscibility which, as will be shown below, affected the carbonated nephelinite magma and permitted the formation of conjugate silicate and carbonate magmas? If the unmixing of liquids is an exothermic process, its onset together with loss of carbonate fraction from the silicate liquid would temporarily raise the liquidus temperatures of the magmas until each had precipitated its more refractory phases, mainly
olivine and/or pyroxene +oxides and a p a t i t e + oxides respectively, until the remaining liquids were again buoyant. In summary, the following magma plumbing is envisaged. A carbonated olivine-rich magma is derived by partial melting of the asthenosphere. This rises through the mantle, gathering lherzolite fragments from the walls and perhaps precipitating some olivine. On passing into the base of the crust, the magma stops temporarily in a reservoir where olivine is precipitated, C02 builds up and the pyroxene stability field expands. The limit for liquid immiscibility is reached and carbonatitic magma separates. Pyroxene crystallizes to yield pyroxenite cumulates and the magma differentiates to phonolite. Influxes of new primitive magma mix with the phonolite producing reversezoned pyroxenes and, in extreme cases, new melts of mugearitic composition. All these magmas can rise directly to the surface and erupt as lavas, but some stop at subvolcanic levels and crystallize there, forming small plutons of ijolite and nepheline syenite.
Nephelinites and carbonatites
However, kimberlites do seem to produce a residual carbonatitic fraction in some cases, such as at Benfontein, S Africa (Dawson & Hawthorne 1973). Gittins (1978) likewise describes some carbonatitic and ultramafic rocks from the Cargill complex, N Ontario, as related by simple magmatic fractionation. The third possibility is that carbonatite and nephelinite magmas can be derived as independent melts in the asthenosphere, a view upheld by Woolley & Jones (1987), but, as Treiman & Essene (1983) point out, carbonatites emanating from partial melting of mantle peridotite would be Na-poor. Derivation of a carbonate melt from the mantle has been shown by the experimental work of Eggler (1978) and Wyllie (1978, 1980). They show that at about 30 kb in a phlogopitebearing mantle with a high CO2/H20 ratio the initial melt could be carbonatitic and associated with kimberlitic or melilitic magmas. Wyllie further showed that at lower CO2/H20 ratios the upper asthenosphere can produce an alkaline ultrabasic silicate partial melt from which an immiscible alkali-rich carbonate liquid would probably separate.
Liquid immiscibility The process of the separation of two immiscible liquids is considered to be central to the interpretation of the evolution of carbonatitic magmatism. Therefore liquid immiscibility must be examined to see whether it best accords with the petrological evidence. First, however, the alternatives for the production of carbonatite magma should be briefly considered, with the assumption that carbonatites are magmatic and not hydrothermal metasomatic deposits (Bulakh & IskozDolinina 1978). There are three possible styles of origin. First, the carbonatite magma is a product of melting of limestone in the crust. This was upheld by Shand (1949) among others, but experimental and geochemical work on synthetic systems, mainly by Wyllie (well summarized in his 1978 paper), has demonstrated the igneous nature of carbonatites. The second possibility is that carbonatite is the product by fractional crystallization of nephelinitic and ijolitic magma (King 1965). The discreteness of many ijolite and carbonatite intrusions and the almost complete lack of mineralogical and chemical continuity between the silicate and carbonate magmas is not in keeping with carbonatite as a residual magmatic fraction from nephelinite magma, a point discussed by Prins (1978). Volcanic rocks midway between nephelinite and carbonatite are unknown except for a few pyroclastic examples. Plutonic rocks midway between ijolite and carbonatite are uncommon, and bimodality can be demonstrated in most complexes. In the few complexes where transitional rock types occur, the structural and textural relations of these rocks in both the field and the laboratory usually show that ijolitic rocks have been invaded by carbonatite, producing hybrids. 900
L1
65
Experimental evidence In their study of possible liquid immiscibility, Koster van Groos & Wyllie (1966, 1968, 1973) observed liquid immiscibility along the join NaA1Si3Os-Na2CO3 above 870~ and at 1 kb or more, but not at atmospheric pressure (Fig. 8). Introducing water into the system (Koster van Groos & Wyllie 1968) Caused the field of liquid immiscibility to shrink. At lower temperatures (about 700-750~ albite is joined by cancrinite on the liquidus. The accompanying H20-COz vapour contained Na20 and SiO2, and such vapours are considered to be the synthetic
[
L 2----~ LI+L2
T~ 800
NO+L2
NC+LI+L 2
_
~
NC+L~
700 Albite
I
20
40 WT %
I 6o
I 8o
Na2CO3
FIG. 8. Highly simplified phase diagram of the join Ab-NC with 10% H20 at 1 kb pressure. The fields of cancrinite and vapour have been omitted. L1 and L2 are immiscibleliquids. (After Koster van Groos & WyUie 1973.)
66
M. J. Le Bas
equivalents of N a - K fenitizing fluids known to emanate from carbonatite magmas. In 1973 Koster van Groos & Wyllie added plagioclase (Abs0 and Abs0) to the runs and observed liquid immiscibility along both joins between the plagioclase and Na2CO3. Cancrinite lay on the solidus and the carbonate liquid was richer in Ca than the silicate liquid which was dominantly albite and nepheline normative. Watkinson & Wyllie (1971) studied the comparable system NaA1SiO4-CaCO3-H20 at 1 kb but found no liquid immiscibility along the join nephelinecalcite. Had higher pressures been used, the experiments of Freestone & Hamilton (1980) indicate that liquid immiscibility would have been observed. Besides revealing the extent of liquid immiscibility in certain silicate-carbonate systems, these experiments showed (Fig. 8) that the silicate melt could have moderate quantities of dissolved carbonates, the proportion varying with water and Ca contents. In contrast, the carbonate melt usually had only about 5~ dissolved silicates. The proportion of silicate in carbonate melts and carbonate in silicate melts decreases as the temperature reduces. These distributions of one component dissolved in another correspond closely to those seen in natural ijolites and carbonatites. Calcite is sometimes observed as a late-stage trace constituent in ijolites, urtites and nepheline syenites, and fluid inclusions trapped in apatites in ijolites also reveal the presence of carbonates early in the history of crystallization of those melts (Rankin & Le Bas 1974a). Introducing K into the phase system likewise permits liquid immiscibility (Wendlandt & Harrison 1979). P and Ti also enhance the occurrence of liquid immiscibilityby increasing the polymerization of the melt (Freestone 1978; Mysen e t al. 1981). P is particularly relevant to carbonatites which are often rich in apatite. Because previous experiments did not contain a pyroxene component, Verwoerd (1978) carried out melting experiments at 2 kb on compositions along the join synthetic ijolite (5acmite +3 nepheline + 2 Na-disilicate) and an Na-Ca-natrocarbonatite end-member in order to obtain synethtic melts close to those of natural compositions. The results confirmed the liquid immiscibility shown by Koster van Groos & Wyllie. Freestone & Hamilton (1980) furthered the study of liquid immiscibility by conducting experiments over the range 0.7-7.6 kb and 9001250~ Using natural nephelinite and phonolite starting materials from Oldoinyo Lengai, together with a simplified synthetic natrocarbonatite, they showed that liquid immiscibility occurs between relatively polymerized and depolymerized liquids
and that both nephelinite and phonolite magmas are immiscible with alkali-rich carbonate melts (Fig. 9). Increasing the pressure widened the immiscibilityfield and increased the partitioning of Mg into the carbonate fraction but decreased that of K. Decreasing the temperature also widened the immiscibilitygap. Most nephelinites and phonolites of carbonatite complexes plot near the solvus for 7.6 kb at 1100~ (Fig. 9). These conditions are interpreted as those closest to the natural ones. Bedson (1983) took the experiments further, extending the pressure range to 25 kb. Amongst various distribution coefficients measured, he found that the rare earths were approximately equally distributed between the silicate and carbonate liquids, but that Ta, and more particularly Hf, were enriched in the silicate liquid. Hence immiscible silicate melts coexisting with carbonate melts should have high Hf/Ta ratios, a feature which would not be consistent with crystal fractionation of carbonate melt from silicate (Bedson 1984). Not enough data exist to test this hypothesis at present. Some early experiments by Brey & Green (1976) suggested the possibility that liquid immiscibility might occur between olivine melilitite and carbonatite, but later work (Brey 1978) showed that this was not the case. Similarly, no evidence has been found that kimberlite and carbonatite are related by liquid immiscibility. Natural evidence
At Fen and Aln6 the ijolitic and carbonatitic components do not crop out as discrete intrusions. Instead, the early ijolites and associated fenites are intruded and refenitized by carbonatites. Since the intrusion of carbonatite is frequently A
NazO + K20
IX iquids
~/ /
II
0.7/ ~.> / / /,i L . ~ "
Si02 + AI205
r~o~_
~ i
wt. %
/SOl
X
\\ \
CaO
FIG. 9. Liquid immiscibilityplot showing silicate fields I-VIII (described in the text p. 68) and the natrocarbonatite field IX around the two-liquid boundaries for 0.7, 3.0 and 7.6 kb at 1100~
67
Nephelinites and carbonatites accompanied by extensive brecciation, the ijolites are fragmented and at Fen, where the original extent of the ijolite intrusion was at least 2 km across, relatively unaltered ijolitic and melteigitic rocks are now present only in a small area (less than 1 km across) around Melteig farm in the SW corner of the complex. The rest are permeated and veined by carbonatite, giving the impression that a complete spectrum of compositions from pure silicate to pure carbonate rocks exists with the carbonate as a late-stage component. Tveitasite, hollaite, kasenite and ringite are local names given to these mixed rocks, but since they are hybrid rocks the names are of little value. In some complexes the ijolite and carbonatite intrusions do not penetrate each other but form discrete intrusions. The type locality of ijolite at Iivaara in NE Finland (Lehijarvi 1960) is one such discrete 2 km wide plug, with no signs of carbonatite at the surface. Carbonatites can also form discrete intrusions, as in the 5 km diameter pipe at Sokli on the west flank of the Kola ultraalkaline province (Vartiainen 1980). The field relations in E Africa show distinct bimodality of the intrusions (Le Bas 1977). Most intrusions combine discrete ijolitic and carbonatitic components with few or no hybrids. The clearest example is the Usaki complex (Le Bas 1977) which is an ijolite ring complex 3 km across with micro-ijolite marginal facies but with no development of carbonatite or calcite ijolite (i.e. an ijolite with more than 10% calcite). A few samples do show late-stage interstitial calcite (about 1%) as well as a few calcite veins, and this calcite is interpreted as being the residue of the few per cent of carbonate remaining dissolved in the silicate melt after the immiscible separation of the carbonate melt. The carbonatite partner of this ijolite is the Wasaki subvolcanic complex which occurs 5 km away to the north. The bimodality is particularly well marked on the rare occasions when extrusive volcanic products are preserved. At Oldoinyo Lengai the lavas are either silicate (nephelinite to phonolite) or carbonate. Parallels of this bimodal behaviour are seen in the extrusive products of all other known volcanic structures, even the newly discovered structure on the Cape Verde Islands (Silva et al. 1981). Fluid inclusions provide another line of evidence that liquid immiscibility can separate carbonate and silicate magmas. During heating and homogenization experiments on fluid inclusions contained in the apatite crystals of E African ijolites, a sequence was observed showing liquid immiscibility actually taking place in the heating stage under the microscope (Rankin & Le Bas 1974b). At 960~ a homogeneous liquid
filled the negative-crystal-shaped cavity of the fluid inclusion, but at 950~ small globules appeared and by 800~ the globules had solidified to silicate glass. At 575~ the remaining liquid solidified and instantaneously crystallized to a strongly birefringent mesh of elongated carbonate crystals. Further evidence of immiscibility comes from fluid inclusions in apatites within ijolite, which have bulk compositions close to natrocarbonatite (Le Bas & Aspden 1981). The ijolite-inclusion tie-line (Fig. 10, Ij-IC) attests that the relation is one of liquid immiscibility. It is interpreted that nephelinite magma was constrained in its crystallization path by being on the solvus and that, as it progressively crystallized, so it lost by immiscibility more of its small content of dissolved carbonate. Since the change in bulk chemical composition of the silicate melt as it migrated along the solvus was governed by loss of Ca (precipitation of diopsidic aegirine-augite and apatite), it moved towards urtite (Fig. 10, Ur). During this process the carbonate liquid became richer in Ca (Fig. 10, UC). Hence it is recognized that not only do droplets ofcarbonatite liquid separate from crystallizing nephelinite magma but also the droplets change composition with fractionation of that magma. One criterion for liquid immiscibility is that, not only must the two liquids be in equilibrium with each other, but so also must any additional phase. Several s6vitic carbonatites contain olivine, e.g. Aln6, Jacupiranga, Kaiserstuhl, Mbeya, Oka and Palaborwa, and, where measured, they lie in the range Fo95_ss. This range is similar to those in co-magmatic olivine ijolites, olivinites and olivine-bearing nephelinites. Such composiA
84
Na20 + K20
t, NCb i
~iIc
/-.// / S[O2+ AI203
'JIj
/ ,liquid
\ CoO
FIG. 10. Plot as Fig. 9 showing joins of silicate host and carbonate fluid inclusions for ijolite (Ij-IC) and urtite (Ur-UC) compared with the solvus for 3 kb at 1100~ NCbl and NCb2 are the span of natrocarbonatite analyses from Oldoinyo Lengai.
68
M. J. Le Bas
tions are in keeping with the proposition that the carbonate and silicate magmas could be related by immiscibility. Liquid immiscibility is also demonstrated by cryptic variation in the apatites in carbonatites and ijolitic rocks. Apatites precipitating from ensuing fractions of silicate magmas take different paths from those in fractionating carbonatite magmas, although all commence at a composition common to both silicate and carbonate hosts (Le Bas & Handley 1979). If the ijolites and carbonatites were related by fractional crystallization, a solid-solution phase crystallizing throughout would be expected to show a continuous variation throughout. Lapin & Vartiainen (1983), in their study of apatites from Finnish orbicular carbonatites, also reached this conclusion.
Relation between magmas and the system (Si + AI)-(Na + K)-Ca The carbonate-saturated system (Si + A1)-(Na + K)-Ca reported by Freestone & Hamilton (1980) accounts for 75%-85~ of the components of ultra-alkaline rocks and is therefore suitable for discussion here. When the compositions of common igneous rocks as well as ultra-alkaline rocks are plotted on this phase system using solvus relations at 1100~ (Fig. 9), a significant distribution is seen. This temperature is chosen because melting experiments by Piotrowski & Edgar (1970) on W Kenyan nephelinites, ijolites and nepheline syenites showed that liquidus temperatures lay in the range 1100-1200~ at 1 arm pressure, with solidus temperatures of about 950-1000~ and liquid immiscibility operates only above solidus temperatures. Figure 9 shows that, apart from some unusual rock types (fields IIa and IIb), most compositions are confined to the (Si + A1) corner with fields II and III lining up along an apparent boundary parallel but lower down in the diagram than the 7.6 kb solvus. Assuming an approximately regular interval between the solvus isobars, the apparent boundary can be taken at about 10 kb. The first significant feature is that it is the nephelinites and ijolites (field II) and phonolites and nepheline syenites (field III) which abut the solvus, and that all the other common rock compositions plot away from the solvus. This distribution supports the proposition that, if experimentally determined tie-lines are used (Freestone & Hamilton 1980), only these strongly alkaline compositions are possible candidates for liquids which could be conjugate with C a - N a - K carbonate liquids (field IX).
A second feature is that olivine nephelinites, melanephelinites and melilite nephelinites (field I) do not apparently lie on the solvus, although they could do so at higher pressures or lower temperatures, both of which enlarge the twoliquid field. The latter case of a lower temperature is petrologically unrealistic, but having mafic magmas of field I on the solvus at higher pressure is quite possible in the conditions envisaged for magma chambers deep in the crust. Even if these nephelinitic compositions were not on the solvus, they could become so after slight fractionation to field II compositions. This distinction perhaps explains why melanephelinites are associated with carbonatites but olivine nephelinites, which seldom fractionate, do not have accompanying carbonatites, it being assumed that the nephelinites were carbonated in the first place. It is further possible that if an olivine nephelinite rose high into the crust and then did fractionate, the magma would not impinge on the solvus and carbonatite could be produced by fractional crystallization. Syenites (field IV) also plot away from the proposed solvus whilst nepheline syenites plot on it, and this relation corresponds to the known absence and presence respectively of carbonatites with these rocks in the field. A third feature is that melilitites (field V) and kimberlites (field VI) lie well away from the proposed solvus, in keeping with the known geology of melilitite and kimberlite occurrences. Likewise basalts, andesites and rhyolites (fields VII and VIII) cannot produce carbonatites by liquid immiscibility. Fields IIa (alnoite) and IIb (urtite) are interesting exceptions. Alnoite is a frequent accessory magma in carbonatitic complexes and is itself commonly strongly carbonated. It is suggested that alnoite could be the fractionated product of olivine-melilite nephelinite which had penetrated high into the crust and from which the late-stage carbonate-rich residue had not separated. Urtite (field IIb) is rare even when compared with the ijolite in which it occurs, usually as minor intrusions at relatively high levels in the ijolitic plutons. It is proposed that urtite can result from liquid immiscibility taking place in an ijolite pluton at mid-crustal levels. The low Ca content of urtite compared with ijolite (Fig. 9) is consistent with loss of a C a - N a rich carbonate fraction by liquid immiscibility. This is further corroborated by the composition of the carbonate fluid inclusions trapped in urtite which are relatively Ca-rich compared with those trapped in ijolite, as shown in Fig. 10. Many of these possibilities must remain speculative until the important constituents Fe, Ti, P, F and oxygen fugacity are incorporated into the system.
69
Nephelinites and carbonatites
TABLE 6. Analyses of typical carbonatites (wt.~)
Carbonatites Despite the fact that nearly all the carbonatites of the world are calcite or dolomite-rich, the primary and parental carbonatitic magmas produced by liquid immiscibility are considered to be strongly alkaline, akin to natrocarbonatite. Twyman & Gittins (1987), however, do not consider natrocarbonatite to be either primary or produced by liquid immiscibility. The close agreement between the experimental work described above leaves no doubt that liquid immiscibility must be significant and that the primary composition is rich in Na + K. The fact that fluid inclusions trapped in early apatites in carbonatites are alkali-rich (Rankin 1975, 1977) lends further support. Most carbonatitic complexes display a wide variety of carbonatites. Field work invariably shows that the earliest intrusions are s6vites, with later alvikites and then ferrocarbonatites. N a - K fenitization is usually confined to the s6vite. Alvikites rarely show fenitization and ferrocarbonatites never show it. S6vites occasionally form diatremes 1-3 km in diameter. More usually they are tens or hundreds of metres across, while alvikites and ferrocarbonatites occur in swarms of dykes and veins.
Natrocarbonatite
1
SiO2 TiO2
AI203 Fe203 MnO MgO CaO Na20 K20 P205 CO2 F C1 SO3 SrO BaO REE
2
3
4
0.05 0.88 0 . 1 6 6 . 1 2 0.01 0.18 0 . 0 7 0 . 6 8 0.11 0.37 0 . 1 7 1.31 0.41 2.62 4.04 7.55 0.48 0.39 0.41 0 . 7 5 0.48 0.31 0.67 12.75 14.43 53.60 51.20 29.03 33.89 0.09 0 . 2 5 0.14 8.39 0.03 0.01 0.79 0.93 3.18 1.52 2 . 6 6 30.53 38.38 39.50 37.03 0.09 2.71 0.06 - . . 3.81 Trace . -0.89 2.88 - 1.35 0.23 0 . 1 0 0.01 0.08 0 . 1 7 0.11 1.26 0.05 0.3 ND c.0.1
5
6
3.24 0.83 0 . 0 0 0.07 0.20 0.65 11.50 11.00 5.18 5.53 10.74 0.36 25.85 43.60 -0.05 -0.06 1.27 0.42 32.62 30.42 --. 0.49 - 0 . 7 3 0.07 2.48 >4.0 2.82 c. 1.5
All analyses sum to 100 __+1.7 (LOI included). 1, Natrocarbonatite lava, 1960 eruption, Oldoinyo Lengai, N Tanzania (Gittins & McKie 1980); 2, s6vite dyke, Tundulu, Malawi (Garson 1965, p. 24); 3, alvikite cone sheet, Homa Mountain, W Kenya (Le Bas 1977, p. 317 (HC629) traces by C. Barber, unpublished); 4, dolomite carbonatite (beforsite dyke), Aln6, Sweden (von Eckermann 1948, p. 122); 5, ferrocarbonatite (high Mg), Kangankunde, Malawi (Garson 1965, p. 54); 6, ferrocarbonatite (low Mg), Homa Mountain, W Kenya (Le Bas 1977, p. 318 (HF15) traces by C. Barber, unpublished).
magma
The chemical composition of parental natrocarbonatite is still uncertain. Table 6, column 1, gives the composition of natrocarbonatite extruded from Oldoinyo Lengai in 1960. This is the only known instance of fresh carbonatite lava, and its composition (CC, 30 wt.~; NC, 58 wt.%; KC, 12 wt.~o) when plotted on Fig. 9 (field IX) indicates that it is the conjugate melt with phonolite magma, as interpreted by Freestone & Hamilton (1980). The conjugate carbonate melt with nephelinite would, according to them, be more calcic, but no natural examples are known except perhaps at Kaiserstuhl (Keller 1981). O and C isotopes confirm the igneous nature of natrocarbonatite and indicate a temperature of crystallization in the range 400-800~ (Suwa et aL 1975). Natrocarbonatite has been described in detail by Dawson (1962), Du Bois et al. (1963), Cooper et al. (1975), McKie & Frankis (1977) and Gittins & McKie (1980). It is porphyritic with large clear platy crystals of nyerereite ((Na,K)2Ca(CO3) 2 solid solution) and large brownish rounded crystals of gregoryite (Na2CO3 with some Ca + K substitution for Na) in a glassy carbonate matrix containing quench needles of nyerereite.
Table 7 gives a modal analysis of a sample collected by N. J. Guest from the 1960 eruption. Older examples of carbonatitic lavas or their pyroclastic equivalents have recently been discovered. Deans & Roberts (1984) described carbonatite tufts and lava clasts from the Miocene volcano at Tinderet in Kenya in which flowaligned calcified nyerereite crystals were identified. At one time they were thought to be pseudomorphs after melilite, and although relict nyerereite has never been found (nor will it be since it is water soluble) further textural study leaves little doubt that the platy cavernous pseudomorphs, largely composed of myriads of small calcite crystals, are after nyerereite. Such pseudomorphs are now also known in lavas at Kerimasi in N Tanzania, and suspected at Homa Mountain and the Ruri Hills, W Kenya, at Kruidfontein, S Africa, and in the Cape Verde Islands. The platy pseudomorphs at Kerimasi volcano described by Mariano & Roeder (1983) were identified on textural features by Hay (1983) as former nyerereite crystals. They coexist with fresh tablet-shaped calcite phenocrysts. Although both pseudomorphs and phenocrysts are com-
7o
M. J. Le Bas
TABLE 7. Modes ofnatrocarbonatites (volume ~ ) Oldoinyo Kerimasi, Lengai, Tanzania Tanzania 82-8-15A2
Kerimasi, Tanzania 82-8-14C6
Kerimasi, Tanzania 83-7-26A
Tinderet, Kenya L. KIP 1
Tinderet, Kenya U. KIP 3
Kaiserstuhl, S Germany, and Fort Portal, W Uganda
A few Abundant
9.3 2.7
Very few Abundant
Abundant None
Phenocrysts Calcite Nyerereite Gregoryite
Nil 38.1 48.0
29.7 67.0 1
7.6 46.0 __
Groundmass Calcite Nyerereite Opaques Apatite Perovskite Pyrochlore
-6.02 1.1 ----
2.8 -0.3 0.2 Trace --
46.33 Trace Trace ---
Interstitial glassa
6.8
--
--
c.50 m
7.1 79.3 0.7 0.4
m
0.5 m
1, not identified in thin section; 2, may include groundmass gregoryite; 3, Cc" N y ~ 1 1 4, includes quench nyerereite. posed of calcite, electron diffraction microprobe analysis (Table 8) distinguishes the typical igneous calcite with high Sr from that in the pseudomorphs which have low Sr but detectable Mg and Na. This distinction, the interpretation that the pseudomorphs cannot be after calcite, the relations known from the C a - N a - K carbonate system (Fig. 11) and the probable liquid immiscibility relations shown in Fig. 9 all confirm the identification. The modal variation of natrocarbonatites (Table 7) shows that these lavas can have a wide range of compositions, assuming that the phenocryst phases are complete and cognate. By using the phase data of the synthetic C a - N a - K carbonate system at I kb (Fig. 11) determined by Cooper et al. (1975), it can be seen that the natrocarbonatite of Oldoinyo Lengai plots in field A on Fig. 11, but the Kerimasi and Tinderet
lavas are more calcic, since both nyerereite and calcite are precipitated, and plot in the field B. Clasts from the Cape Verde Islands also appear to plot in field B. The occurrences of carbonatite lava at Kaiserstuhl in S Germany (Keller 1981) and at Fort Portal in W Uganda (in samples kindly provided by P. H. Nixon) show phenocrysts of calcite alone. This cannot easily be reconciled with the two-liquid field system shown in Fig. 9 and the liquidus surfaces of Fig. 11 unless the lavas were relatively N a poor. It is concluded that a wide variety of natrocarbonatite magmas can exist. Some occur with gregoryite and nyerereite phenocrysts, indicating compositions appropriate to immiscible liquids conjugate with phonolite magma, as at Oldoinyo Lengai. Some have nyerereite phenocrysts alone or with calcite (Table 7), appropriate to immiscible liquids conjugate with nephelinite magma.
TABLE 8. Microprobe analytical data on Kerismasi natrocarbonatites Brown pseudomorphs
Clear calcite crystals 1
MgO CaO S~ Na20
BD 55.1• 0.7• BD
2
BD 54.8• 0.7• BD
3
BD 51.5• 0.6• BD
4
0.8• 48.3• BD 0.5•
5
0.4• 50.6• BD BD
1 and 2, Large phenocrysts in lava (samples 82-8-14C6 and 83-7-26A respectively); 3, groundmass quench prismatic crystal (sample 83-7-26A); 4 and 5, two large pseudomorphs replaced partly by granules of calcite and partly by voids (with 1 and 2 respectively). Each analysis is the mean of four points with standard deviation to 2tr. The energy dispersive microprobe analyses were performed using the Cambridge Microscan Mk.5 at the University of Leicester. BD = below detection limit.
Nephelinites and carbonatites
/
/
NazCOs
/ K2Ca(CO3)2
Gry/(~A"XNyr/~ \\ / ~ ) ~ ~/ .) CaCOs No2Co(C03)2 Na2C%(C05)3
FIG. 11. CC-NC-KC carbonatite system showing field of natrocarbonatites from Oldoinyo Lengai (A) and from Kerimasi and Tinderet (B). Gry and Nyr are plotted at the natural compositions of gregoryite and nyerereite minerals. NY is the end-member nyerereite Na2Ca(CO3)2. FC is fairchildite K2Ca(CO3)2. (After Cooper et al. 1975).
Natrocarbonatites contain few minerals apart from carbonates. The few include oxide and phosphate minerals, but no silicates apart from rare biotite. Deans & Roberts (1984) report 15 wt.~o SiO2 and up to 5~o modal silicates, often biotite, in the Tinderet lavas. Bedson (1983) too usually found 1-5 wt.~ SiO z in his experimental carbonate charges. A120 3 was also present but was even more weakly partitioned into the carbonate melt. Biotite is the most frequent silicate mineral encountered, but only in accessory proportions. In some s6vitic carbonatites, mica and other silicates, particularly mildly sodic augite sometimes rimmed by aegirine, become abundant, and these rocks are the metacarbonatites. Vartiainen (1980) described a collar 2 km wide of metacarbonatite (average SiO2 content, 25 wt.~) around the s6vite core 2 km in diameter at Sokli. Armbrustmacher (1979) described petrographically similar rocks in Colorado, and Robbins & Tysseland (1983) have described silicocarbonatites (metacarbonatites) derived from gabbros at Pollen in N Norway. Metacarbonatites are interpreted to be 'reaction rocks' between magmatic carbonatite and country rock; the resulting 'mixed rock' is often up to 70~o calcite. Similar reactions, but on a smaller scale, are known in nearly all s6vites, and the products are often incorporated as xenoliths in s6vite intrusions. The experimentally determined distribution coefficients of elements between immiscible car-
71
bonate and silicate melts (Freestone & Hamilton 1980; Bedson 1983) help explain some of the observed variability. P partitions strongly into the carbonate melt, and hence nephelinites are often poor in apatite while the early s6vites are rich in apatite. Carbonatite lavas, however, always have less apatite than predicted by the above. Apatite is a dense mineral (specific gravity, 3.1) whereas carbonatite magma has a very low density of 2.2 g cm- 3 (Nesbitt & Kelly 1977), and therefore it is not surprising that apatite attains eruption less commonly. The partition of Mg between silicate and carbonate melts depends mainly on Pco2. At 7.6 kb Mg appears to partition preferentially into the carbonate (the data are poor), but at low pressure (1-3 kb) it partitions into the silicate. The distribution coefficient of K is never far from unity, unlike Na and Ca which both partition strongly into the carbonate melt. This causes a fall in the Na/K ratio of the silicate magma after the separation of carbonatite. Sr also partitions preferentially into the carbonate melt, particularly in the lower-temperature range 600-700~ but the Ca/Sr ratio remains unaltered in all melts (Koster van Groos 1975). Greater pressures increase the partition of REE into carbonate melt, with more light REE than heavy REE. This gives the steep REE pattern typical of carbonatites, but the steepness can vary considerably depending on the temperature at the time of separation (Bedson 1984). Intrusive carbonatites
The intrusion of carbonatite in pipe-like structures usually follows that of the silicate rocks (Fig. 12). The commonest carbonatite is s6vite showing adcumulate texture, and it is normally composed of 90~-100}/o calcite (Table 6). Dolomite may occur instead, particularly in older and more deeply eroded complexes, in keeping with the fact that dolomite is not stable at depths less than about 2 km at carbonatite magmatic temperatures (600-800~ Periclase occurs in the Amba Dongar and Kerimasi shallow-seated carbonatites. Common accessory minerals in carbonatites are fluorapatite, magnetite, biotite, pyrochlore and olivine, and some less-common minerals are aegirine-augite, arfvedsonite, K-feldspar, pyrite, zircon, baddeleyite, perovskite, sphene, quartz, fluorite, baryte, bastnaesite, parisite and monazite. The last six minerals occur mainly in the late-stage carbonatites, the ferrocarbonatites, while the remainder can usually be found in the earlier and more abundant carbonatites. There is a noticeable lack of sodic minerals
M. J. Le Bas
72 0
~,
.
4 0 0 0 t_ |
L
3 0 0 0 [I
NAPAK
,
5 [
-...--~--J~--~
+ + + +'+
metres above sea-level
/ ~ o~\~
~\c 9~J . ~'G'
, ,~o\~~ ~ \ \ ~
.~,~'~ /I --oba b~e-~ ~ '
2ooo t - _ ~ 1000~+
, l .Km
+ -r+
\
~o.~;' . . . . . . . . . . 9
Vole~
I "] -
. . . . . . .
~
UI~ A / t-IIM
~..
LOKUPOI
~eg~,,z,~" ~~ .,e o4 s u b -
-~ .
.-.n,~ ~o~
"oo @
_I
~"
J ""
~
"--.. " - ~
J
- q
+ + + + + /
Fenitized
~.
1/
basement
/ Ijolite
~ Carbonatite
~, Basement
FIG. 12. Cross-section of Napak nephelinite-carbonatite volcanic complex in E Uganda. The structure is typical of such complexes. (After King 1949.) and few potassic ones, considering the abundance of Na and K in parental natrocarbonatite. The Na and K are lost by metasomatic reaction of the early carbonatites with country rock to form fenites. Therefore the carbonatites seen cropping out are not products of the complete magma less any volatiles but have also lost their original complement of alkalis (Table 6). In the case of adcumulate s6vites the loss could have accompanied the escape of the intercumulus material. The loss of alkalis unfortunately masks the original alkali content of the parental natrocarbonatite, and therefore it is not possible to decide whether the carbonate immiscibly separated from a nephelinite or a phonolite melt. However, in some complexes (e.g. Ruri Hills, W Kenya) there is a close temporal association of phonolite and carbonatite intrusions, which could thus be pairs of conjugate immiscible liquids. The immiscible separation of carbonate melt from silicate melt takes place at l l00-1000~ This is well above the liquidus temperature for the carbonate component of the melt, but not above that for olivine and apatite. Both olivine and prismatic apatite therefore precipitate from carbonatite magma and under these conditions form cumulates at the base of the intrusions such as those occurring in Kola. Not only are olivine and apatite much denser (specific gravity of 3.3 and 3.1 respectively) than carbonatite magma (specific gravity, 2.2), but carbonatite magma has an unusually low viscosity of 5x 10-2 P (Treiman & Schedl 1983) permitting efficient gravity settling. If oxidizing conditions prevail, magnetite may also precipitate and accumulate. The carbonate liquidus is encountered at 500600~ This temperature is gauged from homogenization studies on fluid inclusions trapped in apatites (Rankin 1975, 1977). These apatites are not the prismatic crystals which accumulated from the early phase of cooling but are the large ovoid apatite crystals which characterize s6vite. This temperature is in general agreement with the liquidus temperatures determined by Wyllie and others for carbonatite magma. The low
temperatures are also marked by fenitization of the thermal aureole around the carbonatite, in which low albite, orthoclase, microcline, phlogopite and arfvedsonite commonly occur. With cooling of a carbonatite intrusion, fractionation of s6vite to alvikite to ferrocarbonatite takes place, with a final small-scale development of calcite carbonatite which lithologically looks like s6vite but is devoid of the usual accessory mineralogy characteristic of normal s6vite. This sequence is observed in E Africa (Le Bas 1977), in the U.S.S.R (Kapustin 1981), the Chilwa Province of Malawi (Garson 1965) and elsewhere. A variant on this occurs when the carbonatite is dolomitic. In this case the sequence is s6vite to dolomite carbonatite to ferrocarbonatite. These sequences accord with Wyllie's (1965) experiments. Olivine (Foss_90), rarely monticellite and very rarely melilite are restricted to the earlier and deeper calcite carbonatites (s6vites). Apatite and pyrochlore can crystallize throughout the sequence of carbonatites but are most abundant in the early s6vites, with apatite sometimes being so abundant that it forms apatite rock. Such apatite deposits have been mined. Apatite rock also occurs as disrupted schlieren in s6vite and dolomite carbonatite where they are dragged and folded by the magmatic flow. These apatites are almost completely deficient in Sr and REE and can be matched with microfragments of similar textured apatite rock caught up in the nephelinitic lavas. The main apatite in the s6vites occurs as large slightly resorbed crystals which have much the same contents of Sr and La as the apatite phenocrysts in nephelinitic lavas, as is appropriate to the postulated liquid-immiscible relation. Lapin & Vartiainen (1983) similarly describe olivine, apatite and magnetite which have fractionated from carbonatite. Biotite can occur in all intrusive carbonatites, as can pyroxenes and amphiboles, but they are more frequent in the early s6vites and dolomitic carbonatiteso Biotite flakes, often with margins oxidized to opaque oxides, together with magne-
Nephelinites and carbonatites tite phenocrysts can also occur in carbonatite lavas such as those near Fort Portal in SW Uganda. Feldspar, usually potassic but sometimes sodic, occurs in the margins of s6vitic and dolomitic carbonatites and is derived from the enveloping feldspathic fenite. Many but not all s6vites carry 1%-2~ of euhedral magnetite microphenocrysts which often coexist with biotite, apatite and pyrochlore. The magnetite is Ti poor with usually only 1-2 wt.% TiO2, and at Jacupiranga, Brazil, the magnetites in the s6vites are zoned with decreasing contents of Ti towards the rims. MnO and MgO, both of which usually show contents of less than 1 wt.~ and less than 4 wt.~ respectively, also decrease with the zoning. Magnetites in the later intrusive phases of carbonatite have compositions comparable with these outer zones and are also pure magnetite (Prins 1972; Gaspar & Wyllie 1983a). Ilmenite is rare in carbonatites and has very variable Mg, Mn and Nb contents. At Jacupiranga, Brazil, where it coexists with magnetite in the reaction zone betweenjacupirangite (titaniferous pyroxenite) and later s6vite, Gaspar & Wyllie (1983b) have calculated an equilibration temperature of 570-595~ with oxygen fugacities of 10- lS'5-10-19.5 atm, which are similar to the values estimated by Prins (1972) for magnetites from a variety of African carbonatites (540-575~ 10- 23-10- 24 atm). Alvikites are medium grained compared with the coarse-grained s6vites and are less voluminous, although in some complexes they appear to be very abundant because they form numerous but thin minor dykes and cone-sheets which comprise the majority of the outcrop. The mineralogy of alvikites is similar to that of s6vite, with a tendency for alvikite to have fewer accessory oxide and silicate minerals. The principal difference lies in the calcite which, in alvikite, was slightly ferroan when it crystallized. Probe analysis of the calcite often fails to reveal this ferroan character, evidently because the iron has unmixed from the calcite structure as a result of late-stage processes to which all carbonatites are prone. In such cases the iron is visible only as minute black specks scattered throughout the carbonatite quite separate from any magnetite microphenocrysts, and as brown staining along the crystal boundaries of the calcites. Alvikites are also marked by slightly greater contents of Mn, Ba and REE (Table 6). These elements appear to be in solid solution in the carbonate minerals. The apatites in alvikites are also rich in REE as well as Sr. True alvikites usually have a pale brownish colour, reflecting the Mn and Fe contents. Some, however, are white, like s6vite, and these commonly do not
73
contain as much Mn, Fe, Ba and REE as true alvikites and are in fact micros~vites. The distinction in the field is not easy as the brown coloration is often obscured by secondary processes, but trace-element analysis of the rock or probe analysis of the apatite can reveal the distinction. Much of the economic wealth of carbonatites lies in the ferrocarbonatites. The primary carbonate mineral of ferrocarbonatite is ankeritic, but with weathering much of it breaks down to secondary calcite or dolomite rimmed by iron oxides. Ferrocarbonatite occurs in only minor quantities in most complexes, and appears as thin brown dykes and veins with chilled margins and cross-cutting relations against earlier carbonatites. There are many good examples in E Africa, Malawi, Aln6 and the Cape Verde Islands. In other cases ferrocarbonatites form diffuse veins and patches apparently replacing s6vites, and are considered to be a late-stage development of the magmatic ferrocarbonatite. This residue of carbonatite magma appears to have become so rich in volatiles that it no longer behaved as a melt but as a volatile-rich fluid. Brown ferrocarbonatitic material can be seen progressively penetrating coarse calcite and dolomite carbonatites, migrating along crystal boundaries and gradually replacing the calcite and dolomite. The replacement process can go to completion, and if it begins along a fracture, as it usually does, the end-product can be a parallel-sided zone a metre or so wide which can look rather like a crosscutting dyke. This process can be seen replacing s/Svite in Kenya, Malawi and Fen. At Newania, India, and Sarfartoq, W Greenland, it replaces dolomite carbonatite. The apatites in the ferrocarbonatites also provide evidence for the replacement process. The cryptic variation shown by apatite is usually recorded by increases in REE, which reach a maximum in the ferrocarbonatites, but often the apatite in ferrocarbonatite lies compositionally in the alvikite and sometimes s6vite fields. This is consistent with the interpretation that many ferrocarbonatites were formerly alvikites subsequently transformed and mineralized. Mineralization
Four stages of mineralization can be recognized, all following on the magmatic Fe-enrichment (and Mn-enrichment) stage of ferrocarbonatite development. The first stage of mineralization is the production of REE-rich carbonatites from residual Frich magmas (Moiler et al. 1980; Samoylov & Smirnova 1980; Viladkar & Dulski 1986). The
74
M. J. Le Bas
REE minerals segregate in pockets and fractures along sheet-like ferrocarbonatite bodies. The minerals, principally bastnaesite ((Ce,La)FCO3), appear as pale-brown streaks within the darkerbrown to black ferrocarbonatite, and show strong light REE enrichment with La/Yb ratios of 1001000. U - T h radioactivity and associated fluorite or baryte are usually indications of the presence of REE minerals. Nearly all carbonatite complexes show this mineralization to some extent. The largest is at Mountain Pass, California, where Sr-rich carbonatites have more than 10% bastnaesite and other REE fluorcarbonate minerals cut by barytebearing and silicified radioactive veins (Olsen et al. 1954). Experiments by Jones & Wyllie (1983) suggest that the bastnaesite could have been precipitated from a low-temperature carbonatite magma (perhaps about 500~ or less). The Kangankunde carbonatite in Malawi has a similar REE mineralization but with Sr > Ba. Following the formation of the replacement ferrocarbonatites is the second-stage mineralization associated with late-stage fluids and the precipitation of fluorite at lower temperatures (100-200~ (Roedder 1973). Massive fluorite usually forms along fractures, particularly in the roof zones of carbonatitic complexes. It is variously purple, yellow or colourless, and can also occur as scattered crystals in the early carbonatites adjacent to the fractures along which the fluids passed. Evidently, F-rich fluids penetrated into the wall-rocks. At Kruidfontein near Johannesburg and at Amba Dongar, India, the F-rich fluids rose to the top of the carbonatite complex and entered and mineralized the fractured rocks capping the complex. At Kruidfontein the fluids penetrated up into bedded carbonatitic tufts sitting in a volcanic caldera structure and formed fluorite-rich horizons interbedded with the tufts. The third stage is the formation of baryte. It is present in nearly all carbonatite complexes and follows closely on the formation of fluorite with which it commonly occurs but in much smaller quantities. Local concentrations do occur and have been mined (e.g. at Aln6, Sweden). The Ba, F and REE are all mantle derived. They are not products of assimilation of continental crust, as it is sometimes considered for the Curich mineralization at the Palaborwa carbonatite in S Africa, because exactly similar Ba, F and REE mineralization is known in the Cape Verde Islands which are seated directly on oceanic crust (Le Bas 1984). The fourth stage of mineralization is marked by the formation of U - T h minerals together with pervasive silicification. This mineralization is
never very extensive and is characterized by low U/Th ratios which makes some deposits uneconomic (e.g. Loe-Shilman, N W Pakistan (Jan et al. 1981)). This is often followed by late-stage calcite veining and reprecipitation of Fe oxides, usually hematite. The Fe is derived from ferrocarbonatites penetrated by hydrothermal fluids. If the fluids become locally all-pervasive, assisted by incursions of groundwaters, then large hematite deposits can form. The Rodberg iron deposit at Fen, S Norway, is an example of this process (Andersen 1984). These Fe deposits are distinct from the magnetite deposits such as that at Kovdor, Karelia, U.S.S.R., which is a product of primary magmatic precipitation.
Fenitization The products of fenitization are mineralogically syenitic, with the result that they are sometimes misidentified as igneous syenites. The chief points of distinction lie in the texture and structure. Syenitic fenites are often heterogeneous and show low-temperature deformation or shattering. With increasing metamorphic temperatures, schlieren and mechanical flow textures develop, and next to the igneous contact the fenites become coarse grained. The effects of fenitization are more extreme than those of ordinary contact metamorphism because the metasomatism causes strong chemical reactions which can totally obscure the original composition of the rock. As a result there are many theories of fenitic processes (King & Sutherland 1960; Woolley 1982). The key lies in the realization that fenitization is not a single process but is often multiple. The variables involved are (a) the geochemistry of the magma causing the metasomatism, (b) the nature of the fluids leaving the magma, (c) the depth profile of the magma (different reactions develop at different depths), (d) the porosity and structure of the country rocks being fenitized, (e) the mineralogy and geochemistry of the rocks being fenitized and (0 any water-rock reaction.
Syenitic fenites generated by ijolite Fenites are not seen around nephelinitic volcanic pipes and dykes, nor around xenoliths of country rock trapped in nephelinitic magma. Evidently time is insufficient for identifiable metasomatism to take place. However, fenites develop strongly around larger intrusive bodies of nephelinite magma that fractionate and crystallize to ijolite. Fenitic zones can often be identified around ijolite plutons, and Fig. 13 gives the typical
Nephelinites and carbonatites ORIGINAL MINE RA~.,~400
300
METRES 200
I
!
75 0
100 |
CONTACT
GRANODIORITE ZONE IT
ZONE I
=
ZONE 111"
ZONE
I
Quartz ~
Ae~i,rine- ou~lite Mognetite ]lmenite Oligoclose
'
Atbite (An3-s) i
Microcline
Orthoctose
_.
Biotite ,,, ~ Alk(:Ui FeldSpQr
Mesoperthlte & . . . . . o~tipe'rthite iNephetine . . . . .
.........
,
!
~
Pate brown biotite ,,I
; .
.
.
Phlogopite . . . .
.
W
I
I
i_j
tO Hornblende
~, s.0dic a..~hibote (,~Y,)
Sodic ;Qmphibote ' WoliQstonite ~"
~ SHATTERIN~L. w-
STRONG MORTAR
FENIT/ZEO
GRANULATION TEXTURES
GRANOOtoRITE
COMPLETE
& "-
GRANULATION
$YENtTh
FIG. 13. The mineralogical changes across the fenite zones surrounding ijolite in granodiorite. (After Rubie, in Le Bas 1977, p. 72.)
mineralogical changes that take place across the zones. The pyroxene grades from a weakly sodic aegirine-augite near the contact to almost pure aegirine in zone 1, and can form 5%-50~ of the rock. The feldspar near the contact is new hypersolvus alkali feldspar commonly in the range Orgo_6oAb60_40, but in the outer zones the original sub-solvus feldspars of the granitic basement survive. The hypersolvus character of the feldspars and sub-solidus texture of the fenites suggest fenitization temperatures of about 800~ Ijolite intrusion causes uplift of the country rocks. At one time it was thought that the uplift was related to the metasomatic transfer of material from magma to fenite, but Rubie (1982) has shown that in W Kenya the transfer of material, mainly SiOz, was from country rock to magma, and that granodiorite suffered a volume loss of up to 20~ near the contact. This results in the formation of feldspathized ijolite (petrographically equivalent to nepheline syenite) at the margin of ijolite intrusions (Fig. 14). Aluminium appears to have been almost immobile during fenitization, and the K / N a ratio
decreases slightly away from the contact with falling temperature (Rubie & Gunter 1983). Other studies have been based on the assumption that oxygen is immobile, a condition which approximates to a constant-volume fenitization (McKie 1966), but Appleyard & Woolley (1979) based their study of fenitization on a constant-specificgravity model which suggested, at the Borrolan syenite in N W Scotland and the Sokli carbonatite in N Finland, that modest volume increases had taken place. Carbonate 9minerals are rare in fenites around ijolites unless later carbonatite intrusions are present. Evidently the metasomatic solutions causing the fenitization were not carbonate bearing.
Syenitic fenites generated by carbonatite N a - K fenites around carbonatites are much more varied than those around ijolites. Only s6vites, dolomitic carbonatites and rarely micros6vites show fenitization. Alvikites and ferrocarbonatites almost never show fenitization, and this is
76
M. J. Le Bas +
+ +
+
BASEMENT IttLIIII IIIIIIIII1,, Nepheline-poor Nep~;~i;;-po, o:
_jr_
.....
+
--- =
,,
11111IT,iF-_--
x
lllrlll
-
--
•
IIII rr ;i-~-
Nepheline-ll
,,
x
•
'I[II[EIIJV-s'
+
" "
x "
inward d ffus on of Si during fenitization
Ii'/
~X_--~lllrelative
-.~
from less ldesilicified
~- '~1, basement
X
~/N,,~
/,, I
;^~J I
]l~lll(
~
~o.-o,,,o,,e
~'1
x
+
~
+
screen of chilled m c r o - j o te prevents diffusion of Si
X
-
-Jr-
Ix|/i!/
x
aegirine-rich
to
--~-_(!1 'jmore d e s i l i c i f i e d
rich fenite /
+ Jlil~
downward
-', =lllt~'"""~176o, ~,
lJ
I11111
from b a s e m e n t
FIG. 14. Schematic cross-section of ijolite intrusion with chilled margin and influx of silica during fenitization, and the development of feldspathized ijolite (mineralogicallyequivalent to nepheline syenite but texturally distinct) particularly along the upper parts of the margins. taken to indicate that early carbonatite magmas are alkali-rich but that these alkalis are lost during the early stages of carbonatite fractiona1 ~
0 + K20
/ -.6"D~ /
i' ;,,!gO + Fe20T
~k
~
A A
A
9
x X CaO
wt% FIG. 15. Carbonatite plot showing the path of
fractionation of carbonatite magma from natrocarbonatite (#) to s0vite (O) and alvikites ( x ) to beforsites and ferrocarbonatites (A): P, Fort Portal lava; T, estimated Tinderet lava; K, estimated Kerimasi lava (Table 7). (After Le Bas 1981.)
tion (Fig. 15). In some cases even the final phases of s0vite intrusion fail to cause fenitization. Evidently by that stage the s6vite magma has lost its alkalis, and instead could only brecciate the fenites formed earlier. This is contrary to the proposal that the concentration of alkalis increases during carbonatite fractionation (Twyman & Gittins 1987). Depth profiles are important. Fenitization around the topmost parts of a s6vite is strongly potassic, usually with the production of almost pure K-feldspar rock. Early records of this came from Malawi (Dixey et al. 1935), Songwe, Tanzania (Brown 1964), Kaiserstuhl, S Germany (Sutherland 1967) and Zambia (Bailey 1966). At Koga in Swat, N Pakistan, K-feldspar rock (plus rare aegirine) about 50 m thick completely caps a s0vite plug 500m across (Mian, personal communication). The K-feldspar is usually orthoclase Or87_9s, whereupon the rocks are termed orthoclasites (Sutherland 1965). Sometimes the K-rich feldspar is microcline, and Heinrich & Moore (1970) named such cases 'burnt rock' as they had that appearance. Commonly, the only other phase present is 1%-2~ of limonitic iron oxides, with calcite absent. The high K content is character-
Nephelinites and carbonatites istic of these fenites, whereas the K-feldspar developed in feldspathized ijolites and ijolitic fenites is less potassic (Or<87) (Le Bas 1984). When s6vites intrude mafic rocks, the fenites formed are phlogopite rich and not feldspar rich (Gittins et al. 1975), and it has been observed that, when s6vite cuts ijolite containing pyroxenite, the pyroxenite is phlogopitized and the ijolite is feldspathized, indicating that the Mg for the formation of phlogopite comes from the rock being fenitized and not from the carbonatite. Oxygen-isotope analyses on mineral pairs to determine temperature have yet to be made successfully, but X-ray diffraction of the K-rich feldspar indicates a low temperature of formation with obliquities up to 0.6 (Le Bas 1984). Haggerty & Mariano (1983) calculated a temperature of 500-550~ for the fenites around the Parana carbonatite, which agrees with that generally estimated from the mineralogy of carbonatitic fenites. In older and more deeply eroded complexes, the feldspathic fenites developed around so~ites are albite rich, sometimes coexisting with Mgarfvedsonite, aegirine and iron oxides. Such albitites occur around carbonatites in the fenitized quartzites at Turyi and in migmatites at Vuoriyarvi, both on the Kola Peninsula, and in gneisses at Novopoltavskoye between the Azov and Caspian Seas, U.S.S.R (Kapustin 1981). Albitite and aegirine-albite fenites are also formed from grits, granites and migmatites around the Sangu carbonatite in W Tanzania (Coetzee 1963). These albitite and albite-aegirine schists are lithologically similar to the syenitic fenites around the silicate complexes, but they can be distinguished mineralogically. The feldspar of fenites around carbonatites is almost pure albite (Ab>90) or almost pure orthoclase (Or>85), whereas the feldspar in the syenitic fenites around ijolite and syenite is an alkali feldspar Ab60_40 Or4o_60. The compositions reflect the low temperature of carbonatite fenitization (below 550~ in contrast with the higher temperatures of the silicate fenitization (more than 700~ Albitite also occurs as fragments brought up by carbonatitic volcanic breccias, e.g. Amba Dongar, India (Deans et al. 1972), and NyamajiWasaki, W Kenya (Le Bas 1977), indicating albite-rich fenites at depth whilst K-rich fenites are exposed at the surface. A similar zonation occurs around the Novopoltavskoye carbonatite, Azov, where inner zone fenites are albitic but the outer ones are microcline-rich (Kapustin 1982). The N a - K zonation is in accord with Orville's (1963) experiments on alkali-ion exchange between fluid and feldspar at falling low tempera-
77
tures, and finds support from Woolley's worldwide investigations (1982). S6vite and alvikite carry very little F, but the natrocarbonatitic magmas from which they were derived were probably F rich and C1 rich. Natrocarbonatite lava has 1.49-2.93 wt.% F and 0.77-4.15 wt.% C1 (Gittins & McKie 1980). Some minerals in carbonatites also show that they crystallized from halogen-rich magma, and Dawson & Fuge (1980) estimate that F > C1 was usual. In addition to apatite, biotite and Na-amphibole, which are the most important F-bearing minerals, clinohumite (Mg(F,OH)2.4MgzSiO4) has been reported at Cargill in Ontario, Jacupiranga in Brazil, Palaborwa in S Africa, Sokli in Finland and Gardiner in E Greenland (Gittins 1978). Even the residual nephelinitic glasses from Oldoinyo Lengai volcano have appreciable halogen contents (0.25-0.69 wt.% C1), and the F/C1 ratio for the whole rock is 1.6 (Donaldson & Dawson 1978). The F and C1 are evidently mantle derived, probably from phlogopite rather than apatite (Edgar & Arima 1985). However, at Oka, Quebec, Treiman & Essene (1984) calculate that the fluorine fugacity was low (10-44). If carbonatite magma were rich in halogens, particularly F, then all fenitic fluids emanating from it would also be F rich as would the residual fluids. Evidence for the former is seen in the fluorapatites and fluoramphiboles developed in fenites, and for the latter in the fluorite late-stage mineralization. Yardley (1985) also finds evidence from apatites that F might be important in metamorphic fluids. It is considered probable that a fluoride species is the main solvent of the fenitizing fluids.
Conclusions In this assessment of the main features of carbonatite-nephelinite volcanism, the following salient points can be made. The starting point is carbonated nephelinitic magma generated from an enriched mantle source such as that described by Cohen et al. (1984). Such magma rises through the lithosphere, zone refining as it rises, until it stops in the lower crust. Fractionation at that stage would permit olivine and pyroxene cumulates to form. At the same time the melt would fractionate to phonolite, and liquid immiscibility could take place forming a series of conjugate carbonatite magmas (Fig. 16). Nephelinitic magma is dense (about 2.7 g cm-3) and hence is not seen where the crust is composed of thick low-density unaltered sediments. Where nephelinitic magma has the buoyancy to penetrate the crust it can make
78
M. J. Le Bas
MELANEPHELINITE NEPHELINITEPHONOLITE LAVAS&PYROCLASTICS
PSEUDO-
DYKES
Rap,d
t /
IJOLITE
,
Ropid ascent
" "
I --.~
K.FEN!TE~A~ vK~TEE/
FenilizatiOn
IPER,DOTITE
/ /
FERROC,~BONATIT~
/I.SYE
I Slow / aster, I
I
LATE- STAGE CARB~NATITE
/
I:k/1 t oo.o
oscenl
o, .....
"~ "
TRACHYTE
URTITE & WOLL-URTITE LATE- STAGE ~11 ALKALINE FLUIDS/' CALCITE 1~ / IJOLITE -
SYE.,TES \
CARBONATED BRECCIASTUFFS & AGGLOMERATES
CARBONAIIELAVA & PYROCLASTC IS
.a
PQchonohon Low volahle ~ .
silicote melt
Volobie-rJch fluid
I lower crustal magma chamber
HELINITE I LIQUID FRACTIONATES I TO PHONOLITE I IMMISCIBILITY I
CUMULATES
? BASE OF CRUST
CARBONATED NEPHELINITIC MAGMA
PARTIAL MELT OF ENRICHED MANTLE SOURCE
FIG. 16. Schematic flow chart for the evolution of nephelinitic and carbonatitic magmas. (After Le Bas 1977, Fig. 24.2.) pyroxenite-ijolite-urtite plutons, with the ijolite perhaps developing an outer zone and capping of feldspathic ijolite (Fig. 14) as a result of the inward diffusion of silica from the country rocks undergoing fenitization. The country rocks are metasomatized to syenitic fenites with alkali feldspar and Na-pyriboles (Fig. 17). Although natrocarbonatite magmas of various (Na + K)/Ca ratios can be produced by the liquid
immiscibility, depending on whether the conjugate silicate melt is nephelinite or phonolite, the alkali loss to fenitization obscures the differences. All carbonatites follow similar paths producing s6vites, dolomite carbonatites, alvikites and ferrocarbonatites with REE enrichment, with the final production of residual mineralizing fluids which precipitate fluorite, baryte and U - T h minerals. F-rich alkali-bearing fluids cause feni-
Nephelinites and carbonatites
79
J
Cone-sheets~ plones of brecciation end K |enitizetion
'.?3
:"
+ +
/ ~_
.
I/0tite~, ~ + 1 ~ - I * +N'. ;, § i..I/."~ 4-
O,L
__.~.___,
- IN Cl i . r'." .I
FIG. 17. Schematic cross-section across an idealized nephelinite-carbonatite volcanic complex showing the normal structure and sequence of intrusion, the zonal fenitization around the ijolite, the deeper-level albitization (Na), the nearer-surface potassic feldspathization (K) accompanied by brecciation (A) and the emplacement of s6vite (CI), the cone-sheet fracture system of the alvikites (C2) and ferrocarbonatites (C3), and the final development of late-stage carbonate veins (C4) and a mineralized capping above the intrusive centre. (After Le Bas 1977, Fig. 23.1.) tization. Fluids e m a n a t i n g from early carbonatites cause albitization with albite-rich fenites +_ Na-pyriboles being developed at deeper levels around the carbonatite pluton. K-rich feldspathic fenites form towards the top and cap the s6vitic intrusions. They are often a c c o m p a n i e d by
brecciation and are usually followed by an F, Ba and U - T h mineralization. ACKNOWLEDGMENTS" I acknowledge most sincerely the many colleagues who have helped me reach these conclusions.
References ANDERSEN,T. 1984. Secondary processes in carbonatites: petrology of 'rodberg' (hematite-calcitedolomite carbonatite) in the Fen central complex, Telemark (South Norway). Lithos, 17, 227-45. APPLEYARD, E. C. & WOOLLEY,A. R. 1979. Fenitization: an example of the problems of characterizing mass transfer and volume changes. Chem. Geol. 26, 1-15. ARMBRUSTMACHER,T. J. 1979. Replacement and primary carbonatites from the Wet Mountains area, Fremont and Custer Counties, Colorado. Econ. Geol. 74, 888-901.
BAILEY, D. K. (1966) Carbonatite volcanoes and shallow intrusions in Zambia. In: TUTTLE,O. F. & GITTINS, J. (eds) Carbonatites, pp. 127-154. Wiley, New York. --1987. Mantle metasomatism--perspective and prospect. In: FtTTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 1-13. BAKER, B. Iq. 1987. Outline of the petrology of the Kenya rift alkaline province. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 293-311.
8o
M. J. Le Bas
BEDSON, P. 1983. The origin of the carbonatites and their relation to other rocks by liquid immiscibility. Ph.D Thesis, University of Manchester (unpublished). - 1984. Rare earth element distributions between immiscible silicate and carbonate liquids. Prog. Exp. Petrol. NERC 6th Rep., pp. 12-19. BOWEN, N. L. 1924. The Fen Area in Telemark, Norway. Am. J. Sci. 8, 1-11. BREY, G. 1978. Origin of olivine melilitites--chemical and experimental constraints. J. Volcanol. geotherm. Res. 3, 61-88. - - & GREEN, D. H. 1976. Solubility of CO2 in olivine melilitite at high pressures and role of COz in the earth's upper mantle. Contrib. Mineral. Petrol. 5 5 , 217-30. BRISTOW, J. W. & SAGGERSON, E. P. 1983. A general account of Karoo vulcanicity in southern Africa. Geol. Rundsch. 72, 1015-60. BROGGER, W. G. 1921. Die Eruptivgesteine des Kristianiagebietes. IV. Das Fengebeit in Telemark, Norwegen. Vidensk. Skr. 1. Mat-Naturv. Klasse, 1920 (9). BROOKS, C. K. & PRINZLAU,I. 1978. Magma mixing in mafic alkaline volcanic rocks: the evidence from relict phenocryst phases and other inclusions. J. Volcanol. geotherm. Res. 4, 315-31. BROWN, P. E. 1964. The Songwe scarp carbonatite and associated feldspathization in the Mbeya Range, Tanganyika. Q.J. geol. Soc. 120, 223-40. BULAKH, A . G . & ISKOZ-DOLININA, I. P. 1978. Origin of carbonatites in the Central'nyy massif, Turig peninsula. Int. geol. Rev. 20, 822-30. CLAGUE, D. A. & FREY, F. A. 1982. Petrology and trace element geochemistry of the Honolulu Volcanics, Oahu: implications for the oceanic mantle below Hawaii. J. Petrol. 23, 447-504. COETZEE, G. L. 1963. Carbonatites of the Karema depression, western Tanganyika. Trans. geol. Soc. S. Afr. 66, 283-333. COHEN, R. S., O'NIONS, R. K. & DAWSON, J. B. 1984. Isotope geochemistry of xenoliths from East Africa: implications for development of mantle reservoirs and their interaction. Earth planet. Sci. Lett. 68, 209-20. COOPER, A, F., GxYrINS, J. & TUYrLE, O. F. 1975. The system Na2CO2-K2CO3-CaCO 3 at 1 kilobar and its significance in carbonatite petrogenesis. Am. J. Sci. 275, 534-60. DAWSON, J. B. 1962. The geology of Oldoinyo Lengai. Bull. Volcanol. 24, 349-87. --1980. Kimberlites and their Xenoliths, Springer, Berlin. --& FUGE, R. 1980. Halogen content of some African primary carbonatites. Lithos, 13, 139-43. -& HAWTHORNE,J. B. 1973. Magmatic sedimentation and carbonatitic differentiation in kimberlite sills at Benfontein, South Africa. J. geol. Soc. Lond. 12, 61-85. DEANS, T. & ROBERTS, B. 1984. Carbonatite tufts and lava clasts of the Tinderet foothills, Western Kenya: a study of calcified natrocarbonatites. J. geol. Soe. Lond. 141,563-80. - - , SUKHESWALA,R. N., SETHNA,S. F. & VILADKAR,
S. G. 1972. Metasomatic feldspar rock (potash fenites) asssociated with the fluorite deposits and carbonatites of Amba Dongar, Gujarat, India. Trans. Inst. rain. Metall. Sect. B, 81, 1-9. DIXEY, F., SMITH, W. C. & BISSET, C. B. 1935. The Chilwa Series of Southern Nyasaland. Geol. Surv. Nyasaland, Bull. 5 (revised 1955). DONALDSON, C. H. & DAWSON, J. B. 1978. Skeletal crystallization and residual glass compositions in a cellular alkalic pyroxenitic nodule from Oldoinyo Lengai. Contrib. Mineral. Petrol. 67, 139-49. DOWNES, H. 1984. Sr and Nd isotope geochemistry of coexisting alkaline magma series. Cantal, Massif Central, France. Earth planet. Sci. Lett. 69, 32134. DE BOlS, C. G. B., FURST, J., GUEST, N. J. & JENNINGS, G. J. 1963. Fresh natrocarbonatite lava from Oldoinyo L'Engai. Nature, Lond. 197, 445-6. VON ECKERMANN,H. 1948. The alkaline district of Alno Island. Sver. Geol. Unders. Ser. Ca, No. 36. EDGAR, A. D. & ARIMA, M. 1985. Fluorine and chlorine contents of phlogopites crystallized from ultrapotassic rock compositions in high pressure experiments : implication for halogen reservoirs in source regions. Am. Mineral. 70, 529-36. EGGLER, D. H. 1978. The effect of CO2 upon partial melting of peridotite in the system N a 2 0 - C a O A1203-MgO-SiO2-CO2 to 35 kb, with an analysis of melting in a peridotite-H20-CO2 system. Am. J. Sci. 278, 305-43. FODOR, R. V., BAUER, G. R. & KEIL, K. 1982. Ultramafic inclusions and megacrysts in olivine nephelinite. Aitutaki Island. Cook Islands. N.Z. J. Geol. Geophys. 25, 67-76. FREESTONE, I. 1978. Liquid immiscibility in alkaline magmas. Chem. Geol. 23, 115-23. & HAMILTON, D. L. 1980. The role of liquid immiscibility in the genesis of carbonatites--an experimental study. Contrib. Mineral. Petrol. 73, 105-17. GARSON, M. S. 1965. Carbonatites in Southern Malawi. Geol. Surv. Malawi, Bull. 15. GASPAR, J. C. & WYLLIE, P. J. 1983a. Magnetite in the carbonatites from the Jacupiranga complex, Brazil. Am. Mineral. 68, 195-213. - & -1983b. Ilmenite (high Mg, Mn, Nb) in the carbonatites from the Jacupiranga complex, Brazil. Am. Mineral. 68, 960-71. GITTINS, J. 1978. Some observations on the present status of carbonatite studies. Proc. 1st Int. Symp. Carbonatites, Ministerio das Minas e Energia, Departamento Nacional da Produ~ao Mineral, Polos de Caldas, Minas Gerais, Brazil, pp. 10715. -& McKm, D. 1980. Alkalic carbonatite magmas: Oldoinyo Lengai and its wider applicability. Lithos. 13, 213-15. , ALLEN, C. R. & COOPER, A. F. 1975. Phlogopitization of pyroxenite; its bearing on the composition of carbonatite magmas. Geol. Mag. 11,503-7. HAGGERTY, S. E. & MARIANO, A. N. 1983. Strontianloparite and strontio-cherkinite: two new minerals in rheomorphic fenites from the Parana Basin
Nephelinites and carbonatites carbonatites, South America. Contrib. Mineral. Petrol. 84, 365-81. HAY, R. L. 1983. Natrocarbonatite tephra of Kerimasi volcano, Tanzania. Geology, 11,599-602. HEINRICH, E. WM. 1966. The Geology of Carbonatites, Rand McNally, Chicago, I1. --& MOORE, D. G. 1970. Metasomatic potash feldspar rocks associated with igneous alkaline complexes. Can. Mineral. 10, 571-84. JAN, M. Q., KAMAL,M. & QURESHI,A. 1981. Petrography of the Loe Shilman carbonatite complex, Khyber Agency. Geol. Bull. Univ. Peshawar, 14, 29-43. JONES, A. P. & WYLLIE, P. J. 1983. Low-temperature glass quenched from a synthetic, rare earth carbonatite: implications for the origin of the Mountain Pass deposit, California. Econ. Geol. 78, 1721-3. KAPUSTIN, Y. L. 1981. Mineralogy of Carbonatites. Amerind, New Delhi. --1982. Fenitization around carbonatite bodies. Dokl. Akad Nauk SSSR, 249, 126-8. KELLER, J. 1981. Carbonatite volcanism in the Kaiserstuhl alkaline complex: evidence for highly fluid carbonatite melts at the Earth's surface. J. Volcanol, geotherm. Res. 9, 423-31. KING, B. C. 1949. The Napak area of southern Karamoja, Uganda. Geol. Surv. Uganda, Mem. V. --1965. Petrogenesis of the alkaline igneous rock suites of the volcanic and intrusive centres of eastern Uganda. J. Petrol. 6, 67-100. --& SUTHERLAND, D. S. 1960. Alkaline rocks of eastern and southern Africa. Sci. Prog., Lond. 48, 298-321,504-24, 709-20. KOSTER VAN GROOS, A. F. 1975. The distribution of strontium between coexisting silicate and carbonate liquids at elevated pressures and temperatures. Geochim. cosmochim. Acta, 39, 27-34. -d~ WYLLIE, P. J. 1966. Liquid immiscibility in the system Na20-A1203-SiO2-CO2 at pressures to 1 kilobar. Am. J. Sci. 26, 234-55. --& ~ 1968. Liquid immiscibility in the join NaA1Si3Oa-Na2CO3-H20 and its bearing on the genesis of carbonatites. Am. J. Sci. 266, 932-67. --&--1973. Liquid immiscibility in the join NaA1Si308-CaAI2Si2Os-Na2CO3-H20. Am. J. Sci. 273, 465-87. LAPIN, A. V. & VARTIAINEN,H. 1983. Orbicular and spherulitic carbonatites from Sokli and Vuoriyarvi. Lithos, 16, 53-60. LE BAS, M. J. 1971. Per-alkaline volcanism, crustal swelling and rifting. Nature, Lond., Phys. Sci. 230, 85-7. - - 1 9 7 7 . Carbonatite-Nephelinite Volcanism, Wiley, New York. --1978. Nephelinite volcanism at plate interiors. Bull. Volcanol. 41,459-62. --1980. The East African Cenozoic magmatic province. Acad. Naz. Lincei, Atti Convegni Lincei, 47, 111-22. --1981. Carbonatite magmas. Mineral. Mag. 44,
133-40. Oceanic carbonatites. In: KORNPROBST, J.
--1984.
81
(ed.) Kimberlites 1 : Kimberlites and Related Rocks, pp. 169-78. Elsevier, Amsterdam. -1986. Ultra-alkaline magmatism with or without rifting. Tectonophysics (Special Issue on Continental and Oceanic Rift Zones), in press. --& ASPDEN, J. A. 1981. The comparability of carbonatitic fluid inclusions in ijolites with natrocarbonatite lava. Bull. Volcanol. 44, 429-38. --& HANDLEY, C. 1979. Variations in apatite composition in ijolitic and carbonatitic igneous rocks. Nature, Lond. 27, 54-6. LEHIJARVI, M. 1960. The alkaline district of Iivaara, Kuusamo, Finland. Bull. Comm. Geol. Fin. 185. LIPPARD, S. J. 1973. The petrology of phonolites from the Kenya Rift. Lithos, 6, 207-34. McKIE, D. 1966. Fenitization. In: TUTTLE, O. F. & GITTINS, J. (eds) Carbonatites, pp. 261-94. Wiley, New York. - - & FRANKIS,E. J. 1977. Nyerereite: a new volcanic carbonate mineral from Oldoinyo Lengai, Tanzania. Z. Kristallogr. 145, 73-95. MARIANO, A. N. & ROEDER, P. L. 1983. Kerimasi: a neglected carbonatite volcano. J. Geol. 91,449-55. MOLLER, P., MORTEANI,G. & SHLEY, F. 1980. Discussion of REE distribution patterns of carbonatites and alkaline rocks. Lithos, 13, 171-9. MYSEN, B., RYERSON, F. J. & VIRGO, D. 1981. The structural role of phosphorus in silicate melts. Am. Mineral. 66, 106-17. NESmTT, B. E. & KELLY, W. C. 1977. Magmatic and hydrothermal inclusions in carbonatite of the Magnet Cove complex, Arkansas. Contrib. Mineral. Petrol. 63, 271-94. NORRY, M. J., TRUCKLE, P. H., LIPPARD, S. J., HAWKESWORTH, C. J., WEAVER, S. D. & MARRINER, G. F. 1980. Isotopic and trace element evidence from lavas, bearing on mantle heterogeneity beneath Kenya. Phil. Trans. R. Soc. Lond., Ser. A, 297, 259-71. OLSEN, J. C., SHAW, D. R., PRAY, L. C. & SHARP, W. N. 1954. Rare-earth mineral deposits of the Mountain Pass District, San Bernardino County, California. U.S. Geol. Surv. Prof Paper 261. ONUMA, K. & TOHARA, T. 1984. The join CaMgSi206CaA12SiO6-CaCrA1SiO6 with special reference to chrome clinopyroxene and chrome spinel. Lithos, 17, 289-98. ORVILLE, P. M. 1963. Alkali ion exchange between vapour and feldspar phases. Am. J. Sci. 261,20137. PIOTROWSKI, J. M. & EDGAR, A. D. 1970. Melting relations of undersaturated alkaline rocks from S. Greenland compared to those of Africa and Canada. Medd. Gronl. 181 (9). PRINS, P. 1972. Composition of magnetite from carbonatites. Lithos, 5, 227-40. s 1978. Geochemical aspects of the alkaline carbonatite complexes of Damaraland, South West Africa. Proc. 1st Int. Symp. on Carbonatite, Ministerio das Minas e Energia, Departmento Nacional da Produ~ao Mineral, Polos de Caldas, Minas Gerais, Brazil, pp. 179-88. 1981. The geochemical evolution of the alkaline and carbonatite complexes of the Damaraland
82
M. J. Le Bas
igneous Province, South West Africa. Annen. Univ. Stellenbosch, Ser. A1 (Geol.) 3, 145-278. RANKIN, A. H. 1975. Fluid inclusion studies in apatite from carbonatites of the Wasaki area of western Kenya. Lithos, 8, 123-36. -1977. Fluid-inclusion evidence for the formation conditions of apatite from the Tororo carbonatite complex of eastern Uganda. Mineral. Mag. 41, 155-64.
LE BAS, M. J. 1974a. Nahcolite (NaHCO3) in inclusions in apatites from some E. African ijolites and carbonatites. Mineral. Mag. 39, 563-70. - & -1974b. Liquid immiscibility between silicate and carbonate melts in naturally-occurring ijolite magma. Nature, Lond. 2 5 0 , 206-9. ROBBINS, B. &. TYSSELAND, M. 1983. The geology, geochemistry and origin of ultrabasic fenites associated with the Pollen carbonatite (Finnmark, Norway). Chem. Geol. 40, 65-95. ROEDDER, E. 1973. Fluid inclusions from the fluorite deposits associated with carbonatite at Amba Dongar, India and Okorusu, South West Africa. Trans. Inst. Min. Metall., Sect. B, 82, 35-9. RUBIE, D. C. 1982. Mass transfer and volume change during alkali metasomatism at Kisingiri, western Kenya. Lithos, 15, 99-109. - & GUNTER, W. D. 1983. The role of speciation in alkaline igneous fluids during fenite metasomatism. Contrib. Mineral. Petrol. 82, 165-75. SAErHER, E. 1957. The alkaline rock province of the Fen area in southern Norway. Kgl. norske Videns. Selsk. Skr. 1957 (1). SAGGERSON, E. P. 1970. The structural control and genesis of alkaline rocks in central Kenya. Bull. Volcanol. 34, 38-76. SAMOVLOV,V. S. & SMIRNOVA,Ye. V. 1980. Rare earth behaviour in carbonatite formation and the origin of carbonatites. Geochem. Int. 17, 140-52. SnAND, S. J. 1949. Eruptive Rocks. Thomas Murby, London. SILVA, L. C., LE BAS, M. J. & ROBERTSON, A. H. F. 1981. An oceanic carbonatite volcano on Santiago. Cape Verde Islands. Nature, Lond. 294, 664-5. SPENCER, A. B. 1969. Alkalic igneous rocks of the Balcones Province, Texas. J. Petrol. 10, 272-306. STRECKEISEN,A. 1976. To each plutonic rock its proper name. Earth Sci. Rev. 12, 1-33. -1978. IUGS Subcommission on the systematics of igneous rocks; classification and nomenclature of volcanic rocks, lamprophyres, carbonatites and melilitic rocks. Neues Jb. Mineral. Abh. 134 (i), 114. STRONG, D. F. 1972. Petrology of the island of Moheli, western Indian Ocean. Geol. Soc. Am. Bull. 83, 389-406. SUTHERLAND,D. S. 1965. Nomenclature of the potassicfeldspathic rocks associated with carbonatites. Geol. Soc. Am. Bull. 76, 1409-12. -1967. A note on the occurrence of potassium-rich trachytes in the Kaiserstuhl carbonatite complex, West Germany. Mineral. Mag. 36, 334-41. SUWA, K., OANA, S., WADA, H. & OSAKI, S. 1975. - - &
Isotope geochemistry and petrology of African carbonatites. Phys. Chem. Earth, 9, 735-45. THOMPSON,R. N. 1974. Some high-pressure pyroxenes. Mineral. Mag. 39, 768-87. TREIMAN, A. H. & ESSENE, E. J. 1983. Mantle eclogite and carbonate as sources of sodic carbonatites and alkalic magmas. Nature, Lond. 302, 700-3. --&--1984. A periclase-dolomite-calcite carbonatite from the Oka complex, Quebec, and its calculated volatile composition. Contrib. Mineral. Petrol. 85, 149-57. - - & SCHEOL, A. 1983. Properties of carbonatite magma and processes in carbonatite magma chambers. J. geol. 91,437-47. TUTTLE, O. F. & GITTINS, J. 1966. Carbonatites. Wiley, New York. TWYMAN, J. D. & GITTINS,J. 1987. Alkalic carbonatite magmas: parental or derivative. In: FITI'ON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 85-101. TYLER, R. C. & KING, B. C. 1967. The pyroxenes of the alkaline igneous complexes of eastern Uganda. Mineral. Mag. 36, 5-21. VARNE, R. 1968. The petrology of Moroto Mountain, eastern Uganda, and the origin of nephelinites. J. Petrol. 9, 169-90. VARTIAINEN, H. 1980. The petrology, mineralogy and petrochemistry of the Sokli carbonatite massif, northern Finland. Geol. Surv. Finland, Bull. 313. VERWOERD, W. J. 1978. Liquid immiscibility and the carbonatite-ijolite relationship: preliminary data on the join NaFea+Si206-CaCO3 and related compositions. Ann. Rep. geophys. Lab. Wash. 77, 767-74. VILADKAR,S. G. & DULSKI,P. 1986. Rare earth element abundances in carbonatites, alkaline rocks and fenites of the Amba Dongar complex, Gujarat, India. Neues Jb. Mineral. 1986, H.1, 37-48. WATKINSON,D. H. & WYLLIE, P. J. 1971. Experimental study of the composition join NaAISiO4-CaCO3H20 and the genesis of alkalic rock-carbonatite complexes. J. Petrol. 12, 357-78. WENDLANDT, R. F. & HARRISON, W. J. 1979. Rare earth partitioning between immiscible carbonate and silicate liquids and CO2 vapour: results and implications for the formation of light rare earthenriched rocks. Contrib. Mineral. Petrol. 69, 40919. WOOD, D . A . , JORON, J.-L., TREUIL, M., NORRY, M. J. & TARN~Y, J. 1979. Elemental and Sr isotope variations in basic lavas from Iceland and surrounding ocean floor. Contrib. Mineral. Petrol. 70, 319-39. WOOLLEY, m. R. 1982. A discussion of carbonatite evolution and nomenclature, and the generation of sodic and potassic fenites. Mineral. Mag. 46, 1317. - & JONES, G. C. 1987. The petrochemistry of the northern part of the Chilwa Alkaline Province, Malawi. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline lgneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 335-55.
Nephelinites and carbonatites WYLLIE, P. J. 1965. Melting relationships in the system CaO-MgO-CO2-HzO, with petrological applications. J. Petrol. 6, 101-23. -1978. Silicate-carbonate systems with bearing on the origin and crystallization of carb0natites. Proc. 1st Int. Syrup. on Carbonatites, Ministerio das Minas e Energia, Departamento Nacional da
83
Produqao Mineral, Pogos de Caldas, Minas Gerais, Brazil, pp. 61-78. - - - 1980. The origin of kimberlites. J. geophys. Res. 85, 6902-10. YARDLEY,B. 1985. Apatite composition and fugacities of HF and HC1 in metamorphic fluids. Mineral. Mag. 49, 77-9.
M. J. LE Bks, Department of Geology, University of Leicester, Leicester, U,K.
Alkalic carbonatite magmas- parental or derivative? James D. Twyman & John Gittins S U M M A R Y : During the 1950s and 1960s the Tanzanian volcano Oldoinyo Lengai erupted a highly alkalic lava composed almost exclusively of the N a - K - C a carbonate minerals nyerereite and gregoryite and containing about 32~ Na20 and 7% K20. For many years the lava was considered a petrological curiosity but gradually the idea developed that alkalic carbonatite magmas might develop generally during the evolution of commoner carbonatite magmas. In 1981 Le Bas, followed in 1982 by Woolley, ascribed to the Oldoinyo Lengai magma a parental status and erected a scheme whereby other commoner carbonatite rock types are derived from it; the alkalic carbonatite magma is one of two derived by immiscible separation from a nephelinitic magma, the other being ijolitic. The carbonatite liquid has 500~ of superheat and loses alkalis progressively as the magma cools to its liquidus temperature of 400~176 by which stage it has become a calcitic-dolomitic liquid. This liquid then differentiates to produce the commoner types of carbonatite rocks. In rejecting such a scheme we argue as follows: such a scheme would require the magma to be water saturated at the moment of immiscibility, and progressive loss of water and alkalis would induce crystallization thus preventing the formation of a calcite-dolomite magma; the 500~ of superheat is a consequence of the design of the Freestone & Hamilton experiments and does not exist; carbonatite magmas generally begin to crystallize at temperatures greater than 900~ rare-earth-element compositions do not permit derivation of alkalic carbonatite magma from nephelinite magmas by liquid immiscibility; the mafic silicate mineralogy of carbonatites demonstrates alkali enrichment rather than depletion. We propose that the commonest parental magma of carbonatites is a mildly alkalic olivine s6vite composition and that most carbonatite rock types are derived from it by fractional crystallization, allowing the development of cumulates that are well lubricated by intercumulus liquid and capable of much subsequent movement, deformation and re-intrusion as a crystal mush. As this parental magma crystallizes, continued fractionation of alkalideficient and anhydrous minerals increases both its alkali and water contents until water saturation is reached. This in turn causes the development of an aqueous fluid which limits the alkali content of the magma. The composition of the magma at which water saturation develops is controlled by the rate of rise of the magma, and therefore aqueous fluids with various N a : K ratios can develop and control the type of fenitization that occurs. Under plutonic conditions alkalis in excess of the amount that can be fixed as silicate minerals are carried away as fenitizing fluids. Alkali loss does therefore occur, and through the medium of an aqueous fluid, but calcite and/or dolomite crystallize throughout the magmatic cooling history rather than simply at the low-temperature end of this history. The alkalic carbonatite magma of Oldoinyo Lengai type also develops through fractional crystallization of the same alkali-poor olivine s6vite magma, but under essentially dry conditions where the magma is kept liquid by alkalis and halogens, and alkali loss is prevented by the absence of an aqueous phase. Alkalic carbonatite magmas are therefore late differentiates of a more normal mildly-alkalic olivine s6vite magma developed under low water fugacity. They are of very small volume and require long quiescence to develop. They are not to be considered parental to other carbonatite rock types.
Introduction The problem of the c h a r a c t e r and origin of carbonatite m a g m a s is at once tantalizing and enigmatic. For m a n y years few geologists believed in the existence of m a g m a t i c c a r b o n a t e rocks at all, largely because of the high melting t e m p e r a t u r e of CaCO3 and failure to appreciate the effect of H 2 0 , F a n d alkali carbonates in lowering that temperature. T h e melting studies p e r f o r m e d by Wyllie and Tuttle a n d their coworkers b e t w e e n 1957 and the 1970s gradually o v e r c a m e these conceptual difficulties and the
petrological world was won over to the idea of carbonatite m a g m a s . Petrologists in the U. S.S.R., w h e r e there has always been a great propensity for m e t a s o m a t i s m , r e m a i n e d resistant to the idea of carbonatite m a g m a s , but they have n o w j o i n e d the igneous fold. Once the initial barrier of carbonate melting at reasonable t e m p e r a t u r e and pressure was overcome, the course of petrogenetic thought was g o v e r n e d largely by the melting studies of Wyllie and his students at the P e n n s y l v a n i a State U n i v e r s i t y and later at the U n i v e r s i t y of Chicago. T h e d e v e l o p m e n t of calcitic and dolomitic liquids
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 85-94.
85
86
J. D . T w y m a n & J. G i t t i n s
from various ultramafic ultrabasic compositions over a wide range of crustal and mantle pressures was pursued in this work. It must be remembered that this experimental programme was not designed solely to elucidate carbonatite genesis, but was at least as much concerned with the development of kimberlite magmas and magmas that might be parental to the basalts of the Earth's surface and upper crust. Consequently, Na and K did not figure prominently in these experimental systems. The general pattern emerged that calcitic and dolomitic liquids, given adequate H20 and CO2 fugacities, could develop from certain SiO2undersaturated peridotitic compositions such as might commonly be expected in the Earth's mantle. Consequently, they were thought to offer a probable explanation of how carbonatite magmas are formed. However, at the same time some petrologists were suggesting that these experimentally derived liquids were more applicable to the rather special type of carbonatite that is associated with kimberlites and the carbonate that so frequently pervades kimberlite intrusions, and that they might have relatively little relevance to the carbonatites of the classic carbonatitealkalic igneous rock association. Almost at the beginning of the melting studies an interesting turn of events was introduced when in 1960 the Tanzanian volcano Oldoinyo Lengai erupted its unusual N a - C a - K carbonate lava at temperatures too low to be incandescent. Earlier eruptions had, of course, been noted in the 1950s by Guest (1956), but were so unusual that the significance was not widely recognized. The initial tendency, with a few exceptions, was to consider it as a petrological curiosity of little application to the broader problems of carbonatite genesis, but it gradually assumed greater importance in the minds of certain petrologists. Dawson (1964) had originally suggested its importance and he was later supported by Cooper et al. (1975) and Gittins et al. (1975). In each of these papers it was argued that alkalic carbonatite magmas probably develop during the evolution of a carbonatite-igneous rock complex but that in most cases the alkalis are lost to aqueous or aqueous-halide fluids that may be involved in fenitization. The residues were thought to be normal calcitic and dolomitic carbonatites. In general it was felt that there was overwhelming evidence that most carbonatite magmas could not just be calcitic or dolomitic because of the frequent presence within the rocks of sodic pyroxenes, phlogopite-biotite and sodi-potassic amphibole, as well as fenitization adjacent to dykes and other intrusions; extreme examples were the presence of several centimetres of
phlogopite-rich rock in pyroxenite adjacent to carbonatite dykes. The concept emerged, and has now been accepted more widely, that carbonatite rocks do not necessarily represent the composition of the magma from which they crystallized. Le Bas (1981) argued the case that a highly alkalic magma of Oldoinyo Lengai type is parental to the full range of carbonatite rock types. His thesis was that carbonated nephelinite magma separates immiscibly into an ijolitic silicate liquid and an alkalic carbonate liquid at a temperature of 1000~176 The silicate liquid is said to be at its liquidus but the carbonate liquid is at least 500~ above its liquidus, with crystallization not commencing until 400~ 600~ The pressure at which the immiscible separation occurs was not stated but the context seems to imply lower-crustal rather than mantle conditions. The scheme envisaged by Le Bas involves the formation of a progressively lesssodic carbonate magma as alkalis are lost, presumably, although not clearly stated, during cooling through the 500~ superheat range, followed by crystallization differentiation that forms calcitic carbonatites at first and then increasingly more dolomitic and ankeritic carbonatites as the magma cools from its 400~ 600~ liquidus. Le Bas further suggested that the proportion of Ca: Na: K in the alkalic carbonatite magma may be a function of the temperature and pressure at which immiscible separation occurred, a conclusion drawn from the experimental work of Freestone & Hamilton (1980) and which he referred to as 'dynamic un-mixing'. The variable proportion of Na-rich to K-rich fenites was attributed to possible temperature,controlled variations in relative Na and K mobility, which Gittins et al. (1975) had also invoked. The following year Woolley (1982), while supporting the evolutionary concept of the formation of calcitic, dolomitic and ankeritic carbonatites from alkalic magma through alkali loss and fractional crystallization, introduced a modification by suggesting that the aqueous fluid becomes progressively more potassic. In this way he sought to explain the superposition of potassic fenitization on early sodic fenitization observed in some carbonatites. Thus the highly alkalic carbonatite magma of Oldoinyo Lengai was promoted from a stage in the evolution of carbonatite complexes (Cooper et al. 1975) to a parental magma derived immiscibly from a carbonated nephelinite magma. Purportedly, it gives rise, through progressive loss of alkalis followed by fractional crystallization, to the full diversity of carbonatite rock types. It is this parental status and proposed differentiation scheme with which we disagree. We
87
Alkalic carbonatite magmas argue our case by examining progressively the evidence of an immiscible origin, the evidence for or against alkali depletion, rare-earth-element (REE) data, the problems of generating and then abstracting an aqueous fluid phase and problems of magma temperatures.
Critique of the parental-status scheme The Freestone & Hamilton experiments Freestone & Hamilton (1980) established that molten nephelinite and phonolite are each immiscible with molten alkalic carbonatite lava. What was done was the following: a nephelinite and an alkalic carbonatite lava, both of which had erupted from the volcano Oldoinyo Lengai, were mixed together; the mixture was heated to a temperature above the liquidus of both rocks and then quenched in order to establish texturally that two immiscible liquids had coexisted. The experiments were repeated with phonolite and produced similar results. It was concluded that the silicate and carbonate liquids separated in the volcano from a common parent, for which phonolite was preferred to nephelinite. However, the experiments do not prove this. They establish that the two liquids are immiscible at 1 kb but not that they came into being together. There is a weakness in the design of the experiments. If the alkalic carbonatite is a highly evolved rock with a low liquidus temperature (about 500~ as we shall attempt to show, then melting it with a silicate rock whose liquidus is about 1000~ is bound to produce liquids with sharply divergent liquidi. The superheat is a consequence of the experimental design, and the argument is dangerously circular. Two rocks from the same volcano, but which may have been derived by fractional crystallization of two entirely different magmas, are melted together and found to produce immiscible liquids with liquidi 500~ apart. This is held to prove that the two rocks represent the compositions of magmas that were derived immiscibly from a common parent. The experiments prove what is possible but not that the possible happened. It is analogous to calculations that show a reaction to be thermodynamically feasible but not whether it is kinetically possible. We agree that alkalic carbonatite and silicate magmas are mutually immiscible, but we believe that the immiscibility occurred a long way back in the petrogenetic history, and in the mantle.
Alkali depletion or alkali enrichment? The arguments surrounding the alkalic carbonatite question have become rather theoretical with insufficient attention being paid to real rocks. Natural carbonatites can settle the question of whether alkali depletion occurs, for if it does the associated silicate minerals ought to reflect this depletion. Amphiboles from three calcitic and dolomitic carbonatite complexes are shown in Fig. 1. The direction of amphibole evolution is established by compositional zonation and is, from core to rim, an increase in Na + Fe accompanied by a decrease in Ca + Mg. In short, it is the reverse of what is required by the parentalstatus scheme. Rare-earth-element data It is possible to calculate the REE distribution in a carbonate liquid that would be in equilibrium with the Oldoinyo Lengai phonolite by using the partition coefficients published by Wendlandt & Harrison (1979). The result is shown in Fig. 2 where it is compared with the measured patterns of the alkalic carbonatite lava and both the nephelinite and the phonolite. Even after allowing for the fact that the nephelinite and phonolite analyses are not of the highest quality there is a considerable discrepancy between the calculated carbonatite pattern and the measured pattern, 0"7 [ " OLIVINE CARBONATITES l + CLINOHUM,CARBONATITES/
t( ,
~
l-Z+. ~ \
0.5
Co
9 ~%t
0.4
No+K 0.5 0.2 0.1 ol
"'s
~
,
I
l I0
,
I , 20 30
Mg/Fe FIG. 1. Compositions of amphiboles from the Argor and Goldray carbonatites, Ontario, Canada (Twyman 1983) and the Safartoq carbonatite, Greenland (Secher & Larsen 1980). The arrows indicate the direction of compositional change from core to rim. The petrographic varieties represented are those with amphibole alone, amphibole with titanian clinohumite and amphibole with olivine. The general trend of amphibole evolution is enrichment in Na, K and Fe with depletion in Ca and Mg.
88
J. D. Twyman & J. Gittins IO3
\.
,o'
-
--__-~ I I LO Ce
I Nd
I I 1 SmEu Tb
II YbLu
FIG. 2. Chondrite-normalized REE distribution for Oldoinyo Lengai lavas. The calculated composition of the alkalic carbonatite ( - - - - - - ) is based on the distribution coefficients of Wendtlandt & Harrison (1979) for a carbonatite in equilibrium with nephelinite and a phonolite whose measured values are taken from Gerasimovsky et al. (1972). The measured values of the alkalic carbonatite are taken from Twyman (1983) ( - - . . - - ) and Philpotts et al. (1972) (----). and the divergence increases towards the heavy REE. It has been argued already that the melting experiments performed by Freestone & Hamilton (1980) offer no proof that alkalic carbonatite magma was derived from nephelinite magma. The REE data provide further support for this view. The problem of generating an aqueous fluid from the alkalic carbonatite m a g m a
Progressive alkali loss is supposed to have occurred through 'exchange processes [with the] enveloping country rock' and 'by reaction' (Le Bas, 1981, pp. 138, 139). The aqueous fluid that has been invoked by other workers (cited earlier) seems to be inherent in the proposed alkali loss, and so such a fluid must have been generated from the carbonate liquid in some manner. That, in turn, requires the liquid to have reached water saturation. The available possibilities are that the carbonate liquid was saturated at the moment of immiscible separation from its alleged nephelinitic parent, or that it reached saturation during cooling. It is unlikely that the carbonate liquid reached saturation during cooling. The commonest process by which an aqueous fluid is developed from a magma is the progressive crystallization of anhydrous minerals which increases the concentration of water in the remaining liquid. Although the minerals of the Oldoinyo Lengai carbonatite lava are anhydrous, this process is not available
because we are asked to believe that the magma cooled through 500~ of superheat. It was therefore above its liquidus and so could not have been crystallizing. It is possible for a magma to reach water saturation as it rises to a regime of lower pressure but only if the solubility of water in it is relatively low. Solubility data for alkalic carbonatite magmas are incomplete, but it is probable that the solubility is very high and that saturation would not be reached in this way. The remaining alternative, that the carbonate liquid was water saturated at the moment of immiscibility, is also fraught with difficulty. This would require the ijolitic liquid as well as the alkalic carbonatite liquid to be saturated. Yet ijolites, being anhydrous rocks, appear to have crystallized from fairly dry magmas. Furthermore, nephelinitic magmas have fairly low water contents and so, for the alkalic carbonate magma to reach saturation, the available water would have to partition almost exclusively into the carbonate liquid. However, available data on water solubility in nepheline-bearing liquids suggest that the solubility of water in ijolitic magma would not be inconsiderable. The problem of generating an aqueous fluid within the constraints of the parental-status scheme appears to be insuperable, yet without such a fluid the scheme does not work. The effect of continuous removal of water and alkalis from a m a g m a
Suppose that an aqueous fluid could be generated in some manner. What would be the effect of its continued removal from the alkalic carbonatite magma? It has been well established experimentally, through the studies of Wyllie, Gittins and their coworkers, that calcitic and dolomitic liquids are possible primarily through the grace of water and alkalis. If either, or both, are removed, crystallization is the inevitable result. This is especially so since the solubility of water in calcitic and dolomitic liquids is markedly less than in sodi-potassic carbonate liquids. In short the proposed method by which the magma composition changes would be fatal to the magma's continued existence. In addition, under the proposed magmatic evolutionary scheme there is no reason why the Oldoinyo Lengai alkalic carbonatite lava should have erupted at all. And if it did, it ought to have been a vigorous eruption as its aqueous phase escaped and the remaining water in the saturated magma boiled offexplosively. In contrast, the 1960 eruption was quiescent and composed of anhydrous lavas. If the proposed alkali loss was not through the medium of an aqueous fluid but rather by wall-
89
Alkalic carbonatite magmas rock reaction, the same arguments apply. Reaction of such a magma with the enclosing silicate rocks must, inevitably, lead rapidly to crystallization.
Magma temperatures and crystallization temperatures The parental-status scheme has the crystallization of calcitic and dolomitic magmas commencing at 400~176 and is in accord with much popular thinking about the temperature ranges over which carbonatite magmas crystallize. Unfortunately, there is a good deal of misleading thought on this subject. Low temperatures (below 500~ are overemphasized, largely through misuse of the calcite-dolomite geothermometer. The fact that temperatures obtained from measurements of the amount of Mg substituting Ca in calcite that coexists with dolomite are minimum temperatures is often overlooked. If dolomite has exsolved from the calcite electron microprobe analysis cannot generally be used, and even if the calcite is homogeneous it cannot be assumed that there has been no diffusion of Ca and Mg between calcite and dolomite after crystallization, particularly if a fluid phase is present. Published crystallization temperatures are rarely over 550~ and many are as low as 200~ Yet it can be shown that in at least one carbonatite complex (Argor, N Ontario, Canada) a carbonate mineral began to crystallize at a temperature higher than 885~ The bulk composition of coarsely exsolved single crystals (74 CACO322 MgCO3-4 FeCO3) gives an equilibrium temperature of 885~ indicating that the liquidus temperature of the carbonatite magma was at least that high. Recrystallization and re-equilibration during cooling render accurate geothermometry in carbonatites very difficult and have often caused overemphasis on the 500~ range (Gittins 1979). Argor offers proof of the existence of a carbonatite magma at not less than 900~ and we suggest that this is an approximate liquidus temperature for many carbonatite magmas. It is, of course, in sharp contrast with the 400~ 600~ liquidus temperature required by Le Bas for the magma from which calcitic and dolomitic carbonatites crystallized.
The volume requirements of the alkali-depletion scheme A very general idea of the volume of material that would have to be removed from an alkalic carbonatite magma to produce a calcitic-dolomitic carbonatite can be obtained from a simple
subtraction diagram. This does not require knowledge of the fluid phase compositions since we can assume that everything in the fluid came originally from the magma. In Fig. 3 the data are the compositions of the Oldoinyo Lengai alkalic carbonatite lava, the solid composition of the subtracted phase and the average carbonatite of Gold (1966) which, despite the limited value of averages from that period, suffices for the purpose. The composition of the subtracted phase is taken at FeO = 0, and the diagram indicates that 95~ of the alkalic carbonatite magma must be removed to convert it into normal carbonatite. If the calculations are repeated for a subtracted phase with SiO 2 -- 0, the proportion is still 90%. If we are to assume that all normal calciticdolomitic carbonatites are the residue remaining after the removal of 9 0 ~ - 9 5 ~ of the mass of a parental alkalic carbonatite magma the volume requirements become unrealistically enormous, even if it were possible to have a magma still in existence after abstracting so much water and alkalis. Again, if wall-rock reaction rather than fluid involvement is responsible for the alkali loss the proportion of carbonatite to metasomatic encircling rock would have to be vastly different from what is generally observed.
Origin of alkalic carbonatite magma We have attempted to show that alkalic carbonatite magma is not parental to the more normal w t~ < -r
laJ
w
R.~o
o
----Z
Z
4 0 " ""~:
~"
OXIDE IO
0
I 2
5
4
5
6
S iO 2
FIG. 3. Subtraction diagram showing the volume and composition of a fluid phase that must be removed from an alkalic carbonatite to produce a normal carbonatite. See text for further discussion.
J.D. Twyman & J. Gittins
90
calcitic-dolomitic carbonatites. Yet it clearly exists. Where, then, does it come from? We believe that it can be generated by fractional crystallization of a mildly-alkalic olivine s6vite magma which we propose as a fairly common carbonatite parental magma. The arguments in support of such a composition are to be published separately but are based on the magnesian composition of carbonatite olivines (most commonly F095_85), which implies generation of the magma from a very mafic source, and upon the partition coefficients derived by Freestone & Hamilton (1980) from their melting studies. Although we do not believe that the results of their experiments demonstrate what they set out to prove, the data can be used to estimate the N a 2 0 and K 2 0 content of carbonatite liquids in equilibrium with immiscible silicate liquids of
various compositions. We believe that the olivine s6vite magma is the result of immiscible separation from an olivine nephelinite magma at a pressure of approximately 27 kb. The partition coefficients for nephelinite give N a 2 0 + K 2 0 18~, but a more primitive melanephelinite gives a value of about 8%. Our conclusion is that the commonest parent magma that generates the diversity of carbonatite rocks is an olivine s6vite with about 8% N a 2 0 + K20. A computer-modelling scheme has been employed to test whether an Oldoinyo Lengai alkalic carbonatite lava can be produced by fractional crystallization of this mildly-alkalic olivine s6vite magma. The compositions used are an olivine s6vite from the Argor intrusion (Ontario, Canada) with additional alkalis to bring N a 2 0 + K 2 0 up to 8.1~, and the mineral compositions from
TABLE 1. Results of the computer modelling of the differentiation of carbonatite magmas
SiO2 A12Oa TiOz FeO MgO MnO CaO Na20 K20 PEO5 CO2 Total
Weighting factor
Parent olivine s6vite
5.0 1.0 1.0 20.0 5.0 0.1 10.0 50.0 10.0 50.0 5.0
6.56 10.66 1.09 10.66 3.68 0.25 34.50 5.74 2.41 5.92 28.17 100.00
Target alkalitic carbonatite 0.66 0.33 0.11 0.33 1.34 0.16 17.80 33.86 8.18 1.09 36.34 100.00
Calculated target 0.66 0.33 0.12 0.33 1.34 0.14 17.85 33.84 8.30 1.09 36.66 100.44
Difference (%) 0.0 0.0 1.0 0.0 0.2 - 14.6 0.3 - 0.1 1.5 0.0 0.9
Minerals removed ct SiO2 A1203 TiO 2 FeO MgO MnO CaO NazO
1.77 0.68 0.36 51.07
ilm
51.70 48.30
mag
apt
3.05 96.84 0.11 57.10
dt
58.39
42.69 9.93 0.45 16.50 20.05 0.08 0.09
9.52 14.00 1.65 29.34
2.04 8.16 1.74
P2Os Total Solution
bt
0.03 12.57 17.07
K20 CO 2
amph
10.21
42.90 46.13
45.50
100.00
100.00
100.00
100.00
-44.67
-1.57
-6.80
-13.40
100.00 -3.78
100.00 -9.95
ct, calcite; ilm, ilmenite; mag, magnetite; apt, apatite; amph, amphibole; bt, biotite; dt, dolomite. 83.87~ crystallization of parent olivine s6vite will generate a residual alkalic carbonatite.
100.00 -3.70
Alkalic carbonatite magmas the Argot carbonatites. The computer program is a modification by Dr. M. P. Gorton (University of Toronto) of the linear-programming and leastsquares method P E T M I X due to Wright & Doherty (1970). It is essentially a reiterative regression analysis that minimizes the squares of the residuals. The results of the calculation are given in Table 1, and in summary it can be said that 84~ crystallization can generate a lava of Oldoinyo Lengai type by removal of minerals that would form a rock composed of 53% calcite, 4 ~ dolomite, 16~ apatite, 8 ~ magnetite, 2 ~ ilmenite, 12~ biotite and 5 ~ amphibole. This will be recognized as a common type of s6vite. We do not claim a definitive status for the calculation but we believe that it does demonstrate that alkalic liquid can be derived through crystal fractionation of mildly-alkalic olivine s6vite magma. Further support for this concept comes from the chemical compositions of the amphiboles and phlogopite-biotites from the Argor intrusion. Throughout most of the crystallization of the magma there is very little change of A1 relative to Mg/Fe in amphibole (Fig. 4) or of Ti relative to Mg/Fe in biotite (Fig. 5), but in a small number of highly evolved rocks there is a substantial increase amounting to tenfold for A1 and thirteenfold for Ti. Both elements can be treated as trace elements and the curves are strikingly similar to trace-element enrichment curves for liquid predicted by Rayleigh fractionation with bulk crystal-liquid partition coefficients of less than unity. These enrichments can be used to estimate the amount of crystallization required to produce them: about 9 0 ~ - 9 2 % They offer further support for the concept that alkalic carbonatite magmas are produced by fractional crystallization.
An alternative petrogenetic scheme We propose that the alkalic carbonatite magmas known to have erupted from Oldoinyo Lengai, and possibly from other volcanoes (Dawson 1964; Hay & O'Neil 1983, Deans & Roberts 1984; Hay 1984), are of small volume and are derived from a more primitive parental olivine s6vite magma (with about 8 ~ alkalis) by fractional crystallization and gravity separation. Differentiation of this magma can follow more than one course. Development
of normal
carbonatite
The general course of differentiation, whether wet or dry, is toward increasingly alkalic compositions, and this is portrayed by the silicate minerals (principally the amphiboles). Initial magma temperatures are high (close to 1000~
91 -I
0.5
-!0 9 8 7
0.4 AI 0.:5
~
6
5 2-64 CAI
0.2 0.1 0 0
i'. "A~
5 2 " .,,v~- I J t t I I t I J 0 I 2:54 5 6 78 9
Mg/Fe FIG. 4. Compositions of amphiboles from the Argor carbonatite showing variation of A1 with Mg/Fe. Note the dramatic increase in AI as Mg/Fe falls below a value of 2. 15
0.30 0.25 0.20
7.5
CTi o
5
CTi
"
Ti 0.15 0.10 ~o
0.05 0
9
i
i
I
2
9
~ oo
i 5
I .
i
i
4
5
:o
i 6
9
~
.25
I
7
Mg/Fe
FIG. 5. Compositions of biotites from the Argor carbonatite showing variation of Ti with Mg/Fe. Note the dramatic increase in Ti as Mg/Fe fails below 2. and are similar to those of the associated silicate magmas. The alkali content of about 8 ~ and the fairly low water content are sufficient to keep it liquid at high pressure. (A study of the system calcite-dolomite with sodium and potassium carbonate at 1 kb (Beckett 1986) has established that at 8 ~ N a 2 0 the system would have extensive liquid at 900 ~ C.) Crystallization of dominantly anhydrous minerals and, initially, minerals of low alkali content (calcite, dolomite, apatite and Fe-Ti oxide) drives up both the alkali and the water contents which become increasingly concentrated in a diminishing volume of magma and progressively reduce its viscosity. At some stage the magma will become water saturated and an aqueous fluid will separate carrying with it most of the alkalis that have not already been fixed as silicates. The stage at which this occurs will be strongly dependent on the pressure and hence upon the rate at which
92
J. D. Twyman & J. Gittins
the magma has risen through the crust. Thus it is possible for separation of the aqueous phase to occur at various degrees of alkali enrichment according to the rate of magma rise. For example, a rapidly rising magma will not have sufficient time for crystal fractionation to produce a high degree of alkali enrichment before water saturation occurs and so the escaping fluid will carry off very little N a 2 0 + K 2 0 . In contrast, a magma that has risen slowly will have fractionated to an advanced degree, generating a high water content, and so its aqueous fluid will be very alkalic and a powerful fenitizing agent. The wide variations in the extent of fenitization associated with carbonatite complexes can be explained in this way. Woolley's (1982) suggestion that the relative proportion of sodic to potassic fenitization is depth dependent also fits into this scheme. In this scheme magmas of varied alkalic enrichment are possible as a result of fractional crystallization, the degree of enrichment being limited only by the stage at which water saturation develops and an aqueous fluid phase separates carrying off alkalis in excess of the amount that can be fixed as silicates. We suggest that, under plutonic conditions, the process is prevented, by continuous alkali loss, from producing a magma of the composition of the Oldoinyo Lengai lava.
Development of alkalic carbonatites How then does the extreme alkalic carbonatite magma develop? The discussion so far has focussed on two factors that serve to reduce the liquidus temperature of our proposed parental carbonatite magma: increasing water content and increasing alkali content. However, water has the further effect of removing alkalis once water saturation is reached, and so it seems unlikely that the extreme alkalic composition can be reached in a hydrous magma. We pointed out earlier that the aqueous fluid phase will control the alkali content of the magma. Furthermore, the mineralogy of the Oldoinyo Lengai lavas shows that the magma was very dry when erupted. The answer to the dilemma seems to be that the extreme alkalic magma is produced by fractional crystallization under fairly dry conditions. We have already noted that an olivine s6vite with about 8% alkalis will be largely liquid at 900~ but factors other than alkali content can be invoked to reduce the liquidus in such a magma. The Oldoinyo Lengai lava contains as much as 8% F and as much as 4% C1. F is not at all uncommon in carbonatite complexes as indicated by fluorite (see yon Eckermann (1948) on Aln6) and the extensive substitution of F for OH in many carbonatite amphiboles and micas. F
and C1, together with the alkalis, would maintain the parental carbonatite magma in a liquid state throughout its crystallization history. Increase in alkali content, through continued fractionation of alkali-free minerals and a very small percentage of alkali-bearing minerals, is able to continue uninterrupted by alkali loss since an aqueous fluid phase does not develop. A continuum of magmatic compositions from the initial 8% alkalis to the Oldoinyo-Lengai-type extreme of about 37% is possible, but the latter can be reached only if the magma is dry. In this scheme the majority of carbonatite rock types would be formed by cumulus processes as crystals sink through a very fluid magma. They are not cumulates in the passive sense of, for example, gabbro cumulates. The magmatic plumbing system in a carbonatite complex is probably much more dynamic so that the cumulates, with their lubricating interstitial liquid, are readily disturbed, re-intruded and deformed by pulses of new magma. Much of the chaotic intrusive structure of carbonatite complexes can probably be explained in this way.
Conclusions Alkalic carbonatite magma of the type erupted as lava in the Tanzanian volcano Oldoinyo Lengai is not a parental liquid from which other carbonatite magmas and rock types are derived. Rather, it is a late-stage liquid derived through fractional crystallization of an olivine s6vite magma containing about 8% N a z O + K 2 0 . In the parental-status scheme nephelinite magma divides into two immiscible liquids, of which one is ijolitic and the other is an alkalic carbonatite composition with about 500~ of superheat. During cooling, the latter continuously loses alkalis through the medium of an escaping aqueous fluid phase or by wall-rock reaction, and it changes composition until it becomes a calciticdolomitic magma which differentiates further to give rise to a wide variety of residual products. We have argued that, although the details are not brought out in the original paper, the scheme requires that the alkalic carbonatite liquid must be water saturated from the moment of its creation and that progressive loss of water and alkalis will induce crystallization rather than development of a calcitic-dolomitic magma. We have argued that the scheme is not supported by the geothermometry of carbonatite rocks or by REE studies and that mineral zonation in carbonatite amphiboles implies alkali enrichment in the magma rather than alkali depletion.
Alkalic carbonatite magmas The 500~ of superheat is a consequence of experimental design rather than a logical deduction from unconstrained experiments, and does not exist. We propose an alternative petrogenetic scheme in which the parent magma, from which carbonatite rock types are derived, is a mildly-alkalic olivine sSvite composition with about 8 ~ total alkalis. This magma probably originates by immiscible separation from an olivine nephelinite magma deep within the mantle, a proposal that makes less extreme demands on element and component partitioning than does the parentalstatus scheme. During its subsequent rise into and through the crust it undergoes fractional crystallization in which anhydrous and alkalipoor minerals sink, leaving the residual liquid progressively more enriched in alkalis and water. Water saturation is eventually reached but this can occur over a wide range of compositions and pressures governed by the rate at which the magma rises and the time available for differentiation. Rapid rise with little time for fractionation will generate saturation at a relatively low alkali content. Slow rise will generate water saturation at a much later stage and a higher alkali content because of the much higher solubility of water in alkalic carbonatite liquids than in calcitic-dolomitic liquids. The alkali content of the rocks is limited by the amount that can be fixed as silicates and this in turn is limited by the activities of Si and AI, both of which are
93
very low. Separation of an aqueous fluid phase when water saturation is reached controls the alkali content of the remaining magma, and is itself controlled by the solubility of alkali carbonates and halides in supercritical water and by the rate at which anhydrous alkali-poor minerals continue to crystallize. Thus in most plutonic settings the crystallizing magma will lose alkalis at some stage when water saturation is reached; this may be a longcontinued process providing a source of fenitizing fluids. The pre- and post-saturation history of the magma will probably control the N a : K ratio of the fluid and have a profound bearing on the type of fenitization that ensues. A major difference from the parental-status scheme is that the water content of the magma gradually increases through fractional crystallization until saturation is reached, rather than escaping continuously from an already alkalic liquid that is merely cooling but not crystallizing. If the parental olivine s6vite magma has exceptionally low water and high halogen contents, alkali enrichment might reach a very advanced stage and even develop the extreme composition of the Oldoinyo Lengai lava. This will be possible because the magma does not reach water saturation and so alkalis are able to accumulate continuously without being abstracted by an aqueous fluid. Alkalic carbonatite magmas are late derivatives and not primary liquids.
References BECKETT, M. F. 1986. Phase relations in alkali-bearing dolomite carbonatites; effect of alkalinity and fluorine content on the solubility of pyrochlore in carbonatite magma. M.Sc Thesis, University of Toronto, Canada (unpublished). COOPER, A. F., GITTINS,J. & TUTTLE,O. F. 1975. The system NazCO3-KzCO3-CaCO 3 at 1 kilobar and its significance in carbonatite petrogenesis. Am. J. Sci. 275, 534-60. DawsoN, J. B. 1964. Reactivity of the cations in carbonatite magmas. Geol. Assoc. Can. Proc. 15, 103-13. DEANS, T. & ROBERTS,B. 1984. Carbonatite tufts and lava clasts of the Tinderet foothills, western Kenya: a study of calcified natrocarbonatites. J. geol. Soc. Lond. 141, 563-80. VONECKERMANN,H. 1948. The alkaline district of Aln5 Island. Sver. geol. Unders., Ser. C, 36. FREESTONE,I. C. & HAMILTON,D. L. 1980. The role of liquid immiscibility in the genesis of carbonatites--an experimental study. Contrib. Mineral. Petrol. 73, 105-17. GERASIMOVSKY,V. I., ALASHOW,Yu. A. & KARPUSHINA, V. A. 1972. Geochemistry of the rare earth
elements in the extrusive rocks of the East African rift zones. Geochem. Int. 1972, 305-19. GITTINS, J. 1979. Problems inherent in the application of calcite-dolomite geothermometry to carbonatites. Contrib. Mineral. Petrol. 69, 1-4. - - , ALLEN, C. R. & COOPER,A. F. 1975. Phlogopitization of pyroxenite; its bearing on the composition of carbonatite magmas. Geol. Mag. 112, 5037. GOLD, D. P. 1966. The average and typical chemical composition ofcarbonatites. Proc. 4th Gen. Meeting Inter. Mineral. Assoc., 1964, India, 83-9. GUEST, N. J. 1956. The volcanic activity of Oldoinyo L'Engai, 1954. Tanganyika geol. Surv. Rec. 1954, 4, 58-9. HAY, R. L. 1984. Natrocarbonatite tephra of the volcano Kerimasi, northern Tanzania. Geology, 11, 599-602. -& O'NEm, J. R. 1983. Carbonatite tufts in the Laetolil Beds of Tanzania and the Kaiserstuhl in Germany. Contrib. Mineral. Petrol. 82, 403-6. LE BAS, M. J. 1981. Carbonatite magmas. Mineral. Mag. 44, 133-40.
94
J. D. Twyman & J. Gittins
PHILPOTTS, J. A., SCHNETZLER,C. C. & THOMAS,H. H. 1972. Petrogenetic implications of some new geochemical data on eclogitic and ultrabasic inclusions. Geochim. cosmochim. Acta, 36, 113166. SECHER, K. & LARSEN, L. M. 1980. Geology and mineralogy of the Sarfartoq Carbonatite Complex, southern West Greenland. Lithos, 13, 199-212. TWVMAN, J. 1983. The generation, crystallization, and differentiation of carbonatite magmas: evidence from the Argor and Cargill complexes, Ontario. PhD Thesis, University of Toronto (unpublished).
WENDLANDT, R. F. & HARRISON, W. J. 1979. Rare earth partitioning between immiscible carbonate and silicate liquids and CO2 vapor: results and implications for the formation of light rare earthenriched rocks. Contrib. Mineral. Petrol. 69, 40419. WOOLLEY, A. R. 1982. A discussion of carbonatite evolution and nomenclature, and the generation of sodic and potassic fenites. Mineral. Mag. 46, 13-7. WRIGHT, T. L. & DOHERTY, P. C. 1970. A linear programming and least squares computer method for solving petrologic mixing problems. Bull. Geol. Soc. Am. 81, 1995-2008.
JAMES D. TWYMAN* & JOHN GITTINS, Department of Geology, University of Toronto, Toronto M5S 1A1, Canada. * Present address: Arco Oil and Gas Company, Lafayette, LA, U.S.A.
The kimberlite clan: relationship with olivine and leucite lamproites, and inferences for upper-mantle metasomatism J. B. Dawson S U M M A R Y : Kimberlites are rare ultrabasic potassic low-volume melts that originate in the diamond stability field of the upper mantle. They are petrographically complex, and wide
mineralogical and chemical variations suggest that they should be regarded as a group or clan rather than a single narrowly-defined rock type. In S African kimberlites there is a major division between highly-micaceous (Group II) kimberlites that are mineralogically, chemically and isotopically distinct from poorly-micaceous (Group I) kimberlites. Nd and Sr isotope studies of diamondiferous olivine and leucite lamproites from NW Australia indicate that they are isotopically similar to Group II kimberlites but are distinct from Group I. New bulk rock analyses of Group II kimberlites indicate closer chemical similarities with olivine lamproites than previously supposed. and carbonatisation. Kimberlite commonly contains inclusions of upper mantle derived ultramafic rocks. Variable quantities of crusKimberlites are rare ultrabasic potassic igneous tal xenoliths and xenocrysts may also be rocks occupying small vents, sills and dykes. present. Kimberlite may contain diamond but They are petrographically complex because of only as a very rare constituent.' variations in texture (diatreme facies versus The number of qualifications (my italics) in this hypabyssal facies (Dawson & Hawthorne 1970)), definition serves to illustrate the petrographic variability and particularly the range of mineral variations in xenolith and xenocryst content, variable content and proportions of macrocrysts, species developed in the matrix. The mineralogy and variations in the mineralogy of the fine- is also complicated by the fact that some species grained matrix, all of which combine to give a (e.g. calcite and apatite) that are relatively wide range of chemical composition (Dawson common in hypabyssal-facies kimberlites have 1967, 1980). These variations, particularly in not developed to the same extent in diatrememineralogy, have created difficulties in defining facies kimberlites, presumably owing to volatile kimberlite, as exemplified by the recent definition l o s s during diatreme formation; also, certain of kimberlite by Clement et al. (1984): species such as monticellite are extremely suscep'Kimberlite is a volatile-rich potassic, ultratible to calcitization under high Pco2 conditions. basic igneous rock which occurs in small Despite these complications it can be seen that volcanic pipes, dykes and sills. It has a individual kimberlites, although not necessarily distinctively inequigranular texture resulting containing identical matrix minerals, do possess from the presence of macrocrysts set in a some permutation of the minerals listed by finer-grained matrix. This matrix contains, as Clement et al. (1984), and their similar and prominent primary phenocrystal and/or overlapping mineralogies suggest that they should groundmass constituents, olivine and several be regarded as a group or clan of closely-related of the following minerals: phlogopite, carbonrock types rather than a single tightly-defined ate (commonly calcite), serpentine, clinopyrock type. roxene (commonly diopside), monticellite, There is nonetheless a fundamental division apatite, spinels, perovskite and ilmenite. The within the kimberlite clan, recognized as such by macrocrysts are anhedral, mantle-derived, Wagner (1914), between 'lamprophyric' kimberferromagnesian minerals which include olilites having a highly micaceous matrix, and the vine, phlogopite, picroilmenite, chromian spiso-called 'basaltic' variety containing little or no nel, magnesian garnet, clinopyroxene mica in the matrix. (The term 'basaltic' is quite (commonly chromium diopside), and orthopyinappropriate and Dawson (1980) has proposed roxene (commonly enstatite). Olivine is exthat it should be dropped.) In addition to the tremely abundant relative to the other major difference in the phlogopite content of the macrocysts, all of which are not necessarily matrices of these two types of kimberlite, there present. The macrocrysts and relatively-earlyare certain other differences. formed matrix minerals are commonly altered 1 In the case of S Africa, the micaceous by deuteric processes, mainly serpentinisation
Varieties of kimberlite
From:FITTON, J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 95-101.
95
J. B. Dawson
96
kimberlites were intruded 120-200 Ma ago which contrasts with the non-micaceous varieties intruded 80-90 Ma ago. 2 The micaceous kimberlites tend to contain considerable amounts of matrix calcite and apatite. 3 The micaceous kimberlites generally also contain groundmass diopside, which is quite rare in the non-micaceous types (Dawson et al. 1977). 4 Although macrocrystal picro-ilmenite is common in the non-micaceous kimberlite, it is relatively rare in micaceous varieties. 5 The micaceous kimberlites are occasionally accompanied by other high-K minor intrusives, as at the Helam Mine in the Swartruggens area of the western Transvaal where an E-W-trending dyke-swarm contains both diamondiferous kimberlite and non-diamondiferous sanidine-nepheline-lamprophyre (the so-called 'Male' dyke) (Skinner & Scott Smith 1979). Another major difference between the micaceous and non-micaceous kimberlites in S Africa has recently been recognized on isotopic grounds (Smith 1983). The two groups are quite distinct by reason of their U, Pb, Sm, Nd and Sr contents
and isotopic ratios, which are interpreted as reflecting substantial chemical differences in the mantle source rocks for the two groups. The nonmicaceous kimberlites (denoted Group I by Smith) are derived from undifferentiated to slightly depleted sources relative to bulk earth (Sr and Nd systematics) but with high U/Pb ratios. In contrast, the micaceous kimberlites (referred to as Group II by Smith) are derived from an enriched source with high Rb/Sr and Nd/Sm ratios but with low U/Pb ratios; it was suggested by Smith that this enrichment event took place in excess of 103 Ma ago. The isotopic differences between the Group I and Group II kimberlites are well illustrated on a Sr versus Nd isotope plot (Fig. 1) on which the Helam Mine lamprophyre dyke, referred to above, falls close to the micaceous kimberlites. In summary the term 'kimberlite' in S Africa has been used to cover rocks that are distinct petrographically and is 9 and which, furthermore, were intruded at different times. Their common factors are that they are both products of low-volume ultrabasic upper-mantle melts originating in the diamond stability field of the Earth's upper mantle in the late-Jurassic~Sr
-50
0
50
100
150
200
I
]
I
I
i
i
I~MORB I
[] Ohvme " ' lamproltes "
\ ' ~ \ DEPLETED MANTLE (DM) '~'~
9 Leucite lamproites
~ ,
~" Helam lamprophyre
" ..... . a/ ,J,h oc~' K~.~ ~3". ~ o ~"N~. 9'BULTFONTEIN / \ "
/'%
-4
o~J .... /OCEAN ISLANDBASALTS
Gpl ~"-~ \ KIMBERLITES ~ ,, / /
9 Phases in Bultfontein metasomites
~'\
"~
~/(Sr/Nd)gM
_ 2
~'~~Sr/~d)EM
!_Sr/Nd!DM _ 10 ~ (Sr/Nd)EM ~ \ . -12
?-
-
~:r-
~.~AUSTRALIA 9
-'---ZD
Gpll KIMBERLITES
-~.
[] "'0---_.
~ . ~
_
.
//
-16
0.705
I 0.710
I 0.715
ENRICHED MANTLE (EM) I 0.720
87Sr/86Sr (I) FIG. 1. Nd versus Sr isotope plot for kimberlites, lamproites and Bultfontein metasomites. Data for kimberlites and the Helam lamprophyre are from Smith (1983) and those for the Bultfontein metasomites are from Kramers et al. (1983). The lamproite data, the points for depleted and enriched mantle and the depleted/enriched mantle value curves are from McCulloch et aL (1983).
The kimberlite clan early-Cretaceous period. The proposal to regard kimberlite as a clan is particularly appropriate now that it has been shown that differing members have been derived from isotopically distinct areas within the upper mantle.
Relationship between kimberlites and lamproites Although diamond has been mined in the C6te d'Ivoire from dykes described as 'fitzroyites' for some years (Knopf 1970), the discovery of diamonds in olivine- and leucite-bearing lamproites in W Australia has stimulated a number of studies into the relationship between kimberlites and lamproites (Mitchell 1981 ; Atkinson et al. 1984; Jaques et al. 1984; Nixon et al. 1984). In addition, a re-investigation of the diamondiferous 'kimberlite' at Prairie Creek, Arkansas, U.S.A., has shown more mineralogical affinities with W Australian lamproites than with kimberlite (Scott Smith & Skinner 1984). Together, all these studies are unanimous in recognizing distinctions between kimberlites and lamproites on the following grounds. 1 The lamproites contain glass. 2 The lamproite groundmass contains potassic richterite. 3 Compared with typical groundmass micas in kimberlites (Smith et al. 1978) the lamproite micas are richer in Ti, Fe and Na, but poorer in A1. 4 The groundmass diopsides in the lamproites have higher Ti contents than those in micaceous kimberlite (Dawson et al. 1977) (diopside from the Helam Mine lamprophyre resembles that in the lamproites). 5 Compared with groundmass spinels in kimberlites, those in Prairie Creek lamproite have lower A1203 and MgO contents and are relatively restricted in overall compositional range. 6 There is a paucity of garnet and magnesian ilmenite in the W Australian rocks compared with the S African kimberlites. 7 It is claimed that, compared with the average kimberlite analyses by Dawson (1967), the olivine lamproites of Kimberley are characterized by higher SiO2, TiO2, P2Os and particularly K20 contents, but by lower A1203 and FeO and considerably lower CaO and CO2 contents. 8 The mode of emplacement of the lamproites is different from the kimberlites in that some of the lamproite magma intruded to high level and apparently consolidated as lava lakes within craters. The mineralogical differences between kimberlites in general and leucite lamproites are indis-
97
putable, although it should be appreciated that phlogopite (high pressure) is chemically equivalent to the low-pressure combination of leucite and forsterite, and hence the difference could be due to pressure differences at the time of crystallization. Calcite is virtually absent in the lamproites where the main Ca-bearing species are richterite and diopside; diopside is present in Group II kimberlites but contains less TiO2 (typically less than 1~) (Dawson et al. 1977) compared with values generally in excess of 1~ in olivine lamproites (Jaques et al. 1984). The presence of relatively high Ti contents within the silicates (it is also high in the richterite and in groundmass phlogopite) and the relative lack of variety of ilmenite and rutile in the lamproites contrasts with kimberlites where the reverse is true. There are, however, some mineralogical similarities between olivine lamproites and Group II kimberlites. The presence of diopside (although with TiO2 differences) and olivine in two generations is common to both groups, as is abundant perovskite and apatite; moreover, although the rims of lamproite groundmass micas are higher in TiO2 and Fe 3 + but lower in A1203, the cores of these micas are compositionally very similar to the kimberlite groundmass Group II phlogopites (Smith et al. 1978). The main differences are the absence of calcite in the lamproites and the presence in them of glass and richterite; the latter features (together with the relatively quiet emplacement of the lamproite intrusions) are perhaps due to low COz contents in the original magma. However, superseding all these differences is the presence of diamond which gives this group of K-rich, low-volume melts the distinctive fingerprint of having originated in the diamond stability field of the upper mantle. Since the mineralogy is directly related to the bulk chemistry, in addition to the effects of changing oxygen fugacity, temperature and pressure during consolidation, it is also pertinent to examine the suggested differences in bulk composition between kimberlites and lamproites; these differences were recognized by Jaques et al. (1984) from a comparison of lamproite analyses with the average kimberlite analyses published by Dawson (1967). It is proposed here to compare the chemistry of several new analyses of highlymicaceous Group II kimberlites with that of the olivine lamproites since even superficial comparison indicates no overlap with the more siliceous leucite lamproites; it should be noted that the new analyses presented in Table 1 all show higher K20 contents than that of the average micaceous kimberlite (Dawson 1967) used by Jaques et al. (1984) in their comparisons. Table 1 shows the chemistry of Group II kimberlites, the sanidine-
98
J . B . Dawson
leucite lamprophyre from the Helam Mine and W Australian olivine lamproites. The Group II kimberlites are poorer overall in most oxides but are considerably enhanced in CaO and CO2 relative to the Australian rocks, and it is apparent that these other oxides are mainly 'diluted' by modal calcite. Table 2 shows the Group II kimberlites recalculated without CaCO 3 and H 2 0 - compared with an average olivine lamproite and a range of lamproite compositions recalculated on the same basis; oxide ratios of the type used by Jaques et al. (1984) in their comparative study are also given in Table 2 and it should be noted that only the A1203/CaO ratio is affected by the recalculation. In the case of SiO2 the kimberlites have only slightly lower absolute values than the lamproites and have very similar recalculated values. Similarly total iron and MgO are similar in both groups of rock, as are K 2 0 and P205. Furthermore MgO/SiO 2 and KEO/A1203, which were used by Jaques et al. (1984) to distinguish between the two groups, are very similar, and the Group II kimberlites have even higher K 2 0 / N a 2 0 ratios than the lamproites. As observed by Jaques et al. (1984) the lamproites have lower TiO2, Al203 (slightly), CaO and CO2 contents, but in other respects are similar.
Of the trace elements, the lamproites are considerably richer in Ba, Rb and Zr, richer in N b and Ni, and similar in Pb, Zn and Cu compared with the kimberlites; however, they have lower V, Cr, Sr and Y contents (Table 3). These values give rise to relatively high Rb/Sr, T i / N b and Zr/Nb ratios and relatively low K/ Rb, K/Ba and Ca/Sr ratios in olivine lamproite compared with Group II kimberlites; Ti/Zr ratios overlap for the two groups (Table 3). Scott Smith and Skinner (1984) used a N b / Z r plot to illustrate differences between the Prairie Creek and Australian lamproites and the S African kimberlites. Figure 2 is a similar plot incorporating new data from Jaques et al. (1984) and Atkinson et al. (1984) for lamproites, together with the Group II kimberlite data from Table 3. The olivine lamproites lie between the kimberlites and leucite lamproites on the Zr axis. The Group II kimberlites lie within the general field and trend of S African kimberlites in general ( N b / Z r = 2 ) but also lie very close to the position of the diamondiferous Argyle olivine lamproite. In the case of the rare-earth elements (REE) the pattern (light R E E enriched) and abundances for olivine lamproites overlap those for S African Group II kimberlites from Swartruggens (Mitchell & Brunfelt 1975) and the Bellsbank dykes (Fesq et al.
TABLE 1. Analyses o f Group H S African kimberlites compared with N W Australian olivine lamproites (wt,%) I
2
3
4
5
6
7
8
SiO 2 TiO2 A1203 Fe203 FeO MnO MgO CaO Na20 K20 H20 + H20P20 5 CO2 F SO3
36.53 1.58 4.72 1.94 3.86 0.61 11.81 17.40 0.25 3.98 0.91 0.38 0.48 13.3 n.d. 0.74
36.28 1.42 4.16 5.85 3.11 0.21 22.89 7.99 0.14 4.98 4.40 1.52 1.37 3.73 n.d. 0.03
39.26 1.75 4.88 5.59 3.27 0.42 19.81 6.30 0.14 4.37 4.58 2.54 0.71 4.18 0.01 0.09
38.66 1.72 4.89 4.16 4.58 0.16 29.82 5.49 0.17 5.24 4.00 3.18 1.61 2.77 0.03 0.60
37.53 1.50 4.10 3.23 4.46 0.13 23.74 6.33 0.16 4.27 5.46 2.43 1.22 3.42 0.02 0.37
42.82 1.46 5.23 6.42 3.62 0.15 18.70 8.34 1.10 1.58 4.59 4.60 0.75 0.34 NA NA
42.6 3.43 3.96 -8.30* 0.14 25.0 5.05 0.52 4.45 4.96 -1.27 0.18 ---
45.0 3.32 4.84 3.00 4.66 0.12 21.21 4.88 0.46 5.50 3.01 0.67 1.58 0.50 NA NA
Total
98.49
98.08
97.93
98.18
98.37
99.70
--
98.83
* Total iron expressed as FeO 1, BD 1088, Zout en Zuur dyke, Boshof District; 2, BD 1089, New Elands dyke, Boshof District; 3, BD 1102, Gordonia dyke, Theron District; 4, BD 1268, Main dyke, Helam Mine, Swartruggens, Transvaal; 5, BD 1267, Changehouse dyke, Helam Mine; 6, BD 1269, sanidine-leucite lamprophyre dyke ('Male' dyke), Helam Mine; 7, average of 21 Ellendale, W Australia, olivine lamproites (Jaques et al. 1984); 8, magmatic olivine lamproite, AK1 Argyle diamond mine, W Australia (Atkinson et al. 1984) (total includes 0.09 wt.~ BaO). Analyses 1-5 by X-ray fluorescence (analyst, R. Kanaris-Sotiriou) except FeO, H20, CO2 and F (wet methods; analyst, A. Saxby); analysis 6 by wet methods (analyst, J. R. Baldwin).
The kimberlite clan
99
TABLE 2. Analyses of Group H kimberlites and olivine lamproites calculated free of CaC03 and H20-
(wt.% ) 1
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO Na20 KzO HzO § P205 SO3 KzO/A1203 AI203/CaO K20/Na20 MgO/K20 MgO/SiO2
2
53.80 2.32 6.95 2.86 5.68 0.90 17.40 0.71 0.37 5.86 1.34 0.70 1.09 0.8 9.8 15.8 3.0 0.32
3
41.20 1.61 4.72 6.64 3.53 0.24 25.99 3.68 0.16 5.66 4.98 1.55 0.03 1.2 1.3 35.4 4.6 0.63
4
43.86 1.95 5.45 6.24 3.65 0.47 22.13 1.12 0.16 4.88 5.12 0.80 0.10 0.9 4.9 30.5 4.5 0.50
5
43.59 1.94 5.51 4.69 5.16 0.18 23.48 2.22 0.30 5.90 4.51 1.81 0.67 1.1 2.5 19.7 4.0 0.54
42.57 1.70 4.65 3.66 5.06 0.16 26.93 2.25 0.18 4.84 6.19 1.38 0.42 1.0 2.1 26.9 5.6 0.63
6
7
42.8 3.44 3.97 -8.46 0.14 25.10 4.83 0.52 4.47 4.98 1.28 -1.12 0.82 8.5 5.6 0.6
40.71-42.61 2.71-5.77 3.30-4.51 3.80-4.94 2.37-4.98 0.11-0.14 19.04-26.90 4.06-5.61 0.31-0.63 3.46-5.11 3.74-4.93 0.60-1.48 0.9-1.3 0.6-1.1 8.1-12.2 3.7-7.5 0.4-0.7
Analyses 1-5 from Table 1 recalculated on a CaCO 3- and H20-free basis; analysis 6 is the average of 21 Ellendale (W Australia) olivine lamproites (recalculated free of CaCO3 and H 2 0 - and with all Fe as FeO); analysis 7 is a range of six olivine lamproite analyses (Jaques et al. 1984). 1975). H o w e v e r , the closest t r a c e - e l e m e n t similarities b e t w e e n G r o u p II k i m b e r l i t e s a n d olivine l a m p r o i t e s c a n be seen on an N d versus Sr isotope plot (Fig. 1). T h e i m p l i c a t i o n s f r o m this plot are that, like the l a m p r o i t e s , G r o u p II k i m b e r l i t e s are d e r i v e d f r o m a n c i e n t e n r i c h e d m a n t l e and, in this p a r t i c u l a r respect, h a v e m o r e affinities w i t h l a m p r o i t e s t h a n w i t h the G r o u p I kimberlites.
Relevance to upper-mantle metasomatism B o t h the k i m b e r l i t e s a n d the d i a m o n d i f e r o u s l a m p r o i t e s o r i g i n a t e at d e p t h s in excess o f 150 k m in the u p p e r m a n t l e a n d their differences in c h e m i s t r y reflect differences in t h e source regions.
TABLE 3. Trace-element analyses of micaceous kimberlites and N W Australian lamproites (ppm) 1
Cr Ni Ba Pb Zn Cu Rb Sr Y Zr Nb
144 1212 799 1445 47 115 30 142 647 25 189 100
K/Ba K/Rb Rb/Sr Ca/Sr Ti/Nb Ti/Zr Zr/Nb
22.9 223 0.219 192 94 50 1.9
V
2
118 1401 945 2192 26 82 50 185 1618 26 347 96 18.8 162 0.114 35 43 12 3.6
3
167 1631 812 3459 23 65 35 224 865 21 218 110 10.5 176 0.258 52 95 48 2.0
4
5
6
7
151 1266 768 3456 47 87 46 257 836 35 400 156
103 1265 857 1770 40 76 38 198 1133 27 283 107
72 1126 1024 9714 32 75 58 452 1102 16 784 175
19-142 528-1703 673-1500 3867-18281 20-50 61-90 52-93 300-611 959-1245 10-20 564-1214 118-244
13.1 179 0.307 47 57 22 2.6
20.0 159 0.175 40 82 31 2.6
3.4 0.41 31 116 26 4.5
1, BD 1088; 2, BD 1089; 3, BD 1102; 4, BD 1268; 5, BD 1267; 6, 7, mean and range of NW Australian olivine lamproites (Jaques et al. 1984; Table 5, analyses 1-6). The analyst for 1-5 was R. Kanaris-Sotiriou.
J. B. Dawson
I00 6O0
500
9
High-K kimberlites
[]
Kimberley kimberlites
z~pc Prairie Creek lamproites
400
o
Olivine lamproite
9
Leucite lamproite I N ' W
IA
] Australia
Argyle lamproite
E O. O. 300 ..Q Z 200
DD/
DDC'/
~176176I
t~
0I
~
/
O
Oo
OA
,...~,
~
f~pc
5 0I0
I
m
/
I 1000
1500
20100
Zr ppm
FIG. 2. Nb versus Zr plot for kimberlites and lamproites. Data sources: Kimberley kimberlites, Clement (1982) high-K kimberlites, new data (this paper); Argyle lamproite, Atkinson et al. (1984); Prairie Creek, Arkansas, lamproites, Scott Smith & Skinner (1984); NW Australian lamproites, Jaques et al. (1984).
Jaques et al. (1984) envisage a source region of harzburgitic composition (to account for the low AI and Ca in the lamproites) that has been subsequently subjected to K, Ti, Ba, Rb and Zr enrichment. In the case of the Group II kimberlites their source region must have contained some garnet (to account for the higher A1 and Cr relative to lamproites) and carbonate; although the K, Nb, Pb and Sr enrichment was similar to that of the lamproites, the kimberlite source region was not so enriched in Ti, Ba, Rb and Zr. Conversely, the relatively young enrichment event that is reflected in the Group I kimberlites resulted in the addition of less K, but higher Ti, to the Group I source region. All these differences highlight the fact that, although the same groups of elements are added during upper-mantle enrichment, there are specific differences from place to place in the upper mantle and, in the case of S Africa, at different times. The case in S Africa is particularly complex because, in addition to the old Group II and younger Group I enrichment events, the enrichment event that produced the veined and metasomatized xenoliths from the Bultfontein Mine (Kramers et al. 1983; Dawson 1987), although producing an isotopic signature identical with that of the Group I enrichment event (Fig. 1), took place in the amphibole stability field, i.e. at a considerably
shallower depth in the upper mantle than the kimberlite source region.
Conclusions
Recent bulk chemical and isotope data have reinforced the older petrographic distinctions between mica-rich (Group II) and mica-poor (Group I) kimberlites. The Group II kimberlites have closer chemical similarities to olivine lamproites than to Group I kimberlites. Nevertheless, all kimberlites and diamondiferous lamproites originate within the diamond stability field of the upper mantle. Their chemical and isotopic differences point to different degrees of initial melting followed by differing amounts and type of enrichment, at different geologic times, of the source mantle from which the different types of kimberlite and lamproite were derived. As such, they offer further convincing evidence of the temporal and spatial inhomogeneity of the upper mantle. ACKNOWLEDGMENTS: The kimberlite analyses were made by R. Kanaris-Sotiriou, A. Saxby and J. R. Baldwin. The figures were drawn by M. Cooper and the manuscript was typed by P. Mellor. I extend my thanks to all of them.
The kimberlite clan
I 01
References ATKINSON, W. J., HUGHES, F. E. & SMITH, C. B. 1984. A review of the kimberlitic rocks of Western Australia. In: KORNPROBST, J. (ed.) Kimberlites I." Kimberlites and Related Rocks, pp. 195-224. Elsevier, Amsterdam. CLEMENT, C. R. 1982. A comparative geological study of some major kimberlite pipes in the Northern Cape and Orange Free State. PhD Thesis, University of Cape Town (unpublished). , SKINNER, E. M. W. & SCOTT SMITH, B. H. 1984. Kimberlite re-defined. J. Geol. 92, 223-8. DAWSON, J. B. 1967. Geochemistry and origin of kimberlite. In: WYLLIE, P. J. (ed.) Ultramafic and Related Rocks, pp. 269-78. Wiley, New York. -1980. Kimberlites and their Xenoliths, 252 pp. Springer, Berlin. 1987. Metasomatized harzburgites in kimberlite and alkaline magmas: enriched restites and 'flushed' lherzolites. In: MENZIES, M. A. & HAWKESWORTH, C. J. (eds) Mantle Metasomatism. pp. 125-144 Academic Press, London. & HAWTHORNE, J. B. 1970. Intrusion features of some hypabyssal South African kimberlites. Bull Volcanol. 34, 740-57. , SMITH, J. V. & HERVIG, R. L. 1977. Late-stage diopside in kimberlite groundmass. Neues Jb. Mineral. Abh. 1977, 529-43. FESQ, H. W., KABLE, E. J. D. & GURNEY, J. J. 1975. Aspects of the geochemistry of kimberlites from the Premier Mine and other selected South African occurrences with particular reference to the rare earth elements. Phys. Chem. Earth, 9, 687-707. JAQUES, A. U, LEWIS, J. D., SMITH, C. B., GREGORY, G. P., FERGUSON, J., CHAPPELL, B. W. & MCCULLOCH, M. T. 1984. The diamond-bearing ultrapotassic (lamproitic) rocks of the West Kimberley region, Western Australia. In : KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks. pp. 225-54. Elsevier, Amsterdam. KNOPF, D. 1970. Les Kimberlites et les Roches Apparentbes de C6te d'Ivoire, 202 pp. SODEMI, Abidjan. -
-
-
-
KRAMERS, J. D., RODDICK, J. C. M. & DAWSON, J. B. 1983. Trace element and isotope studies on veined, metasomatic and 'MARID' xenoliths from Bultfontein, South Africa. Earth planet. Sci. Lett. 65, 90-106. MCCULLOCH, M. T., JAQUES, A. L., NELSON, D. R. & LEWIS, J. D. 1983. Nd and Sr isotopes in kimberlites and lamproites from Western Australia: an enriched mantle origin. Nature, Lond. 302, 400-3. MITCHELL, R. H., 1981. Titaniferous phlogopites from the leucite lamproites of the West Kimberley area, Western Australia. Contrib. Mineral. Petrol. 76, 243-51. & BRUNFELT, A. O., 1975. Rare earth element geochemistry of kimberlite. Phys. Chem. Earth, 9, 671-86. NIXON, P. H., THIRLWALL, M. F., BUCKLEY, R. & DAVIES, C. J. 1984. Spanish and West Australian lamproites: aspects of whole rock geochemistry. In: KORNPROBST, J. (ed.) Kimberlite I: Kimberlites and Related Rocks, pp. 285-96. Elsevier, Amsterdam. SCOTTSMITH, B. H. & SKINNER, E. M. W. 1984. A new look at Prairie Creek, Arkansas. In : KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 255-83. Elsevier, Amsterdam. SKINNER, E. M. W. & SCOTT SMITH, B. H. 1979. Petrography, mineralogy and geochemistry of kimberlite and associated lamprophyre dykes near Swartruggens, western Transvaal, R.S.A. Extended Abstracts, De Beers Kimberlite Symp. 11, Cambridge. De Beers, Cambridge. SMITH, C. B. 1983. Pb, Sr, and Nd isotopic evidence for sources of southern African Cretaceous kimberlites. Nature, Lond. 304, 51-4. SMITH, J. V., BRENNESHOLTZ,R. & DAWSON, J. B. 1978. Chemistry of micas from kimberlites and xenoliths. I. Micaceous kimberlites. Geochim. cosmochim. Acta, 42, 959-71. WAGNER, P. A. 1914. The Diamond Fields of Southern Africa. 355 pp. Transvaal Leader, Johannesburg. -
-
J. B. DAWSON, Department of Geology, University of Sheffield, Mappin Street, Sheffield S1 3JD, U.K.
Lamproites and other potassium-rich igneous rocks: a review of their occurrence, mineralogy and geochemistry* Steven C. Bergman S U M M A RY : In this paper the geological occurrence, geochemistry and mineralogy of ultrapotassic (KzO/Na20 > 3 (molar ratio)) and perpotassic (K20/A1203 > 1 (molar ratio)) igneous rocks, especially lamproites, are reviewed and discussed in the context of compositionally-similar mantle-derived melts. Lamproites are K- and Mg-rich igneous rocks (typically K20 > 5 wt.%, MgO > 5 wt.%) which possess an exotic and diagnostic mineralogy and geochemistry. Known lamproites occur in 21 major suites or localities in continental regions with a variety of geological and tectonic environments; they range in age from the early Proterozoic dykes at Holsteinsborg, W Greenland, and Chelima, India, and Precambrian pipe at Argyle, W Australia, to the Middle Pleistocene flows and the Recent volcanics of the Leucite Hills, Wyoming, and Gaussberg, Antarctica, respectively. Intrusive and extrusive forms of lamproites include flows, a variety of pyroclastics (welded tufts, piperno, air-faU tufts, volcanic breccias etc.), cinder cones, dykes, sills and diatremes. Whereas kimberlite diatremes tend to be carrot shaped, the shape of olivine lamproite diatremes approximates a sherbet-glass. The recent discovery of diamondiferous lamproites of large volumetric proportion in the E and W Kimberleys, NW Australia, and the reclassification of the diamondiferous micaceous peridotite at Prairie Creek, Arkansas, as a lamproite substantiate their economic importance. The 21 lamproite suites considered here tend to be localized marginal to continental craton Cores in areas that overlie fossil Benioff zones, in contrast with the general occurrence of kimberlites more interior to continental cratons. The petrographic diversity of lamproites has historically hindered the development of a concise and universal classification and nomenclature. Lamproites are distinguished from kimberlites and alkali basalts (and lamprophyres) in terms of mineralogy, mineral chemistry, geochemistry and volcanic extrusive character. Relative to kimberlites, lamproites are enriched in K, Si, Ti, AI, Rb, Sr, Zr and Ba and depleted in CO2, Ca, Mg, Fe, Ni, Co and Cr. Lamproites are characterized by the general presence of phlogopite, diopside, leucite and K-richterite, occasional glass, olivine, sanidine, priderite, perovskite, wadeite, apatite and chrome spinel, and very rare ilmenite. Lamproite amphiboles, diopsides and phlogopites are distinctly depleted in A1~O3 relative to those of nearly all other igneous rocks. Lamproite magmas are produced by the partial melting of old refractory mantle peridotite (approaching a dunite or harzburgite in mineralogy) that was enriched in K-bearing and other incompatibleelement-enriched phases, such as phlogopite and apatite, most probably as a result of some metasomatic event which occurred prior to melting. In contrast with alkali basalt and kimberlite melts which are apparently produced from the partial melting of a CO2-enriched mantle peridotite (i.e. a source with a relatively high CO~/H20 ratio), water is the key volatile species involved with lamproite petrogenesis (source with a low COz/H20 ratio).
Introduction 'Lamproite' designates a K-rich mafic to ultramafic alkaline igneous rock as originally defined by Paul Niggli and P. J. Beger in 1923 in the context of Niggli's detailed chemical system of igneous rock classification. Relative to nearly all other igneous rock types, they possess high K / N a and K/A1 ratios. In addition, lamproite mineralogy and geochemistry are unparalleled. Niggli and Beger recognized two subtypes of the lamproite clan, based on the distinctive compo* This paper is dedicated to the memory of the late Th. G. Sahama who contributed much to the infant stages of the lamproite revolution.
sitions of rocks from the Leucite Hills, W y o m i n g , and SE Spain, a n d included fortunite, verite, orendite, prowersite, wyomingite a n d j u m i l l i t e in the clan. Tr6ger (1935) redefined the term 'lamproite' as the extrusive equivalent of lamprophyres that are rich in K and Mg. In the ensuing five decades, over 150 papers have been published on the subject of lamproites and an exceedingly complex terminology has been generated. Whereas the earliest studies e m p h a s i z e d the two classic localities in Spain and W y o m i n g , it is the purpose of this paper to summarize, compare and contrast the geological, geochemical and mineralogical characteristics of 21 lamproite occurrences on six continents and place these lamproite features in the context of compositionally allied
From: FITTON, J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 103-190.
IO3
Io4
S. C. Bergman
alkaline igneous rocks, including other ultrapotassic rocks, kimberlites, lamprophyres and alkali basalts. While not all K-rich rocks will be reviewed, several of the more important suites, geographically significant localities (because of close proximity to lamproites) or localities which have been previously described as lamproites but are not considered so in the present discussion will be included in this discussion. This paper is not meant to be an exhaustive review of K-rich rocks, but rather a springboard to the K-rich rock literature. Prior to the late 1970s lamproites were thought to represent petrological oddities of limited extent and importance with extreme and exotic mineral assemblages and chemical compositions (e.g. Turner & Verhoogen 1960). However, the discovery of diamond-bearing lamproites at Ellendale and Argyle, on the SW and E margins of the Kimberley craton in N W Australia (Atkinson et al. 1984a, b; Jaques et al. 1984), and the reclassification of the diamond-bearing micaceous peridotite at Prairie Creek, Murfreesboro, Arkansas, from a kimberlite (albeit anomalous) (Miser & Ross 1923; Bolivar 1977; Lewis 1977; Meyer et al. 1977) to a lamproite (Bolivar 1984; Scott-Smith & Skinner 1984a, b) have placed the lamproite clan in an important position in terms of both petrogenesis and economics. Not only must diamond-exploration models be revised to include lamproites but, in addition to kimberlites, certain members of a second group of igneous rocks must originate from within the diamond stability field in the Earth's upper mantle (depths of greater than 150 km) and be explosively transported to the surface in order to preserve diamonds that are unstable in these oxidized magmas. The diamondiferous sandy tuff lamproite at Argyle, W Australia, is among the largest (125 acres and more than 100 million tons) and is the highest grade (about 600 ct per 100 ton) igneous diamond deposit thus far discovered, dwarfing many of the S African kimberlite pipes (typically 20-80 ct per 100 tons in grade and 2060 acres in area) in the quantity of diamonds contained. This paper is organized as follows. After a brief consideration of the igneous geochemistry of K, Na and A1, previous reviews and benchmark papers on lamproites and other ultrapotassic igneous rocks will be briefly summarized. The nomenclature and classification oflamproites will then be discussed, followed by a review of the geological occurrence oflamproite and significant compositionally related suites. The major-, traceelement and isotope geochemistry of lamproites and their mineral chemistry will then be summarized and compared with those of compositionally
related rock types. Finally, after a review of experimental data involving lamproites, current theories on the petrogenesis of lamproites will be presented. It should be noted that, while the literature is presented in an objective forum, many of the author's opinions, which are not necessarily shared by all workers in this constantly debated arena of alkaline rocks, have infiltrated the review, especially in the discussion section.
K, Na and AI in igneous systems Important rock-forming minerals containing substantial proportions of both K and A1 include sanidine, biotite, leucite and rare kalsilite; the molar ratio K/A1 is nearly always less than unity for these minerals. Relative to both A1 and Na, K displays a remarkable regularity in its general distribution and abundance in igneous rocks (Fig. 1). In a bulk-rock sense, K is more incompatible than A1 and Na in basaltic and more differentiated igneous systems subjected to processes involving either partial melting or fractional crystallization. For example, both K/A1 and K/ Na increase in progressing from the 'bulk Earth' (carbonaceous chondrites) or the 'primitive upper mantle' to basalts (either alkaline or tholeiitic) as the result of partial melting, possibly combined with a small amount of fractionation (Fig. 1). These ratios both increase in the generalized sequence basalt, andesite, dacite, each progressively increasing in the degree of fractional crystallization that the given magma has experienced (Fig. 1). For the vast majority of igneous rocks, KzO/ N a 2 0 molar ratios are less than 1.0. As has been suggested by Johannsen (1931, 1933, 1935, 1939), those rocks with KzO/Na20 (molar) ratios of 13 are best termed 'potassic'. Relatively common rocks with K 2 0 / N a 2 0 ratios of about 1-3 include some granites, rhyolites, syenites, trachytes, latites, leucite tephrites, leucite basanites and minettes. Rarer rock types with a 'potassic' character include shonkinites and absarokites. Those rocks with KzO/Na20 ratios in excess of 3 should be termed 'ultrapotassic' and include some minettes, leucite phonolites, rare alkaline rocks (e.g. juvites), kimberlites and lamproites. To describe the relative amounts of the alkalis and A1, Shand (1927) used the term 'peralkaline' for those rocks with the sum of molecular K 2 0 and Na20 in excess of molecular A1203. The numerical value of the ratio (K20 + NazO)/A1203 has been named the agpaitic index or coefficient by Ussing (1912) and has been discussed by Polanski (1949) and Sorensen (1960). Peralkaline
Lamproites and K-rich igneous rocks ........
~g~[,:rE[ r " " 1
. _. . . . . . . . ~S-Y[NITE .,S j PHONOLITE(b- ~ " " " 9 TRACHYTE # #9 II / s 9 NEPHELINE SYENITE S S MONZONITE 9 ADAMELLITE S~ 9 ~TEPHRITE
0.3
,/
no
IS 5
J' S 9 LATITE -" s "~ /// s*~PRESENT U. CRUST (TAYLOR'& McCLENNAN, lg81) TRACHYANDESITE ~"~'0 ~ ii~ ~" ~,~ RHYODACITEiiii: ) /"
// / DEEP SEA PE. . . . . . . . . . r ~ . . . ~..,~, OCL,TM=r~/~ m
0.2
.."
9
.,..,,.
~iiii
ii iiiii::iii ~ I Y
.-"
ARCHEAN U. CRUST J" ~ "~GRANOD:IORI:TEJ .''" (TAYLOR & McCLENNAN, lg81) .'MUGEARITE 9 DACITES 9 TRAC:-HY~A~Xi.T;iiii~Am~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PRESENT / " BASANITE'*D~ ~.~. DEER SEA t E. CRUST (TAYLOR S~,* 9 9 / ICARBONATE ~ & McCLENNAN, Ig~1) ~,~ 9 DIORITE ...... .'" SEDIMENTS | .''HAWAIITE 9 ,~*"'~ 9 .~'~
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.1
OTS (MASS'N, lg67)7~i"t"
DOLERITE,ill' **"4 ~ ~ .jliO.. , , , / ~.,~ 9 ,._.__?..?A,~5 . . . . . . . . v.............. CARB~,/IC~\\\~II~.'~'~-:_~_-" - - ~NO-RITEO DU-~'~--CTB (MASON,1967)
2
AVE. ULTRAMAFIC ROCK (GOLES, 1967) EARTH MANTLE (TAYLOR. 1960)
CHON~'~7~.~g~"l' "'A~IORTHOSITE
T-PRIMITIVE .
0
(a)
.
.
.
I
.
0,5
.
.
.
10
1.
K201 Na20 (WT BASIS)
/ / / / / / /
A
o m rr
9
/ /
1.0 / /
9 NW~
/
/
O
2
0.5
V
/
SB
9 KAM
9
0
'
'
~
'
I
1
,
PEN ~ 2 0
ARG
LH 9 LAMPROITE
8OH
9 CSN 9 MAP AZZ 9 SALOTB / 9 KNX S~'~)9 HLM / CARBON__ MINETrE / NHB(DII 9 9 MUR FOR ]W 9 HM 9 CARBONATITES KIMBERLITE A 9 SANDSTONES t i l ~ N 9 NSW -- SHALES ULTRAMAFIC LAMP LAMPROPHYRE I
(b)
/
WKB ~ 2 2
9 MBAY 9 HP 9 TUSC PPN 9 "L,H*--O t GDW
/
/
BO_.?.•B50
GSB
9 HOL
,
,
9 PRA 9 ENO
117
BB AVERAGE ROCK TYPES 9 LAMPROITE SUITE AVERAGES 9 OTHER POTASSIC/ ULTRAPOTASSIC ROCK SUITES
I
10
5
CHE
!
!
K20 ! Na20 (~rr RATIO)
FIG. 1. (a) Plot of the covariation of K20/A1203 and K 2 0 / N a 2 0 for average igneous rocks and carbonaceous chondrites (Nockolds 1954; Goles 1967; Mason 1967, 1979; LeMaitre 1976) and a few sediments (Turekian & Wedepoh11961) compared with the estimated compositions of various portions of the Earth (continental crust (Clark & Washington 1924; Goldschmidt 1933; Poldervaart 1955; Ronov & Yaroshevsky 1969), primitive mantle (Taylor 1979), and lower, upper and Archaean crust (Taylor & McLennan 1981)). Note that Earth materials show a regular and positive correlation between the degree of'perpotassic' and 'ultrapotassic' character. (b) Same plo t as (a) except that the scale is reduced to include the average compositions of lamproites and related rocks (see text); the shaded region corresponds to the field of Earth materials in (a). Three sedimentary rock compositions are also shown. Lamproites are unparalleled in their degree of K enrichment relative to A1 and Na. See Appendix 3 for rock suite abbreviations. rock types include some rhyolites, trachytes, syenites, granites, minettes, lamproites and kimberlites. Most lamproites and some kimberlites (e.g. the Group II kimberlites of S Africa (Smith 1983a, b) and other m i c a c e o u s kimberlites) are a m o n g the only rock types that are so enriched in K 2 0 that it is present in excess of A1203 (atomic basis). J o h a n n s e n (1931, 1933, 1935, 1939) sug-
gested that the term 'perpotassic' be used to describe these exotic rocks. Therefore most lamproites and some mica-rich kimberlites and minettes share the unique quality of being a m o n g the only rock types that are both ultrapotassic and perpotassic. Those rocks whose elevated K contents are due to near-surface or late-stage m a g m a t i c alteration (such as the potassic altera-
IO6
S . C. B e r g m a n
and contrasts lamproites with lamprophyres and Dawson (1987) compares olivine-rich lamproites with micaceous kimberlites. Foley (1985) considers the oxidation state of lamproite magmas. 'Shoshonites' (i.e. those andesitic intrusive and Previous reviews extrusive rocks with K 2 0 / N a 2 0 near or greater Lamproites and other K-rich alkaline rocks have than 1.0, with slight ~enrichments in CaO and MgO and depletions in FeO and TiO 2 relative to been the subject of many important reviews and contributions, although no-one has attempted a NazO-rich alkali basaltic rocks) and related comprehensive integrated synthesis of tarn~- potassic_ro~s,_~vhich nre minor but ubiquitous components of many island arcs and continental proites. Various aspects of selected potassic and ultrapotassic rock suites have been discussed in margins, are much more aluminous than lammany benchmark papers and books on alkaline proites and have received attention in review and other igneous rocks (Jensen 1908; Daly 1910, papers by Joplin (1965, 1968) and Morrison 1914, 1933; Niggli 1923; Shand 1927; Smyth (1980). Recent contributions to the petrogenesis 1927; Terzaghi 1935; Saether 1950; Turner & of shoshonite magmas include Gest & McBirney Verhoogen 1960; Sobolev 1970; Bardet 1973, (1977), Boccaletti et al. (1978), Pagel & Leterrier 1974, 1977; Carmichaeletal. 1974; Stewart 1979; (1980), Ewart (1982), Keller (1983) and Friend & Janardhan (1984), and references cited therein. Smith 1984b; Venturelli et al. 1985; Foley et al., K-rich rocks occur as local variations within in review). Niggli (1923, pp. 182-5) and Beger (1923, pp. 448-51) were the first to synthesize the various portions of alkaline pyroxenite-bearing ultramafic plutons and gneiss complexes in the early 1920s understanding of ultrapotassic rocks. Their two subtypes of lamproites were character- Oman Mountains, Arabia (Searle 1984), at ized by the following Niggli values (Niggli values Ampipal in the central Nepal Himalayas (Lasare the ratio of the weight per cent oxide to the serre 1976), at Shaki and Okeho in the basement gram molecular weight, normalized to one cation complex of W Nigeria (Oyawoye 1976), in the Batbjerg complex in E Greenland (Gittins et al. • 1000; k = k/(na + k), mg = mg/(mg +fe)). 1980; Brooks et al. 1981), in the Blue Mountain Subtype si al fm c alk k mg c/jm complex, New Zealand (Grapes 1975), in S Murcia 140 16 55 13 16 0.80 0.80 0.23 Greenland (Upton & Thomas 1973) and in the Wyoming 165 18 41 14 27 0.85 0.75 0.35 Mordor complex in central Australia (LangworLacroix (1926) summarized the systematics of thy & Black 1978) among others. The petrogenetic syenitic leucite-bearing rocks including those of relationship between these rocks and lamproites the Leucite Hills and Gaussberg. On the basis of must await more detailed study. Many alkalic-carbonatitic complexes possess petrography and geochemistry, Sahama (1974) differentiated orenditic from kamafugitic rocks fenitized zones which are characterized by rocks in a review of K-rich igneous rocks; the former composed almost entirely of potassium feldspar (Heinrich 1966; McKie 1966; Verwoerd 1966; group corresponds to the lamproites discussed Heinrich & Moore 1970). Whereas these rocks here. A recent book reviewing mineralogical, petrological, geochemical and experimental as- possess some geochemical and mineralogical features common to lamproites, they tend to be pects of rocks that contain modal leucite (Gupta & Yagi 1980) includes a discussion of many of extremely depleted in MgO and it is clear that the lamproite suites synthesized and updated in their petrogenesis and paragenesis are distinct the present work. Wendlandt & Eggler (1980a) from those of lamproites. performed a statistical study of 835 leucitenormative Cenozoic volcanic rocks and recognized two compositionally-distinct sub-groups, one enriched in K20 and A1203 but depleted in Lamproite nomenclature and MgO and another enriched in MgO but depleted classification in K20 and A120 3. Interestingly, lamproites apparently did not fall in either group and were Lamproite terminology has historically been one evidently not included in the original data set of the most complex and disjointed of that for any group of igneous rocks (Table 1). In the first (RKNFSYS; Chayes 1976). Mitchell (1985) half of this century, a nomenclature was develcontributed a comprehensive review of the oped that centred around the mineralogy of mineralogy and mineral chemistry of lamproites individual lamproite suites with little regard to and Wagner & Velde (1986a) discussed the lamproite suites elsewhere in the world. Table 1 mineral chemistry of 10 K-richterite-bearing summarizes the sometimes redundant terms that lamproites from six suites. Rock (1987) compares
tion zones in some fenites and porphyry systems) are excluded from the present discussion.
Lamproites and K-rich igneous rocks
Io7
TABLE 1. Historic summary of lamproite terminology Rock term
Principal mineral and/or phenocryst GGM assemblage opx ol amph cpx phi leuc san
Wyomingite Orendite Madupite Cedricite Mamilite Wolgidite Fitzroyite Verite Jumillite Cancarixite Fortunite Cocite Kajanite Gaussbergite
x
x x x x
x x
x
x x x x x
x x x x x x x x
x x x x
x
x x
x)
x
Rock suite
Leucite Hills, U.S.A.
Original reference
Cross 1897
•
x x x x x x
x
Fitzroy Basin, W. Australia Wade & Prider 1940
•
x ~ Murcia-Almeria, SE Spain ) Murcia-Almeria, SE Spain Murcia-Almeria, SE Spain Coc-Pia, N Vietnam Oele Kajan, Borneo
x
x
x
x x x x
x
x
x
Gaussberg, Antarctica
Osann 1899 Parga-Ponda11935 De Yarza 1893 Lacroix 1933b Brouwer 1909; Lacroix 1926 Lacroix 1926
opx, orthopyroxene; ol, olivine; amph, amphibole; cpx, clinopyroxene; phi, phlogopite; leuc, leucite; san, sanidine; GGM, glassy groundmass.
have been introduced for mineral assemblages characteristic of individual or several lamproite suites. Note, for example, that fitzroyite and wyomingite are equivalent, and cocite and jumilite are mineralogically identical. Unfortunately, the most recent report of the IUGS Subcommission on the Systematics of Igneous Rocks, which specifically emphasized the classification and nomenclature of lamprophyres, carbonatites and other alkaline rocks (Streckeisen 1980), neglected to include lamproites in its discussion. Under the earlier IUGS scheme (Streckeisen 1967), lamproites would fall in the compositional fields of alkali trachyte, phonolite, phonolite foidite or leucitite. Tr6ger (1935) and Johannsen (1931, 1933, 1935, 1939) describe selagites from Italy which some workers (D. Velde, personal communication, 1985) place in the lamproite clan. However, selagites contain about 10~oplagioclase (oligoclase) and therefore do not meet the mineralogical criteria for lamproites discussed in more detail below. Although Middlemost (1975) included lamproites in the basalt clan, lamproites should not be considered as basalts (despite the appearance of basaltic cinder cones and lava flows in many lamproite occurrences such as the Leucite Hills). Lamproites do not contain plagioclase, a phase which is generally regarded as an essential component of basalts. Note, however, that Zirkel (1870) and Rosenbusch (1877) used the term 'leucite basalt' for rocks composed of leucite, augite and olivine which were devoid of plagio-
clase. If olivine lamproites are to be regarded as basalts, kimberlites must also be included in the basalt clan. Lamproites have historically been split up and placed in a variety of sometimes distinct rock groups; these classifications, furthermore, have hindered the development of a clear understanding of the lamproite clan and its petrogenesis. Rosenbusch (1887) classed the Prairie Creek, Arkansas, lamproite with biotite-peridotites but the Leucite Hills rocks with leucite basalts. Niggli (1923) introduced the term 'Mediterranean province type' for potassic alkaline rocks of syenitic to shonkinitic composition; the type locality was the Roman co-magmatic province whose rocks are not lamproites because of their metaluminous to peraluminous character. Nevertheless, some lamproite suites have been included in the Mediterranean magma type (e.g. Turner & Verhoogen 1960). Lacroix (1933) put lamproites in his Division II, Family A: missourites, leucitites and albanites. Tr6ger (1935) placed the Prairie Creek, Arkansas, lamproite in his glimmerite Family (no. 721, olivine-glimmerite). Other lamproites fall into Tr6ger's nepheline syenite (orendite) and shonkinite (cocite, wyomingite, gaussbergite, jumillite) families. Johannsen (1931, 1933, 1935, 1939) detailed the mineralogy and texture of many of the lamproites and ultrapotassic rocks known at that time and placed lamproites and other ultrapotassic rocks in six different families within three classes of his rock-classification scheme. These families
IO8
S. C. B e r g m a n
spanned a wide range of compositions, including feldspathoidal rocks, peridotites, trachytes, perknites and alkali syenites. Williams et al. (1954) included Leucite Hills lamproites in the leucite phonolite petrographic group. Rittmann (1951), who considered both mineralogy and geochemistry in devising his nomenclature for volcanic rocks, placed lamproites in the 'lamproitic phonolite', 'lamproitic leucitite' or'lamproite trachyte' groups. He additionally suggested different terms for established lamproite rock names, i.e. madupite~--pheno-phlogopite-mafitite~-~-lamproite leucitite,verite~pheno-mafitite(phlogopite)~lamproitic trachyte, and wyomingite~pheno-leucitemafitite----lamproitic phonolite. Sahama (1974) classified K-rich alkaline rocks into two groups, lamproitic (or orenditic) and kamafugitic, on the basis of chemistry and petrography. Barton (1979) recognized three sub-groups of K-rich alkaline rocks on the basis of bulk rock and mineral chemistry and petrography: the Leucite Hills type, the Toro-Ankole type and the Roman province type. The 21 lamproite suites considered below fall in both the Leucite Hills and Roman province fields in Barton's scheme; none, however, overlaps the Toro-Ankole field. On the basis of petrographic modal and textural variability, Mitchell (1985) suggested that two broad subdivisions of lamproites be made: phlogopitesanidine-leucite-diopside-lamproites and madupitic lamproites. The former group possesses resorbed phenocrystal phlogopite and includes wyomingites, orendites, fitzroyites, cedricites, verites etc. The latter group is characterized by poikilitic groundmass phlogopite and includes madupites, wolgidites and jumillites. As many lamproites possess an extremely finegrained or glassy groundmass, classification based on mineral abundances alone is doomed to failure because of the uncertainty in the mineral equivalent of the groundmass. Since bulk-rock composition reflects mineral chemistry and abundance of component phases (and vice versa), whole-rock geochemistry can also be used in establishing a classification system for lamproites. This is extremely useful for fine-grained rocks and is essential for glass-rich rocks. Nonethe-less, it is often difficult to pigeon-hole rocks solely on the basis of their geochemistry. ScottSmith & Skinner (1984b) suggested classifying lamproites on the basis of the modal abundance of principal primary minerals, following an analogous scheme proposed by Skinner & Clement (1979) and Clement et al. (1977) for kimberlites. Their six major divisions were phlogopite, K-richterite, olivine, diopside, sanidine and glassy lamproites. Mitchell (1985) has followed their suggestion but has criticized the elimination
of leucite from the list of major phases and the inclusion of glass in a mineralogical classification, two criticisms that are taken into consideration below. Therefore leucite is substituted for glass in the six major sub-divisions proposed by ScottSmith & Skinner (1984b). Minerals of secondary importance can be used as modifiers to the most abundant phases and permit further sub-divisions to these six major classes. Mitchell (1985) also recommended recognizing the texture of phlogopite in the classification, and suggested the following sub-divisions: phlogopite lamproites (in which phlogopite occurs as phenocrysts) and madupitic lamproites (in which phlogopite is poikilitic in the groundmass). Mitchell's mineralogical classification removes many of the redundancies of the classical terminology, and tends to over-simplify the mineralogy of a given rock. Nevertheless, until more is known about the genetic controls on the petrography of lamproites, this classfication scheme is the best alternative.
Lamproite defined The definition of lamproite proposed and used here is based on those of Scott-Smith & Skinner (1984b) and Mitchell (1985) but includes several other parameters. It is suggested that the term 'lamproite' be applied to those intrusive and/or (more commonly) extrusive rocks that fulfil the following mineralogical and/or geochemical constraints. Mineralogical conditions 1 Dominant (more than 5~-30%), although not essential, primary phases (groundmass or phenocryst) include titanian phlogopite (Al-poor), clinopyroxene (Al-poor diopside, more rarely diopsidic augite), Al-poor alkali amphibole (commonly K-Ti-richterite, more rarely K-riebeckite, K-arfvedsonite or other alkali amphiboles), olivine (often in aggregated grains but also euhedral), leucite (Na-poor, commonly Fe-rich) and sanidine (commonly Fe-rich). 2 Accessory (less than 5%-10%) but often characteristic primary phases include spinel (intermediate Cr-Mg-Fe-A1-Ti compositions), priderite ((K,Ba)(Ti,Fe 3+ )8016 ), wadeite (K,~Zr2Si6018), shcherbakovite ((Na,K) (Ba,K) Ti 2Si4014 ), jeppeite ((K,Ba)2 (Ti,Fe)6013 ), apatite, perovskite, sphene, armalcolite ((Mg,Fe)Ti4010), enstatite (rare) and ilmenite (very rare). 3 Alteration or secondary phases comprise analcime, chlorite, quartz, TiO 2 polymorphs,
Lamproites and K-rich igneous rocks carbonate minerals, nontronite (and other zeolites), chlorite, serpentine, barite (Sr-rich), albite and a variety of clay minerals. 4 If a rock carries any of the following, it falls outside the lamproite sensu stricto group: (a) primary plagioclase (late- to post-magmatic secondary albite can occur); (b) melilite or monticellite (places a rock in the more calcic kamafugitic, ultramafic, lamprophyre, kimberlite or basalt clans); (c) kalsilite (requires a SiO2 activity much lower than that observed in lamproites); (d) nepheline, sodalite, nosean and hafiyne (these sodic phases are not compatible with an ultrapotassic assemblage); (e) melanite (despite the TiOz-rich and Al-poor composition oflamproites, melanite contains too much Fe to be stable). 5 Xenoliths and xenocrysts are neither necessary nor sufficient to the definition of lamproites but are mentioned to permit further distinction of lamproites from kimberlites and ultramafic to mafic lamprophyres. Cr-rich pyrope xenocrysts, which are ubiquitous components of kimberlites, are rare components of olivine lamproites (e.g. Ellendale, Argyle, Prairie Creek). 'Dogtooth olivine' aggregated monomineralic xenoliths are a characteristic feature of many olivine lamproites (e.g. Ellendale, Prairie Creek (Scott-Smith and Skinner 1984a, b)) and are distinct from the ubiquitous rounded olivine macrocysts of kimberlites. Spinel and garnet lherzolite xenoliths are extremely rare in lamproites but common in kimberlites. The most common xenoliths found in lamproites (apart from ubiquitous shallowcrustal country-rock fragments) are biotitite, biotite-clinopyroxenite and pyroxenite xenoliths that are most probably cognate. 6 Mineral chemistry (see below) can also be used to distinguish between lamproites and other rocks.
10 9
Occurrences of lamproites and other potassic to ultrapotassic rocks This section summarizes geological and petrological data and literature pertinent to those localities which contain rocks that conform to the constraints discussed above and hence can be called lamproite sensu stricto. Localities containing potassic-ultrapotassic rocks which do not meet the requirements for classification as lamproites will also be considered, and the rationale for not including them in the lamproite group will be given. Tables 2 and 3 abstract the geological parameters of each locality and most of these are plotted in Fig. 2. Note that the abbreviations used to reference each locality will be used throughout this paper and, in particular, that those localities not recognized as containing lamproites sensu stricto are set in square brackets. An alphabetical listing of the abbreviations is given in Appendix 3. Figure 3 illustrates the variety of rock textures observed in the suites from Australia and N America, and Fig. 4 shows representative thin-section photomicrographs from 11 lamproite localities. N America and Greenland
Kimberlites, lamproites, carbonatites and allied rocks are widespread in N America (Fig. 5), however, in contrast with kimberlites and carbonatites, lamproites do not occur in the interior of the craton but are generally located near its margin (Meyer 1976; Turtle & Gittins 1966). The estimated age of the crustal basement to these lamproite localities varies from 1.5 x 103 to 2.5 • 103 Ma. Enoree Vermiculite district, S Carolina ( E V D ) (34 ~ 30' N, 82 ~ W)
Geochemical conditions
When a rock possesses the following general compositional features, it is suggested that it be included in the lamproite clan: K20/A1203 > 0.8 (molar), K 2 0 / N a 2 0 > 4 (molar), Mg number> 70, with general compositional ranges of 45-55 wt.~ SiO2, 4-10 wt.~ A12Oa, 1-5 wt.~ TiO2, 210 wt.% CaO, 5-10 wt.~ K20, 0.2-1.5 wt.~ Na20, 0.5-2.0 wt.% P205 and 1.0-3.0 wt.% BaO. It is recommended that both mineralogy and geochemistry be used to distinguish lamproites from other rocks. However, where mineral data are lacking or impossible to acquire (glassy rocks), these compositions can be used with caution.
Late Proterozoic to Early Palaeozoic vermiculite bodies occur throughout the Appalachians and Piedmont and are best exposed and studied near Enoree, SC, just W of the Kings Mountain Belt (Stewart 1949; McClure 1963; Libby 1975). Palaeozoic metamorphism produced a mineral assemblage of phlogopite, diopside, tremolite (or talc), K-feldspar, apatite, sphene, monazite and zircon. Metamorphism has clearly modified the primary mineralogy and only a few of the major lamproite phases are present. The geochemical character of the EVD rocks, provided that it has not been modified by the metamorphism, is diagnostically lamproitic and nearly identical with the geochemistry of the average lamproite (Tables 4 and 5).
S.C.
I I0
Bergman
O~ ~<:~
~
<.~'5
~z ~
c~
6
o~
.~
'6 ~
~.~ ~
~ .~
~
~
~ '~.~
~ ~':
6
6
g~
~ .~~ O
.~
~
.
-~ 9
~
~
-
~
~.~
~ - s ~ - ~ . ~ . ~ o
~ ' ~
~ ~ ~,.~-~.~
-
0
o~ ~, ,~ _ ~ ,;,.~ 0 ~ t~,0 0~. ~
.~
o.~ ~=
.~.~ ~ o~
6~
~~
=
9~
~
<" ~ ~ : ~
z
o~.~ g~o
.~
~
~ ~ ~
~'~
~
~
c~
~5 o
-,~-~ E'~ E~-
.=
~ ~ ~~
i
1
o~
~
.~~
~
=.~ 9
'~.~ ~ ~
~
~a
m
~ .~-~
!
"<6 0~
oo
2
.~
o
"~
.~
<
.o~ 0.-~
m,
2~
~
r,.)
~o ~
r.)
0 ~
~_
~.o~
~o
~
N
~
z c~
~q 9
~
~
~
~
z
~
b~
Lamproites and K-rich igneous rocks
III
o~ d~
~
~,~
~
e'~
"o ~ =
~
o.~ ~
~
"~
~..~
_
~
~~
~
~
-
,-~ . ~ ~-~
~'~
~,r ~
0
~'=
.
~ "~
"~ ~...'~
,..
~ 0
~.~
~_~
~
-~
~
g
"'~
"~'~'~o
0
'Fo
g
H~, .,-.
.-
~.~,
~ '~
~
~~
o
.~ ~
o~
~
g
g
<
<
6
.o
=~ .~.~ ~ 9.~.
,~
~.~ ~
"~
~ eO
o
~
v
a
~
o :~
o
g
,.-., o
~
~.~
5
e.
~,.0
0
~Z~ f.a~ 0
....
<-z~N
S . C . Bergman
I I2
.>: Zz ~
~.~<
~z
=~z~
-~ ~
.~- ~~~ o=
..t;
-a~
.~~
~
~<
-t,~.~ ~ .~
~
~z
.~-~ ~.~
~
~
-~ ~
~ ' ~ ~
~
~Z<
0
~
r~
0
v,
~ ~-. o~
.
6
>,~
~ 6
~r
~
..'-
~.
~
d
8 .-
.~
~ ~ .~ ;~ .~ ~
~ ~.
.
.
-
.
.
.
=
~
~~
~
~ o o~
~
~
.
F
. .= "=
t,
.~
.~
"'~
r
~= ~
,,.,
o
_
=~
~
~
~o .-
o
.~
.~
0
..=
~
o
.o
8
8
.
9 t'--',
.
.
.
.
t.~
~.
.
a
~
~
~
~
_
~
s
.~
~
.
~
= '-8
4
~d
~f~=Z a~
~
.= v - . _ . o
~ ._.
Lamproites and K-rich igneous rocks
z~5
~
7~
d0.
z~
=~
<~
~ ~3
~<<~ o < ~
~... .
.m
.~.
I
m~
~ ~ -9 o~.~.,~
~
.~'~'~
~_ 2
~.--
",-
~0
.~- ~
"" ..~
~.~ ~
~'H ~
~ C.'~ ,~ e ~ . -o ~
~2
[.,
o
~ ~
<
"~..u ~ =
~
~o
o
..m
<2
2
t.,
{o b
22~
.~
.~
9,-
.~
2oo =
2 2 2 0
o"
~ ~
4 O
O
O
~zz~
.9
S
z
r
~z ~
a]z~
S~
~ .~~ t ~
F~.~
~
~-~<
= o~
o~<
~'
~
I ~4
S . C . Bergman
"
SYNNYR(~
.... " "
O LJ
0 PAMiR OPAMIR GONDWANA~.j ~
c~b.A
"
[]
.
~.
w. K,~B=RL=YS . , , : ~ . . . ~ ,AR~Y,, ~ -,,,
.
-
LAMPRorrE ~OCAUT'~
#
HILLS\I~NO" LPRAIRIE CREEK
'
0 0 T . E R ~T*SS.C, OLTR.,'OTAS~.C ROCK LOCALITY
,~'="U'I-G'~,Ri.~.'I
,
NAVAJO-HOP,B U . E S //~(.r
[ ] COO .PIA
"
MeRCiA ,',LMERIA
LEU
r==~COALFIELDS
""~'~ .
"
]
-~
. "
',.
~,--I"
~
~
I
\\
% o g ..... ~:~
K
--"~J
/ ,~,KO b A \ . ; U"I
. . . . . .
P,,E=TL~,EA
~
.O.,,=~O~LA
=
r
=,SCO ~
~
'
I
"
~
-'*" GAUSSBERG
FIG. 2. World map showing the locations of lamproite sensu stricto suites and some of the compositionally-similar potassic-ultrapotassic rock suites. For a more complete listing, see Tables 2 and 3 and text. Hills Pond, Kansas ( H P ) (37 ~ 40' N, 95 ~ 45' W)
Cretaceous (88 Ma (Zartman et al. 1967)) lamproite plutons, sills and dykes intrude Precambrian granites of the Rose Dome and nearly flatlying Pennsylvanian shales and limestones near Hills Pond, Woodson County, SE Kansas (Twenhofel & Bremer 1928; Knight & Landes 1932) (Fig. 6). About 13 km SW of the Rose Dome, a 'mica peridotite' (madupitic lamproite (P. Berendsen, personal communication, 1984)) of approximately the same age (90-91 Ma (Zartman et al. 1967)) intrudes the Silver City Dome (Bickford et al. 1971; Franks 1966). The HP rocks possess all the earmarks of olivine lamproites and consist of Ti-phlogopite, serpentinized olivine, K-richterite and Ti-diopside in a fine-grained groundmass of serpentine, perovskite, apatite and chrome spinel (Merrill et al. 1977) (Fig. 4). Other petrological and geological contributions concerning the HP intrusives and their xenoliths have been made by Wagner (1954), Franks (1959, 1966), Franks et al. (1971) and Cullers et al. (1985). Fresh rocks have been recovered by drilling; the surface exposure of the HP body has been intensely altered to form a vermiculite body. The petrography of madupitic (K-richterite-diopside) lamproites from the Silver City Dome area are discussed in Berendsen et al. (1985). Other lamproite diatremes occur in the area (H. Coopersmith, personal communication, 1985).
The geochemistry of the HP olivine lamproite is nearly identical to those of the diamondiferous olivine lamproites from Ellendale, Argyle and Prairie Creek. Evidently, the quiescent intrusive character of the Hills Pond magmas, manifested by the preponderance of thin sills, was not conducive to the preservation of diamonds, assuming the magmas were derived from a similar depth and source rock as the three diamondiferous suites mentioned. The HP area is one of the rare occurrences in which lamproite forms welldeveloped sills and the only known occurrence of olivine-lamproite in well-developed sills. Holsteinsborg, W Greenland ( H O L ) (66 ~ 45' N, 53 ~ IV)
Proterozoic lamproite (1200 Ma) and post-tectonic kimberlite (585 Ma) dykes (less than 1-2 m wide) intrude metamorphic rocks 1650-1740 Ma old in the southern part of the Nagssugtoquidian mobile belt on the coast of central W Greenland (Noe-Nygaard & Rambert 1961; Escher & Watterson 1973; Scott 1977, 1979, 1981; Brooks et al. 1978; Thy 1985) (Fig. 7). The lamproites contain phenocrysts of olivine, diopside, Krichterite, Ti-phlogopite and pseudoleucite in a groundmass of Ti-phlogopite, diopside, K-richterite, sanidine, pyrite, a carbonate mineral and priderite (Fig. 4), whereas the kimberlites contain macrocrysts of olivine, phlogopite, picroilmenite and rare pyrope in a groundmass of olivine, Ti-
Lamproites and K-rich igneous rocks
I 15
(a)
(c)
(b)
(e)
(d) FIG. 3. Photographs of lamproite slabs showing typical rock textures (all listed clockwise from top left): (a) olivine lamproites from the Prairie Creek area (PRA) (American Mine, Kimberlite mine, Crater of Diamonds pipe (three samples) and Black Lick); (b) phlogopite lamproites from the Prairie Creek, Crater of Diamonds diatreme (PRA), all from the West Hill, showing tufts and magmatic--autolithic breccias; (c) Fitzroy Basin (WKB) xenolithic/autolithic lamproite breccias (Mount Abbot, Big Spring, 81 mile vent, Mount Ibis, Calwynyardah Pipe and Fishery Hill; center; Ellendale pipe and Wolgidee Hill); (d) Fitzroy Basin (WKB) magmatic and volcanic rocks (Mount North phlogopite lamproite, vesicular 81 mile vent phlogopite lamproite, Ellendale pipe B olivine lamproite, vesicular 81 Mile Vent glassy phlogopite lamproite, altered near-surface Ellendale pipe B; center glassy Oscars Plug phlogopite lamproite); (e) Fitzroy Basin lamproites; (f) Leucite Hills lamproites.
(f)
(Scale: photograph a= 17 • 29 cm)
I I6
S . C . Bergman
(o)
b)
(c)
(d)
(e)
f
(g)
h)
(i/
I j)
(I)
m)
(n)
p (q) r) (s) t) FIG. 4. Photomicrographs (plane light) of various lamproites. Massive magmatic lamproites: (a) Wolgidee Hills wolgidite with perovskite, priderite, shcherbakovite, diopside, K-richterite in view; (b) Machells Pyramid Hill phlogopite lamproite; (c) Oscar Plug (WKB) glassy phlogopite lamproite; (d) Priestly Peak (PP) phlogopite lamproite; (e) Gaussberg (GSB) leucite lamproite; (f) Smoky Butte (SB) phlogopite lamproite; (g) Kamas (KAM) phlogopite lamproite; (h) Spring Butte (LH) phlogopite lamproite; (i) Holsteinsborg (5646) (HOL); (j) Chelima (CHE) phlogopite lamproite. Lamproite breccias: (k) Oscar Plug (WKB) lamproite-sandy tuff-breccia; (1) Boars Tusk (LH) phlogopite lamproite; (m) Argyle AK1 (ARG) sandy tuff; (n) Calwynyardah (WHB) olivine lamproite. Olivine lamproites: (o) Hills Pond (HP); (p) Kimberlite Mine (PRA); (q) Black Lick (PRA); (r) Crater of Diamonds (PRA); (s) Ellendale pipe B with dunite xenolith; (t) Ellendale pipe B with garnet xenocrysts. Scale: each photograph is approximately 4 x 5 mm. phlogopite, apatite, diopside, serpentine, a carbonate mineral, perovskite and spinel (Scott 1977, 1979, 1981; Brooks et al. 1978). So-called 'anomalous' lamprophyre dykes that are extremely rich in diopside with phlogopite, kaersutite and titanomagnetite also occur (Scott 1977, 1979, 1981). Thy (1982) reported an assemblage
of K-richterite-arfvedsonite-riebeckite-actinolite from related dykes. Lherzolite, harzburgite and granulite xenoliths occur in the kimberlite dykes (Scott 1979, 1981). The emplacement of the HOL lamproites is broadly contemporaneous with the emplacement of the Gardar igneous province ((1.1-1.3) x 10 3
Lamproites and K-rich igneous rocks
I 17
~
490
INNER BOUNDARY OF P H A N E R O Z O I C 80,
Q
#l
OROGENIC BELTS
\
t
~.+80
.
• x
AREA UNDERLAIN / B Y ARCHAEAN "~ CRATON
'lL~, '
.(, c.
50~
9 t
\
f'HM At=O A SB
KNX NES
'~,,~i;'" CN "9~ , / &
HP
+ "=411
3 O/k_
4~O ~. 30 ~.
+ 30
15+ 110
CRUSTAL AGES (BILLIONS OF YEARS) 1 = <0.7 2=0.7-1.2 3 = 1.2-1.5
4 = 1.5-2.5 5=>2.5
~MAJOR CRUSTAL PROVINCE BOUNDARY . . . . FRACTURE ZONE ====SEAMOUNT CHAIN, LINEAMENT, OR RIFT ZONE
+15
I(~X)
9 ULTRAMAFIC LAMPROPHYRE o KIMBERLITE LOCALITY / PROVINCE t, LAMPROITE LOCALITY / PROVINCE 9 K-RICH BASALT/LAMPROPHYRE ULTRAPOTASSIC ROCKS
& OTHER
9 LAMPROPHYRE 9 CARBONATITE
FIG. 5. Map of N America showing the locations of lamproite s e n s u s t r i c t o suites, other K-rich rock suites, kimberlites, carbonatites and several notable lamprophyres. Principal geological and tectonic features, i.e. Pacific oceanic crust fracture zones, the New England seamount chain, the mid-continental rift zone, the Colorado lineament zone, the inner boundary of Phanerozoic orogenic belts, the limit of areas presumed to be underlain by Archaean cratonic material and major crustal boundaries (with the approximate basement age after Condie 1976) are also shown. For lamproite and K-rich rock suite abbreviations see Tables 2 and 3 and text. Precambrian crustal provinces: Wy, Wyoming; CN, Central; GR, Grenville; SU, Superior: KA, Kaminak; SL, Slave; CH, Churchill. Pacific fracture zones: ME, Mendicino; PI, Pioneer; MU, Murray; MO, Molokai; CL, Clarion. Other features: NES, New England seamounts, CL, Colorado lineament: MCRS, Mid-continent rift system; OA, Ouachita fold belt.
Bergman
S.C.
I I8 9,w I
9
z
q
I
I
I
I
1 I~ ~
~
~
+1+1+1
+1+1+1+1+1+1+1~
~
~~
- ~ ~
+l+l+l
+1+1+1+1+1+1+1
~
~
M
+1+1+1
I
I
+1+1 ~ m
@ ~
?
9 ~m
+ + I + I + I + I + I + I + I ~ M ~ M N ~ M
+1+1+1
~
~
q
}
+1 +1 ~ +1 +1 e ~
q
. . . .
I
I
1 + 1 + 1 ~ 1 4 1 ~ ~ ~
+1+1+1
~ 1 + 1 + 1 + 1 + 1 + 1 + 1 ~ ~ ~
+1+1+1
r
M
%
q Z 9,,.,.~
9
Z
+ l + l + l + l + I + l + l ~ ~ ~
~ ~ _ _
.
.
.
.
+1+1+1
.
.
:n~_
. . . . . ~.
+l+l+l+l+l+l+l<<<~m<<<~<~
+1+1+1
+ 1 + 1 + 1 + 1 + 1 + 1 + 1 ~ ~ ~
~
-
-
-
-
~
_
+1+1+1
~
_
Z~ <
.{ + I + I + I + I + t + I + I ~ ~ M N ~
[.-,
+1+1+1 . . . .
{ +1
} [,.z.1 ,,-.1 <
E
~
~
~z~
~Z~C
~
Lamproites and K-rich igneous rocks 0
0 ~ 0 ~
+1+1+1+1
I
~ .c:
0 L~ o
~el
o
~
+1+1+1+11
~ee
~
J J
+l+~+l+l+l
I ~
+1+1+1+1+1
t
+1 +1 +1 +1 +t
~+
= 9
~
~
+1+1+1+1+1
~
.
-~-
_
I
~ m ~ - -
.
.
~
.
.
~
+~ +~ § +~ +J
+L+I+I+t+I
M
M
~
M
M
o~
~
~
+1+1+}+1+1 ~
.~~
~
~
.
.
.
.
.
I ~9
S. C. Bergman
I20
TABLW5. Comparison of the average major-element compositions (wt.~) of diamondiferous lamproites and Kimberlites with barren varieties Known diamondiferous
SiO2 TiO2 Al,,O3 FeO* MnO MgO CaO Na20 K20 P205 BaO COE H20 + Mg number n
Known non-diamondiferous
Lamproites
Kimberlites
Lamproites
Kimberlites
47___7 2.9_+ 1 4.8+1 8_+2 0.12_+0.04 23_+6 8+8 0.4-+ 0.3 3.5_+ 1 1.3_+0.7 0.9+0.8 4_+4 3.3_+2 82_+5 46
40+7 2.4_+ 1.7 5_+2 11 _+3 0.18_+0.1 28_+7 10_+8 0.6_+ 0.9 1.5_+ 1.1 0.9_+ 0.8 0.19_+0.2 4.0_+5 7_+4 81 _+6 325
53___6 3.1 ___2 10___2 6_+2 0.10_+0.05 10_+4 6-+3 1.6_+ 1 7.5_+3 1.3 _+0.7 0.6_+0.6 1.7_+2 3_+2 72 +_8 270
36+7 3.0-t-2 5_+4 12+2 0.20_+0.1 27_+8 13_+7 0.6_+ 1.2 1.4_+ 1 0.9_+0.8 0.16_+0.1 8.0_+5 5_+3 78 +_12 100
All oxides, except CO., and H20, are normalized to 100~o on a volatile-free basis to facilitate comparison (see also Dawson 1986). n, number of analyses. Ma (Blaxland 1976a, b)) 700-800 km to the S of HOL as well as a variety of alkaline complexes in Canada (Gittins et al. 1967; Currie et al. 1975), some of which contain pseudoleucite (Watkinson & Chao 1973), and the Kali 'kimberlitic' dyke in N Sweden (Kresten et al. 1977). Proterozoic ((1.01.2) x 103 Ma) potassic ultramafics in S Greenland (Upton & Thomas 1973) are similar in age and general geochemical character to the HOL rocks, but do not fit the definition of lamproite. It should be noted that diamonds have been recovered in kimberlites from the area between Ivigtut and Frederikshab (Andrews & Emeleus 1971) and in heavy mineral concentrates from stream sediments in the Fiskenaesset area (Nielsen 1976) about 500-700 km S of HOL. Over a dozen alkaline rock suites (carbonatites, lamprophyres and kimberlites) occur in a coastal belt S of HOL and range in age from 2.65 x 103 to 0.12x 103 Ma, indicating mantle-derived alkaline magmatic activity recurring in the same region for a time span in excess of half of Earth history (Larsen et al. 1983). Kamas and Moon Canyon, Utah ( K A M ) (40 ~ 40' N, 111 ~ 12' IV)
Middle Tertiary dykes, flows and small plugs of phlogopite-olivine lamproites are spatially associated with more syenitic dykes in Moon Canyon, near Kamas, Scott County, Utah (Fig. 8). They occur near the junction of the W margin of the E - W trending Uinta Arch and the N-S trending Wasatch Mountains (Morris 1953 ; Larsen 1954; Best et al. 1968; Henage 1972) just W of the
Colorado Plateau. Morris (1953) called the rocks melteigites and Best et al. (1968) termed them wyomingites and orendites. Phlogopite and olivine phenocrysts are set in a glassy groundmass of analcime, sanidine, phlogopite, clinopyroxene, K-richterite, ilmenite, pseudobrookite, chromite, a roedderite-like phase and apatite (Wagner & Velde 1986a) (Fig. 4). The dykes range in age from 13 to 40 Ma (Crittenden & Kistler 1966; Best et al. 1968). Henage (1972), on the basis of trace-element data on mineral separates and whole-rock dyke samples, established that the distribution of trace elements was heterogeneous as a result of both crystal separation and volatilephase transfer. According to Boutwell (1912), mica peridotite dykes were also intersected by underground mine workings in the nearby Park City mining district. The K A M dykes and flows are located 250 km SW of the LH lamproites. Leucite Hills, Wyoming ( L H ) (41 ~ 50' N, 109 ~ 10'
w) The 14 Leucite Hills (22 major exposures) consist of cinder cones, lava flows, volcanic necks, dykes, intrusive sheets and pyroclastic rocks which were intruded through and extruded over Tertiary shales, Cretaceous limestones and clastic sedimentary rocks on the northern nose of the Rock Springs uplift in the Green River Basin, Sweetwater County, Wyoming (Fig. 9). The LH are among the youngest lamproites (about 1 Ma old) (Bradley 1964; MacDowell 1966), second only to Gaussberg, and occur over an area of 2000-
L a m p r o i t e s a n d K - r i c h igneous r o c k s
x 21
FIG. 6. (a) Generalized geologicalmap of the Rose Dome Area, Kansas, showing the location and setting of the Hills Pond lamproite (after Jewett 1964); (b) geological map of the Hills Pond lamproite with cross sections AA' and BB' based on drilling (after Wagner 1954). 2500 km 2. Cross (1897), Kemp (1897) and Kemp & Knight (1903) generated the nomenclature for LH that has since been applied to several other lamproite suites (Exeter area dykes, U.K., Kamas, Utah, and Fitzroy Basin, W Australia). More recent geochemical and petrological work includes that of Johnston (1959), Smithson (1959), Yagi & Matsumoto (1966), Carmichael (1967),
Kay & Gast (1973), Ogden et al. (1978), Barton (1979), Ogden (1979), Kuehner (1980), Ogden et al. (1980), Barton & van Bergen (1981), Kuehner et al. (1981), Vollmer et al. (1981, 1984), and Salters & Barton (1985). Petrographic types (Cross 1897) include wyomingite (leucite-phlogopite-diopside lamproiteolivine_+ K-richterite), orendite (leucite-sani-
I22
S . C. [ ] PROTEROZOIC GNEISSES & METASEDIMENTARY ROCKS []
ARCHAEAN GNEISSES PALAEOZOIC KIMBERLITE 9 PROTEROZOIC LAMPROITE FAULTS DOME 0 100 km
Bergman Crater o f Diamonds State Park and associated intrusives, Prairie Creek, Murfreesboro, Arkansas ( P R A ) (34 ~ 2' N, 93 ~ 40' W)
Although the diamondiferous Prairie Creek intrusions in the vicinity of the Crater of Diamonds L (COD) Park, Pike County, Arkansas, were previously thought to possess affinities with 56* 52* kimberlites (Miser & Ross 1923; Scull 1959; -" ' 0' ::::::::::::::::::::::: 48~ Meyer 1976; Bolivar 1977, 1982; Lewis 1977; Meyer et al. 1977; Gogineni et al. 1978; Bolivar EGEOESMIN !!iiiiiiiii~. 69* ~, =================================& Brookins 1979) it is now apparent that they , ~ :iii:::::::i:::iiiii have more lamproite than kimberlite affinities (Scott-Smith & Skinner 1982, 1984a, b, c; Mitchell & Lewis 1983; Bolivar 1984). Note that Carmichael et al. (1974, p. 517) were arguably the HOLSTEINSBOR i ii!iiii)iii! first to suggest this feature. The COD lamproite is a 73 acre diatreme with three different lithological units: a massive hypabyssal olivine.... ~i iiiiiiiii+iiiiii phlogopite lamproite, an intrusive/extrusive volcanic breccia and a phlogopite lamproite tuff ~ ======================================== i!i!iiiiiiiiii (Figs 3, 4 and 10). Only the breccia is presently S U K K E R T O P P E j,.~ ~ i i ::::,o.~ i : : : :::i!!!!! i i ...... ::::!!!~:::::::: believed to contain an economic concentration of diamonds, although the three lithologies have not been tested using modern techniques. At least four other associated lamproite intrusions occur 1-3 km to the NE of COD: the Kimberlite mine, the American mine, Twin Knobs and Black Lick FIG. 7 Generalized geologicalmap of W Greenland prospect. All are diamondiferous except for the showing the relationship between lamproite and last. kimberlite intrusives and structural elements. (After The PRA intrusion was first mentioned by Noe-Nygaard & Rambert 1961 ; Scott 1977, 1981). Powell (1842; cited by Miser & Ross 1923) and received early attention as a micaceous peridotite (Branner & Brackett 1889; Williams 1891). Miser dine-diopside-phlogopite lamproite -t- K-richter- & Ross (1923) were perhaps the first to put the ite) and madupite (diopside-phlogopite lam- PRA intrusive into the geological context of the proite) (see Figs 3 and 4). The last mentioned is better-studied diamondiferous peridotites of typically SiO2-undersaturated whereas the first southern Africa, even though diamonds were two are often SiO2 saturated or oversaturated. initially discovered at COD in 1906. Additional Accessory phases include priderite, wadeite, studies of the geology, geochemistry and minerapatite, perovskite, magnetite (rare), pseudo- alogy of the PRA rocks include those by Miser & brookite, spinel and Sr-barite (Carmichael 1967; Purdue (1929), Ross et al. (1929), Moody (1949), Kuehner et al. 1981). Thoenen et al. (1949), Stone & Sterling (1964), Experimental studies of LH rocks (Yagi & Gregory (1969), Gregory & Tooms (1969), Lewis Matsumoto 1966; Sobolev et al. 1975; Barton et al. (1976), Meyer (1976), Brookins et al. (1979), 1976; Barton & Hamilton 1978, 1979, 1982) are Steele & Wagner (1979) and Waldman et al. summarized below. (1985). PRA diamonds and their mineral incluLherzolite (rare), harzburgite (rare), pyroxenite sions have been the subject of many studies and mica pyroxenite xenoliths occur at nearly all (Langford 1973; Giardini et al. 1974; Giardini & lamproite exposures (Ogden et al. 1978; Kuehner Melton 1975a, b, c) as have been their occluded 1980; Barton & van Bergen 1981) and particularly and included gases (Melton et al. 1972; Melton & in the volcanic necks. Crustal xenoliths occur in Giardini 1976, 1980). Roedder (1984, pp. 509the LH lavas (Kay et al. 1978). Detrital Cr10) disagreed with the interpretations made in pyrope, Cr-diopside and enstatite of ultimate the last three studies. mantle origin but unknown magmatic source (not The PRA lamproites intrude the nearly horithe LH) have been found in ant hills W of the LH zontal Lower Cretaceous Trinity Formation (clay, in the Green River Basin (McCandless 1982, silt, sand, gravels and limestones); peridotite 1984). pebbles occur at the base of the Upper Cretaceous
z:.
:ii
i if!
=======================
L a m p r o i t e s a n d K-rich igneous rocks
123
FIG. 8. Generalized geological map of the W Uinta Mountain area, Utah, showing the geologicalposition of the Kamas and Moon Canyon lamproites. (After Best et al. 1968; Henage 1972; Hintze 1980.)
Tokio formation (Miser & Purdue 1929). Zartman (1977) and Gogineni et al. (1978) suggest an age of intrusion of 97-106 Ma on the basis of K-Ar dating of phlogopite separates. The PRA intrusions occur on the margin of the Ouachita fold belt on the northern edge of the Gulf coastal plain (Wickham et al. 1976). The hypabyssal and breccia phases of the PRA intrusions consist of olivine phenocrysts (often serpentinized) in an extremely fine-grained groundmass of phlogopite, serpentine, perovskite, K-richterite, Cr-spinel, magnetite, diopside, wadeite and priderite (Figs 3 and 4). The epiclastic tuff unit (Scott-Smith & Skinner 1984a, b, c) contains abundant phlogopite, sanidine and quartz in an extremely altered clay-chloritecarbonate matrix. Both the breccia and tuff units contain abundant autolithic and xenolithic (sedimentary rock) fragments. Garnet (pyrope and almandine) is extremely rare and typically only encountered in heavy-mineral concentrates; picroilmenite only occurs as inclusions within garnets and as extremely rare xenocrysts. Coarsegrained lamproite xenoliths consisting of Krichterite, phlogopite, diopside and priderite from the massive hypabyssal facies have been described by Mitchell & Lewis (1983).
Leucite- and pseudoleucite-bearing intrusions (syenites and tinguaites) cut nepheline syenites at Magnet Cove, Arkansas, 90 km NE of PRA (Williams 1891); lamprophyres are widespread in the Ouachita Mountains (Robinson 1976; Mullen & Murphy 1985; Mullen & Petty 1985). Smoky Butte, Montana (SB) (47 ~19'N, 107~3' W)
N-NE-trending dykes and small plugs of lamproite intrude the Tullock member of the Palaeocene Fort Union formation (sandstone and shale) along the axis of a very broad gentlydipping regional syncline over a distance of 3 km just W of Jordan, Montana (Matson 1960) (Fig. 11). The three major rock types at SB contain the following mineral assemblages in addition to glass: (i) sanidine, K-richterite, K-riebeckite and armalcolite; (ii) Ti-phlogopite, diopside, olivine, sanidine and armalcolite; (iii) Ti-phlogopite, diopside, olivine, analcime and armalcolite (Velde 1975; Matson 1960). Velde (1975)noted the presence of priderite and Wagner & Velde (1986a) additionally recognized pseudobrookite and chromite. D. Velde (personal communication, 1985) noted the presence of a new mineral, davanite (K2TiSi601 s), in the SB rocks.
I24
S. C. B e r g m a n
FIG. 9. Generalized geologicalmap of the northern Rock Springs Arch of the Green River Basin showing the location of the Leucite Hills lamproites : BT, Boars Tusk; MAT, Matthew; SM, Steamboat Mountain; BR, Black Rock; OR, Orenda; BAD, Badgers Teeth; ZIR, Zirkel Mesa; EM, Emmons Mesa; HAT, Hatcher Mesa; HAG, Hague; HAL, Hallock; END, Endlich; NP, North Pilot Butte; NT, North Table Butte; ST, South Table Butte; CR, Cross; OS, Osborn; P, Pilot. (After Cross 1897; Kemp & Knight 1903; Love et al. 1955.) The dykes are generally extremely fine grained to glassy in texture and often contain zeolite or carbonate-filled vesicles, but massive coarsegrained (1-2 mm) type (ii) lamproite with residual groundmass glass occurs at the Smoky Butte plug. Bedded welded autolithic lamproite lapilli and piperno (welded agglutinate tufts) in clay-rich matrices overlie and are marginal to the intrusions. Contact intrusive breccias composed of vesicular deformed magmatic lamproite fragments in a clay-rich matrix also occur at the plug and are almost identical with the breccias observed at LH, WKB, MAP and in other lamproite suites (S. C. Bergman, unpublished data).
Mitchell & Hawkesworth (1984) discussed the trace-element and Nd-Sr isotopic geochemistry of the SB intrusions. On the basis of K-Ar dating (Marvin et al. 1980), the SB intrusives are 27 Ma old, relative to the 47-53 Ma ages of the Winnett alnoite (100 km WSW of SB), the Missouri Breaks kimberlites and ultramafic lamprophyres (160 km WNW of SB) (Hearn 1968, 1979; Hearn & McGee 1982, 1984), and the Highwood Mountain shonkinites and minettes (300 km W of SB). Sapphire-bearing minette-like lamprophyres occur at Yogo Gulch in central Montana (e.g. Meyer & Mitchell 1985). Although only obscurely described in the literature, dykes whose occurrence and mineralogy are similar to that of SB
Lamproites and K-rich igneous rocks
125
FIG. 10 (a) Generalized geological map of the Murfreesboro area, Arkansas, showing the positions of lamproite intrusives at Prairie Creek, Twin Knobs, Kimberlite Mine, Black Lick and American Mines; (b) generalized geological map of the Prairie Creek lamproite diatreme; (c) E-W cross-sections of the Prairie Creek lamproite diatreme showing two alternative interpretations (left half of diagram after Bolivar 1977).
I26
S. C.
Bergman
FIG. 11 (a) Map of Montana showing the locations of the Montana petrographic province alkaline rock suites (after Larsen 1940; Hearn 1979; Marvin et al. 1980) (LBM, Little Belt Mountains; HB, Haystack Butte; HM, Highwood Mountains; CAM, Castle Mountains; CZM, Crazy Mountains; BBM, Big Belt Mountains; MR, Madison Range; GM, Gallatin Mountains; ABR, Absaroka Range; SH, Sweet Grass Hills, Haystack Butte, East Butte; BPM, Bearpaw Mountains; RB, Rattlesnake Butte; TB, Twin Butte; LRM, Little Rock Mountains; MBD, Missouri Breaks Diatremes; SB, Smoky Butte; ID, Ingomar Dome Diatreme (Froze-to-Death Butte); PD, Porcupine Dome; WA, Winnett Alnoite; JM, Judith Mountains; MM, Mocassin Mountains); (b) generalized geological map of NE Montana showing the geological environment of the Smoky Butte lamproite intrusives and extrusives (after Ross et al. 1955); (c) geological map of the Smoky Butte intrusives and extrusives (after Matson 1960).
Lamproites and K-rich igneous rocks occur in Rosebud County, Montana, 90-100 km S of SB at T10N, R36E, near the Ingomar Dome at Froze-to-Death Butte (Heald 1926, pp. 26-7; Marvin et al. 1980) and at T10N, R38E, on the Porcupine Dome (Bowen 1915, p. 66). The welldescribed igneous rocks geographically closest to SB are those of the Winnett alnoite sill (Ross 1926a; Marvin et al. 1980). [Central Sierra Nevada, California ( C S N ) (37 ~ 29'N, 119 ~ 19' W ) ]
Ultrapotassic basanitic lavas and plugs (free of plagioclase) and associated potassic olivine and alkali basalts (plagioclase bearing) were erupted from 3.3 to 3.8 Ma ago through the core of the Sierra Nevada batholith (Van Kooten & Peck 1977; Van Kooten 1980; Dodge & Moore 1981). These rocks possess atomic K/Na and K/A1 ratios, MgO contents and Rb contents which are too low and CaO and Na20 contents which are too high to be lamproites. The phlogopites and diopsides of CSN lavas are richer in AI than are those of lamproites. Although the BaO contents of CSN lavas are extreme (0.3-0.6 wt.%) and similar to those of lamproites (0.6 wt.~), the Rb levels are extremely low (less than 100 ppm) for lamproites (typically 200-300 ppm). [Deep Springs Valley, California ( D S V ) (37 ~ 21' N, 118 ~ 03' W ) ]
Four leucite-bearing mafic, olivine melaleucitite, olivine-leucite trachybasalt and potassic basalt plugs and flows (Dodge & Moore 1981) contain plagioclase and have whole-rock K 2 0 / N a 2 0 and K20/A1203 atomic ratios too low to be lamproites. [Highwood Mountains, Montana ( H M ) (47 ~ 31' N, 110 ~ 48' IV)]
I z7
[Navajo-Hopi Buttes, Arizona ( N H B ) (35 ~ 25' N,
11o~ w ) I Dykes, flows, volcanic necks and pyroclastics of potassic and ultrapotassic mafic melts were erupted through Triassic and Cretaceous sediments in two intervals (2-5 and 25-30 Ma ago (Roden et al. 1979)) in the Four Corners area on the SW part of the Colorado plateau (Williams 1936; Anderson et al. 1970; Roden 1981; Rogers et al. 1982). The Navajo rocks are potassic whereas the Hopi Butte localities tend to be sodic. These rocks are not lamproites sensu stricto because they have atomic K 2 0 / N a 2 0 = 1.3, K20/ AlzO3=0.3 and their phlogopites contain 1215.5 wt.~ A1203; however, the presence of analcime-bearing minettes suggests that lamproites may occur as end-members of the suite.
[Seward Peninsula, Alaska ( S E W ) (65 ~ N, 116 ~ IO' W ) ]
A belt of alkaline rocks 350 km long trends NE from the Darby Mountains of the eastern Seward Peninsula to beyond the Kobuk-Selawik lowlands 200 km to the NE (Miller 1970, 1972). The complex is middle Cretaceous in age (97-110 Ma) and consists dominantly of peraluminous potassic nepheline syenites and related rocks with a minor component of peraluminous ultrapotassic rocks. As shown by Miller (1972), the SEW rocks are nearly identical, in terms of both mineralogy and geochemistry, to the HM suite, Montana. Some of the SEW rocks are extremely ultrapotassic, and contain among the highest K20 contents of igneous rocks recorded in the world because of the abundance of kalsilite (atomic K z O / N a 2 0 = 1 3 , 16.6 wt.~ K20: Kobuk Selawik Lowland juvite) (Miller 1972).
The Highwood Mountains sub-province (Larsen et al. 1941 ; O'Brien et al. 1985) displays the best-
developed suite of potassic-ultrapotassic mafic alkaline rocks (see Fig. 11) of any suite within the alkaline petrographic province of central Montana (Weed & Pirsson 1895; Buie 1941; Larsen 1940). The HM rocks occur as flows, plugs, dykes, plutons, sills, lopoliths and laccoliths and were emplaced 50-53 Ma ago (Marvin et al. 1980) about 160 km E of the Rocky Mountain front. Petrographic types of interest include nepheline-normative shonkinite, missourite and minette (mafic phonolite). The HM rocks are not as perpotassic (KzO/A120 3 = 0.4) or ultrapotassic (KzO/NazO,~ 1.5) as typical lamproites (Tables 4 and 5).
[Fortification dyke, Colorado ( F O R ) (40 ~ 48' N, 107 ~ 40' W ) I
Ross (1926b) described a dyke 16 km long (3-6 m wide) trending N60~ about 30 km N of Craig, Colorado, that contains a majority of oligoclase and orthoclase with minor mafics. The FOR dyke, though originally termed 'verite' by Ross (1926b), is not a lamproite because it contains sodic plagioclase and consequently has low K20/ NazO and K20/A1203 ratios. The present author classifies the Fortification dyke as a lamprophyre, but modern diagnostic mineral and chemical studies are needed to say anything further.
S. C. Bergman
128 [ Two Buttes, Colorado ( TB ) (37~
102~
' W) ]
Cenozoic laccoliths and associated dykes of prowersite at Two Buttes in the Arkansas Valley (Gilbert 1896; Cross 1906) are minettes even though they were originally included in Niggli's lamproite clan. Despite an ultrapotassic geochemistry (atomic KzO/Na2O=5), A120 3 contents (12.3~) are slightly high for lamproites. Although the analysis may be suspect, the AI20 3 content of an augite separate is 3.1 wt.~ (Cross 1906), which is much too high for lamproites (see below). [Knox County, Maine ( K N X ) (44 ~1 I' N, 69 ~10' W)]
Prowersites from Appleton Maine (Bastin 1906) contain large phenocrysts of perthitic feldspars. Despite a low A120 3 content (10.6 wt.~) and an ultrapotassic character (atomic K20 / Na20 = 2.5), the KNX rocks are not lamproites or even lamprophyres. [Galore Creek, British Columbia (GAL) (57 ~ 14' N, 131 ~ 30' W ) ]
Ultrapotassic pseudoleucite-bearing syenite porphyries and phonolites occur in the upper Galore Creek Valley, NW British Columbia (Allen et al. 1975). Although altered with abundant evidence for orthoclase replacement, these peraluminous rocks contain elevated K20 contents (9-13 wt.~) and atomic K20/Na20 ratios (about 3). [Spotted Fawn Creek, Yukon (SFC) (64 ~ 22' N, 133 ~ 42' W)]
Pseudoleucite-bearing tinguaites (coarse-grained phonolites occur in the Spotted Fawn Creek area in W central Yukon (Templeman-Kluit 1969). The rock is ultrapotassic, yet metaluminous with atomic K20/Na20 = 3.1 and KzO/A1203 = 0.7. [A illikBay, Labrador (ALL) (59 ~15' W, 55 ~15'N)]
Minette dykes occur as part of the Mesozoic mafic-ultramafic lamprophyre-carbonatite suite of Aillik Bay, Nain Province, Canada (Hawkins 1976; Foley 1982, 1984). These dykes intrude Archaean gneisses and Proterozoic metavolcanics and metasediments and include the following petrographic types: minette, monchiquite, alnoite, aillikite and carbonatite. Although the heretofore understood geochemistry and petrography (Kranck 1953; Hawkins 1976) of the AIL 'minettes' (6~ olivine, 359/00augite, 28~ biotite, 25~ orthoclase and 4 ~ oxide; 44~ SiO2, 4.0 wt.~ TiO2, 9.1~ A1203, 11.0~ FeO*, 10.5~ MgO, 7.5~ CaO, 6.3~ K20, 1.8~/o Na20, 1.1~ P205) displayed characteristics typical of minettes as a
group (cf Rock 1986), recent work has demonstrated the existence of priderite in some of the 'minettes' (K. Collerson, personal communication, 1984). More detailed work is therefore required to establish the true affinities of the AIL dykes, although it is possible that true minettes may carry priderite. [Chino Valley, Arizona ( C H I N ) (35 ~ N, 112 ~ 30'
w)] A 25 Ma old eclogite- and peridotite-xenolithbearing potassic latite (55-60 wt.~ SiO2, atomic K20/Na20 = 1-3 and 4-5 wt.~ MgO) and potassic felsic latites occur at Sullivan Buttes, near Chino Valley, Arizona (Arculus & Smith 1979; Roden et al. 1979; Schulze & Helmstaedt 1979). [Colima Graben, Mexico (COL) (19 ~ 40' N, 103 ~ 40' W) ]
Late-Quaternary K-rich basalts, andesites and related rocks occur in the Colima Graben of the E-W-trending Mexican volcanic belt (Luhr & Carmichael 1980, 1981). These workers have described phlogopite-bearing K-rich basalts which are petrographically similar to minettes. [Other K-rich rock localities]
Brief descriptions of other occurrences of N American K-rich rocks can be found in Glazner & Stork (1976), Shafiqullah et al. (1976), Glazner (1979), Rowell & Edgar (1983), and references cited therein.
Australia
In contrast with the widespread occurrence of diamonds and kimberlites in Australia (Fig. 12), lamproites sensu stricto only occur along the SW and SE flanks of the Kimberley craton in NW Australia. The leucitites of New South Wales are not included in the lamproite clan in this paper because of their anomalous character. The Terowie, S Australia, 'lamproites' reported by Colchester (1982, 1983) are also interpreted as olivine nephelinites (Carr & Olliver 1980; S. C. Bergman, unpublished data). W Kimberleys, W Australia ( W K B ) " ~ 8 ~ N, 124 ~ 45'E)
Over 100 individual occurrences of many petrographic types of lamproite sensu stricto flows, plugs, dykes, plutons, sills, diatremes, cryptovolcanic structures, tufts and other pyroclastics occur on the margin of the Proterozoic King Leopold mobile zone, the Fitzroy Trough and the
Lamproites
and K-rich igneous rocks
I29 [ ] PROTEROZOIC FOLD BELTS []ARCHAEAN
CRATON
RIFT VALLEY
~
CONTINENT-OCEAN
BOUNDARY ( M o ) -----" FRACTURE ZONE
-~- EXTRA-ARCH BASIN ::;:: TASMAN ::::::::::.~, i:i:i ===========BELT ========FOLD ===========:5:::::::::: =======================
22222224~222222;
O DIAMOND
ii!iiiiiii~ili!i!i!ili!i!i!iiiiiii!!i!i!i!iiiiilil
OCCURRENCE &, LAMPROITE ULTRAPOTASSIC ROCK 9 KIMBERLITE | ULTRAMAFIC LAMPROPHYRE 9 CARBONATITE
0 I
400 ~
km
I
FIG. 12. Generalized map of Australia showing the major Archaean cratons, Proterozoic mobile zones, continent-ocean boundary, oceanic crust fracture zones (and other features) and the locations of lamproites, kimberlites, carbonatites, ultramafic lamprophyres and diamond occurrences. (After Garlick 1979, 1983; Stracke etal. 1979; Veevers 1981 ; Atkinson et al. 1984a.)
Lennard Shelf (in three separate fields--Ellendale, Calwynyardah and Noonkanbah), all of which form the Canning Basin on the SW flank of the Kimberley Basin (Fig. 13). Diamonds occur in over 30 separate intrusions (Atkinson 1982; Atkinson et al. 1984b). These intrusions were discovered by Fitzgerald (1907), initially petrographically and chemically described by Simpson (1925), Farquharson (1920, 1922) and Skeats & Richards (1926), and have since received considerable attention in detailed work by Wade (1937), Prider (1939, 1960, 1965, 1982), Wade & Prider (1940), Prider & Cole (1942), Bell & Powell (1970), Derrick & Gellatly (1972), Mason (1977), Atkinson et al. (1982, 1984a, b), Jaques et al. (1982, 1983, 1984a, b), Nixon et al. (1982, 1984), McCulloch et al. (1983a, b), Smith (1984a, b) and Jaques et al. (1986). Although early K - A r and Rb-Sr work erroneously indicated ages of 35 Ma and 250 Ma respectively for WKB rocks (Prider 1960; Kaplan et al. 1967), more recent K - A r and Rb-Sr work on mineral separates limits the age of emplacement of 14 of the WKB lamproites to 17-25 Ma (Wellman 1973; Jaques et al. 1984a; Bergman & Onstott, unpublished data). This ultrapotassic and perpotassic suite consists of a petrological continuum from diamondiferous olivine lamproites to barren leucite lamproites (Atkinson et al. 1983, 1984a; Jaques et al. 1984b,
1986). The WKB olivine lamproites contain olivine megacrysts and aggregates in a finegrained to glassy groundmass of olivine, phlogopite, diopside, apatite, K-richterite, perovskite, wadeite and spinel, whereas the leucite lamproites are composed of phlogopite, diopside, K-richterite and leucite phenocrysts in a groundmass of several of the above phases with spinel, wadeite, priderite, jeppeite, shcherbakovite, perovskite, apatite and rare ilmenite (Figs 3 and 4). Kfeldspar is absent except where it replaces leucite. Wade & Prider's (1940) nomenclature for individual rock types is summarized in Table 1. The WKB lamproites occur on the southern margin of the Precambrian Kimberley block and intrude mostly Mesozoic and Palaeozoic sediments, although a few intrusions cut Proterozoic granitic rocks. The regional structure consists of dominantly NW-trending faults with E-W-trending anticlines and subordinate synclines oriented in an en echelon pattern. Mantle xenoliths are rare in WKB rocks but include dunites and harzburzites (Atkinson et al. 1984a; Jaques et al. 1984b, 1986). Many of the olivine-rich xenoliths display features similar to the aggregates with smaller olivine grains. Hall & Smith (1985), Jaques et al. (1986) and Smith (1984a) discussed aspects of the ARG and WKB diamonds.
S. C. Bergman
I3o 125
124
CRETACEOUS ~ ! CANNING BASIN SEDIMENTARY ROCKS TRIASSIC [ ] SEDIMENTARY ROCKS PERMIAN [ ] SEDIMENTARY ROCKS DEVONIAN [ ] REEF CARBONATES KIMBERLEY BASIN SEDIMENTARY ROCKS (,ARCHAEAN BASEMENT) L. PROTEROZOIC [ ] KING LEOPOLD MOBILE ZONE
126
9 LEUCITE LAMPROITE 9 OLIVINE LAMPROITE FAULT . . . . . ANTICLINE OR SYNCLINE
0 I
50 km i
1
(a)
[]MAGMATIC LAMPROITE COARSELY MICACEOUS [ ~ MAGMATIC LAMPROITE FINELY MICACEOUS []TUFF [ ] S A N D Y TUFF
0 L
I km J B
C /
A' A' ~ /
PLAN
CROSS-ECT,ON
:'it',
,,
0-200mCROSS-SECTIONS A
A'
B
B' C
C' .
.
.
.
.
.
.
.
.
.
i~ i,,:i-::.~i, -"
0 I
(b)
300~m J
40Ore•
[ ] MAGMATIC LAMPROITE DMUDSTON E [ ] TUFF
~oom
200m 300,m (c)
FIG. 13. (a) Generalized geological map of the Canning Basin and King Leopold Mobile Belt, W Australia, indicating the locations of olivine and leucite lamproite intrusives (after Bureau of Mineral Resources, Geology, Geophysics 1976; Atkinson et al. 1984; Jaques et al. 1984a); (b) geological map and cross-sections of the Ellendale olivine lamproite diatreme, Lennard shelf (after Atkinson et al. 1984a); (c) geological map and crosssection of the Calwynyardah olivine lamproite diatreme, Fitzroy Trough (after Atkinson et al. 1984a; Madigan 1983).
Lamproites and K-rich igneous rocks
Creek mobile zone at Lissadell Road and Bow Hill (Atkinson et al. 1984a, b, Jaques et al. 1986).
Argyle, E Kimberleys, W Australia ( A R G ) (16 ~ 40' S, 128 ~ 25' E)
The diamondiferous Argyle AK1 olivine lamproite pipe is located on the eastern flank of the Kimberley craton in the 1940-1800 Ma old Halls Creek mobile zone (Atkinson et al. 1982, 1984a, b; Madigan 1983; Jaques et al. 1986). This intrusive is 125 acres in exposed area and is the richest diamond deposit in the world in terms of grade, containing proven reserves of 61 million tons at 680 ct per 100 tons and additional probable reserves of 14 million tons at 610 ct per 100 tons (Atkinson et al. 1984b) (Fig. 14). An extensive alluvial diamond deposit also occurs adjacent to the Argyle pipe in Smoke Creek and Limestone Creek (Madigan 1983; Meakins 1983). The ARG intrusive is most probably Precambrian in age since it is overlain by Cambrian sediments; it consists predominantly of an olivine lamproite sandy tuff that is intruded by dykes of magmatic olivine lamproite. The massive magmatic phases consist of euhedral olivine phenocrysts (replaced by talc and carbonate) and ragged tetraferriphlogopite microphenocrysts in a fine-grained groundmass of phlogopite, anatase, sphene, perovskite, apatite, Mn-ilmenite, Ti-Mg chromite and sulphides (Fig. 4). The sandy tuff is massive or weakly bedded and contains magmatic lamproite clasts in a groundmass of olivine, ash and xenolithic rounded quartz grains (Fig. 4). Leucite has also been reported in the ARG rocks. Dykes of mica peridotite occur near ARG in the Halls
A
13 ]
[Lake Cargelligo Area, New South Wales ( N S W ) (33 ~ 20' S, 146 ~ 25' E ) ]
Leucite-bearing lava flows and associated scoria cones and pyroclastics were erupted from 10 to 15 Ma ago (Wellman et al. 1970; Cundari et al. 1978) through a peneplaned basement complex of Palaeozoic geosynclinal sediments and granites and an overlying thin veneer of Cenozoic sediments of the central and southern Highlands Fold Belt over a 450 km • 150 km belt with concentrations near Lake Cargelligo, New South Wales (Cundari 1973; Cundari & Ferguson 1982, and references cited therein). The NSW essential mineralogy consists of diopside, olivine, leucite, Fe-Ti oxides, Ti-phlogopite and Ti-richterite with accessory apatite, sanidine and rare nepheline. Although the essential minerals of lamproites are present in the NSW rocks, their compositions deviate slightly from lamproites with Ti-richterite intermediate between the Na and K endmembers (K20/Na20=0.8-3.0 (weight basis)) and high A1203 contents of 1.7-2.4 wt.~; diopsides additionally possess a wide range in AI203 contents (as high as 2.8 wt.%). However, compositions of some phlogopite and diopside phenocryst cores possess typical lamproite characteristics with extremely low A1203 contents (9.7 wt.~ and 0.09 wt.~ respectively).
At31o
A':
[ ] LAMPROITENON SANDY TUFF LAMPROITESANDY TUFF [ ] LAMPROITECONTACT ZONE [ ] PROTEROZOIC SEDIMENTARY ROCKS
w FAULT ~
t
400m
-C
200m
500m J
PLAN
I
|"J 100 C'
I
100 CROSS-SECTIONS
FIG. 14. Generalized geological map and cross-section of the Argyle AKI olivine lamproite diatreme, Halls Creek mobile zone. (After Madigan 1983; Atkinson et al. 1984a).
I32
S. C. Bergman
Although the low CaO, A1203 and Na20 contents of the NSW rocks are typical of lamproites (average of 28 analyses, 9.0 wt.~, 8.7 wt.~ and 1.8 wt.~ respectively), the average atomic KzO/Na20 ratio of 2.1 and atomic K20 / A1203 ratio of 0.6 are rather low for lamproites. TiO2 contents in the NSW rocks are extremely high (range of 3.4-8.0 wt.%; mean, 4.3 wt.~) for lamproites (mean, 2.8 wt.~). The average Mg number of the NSW rocks (66) is somewhat low for lamproites (average, 74). The Sr-Nd isotopic compositions of NSW rocks are anomalous in that they plot close to bulk Earth instead of well within the 'enriched-Sr, depleted-Nd mantle quadrant' as other lamproites on a 143Nd/144Nd versus 87Sr/86Sr plot (R. Mitchell, personal communication, 1984). The NSW rocks are therefore not considered lamproites sensu stricto because of their major-element geochemistry, isotope geochemistry and mineralogy. They are closely allied, however, to the lamproite clan. [Mordor complex, Central Australia ( M O R ) (23 ~ 32'S, 134 ~ 27' E ) ]
The ultramafic to felsic rocks of the Mordor complex in central Australia (Langworthy & Black 1978) are potassic to ultrapotassic (atomic K20/Na20 -- 2-5) and mildly peraluminous (molar (K20 + Na20)/AI203 = 0.3-0.6) and are therefore not lamproites.
Europe Post-tectonic lamproites sensu stricto occur in at least four districts in Europe: in the Hercynides of Cornwall at Pendennis Point, in the Appenides of SE Spain between Murcia and Almeria, in NW Italy and in an isolated occurrence on the island of Corsica. Potassic to ultrapotassic nonlamproite rock suites characterize the Mediterranean area, the most notable being the strongly peraluminous central Italian volcanic province, but also including the Permian Exeter volcanics and Channel Island dykes of the U.K., the Bohemian Massif in Czechoslovakia, the Anatolia province of Turkey, the Spednogorie Zone, Bulgaria, the Hellenic arc in southern Greece, the Aeolian arc, southern Italy, and isolated occurrences in Iran and Norway. Interestingly, both the lamproites and compositionally similar potassic to ultrapotassic rocks of Europe are generally depleted in Ti and Nb (about 1-2 wt.~ TiO2 and less than 40 ppm Nb) relative to these rock types from most other worldwide localities (2-5 wt.~ TiO2 and more than 50 ppm Nb). These compositional features of Ti and Nb depletion are shared by subduction-derived ba-
salts and andesites (Ewart & LeMaitre 1980; Gill 1981 ; Briqueu et al. 1984) and will be discussed in more detail below. Note that minettes are widespread throughout Europe (Velde 1969b). Murcia-Almeria province, S E Spain ( M A P ) (38 ~ 30' N, l ~ W)
Breccia-mantled pipes, plugs, sills, flows and tephra of lamproite were emplaced through Mesozoic and Tertiary sedimentary rocks from 6 to 8 Ma ago (Bellon & Letousey 1977; Bellon 1981 ; Bellon et al. 1981 ; Nobel et al. 1981) at the SE edge of the Betic and Sub-Betic orogenic belts of the Alpine nappes over an area of 15 000 km 2 (Fig. 15). Ironically, these rocks have historically received more attention than any other lamproite suite, and yet they are the least like the other lamproites under present discussion. Niggli (1923) used the MAP suite as a type locality for lamproites. They were initially described by Lewis (1887), Yarza (1895), Osann (1889, 1906) and Washington (1903), and more recently by Meseguer Pardo (1924), Jeremine & Fallot (1929), Fallot & Jeremine (1932), Hernandez-Pacheco (1935), Parga Pondal (1935), San Miguel de la Camara (1935, 1936), San Miguel de la Camara et al. (1952), Fuster & Pedro (1953), Fuster et al. (1954, 1967), Fuster (1956), Fuster & Gastesi (1964), Hernandez-Pacheco (1965), Marinelli & Mittempergher (1966), Borley (1967), Fermoso (1967a, b), Velde (1969b), Fernandez & Hernandez-Pacheco (1972), Pellicer (1973), Caraballo (1975), Lopez Ruiz & Rodriguez Badiola (1980), Nixon et al. (1982, 1984), Venturelli et al. (1984a), Hertogen et al. (1985) and Wagner & Velde (1986a). Petrographic varieties include jumillite (abundant diopside and sanidine with minor phlogopite, K-richterite and olivine), fortunite (abundant hypersthene, phlogopite and sanidine with minor olivine, diopside and K-richterite), verite (abundant phlogopite with minor sanidine, diopside and olivine in a glass-rich groundmass) and cancarixite (abundant sanidine and Krichterite with minor olivine, diopside and hypersthene). Accessory phases include leucite, Crspinel, rutile, apatite, analcime and ilmenite, and alteration phases include quartz, serpentine, carbonate, chlorite and zeolite. Wadeite, priderite or shcherbakovite have not been reported. Compared with other lamproite suites, the MAP rocks are the most enriched in SiO2 and depleted in TiO2 and Nb. They are also the only lamproites that contain apparently primary magmatic orthopyroxene. It is interesting that the MAP lamproites are least like the other lamproites sensu stricto in terms of rock and mineral
Lamproites and K-rich igneous rocks
chemistry, despite the fact that Niggli used them as the type locality for lamproites. The jumillites are the most lamproite-like of all the MAP rocks.
N W l t a l y ( N W I ) (45 ~ 35'N, 7~ 45'E) and Orciatico, Pisa, Italy (PIS) (44 ~ N, 10 ~ E)
Dykes and rare flows of post-Alpine ultrapotassic lamprophyre and associated shoshonitic calcalkaline rocks were intruded N of the Canavese tectonic line in the Sesia-Lanzo and Combin Units (ophiolites, schists) of the internal NW Alps of NW Italy from 29 to 33 Ma ago (De Marco 1959; Krummenacher & Evernden 1960; Carraro & Ferrara 1968; Dal Piaz et al. 1973, 1979; Hunziker 1974; Scheuring et al. 1974; Zingg et al. 1976; Venturelli et al. 1984b). The mineralogy of the NWI ultrapotassic rocks is characterized by an abundance of phlogopite, diopside and sanidine with minor to trace quantities of altered olivine, riebeckite-arfvedsonite, apatite, sphene, Fe-Ti oxide and carbonate minerals (Dal Piaz et al. 1979; Venturelli et al. 1984b). The diopsides and phlogopites from the more ultrapotassic rocks (atomic K20/ NazO = 5-8) are typical of those of lamproites in that they have low A1203 contents (less than 0.3 wt.% and about 12 wt.~ respectively). The lesspotassic rocks (atomic K20/Na20<3.7), however, contain as much as 1.5 wt.~ A1203 in diopsides and 15 wt.~ A1203 in phlogopites. The geochemistries of the more-potassic rocks are also diagnostically lamproitic, with atomic K20/ A1203,~0.9, atomic K20/Na20~4, high MgO contents (9-14 wt.%) and low CaO contents (4.17.7 wt.~) (see Tables 4 and 5). Refractory traceelement levels are lamproitic, with 315-460 ppm Ni and 600-800 ppm Cr; contents of large-ion lithophile (LIL) elements also display lamproite affinities with less than 850 ppm Zr, less than 1240 ppm Sr and less than 570 ppm Rb. The NWI ultrapotassic dykes therefore fulfil the lamproite constraints summarized above, although Velde and Venturelli (personal communication to Bachinski 1985) group many of these rocks in the minette group. Rocks related to lamproites (termed selagites) containing olivine, clinopyroxene, phlogopite, sanidine, ilmenite (Borsi et al. 1967), K-richterite, chromite and apatite were erupted 1-4 Ma ago at Orciatico, Pisa, Italy (Stefanini 1934; Barberi & Innocenti 1967; Wagner & Velde 1986a). However, these rocks apparently do not contain oligoclase, a phase which Tr6ger (1935) and Johannsen (1931, 1933, 1935, 1939) include in selagites.
133
Pendennis ( P E N ) (50 ~ 5' N, 5 ~ 2' W) and Holmeade Farm ( H L M ) (50 ~ 54' N, 3 ~ 28' IV) Cornwall, U.K.
Ultrapotassic dykes, here considered lamproites, occur on the coast of SW England at Pendennis Point where they intrude Palaeozoic slates (Hall 1974, 1982). Analcime-bearing lamprophyres have been reported further inland at Holmeade Farm near Tiverton (Tidmarsh 1932; Knil11969, 1982; Velde 1971; Cosgrove 1972). Knill (1969) likened the HLM rocks to wyomingite and orendites from Leucite Hills and both Cosgrove (1972) and Velde (1971) termed them lamproites despite their minette affinities. The PEN dykes are estimated to be Permo-Carboniferous in age (Hall 1982), whereas HLM has been dated at 281__+11 Ma (Miller & Mohr 1964). The HLM locality falls within the widespread Exeter volcanic series which is treated separately below as a suite of non-lamproite ultrapotassic rocks. The PEN petrography is characterized by sanidine, phlogopite, alkali amphiboles (riebeckite-arfvedsonite), diopside, apatite, barite, spinel, futile, hematite, ilmenite, zircon, quartz, calcite and sulphide. The mineral chemistry of the dykes (low A1203 in diopside, amphibole and phlogopite) and their extremely ultrapotassic and perpotassic character (Table 4) permits their classification as lamproite sensu stricto. The HLM rocks are more difficult to classify with confidence because of a lack of detailed petrographic and mineral chemical data. They fall on the margins of the lamproite chemistry range, with atomic KzO/A1203 =0.6, which is too low for lamproites, and A1203 = 13.6 wt.~, which is too high. Their constituent phases include olivine, phlogopite, sanidine, diopside, anatase, magnetite, apatite, perovskite and analcime that was originally identified as leucite (Tidmarsh 1932; Velde 1969b). However, recent work by Jones & Smith (1985) on the Loxbeare Farm minettes demonstrates that the chemistries of phlogopites (13-15 wt.~ A1203), diopsides (23 wt.~ A1203 ; Mg number, 84-87) and sanidines (Or64_97 ; less than 1.0 wt.~ Fe203) are more typical of the minerals of minettes than of lamproites.
Sisco, Corsica, France (COR) (42 ~ 50' N, 9 ~ 25'E)
A small lamproite sill 13.5-15.4 Ma old intrudes Mesozoic schists W of the hamlet of Sisco, N of Bastia on the northern part of the island of Corsica (Fig. 16) (Prime11963; Velde 1967, 1968; Feraud et al. 1977; Civetta et al. 1978; Bellon 1981). The lamproite (termed porphyry by Primel,
I34
S. C. Bergman
FIG. 15. (a) Map of Spain and NW Africa showing the generalized structure and geology and the distribution of lamproites and other ultrapotassic rocks. (After UNESCO 1971 ; Venturelli et al. 1984a; Vila et al. 1974); (b) location map of SE Spain showing the principal lamproite and other potassic volcanic rock occurrences (after Venturelli et al. 1984a).
FIG. 16. Map of Corsica showing the generalized geology and the location of the Sisco lamproite, the size of which has been grossly exaggerated for clarity. (After Velde 1967; and Geological Survey of Italy 1961.)
Lamproites and K-rich igneous rocks
and minette and lamproite by Velde) contains phenocrysts of phlogopite and altered olivine in a groundmass of sanidine, diopside, phlogopite and richterite-arfvedsonite with accessory sphene, anatase, rutile, ilmenite, chromite, priderite and calcite (Wagner & Velde 1985). Phlogopite (13.8-15 wt.% A1203) and richterite (2.3 wt.~ A1203; 4.7 wt.~ K20) compositions are atypical of lamproites, but the presence of priderite (Velde 1968) suggests lamproite affinities. The rock is SiO2 rich (56-57 wt.%) and MgO poor (6-7 wt.%) for lamproites in general but possesses a high K20 content (10~) and high atomic K20/Na20 and K,O/AI203 atomic ratios (5.4 and 0.9 respectively). [Tuscany ( T U S C ) (42 ~ 15' N, 11 ~ 32' E ) ]
Quaternary volcanic centres at San Venanzo in the Perugia province and at Cuppaello in the Rieti province (both about 50-60 km E of the central volcanic centres of the Latial province) contain kamafugitic lavas (Holm & Munksgaard 1982; Gallo 1984). The geochemistry of the TUSC kamafugitic rocks is very similar to that of the Toro-Ankole province with the exception of slightly lower TiO2, Nb and Sr contents in the TUSC rocks. The relatively high CaO and low SiO2 contents and the presence of kalsilite and melilite in the TUSC rocks distinguish them from lamproites. [Sunnfjord, W Norway ( S U N N ) (61 ~ 20' N, 7~ 20' E) ]
Two hydrothermally altered peralkaline ultrapotassic syenite dykes (Furnes et al. 1982) consist almost entirely of microcline (75~-80~) plus 159/0-209/o phlogopite, but the presence of the K Zr silicate dalyite, which is compositionally similar to wadeite, and the Ba-Ti silicate labuntsovite, which is similar to shcherbakovite, suggests that these rocks have a lamproite affinity. As it is not known to what extent the hydrothermal alteration has modified the primary chemistry and mineralogy, this occurrence is grouped with the lamproite allies. The presence of microcline is consistent with low-temperature reequilibration. [ E Srednogorie, Bulgaria ( S RE ) (42 ~40' N, 26 ~E) ]
Potassic rocks of the SRE include leucite basanites, limburgites, shoshonitic basalts, trachytes and latites (e.g. Boccaletti et al. 1978). Grozdanov (1979) and Stefanova & Boyadjieva (1975) discussed the mineralogy of rocks very close to lamproites in composition from Svidnya in the Sofia.
135
[Bohemian Massif, Czechoslovakia ( B O H ) (49 ~ N, 14 ~ E ) ]
Although most BOH minettes (N~mec 1972, 1973, 1974) contain a typical assemblage of sanidine-phlogopite-augite, one particularly mafic and alkali-rich minette dyke also contains alkali amphibole (N6mec 1972). This rock contains 10.8 wt.~ AI20 3, 8.2 wt.% K20 and 1.0 wt.% Na20 and is geochemically and mineralogically similar to lamproites. Schulze et al. (1985) and N~mec (1985) presented mineral chemistry data on some of the BOH rocks which generally show that these alkaline minettes are distinct from lamproites sensu stricto. However, N~mec's (1985) report of priderite in the BOH rocks indicates the need for further study. Holub (1977) described durbachitic plutons (i.e. syenites) which contain apparently cognate xenoliths of ultrapotassic melasyenites that possess compositions similar to lamproites. These xenoliths contain plagioclase and are therefore not true lamproites. Radulescu (1966) discussed ultrapotassic rocks in the E Carpathians. [Exeter volcanics ( E X V ) (50 ~ 26' N, 3 ~ 19' W) and Channel Island lamprophyres ( C H N ) (49 ~ 13' N, 2~ ' W), U.K.]
The EXV (Hobson 1892; Tidmarsh 1932; Knill 1969, 1982; Jones & Smith 1985) comprise four geochemical groups of rocks (Cosgrove 1972): two are shoshonitic (K20/Na20,~I; K20/ A1203 ~ 0.2) and two are ultrapotassic but are not lamproites ( K 2 0 / N a 2 0 ~ 5 and K20/A1203 ~ 0.5). The HLM lamproite sensu latu discussed above falls in Cosgrove's latter group of EXV. The Channel Island dykes form a suite that extends from Guernsey and Jersey to the coast of Normandy (Lees 1974) and are mainly minettes. Wagner & Velde (1986a) described minette from St. Helier that is compositionally similar to minettes from NHB. They are only mentioned here because of their mildly-ultrapotassic character and spatial association with the EXV. [Other European K-rich rocks]
Although these localities are not related to lamproites, they are mentioned here because of their mildly potassic and alkalic character. Potassic volcanics occur in the Roman Province (RP) and Aeolian Arc, Italy (Appleton 1972; Rosi 1980; Civetta et al. 1981 ; Barberi et al. 1974, 1978), the Laacher See, F.R.G. (Duda & Schminke 1978; Wimmenauer 1974), the Hellenic Arc, Aegean Sea (Innocenti et al. 1982), Anatolia, Turkey (Keller 1983) and NW Iran (Riou et al. 1981).
I36
S. C. Bergman
Africa
The most thoroughly studied African lamproites occur at Bobi and Sequela in the Ivory Coast but, in marked contrast with European or American lamproites, only a limited amount of petrological information is available in the literature. In this regard, it is ironic that Africa is the site of the largest number of petrogenetically related kimberlites for which a wealth of data is available. The kamafugitic ultrapotassic province at ToroAnkole on the E African rift is not a lamproite suite but is perhaps the most thoroughly studied suite of its type. The following lamproite summaries must therefore be disappointingly brief until more thorough studies are published.
Bobi-Sequela, Ivory Coast (BOB) (8 ~ 9' N, 6 ~ 33'
w) Sheared dykes, termed 'meta-kimberlites' by Bardet (1973), intrude Archaean crystalline rocks in the Bobi and Toubabouko regions, Seguela, central Ivory Coast (Knopf 1970; Bardet 1973; Mitchell 1985). These dykes are thought to be 1150-1430 Ma old and diamondiferous in character (Bardet & Vachette 1966; Bardet 1973). The BOB dykes possess a geochemical signature similar to that of lamproites (Table 4); however, it is not known to what extent metamorphism has modified their primary composition. Dykes called fitzroyite and wyomingite by Bardet (1973) occur in the BOB area; they lack pyrope and ilmenite but contain abundant chromite. Intimately associated Proterozoic kimberlites and alnoites followed the strike of older fracture systems during their emplacement. The mineralogy of the BOB 'meta-kimberlite' includes phlogopite, diopside, spinel and talc. Alluvial diamonds from the Seguela area have been traced upstream to dykes of altered fitzroyite and mica-peridotite (Dawson 1967a, b; Knopf 1970).
Luangwa Graben, Zambia ( L U A ) (11 ~ S, 33 ~ E)
Little has been published on the K-Ti-richteritebearing lamproites from the Luangwa Graben (Murray, personal communication to Dawson 1970; Dawson 1980, p. 8; Mitchell 1985). A recent review of diamonds (Wilson 1982) mentions the probability of their existence. Kimberlites in the area have been briefly discussed by Dawson (1970, 1980). The lamproites include dykes, pipes and craters which intrude Karoo sedimentary rocks and are probably Jurassic to Cretaceous in age (B. ScottSmith, personal communication, 1986). Crater
facies lapilli tufts and associated pyroclastic and epiclastic deposits occur in the LUA. All of the characteristic lamproite phases occur (B. ScottSmith, personal communication, 1986), yet wadeite or priderite have not been discovered. The LUA is an exceptional lamproite that occurs in a well-developed continental rift setting. Pniel, Postmasburg, Swartruggens etc. South Africa (PPS) (33 ~ 54' S, 18 ~ 58' E)
Dykes of phlogopite-rich and leucite-bearing 'lamprophyres' are found in association with diamondiferous micaceous kimberlites at the Hellam Mine at Swartruggens, S Africa (Skinner & Scott 1979; Hargraves & Onstott 1980). The Swartruggens 'leucite lamprophyre' contains phlogopite, olivine, diopside, leucite and spinel (Skinner & Scott 1979) and possesses a geochemical composition closer to that of olivine lamproites than that of leucite lamproites (Table 4). Lamproite dykes are also reputed to occur at Pniel (Baster's Mine), Postmasburg and Pielansburg in the Johannesburg and Kimberley areas of S Africa (Dawson 1970, 1980, p. 8; Erlank 1973; L. G. Krol, R. Mitchell and A. J. A. Janse, personal communication, 1982, 1985). Feldspathoid-bearing kimberlites (typically containing nepheline, often out of Sr istopic equilibrium with the host rock (Smith 1983b)) have been reported by De Beers (Skinner, personal communication to Smith 1983b) at Klipfontein (28 ~ 23' S, 24 ~ 08' E) and Poortjie (28 ~ 01' S, 24 ~ 17' E). Unfortunately little information is currently published on their geochemistry and mineralogy. The compositionally related (perhaps identical ?) Group II kimberlites of S Africa are discussed by Smith (1983b) and Dawson (1987). [ Toro-Ankole and Birunga volcanic field, Uganda ( T A N ) (1 ~ S, 29 ~ E ) ]
Potassic-ultrapotassic mafic-ultramafic rocks at Toro-Ankole and Birunga characterize the western branch of the E African rift valley system in SW Uganda. Tufts, lavas, agglomerates and explosion breccias consist of a variety of kamafugitic rock types, ranging from melilite- and kalsilite-bearing katungites through olivine leucitites to K-trachyandesites (Holmes 1950; Bell & Powell 1969; Edgar & Arima 1981 ; Ferguson & Cundari 1982). The more mafic rock types do not contain any plagioclase or K-feldspar. These rocks are not lamproites sensu stricto, despite the fact that some workers (e.g. Vollmer & Norry 1983b) refer to them as such. The TAN rocks are members of the kamafugitic suite of ultrapotassic rocks (Sahama 1974) that are mineralogically and geochemically distinguished from true lamproites
Lamproites and K-rich igneous rocks by their nepheline, melilite and kalsilite, as well as by their elevated A1203 and CaO contents and lower SiO 2 contents (Table 4). [Azzaba, Algeria ( A Z Z ) (36 ~ 43' N, 7~ 5' E) ]
Glassy dykes and extrusive rocks, termed 'potassic olivine trachytes' by Vila et al. (1974), occur near Azzaba, Algeria (see Fig. 15a). Olivine ranges in composition from Fo84 to Fo91 and occurs with augite, sanidine and chromite (Vila et al. 1974; Velde, personal communication, 1985). The AZZ rocks are ultrapotassic (K20/ Na20 = 3.5), but their A1203 contents are rather high for lamproites (12.9%) and the rocks are peraluminous with atomic K20/A1203=0.6. These rocks are therefore best described as lamproites sensu lato until more detailed mineral chemistry and petrographic data are available. [Shaki and Okeho, S W Nigeria ( S H O ) (8 ~ 39'N, 3~
Ultrapotassic, nearly perpotassic syenites and biotite pyroxenites occur in several isolated bodies in the basement complex of W Nigeria (Oyawoye 1976). They consist of microcline, albite, biotite, amphibole, pyroxene, quartz and sphene, and possess atomic KzO/A1203 ratios of 0.6 and atomic K 2 0 / N a 2 0 ratios of 3 with 8.7
wt.% K20. Antarctica
Figure 17 shows the location of the three known lamproite sensu stricto occurrences in Antarctica:
137
Gaussberg, Mount Bayliss and Priestly Peak. These rocks range in age from Palaeozoic to Quaternary and the most recent rocks at Gaussberg occur on the E coast where an extension of the Kerguelen Plateau of the Indian Ocean intersects the continent. The 2000 km long Kerguelen-Gaussberg aseismic ridge has been the site of extensive hot-spot activity, involving the generation of mildly-potassic alkali basalts, for the last 27 Ma (Stephenson 1972; Nougier 1970). All the Antarctic lamproites are rich in TiO 2 (3-5.5 wt.~ TiO2; average lamproite, 2.8 wt.%) and LIL elements (400-1500 ppm Ba, 900-1800 ppm Zr, 1300-3000 ppm Sr, 40-150 ppm Nb and 270-350 ppm Ce). The Antarctic lamproites are also particularly rich in P205 with 1.5-3.3 wt.~ (4-10 modal ~ apatite); they share this trait with the Indian apatite-rich lamproites (see below). It should be noted that E India and E Antarctica were adjacent prior to the breaking of Gondwanaland, which suggests a unique character in the underlying mantle lithosphere. Gaussberg (GSB) (66 ~ 48' S, 89 ~ 19' E)
Leucite lamproite lavas and tephra were erupted from the Gaussberg volcano approximately 56 000 years ago on the coast of Wilhelm II land, E Antarctica (Drygalski 1904; Phillipi 1912; Reinisch 1912; Lacroix 1926; Nockolds 1940; Vyalov & Sobolev 1959; Sheraton & Cundari 1980; Cundari & Ferguson 1982; Tingey et al. 1983). The more recent K - A r work on leucite separates by Tingey et al. supersedes the whole-
FIG. 17. Generalized geology of E Antarctica and the southern Indian Ocean showing the distribution of lamproites and seamounts of the Kerguelen-Heard Island chain. (After UNESCO 1976; Sheraton & England 1980.)
I38
S. C. B e r g m a n
rock K-Ar work of Ravich &Krylov (1964) and Soloviev (1972) which suggested ages of 20 Ma and 9 Ma respectively. In addition, Sheraton & Cundari (1980) and Tingey et al. (1983) described as fission-track work by Gleadow which suggested an age of 25 000+12 000 years. Dort (1972) postulated a late Pleistocene to Recent age on the basis of geomorphology. The GSB lavas have been termed leucitites by most workers. The lavas are nearly aphyric with phenocrysts of olivine, diopside and leucite in a glassy matrix (50~-60~ of the rock) containing quench crystals of leucite, diopside, phlogopite, amphibole, ilmenite and chromite (Sheraton & Cundari 1980) (Fig. 4). The mineral chemistry of the silicate phases is diagnostic of lamproites with low A1203 contents in diopside and phlogopite (averaging 0.4 wt.~ and 6.9 wt.~ respectively). The GSB whole-rock geochemistry is also typical of lamproites (Table 4), and emphasizes its perpotassic and ultrapotassic character. Both cognate (olivine-diopside-leucite-phlogopite) and accidental (spinel lherzolite) xenoliths occur in the volcanics (Sheraton & Cundari 1980). Collerson & McCulloch (1982) discussed the Nd and Sr isotope and rare-earth element geochemistry of the GSB lavas. Mount Bayliss ( M B A Y) (73 ~ 26' S, 62 ~ 50' E)
A dyke 5 m thick, termed 'melasyenite' by Sheraton & England (1980), intruded Archaean felsic rocks during the Silurian (413-430 Ma) in the southern Prince Charles Mountains, MacRobertson Land, approximately 300 km SSW of the Radok Lake alnoites (Trail 1963; Walker & Mond 1971; Sheraton & England 1980; Tingey 1981 ; Sheraton 1983). Medium-grained rocks are composed of microcline, K-richterite, K-arfvedsonite, phlogopite, apatite, anatase, zircon, calcite and Fe-Ti oxide. Finer-grained samples have been recovered from MBAY and these differ slightly from the coarser-grained rocks in that they contain pseudoleucite, more phlogopite, less amphibole and ilmenite. The mineral chemistries ofMBAY phlogopites and amphiboles are similar to those from WKB and LH lamproites. The geochemistries of the MBAY rocks are also typical of lamproites in that they possess atomic ratios K 2 0 / N a 2 0 = 3 . 2 and K20/A1203= 1.1; however, they are more evolved than the average lamproite (Mg number, 74) with an Mg number of 56. Priestly Peak (PP) (67 ~ 11' S, 50 ~ 22' E)
A similar dyke (albeit undated) cuts a Proterozoic dolerite dyke at Priestly Peak in Enderbyland
near the coast (Sheraton & England 1980; Sheraton et al. 1980). The PP dyke is finer grained than that of MBAY and consists predominantly of phlogopite, microcline and K-arfvedsonite with minor amounts of apatite, quartz, rutile, sphene, zircon and barite (Fig. 4). The PP dyke (Mg number, 70) is similar in composition to the GSB lavas (Mg number, 70) in that it is more primitive than the MBAY dyke (Mg number, 56). Asia and Indonesia
The only known occurrences of lamproite sensu stricto in Asia and Indonesia are the Indian dykes at Chelima and the Gondwana Coalfields. Those at Coc Pia, North Vietnam are closely allied to lamproites. Phlogopite-leucite 'basalts' occur in central Kalimantan; however, the petrological data required to characterize them are not available. Potassic-ultrapotassic intrusive and extrusive rock suites have been reported for several areas in the U.S.S.R. In recent discussions with geologists at several Chinese geological institutes it was found that lamproites have not yet been recognized in China (C. Hearn & A. J. A. Janse, personal communication, 1985). Coc Pia and Sin-Cao, Upper Tonkin, N Vietnam (COC) (22 ~ 15' N, 103 ~ 30' E)
These lamproites, termed 'cocites' by Lacroix (1926, 1933a, b), occur near the N Vietnam-Laos border in Indochina. The dykes at Coc Pia cut Mesozoic alkaline syenites and granites and are grey fine-grained rocks consisting of olivine, diopside, phlogopite, magnetite, sanidine and leucite (Lacroix 1933a). According to Fromaget (1933), alkaline dykes also occur at Sin Cao, NE of Laid Chau, where they intrude Triassic sediments. The Sin Cao rocks consist primarily of augite and phlogopite phenocrysts with sanidine and analcime pseudomorphs after leucite in the groundmass. Lacroix (1933a) included the Sin Cao rocks in his cocite group, and noted their chemical and mineralogical similarity to the Spanish jumillites. The geochemistry of the COC rocks is typical of lamproites with the exception of a rather low K20 value (4~-5~) and low atomic K20/Na20 and KEO/A1203 ratios (about 1.0 and 0.4-0.6 respectively). However, these discrepancies can be easily reconciled if the possibility of late- to post-magmatic alteration of leucite to analcime is considered. Lacroix (1933a) and Fromaget (1933) also described shonkinites and nepheline syenites at Pin Chai, S of Pu To and N of Lai Chau in Upper
Lamproites and K-rich igneous rocks
Tonkin. Wagner & Velde (1986b) re-examined the two type specimens from Coc Pia and have noted the absence of Ti-rich oxides and Krichterite, and the presence of the Ti-rich phlogopites titanomagnetite and feldspars (Or80_87). These two type specimens fulfil the mineralogical and compositional requirements for inclusion in the lamproite sensu strieto group; however, feldspar compositions are different from those of true lamproites. Further field work is therefore justified to search for other potential lamproites in the Coc Pia area. Chelima (CHE) (15 ~ 27' N, 78 ~ 42'E) and Gondwana Coalfields (GD W) (24 ~ N, 86 ~ E), India
Proterozoic (1200 Ma old) (Crawford 1969; Crawford & Compston 1973) phlogopite-rich, mafic-ultramafic dykes intrude Proterozoic shales of the Cumbum formation over a 10 km x 6 km area near the Chelima Railway Station on the western margin of the Cuddapah Basin and the eastern margin of the Dharwar craton, Andhra Pradesh, India (Fig. 18) (Appavadhanulu 1962; Sen & Narasimha Rao 1970; Sarma 1983; Bergman & Baker 1984). The two early reports described the dykes as minettes and the rocks as transitional between kimberlites and carbonatites, but the last work suggested, on the basis of rock and mineral chemistry, that they are lamproites. The rocks occur with massive magmatic to glassy breccia phases and consist of abundant foliated phlogopite (with talc and chlorite secondary intergrowths), rutile, apatite, perovskite, ilmenite, serpentine pseudomorphs after olivine and a secondary ferroan dolomite. Phlogopite composition and zoning characteristics are typical of lamproites (see below and compare Mitchell (1985)) with (core --* rim): Mg number, 8 5 4 65; AlzO3, 11 ~ 8 wt.~; TiO2, 5.5--* 3.5 wt.~; BaO, 0.7 4--, 0.3 wt.~; CrzO3, 0.2 ---, 0.0 wt.%. The geochemistry of the dykes is also typical of lamproites when corrected for the dolomite alteration (compare with Table 4): 48 wt.~ SiO2; 5.0 wt.~ A1203; 7.4 wt.~ TiO2; 9.4 wt.~ FeO; 18.9 wt.~ MgO; 4.2 wt.% CaO; 0.2 wt.~ Na20; 3.3 wt.~ K20; 2.5 wt.% PzOs; 560 ppm Cr; 450 ppm Ni; 2500 ppm Ba; 1920 ppm Sr; 160 ppm Rb; 380 ppm La; 700 ppm Ce; 1020 ppm Zr. Sen & Narasimha Rao (1970) noted micro-diamonds in the Chelima dykes; ancient workings, presumably for diamonds, are also present. Alluvial diamonds occur in an adjacent drainage within 6 km of this area (Bergman & Baker, submitted for publication). A widespread suite of Mesozoic lamprophyres and mica peridotites, some of which contain
139
leucite, occur about 1000 km to the N N E of the CHE dykes in the Gondwana Coalfields at Jharia, Bakaro and Raniganj on the N margin of the Singhbum craton (Fox 1930; Ghose 1949; Banerjee 1953; Makherjee 1961 ; Sanyal 1964; Chatterjee 1974; Sathe & Oka 1975; Sarkar et al. 1980). The geochemistry of the leucite-bearing GDW dykes is typical of lamproites (Table 4) and is much more mafic than typical minettes (compare with Rock 1986). Gupta et al. (1983) discussed the petrology of carbonated apatite glimmerites from the Damodar Valley which contain phlogopite, apatite, ankerite and chrome spinel phenocrysts in a groundmass of the same minerals plus rutile, devitrified glass, priderite and pyrite. All these GDW dyke rocks are extremely enriched in P205 (3-6 wt.~), TiO2 (5-7 wt.~) and carbonate minerals (10-20 wt. ~ CO2) compared with typical lamproites and, as a group, probably represent a distinct subtype of the lamproite clan. Indian kimberlite diatremes (840-1200 Ma old) are known in three districts: at Wajrakarur (more than six pipes), in the core of the Dharwar craton (about 250 km W of the Chelima dykes), and at Panna (more than four pipes) and Jungel (more than three pipes) near the SE margin of the Aravalli craton (500-800km W of the Gondwana Coalfield dykes) (Satyanarayana Rao & Phatare 1966; Crawford & Compston 1970, 1973; Marthur & Singh 1971; Paul et al. 1975; Balasubrahmanyan et al. 1978; Paul 1979; Murty 1980; Murty et al. 1980). Alluvial diamonds are widespread throughout central and northern India. Although the lamproites discussed above are widely separated from these three kimberlite occurrences alluvial diamonds and lamproites do occur in the same areas. Further work is clearly required on the Indian lamproites and lamprophyres. Kajan River, Kalimantan ( KAJ) (1 ~ 30' N, 115 ~ 30' E)
Brouwer (1909) reported a phlogopite-leucite basalt on the Kajan River, central-E Kalimantan, containing resorbed phenocrysts of phlogopite with reverse pleochroism, typical of the tetraferriphlogopites from lamproites and kimberlites (Mitchell 1985); additional minerals include leucite, diopside, olivine and Fe-Ti oxides. The KAJ geochemistry, however, is not strictly lamproitic (K20/A1203=0.4; K20/ Na20=2.0), and until more detailed data are available the KAJ rock must be grouped outside the lamproite clan. Lacroix (1926) included the 'kajanite' in his systematic treatment of the syenitic leucitites, but Niggli (1923) was apparently unaware of the kajanite occurrence or
I40
S. C. B e r g m a n
FIG. 18. (a) Generalized geological map of India showing the three Archaean cratons and the distribution of lamproites, kimberlites, lamprophyres and diamonds (after Roy 1962; Roy Chaudhury 1973); (b) geological map of Andhra Pradesh State, India, illustrating the location of the Chelima lamproite, Wajrakrur kimberlite and alluvial diamond occurrences (after Roy 1962; Roy Chaudhury 1973); (c) map showing the distribution of lamproite dikes at Chelima, Andhra Pradesh (after Sarma 1983). neglected it in his treatise on lamproites. Bergman et al. (1985) discussed the mineral and rock chemistry of minettes of non-lamproite affinity in the nearby Karamu River area, central Kalimantan. Bucking (1899) noted a biotite-leucite basalt in the S. Celebes. Indonesian arc volcanics (IA V) (6 ~ S, 111 ~ E)
Leucite basalts, tephrites and potassic andesites characterize isolated occurrences from the active Sunda arc in Java, the Celebes and other islands, and are peraluminous with low atomic K20/ A1203 ratios (0.2-0.3) and elevated A120 3 contents (16-22 wt.~o) (Grubb 1965; Baren 1948; Foden 1983).
Maksimov 1982; Kostyuk 1983). The most ultrapotassic occurrences (those of the Synnyr, Murun and Inagli massifs) do not, apparently, include lamproites. They are petrographically and geochemically similar to the HM, SEW and BRA ultrapotassic suites. Lacroix (1926) made an obscure mention of several 'kajanites' (phlogopite-leucite-diopsideolivine basalts) from Vouyerska Kosa and Sveta Petka, Siberia (with a reference to Raoult and a personal communication from M. Koyitch).
South America Brazil (BRA) (20 ~ S, 48 ~ W)
Baikal-Aldan Belt (BAL) and Pamir, Mongolia (PAM), U.S.S.R.
Potassic-ultrapotassic rocks are widespread along the 1500 km Baikal-Aldan alkalic intrusive belt on the Siberian Platform (Mineyeva 1972;
Shoshonitic volcanics, shonkinite and kamafugitic rocks occur in isolated localities in Brazil (Ulbrich & Gomes 1981, and references cited therein). Associated rock types include ugandite, analcite basalts, limburgite and minette, and
Lamproites and K-rich igneous rocks
although kimberlites and other alkaline rocks are widespread in Brazil (Svisero et al. 1979a, b, 1984), lamproites sensu stricto have not yet been reported.
Age of emplacement Known lamproites range in age from Proterozoic (Argyle, Holsteinsborg and Chelima) to Quaternary (Leucite Hills and Gaussberg). Of the 21 lamproite suites or localities recognized here,
141
nine are Cenozoic, four are Mesozoic, four are Palaeozoic and four are Proterozoic. Therefore, in addition to xenolith-bearing alkali basalts, lamproites are apparently most common in the Cenozoic. This age relationship could result from selective preservation of the extrusive members of recent suites or alternatively from some mantle or tectonic process permitting their formation which has been most active in recent times. The fact that the intrusion of Indian-AntarcticAustralian lamproites (CHE, GDW, GSB,
I42
S. C.
Bergman
MBAY, PP, A R G and WKB) has taken place in the same large region (when corrected for continental drift) over more than a billion years suggests that the former hypothesis may be the most likely explanation. On the basis of the better-studied Cenozoic suites (e.g. WKB and MAP), intrusive-extrusive activity takes place over a time span of 2-10 Ma. In areas such as the Fitzroy Basin the emplacement of primitive diamondiferous olivine lamproite diatremes can occur contemporaneously with that of more differentiated leucite lamproite flows and plugs (Jaques et al. 1984a). Therefore it is possible that diamondiferous olivine lamproite diatremes may be erupted in the Leucite Hills area in the next 10 Ma! However, the WKB suite is the only one in which such contemporaneity has been noted.
Intrusive-Extrusive forms Lamproites have been reported to occur in nearly every igneous form possible, the most common being dykes and flows. The most recent lamproites from GSB and LH display all the volcanological features typical of alkali basaltic eruptives, with classic lava flows, pyroclastic ejecta, cinder cones etc. (Fig. 19(c)). Therefore there is sufficient evidence to suggest that the more eroded slightly older lamproites (e.g. WKB) that occur as dykes and necks also probably possessed extrusives similar to those present at LH and GSB. Lamproite sills occur at HP and COR, and a shallow coarse-grained (grain size of less than 2 cm) lamproite pluton occurs at the 2 km diameter Wolgidee Hills intrusive (WKB). Vesicular rocks are ubiquitous in all but the hypabyssal intrusives; extremely vesicular scoriaceous lavas occur at GSB, LH and MAP among others. Palagonite and agglutinate tufts (welded fall-out tufts equivalent to piperno) and/or bedded pyroclastics occur in the more recent suites (e.g. GSB, LH, MAP, SB, WKB and PRA). Their absence in older suites can be easily attributed to selective erosion of extrusive forms. These extrusive facies are evidence of base surge as well as other volcanic processes (Atkinson et al. 1984b). Autolithic breccias composed of fragments of magmatic lamproite that are often plastically deformed and set in a fine-grained groundmass consisting of both magmatic and xenolithic (pulverized country rock) material characterize the contact zones of nearly every lamproite intrusive (Figs 3 and 4). These autolithic breccias are diagnostic of lamproites and are texturally similar to pyroclastic breccias associated with silicic volcanic centres. Glass-rich rocks are
present in many lamproite sensu strieto occurrences (e.g. LH, GSB, MAP, WKB and SB). The size oflamproite occurrences varies widely from locality to locality, from dykes of limited aerial extent less than 1-2 m wide at HOL, N W I and MBAY, PP and CHE, through small plugs (10-100 m wide) at KAM, MAP and HP, to the volcanic cone 370 m high and 1400 m in diameter at GSB. The estimated total volume of the erupted material at LH, WKB and MAP exceeds 10-100 km 3 in each case. Therefore, whereas lamproites are volumetrically of limited size compared with mafic alkalic igneous rocks in general (Mauna Loa volcano, Hawaii, alone exceeds 100km3), a given eruptive suite can approach the size of a continental alkali basalt volcanic field. Lamproites also occur as diatremes or pipes, and these igneous forms have the greatest potential for possessing a diamondiferous olivine lamproite phase such as those at PRA, WKB and ARG. These lamproite diatremes differ in appearance from the typical kimberlite diatremes (Figs 19(a)and 19(b)). Whereas kimberlite pipes tend to be carrot-shaped and possess 2-3 km of vertical flaring (Hawthorne 1975; Harris 1984), lamproite diatremes are characterized by a funnel or sherbet-glass shape with less than 0.1-0.5 km of vertical flaring. This contrast has important diamond-exploration implications (i.e. volume calculations and magnetic modelling) and can be reconciled by a consideration of the volatile abundances and compositions of the respective intrusive systems (Harris 1984). The larger proportions of both CO2 and H20 (a factor of 23) in kimberlites relative to lamproites produces a relatively-deep (2-3 km) explosive boiling stage in the former intrusives. The volatile budgets of lamproite magmas are dominated by the extremely-soluble HzO so that they boil at shallower depths of less than 1 km. An alternative explanation for the difference in shape between kimberlite and lamproite diatremes involves the relatively-shallower erosion level displayed by the latter. Since the vast majority of kimberlites are Pre-Cenozoic, whereas most lamproites are Cenozoic, this erosion level can be easily rationalized. However, even the older (Palaeozoic) lamproites (e.g. Argyle) display tuffaceous facies.
Tectonic-geological environment Whereas some of the lamproite sensu lato or compositionally related shoshonite localities occur on the margins of continents or in oceanic margins (e.g. Indonesia and the Mediterranean), the 21 lamproite sensu stricto localities described above all occur in continental regions. This
Lamproites a n d K-rich igneous rocks KIMBERLITE
DIATREME
~-
MODEL
.....
.,r--Z
CRATERFACIES~'~,{~'~.-~.7'_~>~<~~ - ~ - ' - - - ' -
LAMPROITE
I43 PIPE M O D E L
__
.........
--'" "
FINE-GRAINED MASSIVE
- ~ - - / ' ~ , . ~'~,~.~ ..... ~;,..~ '~ , ' ~ EROSIONAL ...... / _ _ ~ ' ~ ~ ~ i";'.':i:-!~" ;!!!~~,,~,f - ~ SURFA C E
~--~--~c,~s
~
~.?~-
.
.
.
.
.
.
~_;_C ~
L~++~
S2~_~T_=-- ~
200 m
(ol
CINDER (b)
,-:~_
~:.-q-COARSE GRAINED ~):|1~~ ~ MASSIVE ~-~
~ ~. ~ , . ~ _ L
~
ALKALI BASALT CONE AND FLOW MODEL -~
\,
"
- ' ~ ~ ~~?~--" ~ - . ~ C I N D E R MULTIPLE FLOWS . ~ A S H ~ F, ',~Lj~\
FIG. 19. Idealized cross-sections showing the morphological features of (a) olivine lamproite diatremes, (b) kimberlite diatremes and (c) alkali basalt cinder cones. (After Hawthorne 1975.) continental feature is shared by alkalic rocks in general (Bailey 1974) as well as by leucite-bearing rocks, with the exception of the four oceanic islands discussed by Gupta & Yagi (1980): the Cape Verde Islands, the Kerguelen Islands, the Marquesas Islands and Tristan da Cunha. However, whereas alkaline basaltic magmas with N a > K (weight basis) characterize nearly all well-developed continental rift systems (Bailey 1974; Barberi et al. 1982), with the exception of LUA, lamproites sensu stricto are absent and only
"CONE --,i
"
(c) a few related ultrapotassic rock suites (e.g. LAC, TAN and BAL) occur in well-developed continental rift systems. Lamproites are, in general, the result of postorogenic magmatic phenomena in regions that have experienced collisional orogenesis (with underlying fossil Benioff zones) several tens to many hundred million years prior to their eruption. In contrast, shoshonitic rocks are characteristically more directly linked with subduction. Morrison (1980) found that, while
I44
S. C. Bergman
shonkinites are associated with calc-alkaline rocks of orogenic zones, they tend to be younger and occur over the deeper parts of Benioff zones or in association with block faulting and uplift accompanying the 'flipping' ofa subduction zone. Therefore, whereas shoshonites are temporally associated with the waning stages of subduction, lamproites are clearly post-subduction and apparently require the passage of time in addition to the stabilization of orogenic belts for their formation. It should be noted that Helmstaedt & Gurney (1984) related the kimberlites of S Africa to subduction processes. In contrast with the general restriction of kimberlites to craton interiors (e.g. Janse 1984; Dawson 1980), lamproites generally occur closer to craton margins. In contrast with ocean island chains which young in a certain direction, lamproites do not and they consequently do not owe their presence to hot spots. Note, however, that Edgar (1983) related ultrapotassic magmatism of the Western U.S.A. (including lamproites) to the Yellowstone mantle plume. Several lamproite suites are associated with broad anticlinal or synclinal structures (e.g. WKB, LH and KAM), whereas others intrude relatively undeformed sedimentary-platform sediments resting on Proterozoic basement rocks marginal to the core of the continental craton (e.g. HP, PRA, SB, WKB and LH). Nearly all lamproite suites are associated with surficial lineaments or fault lines which probably connect with well-established deep basement fractures or zones of weakness. Lamproites share this feature with kimberlites and many other mantle-derived extrusive and intrusive rocks. Rowell & Edgar (1983), speculating on the spatial relationships between 15 dated occurrences of Cenozoic K-rich volcanics ( K 2 0 > Na20; SiOz<60 wt.~) in the western U.S.A. and the outer limits of palaeo-arc-related magmatism, suggested that the K-rich volcanism was directly related to deep subduction rather than intra-plate tectonics. Although most of the nonlamproite potassic suites correlated well, none of the lamproites sensu stricto did: they were emplaced long after the ambient western U.S.A. subduction had ceased. Lamproites are therefore controlled by intra-plate tectonic processes, although an association with regions that have experienced collisional orogenesis is indicated by the more recent suites.
Diamondiferous character Except for the CHE lamproite in India which requires more study, there are five known diamondiferous lamproite suites: WKB, ARG,
PRA, LUA and BOB. As the first three have received detailed study, but the latter two have not, the following is based largely on the first three suites. Diamondiferous lamproites all share the following features. 1 They are generally olivine lamproites (mafic to ultramafic) with elevated refractory element contents: 14-30 wt.~ MgO, 35-56 wt.~o SiOz, 26 wt.~ CaO, 0.1-0.8 wt.% Na20, 3-7 wt.% A1203, 400-1300 ppm Ni and 300-1800 ppm Cr (Tables 5 and 6). 2 They are moderately to extremely enriched in incompatible elements: 1-6 wt.~ K20, 100-500 ppm La, 100-600 ppm Ce, 200-11 000 ppm Ba, 100-600 ppm Nb, 100-700 ppm Rb, 300-2000 ppm Sr and 200-1200 ppm Zr; many of these concentrations are in excess of those found in kimberlites. 3 They occur as diatremes with sherbet-glass or funnel-shaped cross-sections, are relatively large in surface area (75-210 acres) and possess multiple intrusive-extrusive phases including a massive magmatic phase, breccias and tufts. 4 They can occur throughout geological history, with the five diamondiferous magmas discovered heretofore erupting in Precambrian, Cretaceous and Miocene times. 5 They are distinctly different from non-diamondiferous varieties of lamproites. The average major- and trace-element compositions of diamondiferous lamproites are compared with those of barren lamproites below (see Tables 5 and 6). Although abundant associated leucite lamproite intrusions are only found in the WKB suite, further mapping is likely to reveal associated differentiated lamproites in the PRA and ARG suites. The eruption of diamondiferous olivine lamproites in the WKB suite is contemporaneous with that of more differentiated leucite lamproites.
Lamprophyre versus iamproite Although the term 'minette' in its present sense, has been used for over 150 years (Bachinski, personal communication, 1984), Gfimbel (1874) introduced the broad textural term 'lamprophyre' or 'shining rock' in reference to a dark dyke rock composed of mafic phenocrysts and groundmass phases. Rosenbusch (1887) gave the term its present significance by defining 'lamprophyres' as dark-coloured dyke rocks of porphyritic texture characterized by a panidiomorphic texture and consisting of a high percentage of mafic minerals as phenocrysts in a fine-grained groundmass composed of the same mafic minerals as well as feldspars and/or feldspathoids. Streckeisen (1980) emphasized the feature that feldspars and/
Lamproites and K-rich igneous rocks
~ =~ ~
~
~
~8
4
_~
~
~
~
~
~
~
~45
~
~
~
~
~
~
~
~
~
~
~
~
o
t....l
~
b
..I,.- ,x.
I46
S. C. Bergman
or feldspathoids, when present, occur only in the groundmass. Rock (1977, 1984, 1985, 1986, 1987) and Wimmenauer (1973) have synthesized the lamprophyre literature and have contributed much to the clarification of many of the problems that have long plagued these mafic-ultramafic rocks. Rock (1986) has chosen to include lamproites as well as kimberlites in the lamprophyre clan. Although the term 'lamprophyre' is extremely broad and includes a host of intermediate, mafic and ultramafic rock types, it is best utilized as a field term, for which it was originally intended, and should not be used to encompass such widely separated rock types as kimberlites and lamproites which possess such distinct evolutions. Clearly, many basalts fulfil the definition of lamprophyre, yet basalts and lamprophyres are petrogenetically distinct (Rock 1987). However, Yagi et al. (1975) described a lamprophyre dike which represents the subvolcanic feeder zone of alkali basalt flows. The original definitions of 'lamprophyre' have emphasized the shallow intrusive dyke character of lamprophyres. Lamproites, as we have seen, characteristically occur as extrusives or shallow intrusives, and it is only the very old occurrences that expose deeper feeder-dyke levels. Clearly, every volcanic edifice can generally be traced to a feeder dyke. As lamproites often possess a feldspathoid (leucite) among the phenocryst population (e.g. SB, WKB etc.), they do not qualify to be called lamprophyres. Tr6ger's (1935) use of the term 'lamproite' as an extrusive equivalent of lamprophyres that are rich in K and Mg (presumably minettes) should be abandoned. As discussed above and below and by Rock (1984, 1987) minettes and lamproites are geochemically and mineralogically distinct, although there are some overlapping features. Therefore minettes and lamproites cannot be related by currently understood processes that could link an intrusive to a coexisting extrusive form. Several, albeit extremely rare, minettes occur as extrusives (e.g. NHB and COL) (see above), but these are not lamproites as we understand them at present. Nevertheless, it is possible that minettes may be nothing more than deep-seated equivalents of extrusive lamproite lavas. The fact that minettes are generally Palaeozoic in age (notable exceptions are the NHB, COL and KAJ minettes of Cenozoic age) and therefore have experienced more erosion than younger lamproites supports this view. More work is required to solve this problem. It is therefore suggested that lamproites as well as kimberlites be excluded from the lamprophyre clan if only to remove some of the inherent genetic biases introduced by associating them.
This opinion is shared by several other workers (J. B. Dawson, personal communication, 1984; R. Mitchell, personal communication, 1985).
Lamproite geochemistry Rather than reproducing individual geochemical analyses, the averages and ranges (represented by standard deviations) of geochemical parameters for various rock suites are discussed in this section, recognizing the inherent biases introduced by such an analysis. The literature sources for individual analyses are tabulated in Appendix 1. Of the 275 analyses available for lamproites, the lamproite group average is weighted toward the MAP, WKB, PRA and LH suites (i.e. 75% of the analyses come from 4 or 20% of the suites). The most apparent feature of the geochemical data is the extreme range in major-, trace-element and isotopic composition of individual lamproite suites and of lamproites as a group.
Major-element geochemistry The average major-element compositions of lamproite suites and other ultrapotassic rock suites are tabulated in Table 4. Univariate frequency-distribution bar charts including all the lamproite major- and minor-element data are illustrated in Fig. 20. These distributions are heterogeneous for all elements and ratios considered, emphasizing the compositional complexity and variability of lamproites. Some distributions are approximately normalor slightly skewed (e.g. K20, P205 and MgO); however, most elements possess frequency distributions that are bimodal or extremely skewed (e.g. SiO2, A1203, TiO2, FeO, CaO, Na20, BaO, H20 and ZrO2). Elemental ratios (e.g. Mg number, K20/ NazO and K20/A1203) display distributions characterized by a high kurtosis. With the exception of a few elements, lamproites overlap in major-element composition with kimberlites and lamprophyres (Table 4). The range in composition of lamproites (expressed as the arithmetic mean_+ 1 standard deviation) overlaps with that of kimberlites for all major elements except K, Si, Mg and A1. However, statistical comparisons of the two groups, lamproites and kimberlites, using a general linear-models procedure, show that they are different in terms of all major and minor elements. Lamproites can be statistically distinguished from lamprophyres, as a group, with respect to all major elements, but most significantly with potassium (general linear model F value, 840) and less so for CaO, FeO*, A1203 (F
Lamproites and K-rich igneous rocks
8oI 6o 5o 3o
FREQUENCY
40 20
I
FREQUENCY
FREQUENCY
40
6
5
io
ib
147
I0
20
_
6
,.'o
21o
s'.o
_
4:0
BaO
6
~b
g
20
Ti02
"
30
30 1
2O
I0
20
I0
6
~
I0 0
I0
20
30
,
0
2o
2O
I0
I0
80
I0
1'5
K20
40 20
Jb
g
~1~
2o
CaO
5
'
5
i
1.5
d o
0
i
0.5 1.0 K20/AI20s
30 3O
zb
Fe20s
ib
'
FeO
~060~
30
I0
.
6
I .'0
210 SZ0
410
40
0
2O I
I
0 60
~
I0
I0
15 20
H 20T
!
I
5
5
~
IO
15
C02
40 20
4'0
s'o
6'o
0
I
2
io
a'o
4o
4
5
30
2O
0
0.I
40
0.2
0,3
MnO
30 2O I0
3 Na20
30
0
5 I0 KzO/No20
15 I0 0
2O I0
o 35
45
55 Si02
65
75
o'.s
;'.o
;.s
0 . 0 6 0.12 0.18 0.24 0 . 3 0 Zr02
2.o
Na20 4- K20 AI20~
FIG. 20. Frequency-distribution diagrams for major- and minor-element contents and various elemental ratios in lamproites (volatile free; weight per cent basis).
S. C. Bergman
I48
values of 300-700). Of all the lamprophyres, minettes approach lamproites the closest; however, when non-diamondiferous lamproites are statistically compared with minettes (general linear models Student t test), the two groups differ in A1203, FeO*, CaO, MgO, Na20 , K20 and TiO2 at the 99.9~ confidence level. Figure 21 illustrates the compositions of all lamproites compared with kimberlites and lamprophyres in two ternary projections used for representing kimberlites and related rocks (Cornelissen & Verwoerd 1975). Although a great deal of overlap occurs in both plots, lamproites occupy a relatively K20-rich field in the K20-MgO-A1203 ternary which is not occupied by either kimberlites or lamprophyres. In the FeO-A1203-MgO projection, lamproites are intermediate between kimberlites and lamprophyres. In fact, with exception of Si, P, Ti, K, Fe and Ca, lamproites are compositionally intermediate between kimberlites and lamprophyres (Table 4). Lamproites AI203
i
',-~< ~",
( ;e
-'_;
9 FeO
\ MgO
(a) K20 A
MgO
~ LAMPROITES . . . . . . . . . LAMPROPHYRES ..... LITES
AI203
(b)
FIG. 21. Ternary diagram showing the compositions of lamproites, lamprophyres and kimberlites projected on (a) the A1203 MgO-FeO* and (b) the K20A1203-MgO plane (weight per cent).
have the highest average K20, TiO2 and SiO 2 and the lowest average CaO and FeO contents of any of these mafic-ultramafic alkalic rock types. Lamproites are distinguished from kamafugitic rocks by distinctly higher SiO 2 and lower A1203 and CaO contents as well as by a higher degree of peralkalinity. Figure 22 and Tables 4 and 5 emphasize this K20-SiO2 distinction between lamprophyres, lamproites and kimberlites as a group. Alkali basalt compositions greatly differ from lamproites in that they are significantly enriched in A12Oa, Na20 and CaO, and depleted in K20 , MgO, SiO2, P205, BaO and ZrO 2 (Tables 4 and 5). Figure 22 illustrates the covariance of selected major elements and their ratios for individual lamproite suites and compares the average compositions of kimberlites, alkali basalts, lamprophyres and several lamproite minerals. With few exceptions, the size of an individual lamproite suite field is positively correlated with the number of analyses of the given suite, a feature which emphasizes the extreme variability in lamproite geochemistry. It should be noted that in nearly every plot, diamondiferous suites overlap with each other in composition. In a plot of the degree ofultrapotassic versus perpotassic character (Fig. 22(a)) most lamproite suites individually show a positive correlation, as displayed by most rock types in Fig. l, although several (e.g. ARG) show a slight negative correlation. As observed in Fig. 22(b), some of the WKB lamproites are the most perpotassic of all lamproites whereas the MAP suites contain some of the least perpotassic lamproites. Figure 22(c) displays the extreme ranges in both CaO and K20 contents for given suites; diamondiferous suites are characterized by the lowest CaO contents. The nature of magmatic differentiation in lamproite suites is documented in a plot of Mg number versus SiO2 (Fig. 22(d)). In addition, the generally primitive character of lamproites is evident; lamproites possess extremely high Mg numbers for their relatively high SiO2 contents compared with nearly all other rock types. Although some suites display decreasing Mg numbers with increasing SiO2 (e.g. WKB, HOL and LH), a trend that is expected to result from the fractionation of mafic phases such as olivine, diopside and phlogopite, many lamproite suites possess trends that display very limited Mg number variation (less than 10) with extremely large changes in SiO2 (e.g. ARG, PRA, KAM and NWI). Therefore some mafic phase fractionation can explain the SiO2-Mg number variations displayed within some suites, but other types of differentiation (e.g. vapourphase transfer) or, alternatively, differences in source composition, pressure and/or temperature
Lamproites and K-rich igneous rocks are required to explain the chemical variations within other lamproite suites. Figure 22(e) illustrates the covariance of TiO2 and K20. Lamproite suites can be divided into two groups according to their TiO2 contents: one group with TiO2 < 2 tool.% includes COR, NWI, PEN, LH and MAP, and the other with TiO2 > 1.5 mol.% includes WKB, ARG, BOB, SB, MBAY, GSB, HP and HOL. Interestingly, the NWI, COR, PEN, LH and MAP suites also share the feature of occurring in a young palaeoorogenic zone and are most probably depleted in TiO2 (as well as Nb etc.) because of a TiOzdepleted source, a feature shared by orogenic andesites and related rocks. The extreme enrichments in TiO 2 displayed by the WKB, CHE, BOB and SB suites must represent TiOz-enriched source rocks.
Normative composition The average CIPW normative compositions for most lamproite suites are given in Table 7, bearing in mind the limitations of the CIPW normative calculations in describing phlogopiteand amphibole-rich rocks. The average lamproite consists of nearly equal amounts (by weight) of salic and femic normative minerals. Almost half (43%) of the 295 lamproites that have been included in this data set are quartz normative; in fact, all suites for which more than five rocks have been analysed contain at least one quartznormative rock. Only eight of the 21 suites contain rocks with normative leucite, and normative K-metasilicate is present in small amounts because of the Al-depleted character of lamproites. All suites but two contain at least a few rocks with normative feldspathoids. Lamproites as a group are characterized by significant normative orthoclase (16-59 wt.%, average 35 wt.%) and femic minerals (33-81 wt.%, average 51 wt.%), but relatively low normative albite (012 wt.%, average 5 wt.%) and anorthite (0-6 wt.%, average 1 wt.%) relative to nearly all other igneous rocks. As with most peralkaline rocks, almost all lamproites contain normative acmite (average 2.5 wt.%). Compared with minettes, lamproites contain similar average amounts of normative quartz, orthoclase and nepheline but significantly lower albite and anorthite, slightly lower clinopyroxene, and slightly higher orthopyroxene, acmite, olivine and leucite. Lamproites as a group are somewhat more enriched in femic minerals than minettes. Despite the common occurrence of m o d a l leucite in lamproites, many lamproite magmas tend to differentiate towards quartz-normative residua. This is based on the natural evidence of
I49
the normative compositions of interstitial glasses and whole-rock samples of glassy lamproites (e.g. GSB (Sheraton & Cundari 1980), LH (Kuehner et al. 1981) and WKB (Wade & Prider 1940; Prider 1982)). In fact many leucite-bearing lamproites are SiO 2 saturated or oversaturated. This differentiation trend can be explained by a small amount of phlogopite fractionation and additionally by the experimental work of Luth (1967) in the system kalsilite-forsterite-silicawater which demonstrated that the phlogopite liquidus surface bridges the forsterite-kalsilite thermal divide at PHzo > 1 kb.
Trace-element chemistry A summary of the trace-element geochemistry of lamproites, compared with that of kimberlites, lamprophyres and alkali basalts, is given in Table 6 and Fig. 23(a). All trace elements show an extreme degree of variance with coefficients of variation typically more than 50% and often more than 100%. If these extreme variations are ignored, some general conclusions can be drawn. On average, lamproites have the highest incompatible-element contents (e.g. Rb, Sr, Ba, U, Zr and rare-earth elements (REE)) and alkali basalts the lowest, whereas kimberlites have the highest compatible-element contents (e.g. Ni, Cr, Co, Sc, Zn and Cu) and alkali basalts generally the lowest. Lamproites are more enriched in both compatible and incompatible elements compared with other ultrapotassic rocks. With respect to the average K/Rb ratios, the various rock groups are ranked as follows: alkali basalts (350), nondiamondiferous lamproites (220), calc-alkaline and alkaline lamprophyres (180), kimberlites (160) and diamondiferous lamproites (85). The average REE contents of lamproites, lamprophyres and kimberlites all overlap to some degree (Table 6 and Fig. 23(a)), and all are markedly light REE enriched and heavy REE depleted compared with alkali basalts. Although a large degree of overlap exists, average lamproite REE contents generally exceed those of kimberlites. The average ratios of light REE to heavy REE (La/Lu)on are ranked as follows: lamproite (100), kimberlite (90), ultramafic lamprophyre (47), alkaline and calc-alkaline lamprophyre (27) and alkali basalt (11), showing that lamproites contain the smallest quantities of heavy REE (relative to light REE) compared with these other rock types. Since the degree of elemental variation for these rock types is so great, overlap is generally observed in comparing the trace-element contents of the various rock groups. Nevertheless, the
S. C. Bergman
15 ~
2.0]
t
DIAMONDIFEROUS SUITES 9 AVE, KIMBERLITE (550) 9 AVE. LAMPROPHYRE (800) ",9 AVE. ALKALI BASALT (4230)
WKB
1.0
"PRA
o
1'o
1
~
(o)
i
30
,o
K20 1 Na20 (moles)
15 -] I / / | / ,
--.,"-9 9 ,,~ f-I O
K20 1 AI203 = 2 J J
10 1
W
K
B
~
DIAMONDIFEROUS SUITES AVE. KIMBERLITE (550) AVE. LAMPROPHYRE (800) AVE. ALKALI BASALT (4230) AVE. LAMPROITE PHLOGOPITE (229) AVE. LAMPROITE AMPHIBOLE (97)
K20/AI203 = 1
=
0
5
(b)
10 AI203 (mole %)
15.
I
DIAMONDIFEROUS SUITES 9 AVE. KIMBERLITE (550) 9 AVE. LAMPROPHYRE (800) AVE. ALKALI BASALT (4230) I-I AVE. LAMPROITE PHLOGOPITE (229) O AVE. LAMPROITE AMPHIBOLE (97)
.:..< .;-..-----
g
PRA
;o ( C )
1'5
CaO (mole %)
COR SB HP KAM GSB MBAY PEN NWl PPN MAP HOL LH ENO CHE NSW
. . . . . . . . . . .
............
............ ...............
.~'/-...--t/.-~q,~ -~',--~,'l ---J, .v~,~vvVV~NV.
Lamproites and K-rich igneous rocks gO-
,
151
ARG
PRA
9 .
~ Z BOB
80-
o r,
70-
E 3 l=
eo-
50-
40-
9 9 ~, I-1 O
DIAMONDIFEROUS AVE. KIMBERLITE (550) AVE. LAMPROPHYRE (800) AVE, ALKALI BASALT (4230) AVE. LAMPROITE PHLOGOPITE (229) AVE. LAMPROITE AMPHIBOLE (97)
;o
3O
(d)
WKBf
"::. . . . . " .... . .....-'/ - ......... .--
~
;o
;o
SIO 2 (mole %)
109 9 zx 17
WKB
DIAMONDIFEROUS SUITES AVE. KIMBERLITE (550) AVE. LAMPROPHYRE (800) AVE. ALKALI BASALT (4230) AVE. LAMPROITE PHLOGOPITE (229) E. LAMPROITE AMPHIBOLE (97)
COR SB HP YAM GSB MBAY PEN NWI PPN MAP HOL LH ENO CHE NSW
~
~
..............
-----~ ~ - - - r ,'----r ---.-.-"~r
_o
'...
"~
;
,~ _
..-~---~
>-~--53
0
( e )
K20 (mole %)
FIG. 22. Plots illustrating the compositional space occupied by various lamproite suites, with average rock and mineral compositions plotted for comparison (see text and Tables 2 and 3 for abbreviations): (a) KzO/A1203 versus K20/Na20; (b) K20 versus A1203; (c) K20 versus CaO; (d) Mg number versus SiOz; (e) TiO2 versus K20.
trace elements that are most useful in distinguishing lamproites from kimberlites are as follows: (a) Rb, Sr, Ba, Zr and the K/Rb ratio; (b) Ni, Cr, Co and the Ni/Cr ratio. Kimberlites generally have Rb < 150 ppm, Sr < 1200 ppm, Ba < 3000 ppm, Z r < 8 0 0 ppm and K/Rb<150, whereas lamproites generally have R b > 150 ppm, Sr> 1200 ppm, Ba > 3000 ppm, Zr > 800 ppm and K/ R b > 150. With respect to the compatible elements, kimberlites generally have Ni > 500 ppm, Cr>800 ppm, Co>50 ppm and Ni/Cr>0.7, whereas lamproites generally have Ni < 500 ppm, Cr<800 ppm C o < 5 0 ppm and Ni/Cr<0.7. In general, the absolute concentrations of both compatible and incompatible elements are larger
in kimberlites and lamproites than in alkali basalts and lamprophyres. Individual lamproite suites show widely-ranging trace-element abundances. The Indian lamproites (CHE, GDW) are the most enriched in REE of all lamproites yet contain typical lamproite signatures with respect to the other trace elements. Lamproite suites possessing the lowest LIL-element contents include KAM, N W I and ARG. On the basis of the available data diamondiferous lamproites display no statistically-significant differences relative to barren lamproites in their trace-element systematics, with the exception of an enrichment in Nb and Ga and in the
S. C.
I52
Bergman
TABLE 7. Average CIP W normative compositions of various lamproite and related rock suites Suite
n
q
ARG 3 BOB 1 CHE 2 COC 2 COR 2 EVD 4 GSB 1 HLM 2 HOL 12 HP 4 KAM 9 LH 24 MAP 99 MBAY 6 NWI 9 PEN 1 PRA 24 SB 13 WKB 51 Lamproite 284 Minette 50 Lamprophyre728
6.1• 4.4 --1.1• ---6.3• -2.8• 0.6• 5.0• 4.3• 0.2• 4.4 5.8• 5.4• 5.5• 3.5• 2.7• 0.6•
n
ol
Suite
ARG 3 BOB 1 CHE 2 COC 2 COR 2 EVD 4 GSB 1 HLM 2 HOL 12 HP 4 KAM 9 LH 24 MAP 99 MBAY 6 NWI 9 1 PEN PRA 24 SB 13 WKB 51 Lamproite 284 Minette 50 Lamprophyre728
17.9• 3.5• 20.9• 1.3• 26.7• 11.3 14.3• 22.7• 25.2• 8.6• 6.5• 7.4• -11.1• -17.4• 0.7• 7.8• 9.9• 6.2• 10.4•
or
25.8• 4.6 15.8• 26.4• 59.1• 21.8• 54.2 47.4• 39.9• 31.6• 36.8• 41.9• 38.0• 49.6• 45.1• 56.1 16.5• 41.8• 38.1• 35.0• 34.5• 13.1•
ab
0.5• . 1.2• 11.3• 5.0• 1.0• -12.5• 0.1• 0.5• 7.0• 0.9• 11.0• . 3.0• 2.4• 7.8• 0.3• 5.4• 13.6• 12.9•
an
lc
0.6• .
-.
--
.
.
.
. 13.4• 0.4 0.1• -4.4• 10.0• -. --0.4• -0.9• 2.0• 0.3• 1.5•
. 2.7• -0.5• 1.5• 0.3• 1.3• 5.6• 14.6•
hm
sp
rt
3.1• -3.4• 3.9• 0.9• 2.8• -2.1• --
5.7•
-7.2 0.3• ---
--------
-5.9 --
1.5• 0.2• 1.0• 0.1• 1.2• 2.8 1.8• -0.6• 1,3• 3.6• 5.4•
2.8• --1.4• 1.8• 1.2• 0.6• --4.7• 0.6• 3.1• 1.7• 1.3• 1.4•
--1.5• 1.7• 0.7• 0.1•
-------0.02• ---
m
0.1• 2.3• 3.5• 1.1• 0.2• 0.03•
0.6• 0.1• 1.0• 0.3• 0.02• 0.1•
ap
2.2• 3.9• -7.4 -4.4• -4.9• 5.6• 2.4• 5.6• 3.1• 3.7 0.5• 2.3• 2.3• 2.5• 0.1• 0.4•
3.0• 4.1 4.9• 1.4• 1.5• 7.0• 3.6 6.6• 2.9• 1.8• 3,0• 3.6• 2.5• 6.5• 2.9• 4.3 2.5• 4.1• 2.6• 2.9• 2.7• 1.7•
cc
2.6• -13.0• 0.5• 2.9• -0.2 0.1• 10.0• -0.1• 0.5• 1.2• 0.8• -6.2 4.4• 3.9• 2.6• 2.2• 2.1• 2.2•
hy
di
1.2• --
-4.2 -1.7• 1.0• 0.3• 1.7• 0.1• 1.1• 1.0• 0.4 0,2• 0.4• 2.6• 0.8• ---
1.3• -0.2• -0.8• 1.5• 1.7• 6.7•
il
ac
0.6• 4.3 . --
. 5.7• . 3.2• ----1.2• 0.6• 0.5•
--
mt
11.3• 1.6• 2.4• 3.8• 6.6 3.5• 6.4• 5.8• 3.0• 2.6• 2.3• 7.3• 2.4• 3.2 3.7• 8.6• 5.0• 4.3• 3.3• 4.4•
ks
.
. 0.7• . 4.8• -6.1• --2.6• 0,6• 2.8• .
ne
5.2• --25.3• 6.4• 14.1• 10.3 0.6• 1.0• 8.7• 9.7• 13.1• 8.3• 3.7• 13.5• -2.4• 4.6• 6.5• 9.8• 14.9• 18.7•
26.9• 25.4 43.3• 15.1•
2.8• 8.5• 23.8• 8.7• 6.0• 14.2• 17.4• 11.4• 18.4 35.2• 15.3• 15.9• 13.7• 6.5• 3.7•
Z Foid
Z Salic
~ Femic
0.6 4.1 -5.7 0.3 16.6 5.7 -3.7 2.5 6.0 12.5 1.0 1.6 3.2 0.4 0.8 0.8 3.8 4.3 2.0 8.4
33.9 54.4 19.0 44.2 65.3 44.1 54.9 66.8 40.6 32.1 55.0 55.5 58.0 54.8 52.4 60.9 26.5 56.6 45.5 48.7 59.0 49.9
66.1 48.6 81.0 55.8 34.7 55.9 45.1 33.4 59.4 67.9 45.0 44.5 42.0 45.2 47.6 39.1 73.5 43.3 54.5 51.3 41.0 50.1
Means • standard deviations are given for n > 2 analyses; weight per cent normative minerals; only those minerals w h i c h occur in significant concentrations (i.e. more than 0.1 wt) are tabulated here. N o r m s are calculated using the reported FeO and Fe203 contents.
compatible elements (Cr, Ni, Co). However, diamondiferous lamproites possess slightly higher average light REE (La, Ce), Rb, Ba and Ta contents and have higher La/Lu (1400 compared with 800), Ba/Sr (6 compared with 3) and much lower K/Rb ratios (85 compared with 220) relative to barren varieties. SB magmas contain among the highest REE, Sr and Ba contents of all lamproite suites, but possess more typical lamproite concentrations of other LIL elements
(e.g. U, Th and Zr) and among the lowest Rb contents of any lamproite. WKB are likewise exceptionally enriched in LIL elements. Stable and radiogenic isotopic geochemistry of ultrapotassic magmas Sr and/or Nd isotopic ratios have been measured in rocks from eight lamproite suites (over 160 samples) and in eight potassic-ultrapotassic rock
Lamproites and K-rich igneous rocks
i
153
i
PN~
/ c~
=, E
...... :~.. . :<
I
.c
_~i,.. o
-
~
~.'.';/../~ ,.>/ /. X
!/
i~
~.i " ~ . . ~
/
~ ~
\
J
~,~/
o
.J
~
o
?-
d
~ ~ . / 9
.
~
.5
w. i
,
,
i
,
,
,
d
i
'
'
,
,
i
,
,
,
,
,
,
,
,
,
d
c~
PN~,~,LIPN~ L
~o_
/
,
G
8
0
/t
"-
Cl
~
~'" ~'o ~c
eg ~am E~
o~
~,
~
~
._~
o
~ , ~
o-~
~
~
~
.
~-
I
0
~.,
~oo~
S. C. Bergman
154
suites other than lamproites (over 200 samples); the results are summarized in Tables 8 and 9 and Fig. 23(b). Whereas ultrapotassic rocks other than lamproites generally fall in the ranges 87Sr/ 86Sr~=0.704-0.708 and 143Nd/144Ndt=0.5122 0.5126 (where t indicates present-day ratios), lamproites are characterized by much more and much less radiogenic compositions respectively, with 8VSr/S6Srt= 0.705-0.718 and 143Nd/ 144Ndt=0.5112-0.5123. Lamproite Nd-Sr isotopic data fall along two district trends on an end-eSr plot (Fig. 23(b)). One trend, which includes the SB and LH suites, is characterized by slightly radiogenic 87Sr/86Sr ratios with extremely nonradiogenic 143Nd/l~4Nd ratios, and another shallow trend, which includes GSB, MAP and WKB rocks, is characterized by extremely radiogenic 87Sr/86Sr ratios and non-radiogenic 143Nd/ 144Nd ratios (Table 8). These data indicate that the source regions of these mantle-derived magmas were characterized by very old rocks (more than (1-3) • 10 3 Ma old) with Rb/Sr and Nd/Sm ratios greater than those of bulk Earth, the former source material possessing slightly lower Rb/Sr ratios or being younger than the latter. Most important is the undeniable conclusion that the source regions of primary mantle-derived dia-
mondiferous melts (e.g. Ellendale) may be extremely old and enriched in Rb relative to Sr and Nd relative to Sm. In the past it has generally been assumed that basalts possessing relatively high 87Sr/86Sr ratios (above 0.7050) (Faure & Powell 1972; Faure 1977) must have been contaminated by radiogenic continental crustal material during their ascent. However, combined with the Sr-Nd isotopic data on some kimberlites and included megacrysts and xenoliths (Menzies & Murthy 1980a, b; Erlank et al. 1982; Hawkesworth et al. 1983; Kramers et al. 1983; Smith 1983a, b), the lamproite data indicate that contamination by continental crust is not required to explain their isotopic systematics and that extremely enriched portions of the sub-continental upper-mantle lithosphere exist. Oxygen isotopic studies have thus far been largely limited to ultrapotassic rocks other than lamproites, with the exception of the LH and GSB suites. Studies by Garlick (1966), Taylor (1968), Barberi et al. (1975, 1978), Taylor & Turi (1976), Turi & Taylor (1976), Taylor et al. (1979, 1984), Ferrara et al. (1980, 1985), Kuehner (1980) and Kyser et al. (1981) demonstrate that unaltered lamproites generally possess slightly higher 8180 ratios (7-12) compared with typical mantle-
TABL~ 8. S r a n d N d isotopic c o m p o s i t i o n s o f l a m p r o i t e s Locality*
87Sr/86Srt n
Yr
WKB
LH MAP PRA NWI GSB SB NSW WKB SB MAP
8
0.7169
13 6 17 53 6 34 1 4 7 4 ~ 15 6 8
0.7142 0.7133 0.7063 0.7057 0,7048 0.7096 O.7O79 0.7187 0.7096
0.7171 0.7061 0.7188
Range
t
n
4SN d / t 44N d t
Yc
Range
0.7047-0.7050 0.712-0.721 0.7104-0.7187 0.7114-0.7165 0.7056-0.7070 0.7053-0.7078 0.7036-0.7058 0.7064-0.7132 -0.7171-0.7221 0.7092 0.7098 0.7059 0.7063 0.7050-0.7057 0.7113-0.7209 0.7059-0.7063 0.7175-0.7208
8 13
0.51206 0.51119
0.5120-0.5123 0.5114-0.5111
19
0.51193
0.5118-0.5121
7 4 9 15 6 8
0.5111
0.51111-0.51117 0.5113-0.5115 0.51252-0.51264 0.5117-0-5121 0.5113-0.5115 0.5112-0.5113
0.51189 0.5114 0.5112
Reference
1 2 3 4 5 2 6 2 7 8 9 10 11 12 13 13 14
* See Tables 2 and 3 and text for abbreviations. 1, Kaplan et al. 1967; 2, Powell & Bell 1970; 3, Nixon et al. 1984; 4, McCulloch et al. 1983a, b; 5, Jaques et al. 1984a; 6, V o l l m e r e t a l . 1984; 7, Bolivar 1977; 8, Smith 1983b; 9, V e n t u r e l l i e t a l . 1984a, b; 10, Collerson & McCulloch 1982; 11, Mitchell & Hawkesworth 1984; 12, Mitchell, personal communication, 1984; 13, Fraser et al. 1985; 14, Nelson et al. 1986.
Lamproites and K-rich igneous rocks
[55
TABLE 9. Sr and N d isotopic compositions o f ultrapotassic-potassic rocks other than lamproites Locality*
Rock type
87Sr/S6Srt n
HM NHB TAN
Shonkiniteetc. Minette, monchiquite etc. Kamafugites, ugandite, kivite, shoshonite etc.
TUSC
Kamafugites
LAC
Basalts
RP
Basalts, intermediate and felsicrocks
CSN
Ultrapotassic andpotassic basalts DSV Leucitite NSW Leucitite S African kimberlites Group I Group II N American kimberlites Group I Finsch II
2
Range
143Nd/144Ndt n
:~
Reference
Range
10 0.7078 0.7072-0.7087 12 0.7062 0.7041-0.7081 104 8 37 6 1 1 9 9 22 13 14 5
0.7059 0.7036-0.7083 0.7047 0.7046-0.7048 0.7069 0.7046-0.7103 0.7035-0.7096 0.7057 0.7054-0.7059 0.7112 -0.7104 0.7054 0.7046-0.7062 __ 0.7044-0.7047 0.7092 0.7085 0.7068-0.7098 0.7097 0.7088-0.7111 0.7064 0.7061-0.7068
2 11
0.7119 0.7119-0.7124 0.7053 0.7050-0.7057
24 28
0.7057 0.7088
0.703-0.706 0.708-0.710
13 0.7948 0.704-0.705 7 0.7100 0.7095-0.7119
4 0.5127 0.51270-0.51272 12 0.51250 0.51211-0.51276
9
__
0.51260-0.51265
8 0.51212 0.51209-0.51216
11 0.5118
0.5116-0.5119
8 0.51276 0.51257-0.51282 6 0.51219 0.51213-0.51234
2 3 4 5 18 6 7 8,9 10 11 12 13 14 14 16 15 15
6 0.51219 0.51217-0.51224
* See Tables 2 and 3 and text for abbreviations. 1, Powell & Bell 1970; 2, Bell & Powell 1969; 3, Vollmer & Norry 1983a; 4, Vollmer & Norry 1983b; 5, Rock 1976; 6, Holm& Munksgaard 1982; 7, Vollmer 1976; 8, Hoefs &Wedepohl 1968; 9, Hurleyetal. 1966; 10, Staudigel & Zindler 1978; 11, Powell & Bell 1974; 12, Cox et al. 1976; 13, Rogers et al. 1985; 14, Van Kooten 1981; 15, Smith 1983b; 16, Fraser et al. 1985; 17, Nelson et al. 1986; 18, Aoki & Kurasawa 1984. derived material and support the following conclusions: (a) in regions such as the Roman Province (RP), where isotopically heavy crust exists (8180---+15 to +25), contamination by crustal material is required to explain the stable isotope compositions, and Sr and Nd isotopic compositions are also probably modified by crustal contamination (Taylor et al. 1984); (b) some lamproites (e.g. GSB) and other ultrapotassic rocks (e.g. TAN) display primary mantle 8180 signatures (about + 6.5) when corrected for nearsurface alteration effects (Taylor et al. 1984); and (c) the source regions of some ultrapotassic magmas do not substantially differ in their 8180 compositions from those of typical mid-ocean ridge basalts (MORBs) (about + 5.5) or kimberlites and their xenoliths ( + 5.5 to + 7.5) (Taylor 1968; Kyser et al. 1981, 1982). It is also possible that lamproite source regions are anomalous in their 8180 systematics compared with other mantle rocks.
Pb isotopic studies are similarly largely restricted to ultrapotassic rocks other than lamproites. Most Pb isotopic data exists for the TAN, RP, CSN and DSV suites (Vollmer 1976, 1977; Van Kooten 1981; Vollmer & Norry 1983a, b) and suggest that, in the first case, a P b - P b model age of 500 Ma dates a mantle-source homogenization event (metasomatism?). Pb isotopic systematics of the RP rocks, however, are most consistent with the mixing of two distinct reservoirs; this interpretation is also consistent with the oxygen isotope data noted above. Van Kooten's (1981) Pb isotope work suggests a rather recent mantle-homogenization event on the basis of large variations in 238U/2~ and 232Th/ :~ ratios with rather restricted ranges in 2~176 2~176 and Z~176 ratios, perhaps as recent as 10 Ma ago. Pb isotopic studies on lamproites from L H and SB are currently under-way at the Open University (e.g. Fraser et al. 1985) and on W K B rocks at the
156
S. C. Bergman
Australian National University (e.g. Nelson et al. 1986) and tentatively indicate model ages of (1.5-3.0) x 103 Ma for these mantle source regions. Further radiogenic and stable isotope studies are clearly required for a better understanding of the petrogenesis of lamproites. For example, are the two trends recognized thus far in terms of eSr-end systematics simply the result of too few samples, and will these two trends develop into one broad trend for more lamproite suites? Are all lamproites characterized by 'enriched' Sr and 'depleted' Nd isotopic signatures in contrast with the vast majority of isotopically 'depleted' signatures commonly observed in alkali basalts and MORB (Hofmann et al. 1979)?
Lamproite mineral chemistry The mineralogy and mineral chemistry of lamproites has been recently reviewed by Mitchell (1985) and Wagner & Velde (1985), to which the reader is referred for detailed accounts. Therefore only the most important features, abstracted from Mitchell's review, will be highlighted in this section, and the mineral chemistry of lamproites will be compared with that of associated rock types. An important feature of lamproite mineral compositions is that they reflect the exotic compositions of the magmas from which the minerals crystallized. A tabulation of literature containing lamproite-mineral chemical data can be found in Appendix 2. The average lamproite and related rock mineral compositions are provided in Table 10.
Olivine Olivine (commonly pseudomorphed by secondary phases) occurs in two generations (large anhedral macrocrysts (dimensions of more than 1 mm) and small strain-free crystals or aggregates (dimension of less than 1 ram)) in the more mafic members (and some felsic members, e.g. MAP) of nearly all lamproite suites. The compositional range observed for both types is Fo77-93 with most about Fo89. NiO contents are typically in the range 0.15-0.6 wt.~ whereas CaO is generally less than 0.2 wt.~. The Mg numbers and CaO and NiO contents of lamproite olivines overlap those observed in kimberlites and lamprophyres but are distinct from those in alkali basalts and related rock types.
Orthopyroxene Orthopyroxenes ranging in composition from En61 to En94 (Cr20 3 ~ 0.3 wt.~; CaO < 1.3 wt. %;
TIO2<0.4 wt.~) occur in the MAP, LH and WKB suites. Their extremely wide range in Mg number and texture (commonly rimmed by phlogopite) suggests that some may be xenocrystic; however, their extremely low A1203 contents (typically less than 1.0 wt.~), are much lower than those observed in orthopyroxenes from marie igneous rocks and suggest some type of lamproite affinity. Therefore many of the Fe-rich orthopyroxenes most probably originate by disaggregation of phlogopite-pyroxene or other cognate xenoliths.
Clinopyroxene Either diopside or salite has been observed in virtually every lamproite; most clinopyroxenes are nearly pure diopside-hedenbergite solid solutions (22-26 wt.~ CaO). The AI203 contents in lamproite clinopyroxenes are remarkably low (average 0.5 wt.~) relative to clinopyroxenes of other marie and ultramafic volcanic rocks (less than 5-10 wt.% A1203), except those in some kimberlites (1-2 wt.~); in fact, many lamproite clinopyroxenes have insufficient A1 to fill all the tetrahedral sites. Anomalously high A1203 contents in some lamproite diopsides (less than 1.0 wt.~) are interpreted as indicating a xenolithic relationship or contamination, based on the compositions of grains in both the groundmass and phenocryst populations of the better-studied localities (i.e. LH, WKB and HOL). Lamproite diopsides typically contain 0.4-3.5 wt.% TiO2 (HP with the highest content, PRA with intermediate content, and MAP and LH with the lowest content), 0.0-1.5 wt.~ Cr203 (LH and HP with the lowest content, and WKB and PRA with the highest content) and 0.0-1.5 wt.% Na20 (generally 0.3-0.4 wt.~). Of all the marie and ultramafic rock types, lamproites contain diopsides with the lowest Na20 contents. It is therefore remarkable that aegirine occurs on the rims of diopside grains, especially those within vesicles, in LH rocks (Kemp & Knight 1903; Carmichael 1967; Kuehner 1980) and other lamproites (Wagner & Velde 1985). Clinopyroxenes found in some kimberlites and some ultramarie lamprophyres are similar in composition to those found in lamproites.
Alkali amphiboles K-Ti-richterite is a common alkali amphibole observed as late-stage groundmass grains in most lamproite suites (e.g. HOL, WKB, LH, SB, HP, PRA and GSB), although K-Ti-Mg-arfvedsonites and K-riebeckites also occur (e.g. at PP,
Lamproites and K-rich igneous rocks TABLE 10.
I57
Summary of mineral chemistry of lamproitephases compared with those of other rock types* Olivines Groundmass, phenocryst phases
n SiO2 TiO2 A1203 Cr203 FeO* MnO NiO MgO CaO Total Mgno.
Xenocryst, xenolith phases
LAMP
LAMPH
KIMB
AOB
ALK
KIMB
AOB
LAMPH
58 40_+1 0.0 0.05 0.03 10_+3 0.17 0.33_+0.1 49 _+2 0.17_+0.2 99.9 89_+3
13 40-+1 0.03 0.03 0.04 12_+5 0.24 0.39_+0.7 47 _+5 0.t6_+0.1 100.1 87_+6
150 41+_1 0.02 0.03 0.06 9_+2 0.11 0.32_+0.1 50 _+2 0.07_+0.8 100.2 91_+3
78 38_+2 0.03 0.03 0.03 22_+9 0.35 0.13_+0.11 39 _+8 0.30_+0.1 99.8 75_+12
14 39 0.07 0.04 0.06 20 0.30 0.24 40 0.42 99.8 77
552 41+2 0.03 0.04 0.04 9_+3 0.11 0.35_+0.1 50_+ 3 0.5_+0.6 100.2 90+3
319 40_+1 0.02 0.17 0.04 12_+5 0.17 0.28_+0.1 47 _+4 0.13_+0.1 100.0 87_+6
51 40_+1 0.07 0.02 0.05 10_+2 0.15 0.34_+0.14 49_+ 2 0.07_+0.6 100.2 90+2
Orthopyroxenes Xenocryst, xenolith phases
Groundmass, phenocryst phases
n SiO2 TiO2 A120~ Cr203 FeO* MnO NiO MgO CaO Na20 Total Mgno.
LAMP
LAMPH
INT
LAMP
LAMPH
KIMB
AOB
7 55_+2 0.27_+0.2 0.9_+0.6 0.29 12_+6 0.27 -30 1.0_+0.4 0.11 99.7 81_+10
7 57_+4 0.31_+0.1 1.31_+0.5 0.9 5_+ 1 0.10 -33 2.2_+0.8 0.12_+0.1 99.6 92_+2
37 53 0.17 0.50 -21 0.56 -24 1.2 -100.1 67
6 56+_2 0.07 1.3_+0.9 0.32_+0.3 7_+6 0.24 0.12 34_+ 5 0.6_+0.3 0.10_+0.8 99.8 89_+10
95 56_+ 1 0.18_+0.1 2.7_+1 0.36_+0.2 7_+2 0.16 0.10 33 _+2 0.9_+0.4 0.15+_0.1 100.4 89+_4
657 57_+2 0.10_+0.1 1.4_+1 0.33_+0.2 6_+42 0.13 0.09 34_+ 3 0.64_+0.4 0.13_+0.1 100.1 90_+6
380 54+_2 0.15_+0.1 3.9_+1.5 0.36_+0.2 8_+4 0.17 0.08 32_+ 3 0.88_+0.7 0.10+0.8 99.8 88_+7
Amphiboles Xenocryst, xenolith phases
Groundmass, phenocryst phases
n SiO 2 TiO2 A1203 ChO3 FeO* MnO NiO MgO CaO Na:O K20 BaO Total Mgno.
LAMP
LAMPH
INT
AOB
ECL
ATP
ALK
KIMB
AOB
LAMPH
118 53+2 4.5+2 0.8_+1.7 0.03_+0.5 8__+7 0.39_+1 0.04_+0.4 17_+5 5-+2 4.9_+1.5 4.2-+1.6 0.17_+0.2 97.1 78_+20
53 40_+3 4.3+1 13_+3 0.02_+0.2 13_+4 0.41_+0.5 -12_+3 11+2 2.5_+1 1.4_+0.6 . . 97.6 62_+13
21 44 3.2 11 -15 0.30 -12 11 1.7 0.8 . 98.8 60
27 40 5.2 13 0.4 12 0.17 -12 11 2.8 1.4 . 97.9 64
91 50 0.2 9 0.3 12 0.12 0.11 13 8 3.6 0.29 . 97.0 65
31 49 1.2 9.2 1.0 3.3 0.07 0.10 20 12 2.4 0.31 . 97.2 91
19 41 2.7 12 0.8 18 0.7 -9 11 3.0 1.4 . 98.0 48
55 49_+7 0.9_+1 7.7_+7 0.6_+0.7 5.2_+4 0.10 0.12 19_+4 9___2 3.3_+0.7 2.7_+1.9 . 97.5 86_+11
219 41_+2 3.7_+1.9 14_+2 0.8-+0.9 9.4-+4 0.16 0.04 14_+3 11_+1 2.9-t-0.6 1.2_+0.6
28 40+2 4.1+1.6 14_+1.5 0.4-+0.6 10-+3 0.17 0.06 13_+3 11_.+0.7 2.4-+0.6 1.5+_0.8
97.9 72+_13
97.6 70+_10
.
S. C. Bergman
I58
TABLE 10. (cont.) Summary of mineral chemistry of lamproitephases compared with those of other rock
types* Clinopyroxenes Groundmass, phenocryst phases LAMP n SiO2 TiO2 A1203 Cr203 FeO* MnO NiO MgO CaO Na20 Total Mgno. wo en fs
LAMPH KIMB
158 128 53• 50+3 1.0+0.6 1.6• 0.51• 4.2_+2.7 0.19_+0.3 0.14• 3.5• 6.8• 0.11_+0.7 0.21• 0.06+0.4 0.12• 17+2 14• 24• 22• 0.46• 0.8• 99.7 99.8 90• 78• 47.5 46.2 47.0 42.4 5.5 11.4
188 54• 0.42• 3.5• 0.8-1-1 4.2+2 0.11• 0.03• 16• 19• 2.2• 99.7 87+8 42.0 50.4 7.5
Xenocryst, xenolith phases
AOB
ECL ATP INT ALK LAMP
KIMB
175 48• 2.4• 6.0• 0.2_+0.2 8.5-+3 0.19• 0.03• 12• 21• 1.0• 99.5 72+13 47.1 38.0 14.9
86 55 0.19 8.2 0.8 6.8 0.5 0.4 9.2 14 6.0 99.6 69 44.2 39.0 16.7
1057 724 54• 51• 0.28•177 4.6_+4.1 6.0• 1.0+0.9 0.7• 3.6• 5.0• 0.10 0.13 0.6 0.4 16• 15• 18• 20+2 2.6• 1.1• 99.9 99.7 88• 84• 42.4 44.3 50.9 47.1 6.7 8.6
115 52 0,39 4.8 0.72 2.8 0.8 0.4 16 22 1.0 99.8 91 46.5 48.8 4.7
31 52 0.57 2.6 0.12 8.5 0.25 0.3 15 21 0.37 99.6 77 42.2 44.2 13.6
135 49 1.1 4.8 0.10 10 0.37 0.6 12 22 1.3 99.7 67 47.2 35.5 17.3
24 54• 0.55• 1.2• 0.9• 4.8• 0.29 0.8 16• 23• 0.87• 100.0 85• 46.6 45.6 7.8
AOB
LAMPH 233 53• 6.0• 0.5+_0.6 5.2• 0.11 0.9 14_+3 18+3 2.5+__1.8 99.8 83• 42.9 47.2 9.9
Spinels Xenocryst, xenolith phases
Groundmass, phenocryst phases
n SiO2 TiO2 A1203 Cr203 FeO* MnO NiO MgO V205 Total
LAMP
AOB
KIMB
LAMPH
25 0.2 9• 3• 37• 41• 0.9• 0.1 • 8• -98
37 0.2 18• 5• 7• 61_+14 0.7+0.4 0.1 • 5• 0.4• 97
253 29 0.2 0.2 10• 10• 9_+12 8• 2 0 + _ 2 3 17• 48+23 59+21 0.6• 0.5• 0.1 • 0.2• 11• 7• 0.5• 0.2• 97 97
ATP
LAMP
AOB
KIMB
LAMPH
45 0.1 0.1 28• 39+10 19_+8 0.3_+0.1 0.1 • 13• -99
13 0.2 1.3• 17+11 45• 22+12 0.6_+0.3 0.2_+0.1 13• 0.3• 99
337 0.2 2.8• 42+20 17+15 21+18 0.3_+0.2 0.4• 17• 0.1• 99
503 0.2 4.3+7 16• 36• 29• 0.4_+0.3 0.2• 13• 0.6• 99
48 0.2 1.6• 39• 23• 19+15 0.2_+0.2 0.4• 17• 0.5 99
Ilrnenites Xenocryst, xenolith phases
Groundmass, phenocryst phases LAMP n TiOz A1203 Cr~O3 FeO* MnO NiO MgO CaO Total
KIMB
4 331 50 • 1 50 + 5 0.2• 0.3• 0.05___0.2 1.2• 44 • 2 37 + 9 1.3• 0.9• 0.09 0.1 _0.1 3.3• 9.2• 0.1 0.15• 99 98
LAMPH
AOB
KIMB
LAMPH
AOB
23 52 • 2 0.5• 0.06• 43 • 3 0.5+0.8 0.01 3.8___4 0.26• 99
11 52 0.4 0.05 40 0.57 0.06 5.1+2 0.38 99
582 49• 0.5• 1.3• 37• 0.6• 0.5• 10• 0.04• 98
23 52• 0.5• 1.I• 36• 0.3• 0.0 9.5• 0.05• 99
54 52• 0.5• 0.06• 37• 2.1• 0.04• 7.6• 0.04• 98
Lamproites and K-rich igneous rocks
159
TABLE 10. (cont.) Summary of mineral chemistry of lamproite phases compared with those of other rock
types* Phlogopites Groundmass, phenocryst phases
n
SiOz TiO 2 A1203 Cr203 FeO* MnO NiO MgO CaO Na20 K20 BaO Total Mg no.
Xenolith, xenocryst phases
LAMP
LAMPH KIMB
INT
ATP
ALK
KIMB
258 41 _+2 5.0_+2 10_+3 0.3_+0.4 6.8_+4 0.06 0.11 22_+3 0.07 0.30 10_+0.6 0.8_+0.6 95.2 85_+10
87 37_+2 4.3_+2 15_+2 0.2_+0.4 11_+4 0.14 0.02 18_+4 0.37 0.42 8.5_+2 0.7_+0.4 94.6 74_+14
10 36 2.9 16 -20 0.41 -12 0.08 0.14 9 0.14 96.0 52
5 39 1.3 15 1.2 2.6 0.12 -24 0.02 0.6 9 -93.2 94
48 37 4.6 14 0.02 15 0.31 ~ 14 0.2 0.6 9 1.3 95.5 62
258 25 40_+2 40-+2 1 . 8 _ + 2 3.3-+2 13_+2 14-+1 0.7_+0.8 0.7+0.6 4.8_+2 7_+4 0.03 0.04 0.15 0.04 24-1-2 21 _+3 0.08 0.14 0.33_+0.3 0.45_+0.3 10_+1 10_+0.5 0.2_+0.2 0.9_+0.5 94.6 95.8 90-+5 84_+10
147 39_+2 2.3_+2 13_+3 0.4_+0.5 7_+4 0.06 0.06 23__.4 0.16 0.20 10_+1 1.7_+1 94.2 85+8
Sanidines (or alkalifeldspar) LAMP n SiO2 AIEO3 FezO3 MgO MnO CaO Na20 K20 BaO
Leucites
LAMPH
28 16 63 _+1 65 _+2 17_+ 1 19_+ 1 2.3-+ 1 0.3_+0.1 0.5_+0.4 0.01 0.02 ,0.01 0.25_+0.4 0.24+_0.3 0.47+0.2 3.6_+5 1 6 _ + 0 . 5 11_+5 1 _+0.7 0.5 99 99
AOB
36 37_+2 4.8_+3 16_+1 0.8_+0.8 8.2+5 0.06 0.08 19_+4 0.07 0.9_+0.3 9_+0.7 1.0_+0.6 95.6 80-+13
Priderites
LAMP SiO2 AIEO3 Fe203 CaO Na20 K20
LAMPH
15 56.0 21.0 1.2 0.01 0.05 21.3 99.6
LAMP SiO2 TiO2 A1203 Cr203 FeO CaO Na:O KzO BaO
20 0.1 71.8-+2 0.5_+0.9 0.5_+0.1 9.1 _+3 0.04 0.38_+0.2 7.7_+1.6 9.8_+5 99.6
Garnets Xenocryst, xenolith phases
n SiO 2 TiO 2 A1203 Cr203 FeO* MnO NiO MgO CaO Na20 Total Mgno.
LAMP
KIMB
LAMPH
AOB
18 41 ___1.5 0.27_+0.2 21___2 3.3_+3 12_+9 1.1+1.3 0.01 18_+6 4.5+1.6 0.01 100.0 72+22
1354 41 +- 1 0.36+_0.6 21+3 3.1_+3 11+5 0.38-+0.2 0.04_+0.5 18_+5 6_+3 0.08_+0.2 100.1 74-+14
119 41 _+ 1 0.33_+0.2 22_+1 1.2_+2 12_+4 0.4-+0.2 0.01 17_+4 6.3_+3 0.6_+0.8 100.3 71+_11
31 41 ___1 0.20_+0.1 23_+1 0.8_+1 11+5 0.35-+0.1 0.0 18_+4 5.5_+1 0.05+_0.1 99.9 73-+14
* LAMP, lamproites; LAMPH, lamprophyres; AOB, alkali olivine basalts; KIMB, kimberlites; INT, intermediate volcanic rocks (e.g. andesites); ALK, alkaline plutonic rocks; ATP, alpine-type peridotites; ECL, eclogites. All major-element analyses in weight per cent. FeO*, total Fe as FeO.
I60
S. C. Bergman
MBAY, SB, WKB, HOL and NWI). The fact that many so-called K-richterites from lamproites possess K20/Na20 molar ratios of less than unity is consistent with this late-stage interpretation. Amphibole also occurs as late-stage euhedral crystals lining vesicles in some lamproite lavas. Lamproite amphibole compositions vary widely with the following typical ranges: Mg number, 40-94; A1203, 0.0-1.5 wt.%; TiO2, 1-7 wt.% (Fig. 24). As with lamproite clinopyroxenes and phlogopites, there is usually insufficient A1 to fill all the tetrahedral sites in lamproite amphiboles. Individual K-richterite grains are typically zoned with rims displaying enrichments in Ti, Na and Fe and depletions in K and Mg relative to the cores. Therefore many of the more evolved amphiboles (e.g. some of those in MAP, PEN, MBAY, WKB, PP, SB and HOL) display K20 / Na20 < 1 (molar). The most potassic richterites (K20/Na20 > 1.5 (molar) occur in the WKB, HP and PRA suites. In contrast with the extremely titaniferous phlogopites in the SB suite (more than 10 wt.% TiO2), SB amphiboles have TiO2 contents (3-4 wt.%) similar to those in the average lamproite. The MnO contents of lamproite amphiboles are similar to those of lamprophyres but are significantly higher than most compositionally-similar rock types. K-richterites are also found in some K-rich ultramafic rocks previously thought to be kimberlites (Baster's mine, Pniel, Barkly West (Erlank 1973); this rock is considered a lamproite in the present study) and in MARID-type xenoliths (Dawson & Smith 1977; Erlank & Richard 1977; Dawson 1979); however, these tend to be depleted in Ti relative to lamproite amphiboles. Relative to nearly all other related rock types, lamproite amphiboles are compositionally unique, being the most depleted in A1 and Ca and enriched in K, Ti and Si. Mason (1977) studied the geochemistry of Krichterites from Wolgidee Hills, Western Australia, and found that they are extremely depleted in REE relative to other amphiboles from alkalicmafic magmatic settings. This REE depletion is in agreement with the interpretation mentioned above, i.e. that K-richterite precipitated relatively late in the crystallization sequence following the crystallization of REE-enriched phases such as apatite.
Phlogopite Ti-rich Al-poor phlogopites and tetraferriphlogopites of widely-varying compositions occur as an essential phase in all lamproites, as both phenocryst and/or groundmass phases. Grains with Al-deficient tetraferriphlogopite margins,
often demonstrating a reverse pleochroic formula, are common in both phenocryst and poikilitic groundmass grains. The characteristics of lamproite phlogopites are generally as follows: Mg numbers, 37-94; TiO2, 2-11 wt.%; Cr203, 0.0-1.5 wt.%; A1203, 1-14 wt.%; SiO 2, 3843 wt.%. Extreme zoning in Ti, A1, Mg and Fe is often observed within a given phenocryst, with grain rims most typically enriched in Ti and Fe and depleted in Mg and A1 relative to the cores. The most AlzO3-depleted phlogopites (1-3 wt.%) occur in the WKB, PRA (Kimberlite mine only) and HP suites (see Fig. 25), whereas the MAP and LH suite phlogopites are relatively A1203 rich (11-14 wt.%). The SB and WKB lavas contain the most TiOz-rich phlogopites (7-12 wt.% and 5-11 wt.% respectively) of all lamproite localities. Lamproite phlogopite compositions are distinct from those in almost all igneous and metamorphic rocks (Fig. 25) with the exception of some minette phlogopites (e.g. Bachinski & Simpson 1984; Velde, personal communication, 1985). The most notable exceptions are some kimberlites and their peridotite xenoliths, especially the MARID-suite types (Dawson & Smith 1977; Erlank & Richard 1977; Dawson 1980). Velde (personal communication, 1985) described phlogopites intermediate between phlogopiteannite-tetraferriphlogopite and ferri-annite from a Jersey minette.
K-feldspar Sanidine or rare microcline (and/or orthoclase) are the only feldspars found in the more aluminous members of various lamproite suites (e.g. HOL, LH, KAM, GSB, SB and MAP). Sanidine is restricted to the groundmass and is a late-stage phase. With the exception of the MAP rocks, lamproite sanidines are extremely K rich (Or>95 ; Na20 = 0-2 wt.%; CaO <0.5 wt.%) with anomalously high Fe203* (typically 0.5-5 wt.%) and BaO (0.2-1.6 wt.%) contents. The extreme Fe-rich character of lamproite sanidines separates lamproites from nearly all other rock types; the sanidines from lamprophyres are usually in the range 0r60_95 and contain about 1.0 wt.% Fe203 (Bachinski & Simpson 1984; Rock 1986). The absence of plagioclase in lamproites provides an effective means for distinguishing lamproites from other ultrapotassic-potassic rocks.
Leucite and analeime Leucite is the only feldspathoid that occurs in lamproites. Euhedral and anhedral leucites are
Lamproites and K-rich igneous rocks
161
~
il
i! <~ ~-~C~ i
i
(% JaW OeO
,9.~
/,,"
..~. /,,0 / ..q',
I,,
l': 2:ili i /I/
o \
A
,
c
c
e~
~ v
~..!
.
~e
~X
9
~o
x
0~ .~=? ..~
K~
li",I I I
1
o (% J.~) CO'iV
~
(% ~
0
~0!I
.
~d~2
I62
S. C. Bergman 20..... MAP'
.
GSB
~ . . . .
10-
"...... ~ ~ : -
~"MBA~jl-"
~.
~ ,,..
~,., /1.4
.." VY-~ ,ZI I
0 Mg number
20~ - , . - ~ " ~ : ' ~ . " : r . . ~. . . . .
.
~
'
~
_
~
\
eo10-
%
Mg number
.... ....... . . . .
" LAMPROITES (229) LAMPROPHYRES (98) ALKALINE INTRUSIVES (32) FKIMBERLITE XENOLITHS (193) t-K G R O U N D M A S S - P H E N - X E N C (166)
12-
10-
A8-
Y 4-
'30
;0
~
do
7'0
do
~
40
MO number
FIG. 25. Phlogopite compositions from lamproites compared with other rock types: (a) At203 versus Mg number as a function of the lamproite suite; (b) A1203 and (c) TiO2 versus Mg number for lamproites and other rock types. (See Table 10 and Appendix 2.) Bold and light-face curves as in Fig. 24.
Lamproites and K-rich igneous rocks
typically weakly anisotropic and twinned (e.g. WKB) or isotropic and twin free (e.g. LH) and are most often altered and pseudomorphed by secondary phases (e.g. sanidine, analcime, quartz and zeolites). Lamproite leucites are commonly non-stoichiometric with a marked excess of Si, Fe 3+ and K. Leucite Fe20 3 contents are as high as 1-2 wt.%. These chemical features separate lamProite leucites from the leucites of other potassic-ultrapotassic rocks (e.g. RP and HM) which possess higher A1 and lower Si, K and Fe contents. Analcime has been recognized in the SB, KAM, LH and MAP lamproite suites and most probably originates by the secondary replacement of leucite, although this view is debatable. Whereas lamproites never display classic pseudoleucite textures (i.e. K-feldsparnepheline intergrowths (Shand 1927)), other potassic-ultrapotassic rocks (e.g. HM, TAN and NSW) commonly contain classic pseudoleucite.
Fe-Ti-Cr-Mg-AI oxides A wide variety of oxide phases in the rhombohedral, cubic and orthorhombic series have been noted in lamproites. Spinels (sensu lato) fall into four compositional groups and occur in rocks from WKB, LH, PRA, GSB, ARG, HP, HOL and MAP, among others: (a) Ti-poor Al-rich Mg chromites; (b) Ti-rich A1-Mg chromites; (c) A1poor Ti-Mg chromites; (d) Mg-Ti-magnetites. Whereas some of these spinel types are commonly found in rocks other than lamproites, lamproite spinels are more Fe rich than kimberlite spinels but are similar to those of other alkaline igneous rocks (Haggerty 1976). Spinel zoning, when present, displays Cr-depleted and Fe 3 +-enriched rims relative to the cores (Mitchell 1985). Lamproite spinels tend to be more Mn rich than those from other alkalic mafic rocks (Table 10). Armalcolite has only been found in SB lavas (Velde 1975), but has been produced experimentally in studies on WKB rocks by Arima & Edgar (1983) (see below). Ilmenite with compositions relatively similar to those of other rock types has been noted in the GSB lavas (4 wt.% MgO (Sheraton & England 1980)), at Sisco (COR) (Velde 1968), at Oscar's Plug (WKB) (S. C. Bergman, unpublished data), at Mount North (WKB) Wagner & Velde 1986a) and at Jumilla (MAP) (2.5-3.5 wt.~ MgO) (Mitchell 1985), but is generally absent. Lamproite ilmenite MnO contents are higher, on the average, than those from compositionally related rock types (except alkali basalt xenolith ilmenites) (see Table 10). Anatase occurs in the altered BOB rocks (Mitchell 1985) and rutile occurs in the CHE lamproites;
163
Van Kooten (1980) and Kuehner (1980) discussed ferro-pseudobrookites from the CSN ultrapotassic rocks and LH lamproites respectively, but these phases are all most probably secondary.
K-Fe-Ti-Ba oxides The natural occurrence of the exotic phases priderite ((K,Ba)(Yi,Fe3+)8016) and jeppeite ((K,Ba)z(Ti,Fe 3+)6013) is almost restricted to lamproites. Priderite has thus far been found in the WKB, COR, LH, KAM, HOL, SB and PRA lamproites, and will most probably be discovered in other suites as more detailed studies are performed. Prider (1939) originally mistook priderite for rutile in the WKB rocks, and Cross (1897), Smithson (1959) and Johnston (1959) mistook it for rutile in LH rocks; the more detailed study of the oxides in the WKB rocks by Norrish (1951) and in the LH rocks by Carmichael (1967) demonstrated its existence. Apart from the priderite-bearing apatite glimmerites in India (Gupta et al. 1983), which are probably lamproites, the only known extra-lamproite priderite occurrence (albeit Ba enriched relative to lamproites) was reported by Zhuravleva et al. (1978) in the olivinite-ijolite members of a carbonatite pluton in the Kola Peninsula, U . S . S . R . K . Collerson (personal communication, 1984) identified priderite in AIL minettes, Labrador. N6mec (1985) described priderite from the K-richer members of Czechoslovakian minettes which may be lamproites. Jeppeite (Bagshaw et al. 1977; Pryce et al. 1984) is apparently restricted to the Wolgidee Hills pluton (WKB) as well as the PRA priderite-bearing phlogopite pyroxenite xenoliths (Mitchell & Lewis 1983). Dubeau & Edgar (1985) presented experimental data on the stability of priderite as a function of pressure and temperature. Much more work on the distribution of priderite-jeppeite in lamproites and minettes is therefore required to understand its paragenesis.
Wadeite Wadeite (K4Zr2Si6018) is a hexagonal ringsilicate which was also originally discovered in the WKB lamproites (Prider 1939), where it occurs as rods up to 1 mm long. Carmichael (1967), Kuehner (1980), Kuehner et al. (1981) and Henage (1972) additionally noted its presence in LH and KAM rocks. B. H. Scott-Smith (personal communication, 1983) discovered wadeite in lamproites from PRA. Tikhonenkov et al. (1960) reported an extra-lamproite occurrence of
I64
S. C. Bergman
wadeite in nepheline-feldspar veins of the Khibina alkaline massif, U.S.S.R. Wadeite is compositionally related to dalyite (K2ZrSi6Ols) and other Zr silicates (zektzerite, sogdianite and darapiosite) which occur in rocks more enriched in SiO2 than lamproites (e.g. hyperalkaline granites (Robins et al. 1983; Thibaut et al. 1972)). Arima & Edgar (1980) studied the stability of pure wadeite and found that, by itself, it can occur under a wide range of uppermantle to crustal P - T conditions.
Shcherbakovite (noonkanbahite) Shcherbakovite ((Na,K)(Ba,K)TiaSi4014) is an unusual phase that was discovered in the WKB wolgidites by Prider (1965). Eskova & Kazakova (1955) initially reported its occurrence in alkalic pegmatites of the Khibina massif, U.S.S.R. Shcherbakovite is closely compositionally related to batisite (Na2BaTi2Si4014), which was initially reported by Kravchenko et al. (i960). Unfortunately, shcherbakovite has not received much attention since Prider's original work and only a limited amount of mineralogical data is available.
Other phases Important and common accessory phases of all lamproites include apatite and perovskite. Sphene (mostly secondary) and zircon are rare accessory minerals. Wagner & Velde (1986a) described rare roedderite-like phases (K, Fe, Mg silicates) in MAP and KAM lamproites. Lamproite secondary phases are widespread and include Sr-rich barite, chlorite, heulandite, nontronite, other zeolites, serpentine, carbonate minerals, quartz and rare albite. As mentioned above, xenocrysts of a wide variety of phases, including diamonds and Cr-rich pyrope, can occur in lamproites. However, lamproite pyropes are more enriched in MnO than are kimberlite and other pyropes, but are otherwise similar in chemistry.
Mineralogy and mineral chemistry: summarizing comments Lamproites contain a diverse suite of minerals, some of which are restricted to lamproites. The diagnostic mineral assemblages (i.e. containing phlogopite, K-richterite, priderite, shcherbakovite and wadeite) and unique mineral compositions (K-rich, Ca- and Al-poor amphiboles, Na- and Al-poor diopsides; Ti-rich, Al-poor phlogopites; Ba- and Fe-rich sanidines; Fe-, K- and Si-rich leucites; the mafic minerals largely possessing
primitive Mg numbers in the range 85-93, although more differentiated phases can occur) shown by lamproites reflect their peralkaline and primitive compositions and provide effective means for distinguishing lamproites from other alkalic mafic--ultramafic magmatic rocks. As discussed at length by Mitchell (1985), the zoning characteristics displayed by lamproite phases place important constraints on both the crystallization history of individual magmas and the petrogenetic relationships between several lamproite magmas. Interestingly, the elevated MnO contents of lamproite amphiboles, ilmenites, spinels and garnets are in marked contrast with the MnO-depleted character of the lamproite whole-rock samples (0.10 wt.% (Table 5)) compared with kimberlites, alkali basalts etc.
Experimental studies on K-rich rocks and their minerals Reviews of experimental studies dealing with Krich rocks and their minerals have been given by Gupta & Yagi (1980) and Gittins (1979). Therefore important and more recent papers on the subject are summarized in this section.
Natural lamproites The high-P-T phase relations of lamproite rock compositions have been determined by Barton (1976), Barton & Hamilton (1978, 1979, 1982) and Arima & Edgar (1983). Barton & Hamilton studied orendites, wyomingites and madupites from the LH and found that only the orendites and wyomingites could equilibrate with a garnet lherzolite assemblage at P > 2 6 kb and Pi_i2o= Ptotal" Madupites probably represent partial melts of phlogopite-pyroxenite or phlogopite-olivinepyroxenite assemblages (Barton & Hamilton 1979). Barton & Hamilton's experiments indicate that the dominant liquidus or near-liquidus phases are leucite, olivine, orthopyroxene, clinopyroxene and garnet at P < 3 0 kb. They also concluded that the peralkalinity of these ultrapotassic magmas could reflect either primary source rock compositions or the selective melting of phlogopite + pyroxene. They postulated that the association of low-SiO: madupites with highSiO2 orendites at LH results from variations in H20 and CO2 and local mineral assemblages in the upper mantle source region. It should be noted that A. Edgar (personal communication, 1985) questions the validity of the phase relationships determined by Barton because of Fe partitioning in platinum capsules. Sobolev et al.
L a m p r o i t e s a n d K-rich igneous rocks (1975) studied the low-P near-liquidus phase equilibria of a LH wyomingite and found that diopside is the sole liquidus phase at temperatures around 1320~ and P = 1 b despite the fact that phlogopite phenocrysts are common and that the 1 b solidus temperature is about 1000~ Arima & Edgar (1983) studied the hydrous (+ COO liquidus and sub-liquidus phase relations of a wolgidite (from the WKB, Mount North) to 40 kb. They found that olivine (at P < 15-24 kb) and orthopyroxene (P > 15-24 kb) occur on the liquidus; phlogopite, rutile, clinopyroxene, armalcolite and priderite follow at lower P and T. Rutile reacts with phlogopite and liquid to produce priderite at P < 15 kb and T < 1010~ and armalcolite reacts to form priderite at T< 1010~ and P < 15 kb; rutile is the high-pressure phase and armalcolite is the high-temperature phase. These phases (priderite, armalcolite and rutile) were only found in runs containing added H20 (13 wt.%) and were absent from runs with low H20 contents (3 wt.~). Their results indicate that wolgidite magma most probably represents the partial melt of a source mantle containing phlogopite, rutile, olivine and orthopyroxene. It is unlikely that a wolgidite-composition melt would be derived from the partial melting of a simple garnet- or spinel-bearing lherzolite mantle, in contrast with the derivation of typical alkali basalts from such a mantle. Other ultrapotassic rock compositions
Edgar et al. (1976, 1979), Edgar (1979), Edgar & Arima (1983), Ryabchikov & Green (1978), Arima & Edgar (1983) and Edgar & Condliffe (1978) investigated the phase behaviour of several TAN rocks (biotite ugandite, biotite mafurite and katungite) in the presence of H20 + CO2 to 40 kb. They found that these melts never equilibrate with an orthopyroxene- or garnet-bearing assemblage and that only diopside, olivine, ilmenite and phlogopite occur on the liquidus, even at high pressure. These three magmas cannot be related to each other by fractionation alone or by partial melting of a single mantle source region. The three magma compositions could be derived by partial melting, at various depths, of clinopyroxenite or peridotite sources with varying degrees of enrichment in K where low H 2 0 / C O 2 ratios produce low enrichments in K. Their experimental findings favour metasomatism of a peridotitic mantle source as a prerequisite for the production of K-rich magmas. Ryabchikov & Green (1978), however, concluded that biotite mafurite, olivine leucitite and ugandite liquids could be produced by the partial melting in the presence of H20 and CO2 of a
I65
lherzolite source mantle locally enriched in phlogopite. Phlogopite occurs on the liquidus at high H20/CO 2 ratios (Xco2=0-0.25), but becomes unstable at higher CO2 contents (Xco2 > 0.25). Cundari & O'Hara (1976) studied the phase equilibria of a leucitite from NSW under anhydrous conditions to 40 kb and determined that garnet did not exist with enstatite above 35 kb, indicating that orthopyroxene did not occur in an anhydrous source peridotite. Thompson (1977) experimentally studied a clinopyroxene leucitite from the RP and found that the phenocryst assemblage could be duplicated at 14 kb and 1260~ indicating a mantle derivation. He postulated that the experimental data were consistent with the view that the apparent crustal contamination (suggested by SrO isotope data) was due to the partial fusion of subducted ocean-floor sediments within the upper mantle. Other experimental work on potassic and ultrapotassic rocks includes that of Dolfi et al. (1976), Lloyd et al. (1985), and Esperanca & Holloway (1985). Compositions in the system N a 2 0 - K 2 0 - M g O AI203-SiO2-H20-CO 2
Schairer & Bowen (1938) studied the anhydrous system KA1Si206-CaMgSi206-SiO2 at 1 atm and established the phase equilibria pertinent to Fe-free K-rich alkaline magmas. They found an extremely large leucite field in the leucitediopside-silica ternary system. However, more recent work on the hydrous analogue of this system (Ruddock & Hamilton 1978a) demonstrates that the leucite field is extremely compressed, which explains the general absence of leucite in minettes. Luth (1967) determined the phase equilibria in the system MgzSiO 4KA1SiO4-SiOz-H20. Bravo & O'Hara (1975) studied the partial melting of a synthetic phlogopite-bearing garnet and spinel lherzolite (CO2 free) and found that partial melts of these solids at 15 and 30 kb had extremely low K20 contents (1-4 wt.~) and were quartz- and/or hypersthene-normative, These liquids share an SiO2-rich compositional feature with lamproites, although they are relatively depleted in K20 relative to lamproites. Modreski & Boettcher (1972, 1973) investigated the stability of phlogopite in a model system and found that it is unstable under oceanic geothermal conditions of P > 25 kb but may persist in sub-continental geothermal conditions to P > 50 kb. Wendlandt (1977a, b) and Wendlandt & Eggler (1980a, b, c) investigated the stability and melting behaviour of sanidine and phlogopite in olivine-
I66
S. C. Bergman
bearing assemblages. They found that sanidine + forsterite breaks down to kalsilite+enstatite along a P-Tlocus including 20 kb at 1000~ and 30 kb at 1300~ indicating that the latter assemblage will be stable along most geotherms (Wendlandt & Eggler 1980c). The melting behaviour of phlogopite in peridotite was found to be controlled by the equilibrium mineral assemblage as well as by the composition of the volatile phase (H20 versus CO2). At P < 1 4 kb the liquid compositions were quartz or enstatite normative, but became leucite to kalsilite normative at higher pressures in the presence of CO2 + H20 vapour. At pressures above 30 kb in the presence of magnesite, phlogopite peridotite melts to produce carbonatitic liquids which become progressively enriched in K20 and SiO2 as pressure increases. Phlogopite is not stable in a peridotite assemblage above 50 kb. Ryabchikov & Boettcher (1981) investigated the solubility of K in an aqueous fluid in equilibrium with phlogopite-pyrope-forsterite gels at 10-30 kb and 1050-1100~ and found that water is capable of dissolving 4-25% K20 and that solute KzO/AI203 ratios are close to unity. Ruddock & Hamilton (1978a) experimentally studied the system leucite-diopside-quartz-H20 to 4 kb and confirmed that, under hydrous conditions, diopside and phlogopite will be the first two liquidus phases followed by sanidine and quartz. This study nicely explains the petrography of minettes and other lamprophyres (however, see Rock (1984, Fig. 8) for an alternative interpretation). Ruddock & Hamilton (1978b) determined the stability of carbonate in an ultrapotassic mafic assemblage and found that a minette composition liquid is saturated in olivine, garnet, orthopyroxene and clinopyroxene at pressures above 15-20 kb and is saturated in carbonate at pressures above 20-30 kb. Schneider (1982) and Schneider & Eggler (1984) determined the high P - T (15-20 kb and 750-1100~ solubility of major-element oxides in H20 and H20-CO2 fluids in equilibrium with jadeite-, amphibole- and phlogopite-peridotite and several individual minerals. They found that the fluid's solute contents were in the range 0.518 wt.% and increased with decreasing CO2 content. H20 or H20 + CO2 fluids in equilibrium with phlogopite-peridotite were found to be peraluminous whereas H20 at- CO2 fluids in equilibrium with amphibole- or jadeite-peridotite ranged from mildly peraluminous to strongly peralkaline in their solute compositions. Solute K/Na ratios were found to be controlled by H20 / CO2 as well as by P, T and the bulk composition of the solid, whereby K/Na increases with increasing CO2/H20 ratio of the fluid.
Petrogenesis A wide variety of models have been advocated for the evolution of K-rich rocks; the reader is referred to previous reviews by Bell & Powell (1969), Sahama (1974) and Gupta & Yagi (1980). Most hypotheses can be placed in one of three broad groups. 1 Partial melting of mantle material that was metasomatically (or otherwise, e.g. by plutonism) enriched in phlogopite and other LIL-elementenriched minor phases (e.g. apatite, zircon, sphene etc.); the melt may be subject to variable amounts of fractionation but does not undergo substantial assimilation of crustal material (Waters 1955; Yagi & Matsumoto 1966; Kay & Gast 1973; Boettcher et al. 1975; Beswick 1976; Gupta et al. 1976; Van Kooten 1980). 2 Assimilation of continental crustal material by 'normal' mantle-derived alkali-basaltic, carbonatitic or alkali-ultramafic melts (Daly 1910, 1933; Shand 1931; Rittman 1933; Larsen 1940; Williams 1936; Holmes 1950; Powell & Bell 1970; Taylor & Turi 1976; Turi & Taylor 1976; Taylor et al. 1984). 3 Evolution involving processes such as zone refining, fractional resorption, gaseous transport etc. in otherwise typical mantle-derived melts (Bowen 1928; Saether 1950; Kennedy 1955; Harris 1957; Prider 1960; Marinelli & Mittempergher 1966; Fuster et al. 1967; Kogarko 1980; Kogarko et al. 1968; Kushiro & Aoki 1968; Harris & Middlemost 1969; Stewart 1979; Wones 1979; Ryabchikov et al. 1982). Sahama (1974) observed that researchers tend to desire hypotheses that attempt to explain the petrogenesis of K-rich rocks in general; however, the geochemical distinctions between lamproites and kamafugitic as well as other K-rich rocks do not permit such a luxury. The tectonic-geological and geochemical distinctions between lamproites, kamafugites, shoshonites and other Krich rocks clearly support this view. The last statement in Sahama's (1974)review of K-rich rocks speculates on a relationship between kimberlites, lamproites and kamafug o ites. Gupta & Yagi (1980) additionally recognized the importance of this speculation. The last 10 years have witnessed major discoveries of diamondiferous lamproites and other developments which make this speculation a reasonable, if not undeniable, view. Although kimberlites and olivine lamproites evolve differently, they possess many similarities which have led several workers to suggest a petrogenetic model for lamproites
Lamproites and K-rich igneous rocks whereby kimberlite magma is differentiated to produce a more potassic SiO2-rich lamproitekamafugite magma (Holmes 1932; Wade & Prider 1940; Scott 1981). Although more data are required to establish an unambiguous petrogenetic picture for lamproites, the data now available make it possible to constrain their petrogenesis to a much greater degree than was possible as little as 10 years ago. In addition, many aspects of the petrogenesis of lamproites have important implications for a variety of aspects of mantle evolution and magma genesis. Lamproites are mantle-derived melts and do not simply result from the melting of recycled continental crust. This is supported by (a) the presence of diamonds in olivine lamproites, (b) the presence of mantle-derived peridotite xenoliths (albeit much rarer than their abundance in alkali basalts and kimberlites) in leucite, phlogopite and olivine lamproites, (c) a primitive (Mg number 74 + 9) mafic-ultramafic major-element composition, (d) an enrichment in the refractory trace elements Co, Cr, Ni and Sc, and (e) the mantle-type pressures and temperatures (P = 2030 kb and T= 1100-1300~ calculated by performing thermodynamic isoactivity calculations (Nicholls & Carmichael 1972; Carmichael et al. 1974, 1977; Barton & Wood 1976) in which lamproite magmas are equilibrated with various source-mantle assemblages. In contrast with being enriched in LIL elements, lamproite magmas possess major-element compositional features typical of partial melts of a refractory (i.e. Mg number, 88-92, diopsidedepleted, harzburgite ?) mantle : they are depleted in the incompatible elements Na, A1 and Ca compared with alkali basalts and other primary melts considered to represent low degrees (less than 10%) of partial melting of a lherzolitecomposition mantle. A depleted diopside-poor source is further supported by the high-P-T experimental studies on lamproite compositions which suggest a non-lherzolite mantle source for lamproite magmas (see above) and also by the characteristically refractory harzburgite and dunite xenolith population that occurs rarely in several lamproite suites (e.g. WKB, LH, MAP). However, there are other ways of interpreting this depletion in certain major elements. Lamproites are phlogopite-rich rocks and many whole-rock analyses of lamproites approach but are more calcic and siliceous than the bulk composition ofphlogopite. Therefore it is possible that lamproites represent the selective fractional fusion of mantle phlogopite and other LILelement enriched phases (e.g. apatite, zircon etc.) diluted by small amounts of peridotite compo-
167
nents (clinopyroxene, orthopyroxene and olivine). Since phlogopite is severely depleted in Na and Ca and moderately depleted in AI compared with basaltic melts, this model nicely explains the bulk chemistry of lamproites. It is further supported by the fact that phlogopite is not expected to survive extensive amounts of partial melting, most of it being consumed at extremely low degrees of melting. The ultrapotassic but variable K/Na ratios of lamproites can be interpreted in a number of ways. K can be enriched relative to Na as a result of (a) source-mantle characteristics, i.e. precipitation of phlogopite (Na absent) via metasomatism in a Na-poor refractory peridotite, (b) fractionation of high K/Na phases (e.g. phlogopite, leucite) and (c) selective volatile-phase transfer of K relative to Na. The experimental work by Schneider & Eggler (1984) and Schneider (1982) contributes important data with which to judge the effectiveness of the latter process, i.e. K/Na in the fluid varies with CO2/H20. However, much more work on it is required to understand vapour-phase transfer of alkalis properly. Since this feature is a local as well as a general feature of lamproites, it most probably results from a combination of all these processes (see discussion in Sahama (1974, p. 106).) Like alkali basalts and lamprophyres, the volatile budgets of leucite-lamproite and phlogopite-lamproite magmas are water-dominated (relative to CO2), whereas those of kimberlites (and olivine-lamproites) possess higher and subequal H20 and CO2 contents. The H20 contents of kimberlites indicated in Tables 4 and 5 are maximum values for the magmas, and the true values are most probably a factor of 2 lower owing to the ubiquitous and often secondary serpentinization that characterizes kimberlite groundmass and xenocrysts. As mentioned above this HzO-rich character provides an explanation for the sherbet-glass shape of lamproite pipes. It is also consistent with the SiO2-rich and variable composition of lamproites. Experimental work by Mysen & Boettcher (1975a, b) and Eggler (1977) has shown that partial melting of a peridotite with a high H20/(CO2 + H 2 0 ) ratio will produce a relatively SiO2-rich melt, eliminating the need to invoke fractionation of low-SiOz mafic minerals to produce SiO2-rich lamproites. Slight variations in the source H 2 0 / ( C O 2 -k-H 2 0 ) ratio would produce variations in the SiO2 contents of partial melts (e.g. Bachinski & Scott 1979). Therefore, whereas kimberlites and alkali olivine basalts are SiO 2 undersaturated and are most probably derived from a mantle source with a low H20/(CO 2 -k-H20) ratio, lamproite sources represent the complementary volatile situation
I68
S. C. B e r g m a n
(high H20/(CO 2+H20)) in addition to other contrasts, i.e. LIL-element enrichments. Of all the alkaline rocks, lamproites are consistently among the most enriched in LILelements. They are also characterized by enrichments in higher-series elements relative to lowerseries elements in the same group (e.g. lower K/ Rb and Sr/Ba) compared with virtually all other rock types. It is now clear that this feature is a mantle phenomenon and does not require crustal contamination for its explanation. The highly variable concentrations of LIL-elements in lamproites is easily explained by a heterogeneous distribution of LIL-element-enriched phases in the source mantle in addition to volatile-phase transfer processes acting during magma ascent. These minerals, most importantly phlogopite but also including other LIL-element-enriched phases such as apatite, monazite, sphene, rutile and zircon, originated from the process of mantle metasomatism, most recently reviewed by Bailey (1982) and Dawson (1984). These phases could be precipitated or otherwise produced from any number of sources, including (a) fluids ascending from a down-going slab which hybridize the overlying mantle wedge, (b) fluids emanating from an ascending plume, (c) 'background' deepmantle fluids ascending throughout the upper mantle and (d) silicate melts of variable origins. As mentioned above, this metasomatism must have been fairly old (10-103 Ma) to generate the 87Sr/86Sr and 143Nd/144Nd ratios of lamproites. The age of this lamproite-source metasomatism contrasts with that suggested for the source regions of alkali basalts and related melts. Since H20 is suggested as the most important 'lamproite volatile', the mantle metasomatism associated with lamproite source regions may be distinct from that associated with alkali basalts (CO2-rich?). If we accept the view that olivine lamproites and leucite lamproites (or phlogopite lamproites) form a cognate petrological continuum, a view that is supported by the time-space-composition relationships present within the WKB and PRA suites, several implications arise. The compositional variations in these suites indicate the existence of differentiation processes more complex and less understood than simple crystal fractionation. Also, the preservation of diamonds within the WKB and PRA suite diatremes indicates that these magmas explosively ascend to the surface from depths of greater than 150 km, i.e. within the diamond stability field in the upper mantle because lamproite melts, which are extremely hydrous, would otherwise oxidize the diamonds. Although this feature is shared by kimberlites, many lamproite diamonds display a
greater degree of graphitization and rounding than those from kimberlites; these features may indicate that lamproite magmas ascend more slowly than kimberlites. Much more work on the stability of diamond in lamproite magmas is required to acquire a better understanding of the ascent characteristics of lamproites. Another implication resulting from the mantle derivation of lamproites involves the radiogenic isotope systems Sm-Nd and Rb-Sr. Some researchers have recently interpreted the isotopic data of K-rich rocks as indicating the operation of continental crust contamination (e.g. TAN (Vollmer & Norry 1983b)) or the mixing of a depleted mantle component (similar to MORB: eNa = + 10 and eSr 40) with an enriched mantle component (eNd= -- 16 and ~Sr= + 240, e.g. WKB (McCulloch et al. 1983a, b)). However, the isotopic data can also be interpreted in the context of a single-mantle reservoir that is heterogeneous but enriched in its Rb/Sr and Nd/ Sm ratios relative to the bulk Earth. The heterogeneities could be related to partial melting or plutonic processes. Small differences in these ratios can produce the time-integrated range in 143Nd/l*4Nd and 87Sr/86Sr ratios observed for lamproites. Regardless of the specific model used to interpret the isotopic data (as all are nonunique), an enriched mantle source is required by all. This 'enriched' source is also represented by Group II kimberlites (Smith 1983a, b) and the diopsides from kimberlite xenoliths (Menzies & Murthy 1980a, b; Basu & Tatsumoto 1980). However, the major-element chemistry of lamproites is characterized by a depletion in some of the major elements that are the first to enter partial melts (Ca, AI and Na) and an enrichment in refractory elements (Mg and Ni) so that the source probably experienced a partial-melting event prior to the enrichment in Nd/Sm and Rb/ Sr (see above). When this is viewed in the light of the Pb isotope data obtained by Fraser et al. (1985) and Nelson et al. (1986) (see above), it is seen that lamproite mantle sources must have been subjected to depletion events followed by enrichment events a long time ago ((1-3)x 103 Ma). This model has interesting implications for Earth history because not only does it suggest the enriched complement to the depleted mantle source represented by MORB and nearly all alkali basalts (except suites such as Kerguelen and Tristan da Cunha), it also supports the existence of Precambrian partial-melting episodes in the mantle and suggests a three-stage model of petrogenesis. If the more recent lamproite suites that occur overlying fossil Benioff zones indicate the general tectonic environment of lamproites, then it is possible that lamproites represent =
- -
Lamproites and K-rich igneous rocks partial melts of sub-continental lithosphereasthenosphere that was (a) initially depleted by the formation of MORB at a spreading centre, (b) moved beneath a continent and (c) later hybridized perhaps by fluids and melts emanating from a down-going slab. All this must have taken place sufficiently long ago to permit the development of time-integrated high 87Sr/86Sr and low 143Nd/14aNd ratios. The most recent melting event that led to the surface emplacement of lamproite magmas could be triggered or enhanced by a number of mechanisms, including transient hot-spots (randomly contacting the unique lamproite-source hybridized mantle; unlike ocean island chains, however, lamproites do not generally form time-space-related linear trends), the build-up of radiogenic heat (due to sources enriched in K, Th and U) and the tectonic conditions of the overlying crust (involving deep fractures). It should be noted that variations of this three-stage model of depletion-enrichmentmelting have been suggested by various workers (Best 1975; Beswick 1976; Thompson et al. 1984). When the lamproite sensu stricto and other ultrapotassic rock localities are plotted on a map with the continents in the positions of 180-200 Ma ago (see Fig. 26), a zonal pattern of the distribution of these K-rich rocks emerges. Two broad belts are evident: one belt 3000 km long includes the Indian, W Australian and Antarctic occurrences, and the other more diffuse belt includes the N American, European and W African localities. Since the mantle source regions
9
~ 7
of lamproites are evidently very old, this pre-drift zonation is perhaps expected and not only permits the possible prediction of other lamproite localities but also has important implications for mantle evolution. Is it possible that the southern hemisphere lamproite belt (representing 1.2x 103 Ma of magmatic activity) delineates a Precambrian (Archaean?) zone of recycling of crustal material (fossil Benioff zone)? While much speculation has filtered into this discussion, it is hoped that a coherent picture of lamproite petrogenesis has emerged. The lamproite revolution is only in its infancy and future studies of these exotic rocks will surely clarify many of the issues only briefly discussed here.
Kimberlite--lamproite relationships Dawson (1987) summarizes the similarities in geochemistry and mineralogy between olivine lamproites and the more micaceous kimberlites (including those in Group II of Smith (1983a, b)) and suggests that olivine lamproites may, in fact, be members of the Group II kimberlites. Smith (1983b) demonstrated Nd-Sr-Pb isotopic similarity between Group II kimberlites and olivine lamproites from PRA and WKB and suggested further that they are one and the same, a view that is additionally held by the present author. Smith (1983b) also suggested that lamproites (as well as Group II kimberlites) were derived by the partial melting of sub-continental lithosphere,
~
~'/~'~ ~
169
LAMPROITES 8~ II ULTRKAsPOTASSIC~/
~'~"~7
9 KIMBERLITES j
EAnt;
~.
\\ \
4-
"')l.".
-
\
,,
4-
FIG. 26. Distribution oflamproites, ultrapotassic rocks and kirnberlites on a map showing a reconstruction of the continents 180-200 Ma ago (Owen 1983). Lamproites fall along two broad belts.
I7O
S. C. Bergman
whereas Group I kimberlites were derived by the partial melting of sub-continental lithosphere or underlying asthenosphere. The fact that MARID-suite xenoliths occur in kimberlites, indicating lamproite-type plutonic-metasomatic processes in sub-continental lithosphere encountered by the host ascending kimberlite magmas, is consistent with both lamproite and kimberlite melts existing in similar areas in the upper mantle.
Relationships between lamproites and MARID and other mica-rich xenoliths The similarity between the geochemistry and mineralogy of lamproites on the one hand and the M A R I D glimmerite and PKP (amphiboleperidotite) suites of xenoliths found in kimberlites on the other suggests some type of relationship. MARID xenoliths, which are also known as 'mantle pegmatites' consist of mica, amphibole, rutile, ilmenite and diopside (Dawson & Smith 1977); in addition, compositionally-similar phases (Ti-mica, K-richterite) are observed as metasomatic alteration products of many peridotite xenoliths from kimberlites (Erlank 1970, 1973; Erlank & Rickard 1977; Erlank et al. 1982; Jones & Smith 1985). Haggerty (1983) and Haggerty et al. (1983) discussed K - B a - C r titanates from metasomatized peridotite xenoliths from S Africa; these phases are not unlike the priderite-jeppeite phases from lamproites. Jones et al. (1985) presented isotopic data for xenoliths from Kimberley, S Africa, which indicated that glimmerite xenoliths were disrupted pegmatitic segregations of Group I kimberlites (terminology of Smith (1983a, b)), whereas M A R I D and PKP xenoliths formed by magmatic-metasomatic processes involving both Group I and Group II components. However, trace-element and isotopic studies on metasomatized peridotite and M A R I D xenoliths from the Bultfontein kimberlite mine, Kimberley, S Africa (Kramers et al. 1983) have demonstrated that the S r - N d - P b isotopic features are consistent with mixing of a metasomatic fluid from a slightly depleted mantle source with mantle material showing a timeintegrated enrichment in incompatible elements. Nevertheless, the dominant (albeit extremely rare) amphibole in kimberlites or in the replacement zones in their included-mantle xenoliths is the K-rich richterite (slightly depleted in Ti relative to lamproite K-richterites) which is the characteristic (and nearly ubiquitous) amphibole in lamproites. The subtle compositional differ-
ences between K-richterites of lamproite, M A R I D xenoliths and metasomatized peridotite xenoliths can be explained by a variable buffering capacity of the host mantle peridotite which may modify the original amphibole composition. The presence of K-richterite, combined with the Tiand phlogopite-rich nature of M A R I D suite xenoliths, further suggests that lamproite-like fluids were the agents of patent metasomatism of kimberlite xenoliths; lamproite melts may be responsible for the precipitation of M A R I D xenoliths. However, much remains to be explored in the M A R I D and metasomatized xenoliths before the extent of this genetic relationship is fully realized. Interestingly, the nature of metasomatic phases (mica and amphibole) in lherzolite and harzburgite xenoliths from alkali basalts (higher A1, and kaersutite or pargasite rather than richterite (Boettcher & O'Neil 1980; Menzies & Murthy 1980b)) substantiates a distinct nature to the mantle source regions of MARIDsuite xenoliths and metasomatized alkali basalt xenoliths.
Conclusions Lamproites are recognized as a coherent but variable and exotic petrographic and chemical group of K-rich igneous rocks which border on and share certain petrogenetic aspects with the alkali basalt, kimberlite, lamprophyre and other K-rich rock clans. Lamproites should not be included in the lamprophyre clan. Lamproites can be effectively distinguished from these other rock types on the basis of mineralogy, mineral chemistry and whole-rock major, trace-element and isotope chemistry. Lamproites are unparalleled by other rock types with respect to their compositions. Known lamproites occur in 21 suites on six continents. Lamproites occur closer to the margins of continents whereas kimberlites are most abundant nearer the craton cores; lamproites often intrude crust that overlies fossil Benioff zones. Lamproites are partial melts of a metasomatized (i.e. phlogopite-, apatite-bearing) but depleted (in Na, AI, Ca) source-mantle peridotite (harzburgite). A three-stage model (depletion-enrichment-melting), at least, is required to explain the evolution of the source regions of lamproite magmas. Lamproites differentiate by processes including crystal fractionation, fractional resorption and volatile-phase transfer, among others. ACKNOWLEDGMENTS: This research was supported by the Anaconda Minerals Company, a division of the Atlantic Richfield
Lamproites and K-rich igneous rocks Corporation, in conjunction with their diamond exploration programme in the U.S. and Kalimantan. I thank R. W. Knostman, I. Gemuts and L. G. Krol for their continued support. Preprints supplied by R. Mitchell, B. Scott-Smith, D. Colchester, J. V. Smith, P. H. Nixon, P. Gregory, D. Velde, R. Vollmer, C. Hawkesworth and M. Menzies are gratefully appreciated. These workers, as well as L. G. Krol, N. R. Baker, M. Skinner, D. Velde, S. Bachinski, P. Berendson, M. Bickford, K. Collerson, F. Albarede, K. Fraser and F. Dodge, participated in many fruitful discussions which served to improve the content of this report. Two days with D. and B. Velde in Ayron proved enlightening. Pat Bickford generously supplied unpublished mineral chemical data on the Hills Pond rocks, Peter Nixon provided samples of the Murcia-Almeria lamproites, Pieter Berendsen provided samples of the Hills Pond lamproites and B. Scott-Smith supplied samples of the Holsteinsborg lamproites. I thank R. Baker, D. Dunn, J. Crann and W. Turner for logistical support and assistance in the field in the U.S.A., Indonesia and Australia. S. Self greatly assisted in the interpretation of the piperno and other pyroclastic rock textures. L. Noodles and Maggie clarified many confusing concepts. An exceedingly constructive review by Sharon Bachinski was invaluable; D. Velde, L. Krol and J. G. Fitton additionally provided useful criticisms on an earlier draft. I thank J. G. Fitton and B. G. J. Upton for only requiring a 25% reduction in the length of the original manuscript. I thank G. W. DeArmond, G. McEntire, D. Schraeder, L. Auvermann and I. Martinez for performing many arduous literature searches and for obtaining many of the obscure references cited in the text. I thank J. Toney and D. J. Henry for much invaluable help in the microprobe work. The rock and mineral chemistry data bases were compiled with the able assistance of V. Mount, S. R. Yang, F. Stiffand E. Kinsel to whom I express sincere thanks. The manuscript was typed by S. Epperson and the graphics were professionally produced by N. Murray.
PRA
SB
(ALL) (CSN) (DSV) (FOR) (GAL) (HM)
(KNX) (NHB)
(SEW) (SFC) (TB) (CHIN)
171
Bolivar 1977; Gogineni et al. 1978; Williams 1891; Scott-Smith & Skinner 1984a, b; Bergman, unpublished data Matson 1960; Velde 1975; Bergman, unpublished data Hawkins 1976 Van Kooten 1980 Dodge & Moore 1981 Ross 1926b Allen et al. 1975 Weed & Pirsson 1896; Pirsson 1905; Osborne & Roberts 1931 ; Wolff 1938; Hurlbut & Griggs 1939; Buie 1941; Tappe 1966; Witkind 1969, 1973; Woods 1975; Bergman, unpublished data Bastin 1906 Williams 1936; Lewis 1973; Schmitt et al. 1974; Rogers et al. 1982 Miller 1972 Templeman-Kluitt 1969 Cross 1906 Arculus & Smith 1979; Schulze & Helmstaedt 1979
Australia WKB
ARG (NSW)
Wade & Proder 1940; Prider 1960, 1982; Atkinson et al. 1984a; Jaques et al. 1984b; Nixon etal. 1984; Bergman, unpublished data Atkinson et al. 1984a Cundari 1973
Europe Osann 1906; Washington 1917; Parga Pondal 1935; Fuster & Pedro 1953; Borley 1967; Fuster et al. 1967; Nixon etal. 1984; Venturelli etal. 1984a Dal Piaz et al. 1979; Venturelli et al. NWI 1984b Hall 1982 PEN Tidmarsh 1932; Knill 1969 HLM Velde 1967 COR (TUSC) Gallo 1984 (SUNN) Furnes et al. 1982 N6mec 1972 (BOH) Duda & Schminke 1978 (LAC) MAP
APPENDIX 1 Sources of whole-rock major and trace-element analytical data in Tables 4-7 and associated figs 20-22
N America EVD HOL HP KAM LH
Libby 1975 Scott 1979 Franks et al. 1971 ; Merrill et al. 1977 Best et al. 1968 ; Bergman, unpublished data Carmichael 1967; Kuehner 1980, Kuehner et al. 1981; Schultz & Cross 1912; Yagi & Matsumoto 1966; Bergman, unpublished data
Africa BOB PPS
Bardet 1973 (Swartruggens only) Skinner & Scott 1979; Bergman, unpublished data
172 (TAN) (AZZ)
S. C.
Bergman
Holmes & Harwood 1937; Holmes 1950; Higazy 1954; Sahama 1974; Mitchell & Bell 1976 Vila et al. 1974
Antarctica
GSB Sheraton & Cundari 1980 MBAY, PP Sheraton & England 1980 Asia and Indonesia
COC Lacroix 1933a, b CHE Bergman & Baker 1984 GDW S a r k a r e t a l . 1980; G u p t a e t a l . 1983 (BAL) Kostyuk 1983 (KAJ) Brouwer 1909 Abbreviations are given in Tables 2 and 3 and in the text. Average compositions
The major- and trace-element data for lamproite sensu stricto are cited above; the major-element data for lamprophyres, kimberlites and alkali basalts were compiled on a computerized data bank that is a combination of the Mutchler et al. (1973) PETROS data bank and a literature compilation by Bergman (File DIROC). Individual references can be obtained by writing to Bergman. Trace-element data for kimberlites and alkali basalts are taken from Wedepohl & Muramatsu (1979) and data for lamprophyres are taken from Rock (1984, 1986). The estimated composition of the primitive mantle comes from Taylor & McLennan (1981) or Mason (1979).
leucite, A analcime, J priderite and jeppeite, A amphibole, I ilmenite, G garnet (suite abbreviations are as given in appendix 3). ARG
Atkinson et al. 1984a (G, H, C); Scott-Smith & Skinner 1984b (P) GSB Sheraton & Cundari 1980 (I, S, H, O, P, C) HOL Scott 1981 (A, S, O, P, C) HP Merrill et al. 1977 (A, P, C); P. Bickford, personal communication 1984 (A, C, P, S); Mitchell 1985 (P, S) LH Carmichael 1967 (A, S, O, P, C, J, F, L); Barton 1979 (L, F, C, P); Kuehner 1980 (A, S, O, P, C); Barton & van Bergen 1981 (H, O, P, C); Kuehner et al. 1981 (A, S, O, P, C); Mitchell 1985 (S, J) MAP Borley 1967 (P, C); Carmichael 1967 (P, F, O, C, A); Fustor et al. 1967 (P); Lopez Ruiz & Rodriguez Badiola 1980 (A, O, P, C); Venturelli et al. 1984a (A, H, S, P, P, C); Mitchell 1985 (P, F, I) MBAY, PP Sheraton & England 1980 (A, P, I,
F) PEN PRA
SB
WKB
APPENDIX 2 Sources of mineral chemical data of Table 10 and Figs 23-25
The mineral chemical data used in these figures and table have been compiled in a computerized data file DIMIC that comprises over 137,000 analyses of phases from kimberlites, lamprophyres, lamproites, alkali basalts and their included xenoliths and megacrysts (S. C. Bergman, unpublished). Since it is unreasonable to tabulate all the data sources here, a copy of the sources can be obtained by writing to Bergman. Only sources containing lamproite mineral chemical data will be summarized. The abbreviations are as follows: P phlogopite, O olivine, C clinopyroxene, H orthopyroxene, S spinel, F sanidine, L
Hall 1982 (A, P, F) Lewis et al. 1976 (G, S, O, P, C); Gogineni et al. 1978 (A, G, S, O, P, C); Mitchell & Lewis 1983 (A, C, P, J); Scott-Smith & Skinner 1984a, c (A, S, O, P, C); Mitchell 1985 (J) Velde 1975 (A, S, P, C); Mitchell 1985 (P, O, N, F, S); Henry & Bergman, unpublished data (A, P, C,O, S) Carmichael 1967 (C, P, A, F, L, J); Mitchell 1981 (P); Mitchell & Lewis 1983 (A, P, C): Atkinson et al. 1984a (G, H, S, C); Jaques et al. 1984a, b (A, G, H, S, O, C); Pryce et al. 1984 (J); Scott-Smith & Skinner 1984b (A, P, S); Mitchell 1985 (P, A, C, L, S,J)
APPENDIX 3 Alphabetical listing of rock suites discussed in the paper
ARG (AZZ) (BAL) BOB (BOH) CHE
Argyle, W Australia Azzaba, Algeria, Africa Baikal Rift-Aldan Shield, U.S.S.R. Bobi and Seguela, Ivory Coast, Africa Bohemian Massif, Czechoslovakia Chelima, Andhra Pradesh, India
Lamproites and K-rich igneous rocks (CHN) (CJ)
(coc) (COL) COR (CSN) DSV EVD (FOR) (GDW) GSB HLM (HM)
HOL HP KAM (KNX) (LAC) LH LUA MAP MBAY (NHB)
(NSW) NWI
Channel Islands, U.K. Campos de Jordao, Brazil Coc Pia, N Vietnam Colima Graben, Mexico Sisco, Corsica, France Central Sierra Nevada, California Deep Spring Valley, California Enoree Vermiculite, S Carolina Fortification Dyke, Colorado Gondwana Coalfields, India Gaussberg, Antarctica Holmeade Farm, U.K. Highwood Mountains, Montana Holsteinsborg, W Greenland Hills Pond, Kansas Kamas, U t a h Knox County, Maine Laacher See Province, Germany Leucite Hills, Wyoming Luangwa Graben, Zambia Murcia-Almeria province, Spain Mount Bayliss, Antarctica N a v a j o - H o p i Buttes, Arizona Lake Cargelligo Area, NSW, Australia N W Italy
PAM PEN PIS PD PP PPS PRA (RP) (SAB) (SAC) SB (SEW) (SRE) (SUNN) (TAN) (TB) (TUSC) WKB
173
Pamir, U.S.S.R. Pendennis, U.K. Orciatico, Pisa, Italy Piedade, Brazil Priestly Peak, Antarctica Postmasburg, Pneil etc. S Africa Prairie Creek, Arkansas Roman Province, Italy Santo Antonia da Barra, Brazil Sacramento, Brazil Smoky Butte, Montana Seward Peninsula, Alaska Srednogorie, Bulgaria Sunnfjord, Norway Toro-Ankole, Birunga, U g a n d a Two Buttes, Colorado Tuscany, Italy W Kimberley, Fitzroy Basin, W Australia
Those suites in parentheses are not lamproites but merely K-rich rocks; those not in parentheses conform to the present definition of lamproite sensu stricto
References ALLEN, D. G., PANTELEYEV,A. & ARMSTRONG,A. T. 1975. Porphyry copper deposits of the Alkalic Suite. Can. Inst. Min. Spec. 15, 402-14. ANDERSON, O. L., BREED, W. J., BREED, C. S., CHIDESTER, A. H., MCGETCHIN, T. R. & SHOEMAKER, E. M. 1970. Field guide to kimberlitebearing diatremes of northern Arizona-southern Utah including stops at Hopi Buttes, Monument Valley and Grand Canyon. Int. Syrup. on Mechanical Properties and Processes of the Mantle, June 2428, 1970, Flagstaff', Arizona, 52 pp. ANDREWS, J. R. & EMELEUS, C. 1971. Preliminary account of kimberlite intrusions from the Frederikshab district SW Greenland. Geol. Surv. Greenland Rep. 31, 26 pp. AOKI, K. I. & KURASAWA,H. 1984. Sr-isotope study of the tephrite series from Nyamuragira volcano, Zaire. Geochem. J. 18, 95-100. APPARADHANULU, K. 1962. Minette and riebeckitekalisyenite dykes in some Upper Cuddapah rocks, Kurnool district, Andhra Pradesh. Rec. geol. Surv. India, 94(2), 303-4. APPLETON, J. D. 1972. Petrogenesis of potassium-rich lavas from the Roccamonfina volcano, Roman Region, Italy. J. Petrol. 13, 425-56. ARCULUS,R. J. & SMITH,D. 1979. Eclogite, pyroxenite and amphibolite inclusions in the Sullivan Buttes latite, Chino Alley, Yavapai County, Arizona. In: BOYD F. R. & MEYER, H. O. A. (eds) The Mantle Sample: Inclusions in Kimberlites and Other Volcanics, Proc. 2nd Int. Kimberlite Conf. Vol. 2. pp. 309-
17. American Geophysical Union, Washington, DC. ARIMA, M. & EDGAR, A. D. 1980. Stability of wadeite (Zr2KzSi6018) under upper mantle conditions: petrological implications. Contrib. Mineral. Petrol. 72, 191-5. & 1981. Substitution mechanisms and solubility of titanium in phlogopites from rocks of probable mantle origin. Contrib. Mineral. Petrol. 77, 288-95. -& - - 1983. High pressure experimental studies of a katungite and their bearing on the genesis of some potassium-rich magmas of the West Branch of the African Rift. J. Petrol. 24, 166-87. ATKINSON, W. J. 1982. Diamond discoveries in the Kimberley, Western Australia. Trans. Inst. Min. Metall. 91A, 135-7. - - , HUGHES,F. E. & SMITH,C. B. 1982. A review of the kimberlitic rocks of Western Australia. Terra Cognita, 2, 204. & -1983. A review of the kimberlitic rocks of Western Australia. Proc 6th Australian Geology ConS, pp. 284-5. Geological Society of Australia, Sydney. & -1984a. A review of the kimberlitic rocks of western Australia. In: KORNPROaST, J. (ed.) Kimberlites I. Kimberlites and Related Rocks, pp. 195-225. Elsevier, Amsterdam. - - , SMITH,C. B. & BOXER,G. L. 1984b. The discovery and evaluation of the Ellendale and Argyle Lamproite diamond deposits, Kimberley region,
x74
S. C. Bergman
Western Australia. Soc. Min. Eng. A I M E Preprint 84-384. BACHINSKI, S. W. & SCOTT, R. B. 1979. Rare-earth and other trace element contents and the origin of minettes (mica-lamprophyres). Geochim. cosmochim. Acta, 43, 93-100. & SIMVSON, E. L. 1984. Ti-phlogopites of the Shaw's Cove minette: a comparison with micas of other lamprophyres, potassic rocks kimberlites and mantle xenoliths. Am. Mineral. 69, 41-56. BAGSHAW, A. N., DORAN, B. H., WHITE, A. H. & WILLIS, m. C. 1977. Crystal structure of a natural K-bearing hexatitanate isostructural with K2Ti6OI 3. Aust. J. Chem. 30, 1195-200. BAILEY, D. K. 1974. Continental rifting and alkaline magmatism. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp. 148-59. Wiley, New York. 1982. Mantle metasomatism--continuing chemical change within the Earth. Nature, Lond. 296, 525-30. BALASEBRAHMANYAN,M. N., MURTY, M. K., PAUL, D. K. & SARKAR, 1978. Potassium-argon ages of Indian kimberlites. J. Geol. Soc. India, 19(12), 5845. BANERJEE, S. 1953. Petrology of the lamprophyres and associated rocks of Raniganj coal field. Indian Mineral. J. 1, 9-29. BARBERI, F., BIZOUARD, H., CAPALDI, G., EERRARA, G., GASPARINI, P., INNOCENTI, F., JORON, J. L., LAMBRET, B., TREUIL, M. & ALLEGRE, C. 1978. Age and nature of basalts from the Tyrrhenian Abyssal Plain. Initial Rep. Deep Sea Drilling Project, 42 (Part 1), 509-14. -& INNOCENTI, F. 1967. Le rocce Selagitiche di arciatico e Montecatini in Val di Cecina. Atti Soc. Tosc. Pisa Sci. Nat. Mem. P. V., Ser. A, 74, 139-80. , FERRARA, G., KELLER, J. & VILLARI, U 974. Evolution of Aeolian arc volcanism (southern Tyrrhenian Sea). Earth planet. Sci. Lett. 21, 26976. --, SANTACROCE, R. & VARET, J. 1982. Chemical aspects of rift magmatism. In : PALMASONG. (ed.) Continental and Oceanic Rifts, pp. 223-58. American Geophysical Union, Washington, DC. BARBERI, M., PENTA, A. & TURI, B. 1975. O and Sr isotope ratios in some ejecta from the Alban Hills volcanic area, Roman comagmatic region. Contrib. Mineral. Petrol. 51, 127-33. BARDET, M. G. 1973. Geologie du Diamant, Vol. 1. Editions BRGM no. 83, Paris. 1974. Geologie du Diamant, Vol. 2. Editions BRGM no. 83, Paris. 1977. Geologie du Diamant, Vol. 3. Editions BRGM no. 83, Paris. & VACHETTE,M. 1966. Determination of the ages of kimberlites of West Africa. 3rd Symp. on African Geology, Rep. GCLO 88, pp. 15-17. Commonwealth Geological Liaison Office. BAREN, F. A. VAN, 1948. On the petrology of the volcanic area of the Geonoeng Moeria (Java). Meded. Alg. Proefstn. Landb. Bultenzorg, Java, 60. BARTON, M. 1976. Melting relations of some ultrapotassic volcanic rocks. Prog. Exp. Petrol., NERC Publ. Ser. D, Vol. 6, pp. 91-4. -
-
-
-
-
-
1979. A comparative study of some minerals occurring in the potassium-rich alkaline rocks of the Leucite Hills, Wyoming, the Vico Volcano, western Italy, and the Toro-Ankole Region, Uganda. Neues Jb. Mineral. Abh., 137(2), 113-34. & VANBERGEN, M. J. 1981. Green clinopyroxenes and associated phases in a potassium-rich lava from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 77, 101-14. & HAMILTON,D. L. 1978. Water-saturated melting relations to 5 kilobars of three Leucite Hills lavas. Contrib. Mineral. Petrol. 66, 41-9. & -1979. The melting relationships of a madupite from the Leucite Hills, Wyoming, to 30 kb. Contrib. Mineral. Petrol. 69, 133-42. &-1982. Water saturated melting experiments bearing upon the origin of potassium-rich magmas. Mineral. Mag. 45, 267-78. & WOOD, B. J. 1976. Equilibration of orendite with lherzolitic mineral assemblages: a combined experimental/theoretical approach. Prog. Exp. Petrol., NERC Publ., Ser. D, Vol. 6, pp. 90-1. BASTIN, E. S. 1906. Some unusual rocks from Maine, J. Geol. 14, 173-87. BASE, m. R. & TATSUMOTO, M. 1980. Nd-isotopes in selected mantle-derived rocks and minerals and their implications for mantle evolution. Contrib. Mineral. Petrol. 75, 43-54. BEGER, P. J. 1923. Der chemismus der lamprophyre. In: NIGGLI, P. (ed.) Gesteins-undMineralProvinzen, Vol. 1. pp. 448-51. Verlag Gebruder Borntraeger, Berlin. BELL, K. & POWELL, J. L. 1969. Strontium isotopic studies of alkalic rocks: the potassium-rich lavas of the Birunga and Toro-Ankole regions, east and central equatorial Africa. J. Petrol. 10, 536-72. & -1970. Strontium isotopic studies of alkalic rocks: the alkalic complexes of eastern Uganda. Geol. Soc. Am. Bull. 81, 3481-90. BELLON, H. 1981. Chronologique radiometrique (KAr) des manifestations magmatiques autour de la Mediterran6e occidentale entre 33 Ma et 1 Ma Proc. Int. Conf. Urbino University, Italy, pp. 34160. -& LETOUSEY, J. 1977. Volcanism related to plate tectonics in the Western and Eastern Mediterranean. In: BYE DUVAL B. & MONTADERTL. (eds) Int. Syrup. on Mediterranean Basins, pp. 165-84. Editions Techniq, Paris. BERENDSEN, P., CULLERS, R. L., MANSKER, W. L. & COLE, G. P. 1985. Late Cretaceous kimberlite and lamproite occurrences in E. Kansas, U.S.A. Geol. Soc. Am. Abstr. Progm. 17, 151. BERGMAN, S. C. & BAKER, N. R. 1984. A new look at the Proterozoic dikes from Chelima, Andhra Pradesh, India: diamondiferous lamproites? Geol. Soc. Am. Abstr. Progm. 16(6), 444. & -The Chelima dikes, Andhra Pradesh, India: diamondiferous lamproites? Submitted to Nature, Lond. - - , DUNN, D. P. & KROL, L. G. 1985. Petrology and geochemistry of the Linhaisai Minette, Karamu River, central Kalimantan. Abstr. Geol. Assoc. Can. Progm. 10, A4. -
-
-
-
-
-
-
-
-
-
Lamproites and K-rich igneous rocks BEST, M. G. 1975. Migration of hydrous fluids in the upper mantle and potassium variation in calcalkaline rocks. Geology, 3, 429-32. --, HENAGE, L. F. & ADAMS, J. A. S. 1968. Mica peridotite, wyomingite, and associated potassic igneous rocks in northeastern Utah. Am. Mineral. 53, 1041-8.
BESWICK,A. E. 1976. K and Rb relations in basalts and other mantle derived materials. Is phlogopite the key? Geochim. cosmochim. Acta, 32, 1167-84. BICKFORD, M. E., MOSE, D. E., WETHERILL,G. W. & FRANKS, P. C. 1971. Metamorphism of Precambrian granitic xenoliths in a mica periodite at Rose Dome, Woodson County, Kansas. Part 1, Rb-Sr isotopic studies. Geol. Soc. Am. Bull. 82, 2863-8. BLAXLAND, A. B. 1976a. Geochronology and isotope geochemistry of the Gardar Alkaline complex, South Greenland. PhD Thesis, University of Edinburgh (unpublished). -1976b. Rb-Sr isotopic evidence for the age and origin of the Ivigtut granite body and associated cryolite body, South Greenland. Econ. Geol. 71, 864-9. BOCCALETTI, M., MANETTI, P., PECCERILLO, A. & STANISHEVA-VASSILEVA,G. 1978. Late Cretaceous high potassium volcanism in E. Srednogorie, Bulgaria. Geol. Soc. Am. Bull. 89, 439-47. BOETTCHER, A . L . , MYSEN, B . O . & MODRESKI, P. J.
1975. Melting in the mantle: phase relationships in natural and synthetic peridotite-H20 and peridotite-HzO-CO2 systems at high pressures. Phys. Chem. Earth, 9, 855-68. - & O'NEIL, J. R. 1980. Stable isotope, chemical and petrographic studies of high-P amphiboles and micas: evidence for metasomatism in the mantle source regions of alkali basalts and kimberlites. Am. J. Sci. 280A, 594-621. BOLIVAR,S. L. 1977. Geochemistry of the Prairie Creek, Arkansas and Elliott County, Kentucky intrusions. PhD Thesis, University of New Mexico (unpublished). - 1982. The Prairie Creek kimberlite, Arkansas. In: MCFARLAND, J. D., III (ed.) Contributions to the Geology of Arkansas, State of Arkansas Geological Commission Misc. Publ. 18, pp. 1-21. - 1984. An overview of the Prairie Creek intrusion, Arkansas. Soc. Min. Eng. A I M E Fall Meet., October 1984, Preprint 84-346. - & BROOKINS,D. G. 1979. Geophysical and Rb-Sr study of the Prairie Creek, AR kimberlite. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds." their Geology, Petrology and Geochemistry, Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 289-99. American Geophysical Union, Washington, DC. BORLEY, G. D. 1967. Potassium-rich volcanic rocks from Southern Spain. Mineral. Mag. 36, 364--79. BORSI, S., FERRARA, G. & TONGIORGI, E. 1967. Determinazioni con il metoda K/Ar delle rocce magmatiche della Toscana. Boll. Soc Geol. Italy, 86, 403-10. BOUTWELL,J. M. 1912. Geology and ore deposits of the Park City District, Utah. U. S. Geol. Surv. Prof Pap. 77, 231.
175
BOWEN, C. F. 1915. Possibilities ofoil in the Porcupine Dome, Rosebud County, Montana. U. S. Geol. Surv. Bull. 621, 61-70. BOWEN, N. L. 1928. Evolution of the Igneous Rocks. Princeton University Press, Princeton, NJ. BRADLEY, W. H. 1964. Geology of Green River Formation and associated Eocene rocks in SW Wyoming and adjacent parts of Colorado and Utah. U.S. Geol. Surv. Prof. Pap. 496A, A1-A86. BRANNER,J. C. & BRACKETT,R. N. 1889. The peridotite of Pike County, Arkansas. Am. J. Sci. 38, 50-9. BRAVO, M. S. & O'HARA, M. J. 1975. Partial melting of phlogopite-bearing synthetic and garnet-lherzolites. Phys. Chem. Earth, 9, 845-54. BRIQUEU, L., BOUGAULT, H. & JORON, J. L. 1984. Quantification of Nb, Ta, Ti and V anomalies in magmas associated with subduction zones: petrogenetic implications. Earth planet. Sci. Lett. 68, 297-308. BROOKINS, D. G., DELLA VALLE, R. S. & BOLIVAR, S. L. 1979. Significance of uranium abundance in United States kimberlites: In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes, and Diamonds: Their Geology, Petrology and Geochemistry: Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 280-8. American Geophysical Union, Washington, DC. BROOKS, C. K., FAWCETT, J. J., GITTINS, J. G. & RUCKLIDGE, J. C. 1981. The Batbjerg complex, east Greenland: a unique ultrapotassic Caledonian intrusion. Can. J. Earth Sci. 18, 274-85. , NOE-NYGGARD,A., REX, D. C. & ROENSBO,J. C. 1978. An occurrence of ultrapotassic dikes in the neighborhood of Holsteinsborg, West Greenland. Bull. geol. Soc. Denmark, 27, 1-8. BROUWER, H. A. 1909. Glimmerleucitbazalt van OostBorneo. Versl. Kon. Akad. Wetensch. Amsterdam wis. Nat. Afd. 18, 85-91. BUCKING, H. 1899. Ber. Naturforsch. Ges. Freiburg, i ( 1 1 ) , 78. Bum, B. F. 1941. Igneous rocks of the Highwood Mountains, Montana, III, Dike and related intrusives. Geol. Soc. Am. Bull. 52, 1754-807. - & STEWART,O. F. 1954. Origin of vermiculite at Tigerville, S. Carolina. Geol. Soc. Am. Bull. 65, 1356-7. BUREAUOF MINERALRESOURCES,GEOLOGYAND GEOPHYSICS, 1976. Geological Map of Australia (1/2,500,000 scale, 4 sheets), Canberra, A.C.T. CARABALLO, J. M. 1975. Geoquimica de las rocas lamproiticas espanolas. PhD Thesis, University of Madrid (unpublished). CARMICHAEL, I. S. E. 1967. The mineralogy and petrology of the volcanic rocks from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 15, 2466. , NICHOLLS, J., SPERA, F. J., WOOD, B. J. & NELSON, S. A. 1977. High temperature properties of silicate liquids: application to the equilibration and ascent of basic magma. Phil. Trans. R. Soc. Lond. Ser. A, 286, 373-431. - - , TURNER, F. S. & VERHOOGEN, J. 1974. Igneous Petrology, McGraw-Hill, New York. CARR, S. G. & OLLIVER,J. G. 1980. Olivine basalt dyke
I76
S. C. Bergman
near Terowie. Qt. Geol. Notes Geol. Surv. S. Aust. 73, 2-7. CARRARO, F. & FERRAR, G. 1968. Alpine 'Tonalite' at Miagliano, Biella (zona-diorite-Kinzigitica). Schweiz Mineral. Petrogr. Mitt. 48, 75-8. CHATTERJEE, S. C. 1974. Petrography of the Igneous and Metamorphic Rocks of India, 391 pp. Macmillan, Bombay. CI-IAYES, F. 1976. Independent exploration of the database and reduction repertoire of the petrographic information system. Carnegie Instn. Wash. Yb. 75, 791. CIVETTA,L., INNOCENTI, F., MANETTI,P., PECCERILLO, A. &POLI, G. 1981. Geochemical characteristics of potassic volcanics from Mts. Ernici (southern Latium, Italy). Contrib. Mineral. Petrol. 78, 37-47. , ORSI, G. & SCANDONE, G. 1978. Eastwards migration of the Tuscan anatectic magmatism due to anticlockwise rotation of the Appenhines, Nature, Lond. 276, 604-6. CLARK, F. W. & W~HINGTON, H. S. 1924. The composition of the earth's crust. Prof. Pap. U.S.G.S. 127. CLEMENT, C. R., SKINNER, E. M. W. & SCOTT, B. H. 1977. Kimberlite redefined. Extended Abstr. 2nd Kimberlite Conf., Santa Fe, NM. COLCHESTER, D. M. 1982. Geology and petrology of some kimberlites near Terowie, S. Australia. M. Sc. Thesis, New South Wales Institute of Technology (unpublished). -1983. The geology and petrology of some kimberlites near Terowie, S. Australia, Prof. 6th Australian Geology Cony., pp. 282-3. Geological Society of Australia, Sydney. COLLERSON,K. D. & MCCULLOCH,M. T. 1982. Nd and Sr isotope geochemistry of leucite-bearing lavas from Gaussberg, Eastern Antarctica. J. geol. Soc. Aust. 30. fONDlE, K. C. 1976. Plate tectonics and crustal evolution. Pergamon Press, Inc., New York. CORNELISSEN, A. K. & VERWOERD, W. J. 1975. The Bushmanland kimberlites and related rocks. Phys. Chem. Earth, 9, 71-80. COSGROVE,M. E. 1972. Geochemistry of the potassiumrich Permian volcanic rocks of Devonshire, England. Contrib. Mineral. Petrol. 36, 155-70. Cox, K. G., HAWKESWORTH,C. J., O'NIONS, R. K. & APPLETON, J. D. 1976. Isotopic evidence for the derivation of some Roman region volcanics from anomalously enriched mantle. Contrib. Mineral. Petrol. 56, 173-80. CRAWFORD, A. R. 1969. India, Ceylon and Pakistan: new age data and comparisons with Australia. Nature, Lond. 223, 380-4. -& COMPSTON, W. 1970. Age of the Vindhyan Systems of Peninsular India. Q. J. Geol. Soc. Lond. 125, 351-71. & -1973. An age of the Cuddapah and Kurnool systems, southern India. J. geol. Soc. Aust. -
-
19,
4.
CRITTENDEN, M. D. & KIRSTLER, R. W. 1966. Isotopic dating of intrusive rocks in the Cottonwood area, Utah. Geol. Soc. Am. Spec. Pap. 101,298-9. CROSS, C. W. 1897. The igneous rocks of the Leucite
Hills and Pilot Butte, Wyoming. Am. J. Sci. 4, 1 1 5 - 4 1 .
1906. Prowersose (syenitic lamprophyre) from Two Buttes, Colorado. J. Geol. 14, 165-72. CULLERS, R. L., RAMAKRISHNAN,S., BERENDSEN, P. & GRIFFIN, T. 1985. Geochemistry and petrogenesis of lamproites, late Cretaceous age, Woodson County, Kansas, U.S.A. Geochim. cosmochim. Acta. 49, 1388-402. CUNDARI, A. 1973. Petrology of the leucite-bearing lavas in New South Wales. J. geol. Soc. Aust. 20, 465-92. & FERGUSON, A. K. 1982. Significance of the pyroxene chemistry from leucite-bearing and related assemblages. Tschermaks Mineral. Petrogr. Mitt. 30, 189-204. & O'HARA, M. J. 1976. Experimental study at atmospheric and high pressures of a mafic leucitite from New South Wales, Australia. Prog. Exp. Petrol, 3rd Rep. NERC Pub., Ser. D, 6, 260-1. --, RENARD, J. G. R. & GLEADOW, A. J. W. 1978. Uranium-potassium relationship and apatite fission track ages for a differentiated leucite suite from New South Wales (Australia). Chem. Geol. 22, 11-20. CURRIE, K. L., CURTIS, L. W. & GITTINS, J. 1975. Petrology of the Red Wine alkaline complexes Central Labrador and a comparison with the Ilimaussaq complex, SW Greenland. Geol. Surv. Can. Pap. 75(1A), 271-80. DAL PIAZ, G. V., HUNZIKER, J. C. & MARTINOTTI,G. 1973. Excursion to the Sesia zone of the Schw. mineral-petrograph gesellschaft, September 30th to October 3rd, 1973. Schweiz Mineral. Petrogr. Mitt. 53, 447-90. --, VENTURELLI, G. & SCOLARI, A. 1979. Calfalkaline to ultrapotassic postcollisional volcanic activity in the internal northwestern Alps. Mem. Sci. Geol. 32. DALY, R. A. 1910. Origin of alkaline rocks. Geol. Soc. Am. Bull. 21, 87-118. 1914. Igneous Rocks and their Origin, 1st edn. McGraw-Hill, New York. -1933. Igneous Rocks and the Depths of the Earth, 598 pp. Hafner, New York. DAWSON, J. B. 1967a. A review of the geology of kimberlite. In: WYLLIE, P. J. (ed.) Ultramafic and Related Rocks, pp. 241-51. Wiley, New York. 1967b. Geochemistry and origin of kimberlite. In : WYLLIE P. J. (ed.) Ultramafic and Related Rocks, pp. 269-78. Wiley, New York. 1970. The structural setting of African kimberlite magmatism. In: CLIFFORD, T. N. • GASS, I. G. (eds) African Magmatism and Tectonics, pp. 32135. Oliver & Boyd, Edinburgh. -1979. New aspects of diamond geology. In: FIELDS, J. E. (ed.) Diamonds, pp. 539-54. Academic Press, New York. 1980. Kimberlites and their Xenoliths, 252 pp. Springer New York. 1984. Contrasting types of upper mantle metasomatism? In: KORNPROBST, J. (ed.)Kimberlites H." The Mantle and Crust-Mantle Relationships pp. 289-94. Elsevier, Amsterdam.
-
-
-
-
-
-
-
-
-
-
Lamproites and K-rich igneous rocks 1987. The kimberlite clan: relationship to olivine and leucite lamproites, and inferences for uppermantle metasomatism. In: LITTON,J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 95-102. & SMITH, J. V. 1977. The MARID (micaamphibole-rutile-ilmenite-diopside) suite of xenoliths in kimberlite. Geochim. cosmochim. Acta, 41, 309-23. DE MARCO, L. 1959. Su alcuni filoni lamprofirici radioattivi del complesso Sesia-Lanzo. Stud. Rich. Div. Geomin. Cnen, 1, 1-30. DE ROEVER, W. P. 1975. Belts with scattered leucite occurrences indicate direction of dip of coeval subduction zones. Petrologie, 1, 37-42. DERRICK, G. M. & GELLATLY,D. C. 1972. New leucite lamproites from the west Kimberley, western Australia. Bur. Min. Resources, Aust., Bull. 125, 103-19. DODGE, F. C. W. & MOORE, J. G. 1981. Late Cenozoic volcanic rocks of the southern Sierra Nevada, California: II. Geochemistry. Geol. Soc. Am. Bull. 92, 1670-761. DOLFI, D., FARINATO,R., TRIGILA, R. & HAMILTON,D. L. 1976. Some experimental and thermodynamic evidences bearing on the origin of potassic volcanics from central Italy. Proc. Int. Congress Thermal Waters, Geothermal Energy and Volcanism of the Mediterranean Area, Vol. 3 Natl. Tech. Univ., Athens, Greece. pp. 43-59. DORT, W. 1972. Late Cenozoic volcanism in Antarctica. In: ADIE, R. J. (ed.) Antarctic Geol. Geophys., pp. 645-52. Universitetsforlaget, Oslo. DRYGALSKI, E. VON 1904. The German Antarctic Expedition. Geogr. J. 24, 129-52. DUBEAU, M. I. & EDGAR, A. D. 1985. The stability of priderite up to 30 kb: inferences on its potential as a reservoir for K, Ba and Ti in mantle sources for lamproitic magmas. Geol. Assoc. Can. Progm. Abstr. 10, A15. DUDA, A. & SCHMINKE' H. V. 1978. Quaternary basanites, melilite nephelinites and tephrites from the Laacher See area (Germany). Neues Jb. Mineral. Abh. 132, 1-33. EDGAR, A. D. 1979. Mineral chemistry and petrogenesis ofultrapotassic-ultramafic volcanic rocks. Contrib. Mineral. Petrol. 71, 171-5. - 1983. Relationship of ultrapotassic magmatism in the western U.S.A. to the Yellowstone plume. Neues Jb. Mineral. Abh. 147, 35-46. & ARIMA, M. 1981. Geochemistry of three potassium-rich lavas from the west branch of the African rift: inferences on their genesis. Neues Jb. Mineral. Monatsh. 539-52. - & -1983. Conditions of phlogopite crystallization in ultrapotassic volcanic rocks. Mineral. Mag. 47, 11-19. - & CONDLIFFE, E. 1978. Derivation of K-rich ultramafic magmas from a peridotitic mantle source. Nature, Lond. 174, 639-40. , , BARNETT,R. L. & SHIRRAN, R. J. 1979. An experimental study of an olivine ugandite magma and mechanisms for the formation of its Kenriched derivatives. J. Petrol. 21,475-97. - -
--,
177
GREEN, D. H. & HIBBERSON, W. O. 1976. Experimental petrology of a highly potassic magma. J. Petrol. 17, 339-56. EGGLER, D. H. 1977. The principle of the zone of invariant vapor composition: an example in the system CaO-MgO-SiO2-COz-H20 and implications for the mantle solidus. Carnegie Instn. Wash. Yb. 76, 428-33. ERLANK, A. J. 1970. Distribution of potassium in mafic and ultramafic nodules. Carnegie Instn. Wash. Yb. 68, 433. 1973. Kimberlite potassium richterite and the distribution of potassium in the upper mantle. Int. Conf. on Kimberlites, Cape Town, Extended Abstracts, pp. 103-6. Univ. Capetown Rondebosch. & RICHARD,R. S. 1977. Potassic richterite bearing peridotites from kimberlite and the evidence they provide for upper mantle metasomatism. Extended Abstr., 2nd Kimberlite Conf., Santa Fe, NM. - - - , ALLSOPP,H . L . , H A W K E S W O R T H , C . J . & MENZIES, M. A. 1982. Chemical and isotopic characterization of upper mantle metasomatism of peridotite nodules from the Bulfontein Kimberlite. Terra Cognita, 2, 261-3. ESCHER, A. 1970. Tectonic~Geological map of Greenland (1/2,500,000 scale), Geological Survey of Greenland, Copenhagen. - • WATTERSON,J. 1973. Kimbedites and associated rocks in the Holsteinsborg-Sondre Stromfjord Region, Central West Greenland. Gronland Geol. Unders. Rapp. 55. ESKOVA, E. M. & KAZAKOVA, M. E. 1955. Shcherbakovite--a new mineral. Am. Mineral. 40, 788. ESPERANCA, S. & HOLLOWAY, J. R. 1985. An experimental study of a mafic minette from Buell park, Arizona. Geol. Assoc. Can. Progm. Abstr. 10, A17. EWART, A. 1982. The mineralogy and petrology of Tertiary-Recent orogenic volcanic rocks: with special reference to the andesitic-basaltic compositional range. In: THORPE, R. S. (ed.) Andesites, pp. 25-95. Wiley, New York. - & LEMAITRE, R. W. 1980. Some regional compositional differences within Tertiary-Recent orogenic magmas. In: LEMAITRE, R. W. & CUNDARI, A. (eds) Chemical Characterization o f Tectonic Province. Chem. Geol. 30, 257-83. FALLOT, P. & JEREMINE, E. 1932. Remarque sur une vari6t6 nouvelle de jumillite et sur l'extension des laves de ce groupe. C. R. Congr. Natl. Soc. Savantes 1929, 1-13. FARQUHARSON,R. A. 1920. Petrologic work. Annu. Rep. geol. Surv. W. Aust. 1919, 42. -1922. Petrologic work. Annu. Rep Geol. Surv. W. Aust. 1921, 56. FAURE, G. 1977. Principles of Isotope Geology, 464 pp. Wiley, New York. -& POWELL, J. L. 1972. Strontium Isotope Geology, 188 pp. Springer, New York. FERAUD, J., FORNARI, M., GEFFROY, J. & LENCK, P. 1977. Min~ralisations arseni6es et ophiolites: le filon a realgar et stibine de Matra et sa place dans le district a Sb-As-Hg de la Corse alpine. Bull. B.R.G.M. 2, II, 91-112. FERGUSON, A. K. ~r CUNDARI, A. 1982. Feldspar
I78
S. C. Bergman
crystallization trends in leucite-bearing and related assemblages. Contrib. Mineral. Petrol. 18, 212-8. FERMOSO, M. L. 1967a. Composicion quimica de las sanidinas de las rocas lamproiticas espanolas. Estud. Geol. 23, 29-30. 1967b. E1 diopsido de las rocas volcanicas de Jumilla. Estud. Geol. 23, 31-3. FERNANDEZ, S . & HERNANDEZ-PACHECO, A. 1972. Lamproitic rocks of Cabezo Negro, Zeneta, Murcia. Estud. Geol. Inst. Invest. Geol. Lucas Mallada, 28, 267-76. FERRARA, G., LAURENZI, M. A., TAYLOR, H. P., TONARINI, S. & TURI, B. 1985. An oxygen and strontium isotope study of the Alban Hills volcanic complex, Italy. Earth planet. Sci. Left. 75, 1328. , MARTINEZ, M. P., TONARINI,S. & TURI, B. 1980. Oxygen and strontium isotopic correlation in volcanic rocks from Mr. Vulsini district (N. Latium, Italy). Int. Conf. on Geochronology, Cosmochronology and Isotopic Geochemistry, Extended Abstracts, pp. 95-6. U.S. Geol. Survey, Washington. FITZGERALD, W. V. 1907. Reports on portions of the Kimberleys (1905-6). W. Australia Parliamentary Paper 19. FODEN, J. D. 1983. The petrology of the calcalkaline lavas of Rindjani Volcano, East Sunda Arc; a model for island arc petrogenesis. J. Petrol. 24, 98130. FOLEY, S. F. 1982. Mineralogy, geochemistry, petrogenesis and structural relationships of the Aillik Bay alkaline intrusive suite, Labrador, Canada. M. Sc. Thesis, Memorial University Newfoundland (unpublished). - 1984. Liquid immiscibility and melt segregation in alkaline lamprophyres from Labrador. Lithos, 17, 127-37. - 1985. The oxidation state of lamproitic magmas. Tschermaks Mineral. Petrol. Mitt., 34, 217-38. , TAYLOR, W. R. & GREEN, D. H. 1986a. The role of fluorine and oxygen fugacity in the genesis of the ultrapotassic rocks. Contrib. Mineral. Petrol., 94, 183-92. , TAYLOR,W. R. & GREEN, D. H. 1986b. The effect of fluorine on phase relationships in the system KA1SiO4-Mg2SiO4-SiO2 at 28 kbar and the solution mechanism of fluorine in silicate melts. Contrib. Mineral. Petrol. 93, 46-55. , VENTURELLI,G., GREEN, D. H. & TOSCANI, L. in review. The ultrapotassic rocks: characteristics, classification and constraints for petrogenetic models. Fox, C. S. 1930. The Jharia Coalfield. Mem. Geol. Surv. India 56, 255. FRANKS, P. C. 1959. Pectolite in mica peridotite, Woodson County, Kansas. Am. Mineral. 44, 10826. 1966. Ozark Precambrian-Paleozoic relations: discussion of igneous rocks exposed in E. Kansas. Bull. Am. Assoc. Petrol. Geol. 50, 1035-42. --, BICKFORD, M. E. & WAGNER, H. C. 1971. Metamorphism of Precambrian granitic xenoliths in a mica peridotite at Rose Dome, Woodson
County Kansas. Part 2, Petrologic and mineralogic studies. Geol. Soc. Am. Bull. 82, 2869-90. FRANTSESSON,E. V. 1968. The Petrology of the Kimberlites, 194 pp. (Translated by D. A. Brown, Australian National University, Canberra, 1970.) FRASER, K. J., HAWKESWORTH, C. J., ERLANK, A. J., MITCHELL, R. H. & SCOTT-SMITH,B. H. 1985. Sr, Nd and Pb isotope and minor element geochemistry of lamproites and kimberlites. Earth Planet. Sci. Lett. 76, 57-70. FRIEND, C. R. L. & JANARDHAN, A. S. 1984. Mineral chemistry of coexisting phases from shonkinitic rocks, Salem, Tamil Nadu, India, Mineral. Mag. 48, 181-93. FROMAGET,J. 1933. Sur la presence de roches intrusives alkalines dans la zone charri6e neotriasique des plateaux calcaires du Tonkin occidental. C. R. Acad. Sci. Paris, 197, 651-3. FURNES, H., MITCHELL, J. G., ROBINS, B., RYAN, P. & SKJERLIE, F. J. 1982. Petrography and geochemistry of peralkaline, ultrapotassic syenite dykes of ' Middle Permian age, Sunnfjord, W. Norway. Norsk. Geol. Tidsskr. 62, 147-59. FUSTER, J. M. 1956. Las erupciones delleniticas del Torciario superior de la Fosa de Vera (provincia de Almeria). Estud. Geol. 20, 299-314. - & DE PEDRO, F. 1953. Estudio petrologico de las rocas volcanicas lamproiticas de Cabezo Maria (Almeria). Estud. Geol. Inst. Lucas Mallada Invest. Geol. Publ. 20, pp. 477-508. - & GASTESl, P. 1964. Estudio petrologico de las rocas lamproiticas de Barqueros (provincia de Murcia). Estud. Geol. 20, 299-331. , SAGREDO, J. & FERMOSO, M. L. 1967. Las rocas lamproiticas del S.E. de Espana. Estud. Geol. 23, 35-69. , IBARROLA, E. & LOBATO, M. P. 1954. Analysis quimacos de rocas espanolas publicados hosta 1952. C.S.LC. Monogr. Inst. Lucas Mallada Invest. Geol. 14. GALLAGHER, M. J. 1963. Lamprophyre dykes from Argyll. Mineral. Mag. 33, 415-30. GALLO, F. 1984. The Kamafugitic rocks of San Venanzo and Cuppaello, Central Italy, Neues .lb. Mineral. Monatsh. 115, 198-210. GARLICK, G. D. 1966. Oxygen isotope fractionation in igneous rocks. Earth planet. Sci. Lett. 1,361-8. GARLICK, H. J. 1979. Australian diamond prospects. The story so far. Indust. Miner. (Feb.) 137, 1729. GATES, O. 1959. Breccia pipes in the Shoshone ranges, Nevada. Econ. Geol. 54, 790-815. GEOLOGICAL SURVEY OF ITALY, 1961. Carta Geologico D'Italia (1/1,000,000 scale, 2 sheets), Rome. GEST, D. E. & MCBIRNEY, A. R. 1977. Genetic relations of shoshonitic and absarokitic magmas, Absaroka Mountains, Wyoming. J. Volcanol. geotherm. Res. 6, 85-104. GHOSE, C. 1949. A petrochemical study oflamprophyres and associated intrusive rocks of the Jharia coalfield. Quart. J. Geol. Min. Metall. Soc. India, 21, 133-47. GIARDINI, A. A. & MELTON, C. E. 1975a. The nature of
Lamproites and K-rich igneous rocks cloud-like inclusions in two Arkansas diamonds. Am. Mineral. 60, 931-3. -& -1975b. Chemical data on a colorless Arkansas diamond and its black amorphous CFe-Ni-S inclusions. Am. Mineral. 60, 934-6. & -1975c. Gases released from natural and synthetic diamonds by crushing under high vacuum at 200~ and their significance to diamond genesis. Fortschr. Mineral. 52, 455-64. , HURST, V. J., MELTON, C. E. & STORMER, J. C. 1974. Biotite as a primary inclusion in diamond: its nature and significance, Am. Mineral. 59, 7839. GILBERT, G. K. 1896. Laccolites in S. E. Colorado. J. Geol. 4, 816-25. GILL, J. B. 1981. Orogenic andesites and plate tectonics. Springer-Verlag, New York. GITTINS, J. 1979. The feldspathoidal alkaline rocks. In: YODER, H. S., Jr. (ed.) The Evolution of the Igneous Rocks, Fiftieth Anniversary Perspectives, pp. 35190. Princeton University Press, Princeton, NJ. , FAWCETT, J. J., BROOKS, C. K. & RUCKLIDGE, J. C. 1980. Intergrowths of nepheline-K-feldspar and kalsilite-K-feldspar: a re-examination of the pseudo-leucite problem. Contrib. Mineral. Petrol. 73, 119-26. , MACINTYRE, R. M. & YORK, D. 1967. The ages of carbonatite complexes in eastern Canada. Can. J: Earth Sci. 4, 651-5. GLAZNER, A. F. 1979. Geochemistry of ultrapotassic metasomatism in the Sleeping Beauty volcanic area, central Mojave Desert, California. Geol. Soc. Am. Abstr. Progm. 11, 79-80. & STORK, A. L. 1976. Ultrapotassic, silicarich Tertiary volcanic necks from the Mojave Desert, California. Geol. Soc. Am. Abstr. Progm. 8, 886. GOGINENI, S. V., MELTON, C. E. & GIARDINI, A. A. 1978. Some petrological aspects of the Prairie Creek diamond-bearing kimberlite diatreme, Arkansas. Contrib. Mineral. Petrol. 66, 251-61. GOLDSCHMIDT, V. M. 1933. Grundlagen der quantitativen Geochemie. Fortschr. Mineral. Krist. Petrog. 112-56. GOLES, G. G. 1967. Trace elements in ultramafic rocks. In: WYLLIE, P. J. (ed.) Ultramafic and related rocks, J. Wiley and Sons, New York, 352-62. GRAPES, R. H. 1975. Petrology of the Blue Mountain Complex, Marlborough, N . Z . J . Petrol. 16, 371428. GREGORY, G. P. 1969. Geochemical dispersion patterns related to kimberlite intrusives in N. America. PhD Thesis, Royal School of Mines, Imperial College, London. (Unpublished) - - & TOOMS,J. S. 1969. Geochemical prospecting for kimberlite. Q. J. Colorado School Mines, 64, 265305. GROZDANOV, L. 1979. Chemistry and genesis of biotite from lamproites in the vicinity of the village of Svidnya, district of Sofia (Bulgaria), Dokl. Bolg. Akad. Nauk, 32, 341-4. GRUBB, P. L. C. 1965. Undersaturated potassic lavas and hypabyssal intrusives in N. Johore. Geol. Mag. 102, 338-46. 1 7 ,
179
Gt2MBEL, C. W. VON, 1874. Die Paliiolithischen Eruptivgesteine des Fichtelgebirges, Franz, Munich. GUPTA, A. K. & YAGI, K. 1980. Petrology and Genesis of Leucite-bearing Rocks, Springer, New York. - - , YAGI, K., HARIYA, Y., ONUMA, K. 1976. Experimental investigation of some synthetic leucite rocks under water pressures. Proc. Jap. Acad. Sci. 52, 469-72. --, LEMAITRE, R. W., HAUKKA, M. T. & YAGI, K. 1983. Geochemical studies on the carbonated apatite glimmerites from Damodar Valley, India. Proc. Jap. Acad. Sci. 59B, 113-16. HAGGERTY, S. E. 1976. Opaque mineral oxides in terrestrial igneous rocks. In: RUMBLE, D. (ed.) Oxides, Mineral. Soc. Am. Short Course Notes, Vol. 3, pp. 101-300. 1983. The mineral chemistry of new titanates from the Jagersfontein Kimberlite, S. Africa: implications for metasomatism in the upper mantle. Geochim. cosmochim. Acta, 47, 1833-854. - - , SMYTH, J. R., ERLANK, A. J., R/CHARD, R. S. & DANCHIN, R. V. 1983. Lindsleyite (Ba) and mathiasite (K): two new chromium-titanates in the crichtonite series from the upper mantle. Am. Mineral. 68, 494-505. HALL, A. 1974. West Cornwall, Geol. Assoc. Guide Book 19. Banham & Co., Colchester, England. 39 PP. 1982. The Pendennis peralkaline minette. Mineral. MaR. 45, 257-66. HALL, A. E. & SMITH, C. B. 1985. Lamproite diamonds: are they different? In: GLOVER, J. & HARRIS,P. G. (ed.) Kimberlite Occurrence and Origin, Univ. W Australia Publ. 8, pp. 167-211. HARGRAVES, R. B. & ONSTOTT,T. C. 1980. Paleomagnetic results from some southern African kimberlites, and their tectonic significance. J. geophys. Res. 85, 3587-97. HARRIS, P. G. 1957. Zone refining and the origin of potassic basalts. Geochim. cosmochim. Acta, 12, 195-208. 1984. Kimberlite volcanism. In: GLOVER, J. E. & HARRIS, P. G. (eds) Kimberlite Occurrence and Origin, Univ. I41.Australia Publ. 8, pp. 125-42. & MIDDLEMOST, E. m. 1969. The evolution of kimberlites. Lithos, 3, 77-88. HAWKESWORTH, C. J., ERLANK, A. J., MARSH, J. S., MENZlES, M. A. & VAN CALSTEREN, P. 1983. Evolution of tile continental lithosphere: evidence from volcanics and xenoliths in southern Africa. In: HAWKESWORTH, C. J. & NORRY, M. J. (eds) Continental Basalts and Mantle Xenoliths, pp. 11139. Shiva, Orpington. - - , FRASER,K. J. & ROGERS, N. W. 1986. Kimberlites and lamproites: extreme products of mantle enrichment processes. Spec. Pub. Geol. Soc. S. Afr. in press. HAWKINS, D. W. 1976. Emplacement, petrology and geochemistry of ultrabasic to basic intrusives at Allik Bay, Labrador. M. Sc. Thesis, Memorial University Newfoundland, St. Johns (unpublished). HAWTHORNE, J. B. 1975. Model of a kimberlite pipe. Phys. Chem. Earth, 9, 1-16. -
-
-
-
-
-
-
-
I
8
0
S. C.
Bergman
HEALD, K. C. 1926. Geology of the Ingomar anticline, Treasure and Rosebud counties, Montana. U.S. Geol. Survey Bull. 786, 1-37. HEARN, B. C. Jr. 1968. Diatremes with kimberlitic affinities in north-central Montana. Science 159, 622-5. 1979. Preliminary map o f diatremes and alkalic ultramatic intrusions, Missouri River Breaks and vicinity, North Central Montana. U.S. Geol Surv. Open File Rep. 79-1128. & MCGEE, E. S. 1982. Garnets in Montana diatremes: a key to prospecting for kimberlites. U.S. Geol. Surv. Open File Rep. 82-722. & -1984. Garnet peridotites from Williams kimberlites, north-central Montana, U.S.A. In: KORNPROBST, J. (ed.) Kimberlites II: The Mantle and Crust-Mantle Relationship, pp. 55-70. Elsevier, Amsterdam. HEINRICH, E. W. 1966. The Geology of Carbonatites, 555 pp. Rand McNally, New York. 555 pp. & MOORE, D. G. Jr. 1970. Metasomatic potash feldspar rocks associated with igneous alkalic complexes. Can. Mineral. 10, 571-84. HELMSTAEDT,H. & GURNEY, J. J. 1984. Kimberlites of southern Africa--are they related to subduction processes ? In: KORNPROBST,J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 425-34. Elsevier, Amsterdam. HENAGE, L. F. 1972. A definitive study of the origin of lamproites. M.S. Thesis, University of Oregon (unpublished). HERNANDEZ-PACHECO, F. 1935. Estudio fisiogral~hico y geologico del territorio comprendido entre Hellin y Cieza. An. Univ. Madrid (Cienc.), 4, 1-38. 1965. Una richterita potasica de rocas volcanicas alcalinas, Sierra de las Cabras (Albacete). Estud. Geol. 20, 265-70. HERTOGEN, J., LOPEZ-RUIZ, J., RODRIGUEZ-BADIOLA, E., DEMAIFFE, D. & WEIS, D. 1985. Petrogenesis of ultrapotassic rocks from S.E. Spain: trace elements and Sr-Pb isotopes. Geol. Assoc. Can. Progm. Abstr. 10, A26. HIGAZY, A. 1954. Trace elements of volcanic and ultrabasic potassic rocks of south-western Uganda and adjoining part of Belgian Congo. Geol. Soc. Am. Bull. 65, 39-70. HINTZE, L. F. 1980. Geological map of Utah (1/500,000 scale, 2 sheets). Utah Geol. Mineral. Surv., Salt Lake City. HOBSON, B. 1892. On the basalts and andesites of Devonshire, known as 'feldspathic traps'. Q. J. Geol. Soc. 48, 496-507. HOEFS, J. & WEDEPOHL, K . H . 1968. Strontium isotope studies on young volcanic rocks from Germany and Italy. Contrib. Mineral. Petrol. 19, 328-38. HOFMANN, A. W., WHITE, W. M. & WHITFORD, D. J. 1978. Geochemical constraints on mantle models: the case for a layered mantle. Yb. Carnegie lnstn. Wash. Yb. 77, 54-61. HOLM, P. M. & MUNKSGAARD, N. C. 1982. Evidence for mantle metasomatism: an oxygen and strontium isotope study of the Vulsinian district, central Italy. Earth planet. Sci. Lett. 60, 376-88.
-
-
-
-
-
-
-
-
-
-
HOLMES, A. 1932. The origin of igneous rocks. Geol. Mag. 69, 543-58. 1950. Petrogenesis of katungite and its associates. Am. Mineral. 35, 772-92. & HARWOOD,H. F. 1932. Petrology of the volcanic fields east and southeast of Ruwenzori, Uganda. Q. J. Geol. Soc. 88, 370-442. -& -1937. The volcanic area of Bufumbira. Pt II. Mere. Geol. Surv. Uganda, 3. HOLUB, F. V. 1977. Petrology of inclusions as a key to petrogenesis of the Durbachitic rocks from Czechoslovakia. Tschermaks Mineral. Petrogr. Mitt. 24, 133-50. HUNZIKER, J. C. 1974. Rb-Sr age determination and the alpine tectonic history of the western Alps. Mem. Ist. Geol. Min. Univ. Padova, 31, 1-54. HURLBUT, C. S. & GRIGGS, D. 1939. Igneous rocks of the Highwood Mountains, Montana: Part I, the Laccoliths. Geol. Soc. Am. Bull. 50, 1043-112. HURLEY, P. M., FAIRBAIRN, H. W. & PINSON, W. H. 1966. Rb-Sr isotopic evidence in the origin of potash-rich lavas of western Italy. Earth planet. Sci. Lett. 1,301-6. INNOCENTI, F., MANETTI,P., MAZZUOLI,R., PASQUARE, G. & VILLARI,L. 1982. Anatolia and NW Iran. In: THORPE, R. S. (ed.) Andesites, pp. 327-49. Wiley, London. JANSE, A. J. A. 1984. Kimberlites--where and when. In: GLOVER, J. E. & HARRIS, P. G. (eds) Kimberlite Occurrence and Origin. A Basis for Conceptual Models in Exploration. Univ. W. Australia Publ. 8, pp. 19-62. Perth, W. Australia. JAQUES, A. L. 1983. The ultrapotassic rocks of the West Kimberley region, W. A. : products of deep-seated intraplate volcanism. Proc. 6th Geological Society of Australia Cony., Vol 9. Geological Society of Australia, Sydney. -& FOLEY, S. F. 1985. The origin of Al-rich spinel inclusions in leucite from the leucite lamproites of western Australia. Am. Mineral., 70, 1143-50. GREGORY, G. P., LEWIS, J. D. & FERGUSON, J. 1982. The ultra-potassic rocks of the W. Kimberley region, W. Australia and a new class of diamondiferous kimberlite. Terra Cognita, 2, 251-2. --, LEWIS, J. D., GREGORY, G. P., FERGUSON, J., SMITH, C. B., CHAPPELL, B. W. & MCCULLOCH, M. T. 1983. The ultrapotassic, diamond-bearing rocks of the West Kimberley region, Western Australia. Proc. 6th Geological Society of Australia Cony., Vol. 9, pp. 286-7. Sydney. & 1984a. The , , , , , diamond bearing ultrapotassic (lamproitic) rocks of the West Kimberley region, Western Australia. In : KORNPROBST,J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 225-54. Elsevier, Amsterdam. -
-
-
-
-
-
& SMITH, C. B. 1986. T h e k i m b e r l i t e s a n d
lamproites of Western Australia. Geol. Surv. Western Australia Bull. 132, Perth, W.A. --, WEBB, A. W., FANNING, C. M., BLACK, L. P., PIDGEON, R. T., FERGUSON, J., SMITH, C. B. & GREGORY, G. P. 1984b. The age of the diamond bearing pipes and associated leucite lamproites of
Lamproites and K-rich igneous rocks the west kimberley region, western Australia. Bur. Mineral. Res. Aust. Geol. Geophys. 9, 1-7. JENSEN, H. I. 1908. The distribution, origin and relationships of alkaline rocks. Proc. Linn. Soc. N.S.W. 33, 491-588. JEREMINE, E. & FALLOT, P. 1929. Sur la pr6sence d'une vari6t6 de jumillite aux environs de Calasparra (Murcia). C. R. Acad. Sci. Paris, 188, 800. JEWETT, J. M. 1964. Geologic map of Kansas (1/ 500,000 scale). State Geol. Surv. Kansas. JOHANSSEN,A. 1931. A Descriptive petrography of Igneous Rocks, Vol. 1, University of Chicago Press, Chicago, IL. -1933. A Descriptive Petrography of Igneous Rocks, Vol. 2, University of Chicago Press, Chicago, IL. -1935. A Descriptive Petrography of Igneous Rocks, Vol. 3, University of Chicago Press, Chicago, IL. -1939. A Descriptive Petrography of Igneous Rocks, Vol. 4, University of Chicago Press, Chicago, IL. JOHNSTONR. H. 1959. Geology of the Northern Leucite Hills, Sweetwater County, Wyoming. M. A. Thesis, University of Wyoming, Laramie, WY (unpublished). JONES, A. P. & SMITH, J. V. 1983. Petrological significance of mineral chemistry in the Agathla Peak and the Thumb minettes, Navajo volcanic field. J. Geol. 91,643-56. - & 1985. Phlogopite and associated minerals from Permian minettes in Devon, S. England. Bull. Geol. Soc. Finland, 57, 89-102. JOPLIN, G. A. 1965. The problem of the potash-rich basaltic rocks. Mineral. Mag. 34, 266-75. - 1968. The shoshonite association: a review. J. Geol. Soc. Aust. 15, 275-94. KAPLAN, G., FAURE, D., ELLOY, R. & HEILAMMER,R. 1967. Contribution ~ l'6tude de l'origine des lamproites. Bull. Centre Rech. PA VSNPA, l, 1539. KAY, R. W. & GAST, P. W. 1973. The rare earth content and origin of alkali-rich basalts. J. Geol. 81, 65382. KAY, S. M., KAY, R. W., HANGAS, J. & SNEDDEN, Z. 1978. Crustal xenoliths from potassic lavas, Leucite hills, Wyoming. Geol. Soc. Am. Abstr. Progm. 7, 432. KELLER, J. 1983. Potassic lavas in the orogenic volcanism of the Mediterranean area. J. Volc. Geoth. Res. 1 8 , 321-35. KEMP, J. F. 1897. The Leucite Hills of Wyoming. Geol. Soc. Am. Bull. 8, 169-82. - - & KNIGHT, W. C. 1903. Leucite Hills of Wyoming. Geol. Soc. Am. Bull. 14, 305-36. KENNEDY, G. C. 1955. Some aspects of the role of water in rock melts. Geol. Soc. Am. Spec. Pap. 62, 489503. KNIGHT, G. L. & LANDES,K. K. 1932. Kansas laccoliths. J. Geol. 40, 1-15. KNILL, D. C. 1969. The Permian igneous rocks of Devon. Bull. Geol. Surv. Gt. Br. 29, 115-38. -1982. Permian volcanism in Southwestern England, In : SUTHERLAND,D. S. (ed.) Igneous Rocks of the British Isles, pp. 329-32. Wiley, London. KNOPF, D. 1970. Les Kirnberlites et les Roches Apparentes de Cote d'Ivoire. Sodemi, Abidjan.
181
KOGARKO, L. N. 1980. Significance of chemical bond polarity in the geochemistry of magmatism. Geokhimiya, 9, 1286-97. , KRIGMAN, L. D. & SHARUDILO, N. S. 1968. Experimental investigations of the effect of the alkalinity of silicate melts on separation of fluorine into the gas phase. Geokhimiya, 8, 948-56. KOSTYUK, V. P. 1983. The potassic alkalic magmatism of the Barkal Aldan Belt. Soy. Geol. Geophys. 24, 31-38. KRAMERS, Ji D., RODDICK, J. C. i . & DAWSON, J. B. 1983. Trace element and isotope studies on veined, metasomatic and " M A R I D " xenoliths from Bultfontein, S. Africa. Earth planet Sci. Lett. 65, 90106. KRANCK, E. 1953. Bedrock geology of the seaboard of Labrador between Domino and Hopedale. Bull. Can. Geol. Surv. 26. KRAVCHENKO, S. M., VLASOVA,E. V. & PINEVICH, V. S. 1960. Batisite, a new mineral, Dokl. Akad. Nauk S.S.S.R. 133, 657-8. KRESTEN, P., PRINTZLAU,1., REX, D., VARTIAINEN,H. & WOOLLEY, A. 1977. New ages of carbonatic and alkaline ultramafic rocks from Sweden and Finland. Geol. Fbren Stockh. Fbrh. 99, 62-65. KRUMMENACHER, D. & EVERNDEN, J. F. 1960. D6termination d'gtge isotopique fautes sur quelques roches des Alpes par le m~thode K-At. Schweiz. Mineral. Petrogr. Mitt. 40, 267-77. KUEHNER, S. M. 1980. Petrogenesis of ultrapotassic rocks, Leucite Hills, Wyoming. M. Sc. Thesis, University of Western Ontario, London, Ontario (unpublished). - - , EDGAR, A. D. & ARIMA, M. 1981. Petrogenesis of the ultrapotassic rocks from the Leucite Hills, Wyoming. Am. Mineral. 66, 663-7. KUSHIRO, I. & AOKI, K. 1968. Origin of some eclogite inclusions in kimberlite. Am. Mineral. 53, 1347-67. KYSER, T. K., O'NEIL, J. R. & CARMICHAEL,I. S. E. 1981. Oxygen isotope thermometry of basic lavas and mantle nodules. Contrib. Mineral. Petrol. 77, 11-23.
& - 1982. Genetic relations among basic lavas and ultramafic nodules: Evidence from oxygen isotope compositions. Contrib. Mineral. Petrol. 81, 88-102. LACROIX, A. 1926. La syst6matique des roches leucitiques: les types de la famille syenitique. C. R. Acad. Sci. Paris, 182, 597-601. 1933a. Contribution /t la connaissance de la composition chimique et mineralogique des roches 6ruptives de l'Indochine. Bull. Serv. Geol. Indochine, 20, 3, 1-208. - 1933b. Les roches 6ruptivespotassiquesleucitiques ou non du Tonkin occidental. C. R. Acad. Sci. Paris, 197, 625-7. LANGFORD, R. E. 1973. A study of the origin of Arkansas diamonds by mass spectrometry. PhD Thesis, University of Georgia, Athens, GA (unpublished). LANGWORTHY,A. P. & BLACK,L. P. 1978. The Mordor complex: a highly differentiated potassic intrusion with kimberlitic affinities in central Australia. Contrib. Mineral. Petrol. 67, 51-62.
I82
S. C.
Bergman
LARSEN, E. S. 1940. Petrographic province of central Montana. Geol. Soc. Am. Bull. 51,887-948. , HURLBUT, C. S., BURGESS, C. H. & BULL, B. F. 1941. Igneous rocks of Highwood Mountains, Montana. Geol. Soc. Am. Bull. 52, 1857-68. LARSEN, L. M., REX, D. C. & SECHER, K. 1983. Age of carbonatites, kimberlites and lamprophyres from southern W. Greenland: recurrent alkaline magmatism during 2500 million years. Lithos, 16, 21521. LARSEN, W. N. 1954. Precambrian geology of the W. Uinta Mountains, Utah, M.S. Thesis, University of Utah, Salt Lake City, UT (unpublished). LASSERRE, J. L. 1976. Amphibolites and alkaline gneisses in the midland formations of Nepal; petrography, geochemistry-geodynamic involvements. In: Jest, C. (ed.) Himalaya ; sciences de la terre, Fr. Cent. Natl. Rech. Sci., Colloq. Int. 268, 2, 213-36. LEES, G. J. 1974. Petrochemistry of the mica lamprophyres (minettes) of Jersey (CI). Proc. Ussher Soc. 3, 149-55. LEMAITRE, R. W. 1976. The chemical variability of some common igneous rocks. J. Petrol. 17, 589637. LEWIS, H. C. 1887. On a diamondiferous peridotite and the genesis of the diamond. Geol. Mag. 4, 22-4. LEWIS, P. H. 1973. Mica pyroxenite inclusions in limburgite, Hopi Buttes Volcanic field, Arizona. unpubl. M. Sci. Thesis, Brigham Young Univ., Utah. LEWIS, R. D. 1977. Mineralogy, petrology and geophysical aspects of Prairie, Creek kimberlite, near Murfreesboro, Arkansas. M. S. Thesis, Department of Geology, Purdue University (unpublished). - - , MEYER, H. O. A., BOLIVAR,S. L. & BROOKINS,D. B. 1976. Mineralogy of the diamond bearing "kimberlite", Murfreesboro, Arkansas LOS, 57, 761 (abstract only). LIBBY, S. C. 1975. Origin of K-ultramafic rocks in the Enoree 'vermiculite' districts, S. Carolina. PhD Thesis, Pennsylvania State University (unpublished). LLOYD, F. E., ARIMA, M. & EDGAR, A. D. 1985. Partial melting of a phlogopite clinopyroxenite nodule from south-west Uganda: an experimental study bearing on the origin of highly potassic continental rift volcanics. Contrib. Mineral. Petrol. 91,321-9. LOPEZ RUIZ, J. & RODRIGUEZ BADIOLA, E. 1980. La region volcanica Neogena del Sureste de Espana. Estud. Geol. 36, 5-63. LOVE, J. D., WEITZ, J. L. & HOSE, R. K. 1955. Geological map of Wyoming (1/500,000 scale) U.S. Geol. Surv., Washington. LUHR, J. F. & CARMICHAEL,I. S. E. 1980, The Colima volcanic complex, Mexico I. Post caldera andesites from Volcan Colima. Contrib. Mineral. Petrol. 71, 343-72. -& -1981. The Colima volcanic complex, Mexico: I1. Late Quaternary cinder cones. Contrib. Mineral. Petrol. 76, 127-47. LUTH, W. C. 1967. Studies in the system KA1SiO4Mg2SiO,-SiO2-H20. I: Inferred phase relations
and petrologic applications. J. Petrol. 8, 372416. MACDOWELL, F. W. 1966. Potassium-Argon dating of Cordilleran intrusives. PhD Thesis, Columbia University, New York (unpublished). MCCANDLESS,T. E. 1982. The mineralogy, morphology and chemistry of detrital minerals of a kimberlitic and eclogitic nature, Green River Basin, Wyoming. M. S. Thesis. University of Utah, Salt Lake City, UT (unpublished). - 1984. Detrital minerals of mantle origin in Green River Basin, Wyoming. Soc. Min. Eng. AIME, Preprint 84-395. MCCLURE, W. C. 1963. Geology of the Enoree area: S. Carolina. M.Sc. Thesis, University of South Carolina, Columbia, SC (unpublished). MCCULLOCH, M. T., JAQUES, A. L., NELSON, D. R. & LEWIS, J. D. 1983a. Nd and Sr isotopes in kimberlites and lamproites from W. Australia: an enriched mantle origin. Nature, Lond. 302, 400-3. , NELSON, D. R., JAQUES, A. L. & LEWIS, J. O. 1983b. Nd and Sr isotopic systematics in kimberlites and lamproites from West Kimberley, Western Australia. Proc. 6th Geological Society o f Australia Cony., Vol. 9, pp. 287-8. Geological Society of Australia, Sydney. McKIE, O. 1966. Fenitization. In: TUTTLE, C. F. & GITTINS, J. (eds) Carbonatites, pp. 261-94. WileyInterscience, New York. MADIGAN, R. 1983. Diamond exploration in Australia. Indiaqua, 35, 27-38. MAKHERJEE, K. K. 1961. Petrology of lamprophyres of the Bokaro coalfield. Qt. J. Geol. Min. Metall. Soc. India, 33, 69-87. MAKSlMOV, E. P. 1982. The Mesozoic magmatism of the Aldan shield as an indicator of its tectonic regime. Soy. Geol. Geophys. 23, 9-14. MANSON, V. 1967. Geochemistry of basaltic rocks: major elements. In : HESS, H. H., & POLDERVAART, A. (eds) Basalts. Wiley, New York, 215-70. MARINELLI, G. & MITTEMPERGHER, M. 1966. On the genesis of some magmas of typical Mediterranean (potassic) suite. Bull. Volcanol. 29, 113-40. MARTHUR, S. M. & SINGH, H. N. 1971. Petrology of the Majgawan pipe rock: Geol. Surv. India Misc. Publ. 19, 78-85. MARVIN, R. F., HEARN, B. C., JR., MEHNERT, H. H., NAESER, C. W., ZARTMAN,R. E. & LINDSEY,D. A. 1980. late Cretaceous-Paleocene igneous activity in north-central Montana. Isochron/West, 29, 525. MASON, B. 1967. Meteorites. Am. Sci. 55, 429-55. - 1977. Elemental distribution in minerals from the Wolgidee Hills intrusion, Western Australia. N. Z. DSIR Bull. 218, 114-20. -1979. Cosmochemistry, Part I, Meteorites. Prof. Pap. U.S.G.S. 440B-1. MATSON, R. E. 1960. Petrography and petrology of Smoky Butte Intrusives, Garfield County Montana. M.S. thesis, Montana State University, Missoula, MT (unpublished). MEAKINS, A. 1983. Geology and genesis of the Argyle alluvial diamond deposits, Kimberley region, W. Australia. In." DAVY, R., BUTTE, C. R. M. &
Lamproites and K-rich igneous rocks BALLINGER,T. A. (eds) Geochemical Exploration in Arid and Deeply Weathered Environments. Geochemistry and Genesis of Ore Deposits Associated with Weathering, Aust. Reg. Meet., Assoc. Expl. Geochem. Perth, Abstracts volume. 54-6. MELTON, C. E. & GIARDINI, A. A. 1976. Experimental evidence that oxygen is the principal impurity in natural diamond. Nature, Lond. 263, 309--10. - & -1980. The isotopic composition of Ar included in an Arkansas diamond and its significance. Geophys. Res. Lett. 7, 461-4. --, Salotti, C. A. & GIARDINI, A. A. 1972. The observation of N2, H20, CO2, CH4 and Ar as impurities in natural diamonds. Am. Mineral. 57, 1518-23. MENZIES, M. & MURTHY, V. R. 1980a. Enriched mantle: Nd and Sr isotopes in diopsides from kimberlite nodules. Nature, Lond. 283, 634-6. - - & - 1980b. Nd and Sr isotope geochemistry of hydrous mantle nodules and their host alkali basalts: implications for local heterogeneities in metasomatically veined mantle. Earth planet. Sci. Lett. 46, 323-34. MERRILL, R . B . , B I C K F O R D , M . E . & I R V I N G , A . J . 1977. The Hills Pond Peridotite, Woodson County, Kansas: A richterite-bearing Cretaceous intrusive with kimberlitic affinities. Extended Abstr. 2nd Kimberlite Conf., Santa Fe, NM. MESEGUER PARDO, J. 1924. Estudio de los yacimlentos de azufre de las provincias de Murcia y Albacete. Bol. Inst. Geol. Esp. 45, 131-214. MEYER, H. O. A. 1976. Kimberlites of the continental United States: a review. J. Geol. 84, 377-403. & MITCHELL, R. H. 1985. Sapphire-bearing lamprophyre from Yogo Gulch, Montana. Geol. Assoc. Can. progm, abstr. 10, A39. --, LEWIS, R. D., BOLIVAR, S. & BROOKINS, D. G. 1977. Prairie Creek Kimberlite, Murfreesboro, Pike County, Arkansas. Extended Abstr. 2nd Int. Kimberlite Conf., Santa Fe, NM. MIDDLEMOST,E. A. K. 1975. The basalt clan. Earth Sci. Rev. 11,337-64. MILLER,J. A. & MOHR, P. A. 1964. K - A r measurements on the granites and some associated rocks from Southwest England. Geol. J. Liverpool, 4, 105-26. MILLER, K. C. 1982. Paleomagnetism of Indian Kimberlites. B. A. Thesis, Princeton University, Princeton, NJ (unpublished). MILLER, T. P. 1970. Petrology of the plutonic rocks of west-central Alaska. U. S. Geol. Surv. Open File Rep. .545. - 1972. Potassium-rich alkaline intrusive rocks of Western Alaska. Geol. Soc. Amer. Bull. 83, 211128. MINEYEVA, I. G. 1972. Principal geochemical characteristics of potassic alkalic rocks. Geochem. Int. 9, 173-9. MISER, H. D. & PURDUE, A. H. 1929. Geology of the De Queen and Caddo Gap quandrangle, Arkansas. U.S. Geol. Survey Bull. 808, 195. - & Ross, C. S. 1923. Diamond-bearing peridotite in Pike County, Arkansas. U. S. Geol. Survey Bull. 735--1,279-322. MIrCHtELL, R. H. 1981. Titaniferous phlogopites from
183
the leucite lamproites of the W. Kimberley Area, W. Australia. Contrib. Mineral. Petrol. 76, 243-51. - 1985. A review of the mineralogy of lamproites. Trans. Geol. Soc. SA, 88, 411-437. - & BELL, K. 1976. Rare earth element geochemistry of potassic lavas from the Birunga and ToroAnkole regions of Uganda, Africa. Contrib. Mineral. Petrol. 58, 293-303. - & HAWKESWORTH, C. J. 1984. Geochemistry of potassic lavas from Smoky Butte, Montana. Geol. Soc. Am. Abstr. Progm. 16, 597. - & Lewis, R. D. 1983. Priderite-bearing xenoliths from the Prairie Creek mica peridotite, Arkansas. Can. Mineral. 21, 59-64. MODRESKI, P. J. & BOETTCHER,A. L. 1972. The stability of phlogopite + enstatite at high pressures: a model for micas in the interior of the earth. Am. J. Sci. 272, 852--69. -& 1973. Phase relationships of phlogopite in the system K 2 0 - M g O - C a O - A 1203-SIO2-H20 to 35 kilobars : a better model for micas in the interior of the earth. Am. J. Sci. 273, 385~414. MOODY, C. L. 1949. Mesozoic igneous rocks of Northern Gulf Coast plain, Bull. Am. Assoc. Petrol. Geol. 33, 1410-28. MORRIS, E. C. 1953. Geology of the Big Piney area, Summit County, Utah. M.S. Thesis, University of Utah, Salt Lake City, UT (unpublished). MORRISON, G. W. 1980. Characteristics and tectonic setting of the shoshonite rock association. Lithos, 13, 97-108. MULLEN, E. D. & MURPHY, S. 1985. Mineralogy and petrology of lamprophyric and carbonatite dikes, central Arkansas. Geol. Soc. Am. Abstr. Progm. 17, 185. - & PETTY, W. B. 1985. Petrologic relations among syenites and lamprophyric rocks, Arkansas alkalic Province. Geol. Assoc. Can. Progm. Abstr. 10, A42. MURTY, M. K. 1980. Tectonics and geochemistry of the diamondiferous kimberlites of the Jungle Valley, Merzapur District, U.P.--an evolutionary model. Proc. Conf. on Diamonds, Geological Survey of India, Bombay. MURTY, Y. G. K., RAO, M. G., MISRA, R. C. & AJIT KUMAR REDDY, T. 1980. Kimberlite diatremes of Andhra Pradesh--their assessment and search for concealed bodies. Prof. Conf. on Diamonds, Geological Survey of India, Bombay. MUTCI-ILER, E. E., ROUGHON, D. J. & LAVIN, O. P. 1976. PETROS: a data bank of major-element chemical analyses of igneous rocks for research and teaching. Comput. Geol. Sci. 2, 51-7. MYSEN B. O. & BOETTCHER,A. L., 1975a. Melting of a hydrous mantle: I. Phase relations of natural peridotite at high pressures and temperatures with controlled activities of water, carbon dioxide, and hydrogen. J. Petrol. 16, 520--48. - & -1975b. Melting of a hydrous mantle: II. Geochemistry of crystals and liquids formed by anatexis of mantle peridotite at high pressures and high temperatures as a function of controlled activities of water, hydrogen, and carbon dioxide. J. Petrol. 16, 549-93. NELSON, D. R., MCCULLOCH,M. T. & SUN, S.-S. 1986.
184
S. C. Bergman
The origins of ultrapotassic rocks as inferred from Sr, Nd and Pb isotopes. Geochim. cosmochim. Acta, 5 0 , 231-46. N~MEC, D. 1972. Micas of the lamprophyres of the Bohemian Massif Neues Jb. Mineral. Abh. 117, 196-216. 1973. Amphibole der lamprophyrischen und lamproiden Gestine in Ostfeil der Bohmischen Masse. Cas. Mineral. Geol. 18, 31-46. -1974. Petrochemistry of the dyke rocks of the Central Bohemian Pluton. Neues J. Mineral. Monatsh, 193-209. - 1985. Ba and its carriers in dyke rocks of the minette series. Geol. Assoc. Canada Progm. Abstr. 10, A43. NICHOLLS, J. & CARMICHAEL, I. S. E. 1972. The equilibrium temperature and pressure of various lava types with spinel- and garnet-peridotites. Am. Mineral. 57, 941-59. NIELSEN, B. L., 1976. Economic minerals diamonds, olivine, garnets in Greenland. In. ESCHER, A. & WATT, W. S. (eds) Geology of Greenland, pp. 4816. Geological Survey of Greenland, Copenhagen. NIGGLI, P. 1923. Gesteins- undMineralprovinzen. Gebruder Borntraeger, Berlin. NIXON, P. H., THIRWALL, M. F. & BUCKLEY F. 1982. Kimberlite-lamproite consanguinity. Terra Cognita 2, 252-54. , , -& DAVIES, C. J. 1984. Spanish and Western Australian lamproites: aspects of whole rock geochemistry. In: KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 285-96. Elsevier, Amsterdam. NOBEL, F . A . , A N D R I E S S E N , P . A . M . , H E B E D A , E . H . PRIEM, H. N.A. & RONDEEL, H. E. 1981. Isotopic dating of the post-Alpine Neogene volcanism in the Betic Cordilleras, southern Spain. Geol. Mijnbouw, 60, 209-14. NOCKOLDS, S. R. 1940. Petrology of rocks from Queen Mary Land. Australian Antarctic Expedition 191114, Rep. Ser. A, IV, Part 2, pp. 15-86. - 1954. Average chemical composition of some igneous rocks. Geol. Soc. Am. Bull. 65, 1007-32. NOE-NYGAARD, A. & RAMBERG, H. 1956. Geological reconnaisance map of the country between latitudes 69~ and 63~ West-Greenland, (1/500,000 scale). Geological Survey of Greenland, Copenhagen. -& -1961. Geological reconnaissance map of the country between latitudes 69~ and 63~ W. Greenland. Meddr. Gronland, 123, 3. NORRISH, K. 1951. Priderite, a new mineral from the leucite-lamproites of the West Kimberley area, Western Australia. Mineral Mag. 24, 496-501. NOUGIER, J. 1970. Contribution ~ l'6tude g6ologique et g6omorphologique des iles Kerguelen (Territoire des Terres Australes et Antarctiques Fran~aises) Com. Natl. Fr. Rech. Antarct., 27, 696 pp. O'BRIEN, H. E., IRVING, A. J. & MCCALLUM,I. S. 1985. Complex zoning of clinopyroxene in shonkinites and mafic phonolites, Highwood Mtns. Montana: evidence for periodic mixing with a K-rich basanitic magma. Geol. Soc. Am. Abstr. Progm. 17, 187.
OGDEN, P. R., Jr. 1979. The geology, major element geochemistry and petrogenesis of the Leucite Hills volcanic rocks, Wyoming. PhD Thesis, University of Wyoming, Laramie, WY (unpublished). --, SPERR, J. T., GUNTER, W. D. & PAJARI, G. E., JR. 1978. Morphology of a recent ultrapotassic volcanic field, Leucite Hills, SW Wyoming. Geol. Soc. Am. Abstr. Progm. 10, 140. --, VOLLMER, R. & SCHILLING, J. G. 1980. Leucite Hills revisited. LOS, 61,412. OSANN, A. 1889. Beitrage Zue Kentniss der Eruptivegesteine des Cabo de Gata I. Z. dch. Geol. Ges. 41, 297-311. 1906. Uber einige Alkaligestine aus Spanien, Festchritte, Rosenbusch, Stuttgart. OSBORNE, F. F. & ROBERTS, E. J. 1931. Differentiation in the Shonkin Sag laccolith, Montana. Am. J. Sci. 5th ser. 22, 331-53. OWEN H. G. 1983. Atlas of continental displacement, 200 million years to the present. Cambridge Univ. Press, Cambridge. OYAWOVE, M. O. 1976. The origin and evolution of the potassic rocks of SW Nigeria. 25 Int. Geological Congr. Abstr., Vol. 21, 10A, pp 424-5. PAGEL, M. & LETERRIER, J. 1980. The subalkaline potassic magmatism of the Ballous Massif (S. Vosges, France); shoshonitic affinity. Lithos 13, 110. PARGA PONDAL, I. 1935. Quisismo de las manifestaclones magmaticas Cenozoicas de la Peninsula Iberica. Trab. Mus. Nac. C. Nat. Ser. Geol. 39. PAUL, O. K., REX, D. C. & HARRIS, P. G. 1975. Chemical characteristics and K-Ar ages of Indian kimberlites. Geol. Soc. Am. Bull. 86, 364-6. PELLICER, M. J. 1973. The petrology and geochemistry of new lamproite rocks near Aljorra, Murcia, Spain. Eustud. Geol. Inst. Invest. Geol. Lucas Mallade, 29, 99-106. PHILLIPI, E. 1912. Geologische Beschreibung des Gassbergs. Deutsche Siidpolar Expedition 19011903. Geogr. Geol. l l , 47-71. PIRSSON, L. V. 1905. Peti-ography and geology of the igneous rocks of the Highwood Mountains, Montana. Bull. U.S. Geol. Surv. Bull. 237. POLANSKI, A. 1949. The alkaline rocks of the EastEuropean Plateau. Bull. Soc. Amis. Sci. Lett. Poznan, Ser. B, 10,119-84. POLDERVAART,A. 1955. Chemistry of the Earth's crust. Geol. Soc. Am. Spec. Pap. 62, 119-44. POWELL, J. L. & BELL, K. 1970. Strontium isotopic studies of alkalic rocks. Localities from Australia, Spain, and the Western United States. Contrib. Mineral. Petrol. 27, 1-10. - & 1974. Isotopic composition of Sr in alkalic rocks. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp 412-21. Wiley, New York. PRIDER, R. T. 1939. Some minerals from the leuciterich rocks of the West Kimberley area, Western Australia. Mineral. Mag. 25, 373-87. - 1960. The leucite lamproites of the Fitzroy Basin, Western Australia. J. Geol. Soc. Aust. 6, 71-118. - 1965. Noonkanbahite, a potassic batisite from the lamproites of Western Australia. Mineral. Mag. 34, 403-5.
185
Lamproites and K-rich igneous rocks 1982. A glassy lamproite from the West Kimberley area, Western Australia. Mineral. Nag. 45, 27982. - & COLE, W. F. 1942. the alteration products of olivine and leucite in the leucite-lamproites from the West Kimberley area, Western Australia. Am. Mineral. 27, 373-84. PRIMAL, L. 1963. Etude g6ologique et m6tallog6nique de la partie m6ridional du cap Corse. PhD Thesis, University of Paris (unpublished). PROSTKA, H. J. 1973. Hybrid origin of the absarokiteshoshonite-banakite series, Absaroka volcanic field, Wyoming. Geol. Soc. Am. Bull. 84, 697-702. PRYCE, M. W., HODGE, L.C. & CRIDDLE, A. J. 1984. Jeppeite, a new K-Ba-Fe titanate from Walgidee Hills, Western Australia. Mineral. Nag. 48, 2636. RADULESCU, D. P. 1966. Rhyolites and secondary ultrapotassic rocks in the subsequent Neogene volcanism from the E. Carpathians. Bull. Volcanol. 29, 425-34. RAVlCH, M. G. & KRYLOV, A. H. 1964. Absolute ages of rocks from E. Antarctica. In." ADIE, R. J. (ed.) Antarctic Geology, pp. 578-89. North-Holland, Amsterdam. REINISCH, R. 1912. Petrographische-Beschreibungder Gaussberg-Gesteins. Deutsch Sfidpolar-Expedition 1901-03. Geogr. Geol. 11, 73-8. RIOU, R., DuPuY, C. & DOSTAL, J. 1981. Geochemistry of co-existing alkaline and calc-alkaline volcanic rocks from Northern Azerbaijan (N.W. Iran). J. Volc. Geoth, Res. 11,253-75. RITTMANN,A. 1933. Die geologisch bedingte Evolution und Differentiation des Somma-Vesuvius magma. Z. Vulkanol. 5, 8-94. -1951. Magmatic character and tectonic position of the Indonesian volcanoes. Nomenclature of volcanic rocks. Bull. Volcanol. 12, 75-109. ROBINS, B., FURNES, H. & RYAN, P. 1983. A new occurrence of dalyite. Mineral. Nag. 47, 93-4. ROBINSON, E. C. 1976. Geochemistry of lamprophyric rocks of the eastern Ouachita Mtns, Arkansas. M.S. Thesis, University of Arkansas, Fayetteville, AR (unpublished). ROCK, N. M. S. 1976. The comparative Sr isotopic composition of alkaline rocks; new data from South Portugal and East Africa. Contrib. Mineral. Petrol. 56, 205-28. 1977. The nature and origin of lamprophyres: some definitions, distinctions and derivations. Earth Sci. Rev. 13, 123-69. - 1984. Nature and origin of calc-alkaline lamprophyres: minettes, vogesites, kersantites, spessartites. Trans. R. Soc. Edinburgh, 74, 193-227. 1986. The nature and origin of lamprophyres: alnoites and allied rocks (ultramafic lamprophyres). J. Petrol. 27, 155-96. - 1987. The nature and origin of lamprophyres : an overview. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 191-226. RODEN, M. F. 1981. Origin of coexisting minette and ultramafic breccia, Navajo volcanic field. Contrib. Mineral. Petrol. 77, 195-206. - -
---,
SMITH, D. & McDOWELL, F. W. 1979. Age and extent of potassic volcanism on the Colorado Plateau. Earth planet. Sci. Lett. 43, 279-84. ROEDDER, E. 1984. Fluid inclusions. Rev. Mineral. 12, 506-511. ROGERS, N. W. BACHINSKI, S. W. HENDERSON, P. & PARRY, S. J. 1982. Origin of potash-rich basic lamprophyres; trace element data from Arizona minettes. Earth planet. Sci. Lett. 57, 305-12. , HAWKESWORTH,
C.
J., PARKER,
R.
J. & MARSH,
J. S. 1985. The geochemistry of potassic lavas from Vulsini, central Italy and implications for mantle enrichment processes beneath the Roman region. Contrib. Mineral. Petrol. 90, 244-57. RONOV, A. B. & YAROSHEVSKY,A. A. 1969. Chemical composition of the Earth's Crust. Am. Geophys. Union Monogr. 13, 37-57. ROSENBUSCH, H. 1877. Mikroskopische Physiographie, 1st edn. Schweizerbart, Stuttgart. 1887. Mikroskopische Physiograpie, 2nd edn. Schweizerbart, Stuttgart. RosI, M. 1980. The island of Stromboli. Soc. Ital. Mineral. Petrol., Rend. 36, 345-68. Ross, C. P., ANDREWS, D. A. & WITKIND, I. J. 1955. Geological map of Montana (1/500,000 scale). Montana Bureau of Mines and Geology, Helena. Ross, C. S. 1926a. Nephelite-hauynite alnoite from Winnett Montana. Am. J. Sci. 11,218-27. -1926b. A Colorado lamprophyre of the verite type, Am. J. Sci. 12, 217-29. - - , MISER, H. O. & STEPHENSON,L. W. 1929. Water laid volcanic rocks of early Upper Cretaceous age in SW Arkansas, SE Oklahoma and NE Texas. U.S. Geol. Surv. Prof Pap. 154F, 175-202. ROWELL, W. F. & EDGAR, A. D. 1983. Cenozoic potassium-rich mafic volcanism in the western U.S.A.: its relationship to deep subduction. J. Geol. 91,338-41. RoY, B. C. 1962. Geological Map of India (1/5,000,000 scale). Geol. Surv. India. RoY CHAUDHURY,M. K. 1973. Geological and mineral map of Andhra Pradesh (1/2,250,000 scale). Geol. Surv. India. RUDDOCK, D. I. & HAMILTON,D. L. 1978a. Stability of carbonate in a simple potassium-rich rock model. Prog. Exp. Petrol. N E R C Publ., Ser. D, 11, 2831. -& -1978b. The system KA1Si206-CaMgSi206-SiO 2 at 4 kb. Prog. Exp. Petrol. NERCPubl., Ser. D, 11, 25-27. RYABCHIKOV, I. D. & BOETTCHER,A. L. 1981. Experimental evidence at high pressure for potassic metasomatism in the mantle of the earth. Am. Mineral. 65, 915-9. - & GREEN, D. H. 1978. The role of CO2 in the petrogenesis of highly potassic magmas. Tr. Inst. Geol. Geofiz. Akad. Nauk S.S.S.R. No 403, 49-64. --, SCHREYER, W. & ABRAHAM, K. 1982. Composition of aqueous fluids in equilibrium with pyroxenes and olivines at mantle pressures and temperatures. Contrib. Mineral. Petrol. 79, 80-4. SAETHER, E. 1950. On the genesis of peralkaline rock provinces. 18th Int. Geology Cortgr., London, Part II, pp. 123-30.
186
S. C. Bergman
SAHAMA, Th. G. 1974. Potassium-rich alkaline rocks. In : SORENSEN,H. (ed.) The Alkaline Rocks, pp. 96109, Wiley, London. SALTERS, V. J. M., & BARTON, M. 1985. The geochemistry of ultrapotassic lavas from The Leucite Hills, Wyoming. LOS, 66, I 109. SAN MIGUEL DE LA CAMERA,M. 1935. Una erupcion de jumillita en la Sierra de las Cabras (Albacete). Bol. R. Soc. Esp. Hist. Nat. 35, 147-54. , ALMELA, A. & FUSTER, J. M. 1952. sobre un volcan de veritas recientemente descubierto en el Miocene de Barqueros (Murcia). Estud. Geol. 7, 411-29. SANYAL,S. P. 1964. Petrology of certain lamprophyres from the Jharia Coalfield Bihar, with a discussion on the differenttiation of the Sudamdih Sill. Misc. Publ. Geol. Surv. India, 8, 27-44. SARKAR, A., PAUL, D. K., BALASUBRAHMANYAN,M. N. & SENGUPTA, N. R. 1980. Lamprophyres from Indian Gondwanas; K - A r ages and chemistry. J. Geol. Soc. India, 21, 188-93. SARMA, B. S. P. 1983. A report on ground magnetic survey over Chelima dyke (Cuddapah basin). Nat. Geophys. Res. Inst., Hyderabad, India, Tech. Rep. 83-210. SATHE, R. V. & S. S. 1975. Petrogenesis oflamprophyres of Mt. Girnar, Saurahtra. Q. J. Geol. Min. Metall. Soc. India, 46, 61-7. SATYANARAYANA RAO, P. & PHADTRE, P. N. 1966. Kimberlite pipe rocks of Wajrakarur Area, Anantapur district, Andhra Pradesh. J. Geol. Soc. India, 7, 118-23. SCHAIRER, J. F. & BOWEN, N. L. 1938. The system leucite-diopside-silica. Am. J. Sci. 235-A, 289-309. SCHEURING, B., AHRENDT, H., HUNZIKER, J. C. & ZINGG, A. 1974. Paleobotanical and Geochronological evidence for the Alpine Age of the metamorphism in the Sesia Zone. Geol. Rundsch. 63, 305-26. SCHMITT, H. H., SWANN, G. A. & SMITH, D. 1974. The Buell Park kimberlite pipe, N. E. Arizona. In: KARLSTRUM, T., SWANN, G. A. & EASTWOOD, R. L. (eds) Geology of Northern Arizona, pp. 672-98. Northern Arizona Univ., Flagstaff. SCHNEIDER, M. E. 1982. Major element compositions of silicates dissolved in supercritical fluids at mantle conditions: implications for mantle metasomatism. M.S. Thesis, Pennsylvania State University (unpublished). & EGGLER, D. N. 1984. Compositions of fluids in equilibrium with peridotite: implications for alkaline magmatism-metasomatism. In: KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 383-94. Elsevier, Amsterdam. SCHULTZ, A. R. & CROSS, W. 1912. Potash-bearing rocks of the Leucite Hills, Sweetwater County, Wyoming. U.S. Geol. Surv. Bull. 512, 1-39. SCHULZE, D. J. & HELMSTAEDT, H. 1979. Garnet pyroxenite and eclogite xenoliths from the Sullivan Buttes latite, Chino Valley, Arizona. In : BOYD, F. R. & MEYER, H. O. A. (eds) The mantle sample: Inclusions in kimberlites and other volcanics. Proc. 2nd Int. Kimberlite Conf., Vol. 2. pp. 318-29. American Geophysical Union, Washington, DC. -
-
--,
SMITH, J. V., & N~MEC, D. 1985. Mica chemistry of lamprophyres from the Bohemian Massif, Czechoslovakia. Neues. Jahrb. Mineral. Abh. 152, 321-34. SCOTT, B. H. 1977. Petrogenesis of kimberlites and associated potassic lamprophyres from central West Greenland. PhD Thesis, University of Edinburgh (unpublished). ,1979. Petrogenesis of kimberlites and associated potassic lamprophyres from central West Greenland. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds. Their Geology, Petrology and Geochemistry. Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 190-205. American Geophysical Union, Washington, DC. 1981. Kimberlite and lamproite dykes from Holsteinsborg, W. Greenland. Meddr Gronland, 4, 3-24. SCOTT-SMITH,B. H. & SKINNER, E. M. W. 1982. A new look at Prairie Creek, Arkansas. Terra Cognita, 2, 210. & -1984a. A new look at Prairie Creek, Arkansas. In: KORNPROBST,J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 255-84. Elsevier, Amsterdam. & -1984b. Diamondiferous lamproites. J. Geol. 92, 433-8. --- & -1984c. Kimberlite and American Mines, near Prairie Creek Arkansas. In: KORNPROBST,J. (ed.) Kimberlites III: Documents Ann. Sci. Univ. Clermont-Ferrand II 74, 27-36. SCULL, B. J. 1959. The age of mineralization in the Ouachita Mountains of Arkansas and Oklahoma. Syrup. on the Geology of the Ouachita Mountains, pp. 62-9. Dallas Geological Society and Ardmore Geological Society. SEARLE, M. P. 1984. Alkaline peridotite pyroxenite and gabbroic intrusions in the Oman Mountains, Arabia. Can. J. Earth Sci. 21,396-406. SEN, S. N. & NARASHIMARAO 1970. Chelima dykes. Proc. 2nd Syrup. on Upper Mantle, Project Session 5, Combined Geosurveys in Dharwar and Cuddapah Basin, Hyderabad, India, pp. 435-9. SHAFIQULLAH,M., DAMON, P. E. & LYNCH, D. J. 1976. Ultrapotassic trachytes at Pichaco Peak in the southern Basin and Range province, Arizona. Geol. Soc. Am. Abstr. Progm. 8, 628. SHAND, S. J. 1927. Eruptive Rocks, l st edn. Thomas Murby, London. -1931. The granite-syenite-limestone complex of Palabora, Transvaal. Trans. Geol. Soc. S. Aft. 67, 81-92. SHERATON, J. W. 1983. Geochemistry of mafic igneous rocks of the northern Prince Charles Mountains, Antarctica. J. Geol. Soc. AuNt. 30, 295-300. & CUNDARI, A. 1980. Leucitites from Gaussberg, Antarctica. Contrib. Mineral. Petrol. 71,417-27. & ENGLAND, R. N. 1980. Highly potassic mafic dykes from Antarctica. J. Geol. Soc. AuNt. 27, 12935. , OFFE, L. A., TINGLY, R. J. & ELLIS, D. J. 1980. Enderbyland, Antarctica-an unusual Precambrian high-grade metamorphic terrain. J. Geol. Soc. AuNt. 27, 1-18. -
-
-
-
-
-
-
-
Lamproites and K-rich igneous rocks SIMPSON, E. S. 1925. Contributions to the mineralogy of Western Australia. Series I. J. R. Soc. W. Aust. 12, 57-66. SKEATS, E. W. & RICHARDS, H. C. 1926. Report of the alkaline rocks of Australia and New Zealand Committee. Aust. Assoc. Adv. Sci. 18, 36-45. SKINNER, E. M. W. & CLEMENT, C. R. 1979. Mineralogical classification of Southern African kimberlites. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds: Their Geology, Petrology and Geochemistry. Proc 2nd Int. Kimberlite Conf., Vol. 1, pp. 129-39. American Geophysical Union, Washington, DC. -& SCOTT, B. H. 1979. Petrography, mineralogy and geochemistry of kimberlite and associated lamprophyre dykes near Swartruggens, W. Transvaal, RSA. Kimberlite Symp. 11, Cambridge. SMITH, C. B. 1983a. Pb, Sr and Nd isotopic evidence for sources of Southern African Cretaceous kimberlites. Nature, Lond. 304, 51-4. -1983b. Rb-Sr, U - P b and S m - N d isotopic studies of kimberlite and selected mantle-derived xenoliths. PhD Thesis, University of Witwaterstrand, Johannesburg (unpublished). -1984a. The genesis of the diamond deposits of the West Kimberley, W. A. In: PURCELL, P. G. (ed.) The Canning Basin, WA. Proc. Geological Society of Australia and Petrology Exploration Society of Australia Symp., Perth, pp. 463-74. -1984b. What is a kimberlite? In: GLOVER, J. E. & HARRIS, P. G. (eds) Kimberlite Occurrence and Origin: a Basisfor Conceptual Models in Exploration. Univ. W. Australia Publ. 8, pp. 1-18. SMITHSON, S. B. 1959. The geology of the southeastern Leucite Hills, Sweetwater County, Wyoming. M.A. thesis, Univ. Wyoming, Laramie (unpublished). SMYTH, C. H. MR. 1927. The genesis of alkaline rocks, Am. Phil. Soc. Proc. 66, 535-80. SOBOLEV, V. S. 1970. On the genesis of leucite-bearing rocks. Dokl. Akad. Nauk S.S.S.R., Earth Sci. Sect, 194, 166-8. , BAZAROVA,T. J. & YAGI, K. 1975. Crystallization temperature of wyomingite from Leucite hills. Contrib. Mineral. Petrol. 49, 301-8. SOLOVlEV, D. A. 1972. Platform magmatic formations of East Antarctica. In: ADIE, R. J. (ed.) Antarctic Geology and Geophysics, pp. 531-8. Universitetsforlaget, Oslo. SORENSEN, H. 1960. On agpaitic rocks. In: OFTEDAHL, C. & HJELMQVIST, S. (eds) Proc. 21st Int. Geol. Congr., Part 13, Petrographic Provinces, lgneous and Metamorphic Rocks, Det Berlingske Bogtrykkeri, Copenhagen pp. 319-27. STAUDIGEL,H. • ZINDLER, A. 1978. Nd and Sr isotope compositions of potassic volcanics from the East Eifel, Germany: implications for mantle source regions. Geol. Soc. Am. Abstr. Prog. 10, 497. STEELE, K. F. & WAGNER, G. H. 1979. Relationship of the Murfreesboro kimberlite and other igneous rocks of Arkansas, USA. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes, and Diamonds. Their Geology, Petrology and Geochemistry. Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp.
187
393-9. American Geophysical Union, Washington, DC. STEFANINI, G. 1934. I1 complesso eruttivo di orciatico e Montecatini inprovincia di Pisa. Atti Soc. Tosc. Sci. Nat. Mem. 44, 224-300. STEFANOVA, M. & BOYADJIEVA, R. 1975. On the geochemistry of Nb and Ta in K-alkaline lamproitic rocks from the village of Svidnja, district of Sofia. Geocham. Min. Petrol. 1, 16-30. STEPHENSON, P. J. 1972. Geochemistry of some Heard Island igneous rocks. In: ADIE, R. J. (ed.) Antarctic Geology and Geophysics, pp. 793-801. Universitetsforlaget, Oslo. STEWART, D. B. 1979. The formation of siliceous potassic glassy rocks. In: YODER, H. S. (ed.) The Evolution of the lgneous Rocks: 50th Anniversary Perspectives, pp. 339-50. Princeton University Press, Princeton, NJ. STEWART, O. F. 1949. Origin and occurrence of vermiculite at Tigerville, S. Carolina. M S Thesis, University of South Carolina, Columbia, SC (unpublished). STONE, C. G. & STERLING, P. J. 1964. Relationship of igneous activity to mineral deposits in Arkansas. State of Arkansas-Ark. Geol. Com., Little Rock. Ar. STRACKE, K. J., FERGUSON, J. & BLACK, L. P. 1979. Structural setting of kimberlites in Southeastern Australia. ln: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites, Diatremes and Diamonds, Proc. 2nd Kimberlite Conf., Vol. 1, pp. 71-91. Am. Geophys. Union, Washington. STRECKEISEN,A. 1967. Classification and nomenclature of igneous rocks--final report of an enquiry. Neues Jb. Mineral. Abh. 107, 144-214. 1980. Classification and nomenclature of volcanic rocks, carbonatites and melilitic rocks. IUGS subcommission on the systematics of igneous rocks. Geol. Rundsch. 69, 194-207. SVISERO, D. P., HASUI, Y. & DRUMMOND, D. 1979a. Geologia de kimberlitos de Alto Paranaiba, Minas Gerais (Geology of kimberlites of the Upper Paranaiba, Minas Gerais). Min. Metall. 42, 34-38. --, MEYER, H. O. A. & HSIAO-MING TSAI 1979b. Kimberlites in Brazil: an initial report. In: BOYD, H. R. & MEYER, H. O. A. (eds) Kimberlites Diatremes, and Diamonds: Their Geology, Petrology and Geochemistry. Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 92-100. American Geophysical Union, Washington, DC. - - , HARALYI,N. L. E. & HASUI, Y. 1984. A note on the geology of some Brazilian kimberlites. J. Geol. 92, 331-8. TAPPE, J. 1966. The chemistry, petrology and structure of the Big Timber complex, Crazy Mountains, Montana. PhD. Thesis, Univ. Cincinnatti (unpubl.). TAYLOR, H. P. 1968. The oxygen isotope geochemistry of igneous rocks. Contrib. Mineral. Petrol. 19, 171. --, GIANETTI, B. & TURI, B. 1979. Oxygen isotope geochemistry of the potassic igneous rocks from the Roccamonfina volcano, Roman comagmatic region. Earth planet. Sci. Lett. 46, 81-106. & TURI, B. 1976. High 180 igneous rocks from the -
-
-
-
I 8 8
S. C.
Bergman
Tuscan magmatic province, Italy. Contrib. Mineral. Petrol. 55, 34-54. & CUNDARI, A. 1984. 180/160 and the chemical relationships in K-rich volcanic rocks from Australia, E. Africa, Antarctica and San Venanzo, Cupaello, Italy. Earth planet. Sci. Lett. 69, 263-76. TAYLOR, S. R. 1979. Lunar and terrestrial potassium and uranium abundances: implications of the fission hypothesis. Proc. lOth Lunar Sci. Conf., 2017-30. -& MCLENNAN, S. M. 1981. The composition and evolution of the continental crust: rare earth element evidence from sedimentary rocks. Phil. Trans. R. Soc. Lond. A301, 381-405. TEMPLEMAN-KLUIT, D. J. 1969. Re-examination of pseudoleucite from Spotted Fawn Creek, West Central Yukon. Can. J. Earth Sci. 6, 55-62. TERZAGHI, R. D. 1935. The origin of the potash-rich rocks. Am. J. Sci. 29, 369. THIBAUT, G., RAOULCARUBA,M. M. & TURCO, G. 1972. Etude exp6rimentale du domaine de formation de la dalyite K2ZrSi6015 et essai de corr61ations p&rologiques. C. R. Acad. Sci. Paris, 274, 792-5. THOENEN, J. R., HILL, R. S., HOWE, E. G. & RUNKE, S. M. 1949. Investigation of the Prairie Creek diamond area, Pike County, Arkansas. U.S. Bur. Mines, Rep. Invest. 45-9. THOMPSON, R. N. 1977. Primary basalts and magma genesis 111. Alban Hills, Roman Comagmatic province, central Italy. Contrib. Mineral. Petrol. 60, 91-108. - 1985. Asthenospheric source of Ugandan ultrapotassic magma. J. Geol. 93, 603-8. , MORRISON,M. A., HENDRY, G. L. & PARRY, S. J. 1984. An assessment of the relative roles of crust and mantle in magma genesis: an elemental approach. Phil. Trans. R. Soc. Lond. Ser. A, 310, 549-90. THY, P. 1982. Richterite-arfvedsonite-riebeckite-actinolite assemblage from MARID dykes associated with ultrapotassic magmatic activity in central West Greenland. Terra Cognita, 2, 247-9. - 1985. Contrasting crystallization trends in ultrapotassic lamproites from central W. Greenland. Geol. Assoc. Canada Progm. Abstr. 10, A63. TIDMARSH, W. G. 1932. The Permian lavas of Devon. Q. J. Geol. Soc. Lond. 8 8 , 712-3. TIKHONENKOV,I. P., KUKHARCHIK,M. V. & PYATENKO, Yu. A. 1960. Wadeite from the Khibiny Massif and the conditions of its formation. Dokl. Akad. Nauk. S.S.S.R. 134, 920-3. TINGEY, R. J. 1981. Geological investigations in Antarctica 1968-1969: The Pryce Bay-Amery Ice Shelf-Prince Charles Mountains area. Aust. Bur. Miner. Resources Geol. Geophys. Rep. 1981/34. - - , McDOUGALL, I. & GLEADOW, A. J. W. 1983. The age and mode of formation of Gaussberg, Antarctica. J. geol. Soc. Aust. 30, 241-24. TRAIL, D. S. 1963. The 1961 geological reconnaissance in the Southern Prince Charles Mountains, Antarctica. Rec. Bur. Miner. Resources Geol. Geophys. Aust. 1963-155 (unpublished). TROGER, W. E. 1935. Spezielle petrographie der Erup-
tivgesteine, Deutschen Mineralogischen Gesellschaft, Berlin. TRONNES, R. G., EDGAR, A. D. & ARIMA, M. 1985. A high pressure-high temperature study of TiO2 solubility in Mg-rich phlogopite: implications to phlogopite chemistry. Geochim. cosmochim. Acta, 49, 2323-9. TUREKIAN, K. K. & WEDEPOHL, K. H. 1961. Distribution of the elements in some major units of the earth's crust. Geol. Soc. Am. Bull. 72, 175-91. TURI, B. & TAYLOR,H. P. 1976. Oxygen isotope studies of potassic volcanic rocks of the Roman Province, central Italy. Contrib. Mineral. Petrol. 55, 1-31. TURNER, F. J. & VERHOOGEN, J. 1960. Igneous and Metamorphic Petrology, pp. 235-56. TUTTLE, O. F. & GITTINS, J. 1966. Carbonatites. WileyInterscience, New York. TWENHOFEL, W. H. & BREMER, W. H. 1928. An extension of the Rose Dome intrusives Kansas. Bull. Am. Assoc. Petrol. Geol. 12, 757-62. ULBRICH, H. H. G. J. & GOMES, C. B. 1981. Alkaline rocks from Continental Brazil. Earth Sci. Rev. 17, 135-54. UNESCO 1976. Geological Worm Atlas., CHOUBERT, G. & FAURE- MURET, A. (compilers), UNESCO, Paris. UPTON, B. G. J. & THOMAS, J. E. 1973. Precambrian potassic ultramafic rocks: South Greenland. J. Petrol. 14, 509-34. USSING, N. V. 1912. Geology of the country around Julianehfib, Greenland. Meddr. Gronland. Geosci. 38, 1 146. VAN KOOTEN, G. K. 1980. Mineralogy, petrology and geochemistry of an ultrapotassic basaltic suite, Central Sierra Nevada, California, U . S . A . J . Petrol. 21, 651-84. - 1981. Pb and Sr systematics of ultrapotassic and basaltic rocks from the Central Sierra Nevada, California. Contrib. Mineral. Petrol. 76, 378-85. - & PECK, D. L. 1977. Highly potassic phonolite of Pliocene age near Merced Peak, central Sierra Nevada, California. Geol. Soc. Am. Abstr. Progm. 9, 1208-9. VEEVERS, J. J. 1981. Australian rifted margins. In: SCRUTTON,R. A. (ed.)DynamicsofPassive Margins. Geodynam. Ser. Vol. 6, Amer. Geophys. Union Washington D.C. pp. 72-89. VELDE, D. 1967. Sur un lamprophyre hyperalcalin potassique : la minette de Sisco (ile de Corse). Bull. Soc. Fr. Minbral. Cristallogr. 90, 214-23. 1968. A new occurrence of priderite. Mineral. Mag. 36 867-70. 1969a. Micas from lamprophyres; kersantites, minettes and lamproites. Bull. Soc. Fr. Min&al. Cristallogr. 92, 203-23. - 1969b. Minettes et kersantites. Une contribution fi l'&ude des lamprophyres. Th&e de Doctorat d'Etat, Facult6 des Sciences, Univ Paris VI (unpublished). 1971. A note on an analcite-bearing lamproite from Devonshire. Geol. Mag. 108, 201-4. -1975. Armalcolite-Ti-phlogopite-diopside-analcite-bearing lamproites from Smoky Butte, Garfield County, Montana. Am. Mineral. 60, 566-73.
Lamproites and K-rich igneous rocks VENTURELLI, G., BALESTRA, G., TOSCANI, L. & CAPEDRI, S. 1985. The ultrapotassic rocks and their geological setting. Geol. Assoc. Can. Progm. Abstr. 10, A65. - - , CAPEDRI, S., DI BATTISTINI,G., CRAWFORD, A., KOGARKO, L. N. 8s CELESTINI, S. 1984a. The ultrapotassic rocks from SE Spain. Lithos, 17, 3754. - - , THORPE, R. S., DAL PIAZ, G. V., DEE MORO, A. & POTTS,P. J. 1984b. Petrogenesis ofcalc-alkaline, shoshonitic and associated ultrapotassic Oligocene volcanic rocks from the north western Alps, Italy. Contrib. Mineral. Petrol. 86, 209-20. VERWOERD, W. J. 1966. Fenitization of basic igneous rocks. In: TVTTLE, C. F. & GITTINS, J. (eds) Carbonatites, pp. 295-307. Wiley-Interscience, New York. VILA, J. M., HERNANDEZ, J. 8s VELDE, D. 1974. Sur la pr6scence d'un ilion de roche lamproitique (trachyte potassique fi olivine) recoupant le flysch de type (Guerrouch entre Azzaba (ex-Jemmapes) et Hammam-Meskoutine dans l'Est du Constantinois (Algerie)). C. R. Acad. Sei. Paris, Skr. D, 278, 2589-91. VOLLMER, R. 1976. Rb-Sr and U-Th-Pb systematics of alkaline rocks; the alkaline rocks from Italy. Geochim. cosmochim. Acta, 40, 283-95. 1977. Isotopic evidence for genetic relation between acid and alkaline rocks in Italy. Contrib. Mineral. Petrol. 60, 109-18. 8s HAWKESWORTH, C. J. 1980. Lead isotopic composition of the potassic rocks from Roccamonfina (South Italy). Earth planet. Sci. Lett 47, 91101. 8s NORRY, M. H. 1983a. Unusual isotopic variations in Nyiragongo nephelinites. Nature, Lond. 301, 141-3. & -1983b. Possible origin of K-rich volcanic rocks from Virunga, East Africa, by metasomatism of continental crustal material: Pb, Nd & Sr isotopic evidence. Earth planet. Sci. Lett. 64, 37486. - - - , OGDEN, P. & SCHILLING, J. G. 1981. Nd isotope geochemistry of Leucite Hills ultrapotassic volcanism. EOS, 62, 431. , KINGSLEY, R. H. 8s WAGGONER, D. G. 19'84. Nd and Sr isotopes in ultrapotassic volcanic rocks from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 87, 359-68. VYALOV, O. S. & SOBOLEV, V. S. 1959. Gaussberg, Antarctica. Int. Geol. Rev. 1, 30-40. WADE, A. 1937. The geological succession in the west Kimberley district of western Australia. Rep. Aust. Assoc. Adv. Sci. 73, 93. & PRIDER, R. T. 1940. The leucite-bearing rocks of the West Kimberley area, Western Australia. Q. J. geol. Soc. Lond. 96, 39-98. WAGNER, C 8s VELDE, D. 1986a. The mineralogy of Krichterite-bearing lamproites. Am. Mineral., 71, 17-37. & -1986b. Lamproites in N. Vietnam? A reexamination of cocites. J. Geol. 94, 770-6. WAGNER, P. A. 1914. The diamond fields of Southern Africa. Transvaal Leader, Johannesburg. -
-
-
-
189
WAGNER, H. C. 1954. Geology of the Fredonia Quadrangle Kansas. U.S. Geol. Surv. Map GQ49. WALDMAN, M. A., MCCANDLESS, T. E. & DUMMETT, H. T. 1985. Geology and mineralogy of the Twin Knobs # 1 lamproite, Pike County, Arkansas. Geol. Soc. Am. Abstr. Progm. 17, 196. WALKER, K. R., 8s MOND, A. 1971. Mica lamprophyre (alnoite) from Radok Lake, Prince Charles Mountains, Antarctica. Aust. Bur. Miner. Resources Geol. Geophys. Ree. 1971/108 (unpublished report). WASHINGTON,H. S. 1917. Chemical analysis of igneous rocks, published from 1884-1913 inclusive. Prof. Pap. U.S. Geol. Surv. 99. WATERS, A. C. 1955. Volcanic rocks and the tectonic cycle. Geol. Soc. Am. Spec. Pap. 62, 703-22. WATKINSON, D. H. & CHAO, G. Y. 1973. Shortite in kimberlite from the Upper Canada gold mine, Ontario. J. Geol. 81, 229-33. WEDEPOHL, K. H. 8s MURAMATSU, Y. 1979. The chemical composition of kimberlites compared with the average composition of three basaltic magma types. In: BOYD, F. R. & MEYER, H. O. A. (eds) Diatremes and Diamonds: Their Geology, Petrology and Geochemistry. Proc. 2nd Int. Kimberlite Conf., Vol. 1, pp. 300-12. American Geophysical Union, Washington, DC. WEED, W. H. & PIRSSON, L. V. 1895. Highwood Mountains of Montana. Geol. Soc. Am. Bull. 6, 389-422. 8s - 1896. Missourite, a new leucite rock from the Highwood Mountains, Montana. Am. J. Sci. 2, 315-23. WELLMAN, P. 1973. Early Miocene potassium-argon age for the Fitzroy lamproites of Western Australia. J. Geol. Soc. Aust. 19, 471-4. , CUNDARI,A. 8s McDOUGALL, J. 1970. Potassiumargon ages for leucite-bearing rocks from New South Wales, Australia. Proc. R. Soc. N.S. W. 103, 103-7. WENDLANDT, R. F. 1977a. The system KA1SiO4-SiO4SIO4-CO2: phase relations involving forsterite, enstatite, sanidine, kalsilite, leucite, vapor, and liquid to 30 kbar. Carnegie Instn. Wash. Yb. 76, 435-41. 1977b. The system K20-MgO-A1203-SiO2H 2 0 - C O 2 : stability of phlogopite as a function of vapor composition at high pressures and temperatures. Carnegie Instn. Wash. Yb. 76, 441. 8s EGGLER, D. H. 1980a. The origins of potassic magmas: I. Melting relations in the systems KA1SiO4-Mg2SiO~-SiO2 and KA1SiO4-MgOSiO2-CO 2 to 30 kb. Am. J. Sci. 280, 385-420. & -1980b. The origin of potassic magmas. II. Stability of phlogopite in natural spinel lherzolite and in the system KA1SiO4-MgO-SiO2H20-COz at high P and T. Am. J. Sci. 280, 421-45. & -1980c. Stability of sanidine+forsterite and its bearing on the genesis of potassic magmas and the distibution of potassium in the upper mantle. Earth planet. Sci. Lett. 51,215-20. WICKHAM, J., ROEDER, D. & BRIGGS, G. 1976. Plate tectonics models for the Ouachita foldbelt. Geology, 4, 173-6. -
-
-
-
-
-
-
-
S. C. Bergman
190
WILLIAMS, H. 1936. Pliocene volcanoes of the NavajoHopi country. Geol. Soc. Am. Bull. 47, 111-72. , TURNER, F. J. & GILBERT, C. M. 1954. Petrography: an introduction to the study of rocks in thin sections. Freeman, San Francisco. WILLIAMS, J. F. 1891. The igneous rocks of Arkansas. Geol. Sum. Arkansas, Report II. WILSON, A. N. 1982. Diamonds, from Birth to Eternity, Gemolog. Institute of America, Santa Monica, CA. WIMMENAUER, W. 1973. Lamprophyre, semilamprophyre and anchibasaltic dike rocks. Fortschr. Mineral. 51, 3-67. -1974. The alkaline province of central Europe and France. In: SORENSEN,H. (ed.) The Alkaline Rocks, pp. 238-71. Wiley, New York. WITKIND, I. J. 1969. Clinophyroxenes from acid, intermediate and basic rocks, Little Belt Mountains, Montana. Am. Mineral. 54, 118-38. - 1973. Igneous rocks and related mineral deposits of the Barker Quadrangle, Little Belt Mountains, Montana. Prof. Pap. U.S. Geol. Surv. 752, 1-58. WOLFF, J. E. 1938. Igneous rocks of the Crazy Mountains, Montana. Geol. Soc. Am. Bull. 49, 1569-628.
WONES, D. P. 1979. The fractional resorption of complex minerals and the formation of strongly femic alkaline rocks. In: YODER, H. S. (ed.) The Evolution of the lgneous Rocks: 50th Anniversary Perspectives, pp. 413-22. Princeton University Press, Princeton, NJ. WOODS, M. J. 1975. Textural and geochemical features
of the Highwood Mountains volcanics, central Montana. PhD Thesis, University of Montana (unpublished). WYOMING STATESURVEYGEOLOGICALMAP 1955. YAGI, K. & MATSUMOTO,H. 1966. Note on the leucitebearing rocks of Leucite Hills, Wyoming, U.S.A. J. Fac. Sci. Hokkaido Univ., 13, 301-12. - - , ISHIKAWA,H. & KOJIMA, M. 1975. Petrology of a lamprophyre sheet in Tanegashima Island, Kagosshima Perfecture, Japan. J. Jap. Assoc. Min. Petrol. Econ. Geol. 70, 213-24. YARZA, R. A. 1895. Roca eruptiva de Fortuna (provincia de Murcia). Bol. Comm. Mapa Geol. Espana, 20, 349-53. ZARTMAN, R. E. 1977. Geochronology of some alkalic rock provinces in eastern and central United States. Ann. Rev. Earth planet. Sci. 5, 257-86. - - , BROCK, M. R., HEYLE, A. V. & THOMAS, H. H. 1967. K-Ar and Rb-Sr ages of some alkalic intrusive rocks from central and eastern US. Am. J. Sci. 265, 848-70. ZHURAVLEYA, L. N., YURKINA, K. V. & RYABEVA,Y. G. 1978. Priderite, first find in the USSR. Dokl. Akad. Nauk S.S.S.R., Earth Sci. 239, 141-3. ZINGG, A., HUNZIKER, J. C., FREY, M. & AHRENDT, H. 1976. Age and degree of metamorphism of the Canavese Zone and the sedimentary cover of the Sesia Zone. Schweiz. Mineral. Petrogr. Mitt. 56, 361-75. ZIRKEL, F. 1870. Untersuchungen fiber die Mikroskopische Zusammensetzung und Struktur der Basaltgesteine, Bonn.
STEVENC. BERGMAN,ARCO Resources Technology, Exploration and Production Research Center, 2300 W. Piano Pkwy, Piano, Tx 75075, U.S.A.
The nature and origin of lamprophyres: an overview Nicholas M. S. Rock S U M MA R Y: Lamprophyres are 'alkaline rocks' because they carry high alkalis at a given percentage of SiOz, together with one or more of normative ne, lc or ac, modal foids, and NaK-Ti-rich amphiboles or pyroxenes. They reach higher whole-rock H20,CO2, Sr and Ba contents than other silicate igneous rocks. Contents of related element subsets in amphiboles (Ti, Ba), K-feldspars (Ba, Fe3), phlogopites (Ti, Ba, Fe 3) and pyroxenes (Ti, A1, Fe 3) include among the highest values known in nature for these minerals. 'Primitive' minerals (diopside, forsterite) commonly coexist with 'evolved' minerals (albite + orthoclase, quartz). Four welldefined 'branches' of the lamprophyre 'clan' have distinctive compositions: calc-alkaline (shoshonitic) lamprophyres (minettes etc.), alone among lamprophyres, have mixed alkalinecalc-alkaline affinities; alkaline lamprophyres (camptonites etc.) are basanitic to nephelinitic and, alone among lamprophyres, usually have Na >>K; ultramafic lamprophyres (aln/Sites etc.) are the most Si poor and Ca rich of silicate igneous rocks, and grade into carbonatites; lamproites (orendites etc.) are uniquely rich in K, Rb, Ba, Th, Mg, Cr and Ni at mainly 'andesitic' SiO2 contents, and carry a suite of diagnostic minerals (wadeite etc.). Each branch comprises at least four rock types which resemble each other much more than rock types of other branches; however, some rock types can be grouped into slightly distinct "families' within one branch (e.g. phlogopitic and madupitic lamproites). A case can be made for including kimberlites as a fifth branch of the lamprophyre clan. Synoptic plots and tables, based on some 5000 major, trace-element and mineral analyses, are presented, to aid identification and classification. Lamprophyres are far more common than generally stated, occurring worldwide in more tectonic settings than many other alkaline rocks and throughout the geological record. They may approximate intratelluric magma compositions. Nearly all represent primitive magmas, and many represent primary magmas. Some represent parental magmas to a wide range of hydrous alkaline intrusive suites: calc-alkaline lamprophyres to potassic pyroxenite-diorite-shonkinite-syenite-granite plutons (Cortlandt); alkaline lamprophyres to hornblendic gabbro-syenite plutons (Monteregian Hills); and ultramafic lamprophyres to ijolite-carbonatite complexes (Fen).
Introduction
Nomenclature
Lamprophyres are the only 'alkaline rocks' covered neither by Sorensen (1974) nor yet in a specialized monograph. Despite Streckeisen's rationalized nomenclature of 1979 and many petrologists' awakening recognition of their global significance, they remain to many others a curiosity (Seeliger 1975) and continue to be treated at best as an 'uncomfortable afterthought' (Hughes 1982) in textbooks. The term 'lamprophyric' often conceals meanings as vague but varied as 'altered', 'biotite-rich' and 'exotic' (Heinrich & Alexander 1975) or even 'not studied in detail'! Yet it is one of few rock names intrinsically conveying its meaning (Greek lampros, porphyros, glistening porphyry or purple rock)--compare locality-based names such as kimberlite, and terms such as granophyre and keratophyre which, sophistically, do not imply porphyritic texture. Furthermore, lamprophyres can often be identified in the field (unlike many fine-grained rocks), and they are among the most widespread of alkaline rocks. This overview is part of a continuing attempt to give them their rightful place in the literature.
Previous confusion perhaps attained its nadir with the Arrow Peak dyke, Montana, variously called 'orthoclase-camptonite' (Rosenbusch 1897), 'minette' (Pirsson 1905), 'leucite-monchiquite' (Beger 1923), 'diopside-lamprophyre' (Knopf 1936) and even 'mafic phonolite' (Buie 1941). A hierarchical classification, expanded in the following ways from Streckeisen (1979), is used here (Fig. 1). 1 As in Rock (1986), Streckeisen's 'melilitic lamprophyres' are renamed 'ultramafic lamprophyres'. 2 As in Rock (1977, 1981), lamproites (the leucite-lamproites of some authors) are regarded as a fourth 'branch' of the lamprophyre 'clan'; they share all the characteristics defined below, and have almost invariably been described as 'lamprophyric' (from Niggli (1923) to Jaques et al. (1984)); the name lamproite, of course, intrinsically implies close affinity with lamprophyres. 3 As in Hughes (1982), and as elaborated in Rock (1987) kimberlites are regarded as a fifth
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline lgneous Rocks, Geological Society Special Publication No. 30, pp. 191-226.
I9I
N. M. S. Rock
I92
m
~
m
~
9
~
g
9
~ ~ ~ "r~ ~ .=.~ ~ ,
Z ~'~
~ ~
._
9
~ -
m-.~
~oo
,-~ ~ . ~
~ 9"-'
9~ - ~ ~ x~ ~_~ ~ ~ o ~.~
oo
~
~
~-
9~
-~-~
~
,-~
.~ ~
o~
~9
~7~: ~ ,~
.~_.~ -~ ~:~.~
,.~
~.
~
~
~l~ F.~ .,,-~
,,.~ ~- ,..~ .c~
.
~.~ o~ ~
~ ~'~
"o .~
<
~
<
.~ ~
~ .~ .~,~
', "O
~
o
~
oL~ ~.~ ~
~
9
~
.,...,
0
~
.
+-4-
.~.~ ,~
.~ ~
e~. ~ . ~
o
=
Nature of lamprophyres branch of the lamprophyre clan (not as a single rock type). Kimberlites must be a subdivision of lamprophyres, and not vice versa, because (a) lamprophyre is a broader 'sack-term' than kimberlite, (b) kimberlites are frequently described as 'lamprophyric' but few lamprophyres are 'kimberlitic', (c) kimberlite is a locality-based name, theoretically with a type-rock, and lamprophyre is more suitable as a clan term, having no type-rock, and (d) lamprophyre has historical precedence. Careful use of this terminology should, in time, overcome much of the lamprophyre 'mystique', provided that (i) the very great differences between the four branches are no longer ignored and (ii) the term 'lamprophyre' is mainly restricted to field use and the rock names in Fig. 1 employed in petrological descriptions. Both changes are necessary because 'lamprophyre' alone covers an inordinate range of compositions from ultramafic to mesotype, ultrabasic to intermediate, Na rich to K rich, melilite to quartz bearing, and peraluminous to perpotassic (Table 1). It is thus even more petrologically vague than 'diabase'. The 'simple descriptive terminology' (biotite-pyroxene-lamprophyre etc.) advocated by Knopf (1936) and Currie (1976) is no less ambiguous for similar reasons.
Purpose of this overview Rock (1977, 1984, 1986) has specificallyreviewed alkaline (AL), calc-alkaline (CAL) and ultramafic (UML) lamprophyres. Gupta & Yagi (1980), Mitchell (1985), Bergman (1987) and Dawson (1987) have covered lamproites (LL). The aim of the present overview is to identify features which distinguish (1) all lamprophyres from 'common' igneous rocks and (2) the different branches from each other. It also indicates how lamprophyres may be fundamental for our understanding of other alkaline rocks. Heterogeneities do exist within each lamprophyre branch. For example, several chemically distinct groups of minettes can be distinguished (Rock 1980, 1984). However, differences within branches are much less than between branches (Metais & Chayes 1964). Moreover, many unimodal (approximately normal or log-normal) element distributions are revealed within each branch by, for example, (i) histogram shapes (Figs. 2 and 7) and (ii) various statistical tests such as kurtosis value, equality of mean and median, and Fillibens's r (Lister 1982). Because such distributions are characteristic of single rock types (LeMaitre 1982), each branch can legiti-
I93
mately be considered as a single petrological entity, at least for comparison with the other branches. Individual rock types will therefore only occasionally be considered hereafter.
Criteria for identification of lamprophyres as a clan A hierarchy of criteria, in roughly decreasing order of importance, is outlined below and explained in later sections; more specific criteria for each branch have been described by Rock (1977, 1984, 1986). Mineral assemblage
Essential phases should include amphiboles, biotites-phlogopites and, often, minerals rich in F, C1, SO3, CO2 and H20 such as halides, carbonates, sulphates and zeolites. Compositional fields of common minerals are delineated in Tables 4-7 and Figs 2-4. High Ba, Fe 3, A1 or Ti in K-feldspars, amphiboles, ph!ogopites and pyroxenes (as appropriate) are characteristic. Mg-rich mafic minerals commonly coexist with alkali feldspars and even with quartz. Cognate minerals absent from all lamprophyres include olivines with mg ~<75, Fe-Mg-Mn- and tremolitic amphiboles (sensu Leake 1978), muscovite, orthopyroxenes, pigeonite and wollastonite. Minerals additionally absent from each individual branch are detailed in Table 1. Texture
The rock should be mesotype to hypermelanocratic and porphyritic (preferably panidiomorphic) with no feldspar or quartz phenocrysts. Feldspathoid or melilite phenocrysts and groundmass olivines are almost always absent. Felsic globular structures and pseudohexagonal or 'castellated' (battlemented) biotites (often with deepbrown rims and pale cores) are characteristic. Most fine-grained mafic dykes with abundant macrocrysts or phenocrysts of biotite or amphibole are lamprophyres (though many lamprophyres only have groundmass amphibole or biotite). Mode of occurrence
The rock should form a dyke, sill, plug, stock, (sub)-volcanic vent or diatreme, or, exceptionally, a margin to a larger intrusion. Associated breccias are very common. Intrusions more than a few square kilometres in outcrop remain unconfirmed. Only a few examples of lamprophyre lavas and pyroclastics are at present known (for
N. M. S. Rock
194
9-
~
~'
~~ss
-.~
.
't
&
=
-=
z
'"
~
~
-~ = . ~
.~=
.~
-~ o
"s"
"~'~
~
~- ,. ~ . -
,.
o ~ = ~ ~v~oooo8
'"-"-~o=
.-'~
~.~<<0~0
~.~.~.
i~
,.4:b
& ~ o
~
~
v
~
~
o~
.-
e
g
gg
gg
o~
~o
oo
~
.
, ~ ~
.
.
~ - - ~ - .-9~
~
.
."t-~ o
.~.~
..~
'
o~
~
.~,,.-~o
~
~
v v ,.u o -u
~
~e~
=o=o = ~ ~ "~
'~
~
~
=o~
b ~..u. ~ ,-.. = = '-2 ~ '-'
~ ,
3
"~
-~
~
~~
o ~,
"~
-~
u ~
"~
=
~
o
.~:
.~
r
<
e ~ " ~ ~
~'=- ~ "~~ "-'~
~
~.~
==
~
-~
~,.,.,~
~
o ~
- ~
z~
~ =~ ~ ' ~o- -~ u ~ ~ o .~- ~ =_~ ~ = I~ ~ ' s
~ e,-_~ ,,1=1 ~
Nature of lamprophyres minettes and lamproites). Lamprophyre dykes are o f t e n c h a r a c t e r i z e d b y s e g m e n t e d f o r m , offsetting, a n d b a n d e d or z o n e d i n t e r n a l s t r u c t u r e s i n d i c a t i n g flow d i f f e r e n t i a t i o n , m u l t i p l e i n t r u s i o n etc. ( U p t o n 1965; C u r r i e & F e r g u s o n 1970; D e l a n e y & P o l l a n d 1981).
Whole-rock composition C o m p o s i t i o n a l fields are set o u t in T a b l e s 8 - 1 0 a n d F i g s 5 - 1 0 , a n d c h e m i c a l s c r e e n s are g i v e n b y R o c k (1977, 1984, 1986); n o o v e r a l l l a m p r o p h y r e s c r e e n is feasible. H i g h levels o f HEO, C O 2 , F, C1, alkalis, Sr, T h , P, B a a n d light r a r e - e a r t h e l e m e n t s ( R E E ) , b u t n e a r - b a s a l t i c levels o f Y, Ti, h e a v y R E E a n d Sc are c h a r a c t e r i s t i c . H i g h H 2 0 a n d COE d o n o t n e c e s s a r i l y i m p l y a l t e r a t i o n o w i n g to high m o d a l c o n t e n t s o f b i o t i t e , a m p h i bole and carbonates.
195
a s s o c i a t i o n s ( T a b l e 2). ' L a m p r o p h y r e s ' f r o m o u t s i d e t h e s e a s s o c i a t i o n s (e.g. D i e t r i c h s o n 1955) n e a r l y a l w a y s v i o l a t e t h e p r e s e n t definitions.
Miscellaneous features A l t e r a t i o n is c h a r a c t e r i s t i c a l l y s e l e c t i v e : if m a f i c ( p h e n o c r y s t ) m i n e r a l s a r e altered, felsic ( g r o u n d m a s s ) m i n e r a l s are o f t e n fresh, a n d v i c e versa. Lamprophyres often exfoliate characteristically, w e a t h e r i n g to a r e d m i c a - r i c h soil. A b u n d a n t a n d v a r i e d x e n o l i t h s a n d x e n o c r y s t s ( T a b l e 12) a r e characteristic (although not diagnostic).
Identification and nomenclature of rock types closely related to lamprophyres Volatile-poor relatives of lamprophyres
Association R o c k s s a t i s f y i n g t h e f o u r p r e v i o u s c r i t e r i a are r e s t r i c t e d to a f e w w e l l - d e f i n e d p e t r o l o g i c a l
TABLE 2.
S o m e A L g r a d e into b a s a l t i c r o c k s o f s i m i l a r b u l k c o m p o s i t i o n , v i a d e c r e a s e o f [ a m p h i b o l e + biotite], loss o f g l o b u l a r t e x t u r e and, o f t e n , a p p e a r -
Summary of known lamprophyre associations Calc-alkaline lamprophyres All rock types
Alkaline lamprophyres
Mainly Mainly spessar- minettes rites
All rock types
Ultramafic Lamproites lamprophyres
Mainly monchiquites
Associations Post-orogenic calc-alkaline granitoid plutons Shoshonitic suites Appinite-breccia pipecomplexes Kimberlites Alkaline rock-carbonatite complexes Alkaline syenite-gabbro complexes Regional dyke-swarms unrelated to central magmatism Localized isolated magmatism
x x x x
. --
---
x __
--
.
.
.
-.
(x) .
.
. --
--
9
__
. --
?
--
?
--
x x
x x
--
--
--
x x
--
--
__
9
_
x x
--
--
(x )
--
(x )
(x)
--
X
X X
-x 9 x x x x
-x 9 x x x x
-x 9 x x x x
x x • -x x x x
x x • -x x x x
x x x ? x x x x
-? ? o x x x
X
Tectonicsettings Oceanic islands Island arcs Major transcurrent faults Orogenic belts Continentalcratons Rift valleys
Partly adapted and expanded from Wimmenauer (1973a). x x , widespread occurrence; x , minor occurrence; ( x ) , rare occurrence; - - , unknown.
?, unconfirmed occurrence;
196
N. M. S. Rock
ance of plagioclase phenocrysts or groundmass olivines. A comprehensive nomenclature covering such gradations, using terms such as 'camptonitic basalt' has been given by Rock (1983). This could be extrapolated to rock types transitional between UML and 'dry' nephelinitesmelilitites ('aln6itic melilitite' (Rock 1986)) and to rare volatile-poor CAL or LL (e.g. Vaniman et al. 1985).
Mafic and felsic-enriched relatives of lamprophyres Widespread mafic-enriched rock types in lamprophyric suites (Rock 1979, 1983) can be covered by terms such as 'picritic monchiquite' and 'melaspessartite'. Such modified terms are perhaps warranted for melanocratic CAL, AL or LL (cf 'madupite') and for UML with fewer light minerals than dark minerals. Unequivocal felsic relatives of LL and UML are not yet recognized. However, intermediate to felsic rock types are commonly associated with CAL (Wimmenauer & Miiller 1985) and have been termed porphyrites or odinites (microdioritic lamprophyric rocks with plagioclase phenocrysts), semi-lamprophyres or malchites (lamprophyres with differentiation index 20-33 (Wimmenauer 1973b)), acidporphyrites (porphyrites with quartz and K-feldspar) and porphyries. A felsic relative of the Navajo minettes (U.S.A.) has been termed both 'felsic minette' and 'sanidine trachyte' (Bachinski & Scott 1980). Mugearitic-trachytic-phonolitic rocks commonly associated with AL have been termed bostonites, hedrumites, heronites, maenites etc. (Scott & Middleton 1983; Laderoute et al. 1985). Although CAL --* porphyrite ~ porphyry and AL ~ bostonite rock series partly mirror calc-alkaline andesite ~ dacite ~ rhyolite and alkaline basalt ~ trachyte series respectively, textural and compositional peculiarities may justify separate terminology (Rock et al. 1986).
Plutonic and volcanic relatives of lamprophyres As recognized by Streckeisen (1979), lamprophyres are not simply textural variants of common plutonic or volcanic rocks, partly because of their exceptionally high contents of volatiles and incompatible elements (Figs 5 and 7). Only AL broadly correspond to any common igneous rocks, having volcanic relatives among alkali basalts, basanites and nephelinites (Rock 1977) and plutonic relatives among kaersutite-bearing gabbroic rocks (Grapes 1975). Of all silicate igneous rocks, UML reach the lowest SiO2 (down to 15%, where they grade into carbonatites) and highest CaO (up to 30%), while LL attain the highest
KzO (up to 13%) combined with very high MgO (up to 27%) but with 'andesitic' SiO 2 contents (Fig. 6). Neither have recognized volcanic or plutonic relatives. CAL only have close relatives among uncommon 'shoshonitic' rocks (Rock 1984).
Discrimination between the four lamprophyre branches Much confusion has been caused by the absence hitherto of discriminatory tables or plots showing the very real differences between the four branches. The tables and plots given here are based on a computerized data base of some 5000 major, trace-element and mineral analyses (Rock 1977, 1984, 1986); this only compiles analyses of rocks called 'lamprophyres' (or one of the rock names in Fig. 1) which actually comply with the present definitions. Through the kind offices of Dr. S. C. Bergman, data have been exchanged between the present data base and his own independently compiled data base. The paucity of analyses contained in only one data base prior to combination suggested that both now contain the vast majority of available lamprophyre data to mid-1986. Statistics presented here and by Bergman (1987) are based on essentially the same data sets, apart from minor differences in emphasis or definition. Following Metais & Chayes (1964), linear discriminant analysis (Davis 1973; LeMaitre 1982) was found to distinguish both whole-rock and mineral compositions efficiently (Table 3). Plots based directly on discriminant functions, however, have three main disadvantages: (1) to many, they are abstract and arcane; (2) ideally, they have to be recalculated for each significant addition to a data base; (3) multigroup discriminant analysis is best for lamprophyres but is least amenable to simple tests of its effectiveness. Twogroup discriminant analysis was therefore performed pairwise, and Figs 3-6 are based on the oxides found to provide the best overall discrimination. Such plots are necessarily compromises since different oxides are effective in different comparisons (Ti discriminates CAL from UML biotites but not, say, CAL from AL biotites). Nevertheless, many plots approach the maximum efficiencies achieved by formal discriminant functions in Table 3, and they provide a good practical aid to correct classification. Use of all the criteria and plots proffered here requires re-naming of many rocks described in the literature. For example 'central complex
Nature of lamprophyres
197
ob o
o
+
E
=
~,~
~
~
m
,~ ~ ~9
~
o ~
~o !
o
_.e
0 f~
~:~ ~
o
~..~
..n
~,-,
~ ~: o
.
~
,
~
.~.~
~
~.~ ~
~ "~ ~
~-~
~
~
~
:
~
.=
ooo.~ .EEEo
I::~'~ "~" ~ o
~
~
9.- ~ c~ u~
~~
. ~ , ~ , ~ "~
o
.o.o.s
=
o
198
N. M. S. Rock
kimberlites' become 'aillikites' (UML) (Rock 1986), while most 'peralkaline minettes' (Velde 1967; Hall 1982), 'leucite-lamprophyres' and associated 'mica-peridotites' (Ghose 1949; Mukherjee 1961; Sarkar et al. 1980) become lamproites. The present view approximates Rosenbusch's (1897) original lamprophyre concept and attributes much confusion to later reassessments (Knopf 1936). Most spatial associations reported between different lamprophyre branches either are coincidental (Kranck 1953), reflect incorrect naming (cf. Tyrrell & Neilson 1938; Baker 1953), or lack age data to confirm consanguinity (Basta et al. 1985). The only common genetic association and gradation between different lamprophyre branches (Table 2) occurs in carbonatite-alkaline provinces (sensu Tuttle & Gittins 1966). Here some or all of AL (mainly monchiquites), UML and, more rarely, CAL (minettes) and LL may be present (e.g. Arkansas, U.S.A.). Each type generally shows its own distinct composition ( e l Tables 8 and 9), although intermediate varieties do sometimes occur. Outside carbonatitic provinces, associations and petrological gradations between different branches are virtually unknown. A note on lamprophyre-kimberfite relations
Except for very rare wholly mica-free types, kimberlites share most characteristics of lamprophyres, and their average geochemistries are in many respects similar (Fig. 5(b)). Even micapoor types may be sufficiently rich in volatile components to be retained within the lamprophyre clan. Kimberlites and lamproites are the only two primary diamond sources yet confirmed, and they intergrade via olivine-lamproites (Scott Smith & Skinner 1984a, b; Bergman 1987; Dawson 1987). Nevertheless, many apparent associations between lamproites and kimberlites have proved fortuitous (Scott 1979, 1981 ; Scott Smith & Skinner 1984b), and lamproites differ from kimberlites in carrying amphibole, glass, leucite and wadeite or related minerals, with higher whole-rock SiO2 and lower Mg, Fe, Cr and Ni (Atkinson et al. 1984, figs 8-10; Jaques et al. 1984, fig. 10; Scott Smith & Skinner 1984a, figs 6-10; Nixon et al. 1984, figs 1-6). Again kimberlites are similar enough to UML in carbonatite complexes to have been repeatedly confused (Mitchell 1979), even though most examples are readily discriminated (Rock 1986). Kimberlites contrast most with CAL and AL in both mineralogy (feldspar content) and chemistry (Mg and Si contents).
Comparativemineralogy Only 'lamprophyres' which comply with the criteria outlined here and by Rock (1977, 1984, 1986) are considered in the remainder of this paper. Considerable space is given to mineralogy because, in view of the doubtful significance of whole-rock compositions (see below), mineral chemistry may provide the most effective way of identifying and distinguishing different lamprophyres. The aim of the present section is to identify both unusual mineral species and compositional fields of common minerals which are characteristic either of lamprophyres collectively or of individual branches. Features such as complex mineral zoning (Scott 1980) and multiple compositional trends within single rock-suites (Mitchell 1981, 1985) are mostly irrelevant for these purposes, but have been considered by Rock (1977, 1984, 1986) and are briefly cited in Table 1 where particularly significant.
Mineralogical features diagnostic of the whole lamprophyre clan
Lamprophyres carry one or more of the following minerals which link them with alkaline rocks: alkali or Ti-rich pyroxenes and amphiboles, BaTi-phlogopites, Ba-K-feldspars, feldspathoids, melanitic garnet, melilite, monticellite, perovskite and silicates of alkalis with Fe, Ti, Zr, Nb etc. Significant contents of minerals rich in F, C1 and SO3 (Ca-Sr-Ba sulphates (Mitchell 1985; N6mec 1985), fluorite and even scapolite (Laderoute et al. 1985)) may be characteristic. Alkali feldspar assemblages are unusual in three ways for fine-grained sub-volcanic rocks known to cool unusually rapidly (Mauger 1984): (1) all structural types of K-feldspar (from high-sanidine to adularia) can occur (Rock 1984); (2) pure K-feldspar not infrequently coexists with pure albite (Rock 1979; Mauger 1982, 1984); (3) alkali feldspars invariably coexist with MR-rich marie minerals. These peculiarities may reflect either the high volatile fugacities in lamprophyre melts, or the disequilibrium or sub-solidus origin of some mineral assemblages (see below). Other striking peculiarities include (1) highMg and/or Mn ilmenites, (2) high-Zn spinels, (3) high-Ti amphiboles, phlogopites and (often) pyroxenes, and (4) high-Ba K-feldspars, amphiboles and phlogopites. Maximum reported Mg, Mn, Ba and Ti contents are among the highest known for the appropriate mineral (Table 4) (cf Deer et al. 1962, 1978). It may eventually be possible to set limits (more than 0.5% BaO or
Nature of lamprophyres
+l+l+l+l+l+l+l+ll+l+l~
e~ t'q e,q e,~ ~ 0
~ 0
+1 +1 +1 +1 +, +, +1 +1
I
0 ~
o ~o o
+, +1 ~. o o.
I99
+1 +1 +1 +1 +1 +1 +1 +1 + l ~1 +1 +1 ~
-- ~
+1 +~ +1 +1 q +~ +1 +~ +1 +1 +1 +1 o
g o
+, +I +~ +, +I ~ ~, ~, I +, +I ~ ~
.
.
.
.
.
.
.
.
.
oo
+' § +, +, +, ~ ~, ~, I +, +, ~ ~ o ~ _ ~ _ ~ - ~
+i+~+t+i~+t+t+i+iT+~+im~ o
o
§ +~ +~ +~ +~
~s~e
o~ ~
.
I +~ +~ +~ § I N ~
eee~ ~o
~
.9 K,
r
+l+i+i§247
~
+l+l+l+l+i+l+l+l+ll+l+i
I~
~.n.n
-~s+,l
+l+l+li
I +1+1+,
I I I
~+1+1+1+1+1+1+1+1+1+1+1+~
r~ ~
,4 gfl
~
++
o
~
200
N. M. S. Rock
more than 6% TiO 2 in amphiboles; more than 3% BaO or more than 8% TiO2 in phlogopites; more than 1.5% BaO in K-feldspars) above which such contents are diagnostic of lamprophyres. These peculiarities are partly but not wholly linked. For example, high mineral Ti contents contrast strikingly with near-basaltic whole-rock Ti contents (Fig. 5(b)). This implies that most Ti in lamprophyres is carried in silicates, and may explain the unusual Mg-Mn-Zn-rich, relatively Ti-poor oxide phases as a corollary. However, high Ti in phlogopites is believed to reflect physical (high-temperature, high-fO2, low-pressure) conditions (Arima & Edgar 1981; Tronnes et al. 1985), whereas high Ti in pyroxenes and amphiboles is believed to reflect chemical (Narich, Si-undersaturated) conditions which apply only to AL and UML. This may explain why AL have high Ti in all mafic silicates, whereas CAL and LL have high-Ti phlogopites but lower-Ti pyroxenes and amphiboles (Table 4; see also below).
40
Calc-alkaline lamprophyres (I0) II I I ~ o II Alkaline Improphyres ~41)
2O
% 2o
2C o
80
84
88
92
mg
Mineral combinations diagnostic of each lamprophyre branch (Table 1)
LL carry a unique suite of K - F e - B a - T i - Z r silicates, including biotite (N~mec 1985), jeppeite, priderite, shcherbakovite and wadeite (Mitchell 1985), which reflect their high K/A1 ratios. Leucite and potassium titanian richterite (sensu Leake 1978) together distinguish LL from other lamprophyres (and all other igneous rocks). Minerals almost confined to UML may include a suite of N a - K - F e sulphides such as bartonite, djerfisherite, erdite and rasvumite (Czamanske et al. 1981). Ti-Zr-garnet, ferri-diopside, melilite, Mg-Mn-spinels, and monticellite distinguish UML from other lamprophyres (Rock 1986). Neither CAL nor AL are known to carry any uniquely diagnostic minerals, but the assemblages [quartz + plagioclase] and [nepheline + plagioclase] respectively distinguish them from other lamprophyres. Differences between cognate minerals common to all lamprophyre branches Olivines (Fig. 2)
Over 200 analyses reveal Fo contents L L > UML > AL. Fo9o + olivines are practically con-
FIG. 2. Distribution of mg (100 Mg/(Mg + Fe 2) at.) ratios in lamprophyre olivines. The numbers of analyses are given in parentheses. The sources are as follows. CAL: Roden & Smith 1979; Allan & Carmichael 1984--10 individual values only, indicated by dots. AL: Rock 1979; McHone & Corneille 1980; Mitchell & Janse 1982; Foley & Malpas 1984; Furnes& Stillman 1987. UML: Walton & Arnold 1970; Kresten & Edelman 1975; Vartiainen et al. 1978; Nikishov et al. 1979; Nixon et al. 1980; Platt & Mitchell 1979, 1982; Foley & Malpas 1984; Meyer & Villar 1984; J. Craven, personal communication. LL: Carmichael 1967; Scott 1979; Kuehner et al. 1981 ; Prider 1982; Scott Smith & Skinner 1984a; Venturelli et al. 1984a; Mitchell 1985. A wider range (without actual data) has been reported by Jaques et al. (1984). The LL histograms are biased by 46 determinations in Scott Smith & Skinner (1984a), but the range is unaffected by excluding these. fined to LL. NiO and CaO contents are generally 0.2%-0.4%, high CaO being noteworthy (Table 4). Clinopyroxenes (Fig. 3(d) )
Nearly 400 analyses cluster around the diopsidesalite-augite boundaries in the pyroxene quadrilateral (Rock 1984, 1986), but A120 3 contents and mg numbers provide reasonable discrimination (Fig. 3(d)). A1203 (like TiO2) contents in AL
FIG. 3. Simple plots discriminating the compositional fields on cognate minerals common to the different lamprophyre branches: (a) biotites; (b) amphiboles; (c) alkali feldspars; (d) clinopyroxenes. Variables are chosen for their high contributions towards linear discriminant functions (Table 3). The sources, which are too numerous to list individually, are available on request. The numbers of analyses are given in Table 4 (total 1200).
Nature of lamprophyres
18-
Calcic
Y
201
Ha~ltlnflslte$.
i
amphiboles
UML§
~
~
u
-~
/
6 Kaersutite
4 I..7.~-.-.-~ 2 14 -
("
':
.....
~.
v
~
~ j,
v
\
i,~
/ : ,
I
Sodic-calcic
o1% o]z 0.4!
•
./'7
" ~-
//
l "
lamproite ',/ L~ (Hall 1982 -~ ~
i/
.hi
~,
"/
,~ 1"
Jrfvedson(teeckermanmtes
/
M .,
l"
4_~i
.,
9
~ - " ~.'7
v
' I' i
v v
~.
LL
/ II 'l
'%~Potassium richterites ..-.~. ~ v ~ .~.~.--'~~ am-phiboles
•
!
v"
I
'
9
0-1 9~
/
'2
"/ 9
-..
....
"
~ v~ i
/
\
n
\
UML'.
~,.
0-04 /
-
I _ ,p
.
.~
,.r I
E;
x~'
,
v ....
40
v
,,~
,;\.
~Iov
0-02 '7
"
70 mg
(a[
<~ M a g n e s i o , ~z r i e - b e c k i t e ~
0~02
vk
0~4
0.1
0 4
I
4
I0
AI20s TiOa
(b)
~vv~
400-
I00
/'
i "v v i ~ !
vy. vv;~
_
:
4~ 9
--
9
2--
t/
./.< O
i
,',/
k
',~
,.J"
pyroxene
% \
~" " " . ~ / l ' ' - i # f ~ l
-
i
0 -
\ ",
~" ~ -
-v-I
• ),,
~x
,~
~'~
, •
0
5O
(c)
Or (mole % )
2 I00
",
I 40
(d)
i
Jumillite lamproRes
" ~ ~ < X l
AL
4 -
AL
/ I
~w r"
I
x>~,x
l
60
"....."
I
l
80 mcj
The individual symbols may represent several overlapping analyses. Total Fe as FeO (except in (c)) owing to large numbers of probe data. Oxides and simple oxide ratios are in weight per cent, and m g numbers are in atomic per cent. Mineral zoning, phenocryst and groundmass variations, and distinct mineral populations are ignored for the purposes of this diagram. Outlying analyses are deliberately excluded from the field-defining lines. Megacrysts and xenocrysts are excluded. Note that the biotite-phlogopite boundary in (a) is only approximate owing to inclusion of Fe203 in the abscissa. Note the logarithmic scales in (b) and (d). In (c) note that alkali feldspars are only an occasional accessory mineral in UML. A few Na-rich pyroxenes are omitted in (d) to preserve scale. Symbols are as follows: O, calc-alkaline lamprophyres (CAL); • alkaline lamprophyres (AL); V , ultramafic lamprophyres (UML); r~, lamproites (LL). For further plots separating CAL and LL minerals, and separating UML from kimberlite minerals, see Rock (1984, 1986).
202
N. M . S. R o c k
and UML pyroxenes include some of the highest known values; some have even been termed 'fassaites' (Durazzo et al. 1984). TiO2 contents vary too widely to provide effective discrimination, but t-tests and M a n n - W h i t n e y tests indicate the order AL > U M L > L L > CAL in central tendency. Na-pyroxenes have been reported in all four branches (Table 4) mainly as groundmass grains or phenocryst rims but occasionally as cores (Mullen & Petty 1985). However, too few analyses are available to draw distinctions. Note that orthopyroxenes are relatively common in a few lamproites but, as in other lamprophyres, they are almost certainly xenocrystic (Mitchell 1985).
Amphiboles (Table 1 ; Fig 3(b)) Over 250 analyses define diverse fields, and the significance of certain gaps between remains unclear. Nevertheless, the following generalizations are warranted (all amphibole names are sensu Leake (1978) and Rock & Leake (1984)). 1 AL amphiboles cover a more restricted range than the other branches; more than 90% of available analyses are 'kaersutites' or 'titanian pargasites', which are also found in some UML. 2 Hastingsitic amphiboles are common t o m a n y CAL, some AL and a few UML. 3 Titanian potassium richterites are unique to LL. 4 Tschermakitic amphiboles are known only from CAL. 5 Ti contents decrease in the general order AL > U M L ~ LL > CAL, but the great majority of lamprophyre amphiboles are titanian. 6 Some amphiboles which have the same names ('arfvedsonite-eckermannite') form more than one compositional group (Fig. 3(b)). 7 Alkali amphiboles are common in LL and U M L and present in some CAL, but are rare in AL; magnesio-riebeckite-arfvedsonites in CAL, UML and LL (Fig. 3(b)) are unreported from AL or indeed from other igneous rocks, being otherwise confined to metasomatic rocks (fenites) and metamorphosed evaporites.
Biotites-phlogopites (Fig. 3(a) ) Over 500 analyses are very diverse. Phenocryst cores and rims, and one or more coexisting groundmass micas, may all have distinct compositions (Nixon et al. 1980, 1985). Synoptic studies of micas (Mitchell 1981, 1985; Bachinski & Simpson 1984) provide further details, but the
following are perhaps the most significant generalizations. 1 A few lamprophyre phlogopites can be matched with those from kimberlites, leucitebearing rocks and mantle-peridotite xenoliths. 2 Nevertheless, the trends of kimberlite and lamprophyre micas are different, especially with respect to A1 and Ti. 3 UML and LL phlogopites often carry extensive tetraferriphlogopite (KMg3Si3Fe3010 (OH)2); LL micas reach the lowest A1 contents known (Fig. 3(a)). 4 Ca-bearing micas have been reported from CAL, AL and U M L (Table 4), but information is insufficient to determine whether this reflects impurities. 5 Only AL micas are commonly biotite (as opposed to phlogopite). 6 Ti contents are in the general order L L > AL > U M L > CAL, although LL/AL and AL/ U M L mean/median values, and all maximum values, are statistically equal.
Feldspars Plagioclases are present only in CAL and AL. Probe and optical data for both cover the range Anl-80, but 'typical' plagioclases are andesineoligoclases. Zoning can reach 50~ An content. Alkali feldspars (Fig. 3(c) and Table 5) are absent from U M L (except in accessory amounts) and from some LL. Sparse data suggest that BaO contents are in the order LL > CAL > AL, with BaO exceeding 1~ for many LL and some CAL K-feldspars. Pure albites have been reported from CAL, AL and UML. The gap in the AL field between Or 1 0 ~ - 3 0 ~ (Fig. 3(c)) may be a sampling effect.
Feldspathoids (Table 5) Analcime can occur in all lamprophyres, ranging from a few modal per cent in some CAL (Roden & Smith 1979) to a dominant phase in some monchiquites (Evans 1901). Most reliable analyses show little deviation from the standard formula, but maximum reported CaO and K 2 0 contents reach 2.14~ and 2.65~ respectively (Furnes & Stillman 1987). Nepheline and sodalite minerals occur only in AL and UML. Analyses are available only for AL (Table 5), where they are richer in Na than the Morozewicz composition but have similar amounts of excess SiO2 (Rock 1978a, 1979). Leucite occurs only in holocrystalline non-ultramafic LL, although analcime and even feldspathic ocelli have been repeatedly misidentified as leucite in other lam-
Nature of lamprophyres
203
TABLE 5. Mean compositions (weight %except for ratios) of feldspathoids, melilite and alkali feldspars in
the different lamprophyre branches Analcime ~
Total analyses4, 5 SiO2 A1203 FeO, MgO CaO Na20 K20 TiO2 (Na + K)/A1 (at.) K/(Na+K) (at.)
Calc-alkaline
Alkaline
Lamproites
1 53.8 23.3 0.79 --10.8 0.45 0.06 0.78 0.026
14 54.1 + 1.68 23.4 ___1.5 0.12+0.08(10) 0.02(2) 0.67 • 0.69,(12) 11.9 +_1.7 0.4 +0.7(12) 0.02+_0.01(4) 0.89 +_0.12(12) 0.02 +-0.004(12)
2 58.74 21,31 0.86 -0.05 9.35 0.32 -? 0.027
Nepheline 2
Melilite3
Alkaline
Ultramafic
21 46.5 _ 1.9 32.4 +_1.0 0.5 +_0.1 0.01(11) 0.80 +_0.24 15.4 +0.4 4.3 +0.9 0.04+_0.06(16) 0.93 + 0.03 0.15+0.03
8 43.4 6.4 2.9 8.8 34.3 3.8 0.1 0.1
Alkali feldspars 6 Calc-alkaline Total analyses4, 5 FeO~ Na20 K20 K/(Na+K)(at.) BaO
Alkaline
26 29 0.72{0.13-2.0}(23) 0.24{0.06-0.71} 2.82{0.27-6.3}(22) 2.9{0.12-10.4} 12.6 {7.0-16.3} 11.8{0.35-16.5} 73.7 {42.5-97.4}(22) 72.7{2.2-98.9} 0.37{0.06-1.2}(8) --
Ultramafic 3 1.73{1.1-2.8} 0.50{0.39-0.57} 15.26{14.8-15.6} 95.2{94.6-96.1} --
Lamproites 47 1.71{0.05-4.3} 0.69{0.02-2.4} 15.2 {11.9-16.8} 93.6 {76.3-99.8} 0.92{0.07-1.9}(26)
1No data for ultramafic lamprophyres; some probe analyses appear to have lost alkalis. 2Absent from calc-alkaline lamprophyres and lamproites; no data for ultramafic lamprophyres. 3Absent from all other lamprophyres. 4 See notes 2 and 4 to Table 4. s Numbers in parentheses indicate the numbers of values where less than the total analyses. 6Figures are means {ranges}; standard deviations are inappropriate given the wide range of Na/K ratios, especially for alkaline lamprophyres.
prophyres (Rock 1984; Mitchell 1985). Pseudoleucite has been confirmed in Greenland lamproites, where it occurs both as macrocrysts and in the matrix (Scott 1981 ; Thy 1985).
doluminescence and isotopic studies are urgently needed before the exact status of lamprophyre carbonates can be assessed. Nevertheless numerous workers (Frenzel 1971) have provided good evidence for primary status.
Carbonates (Fig. 4) Lamprophyres in general carry more carbonates than other silicate igneous rocks; U M L have higher CO2 even than kimberlites (Rock 1986). Probe data have begun to reveal previously unsuspected compositional variety (Fig. 4) including not only the usually reported calcite but also Fe-dolomite and (so far only in U M L and LL) breunnerite. At least four possible origins must be contemplated for such minerals: (1) primary, (2) secondary by carbonation of silicate minerals, and secondary by dolomitization of either (3) primary or (4) secondary calcite (el Le Bas 1984). As primary and secondary carbonates are difficult to distinguish in thin section, catho-
Oxide minerals (Tables 6 and 7) These are rare in LL, where high K/A1 usually causes Fe and Ti to enter priderite and related minerals (priderite itself is not known to crystallize with spinels (Mitchell 1985)). Ilmenites in U M L are unusually rich not only in Mg but also consistently in Mn (Rock 1986). Either or both characteristics may apply to CAL, AL and LL ilmenites, but data are limited (Table 6). Spinels are extremely variable. Phenocryst cores and rims, groundmass spinels, xenocrysts, and primary and secondary spinels in one rock may all be distinct (Nixon et al. 1980; Mitchell
2o4
N. M. S. Rock Co
Dolomite
Mcj
Ankeri'ie
Breunnerite
Fe
FIG. 4. Compositions of carbonate minerals in lamprophyres (Mole per cent) (symbols as in Fig. 3). The sources are as follows. AL: Scott & Middleton 1983; Foley & Malpas 1984. UML: Walton & Arnold 1970; Platt & Mitchell 1982; Foley & Malpas 1984. LL: Hall 1982. No data are available for CAL. Only a few representative calcites are plotted, as most so far analysed are nearly pure CaCO3. 1985). As data are sparse for CAL and AL, only the following tentative generalizations are offered. 1 Some CAL and AL spinels, together with some groundmass spinels in UML, appear to be 'common' igneous types.
2 Some Cr-rich spinels in CAL are similar only to those in boninites among volcanic and subvolcanic rocks (Suzuki & Shiraki 1980). 3 Zn-Mg-chromites in CAL (Wagner & Velde 1985) are similar only to spinels associated with F e - N i sulphide ores; high-Zn spinels are also known from AL (Table 7) and U M L (B. G. J. Upton, personal communication). 4 Some Mg-Cr-rich spinels in U M L and LL overlap with peridotite and kimberlite spinels only. 5 Other U M L spinels are subtly different from kimberlite spinels in both compositional range and element correlations (Mitchell 1979; Scott Smith & Skinner 1984a). 6 As with ilmenites, spinels rich in both Mg and Mn have been found to date only in U M L (Platt & Mitchell 1979, 1982).
Comparative whole-rock geochemistry Major elements
(Tables 8 and 9; Figs 5, 6 and 7)
Figure 6(b) compares lamprophyres with the most recent I U G S alkali-silica classification of volcanic rocks; it also discriminates effectively between the four branches. Apart from a few CAL (mainly kersantites and spessartites), all lamprophyres lie on the alkaline side of the Hawaiian alkaline-tholeiitic divide. In normative composition (Table 8), LL show their unique
TABLE 6. Summary of available ilmenite compositional data for the different lamprophyre branches Calc-alkaline Analyses" 2 1 MgO (~) 0.75-2.6 0.03 MnO (%) -9.17 Source Roden & Wagner & Smith 1979 Velde 1985 Area
Navajo, U.S.A.
Jersey, U.K.
Alkaline 1 1 1.86 0.31 3.00 0.59 Furnes& Scott & Stillman Middleton 1987 1983 Cape Verde Gran, Islands Norway
Lamproites 1 2.80 2.24 Mitchell & Janse 1982 Wawa, Ontario
? <8
3 2.6-3.3
--
1.1-1.9
Jacques et Mitchell al. 1984 1985 W KimJumilla, SE berley, Spain Australia
Ultramafic I Analyses2 3 MgO (%) < 0.4 MnO (~) 0.4-3.2 Source Platt & Mitchell 1979 Area Marathon, Ontario
! 7.3 0.71 von Eckermann 1968 Aln6, Sweden
3 5.1-7.7 3.74-4.45 Vartiainen et al. 1978 Sokli, Finland
3 0.10-0.15 4.7-6.1 Griffin & Taylor 1975 Fen, Norway
1
0.35 2.0 Kresten et al. 1981 Kalix, Sweden
1For plots and compilation of full analyses of UML ilmenites see Rock (1986). 2Most data are for single analyses; ranges are given where more than one analysis is available.
Nature of lamprophyres
20 5
TABLE 7. S u m m a r y o f spinel compositions present in the different lamprophyre branches Area
Reference
No. of analyses
Compositions represented
Calc-alkaline
Colima, Mexico Navajo, U.S.A. Alps, Italy
Allan & Carmichael 1984 Roden & Smith 1979 Wagner & Velde 1985
4 2 3
Ti-magnetite; Mg-chromite Ti-magnetite-ulvospinel Zn-Mg-chromite
Mitchell & Janse 1979 Scott & Middleton 1983 Foley & Malpas 1984 Ohashi 1980 Furnes& Stillman 1987 Rucklidge et al. 1980
2 2 10 1 6 2
Ti-(Mg)-magnetite Ti-(AI-Zn)-magnetite Ti-magnetite Cr-Ti-magnetite Ti-(Mg)-magnetite Ti-magnetite
Malaita, Solomons
Nixon et al. 1980
6
Aland, Finland Aln6, Sweden Ailik, Labrador
Kresten & Edelman 1975 von Eckermann 1968 Foley & Malpas 1984
2 1 10
Fen, Norway
3
Kalix, Sweden Marathon, Canada
Griffin & Taylor 1975; Mitchell 1979 Kresten et al. 1981 Platt & Mitchell 1979
Mg-ulvospinel-magnetite; Cr-ceylonite; Mg-Al-chromite Ti-magnetite Magnetite Ti-magnetite; Mg-A1-Ti-magnetite, MgAl-chromite Ti-magnetite ; Ti-A1-Mg-chromite
Ile Bizard, Canada Tanakh, Siberia
Mitchell 1979 Nikishov et al. 1979
Alkaline
Wawa, Ontario Gran, Norway Aillik, Labrador Setogawa, Japan Cape Verde Islands E Greenland Ultramafic 1
4 17
Magnetite; Ti-magnetite; Ti-picotite Mn-Mg-ulvospinel-magnetite; Ti-AI-Mgchromites; Ti-Al-chromites; Tiulvospinel-magnetites Ti-ulvospinel-magnetites A1-Mg-Cr-Ti-magnetite
Lamproites 2
Leucite Hills, U.S.A. Prairie Creek, Ar, U.S.A. W Kimberley, Australia SE Spain
Carmichael 1967 Scott, Smith & Skinner 1984a; Mitchell 1985 Jaques et al. 1984; Mitchell 1985
Ti-magnetite Ti-Mg-chromite Ti-Mg-Al-chromites; Ti-Mg-chromites; Ti-Cr-magnetite Hercynite; Cr-hercynite; Mg-chromite
Venturelli et al. 1984a; Mitchell 1985
See Rock (1986) for listing of full analyses of UML spinels. 2See Mitchell (1985) for listing of full analyses of LL spinels.
composition by combining high ks (K2SiO3) with abundant ru, and either q or lc. Many U M L are too Si poor and CO2 rich to yield norms, but high la (CazSiO4) is characteristic. Most AL have basanitic norms (i.e. a b + o r + a n + n e + d i ) , although monchiquites often carry lc, ka (KA1SiO4) and la, indicating gradation into foidites and UML. CAL vary from q to ne and from an to ac normative, but this in itself manifests their compositional peculiarity.
Trace elements (Table 10; Figs 5(b), 7-9) Available data are regrettably inhomogeneous-not least in ranges of determined elements and
analytical methods. Basic statistics are compared in Table 10, and spidergrams, whose striking similarities clearly support the concept of a lamprophyre 'clan', are compared in Fig. 5(b). The following are the most significant generalizations. 1 Lamprophyres are linked together and with alkaline rocks by their high average contents of low-field-strength incompatible elements K, Rb, Ba and Sr, but near-basaltic levels of many highfield-strength elements (Ti, Y and heavy REE) and Sc. 2 Negative Nb anomalies are present in LL and especially CAL, but not in AL or UML.
2o6
N. M. S. Rock
TABLE 8. Mean major element composition and range in normative character o f the different lamprophyre branches Calc-alkaline lamprophyresl
Alkaline lamprophyres'
Ultramafic lamprophyres
Lamproites 2
Total analyses ~' 3 754 563 245 293 Major elements (wt. ~ ) (means _ standard deviations (numbers of values where less than the total number of analyses) 3 SiO2 51.5 +4.0 (706) 41.9 +3.6 (561) 30.9 +5.8 49.5 +7.4 A120 3 14.0 +2.3 (679) 13.7 +2.7 (561) 6.8 +3.0 8.5 +2.5 Fe20 3 3.7 +2.3 (723) 5.3 +2.8 (560) 7.9 +4.5 3.7 +2.0 (257) FeO 4.9 + 1.8 (640) 6.5 +2.0 (527) 7.1 +2.6 (229) 4.1 +2.3 (229) FeO, 7.4 +1.9 (751) 10.9 +2.2 13.9 +4.3 6.1 +1.3 MgO 6.9 +2.6 (751) 7.2 +3.0 15.1 __+4.7 11.4 -t-5.6 CaO 6.6 +2.2 (728) 10.6 +2.4 15.0 -+4.8 5.4 +3.0 NazO 2.7 _1.0 (727) 3.2 _1.3 1.1 +1.1 1.3 +0.9 KzO 3.8 +2.0 (731) 2.3 +__1.1 1.8 __+1.0 7.0 +2.6 H,O § 2.7 -t-2.0 (640) 3.6 __+2.3 (513) 3.8 +3.5 (185) 2.8 ___1.9 (241) TiO2 1.3 +0.6 (713) 3.0 +_1.1 (558) 3.6 +1.7 2.7 + l . 5 P2Os 0.71+0.52(622) 0.85+0.47(520) 1.2 -t-0.8 (239) 1.1 +0.5 MnO 0.15+0.14(630) 0.21 +0.09(499) 0.26+0.17(235) 0.09+0.04(284) CO2 2.2 +2.4 (311) 2.7 +2.4 (309) 8.2 +6.9 (162) 3.0 +3.0 (137) Atomic ratios (means + standard deviations (numbers of values where less than the total number of analyses)) (Na+K)/AI 0.62+0.19(678) 0.57+0.14(561) 0.60+0.35 1.17+0.29 K/(Na+K) 47.5 +18.9 (727) 32.6 +14.6 57.1 __+21.6 77.0 +14.7 M g / ( M g + F e 2) 69.7 +_11.1 (640) 64.5 +9.6 (527) 78.1 +8.5 (229) 83.8 +6.9 (229) Mg/(Mg+Fe2)s ~ 68.7 +10.0 (751) 60.9 +9.2 72.6 +8.3 80.5 +6.9 Normative composition4 (CO, ignored in calculations and not treated as C a C O J Complete analyses s 675 561 1996 Percentage of complete analyses with the following character Q> 0 42 * * Hy>0, Q=0 25 3 * OI>O, Ne=Q=O 0 0 * Ne > O, Lc = 0 30 75 82 Lc >O, Ne=O 0 0 0 Ne>0, Lc>0 3 21 14 Ka>O * 18 85 Ka > 0 * 10 79 Ge, Ak>O * * 15 Cm>O 10 * 1 An>O, Cm=O 86 I" 97 Ac>O, Ns=O 2 * 3 Ac>O, Ns>O < 1 * < 1 Ac>O, Ks>O * * 0 Ac >O, Ns> O, Ks >O * * 0 Wo>O <1 5 0 Hm>O 10 7 0 Ru > 0 3 6 0
293 43 21 <1 10 18 8 9 * * * 21 5 20 9 44 1 14 38
* Zero by definition. ~" 100% by definition. 1As an indication of a comparability of pre- and post-1930 analyses, calc-alkaline and alkaline lamprophyre averages were calculated for (a) all available data and (b) post-1930 data only (total analyses 683 and 487 respectively. These two sets of means differed at most by only 1-2 in the first decimal place and are not therefore quoted separately. -"Mean presented merely for completeness and should be treated with extreme caution owing to large compositional differences between the different lamproite families (Bergman 1987). 3 See notes 2 and 4 to Table 4. All analyses screened for conformity to definitions in Streckeisen (1979) and Rock (1977, 1984, 1986). No analyses of tufts, breccias or pyroclastic rocks are included in the statistics. Fe normalized to Fe/total Fe (cation)= 0.30 (Rock 1986). 5 Remainder of 'total analyses' lack a value for one or more of the major oxides essential for calculating a CIPW norm (e.g. SIO2); most are partial analyses with only one to two oxide values. 6The remaining 46 of the 245 complete analyses are too Si poor and CO: rich to calculate CIPW norms. With CO2 ignored, there is insufficient SiO, to form even gehlenite molecule, but COz is too high to be calculated as CaCO 3.
Nature o f lamprophyres
207
TABLE9. Major-element (wt. %) contrasts between contemporaneous lamprophyres of different branches
within the same alkaline province or complex Aillik Bay complex, Labrador, Canada (Foley & Malpas 1984) Lamprophyre branch SiO2 A1203 FeOt MgO CaO Na20 K20 TiO2
AL 2 (19)
UML 2 (7)
33.5+2.4 22.9+2.7 8.3___2.3 3.4-t-1.3 12.5___1.5 10.9___1.1 8.5+1.9 16.0+3.0 15.0+2.1 17.8+2.3 2.7__ 1.1 0.6+0.4 2.0___0.6 2.1+0.5 4.5+1.6 3.0+0.6
Damaraland province Namibia, SW Africa (Visser 1964)
Meimecha-Kotui province, Siberia (Egorov 1970)
CAL 3
AL 3 U M L 3
UML 3
AL 3
48.44 13.73 13.66 2.58 5.80 3.10 3.83 2.25
41.16 16.87 10.50 5.30 8.12 3.00 3.00 3.00
38.00 10.69 13.70 10.12 14.76 5.83 2.60 1.10
44.25 13.54 9.42 7.62 9.58 4.13 2.20 2.88
33.28 18.51 16.08 6.02 12.03 0.55 3.02 2.30
Central Arkansas province, U.S.A. (Steele & Wagner 1979; Scott Smith & Skinner 1984a) 1 AL 3 (14)
UML 3
LL 2 (24)
43.2-t-2.7 39.00 49.9___7.9 13.2-t-2.0 14.15 4.8-t-1.0 10.4-t-2.2 11.02 8.6___2.3 7.8-t-3.1 8.80 24.5+7.0 11.8-1-2.9 17.40 4.7___1.3 1.8+ 1.4 3.20 0.4+__0.3 3.0__ 1.3 1.40 3.1_ 1.2 3.4___0.8 3.26 2.6_+0.7
1Most lamprophyres recorded in the literature have been renamed according to the definitions in the present paper. The Prairie Creek 'kimberlite', for example, is a lamproite (Bergman 1987; Scott Srrtith & Skinner 1984a), and most 'ouachitites' are ultramafic lamprophyres and not alkaline lamprophyres. Means include data from several unpublished theses. 2 Mean ___standard deviation (number of analyses in brackets). 3 Single analysis of individual dyke.
tCorb~
y
~
* Kimberlites
400
-
?
_ r
6
/
oo-
-II-
c9 4
x ...... x o-----13 + ....... + 0-----o
/ !
,oo_
CAL AL UML LL
....
& 9'Common'bosic/.~^,/x kT.2.,-,L 0 (o)
igneous rocks 1 _- __...~....-~Averoge igneous rock " ', I'. i I ~ 2 4 H20
50
* I
6
,i':'//7-
2 o
~.,3
~,o FIG. 5. Comparison of lamprophyre whole-rock compositions with other igneous rocks. Fields of individual analyses would cover much of the plots, so only mean values are shown. The lamprophyre means are taken from Tables 8 and 10. (a) Volatile contents. Shaded field based on LeMaitre's (1976) means for various common basic igneous rocks. Symbols as Fig. 3. Various kimberlite means from Dawson (1980) and Bergman (1987). (b) MORB-normalized 'spidergrams" using the scheme of Pearce (I 982, 1983). Kimberlite (KIMB) and alkali basalt (AOB) means, from Bergman (1987).
*
\",$,, x'l-Z;,
l
t%, I/ x--'~,7*, 7. 0"5 I
I
I
I
I
I
I
I
I
I
I
I
k~"l
I
Sr K Rb Ba Th Nb Ce P Zr Hf Sm Ti Y Yb Sc Cr (b)
208
N. M. S. Rock
!~
(b)
21 (a)
]
i'!'
-
-
r
20 (c)
MgO
40 Si02
60
o
(d) AI203
CoO 9
2b
40
SiOe
60
SiOJlO
TiO z • 4
FIG. 6. Simple plots discriminating the whole-rock compositional fields of lamprophyres and comparing them with recent chemical classifications o f ' c o m m o n ' igneous rocks. The numbers of analyses are given in Table 8 (total, 1850). Fields for each lamprophyre branch enclose 95~ of available analyses to avoid outlying and possibly inaccurate data. The means from Table 8 are shown using the symbols of Fig. 3. All oxides are in weight per cent. The andesite classification fields of Gill (1982) are also shown in (a), and the tholeiite-alkaline divide of Macdonald & Katsura (1964) and the recent IUGS proposal for a definitive total alkali-silica classification of volcanic rocks (LeMaitre 1984) are included in (b). Note that this proposal applies to analyses with original H20 and CO, contents less than 2%, recalculated volatile free, and its comparability with lamprophyre data may therefore be limited. The field labelled 'basanite' covers tephrite to tephriphonolite; 'trachandesite' covers hawaiite, benmoreite etc. (c) and (d) are plots providing better discrimination between branches not so efficiently separated by (a) and (b). CAL and LL cannot be efficiently separated by whole-rock composition alone unless five or more chemical variables are used. For further discriminatory plots see Rock (1984, 1986).
Nature of lamprophyres "OJ
209
'CO
I
o
.o
~=~
~
.0
"~
o
= ~
o
~ . ~~
~
o=N~o
~ 9
='~
I
~.~ ~~ 0
~
.o
~176
~
~ z.~ .~ '~o~
I ~
~
~
~
9 ~.
~
i ~ ~
I
o
,~~ _' ~~ ' ~
~ ~ -
.,..,
~
oo.~-o
-
= ~
= ~ , o ~
I o.= -~ o
9
<-----
~
-/
.~@
~,~
,
~g-
-~
,
~
,~ ~ , ~ .~
~=~..
_
9
senlO^ to ,tuaoJ~d
[.~ . .
~,_~
N . M . S . Rock
2 IO
t--I
~,~
~ ~ r-- ~-~ v~ ~., ~ c~ ~- ~ ~ ~ o
,--~ ---~ e,'~ ~,'-~ " - ~ ~ ,
"..~ " ~
c ~ ,~0 ~ -
9 .
,--~ ~:~ " - ~
+l +1 +1 +1 +1 +1 +1 +1 +1 +l +1 +1 +1 +1 +1 +1
~--c~
~
-
~
~
~'~ I I ~ I ~ e"~,-. ~r'~ ~ , - ~ 1 ' - - c,-~ I P ~ , ~ r,~, , ~ 1 - , . ~.~ ~-~ i ~,-~ ~,~ i
~
~"~r,"ll~
I ~,1 ~
~"--~l"~,--~
+ l +1 +1 + l +1 +i +1 +1
~'~
~
~t" v'~ I ~ ~ ~'N r",l ~ " ~ ' ~ " " ~ , A , r c~l i ~,-~ i 0 o'-, - ~',~
0
~ ' ~ 0
~
I
~
C~" c"~ r "
~
r"
r
~'
~
I I I I I ~ ~ ~ e'~O''-~-~'~ . . . . ~'~
~
0
,~"~
~-.-n O0 ~
I' ~
0
o
P" ~ I ~1
+1 +1 +1 -H -I-I -H +1 -H +1 +1 +1 -H +1 -t-I +1 +1 +1 ("~1 ~'~, '~"l o"~ c ~ ("-1 ~ " , ~
'~
"~0
~
,-~ ~,~ i I t I r'-'~t' ' q " ~ ' ~ ' ~ ~'~'~ l---- o o c,~ ,~1- i . . . . .
r'-- ~t' c ~ v-~ ,.-~
C~ Oe
+1 +1 +1 +1 +1 +1
~1~
+1 +1 +1 ~,~ +1 +1 -H -H +1 +1 +1 -H +1 ~'~1 ~'~ O~ ' ~ ' O0 ~'~ C )
~'
+1 +1 +1
~
~---~ I
C~I/
0,-~
I
"~.~,
I
I
I
" ,~
I
I
+1 +1
.,-~ ~.,~ ~-~.
I
I
I ~-.~-,O~
~.-.,
.~, ,
+1
+1 +t +1 +1 +1 +1 +1 +1 + l +1 +1 +1 +1
o.e
~
+1 +1 +1 +1 +1 +1 +i +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 ~ o
.N
" 0
~'.~
E
~
~" ~ , ,
~
~ , ~ q
or--c~
r
I
e,-,r I ~:) ~'~ ~
I c-,I
~'~ I I o
~'~ ,--~ ~1~ ' ~
~r I
~
I ~1~,, 'q'~ I
'~-" t",l e ~ ~ - "
~ o ~ . , ~ ~ ~'~'-~'~" c,,I
i---./v-~ I ~-.~c~,-~,~-,~r " ~ ~,0 ~'1 ~
~ , i
. ~. ~ . .~-. ~. -.
o ~ - " -
~r~r I ,~,-
I
~'~
9. ~
,~g.
+1 +1 +1 +1 +l +1 +1 +1 +1 +i +1 +1 +1 +1 +1 +1 +l +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1
o
e~ e-i
~ ~ . ~ ~ ~,
Nature o f lamprophyres 'li
I000
zI]
Lamproites
Colc-alloline lamprophyres
400
I00 40
._
(a) t
1
I
~
L
I
I
'
I
1
I
I
I
I
i
Alkoline lomprophyres I000
i
x
f
i
1
[
L
i
f
i
i
i
i
I
i
Ultromafic lamprophyres
_
-
"2
\
400
IO0
--
"2
40
'
I
I
I
Lo Ce Pr Nd
I
I
I
I
I
I
I
I
I
I
Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(d)
I Lo
i
t
/
Ce Pr Nd
I
l
|
I
i
I
~
'
1
Sm Eu Gd Tb DyY Ho Er Tm Yb Lu
FIG. 8. Rare-earth profiles (total 150) normalized to the chondrites of Nakamura (1974). Shaded fields and vertical bars indicate ranges. The means are shown alone for invariable spectra. The methods of analysis (where stated) are as follows: neutron activation (INAA); induced coupled plasma (ICP); X-ray fluorescence (XRF): spark-source mass spectrography (SSMS). The sources for (a) calc-alkaline lamprophyres are given in fig. 3 of Rock (1984). The sources for (b) alkaline lamprophyres are as follows: <~, mean and range of six monchiquites/ camptonites, Monchique, Portugal (Rock 1978a, 1979) (SSMS); Q, mean of three camptonites, Spanish Peaks, U.S.A. (Jahn et al. 1979) (INAA); ]~, mean and range of 18 camptonites, Sunnhordland, Norway (Faerseth et al. 1976) (XRF) (La to Nd only); r-q, one monchiquite, Hopi, AZ, U.S.A. (Nicholls 1969) (XRF) (La to Sm only): ,, one camptonite, Girnar, India (Paul et al. 1977) (INAA); ~-, two monchiquites, Callander Bay, Canada (Cullers & Medaris 1977) (INAA); ~I,, monchiquite (misnamed 'aln6ite'), Oka, Quebec (Eby 1975) (XRF); x, camptonite, Eilean Donain, Scotland (Morrison 1980) (INAA). The sources for (c) lamproites are as follows: 9 'peralkaline minette' (lamproite), Pendennis, England (Hall 1982) (ICP); x, three Leucite Hills lavas, WY, U.S.A. (Kay & Gast 1973) (IN AA) ; ~, mean and range of 48 lamproites, Holsteinsborg, Greenland (Scott 1979) (XRF) (La and Ce only); ~, mean and range for four fitzroyite family lamproites, NW Australia (Nixon et al. 1984) (ICP); ornamented field, range for 15 jumillite family lamproites, SE Spain (N ixon et al. 1984) (ICP). The sources for (d) ultramafic lamprophyres are given in fig. 6 of Rock (1986). Other data (Thompson et al. 1984) are omitted, as they substantially overlap those shown.
3 Sr and Ba c o m m o n l y reach higher concentrations than in other silicate igneous rocks; the highest recorded Sr content (7275 ppm) exceeds Wedepohl's (1978) m a x i m a even for carbonatites.
4 Th, Zr and Rb also reach higher values than in other igneous rocks, except for exotic types of pegmatites and agpaites. 5 Although Y generally has levels characteristic
!
212
N. M. S. Rock 5 Minette and UML fields overlap substantially. 6 There is some overlap between lamprophyre (CAL, UML) and kimberlite fields. 7 On average, lamprophyres show steeper REE profiles than alkali basalts, but have sub-equal contents of the middle REE (around Gd-Er).
of mid-ocean ridge basalt (MORB), it also achieves a few high values (more than 150 ppm in AL and UML). 6 The mean AL spidergram has an almost identical shape to the alkali basalt spidergram, but at enrichments higher by a factor of 1.5-2; this confirms the view (Rock 1977) that AL are volatile- and incompatible-element-enriched alkali basaltic rocks. 7 The mean kimberlite spidergram shows greatest similarities with that of UML.
Isotopic data
Combined Sr and Nd isotopic data are only available for lamproites (Table 11). Sr ratios for AL and UML are consistently low, but CAL display some tendency towards higher values (Fig. 10). Virtually no Pb or stable isotope data are available. Integrated S r - P b - N d - O isotopic studies are urgently needed.
Rare earths (Figs 8 and 9)
Apart from one or two 'anomalous' examples, all lamprophyres show comparable REE spectra, with sub-equal contents of middle REE. Nevertheless the following differences are apparent from Fig. 9.
1 Plagioclase-bearing lamprophyres (spessartires and camptonites) have substantially lower Sm and La/Yb than plagioclase-free types. 2 The order of both light REE enrichment and slope in general is LL > UML > CAL > AL. 3 The light REE thus increase both with increasing K / N a and K-feldspar/plagioclase ratios (cf. Cullers & Graf 1984). 4 Different lamproite families define quite distinct fields.
t
_/
9
I / .~
,Jumillite lamproites ,~~
.....
/
~
Miscellaneous petrological characteristics of the whole lamprophyre clan Xenolith and megacryst/macrocryst assemblages (Table 12)
'
/a
l
20-
t m Camp o ' es I Spessartites ~x-tr,, ~ , ~ M L " ~ ' , Z / / * ~ " X ~ / , ~ ._ F ....~"o . ~ : ~ ' i . . o . . . x . ~ " _ ~.~Fitzroyite /vy..~ lamproites I01- ".'O~.~..o.~..~l:~" i~.' / O
/ 10
Plag~
,,,
20
40
~
__
IO0
200
,
Lu/Yb
FIG. 9. Comparison of lamprophyre REE spectra with those of kimberlites and carbonatites. Symbols as in Fig 3. Note the log-log scales. Sources as in Fig. 8. The majority fields are labelled. Kimberlite field after Nixon et al. (1980).
Lamprophyres characteristically carry abundant crustal xenoliths, which sometimes appear to sample inferred subcrop faithfully (Read et al. 1926). Many lamprophyres, moreover, grade into (or are associated with) breccias, ~zarrying abundant country-rock material (Nixon et al. 1980; Rock 1983). All four lamprophyre branches have also furnished spectacular mantle xenolith suites: CAL, Navajo, U.S.A. (Ehrenberg 1979); AL, Scottish Highlands (Rock 1983); UML, Malaita, Solomons (Dawson et al. 1978); LL, W Kimberley, Australia (Atkinson et al. 1984). The number of recent reports suggests that lamprophyres may even carry proportionately more mantle material than basalts, taking into account relative hostrock abundances. For example, monchiquites have yielded some of the most varied and abundant mantle material among Scottish Permo-Carboniferous alkali basaltic intrusions (Rock 1983; Upton et al. 1983). Curiously,
TABLE 11. Combined Sr and N d &otopic data for lamproites Lamproite source W Australia W Australia Leucite Hills
Reference McCulloch et al. (1983) Nixon et al. (1984) Vollmer et al. (1984)
143Nd/144Nd 0.5110-0.5114 0.5119-0.5121 0.5118-0.5121
87Sr/86Sr 0.7104-0.7187 No data 0.7053-0.7061
Nature of lamprophyres 8-
213
Calc-alkaline lamprophyres (28)
4 I
4--
6 2 3 16
5
21
12-
Alkaline lamprophyres (27)
8 12 9
8-
F.~Tl
I0 13 I4 I
!5
4-
16 -8
II
23
12-
Ultramafic lamprophyres
20
(27)
21
8-
17 18
23
4-
22 I9
25 -
~n
12
8--
Lamproifes (99)
t-
25
424 I
0"705
I
I
0"715
26
177 0"725
e7Sr/e6Sro
whereas the mantle xenolith (?diamond) contents of LL and UML imply depths of origin greater than other igneous rocks (except kimberlites), their contents of leucite and melilite imply crystallization at among the l o w e s t pressures. Varied macrocryst suites are also characteristic (Table 12). Zoned ovoid Ti-phlogopite megacrysts ('eggs') up to 3 cm long (Le Cheminant & Le Cheminant 1985) appear so far to be unique to lamprophyres. Opinions differ as to the cognate or xenocrystic origin of such phases (see references in Table 12).
FIG. 10. Distribution of initial 87Sr/S6Sr ratios in lamprophyres. The numbers in brackets indicate the total numbers of values (isochron intercepts are taken as single values). The numbers within the histograms identify the data sources and localities as follows. Calc-alkaline lamprophyres: 1 kersantites, Massif Central, France (Cantagrel et al. 1970);2 kersantite and spessartite, Sines, Portugal (Canilho 1971); 3 minettes, Spanish Peaks, Colorado (Jahn et al. 1979); 4 minettes, Navajo, U.S.A. (Powell & Bell 1970; Roden 1981); 5 spessartites & kersantites, Cortlandt, U.S.A. (Ratcliffe et al. 1983); 6 minettes, Lake Huron, Canada (Van Schmus 1971); 7 'lamprophyre', probably lamproite, NW Alps (Dal Piaz et al. 1979). Alkaline lamprophyres: 8 camptonites and monchiquites, Monchique, Portugal (Rock 1976); 9 camptonites, Spanish Peaks, U.S.A. (Jahn et al. 1979); 10 monchiquite, Magnet Cove, U.S.A. (Powell et al. 1966); 11 camptonites and monchiquites, New England, U.S.A. (McHone et al. 1976); 12 'lamprophyre' (camptonite), Girnar, India (Paul et al. 1977); 13 camptonite, Predazzo, Italy (Lucchini & Mezzetti 1969); 14 monchiquite, Pyrenees, Spain (Vitrac-Michard et al. 1977); 15 sannaite, Jombo, Kenya (Rock 1976); 16 monchiquites, Hopi, U.S.A. (Powell & Bell 1970). Ultramafic lamprophyres: 17 alnoite, Oka, Canada (Powell et al. 1966); 18 alnoitic rock, Rangwa, Kenya (Rock 1976); 19 damtjernites, Fen, Norway (Griffin & Taylor 1975); 20 aillikite, Ytteroy, Norway (Priem et al. 1968);21 alnoites, Montana, U.S.A. (Powell & Bell 1970); 22 alnoites, Alno, Sweden (Brueckner & Rex 1980); 23 aillikites and alnoites, Frederikshaab, Greenland (Hansen 1981). Lamproites: 24 SE Spain (Powell & Bell 1970); 25 Leucite Hills, U.S.A. (Powell & Bell 1970; Vollmer et al. 1984); 26 W Kimberley, Australia (Powell & Bell 1970); McCulloch et al. 1983; Jaques et al. 1984).
Globular structures
Felsic 'globular structures' (Phillips 1973) ('amygdales', 'ocelli', 'patches', 'varioles' etc. of other workers) are found in all lamprophyre types (summaries in Popov (1972) and Rock (1977, 1984, 1986)). Although similar structures occur in basalts, they are much less widespread. In Scotland, for example, they distinguish PermoCarboniferous AL from Tertiary and Devonian basalts (Bailey et al. 1924). Globular structures differ from their host rock in their lower colour
2 I4
N.M.S.
Rock
TABLE 12. G e n e r a l i z e d a b u n d a n c e o f x e n o l i t h i c a n d m a c r o c r y s t i c / m e g a c r y s t i c m a t e r i a l in the d i f f e r e n t lamprophyre branches
Calc-alkaline lamprophyres 1
Alkaline lamprophyres 2
Ultramafic lamprophyres 3
Lamproites 4
( X )6 --( x )6
X X
X X
X
x x x
x x x
x x x x
x x x x x x x x x x x
x x x x x x x x x
9 x x x x x x x
? x x x x x x x
--
( x )9
?8
x x
x x x
x x --
x x x --
x x x x
x x x x --
(x) -x x (x )
x x
x ? x
Mantle-type ultramafic xenoliths s
Lherzolites, harzburgites Wehrlites, websterites Phlogopite-hornblende-rich rocks Eclogites/garnet-peridotites Other xenolithic material
Al-rich restites Miscellaneous (meta-)sediments Basalt, gabbro, amphibolite Granitoids Deep crustal granulites Country rocks Megacrysts and xenocrysts 7 Diamond Spinel, ilmenite Orthopyro xene, olivine Garnet ( M g - C r )
Kaersutite 1~ Apatite 1~ Clinopyroxene 1~ Sanidine-anorthoclase l~
-x x --
x x x x
x x x x
Quartz, plagioclase Corundum, kyanite, staurolite, sillimanite, A l - F e garnet
x x x x
x x x
x x , widespread; x , recorded; ( x ), very rare; - - , unknown; ?, unconfirmed. 1For sources see table 5 of Rock (1984); additional information in Kramer & Seifert (1984) and Jaques & Perkin (1984). 2Sources: Carstens 1958; Bayrakov 1964; Yeremenko & Shvakova 1969; Dobretsov & Dobretsova 1969; Azambre 1970; Brooks & Rucklidge 1973; Val'ter & Yeremenko 1974; Brooks & Platt 1975; Chapman 1975; Wallace 1975; Janse 1977; Rock 1977, 1983; Faerseth 1978; Leavy 1980; Larsen 1981; Praegel 1981; Mitchell & Janse 1982; Upton et al. 1983. 3 For sources see table 8 of Rock (1986). 4Sources: Carmichael 1967; Kay et al. 1978; Barton & Van Bergen 1981 ; Nixon et al. 1984; Jaques et al. 1984; Bergman 1987. 6 Navajo province, U.S.A. only (e.g. Ehrenberg 1979). vThose generally attributable to the disaggregation of xenoliths are in italics. 8 Aln6, Isle Bizard diatremes (Rock 1986). 9 Wandagee (W. Australia) only currently known example. 10 Probably includes both cognate megacrysts (phenocrysts) and xenocrysts.
index. M a n y o f t h e m g r a d e f r o m s h a r p l y d e f i n e d s u b s p h e r i c a l ocelli, w i t h d e l i m i t i n g t a n g e n t i a l biotite, to v a g u e p a t c h e s , b a r e l y d i f f e r e n t i a t e d f r o m t h e i r host ( R o c k 1979, 1983). The mineral compositions of such structures are q u i t e d i s t i n c t in t h e f o u r l a m p r o p h y r e b r a n c h e s ( T a b l e 1). G i v e n g o o d e x p e r i m e n t a l e v i d e n c e for s i l i c a t e - c a r b o n a t e liquid i m m i s c i bility ( H a m i l t o n et al. 1979), t h e c a r b o n a t e - r i c h s t r u c t u r e s in U M L p r o b a b l y r e p r e s e n t i m m i s c i b l e c a r b o n a t e d r o p l e t s ( ? c a r b o n a t i t e differentiate).
O p i n i o n s differ m u c h m o r e s h a r p l y o v e r t h e f e l d s p a t h i c s t r u c t u r e s in C A L a n d A L , h o w e v e r , with some workers advocating silicate-silicate liquid i m m i s c i b i l i t y ( P h i l p o t t s 1972, 1976,; E b y 1980) a n d o t h e r s a d v o c a t i n g s e g r e g a t i o n o f latestage liquid into vesicles ( C o o p e r 1979; F o l e y 1984; M a u g e r 1984). A p r o f u s i o n o f g e n e t i c suggestions exists for c a r b o n a t e - f e l d s p a r - c h l o r i t e - q u a r t z s t r u c t u r e s in "CAL a n d A L ( R o c k 1984), b u t i d e n t i f i c a t i o n as a m y g d a l e s or varioles has m u c h to c o m m e n d it ( J a f f e 1952; Z i m m e r l e
Nature of lamprophyres 1977). No assessment of these views can be made at least until microprobe and experimental data are available.
Heteromorphism CAL and AL rock types (Fig. l) are heteromorphic, as indicated by the coexistence of nearly isochemical rock types within single dykes, and by both local and global compositional merging (Rock 1977, 1984). Trivial heteromorphism, in which one rock type is merely a chilled version of another, is also found between holocrystalline and glassy lamprophyres (camptonites-hyalomonchiquites, orendites-wyomingites, jumillites-verites) (Rock 1977; Vollmer et aL 1984). Nevertheless, significant heterogeneities within the different branches also occur (Rock 1984, 1986). Heteromorphism is characteristic of lamprophyres, but it is not universal.
Global distribution Distribution through geological time Because they are partly characterized by texture and 'mobile'-element contents, lamprophyres are unlikely to survive even low-grade metamorphism. The older and/or more foliated they are, the more they may be confused with, say, microdior, ites (as in Johnson & Dalziel 1966). Some Precambrian 'lamprophyres' (Watson 1983 and personal communication) and many 'meta-lamprophyres' (Steiner 1984) are very doubtful. Hence the detailed distribution of lamprophyre magmatism through geological time may never be completely defined. Nevertheless, available age data show that lamprophyres are widely distributed from early Precambrian to Recent times (Table 13). As with kimberlites and carbonatites, certain times appear to have been particularly favourable for their emplacement: for example, the PermoTriassic and Jurassic for AL, and the Caledonian and Hercynian orogenies for CAL. LL alone appear to be restricted to a few age bands (Table 13). Some regional lamprophyre dyke-suites span periods of as much as 100 Ma in several distinct tectono-magmatic events (Baxter & Mitchell 1984).
Volumetric abundance Although individual intrusions are of small volume, numerous regional lamprophyre swarms, with thousands of dykes covering many thou-
215
sands of square kilometres, may be volumetrically equivalent to substantial plutons (Rock 1983). Watson (1984) estimated that lamprophyres and associated felsic minor intrusions make up 15% of Scottish Late Caledonian igneous rocks (i.e. several batholith equivalents), and crustal extensions reach 50% locally (Rock et al. 1985, 1986, 1987; Barnes et al. 1986). Although quantitative comparisons are not yet possible, lamprophyres probably have an aggregate global bulk considerably greater than kimberlites, carbonatites or leucite-bearing rocks and at least equal to nepheline syenites. Within lamprophyres, the order of abundance is almost certainly CAL AL >>UML g LL.
Range of tectonic settings (Table 2) Lamprophyres, like other alkaline rocks, are most abundant in continental rifts, failed arms of triple junctions (Brooks & Platt 1975) and on some oceanic islands (De Almeida 1955, 1961 ; Mitchell-Thom~ 1976). However, they also occur widely in orogenic zones and their peripheries (Himalayas (Viterbo & Zanettin 1959); Alps (Dal Piaz et al.1979); Pyrenees (Barrab~ 1952; Rock 1982); Caledonides (Richey 1939); Hercynides (Knill 1982)), in island arcs (Japan (Yagi et al. 1975); Solomons (Nixon et al. 1980)), in passive to destructive continental margins (W U.S.A. (Snaveley & Wagner 1961)), in anomalous uplifted fragments o f oceanic crust (Gorringe Bank (Cornen 1981, 1982); Bermuda (Aumento & Ade-Hall 1973; Aumento et aL 1974)) and near major transcurrent faults (Alpine Fault, New Zealand (Cooper 1979)). Collectively, therefore, lamprophyres are associated not only with intra-plate magmatism, but also with divergent, convergent and even passive plate-margin magmatism.
Petrogenetic significance of lamprophyres Throughout this section, the words primitive, primary and parental are used in the sense of Rhodes (1981). Previous petrogenetic theories for lamprophyres are legion (Rock 1977, 1984, 1986; Bergman 1987). This section concentrates on (1) unifying petrogenetic characteristics of the whole clan and (2) the increasingly recognized possibility that lamprophyres represent primary-mantle melts--by analogy with widely accepted models for alkaline basic to ultrabasic rocks (Frey et al. 1978). Since 'an essential requirement for any primary basalt candidate is that the bulk compo-
216
N. M. S. Rock
.o
~
o e0
0
""
..
o =
~ ~o ~
oov
0 ,.~ ~ ~ . ~
~
,...~
"~ 9
.~:~
"-".~..
~q=l
"
n:~r.~
x x x
x
~5:~zzz
>~_< *
x*
*
x
x
T'T"
x
~,-~l
x x
x
x
~
x
x
"-"
'~ 0 ~ - ' ' - " ~.~ :~,"~
x
x x
x
"
x x
"-"
~
~"
~ ~
~
~ ,~'~,.~
~
x
x
x x
.~
x
x
~
~
~
.~
x
x
~
~
~
~ "~
x
~
~I
"*"~
x
x~'x
~
~D
o
,,-, x:~ ~ C ~
~.
~ ~
, ~ r ~ I= . m
~
*.
X
X
X
X
~
~"~
X
,--,
~ n~
~
~"
0~<~=~
~00 X
P'~
o4 "'~ ,...~
iZ
: ~
C
~l " "
"-'< "~l
.,-,
~
I=
~
0
~
X
X
X
X
X
,.~
.
C eO
X
X
X
X
X
X
X
X
X
~ -r.~
ca . ~ ' ~
9
I~
0
X
X
~u~u~<
Z~Z~:Z~<~,~:O[X
X
~
~'~"~"
m
~ .~l
X
X
~0
X
X
~
X
X
m
Z
~a
X
X
~,
X
0
X
.-~
~.n~ ~.--
X
X
.
~.
~.,~
.... ~l.-~ ~I~.~ I~ ~
X
X
X
X
X
X
~.
~..# ~ ~'~'~ ~
X
X
X
X
X
.~
~.=
~
"~
X
X
~
~
~-
~.=
~ o
_~.
"-~
~=~
~ . ~
0"~
.
~.
~
0
.--
~ ~'--
*
*
0
o
*
"A~
X
;~,
,-~
,~
~
.~
I-,
era ,-1 ~a <
I--.--
. o o > , ~ , ~
=
"~
.=
~
~
.~
~..~ ~-~ ~ > ~....~ ~ ~,~.,
,.~
.~
~
~
~
~
.=
~
X
~a
o4
o
,-,
~
X
o~o
o~nn
_
~'i
'~
/""
"2". c~ .,
C
~
~
9
eao
~,
~
x
' 0
~
~ ~
~
~.~
0""
.
a-~ ~
0
~-.
~_~<
"~,
~
~
~: eO
=~:: ~"~.-C "~ ~ ~ .~
Nature of lamprophyres sition be equivalent to that of a liquid' (Rhodes 1981), we first examine relationships between lamprophyres and their precursor 'magmas'.
Do 'lamprophyremagmas' exist? Owing to their abundant phenocrysts, lamprophyres were assumed by Bowen (1928) never to correspond to fully-liquid magmas. Although lamprophyre phenocrysts need not in fact represent liquidus phases but merely the crystals which grew fastest in such a volatile-rich medium, evidence such as a lack of holohyaline lamprophyres can be used to support Bowen's conclusion (Rock 1984). Lamprophyres might then be attributed to crystal-laden, volatile-charged fluids with only minimal silicate-melt component. These might be generated by gas-phase metasomatism rather than by actual fusion of mantle material, and emplaced by fluidization rather than by normal intrusion (as in Bailey 1984, 1987). However, the lamprophyre combination of [groundmass + phenocrysts + abundant micas, carbonates etc.] could also be taken as a 'frozen' sample representing [melt + suspended crystals + volatile phase]--in fact a 'complete' magma system. Lamprophyres would then be 'closer to intratelluric magma compositions than aphanitic mafic extrusives, the more fashionable yardstick of parental magmas' (Hughes 1982), because they alone retain some representative of the volatile phase. Thus, if it is assumed a priori that primary magmas are fully liquid at the time of generation, natural lamprophyres cannot represent primary magmas. If intratelluric magmas are allowed to be crystal laden and volatile rich, then lamprophyres could uniquely approximate such magmas and many glassy lavas would have to be regarded as fractionated. These alternatives echo the controversy over basalt genesis (i.e. primary origin picrite versus fractionation from primary magma). Two so far largely unaddressed questions are critical here. 1
Are lamprophyre mineralassemblagesprimary ?
Luhr & Carmichael (1981) and Allan & Carmichael (1984 and personal communication) have presented evidence that recent lamprophyres may be erupted as relatively volatile-poor magmas, and that their volatile-rich mineralogy is subsequently acquired by interaction with an aqueous (?deuteric, connate) phase during cooling. The bulk compositions of older lamprophyres may not therefore represent magmas but be determined by syn- or post-magmatic exchanges. 2 A re lamprophyre mineral assemblages in equilibrium ? Commonly coexisting mineral pairs such
217
as analcime with K-feldspar and forsterite or diopside with quartz or alkali feldspars, together with resorbed phenocrysts (especially phlogopite in LL (Mitchell 1985)), may point to disequilibrium. Indeed, many are coming to regard lamprophyre assemblages as 'frozen' mixtures of products and reactants from incompleted reactions (Yoder 1979; Rock 1979, 1986). Unfortunately, while conflicting interpretations of basalt genesis remain unreconciled, and until experimental work is undertaken not only on natural lamprophyres but also on volatile-rich synthetic systems which closely approximate natural lamprophyres, we can only attempt qualitatively to assess whether lamprophyres possess other characteristics consistent with primary status.
Do lamprophyresrepresent primarymagmas? Geochemistry
Among whole-rock parameters widely accepted as indicating primary status are high mg numbers, Ni, Cr and Co. Although acceptable values differ among authors--'primary' mg ranges, for example, from 68%-75% to 70%-83%, and Ni from 240-390 ppm to 200-450 ppm (Rhodes 1981)Fig. 7 shows that many lamprophyres indeed show these 'primary' features. However, many others show lower values and appear to be contaminated or fractionated (Bergman et al. 1985; Laderoute et al. 1985). Absolute contents of MgO around 11%-13% have also been taken (Hart & Davis 1978) to indicate primary status, in which case 10% of analysed CAL, 15% of AL, 20% of UML and 25% of LL would be candidates. However, O'Hara et al. (1975) regarded picritic magmas with 20%-25% MgO as the only primary magmas, in which case only a few lamproites would be candidates. The extreme incompatible-element contents in lamprophyres (Figs 5(b) and 8) present theoretical problems for primary status, for they cannot be generated from mantle materials using known partition coefficients and melting models. Some workers have inferred extreme fractionation of a garnet-bearing assemblage, such as eclogite, to explain steep REE profles (O'Hara et al. 1975). However, similar problems arise with other alkaline basic rocks (Cullers & Graf 1984), and can be reconciled via disequilibrium melting involving a phase rich in REE or a mantle source previously enriched in REE (Venturelli et al. 1984a, b; Hertogen et al. 1985). In view of the inconsistent and sparse Sr-Nd data available (Table 11) and the occurrence of
218
N. M. S. Rock
high Sr ratios in the Australian lamproites (despite their content of diamond and mantle peridotite nodules), the interpretation of unsupported Sr data for other lamprophyres (Fig. 10) must be circumspect. Ratios in AL and UML seem low and consistent enough to rule out both crustal involvement and 'enriched' mantle sources. However, the bias towards higher Sr ratios in CAL has already been interpreted variously in terms of crustal involvement (Powell & Bell 1970) or mantle effects (Roden 1981). Negative spidergram Nb anomalies in CAL and LL (Fig. 5(b)) are characteristic of destructive plate-margin magmatism (Pearce 1982, 1983) and have been interpreted (Thompson et al. 1984) as indicating assimilation of subducted sediment slabs. In that AL and UML have no Nb anomalies, this is at least consistent with the tentative indications of Sr isotopes. However, it is as yet unclear how much weight can be attached to such 'anomalies', given the rather artificial method of data display. Mineralogy
The high Fo content of most lamprophyre olivines (rag number 86-94) (Fig. 2), and the high mg numbers, Cr and Ni of lamprophyre phlogopites have all been taken to support primary status (Jones & Smith 1983; Le Cheminant & Le Cheminant 1985). However, some olivines (especially in CAL and AL) are relatively Fo poor (Fig. 2), and the relative rarity of olivine phenocrysts in CAL could even be taken to deny primary status absolutely. Experimental data
Extensive data are available only for lamproites (Barton & Hamilton 1978, 1979, 1982; Arima & Edgar 1983), where they are consistent with primary origin provided that (1) the mantle source has been previously depleted in garnet and clinopyroxene and enriched in phlogopite and (2) melting occurs under high H20 and F but low CO2 pressures. However, very different melting conditions must apply to UML (probably near kimberlites) (Rock 1986) and to AL (probably near basanites)(Nathan 1968; Rock 1978b). Xenoliths
The abundant mantle-type xenoliths in many lamprophyres (Table 12), and especially the diamond in LL (and ?AL and ?UML), would seem to indicate primary status. However, crystalladen lamprophyres probably have high yield strengths, enhancing ability to carry xenoliths (Sparks et al. 1977; Spera 1980). Xenolith-bearing
lamprophyres might still therefore have fractionated even after xenolith incorporation (cf peridotite-bearing phonolites (Wright 1971)). Larsen ( 1981), for example, argued that one monchiquite had fractionated despite its abundant mantle xenoliths and mg value (71-74) because its olivine and Cr-diopside megacrysts must have crystallized from a magma which was richer still in Mg. Crystallization temperatures and ranges
The consistently late-stage intrusion of lamprophyres and their general lack of contact-metamorphic effects both tend to imply lowtemperature emplacement, which is more consistent with a fractionated condition. However, relatively few dyke-rocks of any composition cause extensive contact metamorphism, and a few lamprophyres have caused pyroxene-hornfels or even sanidinite-facies metamorphism (Rock 1983, 1984; Macdonald et al. 1986). More suggestive still of fractionation or accumulation are the non-cotectic status and long crystallization-temperature ranges of the few natural lamprophyres studied to date (Nathan 1968; Rock 1978b; Esperan~a & Holloway 1985). Tectonic setting
The occurrence of AL and UML on oceanic islands proves that continental-crust involvement is not a prerequisite for their genesis. However, neither CAL nor LL occur in an oceanic setting, and CAL repeatedly show such a close spatial, temporal and almost certainly genetic relationship with granitic plutonism, together with a characteristic abundance of crustal xenoliths (Table 12), that some crustal involvement seems irrefutable in their genesis (Rock 1980, 1984). Although many of the above uncertainties apply to the petrogenesis of other alkaline magmas, they do tend to show the following. 1 Lamprophyres are as diverse genetically as they are compositionally--differing sources and/ or melting conditions can apply even to different examples of one branch within a single igneous complex (Vollmer et al. 1984). 2 Some lamprophyres in each branch approximate primary-mantle melts derived under varying conditions: CAL, Navajo, U.S.A. (Ehrenberg 1979); AL, Wawa, Canada (Mitchell & Janse 1982); UML, Malaita, Solomons (Nixon et al. 1980); LL, W Kimberley, Australia (Jaques et al. 1984). Indeed, LL and UML (along with kimberlites) encompass melts with the deepest origins of all igneous rocks. 3 Other lamprophyres (especially CAL) originate by volatile enrichment, fractionation, mix-
Nature of lamprophyres ing and other multistage processes (Rock 1977, 1984, 1986); Clarke et al. 1983 ; Van Bergen et al. 1983; Macdonald et al. 1986; Nixon et al. 1986).
Differentiation in lamprophyres, and their possible status as parental magmas to hydrous alkaline intrusive rock-suites Whether or not they are primary, lamprophyres repeatedly have the most primitive compositions in a given rock-suite. That they may in turn represent the parental magmas to such suites is supported by several lines of evidence. 1 Availability. In numerous combined major+ minor intrusive alkaline suites (Lovozero and Khibina, U.S.S.R. (Grigor'eva & Savitskij 1979); Borralan, Scotland (Sutherland 1982); Sokli, Finland (Vartiainen et al. 1978)) lamprophyres are the only primitive rocks, the associated major intrusions being exclusively syenitic, ijolitic or carbonatitic. Again, in some entire alkaline provinces (Monteregian and Iberian Provinces (Philpotts 1974; Rock 1982))lamprophyres represent the only primitive liquids intruded throughout the province. 2 M a n t l e xenolith content. In numerous rocksuites (Fen (Griffin 1973); Pyrenees (Azambre 1979)) lamprophyres are the only rocks carrying such xenoliths. 3 Differentation. Single lamprophyre intrusions may carry intermediate to acidic veins, ocelli or schlieren (Mukherjee 1961 ; Philpotts 1972; Rock 1984) or show composite structure (Sathe & Desai 1968 ; Rock 1984, table 8). Entire dyke-suites may show co-magmatic series from primitive lamprophyres to felsic rock types (N~mec 1973; Rock et al. 1986, Barnes et al. 1986).
For these and other reasons, lamprophyres have been equated with parent magmas for the following wide range of hydrous intrusive suites. CAL: 'shoshonitic' hornblende-rich pyroxenite-diorite-norite-granodiorite-granite plutons
219
(Cortlandt, U.S.A. (Ratcliffe et al. 1983)), potassic syenites (Arkansas, U.S.A. (Mullen & Petty 1985); Biella, Italy (Dal Piaz et al. 1979)) and even calc-alkaline granites (Japan (Suzuki & Shiraki 1980); Scottish Caledonides (Macdonald et al. 1986)). AL (camptonites); kaersutite-biotite-rich gabbro-(nepheline) syenite plutons (Monteregian Hills (Philpotts 1974; Bedard et al. 1985)). AL (monchiquites) and UML: ijolite-syenitecarbonatite complexes (Fen (Griffin & Taylor 1975)). One of the most intriguing cases is the Monteregian province (Philpotts 1974), where lamprophyre dykes gradually change from camptonite in the E of the province via monchiquite to aln6ite in the W; the associated plutons change concomitantly from saturated gabbro-syenite via essexite-foyaite to ijolite-carbonatite. The plutonic rocks are cumulates (Eby 1987), and the lamprophyres are among the few candidates for primary magmas. Although Philpotts' view that the three parent lamprophyre types generated the three types of plutonic suites has proved oversimplified (Eby 1985a, b, 1987), there is undoubtedly a genetic connection between representatives of three lamprophyre branches and of all the most globally widespread types of alkaline plutonic suite. In conclusion, lamprophyres are almost always primitive, frequently primary and sometimes parental. They are considerably more abundant and significant than hitherto recognized. This series of reviews will hopefully stimulate new multidisciplinary studies of these fascinating rocks. Stimulus from a more pecuniary direction may well come from the increasing number of deposits of precious stones (Atkinson et al. 1984; Meyer & Mitchell 1985) and metallic minerals (Haughton & Frommerze 1930; Kumar 1968; Burkhalter et al. 1971 ; Rosseykin & Razhmanov 1971 ; Makeyev & Yefimov 1972; Sarkisyan 1973; Shchukin 1974; Daniyelyants & Lyakhov 1975) now reported in (or associated with) lamprophyres !
References References in W European languages are given in those languages. References in other languages are translated into English between square brackets. References have been cited only selectively, as far as possible to expand and update rather than duplicate those already cited by Rock (1977, 1984, 1986) and by other authors in the present volume. ALLAN,J. E. & CARMICHAEL,I. S. E. 1984. Lamprophyric lavas from the Colima graben, Mexico. Contrib. Mineral. Petrol. 88, 203-16. ARIMA, M. & EDGAR, A. D. 1981. Substitution
--
mechanisms and solubility of titanium in phlogopites from rocks ofprobable mantle origin. Contrib. Mineral. Petrol. 77, 288-95. & - - 1983. A high pressure experimentalstudy
220
N. M. S. Rock
on a magnesian-rich leucite-lamproite from the West Kimberley area, Australia: petrogenetic implications. Contrib. Mineral. Petrol. 84, 228 34. ATKINSON, W. J., HUGHES, F. E. & SMITH, C. B. 1984. A review of the kimberlitic rocks of western Australia. In: KORNPROBSTJ. (ed.)Kimberlites I: Kimberlites and Related Rocks, pp. 195-224. Elsevier, Amsterdam. AUMENTO, F. & ADE-HALL, J. M. 1973. Deep-drill1972. Petrology of the Bermuda drill core. EOS 54, 485. --, REYNOLDS, P. M. & GUNN, B. M. 1974. The Bermuda seamount: a reactivated section of an older ocean crust. EOS 55, 455. AZAMBRE, M. B. 1970. Les monchiquites et autres roches basiques intrusives accompagnant les sy6nites n6ph61iniques des Corbi+res. C. R. Acad. Sci. Paris, 271, 641-3. BACHINSKI, S. W. & SCOTT, R. B. 1980. Authors' reply. Geochim. cosmochim. Acta 44, 1389-92. & SIMPSON, E. L. 1984. Ti-phlogopites of the Shaw's Cove minette : a comparison with micas of other lamprophyres, potassic rocks, kimberlites, and mantle xenoliths. Am. Mineral. 69, 41-56. BAILEY, D. K. 1984. Kimberlite: 'The mantle sample' formed by ultrametasomatism. In: KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 323-34. Elsevier, Amsterdam. 1987. Mantle metasomatism--perspective and prospect. In : FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 1-13. BAILEY,E. B., CLOUGH, C. T., WRIGHT, W. B., RICHLY, J. E. & WILSON, G. V. 1924. Tertiary and postTertiary geology of Mull, Loch Aline and Oban. Mem. geol. Surv. Scotland. BAKER, B.H. 1953. The alkaline igneous complex of Jombo Hill. Geol. Surv. Kenya Rep. 24, 32-48. BARNES, R. P., ROCK, N. M. S. & GASKARTH, J. W. 1986. Late Caledonian dyke-swarms of southern Scotland: new data for the Wigtown Peninsula. Geol. J. 21,101-25. BARRABLL. C. 1952. Les roches intrusives fi hornblende brune des Pyr6n6es. 19th Int. Geol. Cong. Fasc., 6, 9-22. BARTON, M. & HAMILTON,D. L. 1978. Water-saturated melting relations to 5 kb of three Leucite Hills lavas. Contrib. Mineral. Petrol. 66, 41-9. -& -1979. The melting relationships of a madupite from the Leucite Hills, Wyoming to 30 kb. Contrib. Mineral. Petrol. 69, 133-42. & 1982. Water-undersaturated melting experiments bearing upon the origin of potassiumrich magmas. Mineral. Mag. 45, 267-78. - & VAN BERGEN, M. J. 1981. Green clinopyroxenes and associated phases in a potassium-rich lava from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 77, 101-14. BASTA,E. Z., EL-SHARKAWY,M. A. & MIKHAIL, M. A. 1985. Petrographical and petrochemical studies on some Egyptian lamprophyres. J. Coll. Sci. King Saud Univ. (Riyadh), 16(1), 137-64. BAXTER, A. N. & MITCHELL, J. G. 1984. Camptonitemonchiquite dyke swarms of Northern Scotland:
age relationships and their implications. Scott. J. Geol. 20, 297-308. BAYRAKOV, V. V. 1964. Occurrence of eclogitic rock xenoliths in a lamprophyre dyke of the Oktyabr'skiy alkalic stock. Dokl. Acad. Sci. U.S.S.R., Earth Sci. Sect. 156, 137-40. BEDARD, J., LUDDEN, J. & DANIS, D. 1985. Silicaoversaturated residua derived by fractionationassimilation from a camptonitic parental magma, the Megantic complex. Geol. Soc. Can. Annu. Meet., Fredericton, NB, Progr. Abstr., A3. BEGER, P. J. 1923. Der Chemismus der Lamprophyre. In: NIGGLI, P. (ed.) Gesteins und Mineralprovinzen. Vol. 1, Einfuhrung, pp. 217-577. Gebriider Borntrager, Berlin. BERGMAN,S. C. 1987. Lamproites and other potassiumrich igneous rocks: a review of their occurrence, mineralogy and geochemistry. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 103-90. - - , DUNN, D. P. & KROL, L. G. 1985. Petrology and geochemistry of the Linhaisai minette, Karamu river, Central Kalimantan. Geol. Soc. Can. Annu. Meet., Federicton, NB, Progr. Abstr., A4. BOWEN, N. L. 1928. Genetic features of alnoitic rocks at Ile Cadieux, Quebec. Am. J. Sci. 3, 1-34. BROOKS,C. K. & PLATT,R. C. 1975. Kaersutite-bearing gabbroic inclusions and the late dike swarm of Kangerdlugssuaq, E. Greenland. Mineral. Mag. 40, 258-83. - - & RUCKLIDGE,J. C. 1973. A Tertiary lamprophyre dyke with high pressure xenoliths and megacrysts from Wiedemanns Fjord, E. Greenland. Contrib. Mineral. Petrol. 42, 197-212. BRUECKNER, H. K. & REX, D. C. 1980. K - A and R b Sr geochronology and Sr isotopic study of the Aln6 alkaline complex, northeastern Sweden. Lithos, 13, 111-20.
BUIE, B. F. 1941. Igneous rocks of the Highwood Mountains, Montana. Part III, Dikes and related intrusives. Geol. Soc. Am. Bull. 57, 1753-808. BURKHALTER,A. J., ARGUELLESALVAREZ,A. & ORTIZ RAMOSA, A. 1971. Estudios sobre los yacimientos de fluorite-galena de serra de Gadar: rocas igneas de Sierra de Gadar. In: 1st Hispano-LusitanianAmerican Economic Geology Congress, MadridLisbon, Vol. 4, pp 51-72. CANILHO, M. H. S. 1971. Estudo geolbgico-petrogr~tfico do macico eruptivo de Sines. Bol. Mus. Lab. Mineral. Geol. Fac. Cikncias Univ. Lisboa, 12, 77161. CANTAGREL, J.-M., VALIZADEH, M. & VIALETTE, Y. 1970. Age des granites, granophyres et kersantites de la r6gion de Thiers dans le Massif Central. C. R. Acad. Sci. Paris, Set. D, 270, 600-3. CARMICHAEL, I. S. E. 1967. Mineralogy and petrology of the volcanic rocks from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 15, 24-66. CARSTENS, H. 1958. The origin of feldspar inclusions in the lamprophyres of Kristiansand, Norway. Nor. geol. Tiddskr. 38, 245-52. CARSWELL, D. A. 1980. Mantle derived lherzolite nodules associated with kimberlite, carbonatite
Nature of lamprophyres and basalt magmatism: a review. Lithos, 13, 12138. CHAPMAN, N. A. 1975. An experimental study of spinel clinopyroxenite xenoliths from the Duncansby Ness vent, Caithness. Contrib. Mineral. Petrol. 51, 223-30. CLARKE, D . B . , M U E C K E , G . K . & PE-PIPER, G. 1983. The lamprophyres of Ubekendt Eijland, West Greenland: products of renewed partial melting or extreme differentiation? Contrib. Mineral. Petrol. 83, 117-227. COOPER, A. F. 1979. Petrology of ocellar lamprophyres from western Otago, New Zealand. J. Petrol. 20, 139-64. CORNEN, G. 1981. Des complexes alcalins similaires en domaines oc~aniques et continental: le banc de Gorringe compar6/t la Serra de Monchique. C. R. Acad. Sci. Paris, 292, 463-6. 1982. Petrology of the alkaline volcanism of Gorringe bank (SW Portugal). Marine Geol. 47, 101-30. CULLERS,R. L. & GRAF, J. L. 1984. Rare-earth elements in igneous rocks of the continental crust: predominantly basic and ultrabasic rocks. In : HENDERSON, P. (ed.) Rare-earth Element Geochemistry, pp 23774, Elsevier, Amsterdam. & MEDARIS, L. G. 1977. Rare-earth elements in carbonatite and cogenetic alkaline rocks: examples from Seabrook Lake and Callander Bay, Ontario. Contrib. Mineral. Petrol. 65, 143-53. CURRIE, K. L. 1976. The alkaline rocks of Canada. Bull. geol. Surv. Can., 239. - - & FERGUSON,J. 1970. The mechanism of intrusion of lamprophyre dikes as indicated by 'offsetting' of dikes. Tectonophysics, 9, 525-35. CZAMANSKE, G. K., ERD, R. C., LEONARD, B. F. & CLARK, J. R. 1981. Bartonite, a new potassium sulfide mineral. Am. Mineral. 66, 369-75. DAL PIAZ, G. V., VENTURELLI,G. & SCOLARI,A. 1979. Calc-alkaline to ultrapotassic postcollisional activity in the internal northwestern Alps. Mere. Ist. Geol. Unit,. Padova, 32, 4-16. DANIYELYANTS, S. Y. & LYAKHOV, Y. V. 1975. [New data on the relative age of lamprophyres and ores in the Darasun deposit, E. Transbaikal.] Vyssh. Uchebn. Zaved. Izv. Geol. Razved. 5, 94-103 (in Russian). DAVIS, J. C. 1973. Statistics and Data Analysis in Geology, Wiley, New York. DAWSON, J. B. 1980. Kimberlites and their Xenoliths, Springer, Berlin. 1987. The kimberlite clan: relationship with olivine and leucite lamproites, and inferences for upper-mantle matasomatism. In: FITTON, J. G. & UPTON, B. G. J. (eds)Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 95-101. , DEt,ANEY, J. S. & SMITH, J. V. 1978. Aspects of the mineralogy of alnoitic breccia, Malaita, Solomon Is. : comparison with continental kimberlites. Contrib. Mineral. Petrol. 67, 189-94. DE ALMEIDA, M. A. 1955. Geologia e petrologia do arquip61ago de Fernando de Noronha. Monogr. Div. Geol. Mineral. Brazil, 13. -
-
221
1961. Geologia e petrologia da ilha da Trindade. Monogr. Div. Geol. Mineral. Brazil, 18. DEER, W. A., HOWIE, R. A. & ZUSSMAN,J. 1962. Rockforming Minerals, Longmans, Harlow, Essex. , & -1978. Rock-forming Minerals, 2nd edn., Longmans, Harlow, Essex. DELANEY, P. T. & POLLARD, D. D. 1981. Deformation of host rocks and flow of magma during growth of minette dikes and breccia-bearing intrusions near Ship Rock, New Mexico. Prof. Pap. U.S. geol. Surv. 1202. DIETRICHSON,B. 1955. Spessartite and pseudotachylite from the thrusting-zone of E. Jotunheim. Nor. geol. Unders. 191, 30-55. DOBRETSOV, G. L. & DOBRETSOVA, T. G. 1969. [Inclusions of deep-seated origin in camptonites of southern Dzhungaria and their genetic significance.] Akad. Nauk. U.S.S.R. Sib. Otd. Instn. Volkanol. Mosk. 40-50 (in Russian). DURAZZO, A., TAYLOR, L. A. & SI-IERVA~S,J. W. 1984. Ultramafic lamprophyre in a carbonate platform environment, Mt. Queglia, Abruzzo, Italy. Neues J. Mineral. Abh. 150, 199-217. EBY, G. N. 1975. Abundance and distribution of the rare-earth elements and yttrium in the rocks and minerals of the Oka carbonatite complex, Quebec. Geochim. cosmochim. Acta, 39, 597-620. 1980. Minor and trace element partitioning between immiscible ocelli-matrix pairs from lamprophyre dykes and sills, Monteregian Hills petrographic province, Quebec. Contrib. Mineral. Petrol. 75, 269-78. - 1985a. Mafic alkaline rocks of the Monteregian Hills, Quebec. Geol. Soc. Can. Annu. Meet. Fredericton, NB, Progr. Abstr., A15. 1985b. Chemistry and age relations of mafic alkaline dykes from the Monteregian Hills and younger White Mountain igneous provinces. Can. J. earth Sci., 22, 1103-11. - 1987. The Monteregian Hills and White Mountain alkaline igneous provinces, eastern North America. In: FITTON, J. G. & UPTON, B. G. J. (eds) Igneous Alkaline Rocks, Geol. Soc. Spec. Publ. 30, pp. 433 47. ECKERMANN, H. VON 1968. The chemical and optical properties of some minerals of the Aln6 alkaline rocks. Ark. MineraL-Geol. 5, 93-210. EGOROV, L. S. 1970. Carbonatites and ultrabasicalkaline rocks of the Maimecha-Kotui region, N. Siberia. Lithos, 3, 341-59. EHRENBERG, S. 1979. Garnetiferous ultramafic inclusions in minette from the Navajo volcanic field. In: BOYD, F. R. & MEYER, H. A. O. (eds) The Mantle Sample, pp. 330-44. American Geophysical Union, Washington. ESPERANgA, S. & HOLLOWAY, J. R. 1985. An experimental study of a mafic minette from Buell Park, Arizona. Geol. Soc. Can. Annu. Meet., Fredericton, NB, Progr. Abstr., A17. EVANS,J. W. 1901. A monchiquite from Mount Girnar, Junagarh. Q. J. geol. Soc. Lond. 57, 38-52. FAERS~TI-I, R. B. 1978. Mantle-derived lherzolite xenoliths and megacrysts from Permo-Triassic dykes, - -
222
N. M. S. Rock
Sunnhordland, western Norway. Lithos, 11, 2335. --, MACINTYRE, R. M. & NATERSTAD, J. 1976. Mesozoic alkaline dykes in the Sunnhordland region, western Norway. Lithos, 9, 331-45. FOLEY, S. F. 1984. Liquid immiscibility and melt segregation in alkaline lamprophyres from Labrador. Lithos, 17, 127-38. - & MALPAS, J. 1986. Petrology of the alkalineultrabasic lamprophyres of Aillik Bay, Labrador. Can. J. earth Sci., in press. FRENZEL, G. 1971. Die Mineralparagenese der Albersweiler Lamprophyre. Neues J. Mineral. Abh. 115, 164-91. FREY, F. H., GREEN, D. H. & ROY S. D. 1978. Integrated models of basalt petrogenesis: a study of quartz tholeiites to olivine melilitites from SE Australia utilizing geochemical and experimental petrological data. J. Petrol. 19, 463-513. FURNES, H. & STILLMAN,C. J. 1987. The geochemistry and petrology of an alkaline lamprophyre sheet intrusion complex on Maio, Cape Verde Republic, J. geol. Soc. Lond., 144, 227-41. GHOSE, C. 1949. Petrochemical study of the lamprophyres and associated intrusive rocks of the Jharia coalfield. Q. J. Geol. Min. Metall. Soc. India, 21(4), 133-47. GILL, J. B. 1982. Orogenic Andesites and Plate Tectonics. Springer, Berlin. GRAPES, R. H. 1975. Petrology of the Blue Mountain complex, Marlborough, N . Z . J . Petrol. 16, 371428. GRIFFIN, W. L. 1973. Lherzolite nodules from the Fen alkaline complex, Norway. Contrib. Mineral. Petrol. 38, 135-46. -& TAYLOR, P. N. 1975. The Fen damkjernite: petrology of a 'central complex kimberlite'. Phys. Chem. Earth, 9, 163-77. GRIGOR'EVA, L. V. & SAVlTSKIJ,A. V. 1979. [New data on alkaline lamprophyres of the NE border of the Khibiny massif.] Dokl. Akad. Nauk S.S.S.R. 248, 1400-03 (in Russian). GUPTA, A. K. & YAGI, K. 1980. Petrology and Genesis of Leucite-bearing Rocks, Springer, Berlin. HALL, A. 1982. The Pendennis peralkaline minette. Mineral. Mag. 45, 257-66. HAMILTON, D. L., FREESTONE,I. C., DAWSON, J. B. & DONALDSON,C. n. 1979. Origin of carbonatites by liquid immiscibility. Nature, Lond. 279, 52-4. HANSEN, K. 1981. Systematic Sr-isotopic variation in alkaline rocks from west Greenland. Lithos, 14, 183-8. HART, S. R. & DAVIS, K. E. 1978. Nickel partitioning between olivine and silicate melt. Earth planet. Sci. Lett. 40, 203-19. HAUGHTON, S. H. & FROMMURZE,H. F. 1930. Notes on the occurrence of argentiferous galena near Aiais, SW Africa. Trans. geol. Soc. S. Aft. 32, 89-102. HEINRICH, E. W. & ALEXANDER,D. H. 1975. Exotic and prosaic lamprophyres, McClure Mtn. Province, Colorado. Geol. Soc. Am. Abstr. Progr. 7(6), 777-8. HERTOGEN,J., LOPEZ-RUIZ,J. & RODRIGUEZ-BADIOLA, E. 1985. Petrogenesis of ultrapotassic volcanic
rocks from S.E. Spain: trace elements and Sr-Pb isotopes. Geol. Soc. Can. Annu. Meet., Fredericton, NB, Progr. Abstr., A26. HUGHES, C. J. 198Z Igneous Petrology. Elsevier, Amsterdam. JAFFE, H. F. 1952. Amygdalar camptonite dykes from Mount Jo, Mount Marcy Quadrangle, Essex County, New York. Am. Mineral. 38, 1065-77. JAHN, B . , SUN, S.-S. & NESBITT, R. W. 1979. REE distribution and petrogenesis of the Spanish Peaks igneous complex, Colorado. Contrib. Mineral. Petrol. 70, 281-98. JANSE, A. J. A. 1977. Harzburgite nodules from a lamprophyre near Wawa, Ontario, Canada. 2nd Int. Kimberlite Conf., Extended Abstr., Santa Fe. JAQUES, m. L., LEWIS, J. D., SMITH, C. B., GREGORY, G. P., FERGUSON, J., CHAPPELL, B. W. & MCCULLOCH, M. T. 1984. The diamond-bearing ultrapotassic (lamproitic) rocks of the West Kimberley region, Western Australia. In : KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 225-54. Elsevier, Amsterdam. - & PERKIN, D. J. 1984. A mica, pyroxene and ilmenite-megacryst bearing lamprophyre from Mount Wooloma, northern NSW. B M R J. Aust. Geol. Geophys. 9, 33-40. JOHNSON, M. R. W. & DALZIEL, I. W. D. 1966. Metamorphosed lamprophyres and the late thermal history of the Moines. Geol. Mag. 103, 240-9. JONES, A. P. & SMITH, J. V. 1983. Petrological significance of mineral chemistry in the Agathla peak and The Thumb minettes, Navajo volcanic field. J. Geol. 91,643-56. KAY, R. W. & GAST, P. W. 1973. REE content of alkalirich basalts. J. Geol. 81, 653-82. KAY, S. M., KAY, R. W., HANGAS, J. & SNEDDEN,T. 1978. Crustal xenoliths from potassic lavas, Leucite Hills, Wyoming. Geol. Soc. Am. Abstr. Progr. 10(7), 432. KNILL, D. C. 1982. Permian volcanism in SW England. In: SUTHERLAND,D. S. (ed.) Igneous Rocks of the British Isles, pp. 329-32. Wiley, New York. KNOPF, A. 1936. Igneous geology of the Spanish Peaks region. Bull. geol. Soc. Am. 47, 1728-84. KRAMER, W. & SEIFERT, W. 1984. Xenolithe, Lamprophyre und Kruste-Mantel-Beziehungen. Forsch. Reithe C (Leipz.), 389, 38-49. KRANCK, E. 1953. Bedrock geology of the seaboard of Labrador between Domino Run and Hopedale. Bull. geol. Surv. Can. 26. KRESTEN, P. & EDELMAN, N. 1975. A boulder of ouachitite from Inderskar, Aland Is. Bull. geol. Surv. Finl. 47, 167-9. - - , AHMAN, E. & BRUNFELT, A. O. 1981. Alkaline ultramafic lamprophyres and associated carbonatite dykes from the Kalix area, Northern Sweden. Geol. Rundsch. 70, 1215-31. KUEHNER, S. M., EDGAR, m. D. & ARIMA, M. 1981. Petrogenesis of the ultrapotassic rocks from the Leucite Hills, Wyoming. Am. Mineral. 66, 663-77. KUMAR,J. C. N. 1968. [Spessartite and diorite-porphyry dikes in the Elbrus polymetallic deposit.] Vyssch. Ucheb. Zaved. Isv. Geol. Razved. 3, 161-5 (in Russian).
Nature of lamprophyres
223
petrologic characteristics of lamprophyres in one LADEROUTE, D. G., EVANS, N. J. & MITCHELL, R. H. of the tin ore areas in the central Chukchi 1985. Lamprophyres associated with the Coldwell peninsula.] Soy. Geol. 12, 105-12 (in Russian). alkaline complex, N.W. Ontario. Geol. Soc. Can. MAUGER, R. L. 1982. Petrology of a Mesozoic biotiteAnnu. Meet., Fredericton NB, Progr. Abstr., A33. lamprophyre near Concord, N. Carolina. Geol. LARSEN, J. G. 1981. Medium pressure crystallisation of Soc. Am. Abstr. Progr. 14(1-2), 38. a monchiquitic magma--evidence from mega1984. Origin of ocelli in an early Carboniferous crysts of Drever's block, Ubekendt Eijland, W. - minette near Concord, N Carolina. Geol. Soc. Am. Greenland. Lithos, 14, 241-62. Abstr. Progr. 16(3), 179. LEAKE, B. E. 1978. Nomenclature of amphiboles. METALS, D. & CnAYES, F. 1964. Classification of Mineral. Mag. 42, 533-63. lamprophyres: a possible petrographic application LEAVY, B. D. 1980. Mantle-derived lamprophyre dykes of multigroup discriminant function analysis. and included ultramafic nodules from SE New Carnegie lnstn. Wash. Yb. 63, 182-5. England. Geol. Soc. Am. Abstr. Progr. 12(2), 69. LE BAN, M. J. 1984. Oceanic carbonatites. In: KORN- MEYER, H. O. A. & MITCHELL, R. H. 1985. Sapphirebearing lamprophyre from Yogo Gulch, Montana. PROBST, J. (ed.) Kimberlites I: Kimberlites and Geol. Soc. Annu. Meet. Fredericton, NB, Progr. Related Rocks, pp. 169-80. Elsevier, Amsterdam. Abstr., A39. LE CHEMINANT,A. N. & LE CHEMINANT, G. M. 1985. & VILLAR, L. M. 1984. An aln6ite in the Sierras Phlogopite from 1.8 Ga lamprophyres and trachy- - Subandinas, northern Argentina. J. Geol. 92, 741andesites, district of Keewatin: petrologic impli51. cations. Geol. Soc. Can. Annu. Meet. Fredericton, MITCHELL, R. H. 1979. The alleged kimberliteNB, Progr. Abstr., A33. carbonatite relationship: additional contrary miLE MAITRE, R. W. 1976. The chemical variability of neralogical evidence. Am. J. Sci. 279, 570-89. some common igneous rocks. J. Petrol. 17, 589-1981. Titaniferous phlogopites from the leucite 637. lamproites of the W Kimberley area, Western -1982. Numerical Petrology. Elsevier, Amsterdam. Australia. Contrib. Mineral. Petrol. 76, 243-51. - 1984. A proposal by the IUGS Subcommission on 1985. A review of the mineralogy of lamproites. the systematics of igneous rocks for a chemical - Trans. geol. Soc. S. Afr., 88, 411-437. classification of volcanic rocks based on the total & JANSE, A. J. A. 1982. A harzburgite-bearing alkali silica (TAS) diagram. Austr. J. earth Sci. 31, - monchiquite from Wawa, Ontario. Can. Mineral. 243-55. 20,211-16. LISTER, B. 1982. Evaluation of analytical data: a practical guide for geoanalysts. Geostand. News. MITCHELL-THOM~, R. C. 1976. Geology of the Middle Atlantic Islands. Gebrfider Borntraeger, Berlin. 6(2), 175-205. LUCCHINI, F. & MEZZETTI, R. 1969. Megacristalli di MORRISON, M. A., THOMPSON, R. N., GIBSON, I. L. & MARRINER, G. F. 1980. Lateral chemical heterosanidino nella camptonite di Imana (Predazzo), geneity in the Palaeocene upper mantle beneath Mineral. Petrogr. Acta, 15, 185-90. the Scottish Hebrides. Phil. Trans. R. Soc. Lond., LUHR, J. F. & CARMICaAEL,I. S. E. 1981. The Colima SeT. A, 297, 229-44. volcanic complex, Mexico: Part II. Late Quaternary cinder cones. Contrib. Mineral. Petrol. 76, MUKHERJEE, K. K. 1961. Petrologyofthe lamprophyres of the Bokaro coalfield. Q. J. Geol. Min. Metall. 127-47. Soc. India, 33(2), 69-87. MCCULLOCH, M. T., JAQUES, A. L., NELSON, D. R. & LEWIS, J. D. 1983. Nd and Sr isotopes in MULLEN, E. D. & PETTY, W. B. 1985. Petrologic relations among syenites and lamprophyric rocks, Kimberlites and lamproites from Western AusArkansas alkalic province. Geol. Soc. Can. Annu. tralia: an enriched mantle origin. Nature, Lond. Meet., Fredericton, NB, Progr. Abstr., A42. 302, 400-3. MACDONALD, G. A. & KATSURA, T. 1964. Chemical NAKAMURA, N. 1974. Determination of REE, Ba, Mg, Na and K in carbonaceous and ordinary chonComposition of Hawaiian lavas. J. Petrol. 5, 82drites. Geochim. cosmoc~(m. Acta, 38, 757-75. 133. MACDONALD, R., THORPE, R. S. & GASKARTH, J. W. NATHAN, S. 1968. Experin~e)atal crystallisation of a lamprophyre glass. N.Z. J. Geol. Geophys. 11, 1985. Multi-source origin for lamprophyres of N. 1191-7. England. Mineral. Mag., 49, 485-94. - - , ROCK, N. M. S., RUNDLE, C. C. & RUSSELL,O. J. NEMEC, D. 1973. Differentiation series of minettes in the central Bohemian pluton. J. Geol. 81,632-42. 1986. Relationships between late Caledonian lam-1985. Ba and its carriers in dyke rocks of the prophyric, syenitic, and granitic magmas in a minette series. Geol. Soc. Can. Annu. Meet., differentiated dyke, southern Scotland. Mineral. Fredericton, NB, Progr. Abstr., A43. Mag. .50, 547-57. MCHONE, J. G. & CORNEILLE,E. S. 1980. Alkalic dykes NICHOLLS, J. W. 1969. Studies of the volcanic petrology of the Navajo-Hopi area, Arizona. PhD Thesis, of the Lake Champlain valley. Vermont. Geol. 1, University of California, Berkeley, CA. (unpub16-21. lished). , BLACK, W. W. & PEGRAM, W. J. 1976. Origin of NIGGLI, P. 1923. Gesteins und Mineralprovinzen. Gebriinorthwestern New England lamprophyres: Sr isotopes and geochemical evidence. Geol. Soc. Am. der Borntraeger, Berlin. Abstr. Progr. 8(2), 227-8. NIKISHOV, K. N . , POBEREZHSKIJ, V. A . & MAKHOTKO, V. F. 1979. Composition of olivine and monticellite MAKEYEV, B. V. & YEFIMOV, V. F. 1972. [Some
224
N. M. S. Rock
in monticellite-aln6ites of Siberia. Dokl. Acad. Sci. U.S.S.R. 244, 116-21. NIXON, P. H., MITCHELL, R. H. & ROGERS, N. W. 1980. Petrogenesis of aln6itic rocks from Malaita, Solomon Islands. Mineral. Mag. 43, 587-96. , THIRLWALL,M. F., BUCKLEY, F. & DAVIES, C. J. 1984. Spanish and Western Australian lamproites: aspects of whole-rock geochemistry. In: KORNPROST,J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 285-96. Elsevier, Amsterdam. , REX, D. C. & CONDLIFFE, E. 1986. A note on the age and petrogenesis of lamprophyre dykes of the Cautley area, Yorkshire Dales. Proc. Leeds geol. Assoc., in press. O'HARA, M. J., SAUNDERS, M. J. & MERCY, E. L. P. 1975. Garnet-peridotite, primary ultrabasic magma and eclogite: interpretation of upper mantle processes in kimberlite. Phys. Chem. Earth, 9, 571-604. OHASHI, F. 1980. An alkali olivine basalt and its related rocks from the Setogawa group, Shizuoka prefecture. J. geol. Soc. Japan, 86(12), 779-815. PAUL, D. K., POTTS, P. J., REX, D. C. & BECKINSALE, R. D. 1977. Geochemical and petrogenetic study of the Girnar igneous complex, Deccan volcanic province, India. Bull. geol. Soc. Am. 88, 227-34. PEARCE, J. A. 1982. Trace element characteristics of lavas from destructive plate boundaries. In: THORPE, R. S. (ed.) Andesites, pp. 525-48. Wiley, New York. 1983. Role of the sub-continental lithosphere in magma genesis at active contintental margins. In : HAWKESWORTH, C. J. & NORRY, M. J. (eds) Continental Basalts and Mantle Xenoliths, pp. 23049. Shiva, Orpington. PHILLIPS, W. J. 1973. Interpretation of crystalline spheroidal structures in igneous rocks. Lithos, 6, 235-44. PHILPOTTS, A. R. 1972. Density, surface tension and viscosity of the immiscible phase in a basic alkaline magma. Lithos, 5, 1-18. -1974. The Monteregian Hills. In: SORENSEN H. (ed.) The Alkaline Rocks, pp. 293-310. Wiley, New York. 1976. Silicate liquid immiscibility--its probable extent and petrogenetic significance. Am. J. Sci. 276, 1147-77. PIRSSON, L. V. 1905. Petrography and geology of the igneous rocks of the Highwood Mountains, Montana. U.S. Geol. Surv. Bull. 237. PLATT, R. G. & MITCHELL, R. H. 1979. The Marathon dykes. I: Zirconium-rich titanian garnets and manganoan magnesian ulvospinel-magnetite spinels. Am. Mineral. 64, 546-50. -& -1982. The Marathon dykes: ultrabasic lamprophyres from the vicinity of McKellar harbour, NW Ontario. Am. Mineral. 67, 907-16. PoPov, V. S. 1972. [Globular structure of lamprophyres.] Zap. Vses. Mineral. Obshch. 106(6), 37079 (in Russian). POWELL, J. L. & BELL, K. 1970. Strontium isotopic studies of alkalic rocks: localities from Australia, Spain and the western United States. Contrib. Mineral. Petrol. 27, 1-10. -
-
-
-
--,
HURLEY, P. M. & FAIRBAIRN, H. W. 1966. The strontium isotopic composition of carbonatites. In: TUTTLE,O. F. & GITTINS, J. (eds) Carbonatites, pp. 365-78. Wiley, New York. PRAEGEL, N.-O. 1981. Origin of ultramafic inclusions and megacrysts in a monchiquite dyke at Streap, Inverness-shire, Scotland. Lithos, 14, 305-22. PRIDER, R. T. 1982. A glassy lamproite from the West Kimberley area, Western Australia. Mineral. Mag. 45, 279-82. PRIEM, H. N. A., VERSCHURE,R. H., BOELRIJK, N. A., HEBEDA, E. H. & THORKILDSEN, C. D. 1968. R b Sr and K - A t age measurements on phlogopitic biotite from the ultrabasic lamprophyre dike on Ytteroy. Nor. geol. Tidsskr. 48, 319-21. RATCLIEFE, N. M., ARMSTRONG, R. L., MOSE, D. G., SENESCHAL,R., WILLIAMS,N. & BAIAMONTE,M. J. 1983. Emplacement history and tectonic significance of the Cortlandt complex, related plutons and dike swarms in the Taconide zone of New York, based on K-Ar and Rb-Sr investigations. Am. J. Sci. 282, 358-90. READ, H. H., PHEMISTER,J. & ROSS, G, 1926. Geology of Strath Oykell and lower Loch Shin. Mere. Geol. Surv. Scotl. RHODES, J. M. 1981. Characteristics of primary basaltic magmas. In: BVSP. Basaltic Volcanism on the TerrestrialPlanets, pp. 409-52. Pergamon, Oxford. RICHLY, J. E. 1939. The dykes of Scotland. Trans. Edinb. geol. Soc. 13, 393-435. ROCK, N. M. S. 1976. The comparative strontium isotopic composition of alkaline rocks; new data from southern Portugal and East Africa. Contrib. Mineral. Petrol. 56, 205-28. 1977. The nature and origin of lamprophyres: some definitions, distinctions and derivations. Earth Sci. Rev. 13, 123-69. 1978a. Petrology and petrogenesis of the Monchique alkaline complex, Portugal. J. Petrol. 19, 171-214. 1978b. The effect of CO2 on the liquidus relations of alkali basalts and its relevance to the nephelinite problem. N E R C Prog. Exp. Petrol. Publ. Ser. D, No. 1 l, pp. 164-70. 1979. Petrology and origin of the type monchiquites and associated lamprophyres of Serra de Monchique, Portugal. Trans. R. Soc. Edinb. 70, 149-70. -1980. Rare-earth and other trace element contents and the origin of minettes; critical comment on a paper by Bachinski & Scott (1979). Geochim. cosmochim. Acta, 44, 1385-8. -1981. How should igneous rocks be grouped? Geol. Mag. 118, 449-61. 1982. The Late Cretaceous alkaline igneous province in the Iberian Peninsula, and its tectonic significance. Lithos, 15, 111-31. 1983. The Permo-Carboniferous camptonitemonchiquite dyke-suite of the Scottish Highlands & Islands: distribution, field and petrological aspects. Inst. Geol. Sci. Rep. 82/14. 1984. The nature and origin of calc-alkaline lamprophyres : minettes, vogesites, kersantites and spessartites. Trans. R. Soc. Edinb. 74, 193-227. -
-
-
-
-
-
Nature of lamprophyres
-
1986. The nature and origin of ultramafic lamprophyres: aln6ites and allied rocks. J. Petrol., 27, 155-196. 1987. Kimberlites as varieties of lamprophyres. Proc. 4th Int. Kimb. Conf. (in press). & LEAKE, B. E. 1984. The International Mineralogical Association amphibole nomenclature scheme: computerisation and its consequences. Mineral. Mag. 48, 211-27. , BARNES;R. P., GASKARTH,J. W. & MACDONALD, R. 1985. Regional late Caledonian dyke-swarms of Southern Scotland. Geol. Soc. Can. Annu. Meet., Fredericton, NB, Progr. Abstr., A52. , GASKARTH, J. W. & RUNDLE, C. C. 1986. Late Caledonian dyke-swarms of southern Scotland : a regional zone of primitive K-rich lamprophyres and associated vents. J. Geol. 94, 505-22. --, COOPER, C. & GASKARTH, J. W. 1987. Late Caledonian subvolcanic vents and associated dykes in the Kirkcudbright area, SW Scotland. Proc. Yorks. Geol. Soc. (in press). RODEN, M. F. 1981. Origin of coexisting minette and ultramafic breccia, Navajo volcanic field. Contrib. Mineral. Petrol. 77, 195-206. -& SMITH, D. 1979. Field geology, chemistry and petrology of Buell Park minette diatreme, Apache county, Arizona. In : BOYD, F. R. & MEYER, H. A. O. (eds) Kimberlites, Diatremes and Diamonds, pp. 364-81. American Geophysical Union, Washington. ROSENBUSCH, H. 1897. Mikroskopische Physiographie der Mineralian und Gesteinen. Vol. II, Part I. Schweizbartsche, Berlin. ROSSEYKIN,L. V. & RAXHMANOV,A. M. 1971. [Relation of lamprophyre dikes to the ore mineralisation in the Maykurian deposit.] Dokl. Akad. Nauk Tadzh. S.S.R. 14, 44-7 (in Russian). RUCKLIDGE, J. C., BROOKS, C. K. & NIELSEN, T. F. D. 1980. Petrology of coastal dykes at Tugtilik, SE Greenland. Greenland Geosci. 3. SARKAR, A., PAUL, D. K., BALASUBRAHMANYAN,M. N. SENGUPTA, N. R. 1980. Lamprophyres from Indian Gondwanas--K-Ar ages and chemistry. J. geol. Soc. India, 21,188-93. SARKISYAN, G. A. 1973. [Lamprophyre dikes of the Azatek gold-antimony deposit.] Dokl. Akad. Nauk Arm. S.S.R. 57, 84-8 (in Russian). SATHE, R. V. & DESAI, S. D. 1968. On lamprophyrenepheline syenite composite intrusions of Mt. Girnar, Saurashtra. J. Sci. Technol. Poona Univ. 34, 77-80. SCOTT, B. H. 1979. Petrogenesis of kimberlites and associated potassic lamprophyres from central W Greenland. In: BOYD, F. R. & MEYER, H. A. O. (eds) Kimberlites, Diatremes and Diamonds, pp 190205. American Geophysical Union, Washington. 1981. Kimberlite and lamproite dykes from Holsteinsborg, W Greenland. Greenland Geosci. 4. SCOTT, P. W. 1980. Zoned pyroxenes and amphiboles from camptonites near Gran, Oslo region. Mineral. Mag. 43, 913-17. & MIDDLETON, R. 1983. Camptonite and maenite sills near Gran, Hadeland, Oslo region, Norway. Nor. Geol. Unders. 389, 1-26. -
-
-
-
-
225
SCOTTSMITH, B. H. & SKINNER,E. M. W. 1984a. A new look at Prairie Creek, Arkansas. In: KORNPROBST, J. (ed.) Kimberlites I: Kimberlites and Related Rocks, pp. 255-84. Elsevier, Amsterdam. & 1984b. Diamoniferous lamproites. J. Geol. 92, 433-8. SEELIGER, E. 1975. Was sind lamprophyre? Aufschluss, 26, 4-7. SHCHUKIN, S. I. 1974. [Lamprophyres and ores.] Geol. Rud. Mestorozhd. 16(6), 97-101 (in Russian). SNAVELEY,P. D. & WAGNER, H. C. 1961. Differentiated gabbroic sills and associated alkalic rocks in the central part of the Oregon coast range, Oregon. Res. U.S. Geol. Surv. 1961, Short Pap. geol. hydrol. Sci. Artic. 293-435, D156-60. SORENSEN, H. (ed.) 1974. The Alkaline Rocks. Wiley, New York. SPARKS, R. S. J., PINKERTON, H. & MACDONALD, R. 1977. The transport of xenoliths in magmas. Earth planet. Sci. Lett. 35, 234-8. SPERA, F. J. 1980. Aspects of magma transport. In: HARGREAVES, R. B. (ed.) Physics of Magmatic Processes, pp. 265-323. Princeton University Press, Princeton, NJ. STEELE K. F. & WAGNER G. H. 1979. Relationship of the Murfreesboro Kimberlite and other igneous rocks of Arkansas, U.S.A. In: BOYD, F. R. & MEYER, H. A. O. (eds) Kimberlites, Diatremes and Diamonds, pp. 393-9. American Geophysical Union, Washington. STEINER, H. 1984. Mineralogische, petrographische, geochemische und isotopengeologische Untersuchungen an einem Meta-Lamprophyre und seinem granodioritischen Nebengestein (MatorelloGneis) aus der Maggia-Decke. Schweiz. mineral. petrogr. Mitt. 64, 261-71. STRECKEISEN,A. 1979. Classification and nomenclature of volcanic rocks, lamprophyres, carbonatites and melilitic rocks. Geology, 7, 331-5 ; Neues J. Mineral. Abh. 134, 1-14. SUTHERLAND, D. S. 1982. Igneous Rocks of the British Isles, Wiley, New York. SUZUKI, K. & SHIRAKI, K. 1980. Chromite-bearing spessartites from Kasuga-Mura, Japan, and their bearing on possible mantle origin andesite. Contrib. Mineral. Petrol. 71, 313-22. THOMPSON, R. N., MORRISON, M. A., HENDRY, G. L. & PARRY, S. J. 1984. An assessment of the relative roles of crust and mantle in magma genesis: an elemental approach. Phil. Trans. R. Soc. Lond., Ser. A, 310, 549-90. THY, P. 1985. Contrasting crystallisation trends in ultrapotassic lamproites from Central West Greenland. Geol. Soc. Can. Annu. Meet., Fredericton, NB, Progr. Abstr., A63. TRONNES, R. G., EDGAR, A. D. & ARIMA, M. 1985. Titanium solubility in phlogopite: an experimental study on its potential as a P-T indicator for upper mantle-lower crustal rocks. Geol. Soc. Can. Annu. Meet., Fredericton, NB, Progr. Abstr., A63. TUTTLE, O. F. & GITTINS, J. 1966. Carbonatites. Wiley, New York. TYRRELL, G. W. & NEILSON, m. T. 1938. Igneous rocks from the neighbourhood of Mount Jombo and the -
-
226
N. M. S. Rock
Sabaki River. Monogr. Geol. Dep. Hunterian Mus. Glasgow V, Chapter IX, pp 108-16. UPTON, B. G. J. 1965. The petrology of a camptonite sill in south Greenland. Meddr Gronland, 169(11). --, ASPEN, P. & CHAPMAN, N. A. 1983. The upper mantle and deep crust beneath the British Isles: evidence from inclusions in volcanic rocks. J. geol. Soc. Lond. 140, 105-21. VAL'TER, A. A. & YEREMENKO, G. K. 1973. [Genesis and composition of augite megacrysts from camptonite dikes of the Eastern Azov region.] Dopov. Akad. Nauk Ukr. R.S.R., Ser. B, (10), 837-9 (in Russian ; English summary). - - - - & - - - - 1974. [Megacrysts of alkali feldspar in camptonite of the Azov region.] Dokl. Akad. Sci. U.S.S.R. Earth Sci. Sect. 217, 129-32 (in Russian). VAN BERGEN, M. J., GHEZZO, C. & RICCI, C. A. 1983. Minette inclusions in the rhyodacitic laMas of Mt. Amiata (central Italy): mineralogical and geochemical evidence of mixing between Tuscan and Roman-type magmas. J. Volcanol. geotherm. Res. 19, 1-35. VANIMAN, D., LAUGHLIN, A. W. & GLADNEY, E. S. 1985. Navajo minettes in the Cerros de las Mujeres, New Mexico. Earth planet. Sci. Lett. 74, 69-80. VAN SCHMUS, W. R. 1971. Ages of lamprophyre dikes and of the Mongowin pluton, N. shore of lake Huron, Ontario. Can. J. earth Sci. 8, 1203-9. VARTIAINEN, H., KRESTEN, P. & KAFKAS, Y. 1978. Alkaline lamprophyres from the Sokli complex, Northern Finland. Bull. geol. Soc. Finl. 50(1--2), 59-68. VELDE, D. 1967. Sur un lamprophyre hyperalcalin postassique. La minette de Sisco (Iie de Corse). Bull. Soc. fr. Mineral. Cristall. 90, 214-23. VENTURELLI, G., CAPEDRI, S., DI BATTISTINI, G. n., CRAWFORD, A., KOGARKO, L. N. & CELESTINI,S. 1984a. The ultrapotassic rocks of SE Spain. Lithos, 17, 37-54. - - - , THORPE, R. S., DAL PIAZ, G. V., DEL MORO, A. & POTTS, P. J. 1984b. Petrogenesis ofcalc-alkaline, shoshonitic and associated ultrapotassic Oligocene volcanic rocks from the Northwestern Alps, Italy. Contrib. Mineral. Petrol. 86, 209-20. VISSER, J. N. J. 1964. Analyses of rocks, minerals and ores. Geol. Surv. S. Air. Handbook 5. VITERBO, C. & ZANETTIN,B. 1959. I filoni lamprofirici dell'Alto Baltoro. Atti Mem. Acad. Patav. Sci. Lett. Arti, C1. Sci. Mat. Nat. 71, 3-39. VITRAC-MICHARD, A., ALBAREDE, F. & AZAMBRE, B. 1977. Age Rb-Sr et 39Ar-*~ de la syenite n~ph~linique de Fitou. Bull. Soc. p. Mineral. Cristall. 100, 251-4. VOLLMER, R., OGDEN, P., SCHILLING,J.-G., KINGSLEY,
R. H. & WAGGONER, D. G. 1984. Nd and Sr isotopes in ultrapotassic volcanic rocks from the Leucite Hills, Wyoming. Contrib. Mineral. Petrol. 87, 359-68. WAGNER, C. & VELDE, D. 1985. Mineralogy of two peralkaline, arfvedsonite-bearing minettes. A new occurrence of Zn-rich chromite. Bull. Min&al. 108, 173-87. WALLACE, R. C. 1975. Mineralogy and petrology of xenoliths in a diatreme from South Westland, N.Z. Contrib. Mineral. Petrol. 49, 191-9. WALTON, B. J. & ARNOLD, A. R. 1970. Plutonic nodules in lamprophyric carbonatite dykes from Frederikshaab, SW Greenland. Meddr. Gronland 190(5); Gronl. geol. Unders. 91, 1-25. WATSON, J. 1983. Lewisian. In: CRAIG, G. Y. (ed.) Geology of Scotland, pp. 23--48. Scottish Academic Press, Edinburgh. - 1984. The ending of the Caledonian orogeny in Scotland. J. geol. Soc. Lond. 141,193-214. WEDEPOHL, K. 1978. Handbook of Geochemistry. Springer, Berlin. WIMMENAUER, W. 1973a. Granites et lamprophyres. Bull. Soc. Geol. Fr. 15(3), 195-8. 1973b. Lamprophyre, Semilamprophyre und anchibasaltische Ganggesteine. Fortschr. Mineral. 51, 3-67. -& MULLER, H. 1985. Intermediate and acid dyke rocks of lamprophyric affinity. Geol. Soc. Can. Annu. Meet., Fredericton, NB, Progr. Abstr., A69. WRIGHT, A. E. & BOWLS, D. R. 1979. Geochemistry of the appinite suite. In : HARRIS,A. L., HOLLAND, C. H. & LEAKE, B. E. (eds) Caledonian of the British Isles Reviewed. Geol. Soc. Lond. Spec. Pub. 8, pp. 669 704. WRIGHT, J. B. 1971. The phonolite-trachyte spectrum. Lithos, 4, 1-5. MAGI, K., ISHIKAWA,H. • KOJIMA, M. 1975. Petrology of a lamprophyre sheet in Tanegashima Island, Kagoshima prefecture, Japan. J. Jap. Assoc. Mineral. Petrol. Econ. Geol. 70, 213-24. YEREMENKO, G. K. & SHVAKOVA, A. A. 1969. [Titanium-bearing garnet from the xenoliths of monchiquite dikes of the Azov sea area.] Dopov. Akad. Nauk Ukr. R.S.R., Ser. B, (11)971-4 (in Russian). YODER, H. S. 1979, Melilite-bearing rocks and related lamprophyres. In: YODER, H. S. (ed.) Evolution of the Igneous Rocks: 50th Anniversary Perspectives, pp. 391-411. Princeton University Press, Princeton, NJ. ZIMMERLE, W. 1977. Sph~irische Texturen in Lamprophyren aus der kristallinen Umrahmung des Oberrheingrabens. Oberrhein. geol. Abh. 26(1-2), 55-113.
NICHOLAS M. S. ROCK, Dept. Geology, University of Western Australia, Nedlands 6009, Western Australia.
Hawaiian alkaline volcanism David A. Clague SUMMA R Y:Hawaiian volcanoes erupt alkalic lavas during three of four eruptive stages. Stage 1 is characterized by submarine eruption of alkalic basalt and basanite followed by tholeiitic basalt which is exclusively erupted during stage 2. Stage 3 consists ofa post-calderacollapse capping of alkalic basalt and related differentiates. Stage 4 consists of post-erosional eruptions of alkalic basalt, basanite, nephelinite and melilitite. Eruption rates and volumes increase from stage 1 to stage 2 and then decline through stages 3 and 4. Stage 1 and 4 lavas contain deep-seated lherzolite xenoliths (even some garnet-bearing xenoliths in stage 4 lavas) whereas stage 2 lavas contain only very shallow cumulates, veins, dykes and sills; stage 3 lavas contain mainly dunite, wehrlite and various cumulates. The xenolith assemblages reflect the development of magma storage reservoirs which act as hydraulic filters in removing any xenoliths entrained below the magma storage reservoirs. The development of the magma storage system is a direct response to varying magma supply rates during the four eruptive stages. Trace-element, isotopic and rare-gas data all indicate that Hawaiian lavas are generated from heterogeneous sources consisting of at least three distinct compositions. Mixing of the sources or of partial melts derived from these sources is inevitable ; however, the characteristics of the mixing end-members are not well constrained. In a general way, early and late lavas are generated by small degrees of melting of sources enriched with less incompatible elements, whereas the voluminous shield tholeiites are generated by large degrees of melting of sources enriched with more incompatible elements.
Geological setting
Hawaiian eruption sequence
The Hawaiian Islands are the youngest part of a volcanic chain which extends nearly 6000 km across the north Pacific Ocean (Fig. I). This unique geological feature consists of more than 107 volcanoes with a combined volume of slightly greater than 1 million km 3 (Bargar & Jackson 1974). The chain is age progressive (Fig. 2) with volcanoes which are still active to the SE; the volcanoes at the north-western end are about 7580 Ma old (Dalrymple et al. 1981). Hawaiian volcanoes erupt lavas of distinct chemical compositions during different stages in their evolution and growth (Table 1). Three of these stages are characterized by eruption of alkalic to strongly-alkalic lavas. However, the lavas erupted during these alkalic stages probably represent less than 5~/ of the total volume of Hawaiian volcanoes. The remainder is composed of tholeiitic lavas that form the shield-shaped edifices for which the islands are well known. The distribution of the alkalic lavas of the Hawaiian Islands and Ridge and the Emperor Seamounts is reviewed in this paper, and their sequence, timing, volume and compositions, the distribution of the enclosed xenoliths and models for the generation of alkalic lavas in the mantle beneath Hawaii are discussed.
Hawaiian volcanoes evolve through four discrete eruptive stages starting with a pre-shield alkalic stage followed by a tholeiitic shield-building stage, a post-caldera or post-shield alkalic stage and a post-erosional or rejuvenated alkalic stage (Fig. 3). The three latest stages occur widely and are well studied and documented (Stearns 1940; Macdonald & Katsura 1964; Macdonald 1968), but the pre-shield alkalic stage, which includes the early phase of the submarine history of the volcano, has only been described recently from one volcano (Moore et al. 1982). Insight into the pre-shield stage has come from recent studies of Loihi Seamount, a small submarine volcano located about 30 km off the SE coast of Hawaii. The location, small size, seismic activity, fresh glassy lavas and active hydrothermal discharge all indicate that Loihi is an active volcano and the youngest in the linear HawaiianEmperor volcanic chain. Some of the older lavas recovered from Loihi Seamount are alkalic basalt and basanite, whereas most of the youngest lavas are tholeiitic and transitional basalt. This observation led Moore et al. (1982) to conclude,that Loihi Seamount, and perhaps all Hawaiian volcanoes, initially erupted alkalic lavas. In the shield-building stage tholeiitic basalt
Front: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 227-252.
227
228
D.A. 1620 24 o
Clague 160 o
158 o I
I
500
"~176
1560 I
154 o I
',.',00~------'i
20 o
I
I
I
I
400
DR BEND
Se~Ua~;uU~t 30 30oa -
-~/~ ~Colahan Se. . . . . . ~~'~
o ~
~ w o y ,. Pea;::nd. . . . . .
v ~..~ ~
,~x(j_lq
F..... Frigo,e
Necker l ~ - - ~ 9
I
I
1700
1800
I 170 ~
.~Kouai~50oh Hawaiior _~-r~e~) ~ . , t o n , , . ~ . Deep
I 160 ~
FIo. i. Location map of the Hawaiian-Emperor volcanic chain in the north-central Pacific Ocean. The bathymetry of the principal Hawaiian Islands and Loihi Seamount is shown in the inset (after Chase et al. 1970). Note the striking linearity of the chain and the pronounced bend between the Hawaiian and Emperor sections of the chain. Contour interval 1 km.
flows construct the main volcanic edifice in a relatively short time, perhaps 1 Ma or less (Jackson et al. 1972). Most of the volume of an individual volcano (probably more than 95%) is formed from these voluminous eruptions. The later part of the shield-building stage may include
caldera collapse and the eruption of calderafilling tholeiitic, transitional or alkalic Iavas. In the post-caldera stage a relatively thin cap of alkalic basalt and associated differentiated lavas cover the main shield. These alkalic lavas account for less than 1~ of the total volume of
Hawaiian alkaline volcanism 0
K - A r or 4 ~
9
M o s t reliable K - A r and 4 ~
z~
Fossil a g e s
229
ages ages
0
60 90
0
(3 40 --
<
- - -
HAWAIIAN
RIDGE
~1~
9
-EMPEROR
SEAMOUNTS~
O~ 0
o
o 9
~o I 0
i
2000 DISTANCE
I 4000
F R O M LOIHI S E A M O U N T
6000 (km)
FIG. 2. Plot of the age of the volcanoes in the Hawaiian-Emperor volcanic chain as a function of distance along the chain. For discussion of individual ages plotted see Clague & Dalrymple (1987). Note that in general the volcanoes of the chain increase in age to the N and W away from the Loihi Seamount, Kilauea and Mauna Loa.
TABLE 1.
Hawaiian eruptive products
Eruptive stage
Rock types
Eruption rate
Volume
Post-erosional stage
Alkalic basalt Basanite Nephelinite Nepheline-melilitite
Very low
< 1%
Post-caldera stage
Alkalic basalt Transitional basalt Ankaramite Hawaiite Mugearite Benmoreite Trachyte
Low
--~1%
Shield-building stage
Tholeiitic basalt Tholeiitic picrite Alkalic basalt (?)*
High
95%-98%
Pre-shield stage
Basanite Alkalic basalt Transitional basalt Tholeiitic basalt (?)*
Low
~ 3%
* Wright & Helz (1987) suggest that the shield stage may include rare intercalated alkalic basalt and that the pre-shield stage includes tholeiite. I suspect that tholeiite and alkalic basalt occur intercalated during the transitions from pre-shield stages and from shield stage to post-caldera stages, but that during the main shield-building stage only tholeiitic lavas are erupted. From Clague & Dalrymple 1987.
230
D. A. Clague Post-erosional alkalic . P o s t- c a Id e r a a Ik a li c ~
Ocean
":'('!i::::'::i'ii"i':~'ii'i"i'i':':
ROCK TYPES
~ ~ ' ....~m.",-d:.~,.,-~:~.~.:-..;.:.? ~-.-,:, ~::z:::::;
AIkalic bas alt
ii![i~i i
Nephelinite Melilitite
f.:3:.,:~.:.'...: ?:.:. >.f,
Hawaiite
:i'i:~:':~:i~'::'i"i:~'~-'i:i'i~':~:i~'i~'i"i~~':::::':i'i'i'i
ESTIMATEDVOLUME
<<1%
Alkalic basal! 1%
TMug r aceharite yte Tholeiitic basalt Tholeiilic picrite
97-98%
~ Transitionalbasalt Alkalic basalt Basanite
1-2%
FIG. 3. Cross-sectionthrough a hypothetical Hawaiian volcano showing the age and volume relations of the four eruptive stages. The main rock types and their relative volumes are also shown. the volcano and are erupted soon after cessation of the shield-building stage. The final stage occurs after as much as several million years of volcanic quiescence and erosion. During this post-erosional stage, very small amounts (less than 1%) of SiOz-poor lavas may erupt from satellite vents. An individual volcano may become extinct before this eruptive cycle is complete, but the general sequence is typical of the well-studied Hawaiian volcanoes. The rock types known to occur on each of the volcanoes in the HawaiianEmperor volcanic chain and their inferred eruptive stages are summarized in Table 2. Evidence for the pre-shield alkalic stage exists only at Loihi Seamount. If this stage is present at all Hawaiian volcanoes, it is completely buried by later shieldbuilding tholeiitic lavas. The post-caldera alkalic stage lavas occur relatively late in the eruptive sequence. A number of the volcanoes have erupted predominantly mugearite whereas others have erupted predominantly hawaiite; these are called Kohala type and Haleakala type respectively (Macdonald & Katsura 1962). Farther W along the Hawaiian Ridge and in the Emperor Seamounts, lavas thought to have been erupted during the post-caldera alkalic stage are quite common. Post-erosional alkalic-stage lavas occur on Haleakala, Maui, E Molokai, the Koolau and Waianae volcanoes on Oahu, Kauai, Niihau, Kaula Island and perhaps Kahoolawe. To the W along the Hawaiian Ridge, lavas thought to be post-erosional are scarce but have been recovered from five seamounts and islets. Post-erosional lavas have not been recovered from the Emperor Seamounts.
The assignment of lavas to volcanic stages is difficult for samples recovered by drilling and dredging since the field relations are essentially unknown. We have used geochemical and mineralogical criteria developed from lavas of known volcanic stages in the principal islands (Clague et al. 1980; Frey & Clague 1983) to assign samples recovered from the Hawaiian Ridge and Emperor Seamounts. In general, lavas from the different stages are distinct rock types with the exception of alkalic basalt which can occur in any of the three alkalic stages. At Ojin Seamount and Suiko Seamount drilling has recovered alkalic lavas overlying tholeiitic lavas, as in the Hawaiian Islands. In both cases the overlying lavas are interpreted to be postcaldera-stage alkalic lavas (Kirkpatrick et al. 1980). These observations suggest that the sequence of stages observed in the Hawaiian Islands has occurred repeatedly throughout the nearly 80 Ma history of the Hawaiian-Emperor volcanic chain.
The pre-shield alkalic stage Loihi Seamount is located to the SE of Kilauea and Mauna Loa volcanoes and lies 30km offshore. The seamount rises to within 950 m of the sea surface. Two distinct rift zones extend N and S from a summit caldera 2.8 km wide and 3.7 km long (Malahoff et al. 1982). Moore et al. (1979) describe tholeiitic basalt dredged from Loihi Seamount in 1978 that is fresh and glassy. The young age of the lavas, coupled with seismicity (Klein 1982) and active hydrothermal discharge above the volcano (Malahoff et al. 1982; Horibe et al. 1983; Kim & Craig 1983),
Hawaiian alkaline volcanism TABLE 2.
23I
Rock types and inferred volcanic stages represented along the Hawaiian-Emperor chain Shield-stage 1
2
Kilauea Mauna Loa Hualalai Mauna Kea Kohala
x x x x x
x x x x
Haleakala Kahoolawe W Maui E Molokai Lanai
x x x x x
x x x x
W Molokai Koolau Waianae Kauai Niihau
x x x x x
x x x x x
Kaula Nihoa No. 19 No. 20 No. 21
x x
Necker La Perouse Pinnacles Gardner Pinnacles Brooks Bank St. Rogatein Bank Laysan Island Northampton Bank Pioneer Bank Pearl and Hermes Reef Ladd Bank
Post-erosional alkalic stage
Post-caldera alkalic stage 3
4
5
X
X
X
X
6
7
X
X
9
X
X
X
11
X
X
12
13
14
X
X
X
X
x X
10
X X
X
8
x? X
X
X
X
X
X X
X
X
X
X
X
X
X
X
X X X
X X
X X X
X
x x
x x x x x
x?
x
x?
x X
• •
X
Midway Island Nero Bank No. 57 No. 63 Colahan Seamount
x
Abbott Seamount Daikakuji Yuryaku Kimmei Koko (S)
x (T) •
Koko (N) Ojin Jingu Nintoku Suiko (No. 90)
x x
Suiko Meiji
x x
X
x X
X
X
X
x X*
x (T)
x*
x
X
X
X
X
X
x?
x
From Clague & Dalrymple 1987. 1, Tholeiite; 2, tholeiitic picrite; 3, rhyodacite; 4, alkalic basalt; 5, ankaramite; 6, hawaiite; 7, mugearite; 8, trachyte; 9, phonolite; 10, alkalic basalt; 11, basanite; 12, nephelinite; 13, nepheline-melilitite; 14, tephrite. (T) indicates transitional basalt. * Dredges from seamount no. 63 and Colahan Seamount recovered ankaramite, tephrite, and amphibole-bearing hawaiite that are probably all post-erosional stage lavas.
23 Z
D. A. Clague
indicate that Loihi Seamount is an active volcano and the youngest edifice in the H a w a i i a n Emperor volcanic chain. Moore et al. (1982) describe fresh glassy pillow lavas dredged from Loihi Seamount in 1981 that include basanite, alkalic basalt, basalt transitional between alkalic and tholeiitic basalt, and tholeiitic basalt. They also demonstrated, using thicknesses of palagonite, that the alkalic lavas are generally older than the tholeiitic lavas and suggested that Loihi Seamount, and presumably all Hawaiian volcanoes, have a pre-shield alkalic stage characterized by infrequent small-volume eruptions of alkalic lavas. Representative analyses of the Loihi lavas are given in Table 3. I selected a subset of the Loihi lavas for detailed analysis by a group of investigators for trace elements, radiogenic and stable isotopes and rare gases. Results from many of these studies are presented in a special volume edited by Craig (1983). The results of these detailed studies have broad implications for the generation of Hawaiian alkaline magmas. The broad range of rock types recovered from Loihi Seamount is shown in Fig. 4. Unlike most other Hawaiian volcanoes, the compositional range is continuous from tholeiite to alkalic basalt rather than plotting as discrete tholeiitic and alkalic clusters of analyses.
The lava types recovered have minor-element, trace-element and isotopic characteristics similar to those of sub-aerial Hawaiian lavas (Frey & Clague 1983; Lanphere 1983; Staudigel et al. 1984). These similarities confirm that Loihi Seamount is indeed a Hawaiian volcano. These same data also indicate that the lavas from Loihi Seamount are generated from heterogeneous mantle sources and that the lavas have equilibrated within the garnet stability field (Frey & Clague 1983). Trace-element ratio-ratio plots define linear arrays indicative of mixing but do not identify two end-members; this implies that more than two mixing components are required (Frey & Clague 1983). Radiogenic isotope studies indicate that three or more source compositions are involved in the generation of these lavas (Staudigel et al. 1984). In addition, there is no obvious correlation between radiogenic isotopic ratios of Pb, Nd or Sr and the bulk composition of the lavas or their eruptive ages (Lanphere 1983 ; Staudigel et al. 1984). Helium, argon and xenon isotopic data for glass from the Loihi lavas show that the Loihi samples are strongly enriched in 3He/4He compared with mid-ocean ridge basalt, which is interpreted as indicating a primitive u n d e g a s s e d mantle source for the volatile elements (All6gre et al. 1983 ; Kaneoka et al. 1983 ; Kurz et al. 1983;
TABLE 3. Major-element analyses (weight %) o f Loihi Seamount lavas Tholeiitic and transitional basalts 29-10
23-3
21-2
31-12
4 8 . 4 4 7 . 6 4 6 . 9 46.9 48.9 12.4 12.4 12.5 12.5 14.3 2.02 3.41 3.22 2.78 3.64 10.51 8.63 9.25 9.83 8.96 10.0 7.92 9 . 6 3 9.86 4.65 9.47 10.6 12.6 11.0 11.4 2.19 2.30 2.40 2.35 3.50 0 . 3 4 0 . 4 9 0.55 0.49 0.88 2.37 2.83 2.63 2.34 3.56 0.23 0.26 0.27 0.25 0.46 0.18 0.18 0.19 0.18 0.19 0.34 0.68 0.51 0.52 0.70 0.35 0.40 0 . 1 6 <0.01 0.14 <0.01 <0.01 <0.01 <0.01 <0.01
44.4 10.6 2.24 10.22 16.3 10.1 1.91 0.52 1.68 0.19 0.19 0.57 0.03 0.01
SiO2 A1203 Fe203 FeO MgO CaO Na20 K20 TiO2 P205 MnO H:O + H20CO2
48.7 12.3 2.06 9.57 10.7 10.6 1.96 0.32 2.22 0.20 0.17 0.36 0.01 <0.01
Total
99.2
99.9
Quartz -Hypersthene 22.2 Nepheline --
-17.2 --
1 8 - 8 25-4
99.7 -
99.2
-
8.7 .
-
-
9 9 . 4 99.3 -
6.7 .
24-7
Alkalic basalts
.
-
3.9 .
0.7 10.4
27-4
1 7 - 2 20-4
1 7 - 1 7 15-4
43.1 46.3 45.7 9.49 13.6 13.2 2.68 3.65 3.50 10.19 8.50 8.55 18.5 6.58 7.08 9.55 11.8 12.0 1.93 2.75 3.13 0 . 6 2 0.91 0.99 1.94 3.27 3.37 0.23 0 . 3 8 0.45 0 . 1 9 0 . 1 7 0.16 0 . 7 0 1 . 1 5 0.96 0 . 0 6 0 . 2 0 0.14 <0.01 <0.01 <0.01
42.4 44.7 10.8 14.6 3.99 4.29 9.73 7.75 ll.9 5.52 11.1 10.5 3.13 3.69 1.23 1.61 3.03 3.74 0.47 0.59 0.18 0.17 0.96 1.35 0.15 0.70 <0.01 0.01
99.0 . . 2.4
Basanite
9 9 . 2 99.3 . .
. . 4.7
. . 0.4
99.2
99. l
99.2
4.4
11.8
7.0
From Frey & Clague 1983. Major-element abundances determined by X-ray fluorescence (analysts: J. S. Wahlberg, J. Taggart and J. Baker, U.S. Geological Survey, Denver, CO). Analyses for FeO, H20 +, H20- and CO2 by classical techniques (analysts: H. Neiman and E. Engleman, U.S. Geological Survey, Denver, CO).
Hawaiian alkaline volcanism
233 /
Alkalic b a s a l t J
Thole,re 8
~
Basanite
~]
Transitional
ALKALIC ~
FIELD /
~6
J
%4 2
0
~ 34
I 38
t
I 42
--am
I 46
J
l 50
t
I 54
i
I 58
I
I 62
i 66
S i O 2 ( w t %)
FiG. 4. Alkali-silica diagram for lavas from Loihi Seamount. Fields enclose whole-rock analyses of tholeiitic, transitional and alkalic basalts and basanite. The Loihi lavas are a continuum of compositions from alkalic basalt to tholeiite. Analyses have been published for 12 samples (Frey & Clague 1983); the other 18 are unpublished data of D. A. Clague & J. G. Moore. Rison & Craig 1983). This mantle source has been isolated for much of Earth history and may represent the 'plume' component in Hawaiian lavas. Another lower 3He/4He component resides in olivine xenocrysts and dunite xenoliths and apparently contaminates the 3He/4He of the alkalic lavas. The volcanoes on Hawaii show an inverse correlation of 3He/~He with the volume of the volcano, suggesting that the melts producing smaller edifices are generated from bulk sources having larger percentages of the primitive undegassed mantle component. The Loihi lavas are notably vesicular considering their depth of eruption (1-2.6 kin). Moore & Clague (1982) note that, during vesicle formation, S, CO2 and perhaps up to 0.2 wt.% H20 are lost from the most vesicular alkalic melts. The alkalic lavas have higher FezOJFeO ratios than the tholeiitic lavas, indicating eruption under higher oxygen fugacity. The higher oxygen fugacity could result from loss of hydrogen from the alkalic lavas during vesiculation. Water behaves as an incompatible element in the Loihi lavas and has a positive correlation with K20 such that H20 + --0.9K20. The high gas content of these early alkalic lavas provides a mechanism for initiation of volcanism on the sea-floor and for buoyant transport of magma through the thick relatively cool lithosphere (Moore & Clague 1981). A few of the most alkalic flows from Loihi Seamount contain ultramafic xenoliths. Most of these xenoliths are dunite, but rare lherzolite and olivine cumulates are also present. The xenoliths contain abundant CO2-rich inclusions that were
trapped at depths of 10-18 km. Other decrepitated inclusions indicate that at least some of the xenoliths originated even deeper (Roedder 1983). The mineral chemistry of these dunite xenoliths (Clague et al. 1982b) is similar to that of dunite from the post-erosional Honolulu Volcanics on Oahu which are thought to be recrystallized cumulates related to Hawaiian lavas (Sen 1983). The lherzolite xenoliths occur in several of the older and most primitive alkalic lavas. The lava flows are apparently of small volume since 65 chemically distinct flows were recovered in 18 dredges. Thus the early stages of the growth of Hawaiian volcanoes are characterized by small-volume eruptions of basalt with a large compositional range including alkalic and tholeiitic lavas. Loihi Seamount has already evolved to the point where intrusion of magma occurs frequently enough to sustain a shallow-level magma chamber, believed to underlie the summit caldera, in which many of the younger lavas have differentiated and perhaps mixed.
Tholeiitic shield-building stage The lavas erupted during this stage are tholeiitic basalt and tholeiitic picrite. Occasionally, a pocket oftholeiitic magma is isolated long enough to differentiate to basaltic andesite, andesite (icelandite) and rhyodacite (Sinton 1981). The tholeiitic lavas from each volcano are geochemically distinct (Wright 1971 ; Leeman et al. 1977; Basaltic Volcanism Study Group 1981; Budahn & Schmidt 1984). These sub-alkalic lavas are discussed in detail in these references and they
234
D. A . C l a g u e Likewise, post-caldera alkalic-stage lavas have been recovered from eight of eleven seamounts sampled in the Emperor Seamounts. Analyses of post-caldera alkalic-stage lavas from the principal islands are presented by Macdonald (1968) and Macdonald & Katsura (1964). Several of the volcanoes have been studied in some detail, including Hualalai volcano (Clague et al. 1980), Kohala (Feigenson et al. 1983; Lanphere & Frey 1987), Haleakala (Chen & Frey 1983, 1985) and E Molokai (Beeson 1976; Clague & Beeson 1980; Clague et al. 1983). Additional detailed studies of Mauna Kea (West & Garcia 1982; Wise 1982), and W Molakai (Clague, unpublished data) are in progress. Representative analyses are given in Table 4 and plotted on an alkali-silica diagram in Fig. 5. Samples recovered from the Hawaiian Ridge and Leeward Islands include hawaiite and mugearite cobbles from Laysan Island (Dalrymple et al. 1981) and Midway Atoll (Dalrymple et al. 1977) and miscellaneous samples of alkalic basalt, hawaiite, mugearite and phonolite from other volcanoes (Clague 1974; Garcia et al. in press). Samples from the Emperor Seamounts are alkalic basalt, hawaiite or mugearite, except for a large collection of cobbles from Koko Seamount that includes alkalic basalt, hawaiite, mugearite, benmoreite, sanidine trachyte, anorthoclase trachyte and phonolite (Clague & Greenslate 1972; Clague 1974). These diverse lavas can be related to one another by crystal fractionation of augite, olivine, plagioclase, ilmenite, apatite, titanomagnetite and sanidine.
will not be discussed further here. Throughout this eruptive stage, intrusion of magma occurs frequently enough to sustain a shallow-level magma chamber. It is also probable that a deeper magma storage site develops near the base of the oceanic crust. Post-caldera alkalic stage Lavas from the post-caldera alkalic stage consist mainly of alkalic basalt, hawaiite and mugearite with smaller amounts of ankaramite, trachyte and phonolite. These flows form a thin, usually discontinuous, cap on the tholeiitic shield. Eruption of these alkalic lavas commonly follows collapse of the summit caldera and can occur as part of of the caldera fill or on tholeiitic calderafilling lavas. The eruptions are commonly explosive and form cinder cones that dot the summit regions of the volcanoes. The vents are commonly concentrated in and around the caldera and along the rift zones established during the shieldbuilding stage. The hiatus between the tholeiitic shield stage and the post-caldera alkalic stage is generally of short duration, if it exists at all. The post-caldera alkalic stage has yet to develop on Loihi, Kilauea and Mauna Loa volcanoes. It is present on all the other volcanoes in the Hawaiian Islands except Lanai and Koolau on Oahu (Table 2) although the volumes present on Kauai, Niihau, Kahoolawe and W Molokai are small. In the Leeward Islands and the Hawaiian Ridge, lavas thought to have erupted during the post-caldera alkalic stage are common.
TABLE 4. Major-element compositions (weight %) of representative Hawaiian post-caldera alkalic-stage lavas Volcano Volcanic unit Rock type Sample no.
Mauna Kea
Hualalai Hualalai AOB
Waawaa T
65HU-101 HAW-152
Hamakua AOB
Kohala
Haleakala
Hawi
Kula
Laupahoehoe A
H
C 773 C-673
C-793
AOB
H
M
B
KHll 4 MG9a4MG4b4KHI 4
SiO2 A120 3 Fe20 3 FeO MgO CaO Na_,O K,O TiO, P20 s MnO H20 + H20CO 2
46.62 14.72 2.00 10.71 9.10 10.04 2.82 0.91 2.37 0.27 0.19 0.00 0.04 0.01
62.99 18.04 4.42 -0.32 0.71 7.19 4.87 0.45 0.13 0.29 1.31 -0.11
46.65 14.72 3.63 8.98 7.96 10.14 2.53 0.80 3.22 0.33 0.18 0.63 0.51
44.54 13.52 4.47 7.76 11.92 10.71 1.69 0.59 2.60 0.32 0.18 0.52 0.77
49.79 18.37 6.03 4.82 4.01 6.68 4.58 2.08 2.51 0.56 0.22 0.32 0.24
47.99 15.97
Total
99.88
100.83
100.28
99.59
100.21
100.76
12.31 6.49 10.70 2.80 0.80 3.08 0.42 0.20 . .
51.13 13.61 -13.11 4.45 8.19 3.02 1.10 3.50 0.66 0.26 . . . .
53.18 18.15
99.03
99.31
9.66 2.89 5.40 4.04 2.11 2.14 1.48 0.26 . .
AOB
H
M
C -140s C -1435 C-1415
59.04 18.37 -6.26 1.73 3.39 6.07 2.71 1.39 0.59 0.23
44.72 13.86 5.69 9.82 5.07 10.89 3.07 1.09 3.96 0.61 0.23 0.40 0.33
45.66 17.01 4.96 7.73 4.60 8.09 5.01 1.55 3.41 0.75 0.20 0.37 0.40
51.90 17.10 3.38 6.64 2.26 5.70 6.65 2.72 2.11 0.84 0.17 0.34 0.29
99.78
99.74
99.74
100.10
Hawaiian alkaline volcanism
235
TABLE4--Continued Volcano
W. Maui
E. Molokai
W. Molokai
Niihau
Volcanic unit
Honolua
E. Molokai
W. Molokai
Paniau
H
AOB
Rock type
Sample no.
H
M
B
C-1534 C-1554 C-1514
SiO2 Al:O3 FezO3 FeO MgO CaO Na20 KzO TiO2 P205 MnO H20 + H20CO2
47.74 13.35 4.26 10.35 5.09 9.90 2.78 0.61 3.69 0.59 0.19 0.56 0.62 .
55.26 18,16 4.54 4.04 1.97 3,81 5.58 2.45 1,51 1.06 0.23 0,66 0.77 .
56.08 18.75 6.59 1.02 0.95 2.82 5.57 2.69 1.40 0.69 0.15 1.45 1.39 .
Total
99.73
100.04
99.55
Volcano
Midway Atoll
Rock type
AOB
H
M
H
M
KLPA-286 KLPA-296 KLPA-336
.
45.64 14.04 4.09 10.16 8.11 9.17 2.96 0.98 3.30 0.59 0.20 0.36 0.26 .
48.15 17.20 4.52 7.20 3.78 9.36 3.66 1.12 3.40 0.62 0.19 0.20 0.28 .
51,40 17.67 5.11 4.95 2,97 6,04 5.08 2.00 2,34 1.09 0.21 0.38 0.41
99.97
99.72
99.69
AOB
H
M
B
SiO2 A1203 F%O 3 FeO MgO CaO Na20 K20 TiOz P205 MnO H2 O+ H20CO2
48.60 15.26 8.01 2.84 5.20 7.94 3.54 1.68 2.69 0.39 0.35 2.17 1.26 --
51.04 16.44 8.28 1.87 2.72 5.25 4.46 2.64 2.40 0.85 0.27 1.83 0.93 --
44.4 15.2 -11.9 5.3 8.5 3.2 1.7 4.3 1.0 0.13 ----
45.6 15.7 7.2 6.3 3.5 6.7 4.4 1.8 2.7 2.5 0.17 1.49 0.64 0.01
47.4 16.1 9.5 3.0 2.4 6.8 4.6 2.4 2.5 2.1 0.14 0.91 0.89 0.05
52.7 17.7 3.2 6.3 1.2 2.8 5.4 3.5 0.91 1.30 0.03 1.50 1.30 0.08
Total
99.93
98.98
95.7
98.9
98.9
97.9
M
7ON11-237
46.4 15.5 7.35 6.7 3.81 6,73 4.10 1;31 4.15 0.79 0.t5 1.2 0.62 <0.01
47.46 15.22 4.55 8.25 6.05 9.97 3.08 0.48 3.31 0.40 0.15 0.54 0.55 0.01
46.76 14.81 -11.81 5.27 9.86 3.68 1.29 3.88 0.62 0.15 ---
50.38 17.16 -8.47 2.97 6.13 4.37 2.61 2.11 0.70 0.13 ---
--
--
100.02
98.15
95.03
98,8
D1-2 s D I - l l s
Ojin Nintoku Suiko Seamount Seamount Seamount T
11899 1 1 8 6 9
H
PH-37
KokoSeamount
Sampleno.
Laysan Is.
T
P
43-821~ 43 721~ 43-4 l~ 43-67 l~ 43 86 l~ 43-781~ 43-701~ 59.8 18.0 4.8 2.2 0.5 1.9 5.8 5.6 0.40 0.16 0.16 0.76 0.44 0.10 100.7
H
AOB
AOB
Avg 11
Avg 11
Avg 11
46.3 16.5 6.3 6.2 5.1 8.6 3.6 1.2 2.7 0.50 0.15 1.5 1.3 0.10
45.5 14.3 6.2 7.8 5.9 9.3 3.4 0.87 3.0 0.39 0.19 1.1 1.0 0.05
63.9 17.5 2.8 0.4 0.1 0.4 7.l 5.5 0.02 0.04 0.09 0.44 0.30 0.08
59.2 19.2 2.8 0.3 0.1 0.4 8.6 4.9 0.02 0.05 0.18 2.47 0.27 0.17
47.8 15.9 8.0 4.1 3.8 6.6 4.3 1.7 2.9 1.2 0.10 1.2 2.1 0.06
98.7
98.7
99.8
100.1
100.0
Rock types : AOB, alkalic olivine basalt; A, ankaramite; H, hawaiite; M, mugearite; B, benmoreite; T, trachyte; P, phonolite. Sources: 1, Clague et al. 1980; 2, Basaltic Volcanism Study Project 1981; 3, Macdonald 1968; 4, Feigenson e t al. 1983; 5, Macdonald & Powers 1968 ; 6, Beeson 1976; 7, Clague, unpublished data; 8, Dalrymple e t al. 1977; 9, Dalrymple et al. 1981 ; 10, Clague 1974; 11, Kirkpatrick et al. 1980.
The compositions of many of the lavas erupted during the post-caldera alkalic stage reflect differentiation due to crystal fractionation. The exact site of magma storage and differentiation is less clear, although evidence from Hualalai Volcano in particular suggests that fractionation leading to hawaiite occurs in a magma storage zone at greater than 15 km depth rather than in a shallow sub-caldera magma chamber. The key evidence supporting this conclusion is that fractionated lavas contain recrystallized dunite, and
wehrlite xenoliths contain C O 2 vapour inclusions trapped at a depth of about 15 km or greater (Roedder 1965; Clague et al. 1980; Kirby & Green 1980). The presence of these xenoliths in fractionated magmas implies that the parental magmas fractionated at depths greater than 15 km prior to incorporating the xenoliths during migration to the surface. At shallower levels these fractionated lavas also incorporate cumulate gabbroic xenoliths related to oceanic crust, the earlier tholeiitic shield-building stage and earlier
236
D. A. Clague p
,
,
,
!
,
i
,
i
,
i
,
,
,
Kohala Hualalai Mauna Kea
I0
Alkalic field
,
/
o
+ o z
6 oleiitic field
4
-
Tholeiitic field L~
Haleakala
./
0 (a)
J ~
,
i
i
I
I
L
I
I
i
I
L
(b)
14 ~]
fl
Koko
M i d w a y atoll [5555] Laysan I s l a n d
/ ~
Suiko
3
oN
l/
(
~
--
2 /
I0 [Z]
Alkalic field 8
+ o
6
z
4
5'
leiitic field
/
A///~
~
.~ ~"
{r
i ,-~
54
38
42 ' 416
Alkalic field
5~0
i
54
/
~
West M o l o k a i
~
Niihau
5~8
i
6~2
Q_b~q:~
i
Si02 (wt.%)
~
~ ~ T h o l e i
East Molokai
E'~ Waianae
~ /
0
Nintoku
i
66
34
I ~
38
I
42
itic field
J
4'6
i
I
5o
i
5'4
I
I
58
i
~2
L
ee
Si02 (wt.%) (d)
FIG. 5. Alkali-silica diagrams for lavas from the post-caldera alkalic stage: (a) lavas from Hualalai, Mauna Kea and Kohala volcanoes on the island of Hawaii (data from Macdonald & Katsura (1964), Macdonald (1968), Clague et al. (1980), Feigenson et al. (1983) and Lanphere & Frey (1987)); (b) lavas from Haleakala and W Maui (data from Macdonald & K atsura (1964), Macdonald (1968) and Macdonald & Powers (1968)); (c) lavas from E and W Molokai, Waianae and Niihau (data for Waianae from Macdonald (1968), for E Molokai from Beeson (1976) and for W Molokai and Niihau from Clague (unpublished data)); (d) lavas from Midway Atoll (Dalrymple et al. 1977), Laysan Island (Dalrymple et al. 1981), Koko Seamount (Clague 1974) and Ojin, Nintoku and Suiko Seamounts (Kirkpatrick et al. 1980). alkalic magmas of the post-caldera alkalic stage (Clague, unpublished data). The eruptions are generally less voluminous and less frequent than eruptions of the tholeiitic shield-building stage, although individual eruptions may be comparable in volume with the largest tholeiitic-shield-stage eruptions (Macdonald & Abbott 1970). In general the post-caldera alkalic-stage lavas have less radiogenic Sr and more radiogenic Nd isotopic ratios than the underlying tholeiites (Clague 1982; C h e n & Frey 1983; Clague et al. 1983; Lanphere & Frey 1987) but have greater enrichment of incompatible trace elements and higher ratios of light rare-earth elements to heavy rare-earth elements. The enrichment of incom-
patible trace elements leads to post-caldera alkalic-stage lavas with higher Rb/Sr ratios and lower SVSr/86Sr ratios than the underlying tholeiites. These inverse correlations between isotopic ratios and their respective parent-todaughter abundance ratios can be explained by mixing of melts derived from enriched and depleted sources (Chen & Frey 1983, 1985) and will be discussed in a later section. The post-caldera alkalic-stage lavas contain a variety of xenoliths including dunite, wehrlite, various cumulate rocks and fragments of dykes, sills and veins. In general, the percentages of dunite and wehrlite decrease in the more differentiated lavas whereas the percentages of dykes, sills and cumulates all increase. In addition, the
Hawaiian
alkaline volcanism
cumulate xenoliths in the more differentiated lavas are dominantly pyroxene cumulates whereas those in alkalic basalt are dominantly olivine cumulates (Jackson 1968). Rare-gas compositions of dunite are similar to those in midocean ridge basalt and suggest that these dunite xenoliths may be related to the formation of oceanic crustal rocks (Kaneoka & Takaoka 1978, 1980; Kurz et al. 1983). The xenolith assemblages are notable for their lack of lherzolite or any type of garnet-bearing peridotite. Post-erosional alkalic stage Following the post-caldera alkalic-stage Hawaiian volcanoes are dormant for up to 2.5 Ma during which time deep erosional valleys are cut in the tholeiitic shield and overlying post-caldera alkalic cap. On many of the volcanoes this dormant period is ended by eruption of very small volumes of alkalic to strongly-alkalic lavas, generally from vents unrelated to the pre-existing rift system. The youngest volcano to have posterosional alkalic-stage eruptions is Haleakala. Further W, Kahoolawe, W Maui, E Molokai, Koolau and Waianae on Oahu, Kauai, Niihau and Kaula Island have post-erosional lavas, although only Haleakala, Koolau, Kauai and Niihau had more than a few small post-erosional eruptions. Lavas erupted during this stage are dominantly alkalic basalt and basanite, but nephelinite and nepheline melilitite are common among the posterosional Honolulu Volcanics on the Koolau shield on Oahu, and the Koloa Volcanics on Kauai (Winchell 1947; Macdonald 1968; Jackson & Wright 1970; Clague & Frey 1982; Clague et al. 1982; Chen& Frey 1983). Many of these lavas are primitive as shown by their high MgO and Ni contents and the presence of mantle xenoliths. Representative analyses are given in Table 5 and plotted on an alkali-silica diagram in Fig. 6. The erosional period between the post-caldera and post-erosional alkalic stages is not of random duration (Fig. 7) but increases systematically from less than 0.4 Ma at Haleakala to about 2.5 Ma at Niihau (Clague et al. 1982a). Jackson & Wright (1970) proposed that the Honolulu Volcanics might have been generated by uplift as Oahu passed over the Hawaiian Arch. They argued that the Hawaiian Arch, an isostatic response to volcanic loading of the oceanic crust, followed the active volcanic centre by several hundred kilometres and several million years. Clague & Dalrymple (1987) showed that the rate of volcanic migration along the chain is increasing and that the post-erosional volcanic stages follow the formation of the shields not by a constant time interval but by a constant distance
237
of 190_ 30 km. It seems likely that the energy source for post-erosional magma generation is the rapid change from subsidence to uplift as the volcanoes override the Hawaiian Arch. Post-erosional lavas are quite rare from the seamounts to the W of the principal Hawaiian Islands (Table 2), possibly reflecting the lack of large volcanic edifices capable of flexing the lithosphere sufficiently to provide the energy source for magma generation but also reflecting the small chance of recovering these lavas which cover only small areas of the principal islands. The lack of post-erosional lavas from the Emperor Seamounts probably reflects the wide spacing of volcanoes in the Emperor Seamounts and the young thin oceanic crust on which these volcanoes were built. By the time the next younger volcano formed, the previously constructed volcano was already beyond the zone of flexure. The young oceanic crust beneath the Emperor Seamounts results in a smaller distance between the load and the surrounding arch because the lithosphere is less rigid. Post-erosionallavas from the Honolulu Volcanics, the Koloa Volcanics and Kaula Island contain abundant xenoliths of spinel lherzolite and dunite (Jackson 1968; Garcia et al., in press). In addition, several vents in the Honolulu Volcanics, notably Salt Lake Crater, and the tuff at Kaula Island, contain garnet-bearing peridotite (Jackson & Wright 1970; Garcia et all, in press). Xenoliths are rare or absent in post-erosional lavas from Niihau, E Molokai, W Maui and Haleakala. The garnet-bearing peridotite xenoliths originate at depths greater than 50 km and the lherzolite originates at depths of the order of 30-50km (Sen 1983). Clague & Frey (1982) conclude that the varieties of xenoliths in the Honolulu Volcanics lavas are all unrelated to the magmas that brought them to the surface. Frey (1980, 1984) and Frey & Roden (1987) suggest that the spinel lherzolite xenoliths represent depleted oceanic lithosphere modified by processes related to Hawaiian magmatism. The presence of relatively deep-seated xenoliths in some of the post-erosional lavas suggests that (1) the host magmas continuously ascend more rapidly than the xenoliths sink and (2) the host magmas cannot therefore slow or stop in any magma storage reservoirs between the depth at which they incorporate the xenoliths and the surface. In several detailed studies of post-erosional lavas, particularly the Honolulu Volcanics and the Hana Volcanics on Haleakala, it is concluded that the mantle source for these lavas is enriched in incompatible elements and that the enrichment event occurred relatively recently (Clague & Frey
238
Clague
D. A.
TABLE 5. Major-element composition (weight %) o f representative Hawaiian post-erosional alkalic-stage lavas Volcano Volcanic unit
Haleakala Hana
W Maui Lahaina
E Molokai Kalaupapa
Niihau Kiekie
Ladd Bank
Rock type
N
B
B
AOB
B
AOB
AOB
N
Sample no.
C1391
C1291
C-1152
C-1302
MoE23
74KAL-14
C-1682
H1G-23-16
SiO2 A1203 FezO3 FeO MgO CaO Na20 K20 TiO2 P205 MnO H20 § H20CO2
41.55 14.13 3.99 11.36 6.04 11.79 3.84 1.67 4.10 0.63 0.19 0.05 0.23 .
42.53 12.43 2.68 11.23 12.17 11.80 2.35 0.81 2.92 0.55 0.16 0.03 0.17
44.07 13.68 4.01 8.35 11.74 10.22 3.45 1.35 2.16 0.40 0.20 0.20 0.38 .
44.02 12.41 2.14 10.99 13.40 10.95 2.39 0.62 2.05 0.45 0.20 0.17 0.45
43.02 11.82 5.26 9.19 13.76 10.37 2.74 0.74 2.17 0.33 0.16 0.88 0.21
45.8 13.7 3.68 9.11 10.4 10.9 2.29 0.27 1.59 0.20 0.19 0.57 0.61 0.19
44.67 15.17 1.98 10.39 11.65 11.10 2.27 0.27 1.01 0.16 0.18 1.30 0.43 --
39.7 13.2 -13.45 9.20 12.16 2.61 1.13 3.73 0.59 0.15 ----
Total
99.57
100.24
100.65
99.50
100.58
96.0
.
. 99.83
Volcano Volcanic unit Rock type Sample no.
100.21
.
Koolau Honolulu NM
N
65MOLL-25 68BP-25
Kauai Koloa B
AOB
65PAL-15 65KAL-15
NM
N
B
AOB
C-1892
C-186 z
C-1882
C-1932
44.34 10.37 4.14 9.74 11.09 12.67 2.65 0.96 2.38 0.97 0.17 0.16 0.39 .
45.42 12.39 3.22 9.04 11.54 11.72 2.56 0.80 1.98 0.73 0.16 0.48 0.35
100.03
100.39
SiO2 A1203 Fe203 FeO MgO CaO NaEO K:O TiO2 P205 MnO H20 § H20CO2
35.52 11.01 9.82 7.13 11.46 12.17 4.84 1.72 2.78 1.05 0.25 0.58 0.63 0.26
40.15 12.01 5.80 7.20 13.27 12.73 4.13 1.06 1.90 0.98 0.23 0.14 0.14 0.03
42.27 12.11 4.25 9.17 12.96 10.57 3.62 0.87 2.42 0.69 0.21 0.33 0.31 0.01
45.14 13.37 3.85 9.00 11.46 10.39 2.94 0.65 1.97 0.35 0.20 0.19 0.15 0.05
39.91 9.13 5.29 8.16 15.88 11.97 3.16 1.53 2.76 0.81 0.18 0.54 0.63 .
Total
99.33
99.95
99.91
99.86
99.95
41.86 11.04 4.42 9.51 12.00 11.86 3.06 1.20 3.24 0.50 0.17 0.52 0.59 . . 99.97
Rock types: AOB, alkalic olivine basalt; B, basanite; N, nephelinite; NM, nepheline-melilitite. Sources: 1, Macdonald & Powers 1968; 2, Macdonald 1968; 3, Naughton et al. 1980; 4, Clague et al. 1982a; 5, Clague & Frey 1982; 6, Clague 1974. 1982; C h e n & F r e y 1983; R o d e n et al. 1984). N u m e r o u s studies h a v e s h o w n that the posterosional lavas h a v e the least r a d i o g e n i c isotopic ratios o f Sr a n d the m o s t r a d i o g e n i c isotopes o f N d , but h a v e h i g h a b u n d a n c e s o f i n c o m p a t i b l e trace e l e m e n t s ( L a n p h e r e & D a l r y m p l e 1980; Clague & F r e y 1982; C l a g u e et al. 1982a; C h e n & Frey 1983; Stille et al. 1983; U n r u h et al. 1983; F e i g e n s o n 1984; H e g n e r et al. 1984; R o d e n et al. 1984). T h i s is the s a m e inverse r e l a t i o n s h i p b e t w e e n isotopic ratios a n d their r e s p e c t i v e
p a r e n t - t o - d a u g h t e r a b u n d a n c e ratios n o t e d for post-caldera alkalic lavas a n d tholeiitic lavas. T h e s e r e l a t i o n s h i p s require c o m p l e x p e t r o g e n e t i c m i x i n g m o d e l s t h a t will be d i s c u s s e d in a later section.
Experimental constraintsmdepth of origin C l a g u e & Frey (1982) r e v i e w e d the e x p e r i m e n t a l data bearing on the generation of Hawaiian
239
Hawaiian alkaline volcanism
8
"~
Co
+
4
Alkalic field
Alkalic field
/f1/~
/
oa Volconics
//
/f/~'~J
C)
~
Honolulu V o l c a n i c s Hana V o l c a n i c s
E:Z] Lahaina V o l c a n i c s
2
Tholeiitic field E~Kalaupapa Basalt
0
~
54
58
I
42
i
I
46
i
I
50
i
I
54
i
I
58
i
62
/
\ ~ -~/"
/ ~//~/~f ', ~. J/ ~ ' ~ ' ~ ~ / / ] / ~ / ~ / Th:leiitic i
66
54
SiO 2 (wt.%)
,/"
58
I
42
L
EEB .--, '- J v
field
I
46
EEE Kauta Island
A Seamount63
i
50
Seamount 57 Ladd B a n k (Seamount 51) Seamount 2 0
I
54
i
I
i
58
I
62
i
66
Si02(wt.%)
(a)
(b)
FIG. 6. Alkali-silica diagrams for lavas of the post-erosional alkalic stage: (a) data for the Hana Volcanics on Haleakala from Macdonald & Powers (1968), the Lahaina Volcanics on W Maui from Macdonald (1968), the Kalaupapa Basalt on E Molokai from Clague et al. (1982a) and the Honolulu Volcanics on Koolau volcano, Oahu, from Clague & Frey (1982); (b) data for the Koloa Volcanics on Kauai from Macdonald (1968), Feigenson (1984), and D. A. Clague (unpublished data), the Kiekie Basalt on Niihau from Macdonald (1968) and D. A. Clague, G. B. Dalrymple, M. H. Beeson & E. D. Jackson (unpublished data), from Kaula Island from Garcia et al. (in press) and from seamounts 20, 57 and 63 and Ladd Bank from Clague (1974). Note that few of the posterosional lavas are differentiated. Lavas from seamount 63 display a differentiation trend similar to post-calderastage lavas; however, they contain amphibole and clinopyroxene compositionally similar to that in post-erosional stage lavas from the Hawaiian Islands. alkalic basalt to nepheline melilitite. Brey (1978) and Yoder (1979) give more detailed reviews and discussion of the experimental data. There is a general consensus that an increasing C 0 2 / H 2 0 ratio leads to increasingly silica-undersaturated
6
5
3 <
2
---
v
T
T
A
[ i
v
-~ o
= <_ ~
o 0 z~
<
o0
/ ~ ~ / o /~ o ~ .N.~ e,i
<' ~
g
4
-
~
< ~- < ,..' ~o ,,,
< "'
1.1.1~ T / -
T
/ ~
/o.,,~247
-
" /.~'~,,.~-/I T/R~.~ - / I
I I
Y ~
c~ / ~/
P II ',
//
:~]~ [/~ ~ "~ <0.4-,_ .~" I P? ? I ? I 100 200 aoo 400
iI
10
I P
-~ ~
zA
I
500
DISTANCE TO KILAUEA (km)
Fro. 7. Age of tholeiitic (T), post-calderaalkalic (A) and post-erosional (P) stages for the Hawaiian Islands that have both age data and post-erosional lavas plotted as a function of distance to Kilauea. The duration of the quiescent period decreases systematically from 2.5 Ma on Niihau to less than 0.4 Ma on Haleakala. (From Clague et al. (1982a) where the data sources are cited.)
I
600
melts, largely because of the expansion of the primary-phase field of orthopyroxene, an effect also caused by increased pressure. However, no consensus exists on the exact compositions produced from the peridotite-CO2-H20 system as the degree of melting increases (Eggler 1978; Wyllie 1978, 1979). The transition from hypersthene-normative to nepheline-normative melts is poorly constrained, but Wyllie (1979) suggests that the transition to larnite-normative liquids occurs near 2.5-3.5 GPa, depending on the CO2/ H 2 0 ratio. Since all the alkalic lavas have rareearth patterns suggesting equilibration with residual garnet, it is concluded that the m i n i m u m depth of equilibration is about 47 km whereas the more silica-undersaturated post-erosional lavas are probably equilibrated at depths of 75110 km (Clague & Frey 1982; Sen 1983). It is interesting to note that these depth estimates are near the base of the oceanic lithosphere which is estimatedtobe90kmthickforthe90-100Ma old crust beneath Hawaii (Detrick & Crough 1978). There is also a general consensus that the alkalic lavas all originated deeper than the tholeiitic lavas which produce seismic swarms and tremors as deep as 60-70 km (e.g. Klein 1982).
Isotopic constraints on mantle sources
Abundant isotopic data exist for post-caldera and post-erosional alkalic lavas from Hawaii. However, most of the older data are of low precision
24o
D. A. Clague i
i
T
i
[~1 MORB 0.51:52
~
MORB Transitional
Haleakala
E[~] West Maul [Z] East M o l o k a i Mauna K e a Waianae Hualalai
Alkalic
0"5151
~
Tholeiite
0"5150 -g z 0"5129
\ \\\
~\\
O. 5128 0.5127
\\
,\
the\ \
Bulk ear
] (a)
0-7050
\ \~,
\\ \
\\
\ \\\
\
0.5126 0.7020
\\ \\\
0-7040 ~ x ~ 0 . 7 ' 0 5 0 87Sr/86Sr
i
0"71030
0"7( )20
0"7040
0"7050
87Sr/86Sr (b)
eL% \ 0-5129
x
\
E]Z] MORB Hana Volcanics [Z~ Lahaina Volconics [~Z Koloa Volcanics HonoluluVolcanics m KalaupapaBasalt
\\\ \\\\\\\
0-5128 0"5127
\,,
0.5126 0"7020
0"7050
(c)
0.7040 87Sr/86Sr
Bulk
earth "~
xI
0.7050
FIG. 8. 143Nd/l~4Nd versus 87Sr/S6Sr diagrams for (a) the pre-shield stage, (b) the post-caldera alkalic stage and (c) the post-erosional alkalic stage. In a general way the post-erosional lavas have sources with the least radiogenic Sr, the pre-shield and post-caldera lavas have sources with more radiogenic Sr and Hawaiian tholeiites have sources with the most radiogenic Sr. The reverse is true for Nd isotopes. However, considerable overlap occurs, indicating that the source compositions are not homogeneous for all Hawaiian tholeiites, nor for all Hawaiian post-erosional alkalic-stage lavas, nor for all Hawaiian post-caldera-stage lavas. See text for discussion. Data sources are Staudigel et al. (1984) for Loihi Seamount, Chen & Frey (1983, 1985) for the Hana and Kula Volcanics, Stille et al. (1983) for the Waianae Volcanics, Stille et al. (1983) and Roden et al. (1984) for the Honolulu Volcanics, Feigenson (1984) for the Koloa Volcanics, Frey et al. (1984) for Mauna Kea, D. A. Clague, F. A. Frey & A. Kennedy (unpublished data) for E Molokai Volcanics and the Kalaupapa Basalt, and Tatsumoto et al. (1987) for the Lahaina Volcanics and Honolua Basalt of W Maui. and were usually obtained on samples for which no other g e o c h e m i c a l data have been published. More recent studies include integrated isotopic analyses of Sr, N d and occasionally Pb and Hf. Available data for STSr/S%r and 143Nd/144Nd are plotted in Fig. 8 for alkalic lavas from all three eruptive states. In general, the pre-shield and post-caldera alkalic lavas are isotopically similar, but the post-
erosional lavas have less radiogenic Sr and more radiogenic N d isotopic ratios. H a w a i i a n tholeiitic lavas have isotopic values that overlap those of the post-caldera lavas but generally have more radiogenic Sr and less radiogenic N d ratios and occupy the portion of the H a w a i i a n field towards bulk Earth values (Staudigel et al. 1984). Some H a w a i i a n volcanoes show a systematic variation in, for example, Sr isotope ratios as a function of
Hawaiian
2 41
alkaline volcanism
stratigraphic position from more radiogenic in further propose that the Kilauea end-member the tholeiites to progressively less radiogenic in represents the deep-seated mantle plume and that post-caldera and post-erosional lavas (Chen & the remaining two end-members may reside Frey 1983; Clague et al. 1983; Feigenson 1984; within the lithosphere. White & Hofmann (1982) Lanphere & Frey, in press). The arrays of Sr and offer the alternative view that one of the three Nd isotopic data can be explained by the mixing mixing components is recycled crust, whereas the of two components, one similar to bulk Earth and remaining sources are the same depleted and the other more similar to the depleted source of undepleted sources advocated by Chen & Frey mid-ocean ridge basalt. This relatively simple (1983,1985). model is complicated by the fact that the lavas with the least radiogenic Sr and the most radiogenic Nd isotopic ratios also have the most incompatible-element enrichment (Chen & Frey Rare-gas constraints on mantle 1983). These relations require more complex sources mixing of those two sources and small-percentage partial melts derived from at least one of the Isotopic analyses of rare gases, particularly argon sources. C h e n & Frey (1985) present an array of and helium, indicate complex mixing relations between three or more sources which Kaneoka models of this type and geochemical-balance calculations which suggest that small-percentage and Takaoka (1980) infer to be an undegassed partial melts of a depleted source similar to that plume source, a mid-ocean ridge degassed source of mid-ocean ridge basalt mix with enriched and an atmospheric component (Fig. 9). They mantle to produce the sources for tholeiitic and infer that the characteristics of these sources are as follows: the plume source has 3He/4He = 5 x alkalic lavas from Haleakala. 10-5 (36 times atmospheric) and 4~ Pb isotopic data (Tatsumoto 1978; Sun 1980; Stille et al. 1983; Staudigel et al. 1984) indicate 36Ar=400; the ridge source has 3He/aHe = 1 • 10 -5 (seven times atmospheric) and 4~ that even the relatively complex models proposed 36Ar= 16 000; and the atmospheric source has by Chen & Frey (1983, 1985) are still inadequate = 295.5. in that at least three isotopically distinct sources 3He/4He = 1.384 x 10-6 and 4~ Most data from the Hawaiian Islands are for are required to explain the Pb data. Tatsumoto (1978) demonstrated that the volcanoes of the tholeiitic basalt or xenoliths enclosed in alkalic Loa trend (Mauna Loa and Hualalai) have more basalt. However, some data, mostly 3He/4He radiogenic Pb than the volcanoes of the Kea ratios, exist for alkalic lavas from the pre-shield trend (Mauna Kea, Kilauea and Kohala). In alkalic stage (Kaneoka et al. 1983; Kurz et al. addition, the Pb isotopic ratios of tholeiitic and 1983; Rison & Craig 1983), the post-caldera alkalic lavas of the post-caldera alkalic stage are similar for the Kohala and Mauna Kea volcanoes. Haleakala e Hualalai dunite Tatsumoto suggested that this last observation o Kilauea tholeiite • Salt Lake tuff (xenoliths) indicated 'a close relationship between the tholv Hualalai tholeiite 9 Loihi lavas IOO 9 Loihi dunite o Loihi olivine phenocrysts eiitic and alkalic basalt (post-caldera) sources of 10-4 7o 5o each volcano' (Tatsumoto 1978, p. 71). Tatsu4o moto also proposed that one of the mixing end5o T~T~ ,. n . n ~ . ' ~ \ members of the source was the oceanic lithos20 < r'r phere. Lanphere et al. (1980) suggested that the ~o oceanic lithosphere was the source of the lavas of ~z I0 -5 the Hawaiian-Emperor volcanic chain and that the underlying plume provided mainly the heat required to melt the overlying lithosphere. Recent integrated isotopic studies of lavas from Loihi Seamount (Staudigel et al. 1984) and 10-6 I0 3 I0 4 10 5 10 2 Oahu (Stille et al. 1983) suggest that the three 4OAr/36Ar source end-members are (1) a highly depleted source, (2) a slightly depleted source and (3) a plot showing source intermediate between (1) and (2) which FIG. 9.3He/aHe v e r s u s 4~ Hawaiian data from Kaneoka & Takaoka (1978, may be the deeper source. Staudigel et al. (1984) 1980), Kyser & Rison (1982), All~gre et al. (1983) and propose that these isotopic end-members are Kaneoka et al. (1983). The mixing end-members most closely represented by the post-erosional shown are the atmosphere (A), a degassed mid-ocean Honolulu Volcanics, the shield-stage Koolau ridge source (R) and an undegassed plume (P). Mixing lavas and the shield-stage Kilauea lavas. They models shown are from Kaneoka et al. (1983). i
i
L
I
i
242
D. A.
alkalic stage (Kaneoka and Takaoka 1980; Kurz et al. 1983) and the post-erosional stage (Kyser & Rison 1982; Rison and Craig 1983). The alkalic lavas from Loihi Seamount have 3He/4Heatm = 21-24, whereas post-erosional lavas from the Hana Volcanics on Haleakala and the Honolulu Volcanics on Oahu have (3He/4He)atm=8-14. Post-caldera alkalic lavas from the Kula Volcanics on Haleakala have (3He/4He)atm=27-32, whereas a single post-caldera alkalic lava from Mauna Kea has (3He/4He)atm = 7.9. It is difficult to reconcile these variations in 3He/4He, and similarly large variations in 4~ with simple mixing models. Hawaiian lavas with the highest 3He/4He isotopic ratios, and therefore the lavas believed to be derived from the most primitive undegassed sources, are post-caldera alkalic lavas from Haleakala and tholeiitic lavas from Loihi Seamount. Similarly, Hawaiian samples with 3He/4He ratios most similar to that of mid-ocean ridge basalt, and therefore believed to be derived from the most degassed sources, include post-caldera alkalic lavas from Mauna Kea, post-erosional alkalic lavas from Haleakala and Oahu, dunite xenoliths from post-caldera alkalic lavas on Hualalai (Kaneoka & Takaoka 1980; Kyser & Rison 1982; Kurz et al. 1983; Rison & Craig 1983) and a variety of xenoliths from post-erosional alkalic lavas from Oahu (Kaneoka & Takaoka 1980; Kyser & Rison 1982). Interpretation of rare-gas data is complicated by a number of factors (see discussion in Kaneoka (1981)) but the most striking observation is that helium isotopic ratios equilibrate rapidly between melt and xenocrysts (or xenoliths) owing to high rates of diffusion at magmatic temperatures, whereas the heavier-rare-gas isotopic ratios commonly display disequilibrium between coexisting quenched melt (glass) and enclosed xenocrysts or xenoliths (Kaneoka et al. 1983). This observation suggests that 3He/4He ratios determined for melts probably represent the cumulative history of the melt including its source characteristics and the characteristics of the mantle and crust through which it migrates. The rare-gas isotopic ratios observed in the lavas must therefore reflect the ratios and abundances of gases of both the source and the intervening oceanic lithosphere as well as the percentage partial melting of the source and the ascent velocity of the melt. Models that explain the rare-gas characteristics of Hawaiian lavas will probably be similar to the complex partial-melting and mixing presented by Chen & Frey (1983, 1985) to explain trace-element and Sr- and Nd-isotope characteristics of lavas from Haleakala. These models will be discussed in the following section. The available data, particularly for the Loihi Seamount, suggest that those
Clague samples with the most radiogenic Sr and the least radiogenic Nd isotopic ratios also have the highest 3He/4He ratios indicating the least degassed source. These data suggest that the hypothesized plume source both is the least degassed and has the most radiogenic Sr and the least radiogenic Nd of the multiple sources involved in the generation of Hawaiian magmas.
T r a c e - e l e m e n t c o n s t r a i n t s on m a g m a sources Voluminous trace-element data exist for Hawaiian lavas and numerous additional traceelement and isotopic studies are currently under way. As with the isotopic data, much of the earlier data were obtained for poorly located samples or for samples that had not been analysed for major elements. Few of these early studies attempted to evaluate coherent suites of samples from single volcanoes and fewer still evaluated known stratigraphic sequences of samples from single volcanoes. In the past 5 years integrated studies of stratigraphically controlled samples have shown that much of the variation between volcanoes in Hawaii also occurs within the lavas from single volcanoes. Detailed trace-element studies have been published for post-caldera alkalic lavas from E Molokai (Clague & Beeson 1980; Clague et al. 1983), Hualalai (Clague et al. 1980), Haleakala (Chen & Frey 1983, 1985), Kohala (Feigenson et al. 1983; Lanphere & Frey 1987) and Mauna Kea (Frey et al. 1984). All Hawaiian alkalic lavas analysed so far have rare-earth patterns that are enriched in the light rare earths, generally to a much greater degree, than Hawaiian tholeiites. Figure 10 shows chondrite-normalized rare-earth patterns for alkalic lavas from the pre-shield, post-caldera and posterosional alkalic stages. The rare-earth data present the classic problem of the generation of strongly enriched rare-earth patterns in alkalic magmas from worldwide locations (Kay & Gast 1972). In order to generate these patterns from chondritic source compositions, extremely small degrees of partial melting are required. More recent models propose an enriched source, thereby allowing the magmas to be generated by higher percentages of partial melting (see Clague & Frey (1982) for a discussion). Other trace-element data show that the lavas from the different eruptive stages cannot be generated from similar sources. In particular, the post-erosional lavas are derived from sources with lower Zr/P205 and K/Ba ratios and variable KzO/P20 5 ratios (see Figs 11 and 12). The sources
Hawaiian
alkaline
for the post-erosional lavas are apparently more depleted than those for the post-caldera or preshield lavas. A detailed study of the pre-shield alkalic (and tholeiitic) lavas from Loihi Seamont (Frey & Clague 1983) indicates that these lavas are geochemically similar to the post-caldera alkalic lavas erupted along the Hawaiian-Emperor volcanic chain. However, most of the alkalic samples from Loihi are relatively unfractionated whereas those from the post-caldera stage have commonly evolved to hawaiite, mugearite and, more rarely, trachyte. Trace-element ratios, such as K/Ba and Zr/P, isotopic ratios and rare-earth abundance patterns are similar for the pre-shield and postcaldera alkalic lavas (Frey & Clague 1983). On the basis of K/Ba ratios, the mantle residuum for the E Molokai alkalic basalts retains phlogopite whereas that for transitional lavas does not (Clague & Beeson 1980). Clague et al. (1980) proposed that the source of Hualalai alkalic basalts is garnet bearing and relatively enriched in the light rare-earth elements. Feigenson et al. (1983) proposed that the source for alkalic basalt from Kohala has fiat to slightly-convex-upward chondrite-normalized rare-earth patterns and argued that a metasomatic pre-enrichment of the source was not required, Their model is significantly different from that of Chen & Frey (1983, 1985) which explains the trace-element and isotopic characteristics of the shield tholeiites, the post-caldera alkalic lavas and the posterosional alkalic lavas from Haleakala. Clague & Frey (1982) presented a detailed trace-element analysis of the post-erosional Honolulu Volcanics on Oahu and concluded that lavas ranging from nepheline-melilititeto alkalic basalt were generated by 2~-11X partial melting of a homogeneous garnet (less than 10%) lherzolite source that was carbon bearing. The source had recently been enriched and had a chondritenormalized La/Yb ratio of 4.4. During the melting process phlogopite, amphibole and a Ti-rich phase (oxide?) remain in the residuum, but apatite is completely melted. This model emphasizes the bulk composition of the source rocks and the physical conditions of the melting process. Other studies of post-erosional lavas are less comprehensive but include evaluation of the Kalaupapa Basalt on E Molokai (Clague et al. 1982a), the Koloa Volcanics on Kauai (Feigenson 1984) and the Hana Volcanics on Haleakala (Chen & Frey 1983, 1985). The Feigenson (1984) model is similar to that of Chen & Frey (1983), except that the primitive-mantle-mixing endmember is replaced by a combination of ancient subducted crust and primitive mantle.
volcanism
243
The Chen& Frey (1983, 1985) model proposes mixing of partial melts derived from an enriched mantle composition and very small percentage partial melts derived from a depleted mid-ocean ridge basalt source (Fig. 13). The model is specifically designed to explain the inverse correlations of radiogenic isotopic ratios and their respective parent-to-daughter abundance ratios. In other words, at Haleakala the lavas with the most radiogenic 878r/86Sr ratios (the tholeiites) have the lowest Rb/Sr ratios whereas the lavas with the least radiogenic 87Sr/86Sr ratios (the post-erosional alkalic lavas) have the highest Rb/ Sr ratios. This same pattern has now been documented for E Molokai (Clague et al. 1983) Kohala (Lanphere & Frey 1987), Hualalai (Clague 1982), Koolau (Roden et al. 1984) and Kauai (Feigenson 1984). The pattern does not appear to be universal for Hawaiian volcanoes as tholeiitic and post-caldera alkalic lavas from both Mauna Kea (Frey et al. 1984) and Waianae (Stille et al. 1983) have identical 87Sr/S6Sr and 143Nd/a44Nd isotopic ratios. Chen & Frey (1983) propose a geodynamic model in which partial melts of the deeper enriched-mantle source extract very small percentage partial melts of their wall-rocks, which are assumed to be a depleted mid-ocean ridge basalt source, during their ascent. An alternative model (Fig. 14) is that small-percentage partial melts of the enriched source are mixed with depleted mid-ocean ridge source (Clague et al. 1983) (this model was also discussed and rejected by C h e n & Frey (1985)). This alternative is the equivalent of a metasomatic model wherein the small-percentage partial melts of the enriched source are added to the depleted lower lithosphere which is then partially melted to create the various Hawaiian lavas. Roden et al. (1984) combine the Chen & Frey (1983) model to generate the source compositions of the Clague & Frey (1982) model for generation of the Honolulu Volcanics post-erosional lavas. The isotopic homogeneity of the Honolulu Volcanics lavas requires that the source of the nepheline-melilitite to alkalic basalt must have been created by mixing prior to the partial melting that generated the lavas. This is a rigid constraint since the Chen & Frey (1983) model forms each magma batch by partial melting of enriched mantle followed by mixing with smallpercentage melts derived from the wall-rocks. It seems improbable that the mixing of variablepercentage partial melts of the enriched mantle and variable-percentage partial melts of the wallrock invariably occur in proportions which will yield isotopically homogeneous lavas. The compositions of the Honolulu Volcanics lavas thus
D. A. Clague
244
,~176 l 20
8C 6O
!Mauna
40
~ rr
20
~o
2
20 Haleakala
I0
7 15
,a&
4~ Sm~u&'g
40 -/m'gLu
East Molokai!
IO 7
30
(a)
20
70 50
La Ce "
Cbl) IOC
IOC
70 50
50
La Ce (b2)
_~0
T~n Yb Lu
Nd
s" ~u & -~
Tm Yb Lu
z45
Hawaiian alkaline volcanism 5OO
i
i
i
i
K6k6 seamount
300 5OO
Phonol ire
300
Sanidine 300 200 Anorthoclase trachyte
200
20
lugearite ant ;enmorei te
120 i O0 -k'------....~ Hawa iite
"10
30
5 10 5 L
La Ce (b3)
I
Nd
L
Sm Eu Gd Tb
Ho
I
~L~
tm Yb Lu
I
La Ce (c)
1
Nd
I
I
Sm Eu Gd'(b
14o
TmYb Lu
FIG. 10. Chondrite-normalized rare-earth diagrams. (a) Lavas from Loihi Seamount (Frey & Clague 1983; D. A. Clague, unpublished data). The Loihi lavas show increasing relative light-rare-earth enrichment from the tholeiites through the transitional lavas to the alkalic lavas. (b) Lavas from the post-caldera alkalic stage from Mauna Kea (Frey & Roden 1987), Kohala (Feigenson et al. 1983; Lanphere & Frey 1987), Hualalai (Clague e t al. 1980; Basaltic Volcanism Study Project 1981), the K ula Formation of Haleakala (Chen & Frey 1983), Niihau (D. A. Clague, G. B. Dalrymple, M. H. Beeson & E. D. Jackson, unpublished data), Midway Island (F. A. Frey & D. A. Clague, unpublished data), Koko Seamount (D. A. Clague, unpublished data) and Ojin, Nintoku and Suiko Seamount (Bence et al. 1980). Rock types from Koko Seamount lavas are shown separately to emphasize the effects of crystal fractionation including the progressive development of a Eu anomaly due to sanidine fractionation and the curvature in the heavy-rare-earth elements that reflects apatite fractionation. In general the post-caldera lavas from along the chain have patterns of similar shape whose absolute enrichments are due to varying degrees of crystal fractionation. (c) kavas from the post-erosional alkalic stage including the Hana Volcanics (Chen & Frey 1983), the Kalaupapa Basalt (Clague et al. 1982a), the Honolulu Volcanics (Clague & Frey 1982), the Koloa Volcanics (Feigenson 1984; D. A. Clague, unpublished data) and the Kiekie Basalt (D. A. Clague, G. B. Dalrymple, M. H. Beeson & E. D. Jackson, unpublished data). The post-erosional lavas generally have greater relative light-rareearth enrichment and lower absolute heavy-rare-earth abundances than those of the post-caldera alkalic lavas.
246
D. A. Clague
500
individual volcanoes. These models can then be expanded to explain the differences between volcanoes in the Hawaiian Islands and along the Hawaiian-Emperor volcanic chain. It is also clear that the origin of alkalic and stronglyalkalic lavas in Hawaii cannot be determined independently from the origin of the more voluminous tholeiitic lavas.
East M o l o ~
400
:7
////
300 Loti
~ill
l 1
Volume relationsmmagma storage at depthmxenolith assemblages
2OO
100
. Hualal
04
0.8
1.2
Pe 9 ( wt%l
FIG. 11. Zr-P205 plot showing that the pre-shield Loihi lavas (horizontally shaded field) have similar Zr/ P2Os ratios to those of the post-caldera lavas (stippled fields) of the Hualalai, Niihau and E Molokai Volcanics, the Kula Volcanics of Haleakala and the Paniau Basalt of Niihau, whereas the post-erosional lavas (diagonally shaded fields) of the Koloa Volcanics, Kiekie Basalt, Honolulu Volcanics and the Kalaupapa Basalt have distinctly lower Zr/P205 ratios. The Kauai field shown is for the tholeiites of the shield stage and shows the similarity of the sources for the shield stage and post-caldera alkalic lavas. (From Frey & Clague (1983) modified after a figure in Clague et al. (1980.))
lend support to the metasomatic model proposed by Clague et al. (1983). These various interpretations illustrate the problems still to be resolved in unravelling the mantle source characteristics of basaltic magmas in general and Hawaiian basalts in particular. It seems possible that the three mixing end-members include two enriched sources, one the deep plume and the other a source more like that beneath Iceland, and a depleted mid-ocean ridge source. The depleted mid-ocean ridge and the enriched ridge sources probably reside in a heterogeneous lithosphere formed at the E Pacific Rise about 90 Ma ago. The main difference between the two enriched sources is the 90 Ma of radioactive decay that the ridge source has undergone since formation. Successful mixing models will have to integrate Sr, Nd and Pb isotopic data with major- and trace-element data for stratigraphically controlled samples from
The volumes of post-caldera and post-erosional lavas on individual volcanoes are inversely correlated. Thus, moderate to large volumes of post-caldera alkalic lavas occur on the Mauna Kea, Hualalai, Kohala, Haleakala, W Maui, E Molokai and Waianae volcanoes, whereas moderate to large volumes of post-erosional alkalic lavas occur on Haleakala, Koolau, Kauai and Niihau only. Thus Haleakala is the only volcano to have relatively large volumes of both postcaldera and post-erosional lavas. Elsewhere, if post-erosional lavas are abundant, post-caldera stage lavas are either absent or present in exceedingly small volumes. This relationship suggests that the source regions for these alkalic stages are related to one another. If the mantle source produces abundant post-caldera alkalic magmas, there is little remaining to contribute to post-erosional alkalic magmas. However, if little or no post-caldera alkalic magma is produced from the mantle source, abundant post-erosional alkalic magmas may be produced. In a general way the volumes of lavas produced during each stage reflect the degree of partial melting of the mantle sources. Thus Hawaiian volcanism commences with rather small-volume small-percentage partial melts that erupt infrequently. These are the early pre-shield alkalicstage lavas. This is followed by large-volume large-percentage partial melts that erupt frequently. These are the tholeiitic-shield-stage lavas. Post-caldera alkalic-stage lavas are once again small volumes of small-percentage partial melts that erupt relatively infrequently. The strongly-alkalic post-erosional lavas represent very small volumes of very small-percentage partial melts that erupt very infrequently. These relationships are shown diagrammatically in Fig. 15. The magma storage system within and beneath the growing volcano also evolves as a function of eruption rate so that during the earliest infrequent alkalic eruptions no magma storage system can be maintained. As the eruption rate increases, a
Hawaiian alkaline volcanism 1000
t
Honolulu~ K o l o a ~ / ~
700 F
~
2000
Kula
~
Ea~tokai
300 [KiekieN~.~~ - ~ ~ |Kaulapapa .,,~x\'~ ~Y.[
247 K()KO (A)
Iooo 800
Hualalai (A)
~-//L AYSAN (A)
400 300
i'/~-
MOLOKAI (A)
80 60
~
Hualalai(T)
0.1
'
HUALALAI (A)
'--SUIKO(.6,)
20Oi00
O lO IS
6 J IN (A)
~ - - ~
NAPALI (T)
4-0
I O2
Ol
I 1 I 01610181I I 04 I(3. 20 K (wf.%)
I I I I 40 60
(b)
{a)
013 '01.5 '017' 'II.0 K [wt%)
'
310 '510
FIG. 12. (a) K-Ba plot showing that the pre-shield Loihi lavas (horizontally shaded fields) have similar K/Ba ratios to those of the lavas of the post-caldera alkalic stage (stippled fields) and the tholeiitic shield stage of Hualalai, whereas the post-erosional lavas (diagonally ruled fields) have distinctly lower K/Ba ratios indicating derivation from more depleted sources. (From Frey & Clague (1983) modified after a figure in Clague et al. (1980).) (b) K-Ba plot showing fields for tholeiitic lavas from Kauai (Napali Member of the Waimea Canyon Basalt (D. A. Clague, unpublished data)) and the Suiko and Ojin Seamounts (Kirkpatrick et al. 1980). Also shown are the fields for post-caldera alkalic-stage lavas from Suiko, Nintoku and Ojin Seamounts (Kirkpatrick et al. 1980), Laysan Island (Dalrymple et al. 198 !) and Koko Seamount (Clague 1974). These data demonstrate that the K/Ba ratios of post-caldera alkalic-stage lavas and their source rocks have been constant during the last 64 Ma of activity along the Hawaiian-Emperor volcanic chain.
0"58 0'54 0"50 L~ 0'46
-
~bs 0"O5 ['~,-.: ~O I i'x~ %%
~1 ~
Upper E. Molokai . . . . . I 0"1L ! ' ~ ' ;4 Lines of constont ~ I~'~" Ir~'x~'bc ~'~.% mixing pr~176176 _~ I i '!"~x.i ~~x %' o "-"'c-["~lHnna, ~ oL , I ', ""4.~"h ~,..Honolulu I~.b< ~ rn . . . . . .<.% ~ 0"3~'--.~ ~ ~'Loih'
J
o-4
0'34
, o,
0"7020
II
% cb~
o'
l,
3 i
Mixing lines
L nes of constant mixing proportions
0.58
Mixing lines
O.52r
! '
/
o.2
~/ /i ~U/// / ~
~
,/~ i
I // //~
/'~ .,-"/
f.~ "~--~
I
'*6
0"3
I
~ "5 "~ :.,=
I
' Honomonu o=onu
o ce . . . .
i ....
0.70.30 0.7040 87Sr/86Sr
J , I
0-7050
FIG. 13. Plot of La/Ce versus 87Sr/S6Sr for various lavas from the Hawaiian Islands. The mixing lines are from Chen & Frey (1983) and show models in which small-percentage partial melts of a mid-ocean ridge source are mixed with a primitive bulk-Earth composition source. The broken lines are lines of constant mixing proportions. In general, Hawaiian lavas can be generated from sources composed of 0.5%-3% of 0.1%- 1% partial melts of a depleted source mixed with 99.5%-97% bulk-Earth source. See text for discussion.
I
M
0.702
i
i
0"705 0.704 8ZSr/86Sr
i
0.70{
FIG. 14. Plot of La/Ce versus 87Sr/86Sr for various lavas from the Hawaiian Islands. This model (Clague et al. 1983) has the opposite mixing relations to those shown in Fig. 13. In this model small-percentage partial melts of the bulk-Earth source are mixed with the depleted mid-ocean ridge source. This model has a positive correlation between the percentage partial melting of the bulk-Earth source and the amount of melt mixed with the depleted source, whereas the model shown in Fig. 13 has an inverse correlation between these parameters.
248 108
D. A. Clague i0 0
Volcanic stages ~_lf-2~-3"LErosion I 4
Stage I
Rock .ty ~__ss
gI
Tholeiitic pictrite
I Pre-shield alkolic
iO-~j 107
3o
5 Post-caldera
alkalic 4 Post-erosional alkalic
20~
g
io-21
o~ Transitional basalt
i,h
106 _
io-31 105 104
I0
10-41
io-SI
o
~;
3 4 Time (Ma)
'
5
0
Alkalir basalt Basanite Nephelinite
_ Spinel 30 Iherzolite X ---X E 40 --
~- 5o 123
FIG. 15. Plot showing the eruption rates, eruption frequency and percentage partial melting during the evolutionary stages of a hypothetical Hawaiian volcano. During stage 1, the pre-shield alkalic stage, eruption rates are low and eruptions are infrequent, perhaps 1 in 100 years. The lavas erupted are smallpercentage melts (less than 15%) of relatively depleted sources and consist of basanite to transitional basalt. This stage is shown as lasting for 250 000 years but may be considerably longer. The earliest lavas are unsampled but may be nephelinite, melilitite or perhaps even more alkaline compositions. Stage 2, the tholeiitic shield stage, is characterized by high eruption rates and frequent eruptions of largepercentage partial melts (15% 40%) generated from relatively enriched sources. The eruption frequency and rate shown are based on estimates for Kilauea (Dzurisin et al. 1984). The maximum percentage of partial melting is taken from Wright & Helz (1987). Stage 3, the post-caldera alkalic stage, is characterized by eruptions every 100 to 1000 years at moderate to low eruption rates. The lavas are generated by 10%15% partial melting of relatively depleted sources and consist of transitional to alkalic basalt and related differentiates. Stage 4 follows a period of erosion and volcanic quiescence, this erosional period is between 0.4 and 2.5 Ma long. The post-erosional alkalic stage is also of variable duration but can be as long as 2.0 Ma as shown here. Eruptions are very infrequent, perhaps 1 every 10 000 to 100 000 years, and have very low eruptive rates. The lavas are generated by less than 10~ partial melting of sources nearly as depleted as the source of mid-ocean ridge basalt. The erupted lavas include alkalic basalt, basanite, nephelinite and melilitite.
shallow m a g m a storage system develops w h i c h is c h a r a c t e r i z e d by an overlying caldera. Loihi S e a m o u n t has already evolved to this stage. D u r i n g the period with the greatest e r u p t i o n rates, not only is m a g m a continuously stored in this shallow m a g m a reservoir but a n o t h e r d e e p e r staging area a p p a r e n t l y develops. D u r i n g the
70
S )inel Iherzolite
S )inel Iherzolite
Spinel Iherzolite
I -
i~
Garnet Iherzoite
Melilitite
60
Stage 4
X . . . .
20
"~ Tholeiite
o "..IT_ P, 3
~
Stage 3
O~n~crust
2 Tholeiitic shield "C
Stage 2
5 SL
40
I
Garnet Iherzoite
Garnet Iherzoite
-
_
Garnet Iherzoite I
\Magma /I \,~accumulationl
j
I
J
/,I
2 \\
Magma 80 - -// \\ accumulation
i Magma I / i Xaccumulation! ' ' /~1 ~ X--X
90 I00"
Magma (/i~ accumulation/.//!~
FIG. 16. Cross-sections showing the structure beneath a typical Hawaiian volcano during the four eruptive stages. The stability fields for spinel lherzolite and garnet lherzolite are from Sen (1983). The lines denoted by x - x show the maximum depth from which xenoliths are brought to the surface. Zones of magma accumulation are shown. The ocean crust is shown flexing beneath the accumulating load of the volcano and thickening due to dyke injection. Stage 1 lavas erupt onto ocean crust and come from perhaps 80 km deep and carry rare xenoliths of spinel lherzolite to the surface. No shallow storage zones exist. Stage 2 lavas come from 65-70 km deep (the depth of the deepest earthquakes (Klein 1982)) but pass through two magma storage zones, a deep zone (D) at the base of the thickened down-warped crust and a shallow zone (S) within the volcano and underlying a summit caldera. The magma storage zones filter out all xenoliths from deeper levels so that the erupted lavas contain only xenoliths derived from above 3 + km. Stage 3 lavas come from 75-85 km but pass through a deep magma storage zone at the base of the crust. The shallow storage chamber apparently solidifies as the eruption rate declines at the end of stage 2 or the beginning of stage 3 (see Fig. 14). Xenoliths from above the deep storage zone are carried to the surface and include dunite, wehrlite and a variety of gabbroic rocks. Stage 4 lavas come from depths of around 100 km and move directly to the surface. Xenoliths from within the garnet stability field are occasionally brought to the surface; these probably originate at depths of about 80 km. The lavas from stages 2 and 3, and the late parts of stage 1, erupt from distinct rift zones whereas those from stage 4, and perhaps the early part of stage 1, erupt randomly.
Hawaiian alkal&e volcanism
249
TABLE 6. Distribution of Hawaiian xenofiths* Eruptive stage
Lava types
Dykes, sills, veins
Cumulates
Metamorphic rocks
Post-erosional alkalic
Alkalic basalt Basanite Nephelinite Nepheline-melilitite Alkalic basalt
< 1%
1%, olivine cumulates dominant
99%, dunite ,~ wehrlite,~ lherzolite > harzburgite, garnet peridotite locally
Post-caldera alkalic
Ankaramite
Hawaiite Mugearite Trachyte
3%, veins dominant 35%, olivine 62%, dunite >>wehrlite cumulates > pyroxene cumulates 14%, dykes and sills 53%, pyroxene 33%, dunite>>wehrlite dominant cumulates > olivine cumulates 14%, dykes and sills 57%, pyroxene 29~, dunite >>wehrlite dominant cumulates > olivine cumulates
Shield-building
Tholeiite
75%, dykes and sills 25%, olivine dominant cumulates dominant
None
Pre-shield alkalic (Loihi)
Alkalic basalt Basanite
None
99%, dunite >>lherzolite
1%, olivine cumulates
* Data for all but the pre-shield alkalic stage from Jackson (1968). The percentages indicate the relative percentages of dykes, sills and veins, cumulates and metamorphic rocks. The symbols ~, >, >> indicate the relative abundance of each lithology. post-caldera alkalic stage, as eruption rates wane, the shallow magma reservoir can no longer be maintained and it crystallizes. Magmas continue to slow their ascent and to differentiate within the deeper staging area, probably located near the base of the oceanic crust at a depth of about 20 km. By the time of the post-erosional alkalic stage, the magmas ascend rapidly from the mantle without slowing or stopping in any magma storage zones. The development of magma storage or staging areas is schematically shown in Fig. 16. The main evidence for the preceding model is the assemblages of xenoliths that occur in lavas of the different eruptive stages (Table 6). Although it would be useful if these xenolith populations reflected the mantle through which the lavas ascend, it seems more likely that they reflect the development of magma storage reservoirs which, once formed, act as hydraulic filters that remove xenoliths carried up from greater depths (Clague & Dalrymple, 1987). Thus the presence of lherzolite in lavas of only the preshield and post-erosional alkalic stages reflects the lack of shallow reservoirs during these times of low eruption rates. The tholeiitic lavas apparently pass through a double filter and generally contain only fragments of veins, dykes and shallow olivine cumulates. The post-caldera alkalic-stage lavas erupt after the shallow reservoir
has crystallized, but the deeper staging area still filters out any spinel lherzolite or garnet-bearing deep-origin xenoliths. The lavas, however, do pick up a shallower xenolith assemblage ofdunite, wehrlite and gabbro that includes rocks that are part of the oceanic crust and cumulates from earlier eruptive stages of the volcano.
Conclusions Hawaiian volcanoes are constructed of lavas erupted during four sequential stages. Stage 1 is characterized by submarine eruption of alkalic basalt and basanite followed by eruption of tholeiitic lavas. Stage 2 is characterized by eruption of enormous volumes of tholeiitic basalt which form the shield volcanoes. Stage 3 follows collapse of the summit caldera and consists of a cap of alkalic basalt and related differentiates which commonly erupt from vents aligned along the rift zones developed during stage 2. Stage 4 follows a period of volcanic quiescence of as much as 2.5 Ma and is characterized by infrequent small-volume eruptions of nephelinemelilitite, nephelinite, basanite and alkalic basalt from vents unassociated with pre-existing structures. Early stage 1 and stage 4 lavas are relatively primitive and contain deep-origin xenoliths of
z5o
D. A. Clague
lherzolite. Stage 2 and 3 lavas are generally m o d e r a t e l y differentiated indicating substantial residence time in sub-caldera and perhaps d e e p e r m a g m a storage sites. E r u p t i o n rates and xenolith populations in the lavas of the four stages enable a model to be constructed in w h i c h stage 1 lavas are small volumes of small-percentage partial melts that erupt infrequently; no m a g m a storage system is m a i n t a i n e d and deep-seated xenoliths are brought to the surface. Stage 2 lavas are large volumes of large-percentage partial melts that erupt frequently; two m a g m a storage zones are m a i n t a i n e d b e n e a t h the caldera and b e n e a t h the base of the oceanic lithosphere. These storage zones filter out all deep-origin and crustal xenoliths. Stage 3 lavas are again small volumes of small-percentage partial melts that erupt infrequently; the shallow storage zone is no longer m a i n t a i n e d but the deeper zone still filters out deep-origin xenoliths. Crustal xenoliths are c o r n -
m o n in these lavas. Stage 4 lavas are very small volumes of very-small-percentage partial melts that erupt infrequently following a long volcanic quiescence. As for early stage 1 lavas, no m a g m a storage zones are m a i n t a i n e d and deep-origin xenoliths are brought to the surface. The m a g m a sources are not h o m o g e n e o u s through these eruptive stages. In general the early stage 1 and stage 4 lavas are derived from the most depleted sources and the stage 2 lavas from the most enriched source. Rare-gas contents of H a w a i i a n lavas indicate that at least one of the source compositions involved in g e n e r a t i o n of the lavas is a primitive, p r e s u m a b l y deep, undegassed c o m p o n e n t . Various mixing models can be constructed to explain the source compositions, but the characteristics of the e n d - m e m bers are not well constrained. At least three c o m p o n e n t s are required by radiogenic isotopic data.
References ALLEGRE, C. J., STAUDACHER,T., SARDA, P. & KURZ, M. 1983. Constraints on evolution of Earth's mantle from rare gas systematics. Nature, Lond. 303, 762-6. BARGER, K. E. & JACKSON, E. n. 1974. Calculated volumes of individual shield volcanoes along the Hawaiian-Emperor chain. U.S. Geol. Surv. J. Res. 2, 545-50. BASALTIC VOLCANISM STUDY GROUP. 1981. Basaltic Volcanism on the Terrestrial Planets, Pergamon Press, Oxford. BEESON, M. H. 1976. Petrology, mineralogy and geochemistry of the East Molokai Volcanic Series, Hawaii. U.S. Geol. Surv, Prof. Pap. 961, 53 pp. BENCE, A. E., TAYLOR, S. R. & FISK, M. 1980. Majorand trace-element geochemistry of basalts from Ojin, Nintoku, and Suiko Seamounts of the Emperor Seamount Chain: DSDP-180D Leg 55. In JACKSON, E. D., KOIZUMI,I., et al. (eds.) Initial Reports of the Deep Sea Drilling Project, 55, pp. 599-606. U.S. Government Printing Office, Washington DC. BREY, G. 1978. Origin of olivine melilitites--chemical and experimental constraints. J. Volcanol. geotherm. Res. 3, 61-88. BUDAHN, J. R. & SCHMITT, R. A. 1984. Petrogenic modelling of Hawaiian tholeiitic basalts. A geochemical approach. Geochim. cosmochim. Aeta, 49, 67 87. CHASE,T. E., MENARD,H. W. & MAMMERICKX,J. 1970. Bathymetry of the north Pacific. Scripps Institution of Oceanography Charts 2, 7, 8, Scripps Institution of Oceanography, La Jolla, CA. CHEN, C.-Y. & FREY, F. A. 1983. Origin of Hawaiian tholeiites and alkalic basalt. Nature, Lond. 302, 785-9. - - & - - 1985. Trace element and isotope geochemistry oflavas from Haleakala Volcano, East Maui:
implications for the origin of Hawaiian basalts. J. geophys. Res., 90, 8743-68. CLAGUE, D. A. 1974. The Hawaiian-Emperor seamount chain: its origin, petrology, and implications for plate tectonics. PhD Thesis, University of Calfornia, San Diego, CA (unpublished). - 1982. Petrology of tholeiitic basalt from Hualalai Volcano, Hawaii. Eos, 63, 1138. - & BEESON, M. H. 1980. Trace element geochemistry of the East Molokai Volcanic Series, Hawaii. Am. J. Sci. 280A, 820-44. -& DALRYMPLE, G. B. 1987. The geology of the Hawaiian-Emperor volcanic chain. U.S. Geol. Surv. Prof. Pap. 1350, pp. 5-55. - & FREY, F. A. 1982. Petrology and trace element geochemistry of the Honolulu Volcanics, Oahu. Implications of the oceanic mantle beneath Hawaii. J. Petrol. 23, 447-504. -& GREENSLATI~,J. 1972. Alkalic volcanic suite from the Emperor Seamounts. Geol. Soc. Am. Abstr. 4, 136. - - , DAO=GONG, C., MURNANE, R., BEESON, M. H., LANPHERE, M . A . , D A L R Y M P L E , G . B . , FRIESEN, W. & HOLCOMB, R. T. 1982a. Age and petrology of the Kalaupapa basalt, Molokai, Hawaii. Pacific Sci. 36, 411-20. - - , FREY, F. A. & BEESON, M. H. 1983. Rare-earth element and Sr isotopic evidence for the origin of the East Molokai volcanics, Hawaii. Eos, 64, 902. - - , JACKSON,E. D. & WRIGHT, T. L. 1980. Petrology of Hualalai volcano Hawaii. Implication for mantle composition. Bull. Volcanol. 434, 641-56. - - , MOORE, J. G. & NORMARK, W. R. 1982b. Loihi Seamount: the youngest Hawaiian volcano. Generation of Major Basalt Types, 1AVCEI-IAGC Scientific Assembly, Reykjavik, Iceland, p. 42. CRAIG, H. 1983. Introduction. Earth planet. Sci. Left. 66, 334-6.
Hawaiian alkaline volcanism DALRYMPLE, G. B., CLAGUE, D. A., GARCIA, M. O. & BRIGHT, S. W. 1981. Petrology and K - A r ages of dredged samples from Laysan Island and Northampton Bank volcanoes, Hawaiian Ridge, and evolution of the Hawaiian-Emperor chain. Part II. Geol. soc. Am. Bull. 92, 884-933 (Summary in Part I, 92, 315-8). Midwa& Y LANPHERE, M. A. 1977. Revised age for volcano, Hawaiian volcanic chain. Earth planet. Sci. Lett. 37, 107-16. DETRICK, R. A. & CROUGH, S. T. 1978. Island subsidence, hot spots, and lithospheric thinning. J. geophy. Res. 83, 1236-44. DZURISIN, D., KOYANAGI,R. Y. & ENGLISH, T. T. 1984. Magma supply and storage at Kilauea volcano, Hawaii, 1956-1983: J. Volcanol. geotherm. Res. 21, 177-206. EGGLER, D. H. 1978. The effect of CO2 upon partial melting of peridotite in the system N a 2 0 - C a O AI203-MgO-SiO2-CO2 to 35 kb, with an analysis of melting in a peridotite-H20-CO2 system. Am. J. Sci. 278, 305-43. FEIGENSON, M. D. 1984. Geochemistry of Kauai Volcanics and a mixing model for the origin of Hawaiian alkalic basalts. Contrib. Mineral. Petrol. 87, 109-19. , HOFMANN, A. W. & SPERA, F. J. 1983. Case studies on the origin of basalt II. The transition from tholeiitic to alkalic volcanism on Kohala Volcano, Hawaii. Contrib. Mineral. Petrol. 84, 390-405. FREY, F. A. 1980. The origin of pyroxenites and garnet pyroxenites from Salt Lake Crater, Oahu, Hawaii: trace element evidence. Am. J. Sci. 280A, 427-49. - 1984. Rare earth element abundances in upper mantle rocks. In : HENDERSON, P. (ed.) Rare Earth Element Geochemistry, pp. 153-203. Elsevier, Amsterdam. -& CLAGUE, D. A. 1983. Geochemistry of diverse basalt types from Loihi Seamount, Hawaii: petrogenetic implications. Earth planet. Sci. Lett. 66, 337-55. -& RODEN, M. F. 1987. The mantle source for the Hawaiian Islands: constraints from the lavas and ultramafic inclusion. In ." MENZIES, M. & HAWKESWORTH, C. (eds) Mantle Metasomatism, pp. 42363. Academic Press, New York. , K E N N E D Y , G . , GARCIA, M . O., WEST, H. B., WISE, W. S. & KWON, S.-T. 1984. Mauna Kea Volcano, Hawaii: Lava compositions and their evolution. Eos, 65, 300. GARCIA, M. O., GROOMS, D. G. & NAUGHTON, J. J. Petrology and geochronology of volcanic rocks from seamounts along and near the Hawaiian Ridge. Contrib. Mineral. Petrol., in press. HEGNER, E., UNRUH, D. & TATSUMOTO,M. 1984. Sr, Nd, and Pb isotopic systematics in the West Maui Volcanic Suite, Hawaii. Eos, 65, 1138. HORIBE, Y., KIM, K. & CRAIG, H. 1983. Off-ridge submarine hydrothermal vents: back-arc spreading centers and hotspot seamounts. Eos, 64, 724. JACKSON, E. D. 1968. The character of the lower crust and upper mantle beneath the Hawaiian Islands. Proc. 23rd lnt. Geology Congr., Prague, Vol. 1, pp. 135-50.
2 51
d~r WRIGHT, T. L. 1970. Xenoliths in the Honolulu Volcanic Series, Hawaii. J. Petrol. 11,405-30. --, SILVER, E. A. & DALRYMPLE, G. B. 1972. Hawaiian-Emperor chain and its relation to Cenozoic circumpacific tectonics. Geol. Soc. Am. Bull. 83, 601-18. KANEOKA, I. 1981. Noble gas constraints on the layered structure of the mantle. Rock Magn. Paleogeophys. 8, 94-9. -• TAKAOKA,N. 1978. Excess 129Xe and high 4He/ 3He ratios in olivine phenocrysts of Kapuho lava and xenolithic dunites from Hawaii. Earth planet. Sci. Lett. 39, 382-6. - & -1980. Rare gas isotopes in Hawaiian ultramafic nodules and volcanic rocks: Constraints on genetic relationships. Science, 208, 1366-68. , & CLAGUE, D. A. 1983. Noble gas systematics for coexisting glass and olivine crystals in basalts and dunite xenoliths from Loihi Seamount. Earth planet. Sci. Lett. 66, 427-37. KAY, R. W. & GAST, P. 1972. The rare earth content and origin of alkali-rich basalts. J. Geol. 81, 65382. KIM, K. & CRAIG, H. 1983. Methane: a "real-time" tracer for submarine hydrothermal systems. Los, 64, 724. KIRBY, S. H. & GREEN, H. W. 1980. Dunite xenoliths from Hualalai volcano: evidence for mantle diapiric flow beneath the island of Hawaii. Am. J. Sci. 280A, 550-75. KIRKPATRICK, R. J., CLAGUE, D. A. & FRIESEN, W. 1980. Petrology and geochemistry of volcanic rocks, DSDP Leg 55, Emperor seamount chain. In . JACKSON, E. D., KOIZUMI, I., et al. (eds) Initial Reports of the Deep Sea Drilling Project 55, pp. 50957. U.S. Government Printing Office, Washington, DC. KLEIN, F. W. 1982. Earthquakes at Loihi Submarine volcano and the Hawaiian hot spot. J. geophys. Res. 87, 7719-26. KURZ, M. D., JENKINS, W. J., HART, S. R. & CLAGUE, D. 1983. Helium isotopic variations in volcanic rocks from Loihi Seamount and the island of Hawaii. Earth planet. Sci. Lett. 66, 388-406. KYSER, T. K. & RISON, W. 1982. Systematics of rare gas isotopes in basic lavas and ultramafic xenoliths. J. geophys. Res. 87, 5611-30. LANPHERE, M. A. 1983. 875r/S6Sr ratios for basalt from Loihi Seamount, Hawaii. Earth planet. Sci. Lett. 66, 380-7. - & DALRYMPLE, G. B. 1980. Age and strontium isotopic compositions of the Honolulu Volcanic Series, Oahu, Hawaii. Am. J. Sci. 280A, 736-51. & FREY, F. A. 1987. Geochemical evolution of Kohala Volcano, Hawaii. Contrib. Mineral. Petrol., 95, 100-13. , DALRYMPLE,G. B. & CLAGUE, D. A. 1980. R b Sr systematics of basalts from the HawaiianEmperor volcanic chain. In: JACKSON, E. D., KOIZUMI, I. et al. (eds) lnitial Reports o f the Deep Sea Drilling Project 55, pp. 695-706. U.S. Government Printing Office, Washington, DC. LEEMAN, W . P., MURALI, A . V . , MA, M.-S. & SCHMITT, R. A. 1977. Mineral constitution of mantle source
2 52
D.A.
regions for Hawaiian basalts-rare earth element evidence for mantle heterogeneity. In: DICK, H. J. B. (ed.) Magma Genesis, Bull. Oregon Dep. Geol. Miner. Ind. 96, pp. 169-83. MACDONALD, G. A. 1968. Composition and origin of Hawaiian lavas. In: COATS, R. E., HAY, R. L. & ANDERSON, C. A. (eds) Studies in Volcanology, Geol. Soc. Am. Mere. 116, pp. 477-522. -& ABBOTT, A. T. 1970. Volcanoes in the Sea, University of Hawaii Press, Honolulu, HI. & KATSURA,T. 1962. Relationship of petrographic suites in Hawaii. In : MACDONALD,G. A. & KUNO, H. (eds) The Crust of the Pacific Basin, Am. Geophys. Union Monogr. 6, pp. 187-95. & -1964. Chemical composition of Hawaiian lavas. J. Petrol. 5, 82-133. & POWERS, H. A. 1968. A further contribution to the petrology of Haleakala volcano, Hawaii. Bull. Geol. Soc. Am. 79, 877-88. MALAHOFF, A., MCMURTY, G. M., WILTSHIRE,J. C. & YEn, H.-W. 1982. Geology and geochemistry of hydrothermal deposits from active submarine volcano Loihi, Hawaii. Nature, Lond. 289, 234-9. MOORE, J. G. & CLAGUE, D. A. 1981. Loihi seamount lavas: volatile contents. Eos, 62, 1083. -& CLAGUE, O. A. 1982. Volatile contents of Loihi lavas, Hawaii. Generation of Major Basalt Types, IA VCEI-IAGC Scientific Assembly, Reykjavik, Iceland, p. 100. & NORMARK, W. R. 1982. Diverse basalt types from Loihi seamount, Hawaii. Geology 10, 88-92. --, NORMARK, W. R. & LIPMAN, P. W. 1979. Loihi seamount--a young submarine Hawaiian volcano. Hawaii Symp. on Intraplate Volcanism and Submarine Volcanism, Abstracts, p. 127. Hilo. Hawaii. NAUGHTON, J. J., MACDONALD, G. m. & GREENBERG, V. A. 1980. Some additional potassium-argon ages of Hawaiian rocks: the Maui Volcanic Complex of Molokai, Maui, Lanai and Kahoolawe. J. Volcanol. geotherm. Res. 7, 339-55. RISON, W. & CRAIG, H. 1983. Helium isotopes and mantle volatiles in Loihi Seamount and Hawaiian island basalts and xenoliths. Earth planet. Sci. Lett. 66, 407-26. RODEN, M. F., FREY, F. A. & CLAGUE, O. A. 1984. Geochemistry of tholeiitic and alkalic lavas from the Koolau Range, Oahu, Hawaii: implications for Hawaiian volcanism. Earth planet. Sci. Lett. 69, 141-58. ROEDDER, E. 1965. Liquid CO2 inclusions in olivinebearing nodules and phenocrysts for basalts. Am. Mineral. 50, 1746-82. 1983. Geobarometry of ultramafic xenoliths from Loihi seamount, Hawaii on the basis of CO2 inclusions in olivine. Earth planet. Sci. Lett. 66, 369-79. SEN, G. 1983. A petrologic model for the constitution of the upper mantle and crust of the Koolau Shield, -
-
-
-
-
-
-
-
-
-
Clague Oahu, Hawaii, and Hawaiian magmatism. Earth planet. Sci. Lett. 62, 215-28. SINTON, J. M. 1981. Silicic lavas of the Waianae volcano, Oahu, Hawaii. Eos, 62, 1068. STAUDIGEL, H., ZINDLER, A., HART, S. R., LESLIE, T., CHEN, C. Y. & CLAGUE, D. 1984. The isotopic systematics of a juvenile intraplate volcano: Pb, Nd, and Sr isotope ratios of basalts from Loihi seamount, Hawaii. Earth planet. Sci. Lett. 69, 1329. STEARNS,H. T. 1940. Four-phase volcanism in Hawaii. Geol. Soc. Am. Bull. 51, 1947-1948. STILLE, P., UNRUH, D. M. 86 TATSUMOTO,M. 1983. Pb, St, Nd and Hfisotopic evidence of multiple sources for Oahu, Hawaii basalts. Nature, Lond. 304, 259. SUN, S.-S. 1980. Lead isotopic study of young volcanic rocks from mid-ocean ridges, ocean islands, and island arcs. Phil. Trans. R. Soc. Lond., Ser. A, 297, 409-45. TATSUMOTO, M. 1978. Isotopic composition of lead in oceanic basalt and its implication to mantle evolution. Earth planet. Sci. Lett. 38, 63-87. , HEGNER, E. & UNRUH, D. M. 1987. Origin of the West Maui Volcanics inferred from Pb, Sr and Nd isotopes and a multi-component model for oceanic basalts. U.S. Geol. Surv. Prof Pap. 1350, pp. 72344. UNRUH, D. M., STILLE, P. & TATSUMOTO, M. 1983. Isotopic heterogeneity among Hawaiian-Emperor basalts. Eos, 64, 347-8. WEST, H. B. & GARCIA, M. O. 1982. Composition and evolution of the alkalic cap stage lavas from Mauna Kea Volcano, Hawaii. Eos, 63, 1138. WHITE, W. M. & HOFMANN, A. W. 1982. Sr and Nd isotope geochemistry of oceanic basalts and mantle evolution. Nature, Lond. 296, 821-5. WINCHELL, H. 1947. Honolulu Series, Oahu, Hawaii. Bull. Geol. Soc. Am. 58, 1-48. WISE, W. S. 1982. A volume-time framework for the evolution of Mauna Kea volcano, Hawaii. Eos, 63, 1137. WRIGHT, T. L. 1971. Chemistry of Kilauea and Mauna Loa lava in space and time. U.S. Geol. Surv. Prof. Pap. 735, 40 pp. & HELZ, R. T. 1987. Recent advances in Hawaiian petrology and geochemistry. U.S. Geol. Surv. Prof. Pap. 1350, pp. 625-41. WYLLIE, P. J. 1978. Mantle fluid composition buffered in peridotite-CO2-H20 by carbonates, amphibole, and phlogopite. J. Geol. 86, 687-713. 1979. Magmas and volatile components. Am. Mineral. 64, 469-500. YODER, H. S. 1979. Melilite-bearing rocks and related lamprophyres. In: YODER, H. S., Jr., (ed) The Evolution of the Igneous Rocks--Fiftieth Anniversary Perspectives, pp. 391-411. Princeton University Press, Princeton, NJ.
DAVID A. CLAGUE, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025, U.S.A.
Geochemistry of ocean island basalts from the South Atlantic: Ascension, Bouvet, St. Helena, Gough and Tristan da Cunha Barry L. Weaver, David A. Wood, John Tarney and Jean Louis Joron S U M M A R Y: Basaltic and hawaiitic lavas from the S Atlantic Ocean islands of Ascension, Bouvet, St. Helena, Gough and Tristan da Cunha have been analysed for major elements and a wide range of trace elements. There is marked chemical (particularly trace-element) diversity between these islands which parallels observed Pb, Sr and Nd isotopic variations, and implies considerable large-scale heterogeneity in the source regions for ocean island volcanism. Trace-element and isotopic variation within individual islands suggests significant small-scale source heterogeneity. Abundance ratios between highly-incompatibletrace elements (Rb, Ba, Th, U, K, Ta, Nb and La) appear not to be fractionated during partial melting (except at low degrees of melting in the production of Tristan da Cunha lavas) and can be used to infer source characteristics. Lavas from the islands of Ascension, Bouvet and St. Helena have comparable highlyincompatible trace-element ratios (e.g. La/Nb, Th/Ta, La/Th, Th/U, Ba/La and Ba/Nb), with the exception of the apparent depletion of Rb and K relative to other highlyincompatible elements in St. Helena lavas. Tristan da Cunha and Gough lavas have similar La/Th ratios but much higher La/Nb, Th/Ta, Th/U, Ba/La and Ba/Nb ratios than do lavas from the other islands. These differences reflect depletion in Nb and Ta and enrichment in Ba relative to other highly-incompatibleelements in Gough lavas. The anomalous behaviour (enrichment) of Nb and Ta compared with other highlyincompatible trace elements, and specific consideration of relations between Ba, La and Nb, suggest that ancient subducted ocean crust, rather than primitive or depleted mantle, is the main component of the mantle source for ocean island basalts. Gough, Tristan da Cunha and Walvis Ridge lavas seem to be derived from such a source which has been contaminated by a component with high Ba/Nb, La/Nb and Ba/La ratios. Correlation between Ba/Nb and La/Nb in Gough basalts and hawaiites implies that these lavas contain variable proportions of this component. Pelagic sediment has the high Ba/Nb, La/Nb and Ba/La ratios required of this contaminant, and a few per cent of such material in the mantle source for Gough, Tristan da Cunha and Walvis Ridge lavas would account for their trace-element chemistry.
Introduction Ocean island basalts (OIBs) are generally regarded as being derived from chemically-anomalous mantle sources, and are associated with hot-spot activity (Morgan 1971, 1972). The strong enrichment in incompatible trace elements evident in these volcanics (e.g. Wood et al. 1981) represents a relatively recent event in the source region, as isotopic constraints demand that the magmas are derived from mantle sources with a time-integrated depletion in, for instance, Rb relative to Sr and Nd relative to Sm (O'Nions et al. 1977). The relationship of this source to the continental crust and the mantle source for 'normal' (depleted) mid-ocean ridge basalts (NMORB) is of considerable interest, and has been extensively studied using isotope systematics (O'Nions et al. 1977; Sun 1980; White & Hofmann 1982; Zindler et al. 1982; White 1985). Four main hypotheses have been advanced to explain the origin of chemically-anomalous OIB. These hypotheses suggest that the origins of these volcanics are as follows:
1 From primordial lower mantle which has undergone little or no chemical fractionation since shortly after the formation of the Earth (Schilling 1973; Sun & Hanson 1975; Dupr6 & All6gre 1980; All6gre 1982). 2 From subducted ocean crust which has had a considerable residence time at the boundary between the upper and lower mantle (Hofmann & White 1982; Ringwood 1982; Vollmer 1983; White 1985). 3 From material which was originally part of the subcontinental mantle, but which is now resident in the deep asthenosphere (McKenzie & O'Nions 1983; Cohen et al. 1984). 4 From a mantle source which has been contaminated by recycled continental crustal material (Hawkesworth et al. 1979; Cohen & O'Nions 1982; White 1985). However, relatively few comprehensive traceelement data are available for ocean island lavas which might provide additional important constraints on these possibilities. Here we report major and detailed trace-element data for OIBs
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 253-267.
253
B . L . W e a v e r e t al.
254
from five islands in the S Atlantic Ocean: Ascension, Bouvet, St. Helena, Gough and Tristan da Cunha (Fig. 1). These ocean islands are well characterized isotopically (O'Nions & Pankhurst 1974; O'Nions et al. 1977; Sun 1980; Cohen & O'Nions 1982; White & Hofmann 1982), and display considerable inter-island (and often intra-island) isotopic heterogeneity (Sun 1980; Cohen & O'Nions 1982). They are individually representative of islands which are both typical and extreme in terms of the isotopic variation observed in ocean island volcanics. Gough and Tristan da Cunha lavas have Nd and Sr isotope compositions which fall on the mantle array and lie close to supposed bulk Earth 143Nd/144Nd and 87Sr/S%r ratios (O'Nions et al. 1977), although He isotopes are suggestive of a non-primitive source (Kurz et al. 1982). Compared with MORB and other OIBs from the Atlantic Ocean, Gough and Tristan da Cunha lavas display anomalous 2~176 2~ 2~ and 2~176 ratios, having non-radiogenic 2~176 ratios and somewhat radiogenic :~176 ratios. However, St. Helena lavas plot somewhat to the left of the Nd-Sr mantle array with 143Nd/144Nd ratios greater than and 87Sr/86Sr ratios less than bulk Earth values. Additionally, St. Helena lavas have unusually radiogenic Pb isotope compositions (Sun 1980). Ascension and Bouvet lavas might be termed more isotopically typical of ocean island volcanics; they plot on (or in the case of Ascension possibly slightly to the left of) the Nd-Sr mantle
C *o
F.
A
'"
~
FIG. 1. Map of the main ocean islands in the Atlantic Ocean to the S of latitude 40~ : M, Madeira; C, Canary Islands; CV, Cape Verde Islands; F, Fernando de Noronha; A, Ascension; SH, St. Helena; T, Trindade; TC, Tristan da Cunha; G, Gough; DT, Discovery Tablemount; B, Bouvet.
array, although Bouvet lavas have less radiogenic Nd and more radiogenic Sr than do Ascension lavas. Pb isotope ratios for Ascension and Bouvet lava,s are comparable with those of the majority of Atlantic Ocean islands (e.g. the Canary Islands, the Azores, the Cape Verde Islands and Fernando de Noronha (Sun 1980)).
Sampling and analytical techniques Samples from Ascension (Daly 1925; Harris 1983), Bouvet (Verwoerd et al. 1976; Imsland et al. 1977; LeRoex & Erlank 1982), St. Helena (Daly 1927; Baker 1969) and Tristan da Cunha (Baker et al. 1964) were obtained from the collections of the British Museum (Natural History), London. The samples from Gough Island were obtained from the collections of LeMaitre (1962) at the Department of Earth Sciences, University of Cambridge. Relevant geological details for each island, some petrographic data and a limited amount of geochemical data are available in the above cited publications. In all instances samples were chosen to cover the full compositional range of lavas erupted on each island, although only data for basaltic and hawaiitic rocks are reported here. Data for the major elements and the trace elements Cr, V, Ni, Zn, Rb, Sr, Ba, Zr, Nb, Y, Th, La, Ce and Nd were obtained by X-ray fluorescence (XRF) analysis at the University of Leicester using a Philips PWl400 automatic spectrometer. Major-element determinations were made using an Rh X-ray tube on glass fusion beads utilizing a lithium tetraborate-lithium metaborate flux and with calibrations produced against international standards. Major-element data have been corrected to take account of weight loss on ignition of the sample at 900~ Trace-element concentrations were determined on pressed powder pellets using calibrations produced against international standards and standard spikes. Mass absorption corrections were applied using the intensity of the Rh Kct Compton scatter peak for the elements Ni, Zn, Rb, Sr, Y, Zr, Nb and Th and by using the intensity of the W L0r Rayleigh scatter peak combined with Fe K~ counts to cross the Fe absorption edge for the elements Cr, V, Ba, La, Ce and Nd. A more detailed description of the analytical procedures is given by Weaver et al. (1985). Data for the the trace elements Cr, Ni, Sc, Co, Cs, Rb, Ba, Zr, La, Ce, Eu, Tb, Hf, Ta, Th and U were obtained by epithermal neutron activation analysis (ENAA) at C.N.R.S., Universit6 Pierre et Marie Curie, Paris. Full details of the
Geochemistry of ocean island basalts analytical procedure, the precision and the accuracy are given by Chayla et al. (1973) and Jaffrezic et al. (1980). Where there is overlap in the tabulated data (Table 1) between the two analytical techniques we have reported the XRF data for the elements Cr, Ni, Rb, Ba, Zr and Ce and the NAA data for Th and La. A standard Fe203/FeO ratio of 0.22 was used in the calculation of CIPW normative compositions.
Geochemistry Major elements Representative analyses of lavas from the islands of Ascension, Bouvet, St. Helena, Gough and Tristan da Cunha are presented in Table 1. In this paper we have attempted to select only those lavas which by chemical criteria appear to have undergone limited low-pressure crystal fractionation. Our main criteria in selecting samples have been based on uniformity of trace-element ratios rather than major-element composition. Essentially, only those lavas which have a basaltic or hawaiitic major-element chemistry are considered. In Fig. 2 normative compositional data for basalts and hawaiites from the islands are plotted in terms of the components ne-di-ol-hy-qz. Considerable variation in normative composition exists between islands, and often within individual islands. Bouvet basalts and hawaiites are uniformly ol-di-hy normative and have relatively high di/ol and di/hy ratios (Fig. 2). St. Helena basalts and hawaiites (with the exception of one sample) are ne normative, while Ascension basalts and hawaiites are somewhat variable in composition, ranging from ne normative to moderately Di
/
Ol
k
Hy
FIG. 2. Plot of normative components ne-ol-di-hy-qz in basalts and hawaiites from Bouvet (I~), Ascension (-k), Gough ( 9 St. Helena (D) and Tristan da Cunha (0). All normative compositions were calculated using an Fe203/FeO ratio of 0.22.
255
hy normative (Fig. 2). There is an interesting contrast between the chemistry of Gough and Tristan da Cunha basalts and hawaiites. The Tristan da Cunha samples range from moderately to mildly ne normative, with some samples straddling the di-ol join with approximately constant ol/di ratios. Gough samples, however, are extremely variable in normative mineralogy, ranging from hawaiites with a negligible amount of normative ne (essentially lying on the di-ol join) through basalts and hawaiites with quite variable proportions of ol, di and hy to a hawaiite with a small amount of normative qz (Fig. 2). Clearly, few (if any) of the samples considered in this study are representative of primary mantle melts; in general Mg numbers (100 Mg/(Mg+ FEZ+)) and Ni and Cr contents (e.g. Table 1) are low, and those lavas with Mg numbers which might be appropriate for primary mantle melts are generally olivine- and/or clinopyroxenephyric. Indeed, in these islands the major-element compositions are largely controlled by removal or accumulation of the assemblage olivine+ clinopyroxene + plagioclase + Fe-Ti oxide in basaltic and hawaiitic lavas (Zielinski & Frey 1970; LeRoex & Erlank 1982; Harris 1983; LeRoex 1985).
Trace elements In this section we are primarily concerned with assessing the compositional variation in the mantle sources for OIBs. The most useful elements for this purpose are those trace elements which have very low bulk crystal-liquid partition coefficients D during both partial melting and fractional crystallization processes. In particular, ratios between such elements should be representative of the ratios in the mantle source regions. Cs, Rb, Ba, Th, U, K, Ta, Nb and La are highly-incompatible trace elements for which D is very low (less than or equal to 0.01), and the behaviour of these elements will be emphasized. In evolved compositions (mugearites and trachytes) from each island ratios between some of these trace elements become influenced by lowpressure crystal-liquid processes. In the less evolved basaltic and hawaiitic compositions considered here there is no evidence for highlycompatible trace-element fractionation attributable to low-pressure crystallization. However, the degree to which, in particular, alteration and partial melting are capable of affecting abundance ratios needs to be assessed.
Alteration In general the abundances of the large-ion lithophile (LIL) elements Cs, Rb, Ba, Th, U, K
256
B. L. W e a v e r et al. . . . .
~
~
~
~o
I ~ . . . .
~9
~
I~
P"~.CN r
,:~r"q
r~..,r
,#
P,,
~ ~
~
~
~
~
~
~
~
~.,~
g ra
ial
.<
~ 9 ~ o o o O o ~
~ ~
Geochemistry o f ocean &land basalts and Sr are susceptible to variation owing to lowtemperature alteration processes (Hart et al. 1974; Clague & Frey 1982). Element mobility accompanying sub-aerial alteration and hydration is evident in a number of the ocean island lavas studied, although the effects are not always consistent. Cs abundances are particularly unreliable, and can be either enriched or depleted relative to immobile trace elements (Nb, Ta, La) in both altered and otherwise fresh lavas. U is moderately mobile upon alteration, usually with a loss of U leading to an increase in Th/U ratio in altered lavas (Fig. 3). K and Rb appear to be somewhat less mobile than Cs and U, the mobility
257
of K and Rb generally resulting in an increase in the K / R b ratio. There is also some evidence that Ba and Th may display more limited mobility. These effects clearly contribute to the scatter evident in the LIL element plots of Fig. 3, and it is necessary to use care in the interpretation of the behaviour of some of these elements.
Partial melting For trace elements to be good indicators of mantle source composition they must not be strongly fractionated during partial melting (e.g. by low degrees of melting or owing to the effect of K2o
L8
y/
/
fS
ell,..- /
TI1
Th I
I
I
I
I
I
I
I
[
1
. []
.
~ '*
[] ~
i Th I
I
I
9
i
i
T h
Ba lOOO
Rb
:/y /
800
600
/ 400
,sJ
[]4
/
Z41~*/S"~ _ L J ~
g4v" /
Th L 2
I 4
I 6
L 8
10
.~ w /
I 2
//:/
2oo Th I 4
i 6
i 8
1,, 10
FiG. 3. Plots of Th against other highly-incompatibletrace elements for the islands of Ascension ( , ) , Bouvet (1), St. Helena (F1), Gough (O) and Tristan da Cunha (0).
258
B. L. Weaver et al.
residual mineralogy). This is clearly a problem with alkali basalts, as the degree of melting may be less than 5~o-10~ (Clague & Frey 1982). At these low degrees of melting, phases other than olivine, orthopyroxene, clinopyroxene and garnet (such as amphibole and/or phlogopite), which are capable of fractionating Ba, K, Rb and Th from other highly-incompatible trace elements (Philpotts & Schnetzler 1970), may be residual. From the systematic behaviour of highlyincompatible trace elements outlined below we believe that for Ascension, Bouvet, St. Helena and Gough there is little chemical fractionation attributable to partial melting. However, anomalous values for some trace-element ratios in Tristan da Cunha lavas suggest that these ratios may not be representative of the source. The geographical proximity of Gough and Tristan da Cunha implies that these islands are related to the activity of the same hot spot, and indeed Nd, Sr, Pb and He isotopes indicate a chemically similar source for their lavas. However, the strongly ne-normative nature of Tristan da Cunha lavas relative to Gough lavas (Fig. 2) may be the result of a smaller degree of partial melting. Additionally, significant differences in rare-earthelement (REE) fractionation between the two islands (as expressed by Cen/Yn ratios which have values of 8.0 and 10.8 for Gough and Tristan da Cunha respectively (Table 2)) could be interpreted in terms of a lower degree of partial melting in the production of the Tristan da Cunha lavas. Similar 143Nd/144Nd ratios (Cohen & O'Nions 1982; White & Hofmann 1982) indicate that the source for Tristan da Cunha lavas has had a similar time-integrated Sm/Nd ratio to that of the source for Gough lavas. Relationships amongst the trace elements Rb, Ba, K and Th are somewhat ambiguous regarding the possible role of amphibole and/or phlogopite
as residualphases in the mantle source for Tristan da Cunha lavas. Partition-coefficient data suggest that DR > ORb ~ DBa for amphibole, whereas DK ,~ DRb > DBa for phlogopite (Philpotts & Schnetzler 1970). T h / N b and Th/Ta ratios (Fig. 3; Table 2) are not fractionated signficantly during different degrees of melting in the production of Gough and Tristan da Cunha lavas. Variations in Ba/ Nb, K / N b and R b / N b ratios (Table 2) demonstrate that Ba and K are equally partitioned into the residue in the production of Tristan da Cunha lavas, but that Rb behaves in a more incompatible fashion. However, for residual amphibole and/or phlogopite Ba should behave more incompatibly than K. Clearly it is not possible to constrain the residual mantle mineralogy tightly during the partial-melting event which generated the Tristan da Cunha lavas, except that phases retaining K, Ba and (to a lesser extent) Rb appear to be required. Source characteristics Abundances of La, Ta, Rb, K, Ba and U are plotted against Th (used as a general index of fractionation) in Fig. 3. In addition, selected element ratios (with associated l a standard deviations) for each island are tabulated in Table 2. Mean La/Th ratios are constant for the islands with the exception of Tristan da Cunha, which has a lower La/Th ratio owing to slightly more compatible behaviour of La during the low degree of partial melting. A rather larger variation in Th/Ta ratios between the islands is evident, with Gough and Tristan da Cunha having higher Th/ Ta ratios (1.47 and 1.50 respectively) than those of lavas from Ascension, Bouvet and St. Helena (range 1.00-1.09). Plots of Rb, K, Ba and U against Th display somewhat more scatter than those of Ta and La, largely owing to alteration
TABLE 2. Means and l a standard deviations for selected element ratios in basaltic and hawaiitic lavas from Gough, Tristan da Cunha, St. Helena, Ascension and Bouvet Islands Ce,/Y, Zr/Nb La/Th Ba/Th Rb/Th Th/U Th/Ta La/Nb Th/Nb Ba/Nb K/Nb Rb/Nb Ba/La Gough
8.0 0.9
6.8 0.8
9.1 0.6
154 20
9.5 0.9
4.86 0.28
1.47 0.15
0 . 9 7 0.105 16.1 0.13 0.009 2.8
432 47
0.99 0.15
16.6 1.6
Tristanda Cunha
10.8 1.2
4.2 0.4
7.8 0.8
103 12
8.2 1.0
4.50 0.08
1.50 0.12
0.86 0.108 11.4 0 . 0 6 0.007 0.8
307 16
0.88 0.09
13.2 0.7
St. Helena
7.3 0.9
4.5 0.2
8.9 0.5
77 8
4.9 0.5
3.78 0.10
1.09 0.07
0 . 6 9 0.078 0 . 0 3 0.005
5.9 0.2
179 9
0.38 0.04
8.7 0.7
Ascension
4.6 0.5
5.3 0.4
8.7 1.0
92 12
6.9 0.8
3.54 0.17
1.05 0.09
0.65 0.04
0.073 0.006
6.8 0.5
230 17
0.46 0.05
10.3 1.2
Bouvet
3.9 0.1
6.9 0.3
9.3 0.4
83 4
7.1 0.4
3.68 0.19
1.00 0.03
0 . 6 8 0.073 0.04 0.003
6.0 0.3
272 13
0.52 0.02
8.9 0.4
Geochemistry of ocean island basalts effects. However, it is apparent that St. Helena lavas have lower Rb/Th ratios (average 4.9) than those of Ascension and Bouvet lavas (average ratios 6.9 and 7.1), which are in turn lower than the Rb/Th ratios in Gough lavas (average 9.5). A very similar pattern is evident for K/Th ratios, with the exception that Bouvet lavas have a higher K/Th ratio than do Ascension lavas. Both Gough and Tristan da Cunha lavas (when altered samples are excluded) have considerably higher Th/U ratios (average 4.86 and 4.50 respectively) than those of Ascension, Bouvet and St. Helena lavas (range of average ratios, 3.54-3.78). There is high degree of variation in Ba/Th ratios, often within each island, but Ascension, Bouvet and St. Helena lavas have rather consistent ratios of 77-92, whilst the Ba/Th ratio of 154 in Gough lavas is high. It is evident from the foregoing that, in general, the trace-element characteristics of the lavas from Gough and Tristan da Cunha (where fractionation during melting affects some element ratios) are very different from the characteristics of the lavas from Ascension, Bouvet and St. Helena. Within the latter three islands the St. Helena lavas appear to be somewhat depleted in K and Rb relative to other highly-incompatible trace elements. It is also apparent that within each island there is often a degree of heterogeneity in element ratios (expressed by the standard deviation) which cannot be simply attributed to analytical uncertainty or alteration effects. This observation is in accord with the well-documented isotopic heterogeneity within lavas from a single island (Sun 1980; Cohen & O'Nions 1982). Figure 4 is a plot of Th against Zr, a less incompatible element that is likely to be fractionated from highly-incompatible trace elements (e.g. Th) at low to moderate degrees of partial
400
200 /
/
-
121
Th
J
41
J
6L
I
8J
I
110
L
FIG. 4. Plot of Th against Zr for basalts and hawaiites from the islands of Ascension ( , ) , Bouvet (*), St. Helena (U]), Gough (C)) and Tristan da Cunha ($).
259
melting. The lavas of Bouvet and Ascension define good linear trends on this diagram which intersect the Zr axis near the origin, indicating slightly more compatible behaviour of Zr than Th during melting. A rather wide variation in Zr/Th ratios in Gough lavas results in an imprecise correlation between Zr and Th, with an extrapolated trend which intersects the Zr axis at a rather higher Zr value implying a significantly more compatible behaviour of Zr than Th during partial melting. Similarly, good correlation of Zr and Th abundances in lavas from St. Helena and Tristan da Cunha have high intercepts on the Zr axis (particularly in the case of Tristan da Cunha). Although difficult to quantify, owing to uncertainties in the degree of relative light REE enrichment in mantle source, high Ce,/Y, ratios for St. Helena, Gough and Tristan da Cunha lavas (7.3, 8.0 and 10.8 respectively) suggest a lower degree of partial melting in the generation of these lavas than for the Ascension and Bouvet lavas (with mean Cen/Y, ratios of 4.6 and 3.9 respectively). It is also clear that for the fomer three islands Zr does not behave as a highly-incompatible element during partial melting. Thus, depending upon the degree of melting, use of ratios employing Zr (e.g. Zr/Nb ratios) will not necessarily reflect source characteristics for OIBs. This becomes apparent from a plot of Zr against Nb (Fig. 5) where Tristan da Cunha lavas display a strong covariation of Zr and Nb which has a considerable intercept on the Zr axis. The distinctive trace-element characteristics of Gough and Tristan da Cunha as opposed to Ascension, Bouvet and St. Helena basalts and hawaiites is well displayed on plots of La, K20 and (particularly) Ba against Nb (Fig. 5). Average La/Nb ratios in Ascension, Bouvet and St. Helena lavas are rather uniform at 0.65-0.69 (Fig. 5; Table 2), but significantly higher at 0.97 in Gough lavas (Fig. 5; Table 2). There is some variation of K/Nb ratios between Ascension, Bouvet and St. Helena lavas with, as already noted, St. Helena basalts and hawaiites (K/Nb = 179) being somewhat depleted in K (Fig. 5) compared with Ascension and Bouvet basalts and hawaiites (K/ Nb values of 230 and 272). The K/Nb ratio of Gough lavas is high at 432 (Fig. 5; Table 2). Extremely large differences in Ba/Nb ratios are evident between lavas from the different islands (Fig. 5); Ascension, Bouvet and St. Helena lavas have consistent Ba/Nb ratios (Fig. 5) of 5.9-6.8, whereas the average Ba/Nb ratio in Gough lavas (Fig. 5) is very high at 16.1 but with considerable dispersion of values about the average ratio (Table 2). Representative plots of incompatible element abundances normalized to an estimated primor-
260
B . L . W e a v e r et al. La
Zr
80
400
60 300 }
40 200
Nb
~/
I
1
I
I
Nb
I
1000
I
I
I
I 40
I 60
I 80
I
K20
800
600
// 400
///
/
200
Nb I 20
410
I
1
I
60
80
100
I 20
I 100
FIG. 5. Plots of Nb against other incompatible trace elements for basalts and hawaiites from the islands of Ascension (~-), Bouvet (#), St. Helena (R), Gough (O) and Tristan da Cunha (0). dial mantle composition (Wood et al. 1981) for basaltic and hawaiitic lavas from Ascension, Bouvet, St. Helena, Gough and Tristan da Cunha appear in Fig. 6. The order in which the elements are plotted corresponds to decreasing D from right to left in the diagram for equilibration of basaltic melt with four-phase lherzolite (Wood 1979). These diagrams illustrate some interesting features of the trace-element chemistry in addition to those described above. The general patterns are of increasing normalized abundance from right to left in the diagrams, although amongst the highly-incompatible elements Cs-K there is a general decrease in normalized abundance with increasing incompatiblity (Fig. 6). The anomalous behaviour of Sr relative to Ce and Nd is due to plagioclase fractionation. Hf abundances define consistent negative anomalies relative to Zr and P, and Zr/Hfratios are constant between all five islands (range of average ratios from 43 to 45). These values are considerably higher than the primordial Zr/Hf ratio of 31
(Wood 1979), but similar to Zr/Hf ratios observed in other alkali basalt suites (Frey et al. 1978; Clague & Frey 1982). Maximum normalized abundances occur for Nb and Ta (Fig. 6), and Nb/Ta ratios are very uniform (within analytical error) both within and between islands (range of average Nb/Ta ratios from 13.7 to 14.1). Gough lavas are notable for the similar normalized abundances of Nb, Ta, La, K and Ba (Fig. 6). From Nb to Cs there is a general decrease in normalized highly-incompatible element abundances, with the exception of the islands of Tristan da Cunha and (particularly) Gough where Ba is enriched (producing a positive Ba spike) relative to other highly-incompatible LIL elements (e.g. Cs, Rb, Th, U and K). The two most striking features of Fig. 6 are the large positive anomaly present for Nb and Ta in Ascension, Bouvet and St. Helena lavas, and the disappearance of this anomaly in Gough (and to an extent Tristan da Cunha) lavas with the simultaneous development of a positive Ba anomaly. The
Geochemistry of ocean island basalts 100
~
=,
~
/
•
ST. HELENA BOUVET
261
~~.~SCENSION
o lo
i
i
L
i
i
i
i
i
i
i
I
i
I
I
I
L
I
I
I
I
i
L
I
I
l
I
I
GOUGH I-Z <
0 n-
10
, CS
, , R b Ba
, Th
J U
: K
, Ta
, , L Nb L a C e
L Sr
J Nd
L P
H'f
' Zr
~ Eu
' TI
b" T
~ Y
J Cs
R"
' b Ba
vL h
, U
, K
~ Ta
, , ~ Nb La Ce
J Sr
.'d
' P
' Hf
' Zr
J Eu
' xJb Ti
' Y
F]6.6. Plots of incompatible element abundances normalized to primordial mantle (Wood et al. 1981) for selected basalts and hawaiites from the islands of Ascension, Bouvet (stippled field), St. Helena, Gough and Tristan da Cunha. magnitude of the N b - T a anomaly (as expressed by La/Nb and Th/Ta ratios in Table 2) is constant in Ascension, Bouvet and St. Helena lavas, and typical of OIBs from other islands in the Atlantic Ocean (B. L. Weaver et al., unpublished data) as well as of continental alkali basalts (e.g. Thompson et al. 1984).
Discussion Amongst the highly-incompatible trace elements it is evident that the behaviour of, in particular, Ba, Nb and Ta may provide strong constraints on potential source reservoirs for OIBs. For instance, partial melting (especially at low degrees of melting) of primordial (or primitive) mantle should produce melts with highly-incompatible trace-element (Cs-La) abundances which, when plotted on normalized diagrams such as Fig. 6, would be either unfractionated or, more likely, slightly enriched in more highly-incompatible elements (e.g. Cs) over less highly-incompatible elements (e.g. La). It would not be possible to produce significant N b - T a or Ba anomalies (Fig. 6) from such a primordial source. Furthermore, the depletion of Cs and Rb relative to K and La evident from Fig. 6 argues for derivation of OIBs from a source which is itself depleted in more highly-incompatible trace elements. It
would therefore seem that chemically primitive mantle (such as the lower mantle (All6gre 1982)) is not a suitable source for the production of OIBs. In models for the chemical evolution of the crust-mantle system the continental crust and the mantle source for N-MORB are considered to be complementary reservoirs (e.g. Jacobsen & Wasserburg 1979; O'Nions et al. 1979; All6gre et al. 1983; All6gre & Rousseau 1984). To a first approximation those incompatible trace elements which are depleted in N-MORB are enriched in the continental crust. However, this is clearly not the case when consideration is given to the chemical budgets for Nb and Ta. Subductionzone-related magmas, and indeed the bulk continental crust itself, are strongly depleted in Nb and Ta relative to other highly-incompatible trace elements (e.g. Saunders et al. 1980; Thompson et al. 1984). For example, continental margin basalts and andesites typically have La/Nb ratios in the range 2.2-4.7 (data from the compilations of Ewart (1982)), while estimates of the bulk composition of the continental crust (Taylor & McLennan 1981 ; Weaver & Tarney 1984) suggest that the crust itself is similarly depleted, with La/ Nb ratios of 1.7-2.2. Island-arc volcanics have equally high La/Nb ratios (e.g. Tarney et al. 1981). Estimates of the composition of typical NMORB (Sun 1980; Tarney et al. 1981 ; Pearce
262
B . L . W e a v e r e t al.
1983) yield a La/Nb ratio of close to 1.0 which, considering the very similar D values for La and Nb during MORB genesis (Briqueu et al. 1984), represents the La/Nb ratio in the MORB reservoir. A La/Nb ratio of 0.83-0.94 is likely for the chondritic Earth or primordial mantle (Sun 1980; Taylor & McLennan 1981; Thompson et al. 1984). Thus the production of continental crust with La/Nb > 2 from primordial mantle with La/ Nb ~ 0.9 requires a volumetrically large complementary reservoir with a low (considerably less than 1) La/Nb ratio; the N-MORB source does not represent an appropriate reservoir. However, this reservoir would appear to be the source for ocean island (and continental) alkali volcanics with low La/Nb ratios (about 0.7). It has been suggested that oceanic and continental alkali basalts could be the product of partial melting of subducted ocean crust which has been resident in the deep mantle for a considerable period of time (Hofmann & White 1982; Ringwood 1982). Such a model is consistent with Pb isotope data for OIBs which imply derivation from source regions that have maintained high U/Pb ratios (typical of altered oceanic crust) for periods of time of the order of (1.52.0) x 103 Ma (Tatsumoto 1978; Sun 1980; Chase 1981). With regard to the behaviour of Nb and Ta, a key question is: what chemical modification occurs in oceanic crust during subduction ? Traceelement and isotopic considerations of the chemistry of subduction-related magmas invariably require a component derived from the subducted ocean crust (Kay 1984; O'Nions 1984; Wilson & Davidson 1984). This component either could be a fluid produced by dehydration of the slab which carries LIL elements (and La, Ce?) into the overlying mantle wedge, or may involve a degree of partial melting of the slab, again under hydrous conditions (Kay 1984; Wilson & Davidson 1984; Wyllie 1984). The effect of these processes on the trace-element chemistry of the subducted ocean crust would be that LIL elements would be strongly depleted in subducting N-MORB-type ocean crust. However, it is also likely that the conditions pertaining during dehydration (and melting ?) of the ocean crust would stabilize minor phases (especially sphene (Hellman & Green 1979)) in the ocean crust which would retain high-field-strength elements such as Nb and Ta (Saunders et al. 1980; Briqueu et al. 1984; Thompson et al. 1984). Thus Nb and Ta would become strongly decoupled in their geochemical behaviour from other highly-incompatible trace elements, and the subducted ocean crust would develop low La/Nb and Ba/Nb ratios whilst being severely depleted in LIL elements. Unfortunately, lack of knowledge of phase relations
and partition coefficients for the probable mineralogy of the ocean crust in the deep mantle (Ringwood 1982) precludes modelling of subsequent partial melting of this material. It is perhaps probable that moderate degrees of melting (10~-15~) would produce undersaturated liquids with strong light REE enrichment (Hofmann & White 1982), positive N b - T a anomalies and increasing relative depletion in the more highly-incompatible trace elements reflecting the composition of the old ocean crust source. Such melts would mix with the mantle to form an enriched hot-spot component, the composition of which is strongly reflected in the chemistry of lavas from Ascension, Bouvet and St. Helena. This component might further interact and mix with the depleted mantle source of N-MORB (producing transitional or T-type MORB), or could invade and vein (metasomatize) depleted upper-mantle peridotite (Menzies 1983; Hawkesworth et al. 1984) which might itself be subsequently melted to produce alkali basalts. These general features are demonstrated in a plot of Ba/Nb against La/Nb in Fig. 7. As Nb is used as a denominator in the ratios plotted on both axes, mixing trends will be straight lines on
19 9
A
Ba/Nb
15
/~527
~30A I 0.6
N-MORB I 0.8
I 1.0
] 1.2
La/Nb
FIG. 7. Plot of Ba/Nb against La/Nb ratios for basalts and hawaiites from the islands of Ascension (~), Bouvet (,), St. Helena ([-]), Gough (O) and Tristan da Cunha (0). Also plotted are average ratios for basalts from DSDP Holes 525A, 527, 528 and 530A from the Walvis Ridge (data from Humphris and Thompson (1983)), and fields showing the average Ba/Nb and La/ Nb ratios in N-MORB and primordial mantle.
Geochemistry of ocean island basalts
this diagram. Since DBa~
263
ditions of partial melting where DNb,~DLa. Rather, the chemistry is best explained by mixing between a 'typical' hot-spot component (represented by Site 530A basalt or Ascension, Bouvet and St. Helena lavas) and a component with high Ba/Nb and La/Nb ratios. It is clear from the Ba/ Nb-La/Nb relations in Fig. 7 that an N-MORB component or source is not a suitable endmember in any mixing model, as pointed out by Richardson et al. (1982) for Walvis Ridge lavas and LeRoex (1985) for Gough lavas. The wide range in, and good correlation between, Ba/Nb and La/Nb ratios in Gough lavas (Fig. 7) suggests variable proportions of the high Ba/Nb and La/ Nb ratio components in these lavas. This trend is not simply the result of dilution of Nb causing the Ba/Nb and La/Nb ratios to increase; with decreasing magnitude of the N b - T a anomaly (e.g. increasing La/Nb ratio) Ba abundances increase relative to other highly-incompatible trace elements such that there are reasonable correlations between La/Nb, Ba/Th and Ba/La ratios. Thus the end-member with the high Ba/ Nb and La/Nb ratios must be a component which itself has a negative Nb anomaly but a positive Ba anomaly. Such chemical characteristics are typical of pelagic sediments, which are strongly enriched in Ba relative to Th, Rb and Nb and in La relative to Nb (Hole et al. 1984; Kay 1984; Thompson et al. 1984). Terrigenous sediments are also a possible contaminant, but typically have rather low La/Nb, Ba/Nb and Th/Nb ratios which makes them unsuitable end-members in mixing models. Hence the trace-element data are qualitatively suggestive (as are the He isotope data (Kurz et al. 1982)) of a (pelagic) sedimentary component in the Gough, Tristan da Cunha and Walvis Ridge lavas. We can make some attempt at quantifying the mixing process using Ba/Nb and La/Nb ratios, and approach the question of low-pressure mixing of the common parental ocean island magma type with pelagic sediment versus high-pressure mixing of components in the source region of some OIBs. A parental (or primary) OIB would have Ba/Nb and La/Nb ratios of 6.5 and 0.68 (typical of Ascension, Bouvet and St. Helena lavas (Table 2)), and might have a Nb content (corrected for the effects of low-pressure crystal fractionation) of 25 ppm. For an estimate of the composition of pelagic sediment we use the Pacific authigenic weighted mean sediment (PAWMS) of Hole et al. (1984), which has Ba/ Nb and La/Nb ratios of 1070 and 20.6 respectively and an Nb concentration of 1.25 ppm. A mixture of 20~ PAWMS + 80~ parental OIB is required to generate high Ba/Nb and La/Nb ratios (19.6 and 0.93 respectively) which approximate to the
264
B . L . W e a v e r e t al.
values for these ratios in Gough lavas (Table 2). Varying the proportions of nannofossil ooze, diatom ooze, ferruginous clay and pelagic clay in the sediment (PAWMS comprises 95~ nannofossil ooze plus 5~ ferruginous clay) has little effect on element ratios in the sediment or on the amount of pelagic sediment required as a contaminant. Addition of such large proportions of pelagic sediment would have a drastic effect on the major-element composition of the mixture for which there is little evidence. Also, why should such low-pressure mixing (presumably associated with a sub-volcano magma chamber) be apparently restricted geographically to the islands of Gough and Tristan da Cunha in the South Atlantic and to islands in the Indian Ocean (Dupr6 & All6gre 1983)? However, if mixing occurs in the source regions of some OIB the endmembers would be pelagic sediment and subducted ocean crust (the main mantle source component for OIB). The subducted ocean crust would have a Nb content of approximately 3 ppm and Ba/Nb and La/Nb ratios similar to those of the hot-spot material. In consideration of pelagic sediment older than Mesozoic, nannofossil and diatomaceous oozes would not be a component. For a ratio of ferruginous clay (Hole et al. 1984) to pelagic clay (Li 1982) of 2:8, the ancient pelagic sediment might have a Nb content of 12.3 ppm and La/Nb and Ba/Nb ratios of 5.7 and 341 respectively. In this case a mixture of 1~ pelagic sediment plus 99~ OIB source would have a La/Nb ratio of 0.92 and a Ba/Nb ratio of 22, similar to the values for these ratios in Gough lavas. Such a situation is compatible with the injection of a small proportion of sediment into the deep mantle along with subducted ocean crust. Additional constraints on mixing processes in Gough, Tristan da Cunha and Walvis Ridge lavas come from Pb, Sr and Nd isotope data (Sun 1980; Cohen & O'Nions 1982; Richardson et al. 1982). Isotopic trends within Walvis Ridge basalts suggest a limiting composition for the enriched end-member with radiogenic Sr (8~Sr/ 86Sr=0.7058) and non-radiogenic Nd (143Nd/ 144Nd = 0.5122) and Pb (e.g. 2~176 = 17.1) (Richardson et al. 1982). Amongst the Walvis Ridge and Gough and Tristan da Cunha lavas, Site 525A basalts have isotopic characteristics suggesting the greatest proportion of this enriched end-member, which from the Ba/Nb-La/Nb trend in Fig. 7 is equated with the high Ba/Nb and La/Nb ratio component. The isotopic composition of the depleted end-member is not well constrained by the Pb, Sr and Nd data (Richardson et al. 1982). However, Ba/Nb-La/Nb relations (Fig. 7) imply a hot-spot (rather than
MORB) end-member represented by Site 530A basalts, for which isotope data are lacking. The trend defined by Walvis Ridge and Tristan da Cunha lavas on a 2~176176176 diagram (Richardson et al. 1982) suggests that this depleted end-member could have 2~176 and 2~176 ratios similar to those of Ascension and Bouvet lavas. If such a hot-spot component has a 2~176 ratio of approximately 19.5, then about 40~ of the Pb in Gough and Tristan da Cunha (and about 70~ of the Pb in Site 525A) lavas would be derived from the component with high Ba/Nb and La/Nb ratios. A 2~176 ratio for the hot-spot component in excess of those of Ascension and Bouvet lavas would be required, however, as the 2~176 2~176 trend for Gough, Tristan da Cunha and Walvis Ridge lavas (Richardson et al. 1982) is parallel to that for MORB and other OIBs from the Atlantic. Thus the hot-spot component may not have isotopic compositions directly comparable with those of other islands in the S Atlantic. The isotopic composition of the limiting enriched end-member is not apparently compatible with that of typical pelagic sediment. Although non-radiogenic Nd and radiogenic Sr are consistent with the characteristics of pelagic sediment, non-radiogenic Pb is not. Modern Atlantic pelagic sediments generally have rather radiogenic Pb (2~176 ~ 19, 2~176 ~ 15.7 and 2~176 ~ 39.1 (Sun 1980)). However, if the model for derivation of OIB dominantly from old subducted ocean crust with the possibility of contamination occurring in the source is accepted, the sedimentary component of interest will be that subducted with the ocean crust at some considerable time (of the order of (1.52.0) x 103 Ma) in the past. The initial isotopic composition of this sediment is obviously uncertain, but it is likely to have had, relative to upper mantle at that time, radiogenic Sr and Pb and non-radiogenic Nd. The pelagic sediment would have had a Sm/Nd ratio less than the chondritic value (Sm/Nd ~ 0.20), a low Rb/Sr ratio (Rb/Sr 0.05) and very low U/Pb and Th/Pb ratios (# ~ 2 owing to high Pb contents in pelagic sediments of up to 50 ppm (Cohens & O'Nions 1982)). These arguments assume that the sediment maintains its chemical integrity through the subduction process. This is unlikely, but the chemical effects associated with dehydration and/or partial melting of sediment during subduction are very poorly constrained. Subsequent to subduction, growth of radiogenic Nd, Sr and Pb would be inhibited (strongly so in the case of Pb) in this material, and could result in the present isotopic composition predicted for the end-member with high Ba/ Nb and La/Nb ratios observed in the Gough,
Geochemistry of ocean island basalts Tristan da C u n h a and Walvis Ridge lavas. M u c h of the foregoing must r e m a i n speculative until the composition of this c o m p o n e n t can be more rigorously characterized. The variation of tracee l e m e n t ratios (Fig. 7) in G o u g h lavas suggests that a c o m b i n e d detailed trace-element and isotopic study of this island m i g h t be of particular i m p o r t a n c e in a p p r o a c h i n g this problem. In modelling the petrogenesis of Walvis Ridge and G o u g h lavas, R i c h a r d s o n et al. (1982) and L e R o e x (1985) ascribed the unusual isotopic and trace-element characteristics of these lavas to a ' d o u b l e - e n r i c h m e n t ' event in the m a n t l e source
265
region. W e would equate the first trace-element ' e n r i c h m e n t ' event to a melt c o m p o n e n t derived d o m i n a n t l y from old subducted o c e a n crust in the source region for OIB. This event produces OIB with the r a t h e r u n i f o r m trace-element composition of Ascension, Bouvet a n d St. H e l e n a lavas. The second ' e n r i c h m e n t ' event observable in the geochemistry of Gough, Tristan da C u n h a and Walvis Ridge lavas can be e q u a t e d with c o n t a m i n a t i o n of the OIB source region by a small a m o u n t (1%-2%) of ancient pelagic sedim e n t w h i c h was subducted into the deep m a n t l e along with the o c e a n crust.
References ALLI~GRE,C. J. 1982. Chemical geodynamics. Tectonophysics, 81, 109-32. ROUSSEAU,D. 1984. The growth of the continent through geological time studied by Nd isotope analysis of shales. Earth planet. Sci. Lett. 67, 1934. HART, S. R. & MINSTER, J.-F. 1983. Chemical structure and evolution of the mantle and continents determined by inversion of Nd and Sr isotopic data, I. Theoretical methods. Earth planet. Sci. Lett. 66, 177-90. BAKER, I. 1969. Petrology of the volcanic rocks of Saint Helena Island, South Atlantic. Geol. Soc. Am. Bull. 80, 1283-310. BAKER, P. E., GASS, I. G., HARRIS, P. G. & LEMAITRE, R. W. 1964. The Volcanological report of the Royal Society expedition to Tristan da Cunha. Phil. Trans. R. Soc. Lond., Ser. A, 256, 439-578. BRIQUEU, L., BOUGAULT, H. & JORON, J. L. 1984. Quantification of Nb, Ta, Ti and V anomalies in magmas associated with subduction zones: petrogenetic implications. Earth planet. Sci. Lett. 68, 297-308. CHASE, C. G. 1981. Oceanic island Pb: two-stage histories and mantle evolution. Earth planet. Sci. Lett. 52, 277-84. CHAYLA, B., JAFFREZIC, H. • JORON, J. L. 1973. Analyse par activation dans les mentions 6pithermique. Application ~t la d6termination d'616ments en trace dans les roches. C. R. Acad. Sci. Paris, 277, 273-5. CLAGUE, D. A. 8z FREY, F. A. 1982. Petrology and trace element geochemistry of the Honolulu volcanics, Oahu: implications for the oceanic mantle below Hawaii. J. Petrol. 23, 447-504. COHEN, R. S. & O'NIONS, R. K. 1982. Identification of recycled continental material in the mantle from Sr, Nd and Pb isotope investigations. Earth planet. Sci. Lett. 61, 73-84. -- - & DAWSON,J. B. 1984. Isotope geochemistry 'of xenoliths from East Africa: implications for development of mantle reservoirs and their interaction. Earth planet. Sci. Lett. 68, 209-20. DALY, R. A. 1925. The geology of Ascension Island. Proc. Am. Acad. Arts Sci. 60, 1-80. -
-
-
-
&
--
1927. The geology of Saint Helena Island. Proc. Am Acad. Arts Sci. 62, 31-92. DUPR~, B. & ALL~GRE, C. J. 1980. Pb-Sr-Nd isotopic correlation and the chemistry of the North Atlantic mantle. Nature, Lond. 286, 17-22. & -1983. Pb-Sr isotope variation in Indian Ocean basalts and mixing phenomena. Nature, Lond. 303, 142-6. EWART, A. 1982. The mineralogy and petrology of Tertiary-Recent orogenic volcanic rocks: with special reference to the andesitic-basaltic compositional range. In: THORPE, R. S. (ed.) Orogenic Andesites, pp. 25-95. Wiley, London. FREY, F. A., GREEN, D. H. & ROY, S. D. 1978. Integrated models of basalt petrogenesis : A study of quartz tholeiites to olivine melilitites from South Eastern Australia utilizing geochemical and experimental petrological data. J. Petrol. 19, 463-513. HARRIS, C. 1983. The petrology of lavas and associated plutonic inclusions of Ascension Island. J. Petrol. 24, 424-70. HART, S. R., ERLANK, A. J. & KABLE, E. J. D. 1974. Sea floor basalt alteration: some chemical and Sr isotopic effects. Contrib. Mineral. Petrol. 44, 21930. HAWKESWORTH, C. J., NORRY, M. J., RODDICK, J. C. & VOLLMER, R. 1979. 143Nd/14r and 87Sr/86Sr ratios from the Azores and their significance in LIL-element enriched mantle. Nature, Lond. 280, 28 31. , ROGERS, N. W., VAN CALSTEREN, P. W. C. & MENZlES, M. A. 1984. Mantle enrichment processes. Nature, Lond. 311,331-35. HELLMAN, P. L. & GREEN, T. H. 1979. The role of sphene as an accessory phase in the high pressure partial melting of hydrous mafic compositions. Earthplanet. Sci. Lett. 42, 191-201. HOFMANN, A. W. & WHITE, W. M. 1982. Mantle plumes from ancient oceanic crust. Earth planet. Sci. Lett. 57, 421-36. HOLE, M. J., SAUNDERS,A. D., MARRINER, G. F. & TARNEY, J. 1984. Subduction of pelagic sediments: implications for the origin of Ce-anomalous basalts from the Mariana Islands. J. geol. Soc. Lond. 141, 453-72. -
-
266
B. L. Weaver et al.
HUMPHRIS, S. E. & THOMPSON,G. 1983. Geochemistry of rare earth elements in basalts from the Walvis Ridge: implications for its origin and evolution. Earth planet. Sci. Lett. 66, 223-42. IMSLAND, P., LARSEN,J. G., PRESTVIK, T. & SIGMOND, E. M. 1977. The geology and petrology of Bouvetoya, South Atlantic Ocean. Lithos, 10, 213-34. JACOBSEN, S. B. & WASSERBURG,G. J. 1979. The mean age of mantle and crustal reservoirs. J. geophys. Res. 84, 7411-27. JAFFREZlC, H., JORON, J. L. TREUIL, M. & WOOD, D. A. 1980. A study of the precision attained by neutron activation analysis using international standard rocks GS-N and BCR-I as examples. J. radioanal. Chem. 55, 417-25. KAY, R. W. 1984. Elemental abundances relevant to identification of magma sources. Phil. Trans. R. Soc. Lond., Ser. A, 310, 535-47. KURZ, M. D., JENKINS, W. J. & HART, S. R. 1982. Helium isotopic systematics of oceanic islands and mantle heterogeneity. Nature, Lond. 297, 43-7. LEMAITRE, R. W. 1962. Petrology of volcanic rocks, Gough Island, South Atlantic. Geol. Soc. Am. Bull. 73, 1309-40. LEROEX, A. P. 1985. Geochemistry, mineralogy and magmatic evolution of the basaltic and trachytic lavas from Gough Island, South Atlantic. J. Petrol. 26, 149-86. & ERLANK, A. J. 1982. Quantitative evaluation of fractional crystallisation in Bouvet Island lavas. J. Volcanol. geotherm. Res. 13, 309 38. --, DICK, H. J. B., ERLANK, A. J., REID, A. M., FREY, F. A. & HART, S. R. 1983. Geochemistry, mineralogy and petrogenesis of lavas erupted along the Southwest Indian Ridge between the Bouvet triple junction and 11 degrees East. J. Petrol. 24, 267-318. LI, Y.-H. 1982. A brief discussion of the mean oceanic residence time of elements. Geochim. cosmochim. Acta, 46, 2671-75. MCKENZIE, D. & O'NIONS, R. K. 1983. Mantle reservoirs and ocean island basalts. Nature, Lond. 301,229-31. MENZIES, M. A. 1983. Mantle ultramafic xenoliths in alkaline magmas: evidence for mantle heterogeneity modified by magmatic activity. In . HAWKESWORTH, C. J. & NORRY, M. J. (eds) Continental Basalts and Mantle Xenoliths, pp. 92-110. Shiva, Orpington. MORGAN, J. 1971. Convection plumes in the lower mantle. Nature, Lond. 230, 42-3. -1972. Plate motions and deep mantle convection. Geol. Soc. Am. Mem. 132, 7-22. O'NIoNS, R. K. 1984. Isotopic abundances relevant to the identification of magma sources. Phil. Trans. R. Soc. Lond., Set. A, 310, 591-603. & PANKHURST, R. J. 1974. Petrogenetic significance of isotope and trace element variations in volcanic rocks from the Mid-Atlantic. J. Petrol. 15, 603-34. , EVENSEN, N. M. & HAMILTON, P. J. 1979. Geochemical modelling of mantle differentiation and crustal growth. J. geophys. Res. 84, 6091101. -
-
-
-
--,
HAMILTON, P. J. & EVENSEN, N. M. 1977. Variations in 143Nd/144Nd and 87Sr/S6Sr ratios in oceanic basalts. Earth planet. Sci. Lett. 34, 13--22. PEARCE, J. A. 1983. Role of sub-continental lithosphere in magma genesis at active continental margins. In: HAWKESWORTH, C. J. & NORRY, M. J. (eds) Continental Basalts and Mantle Xenoliths, pp. 230-49. Shiva, Orpington. PHILPOTTS, J. A. & SCHNETZLER, C. C. 1970. Phenocryst-matrix partition coefficients for K, Rb, Sr and Ba, with applications to anorthosite and basalt genesis. Geochim. cosmochim. Acta, 34, 307-22. RICHARDSON, S. H., ERLANK, A. J., DUNCAN, A. R. & REID, D. L. 1982. Correlated Nd, Sr and Pb isotope variation in Walvis Ridge basalts and implications for the evolution of their mantle source. Earth planet. Sci. Lett. 59, 327-42. RINGWOOD, A. E. 1982. Phase transformations and differentiation in subducted lithosphere: implications for mantle dynamics, basalt petrogenesis, and crustal evolution. J. Geol. 90, 611-43. SAUNDERS, A. D., TARNEY, J. • WEAVER, S. D. 1980. Transverse geochemical variations across the Antarctic Peninsula: implications for the genesis of calc-alkaline magmas. Earth planet. Sci. Lett. 46, 344-60. SCHILLING, J. G. 1973. Iceland and mantle plume. Geochemical evidence along Reykjanes Ridge. Nature, Lond. 242, 565-71. SUN, S. S. 1980. Lead isotopic study of young volcanic rocks from mid-ocean ridges, ocean islands and island arcs. Phil. Trans. R. Soc. Lond., Ser. A, 297, 409-45. & HANSON, G. N. 1975. Evolution of the mantle: geochemical evidence from alkali basalt. Geology, 3, 297-302. TARNEY, J., SAUNDERS,A. D., MATTEY, D. P., WOOD, D. A. & MARSH, N. G. 1981. Geochemical aspects of back-arc spreading in the Scotia Sea and western Pacific. Phil. Trans. R. Soc. Lond., Ser. A, 300, 263-85. TATSUMOTO, M. 1978. Isotopic composition of lead in oceanic basalt and its implication to mantle evolution. Earth planet. Sci. Lett. 38, 63-87. TAYLOR, S. R. & MCLENNAN, S. M. 1981. The composition and evolution of the continental crust: rare earth element evidence from sedimentary rocks. Phil. Trans. R. Soc. Lond., Ser. A, 301, 38199. THOMPSON, R. N., MORRISON, M. A., HENDRY, G. L. & PARRY, S. J. 1984. An assessment of the relative roles of crust and mantle in magma genesis: an elemental approach. Phil. Trans. R. Soc. Lond., Ser. A, 310, 549-90. VERWOERD, W. J., ERLANK, A. J. & KABLE, E. J. D. 1976. Geology and geochemistry of Bouvet Island. In: GONZ~,LES-FERRAN, O. (ed.) Andean and Antarctic Volcanology Problems, pp. 203-37. International Association of Volcanology and Chemistry of the Earth's Interior, Rome. VOLLMER, R. 1983. Earth degassing, mantle metasomatism, and isotopic evolution of the mantle. Geology, 11,452-4. -
-
Geochemistry of ocean island basalts WEAVER, B. L. & TARNEY,J. 1984. Empirical approach to estimating the composition of the continental crust. Nature, Lond. 310, 575-7. , • SAUNDERS,A. D. 1985. Geochemistry and mineralogy of basalts recovered from the Central North Atlantic. Initial Reports of the Deep Sea Drilling Project 82, pp. 395-419. U.S. Government Printing Office, Washington, DC. WHITE, W. M. 1985. Sources of oceanic basalts: radiogenic isotope evidence. Geology, 13, 115-18. & HOFMANN, A. W. 1982. Sr and Nd isotope geochemistry of oceanic basalts and mantle evolution. Nature, Lond. 296, 821-5. WILSON, M. & DAVIDSON,J. P. 1984. The relative roles of crust and upper mantle in the generation of oceanic island arc magmas. Phil. Trans. R. Soc. Lond., Ser. A, 310, 661-74. WOOD, D. A. 1979. A variably veined sub-oceanic upper mantle--genetic significance for mid-ocean -
-
267
ridge basalts from geochemical evidence. Geology, 7, 499-503. , TARNEY,J. & WEAVER, B. L. 1981. Trace element variations in Atlantic Ocean basalts and Proterozoic dykes from northwest Scotland: their bearing upon the nature and geochemical evolution of the upper mantle. Tectonophysics, 75, 91-112. WYLLIE, P. J. 1984. Constraints imposed by experimental petrology on possible and impossible magma sources and products. Phil. Trans. R. Soc. Lond., Ser. A, 310, 439-56. ZIELINSKI, R. A. & FREY, F. A. 1970. Gough Island: evaluation of a fractional crystallisation model. Contrib. Mineral, Petrol. 29, 242-54. ZINDLER, A., JAGOUTZ, E. & GOLDSrEIN, S. 1982. Nd, Sr and Pb isotopic systematics in a three-component mantle: a new perspective. Nature, Lond. 298, 519-23.
BARRY L. WEAVER, School of Geology and Geophysics, University of Oklahoma, Norman, OK 73019, U.S.A. DAVID A. WOOD, Amoco Europe and West Africa Inc., Amoco House, 1 Stephen Street, Tottenham Court Road, London W 1P 2AU, UK JOHN TARNEY, Department of Geology, University of Leicester, Leicester LE1 7RH, U.K. JEAN LOUISJORON, Group des Sciences de la Terre, Laboratoire Pierre Sue, C.N.R.S., Centre d'Etudes Nucl6aires de Saclay, BP 2, 91190 Gif-sur-Yvette, France.
M a g m a and fluid evolution in the lavas and associated granite xenoliths of Ascension Island Chris Harris and Simon M.F. Sheppard S U M M A R Y : The mildly alkaline lavas of Ascension Island evolved in a H20undersaturated environment. H20 comprises less than 0.4 wt.% of glassy equivalents of the most evolved lavas (comendites) for which there is no evidence that it has been lost by degassing. In contrast, peralkaline granite xenoliths which are chemically similar to the evolved lavas contain silicate glass and fluid inclusions which demonstrate that the original granite body crystallized from a water-saturated magma. Microprobe analysis of the glass inclusions gives low totals suggesting that the H20 content was high (5-8 wt.%). Amphibole 6D values for the granite xenoliths are significantly different from the whole-rock 6D values of the comenditic obsidians while whole-rock 8180 values are similar. It is suggested that the high water content in the original granite magma is due to sea-water addition and that this explains these differences. Sea-water interaction with high-level magma chambers on Ascension may have been the cause of the extensive pyroclastic volcanism which occurred on the island.
Introduction Ascension Island is one of the closest exposed volcanoes to the mid-Atlantic Ridge in the S Atlantic Ocean. It is situated on oceanic crust of age 7 Ma (van Andel et al. 1978), but the oldest dated rocks on the island are 1.5 Ma in age (Harris et al. 1982). There appears to be no connection between the island's location and any special tectonic features and there is no physical evidence that the island is part of a hot-spot track (van Andel et al. 1973). The geology of Ascension has been described in detail by Daly (1925) and Atkins et al. (1964), while the petrology and isotope geochemistry (Sr, Pb, O and H) w a s discussed by Harris (1983), Harris et al. (1982a, b, 1983) and Sheppard & Harris (1985). About half of the island's exposed surface is covered by lava, and the remainder is covered by a variety of pyroclastic deposits. Associated predominantly with the pyroclastics is a suite of plutonic xenoliths including granites. The chemical and fluid evolution preserved in both lavas and plutonic xenoliths are compared in this brief review. Particular emphasis is placed on the most evolved compositions.
Composition and evolution of the lavas and plutonic xenoliths The lavas range from mildly alkaline basalts to peralkaline rhyolites (comendites) with up to 1.5% normative sodium metasilicate (Harris 1983, and unpublished data). Major- and traceelement variations suggest that the comendites
were derived from the alkali basalts by extensive crystal fractionation (about 80% (Harris 1983)). Taken as a whole, the minerals of the plutonic xenoliths show a similar compositional range to the phenocrysts in the lavas (Harris 1983). The less evolved plutonic rocks have textural features which suggest that they are cumulates. The more evolved plutonic rocks (monzonites, quartz syenites and peralkaline granites) are not obviously cumulates and are closely similar in whole-rock chemistry to the trachytes and comendites. They would appear to represent slowly cooled equivalents of magmas similar to the evolved lavas which did not reach the surface. The most evolved plutonic xenoliths are granites which appear to have been derived from a single plutonic body. They are remarkable in that they contain the rare minerals dalyite (K2ZrSi6Ols) and vlasovite (Na2ZrSi4Oll), and are among the few granites described to contain eudialyte (van Tassel 1952; Fleet & Cann 1967; Harris et al. 1982a). Dalyite and vlasovite are indicative of the peralkalinity of these granite xenoliths.
Fluid evolution in the lavas and granite xenoliths There is no evidence, such as amphibole phenocrysts, to suggest that the basalt to comendite lavas were water rich. The water content of the comenditic obsidians from Ascension is less than 0.4 wt.% (Sheppard & Harris, 1985). These obsidians are wholly glassy, non-vesicular and have relatively high C1 and F contents (Table 1) which are consistent with them not having lost signifi-
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 269-272.
269
C. Harris & S. M. F. Sheppard
z7o
TABLE 1. Selected major- and trace-element analyses
SiOz TiO2 A1203 FeO* MnO MgO CaO Na20 K20 LOI(wt.%) H20 + C1 (ppm) F Nb
1
2
74.39 0.17 10.68 4.38 0.03 0.03 0.20 5.52 4.60
73.36 0.27 12.26 3.42 0.14 0.18 0.26 5.47 4.62
------
0.76 0.36 2500 2100 195
3
4
73.20 75.20 0.30 0.16 12.40 11.04 3.46 3.92 0 . 1 3 0.13 0 . 1 6 0.00 0 . 2 0 0.00 5 . 5 2 4.91 4.62 4.64
5
74.46 0.25 10.98 4.32 0.12 0.13 0.00 5.26 4.48
0.62 0.12 0.56 0.31 0.18 2800 167 212 2100 1700 9OO 199 233 241
FeO*, total iron as FeO; LOI, loss on ignition. Nb by X-ray fluorescence (Oxford); C1 and F by wet chemical methods (CRPG, Nancy). 1, silicate melt inclusion in granite H30(1) recalculated to 100% (original total, 91.22 wt. %); 2, OB6 (obsidian from Middleton's Ridge, Ascension); 3, OB7 (obsidian from Middleton's Ridge, Ascension); 4, H30(1) (granite from Five Mile Post, Ascension); 5, H30(2) (granite from Five Mile Post, Ascension). Analysis 1 by microprobe, University of Nancy (Harris, 1986). Analyses 2-5 by inductively-coupled plasma emission spectrometry, CRPG, Nancy (Harris, 1986, and unpublished data); major-element oxides recalculated to 100%. H20 + by manometric measurement of H2 gas extracted during hydrogen isotope analysis. cant H 2 0 by degassing. Ascension lavas as a whole are therefore typical of oversaturated peralkaline series, as they are very water undersaturated (Bailey & Cooper 1978). Fluid inclusions in Ascension granite xenoliths were first described by Roedder & Coombs (1967) and fluid inclusions in the granites and other plutonic xenoliths from Ascension have been further studied by Harris (1986). The most important findings are as follows. 1 The granites contain inclusions of silicate glass and highly saline aqueous fluid, which in some cases occur within the same cavity. This shows that magma and saline fluid coexisted during crystallization (Roedder & Coombs 1967). 2 Low-temperature (below 300~ fluid circulation was minimal. 3 Vapour-rich inclusions contain CO2 and H20 and represent vapour which coexisted with the saline solutions (Roedder & Coombs 1967). 4 The silicate glass inclusions have small contraction bubbles which homogenize between 715 and 790~ (mean of seven runs, 763~ Micro-
probe analyses of the glass have totals which are consistently between 90 and 92 wt.%. Allowing for possible alkali loss under the probe beam it is suggested that H20 is high (probably between 5 and 8 wt.% (Harris, 1986)). A value of 5 wt.% H 2 0 would correspond to a trapping pressure of about 1.5 kb (Dingwell et al. 1984). 5 There is no strong correlation between the salinity of the high-density saline inclusions (liquid, halite cube and vapour bubble) and the temperature of homogenization of the inclusion. However, those inclusions with the lowest salinities (30-50 wt.% NaC1) are not homogenized at 570~ (the upper temperature limit of the stage). The fluid inclusion suite was interpreted by Roedder & Coombs (1967) to be the result of heterogeneous trapping of saline liquid and aqueous vapour during fluid immiscibility. It is suggested by Harris (1986) that fluid immiscibility and loss of the vapour produced increased the salinity in the remaining saline fluid. Thus the lowest-salinity inclusions are interpreted as fluids which were parental to the higher-salinity fluid inclusions.
Discussion The evidence summarized in the previous section suggests that the evolved lavas were always dry but the original granite body crystallized from a water-saturated magma. There is no textural evidence that the high H 2 0 and its eventual saturation occurred only at the end stages of crystallization of the granite. If this were the case the glass inclusions should show a range of H20 contents. If the original melt from which the granite crystallized contained about 5 wt.% HzO and a similar CI content to the comendite obsidians, the salinity of the fluid produced by exsolution of all the dissolved H 2 0 would be low (less than 10 wt.% NaC1) even if all the dissolved C1 were partitioned into the fluid. The high salinity of fluid inclusions suggests that considerable loss of H 2 0 from the system as vapour must have occurred. Two mechanisms by which the high water content in the original granite magma could have been generated are considered below. 1 High degrees of crystallization of comendite magma led to eventual water saturation. This would require greater than 90% fractionation of comendite magma with initial H 2 0 = 0 . 4 w t . ~ and is not consistent with the small differences in major- and trace-element chemistry between the comendite obsidians and the granite xenoliths (Table 1).
Magma andfluid evolution 2 Water was added from some outside source; in this case sea-water or meteoric water are the most obvious sources. The latter would not seem likely owing to the low rainfall and high evaporation rates on Ascension.
271
constituents and there is no evidence that fO2 increased during crystallization of the granite. This should have been the case if significant H 2 loss occurred. Direct addition of about 5 wt.% H20 to comendite magma would decrease its 61so slightly and this may be why the granites show a slightly greater range in 61sO than the comendite obsidians (Sheppard & Harris, 1985).
Stable isotope constraints on fluid evolution These data are discussed by Sheppard & Harris (1985). The granite whole-rock 81 SO values range from 6.4%o to 6.8%0 and are comparable with those of the obsidians (6.7%o_+0.1; n=3). The range in the lava suite as a whole is from +6.0%o to +6.8%o. Separated amphiboles from the granites have 6D values between - 53%o and -58%o which is rather too high for primary magmatic amphiboles and some 20%o higher than the comendite obsidian whole-rock 6D (-76%o to -83%o). Accepted primary magmatic 6D values for amphiboles and basalts are between -70%o and -85%o (e.g. Boettcher & O'Neil 1980). Factors affecting the final 6D value of these amphiboles are (i) the original 6D of the magma, (ii) the hydrogen isotope fractionation factor between amphibole and fluid, (iii) the amount of fluid lost during degassing and (iv) the speciation of hydrogen in the fluid (e.g. H2, H20, CH4 etc.). Taylor et al. (1983) showed that during degassing of H20 from silicic magmas, the residual melt is depleted in deuterium. This depletion is most pronounced when the fraction of H20 remaining is small (el Rayleigh fractionation). The fraction remaining in the Ascension granites is small (for granite H30(1) H20 + = 0.18 wt.% (Table 1)) and, given the combined effect of an amphibole-fluid fractionation factor (1000 In 0d,mph..2o) of about -20%o (Graham et al. 1984) and open-system degassing, the original fluid 6D was estimated by Sheppard & Harris (1985) to have been close to zero. This is consistent with the fluid's having been derived directly from sea-water. Water may have been derived from dehydration of the basalts, hydrothermally altered by seawater, which presumably make up part of the volcanic cone, but this is not considered to be the principal source of the water (Sheppard & Harris, 1985). Differential loss of H 2 relative to H20 could explain the difference between the 6D of amphibole in the Ascension granite and the accepted primary magmatic values. However, under the prevailing fO2 conditions of the crystallizing magma, on the HM side of the FMQ buffer (Harris 1983), both species are trace
Conclusions and speculations Fluid inclusion data show that granite xenoliths from Ascension Island, unlike the co-genetic comendite lavas, crystallized from a watersaturated magma. Hydrogen isotope data suggest that sea-water was the probable source of the fluid rather than the concentration of primary magmatic water after high degrees of crystallization. Contamination of the granite magmas by sea-water could explain the anomalously high STSr/86Sr ratios reported by Harris et al. (1982a, 1983) for these rocks. Mechanisms by which sea-water addition took place are not well constrained by the data. Any magma chamber in which fractionation took place was probably sited within the volcanic cone and water may have been derived from hydrated wall-rock basalt. An alternative mechanism is direct addition via a caldera. In both these cases (especially the latter) the tendency would have been to produce explosive pyroclastic eruptions because, unlike dry acid magmas, the d P / d T of wet acid peralkaline magmas is negative (Bailey & Cooper 1978). The relatively high proportion of pyroclastics to lava on Ascension (47% by area (Atkins et al. 1964)) may be the result of seawater addition. It is suggested that two types of eruption occurred on Ascension. One was dry and dominated by lava, and the other was pyroclastic dominated and triggered by sea-water interaction. The granite xenoliths of Ascension are the slowly cooled equivalents of the latter. Field and laboratory study of the pyroclastic deposits of Ascension, which have so far been neglected, are needed to substantiate this model. ACKNOWLEDGMENTS: C. Harris is grateful to the
Natural Environment Research Council and the Royal Society for financial support. Many colleagues in Oxford and Nancy provided technical help and advice. We are particularly grateful to Brian Atkins, David Bell and Pierre Coget for their help and encouragement. Tony Ewart, J. R. Cann, E. Roedder, Stuart Smith and Andy Duncan provided valuable comments on earlier versions of the paper.
272
C. Harris & S. M. F. Sheppard
References ATKINS, F. B., BAKER, P. E., BELL, J. D. & SMITH, D. G. W. 1964. Oxford expedition to Ascension Island, 1964. Nature, Lond. 204, 721-4. BAILEY, D. K. & COOPER, J. P. 1978. Comparison of the crystallization of pantelleritic obsidian under hydrous and anhydrous conditions. Progress in Experimental Petrology, NERC Progress Report Dll, pp. 220-33. BOETTCHER, A. L. & O'NEIL J. R. 1980. Stable isotope, chemical and petrographic studies of high-pressure amphiboles and micas: evidence for mantle metasomatism in the mantle source region of alkali basalts. Am. J. Sci. 280A, 594-621. DALY, R. A. 1925. The geology of Ascension Island. Am. Acad. Arts Sci. Proc. 60, 1-80. DINGWELL, D. B., HARRIS, D. M. & SCARFE, C. M. 1984. The solubility of H20 in melts in the system SiO2-A1203-Na20-K20 at 1 to 2 kbars. J. Geol. 92, 387 95. FLEET, S. G. & CANN, J. R. 1967. Vlasovite: a second occurrence and a trictinic to monoclinic inversion. Mineral. Mag. 36, 233-41. GRAHAM, C. M., HARMON, R. S. & SHEPPARD, S. M. F. 1984. Experimental hydrogen isotope studies: hydrogen isotope exchange between amphibole and water. Am. Mineral. 69, 128-38. HARRIS, C. 1983. The petrology of lavas and associated plutonic inclusions of Ascension Island. J. Petrol. 24, 424-70. 91986. A quantitative study ofmagmatic inclusions in the plutonic ejecta of Ascension Island. J. Petrol., 27, 251-76.
---,
BELL, J. D. & ATKINS, F. B. 1982a. Isotopic composition of lead and strontium in lavas and coarse-grained blocks of Ascension Island. Earth planet Sci. Lett. 60, 79-85. -- & --. 1983. Isotopic composition of lead and strontium in lavas and coarse-grained blocks of Ascension Island--an addendum. Earth planet Sci. Lett. 63, 139-41. , CRESSEY, G., BELL, J. D., ATKINS, F. B. & BESWETHERICK, S. 1982b. An occurrence of rareearth rich eudialyte from Ascension Island. Mineral. Mag. 46, 421-5. ROEDDER, E. & COOMBS, D. S. 1967. Immiscibility in granitic melts indicated by fluid inclusions in ejected granite blocks of Ascension Island. J. Petrol. 8, 417-5 I. SHEPPARD, S. M. F. & HARRIS, C. 1985. Hydrogen and oxygen isotope composition of Ascension Island lavas and granites : variation with crystal fractionation and interaction with sea-water. Contrib. Mineral. Petrol., 91, 74-8 I. TAYLOR, B. E., EICHELBERGER, J. C. & WESTRICH, H. R. 1983. Hydrogen isotopic evidence of rhyolite magma degassing during shallow intrusion and eruption. Nature, Lond. 306, 541-5. VAN ANDEL, T.-J., REA, D. K., VON HERZEN, R. P. & HOSKINS, H. 1973. Ascension Fracture Zone, Ascension Island and the mid-Atlantic Ridge. Geol. Soc. Am. Bull. 84, 1527-46. VAN TASSEL, R. 1952. Daylite: a new potassium zirconium silicate from Ascension Island 9Mineral. Mag. 29, 850-7.
CHRIS HARRIS*, Department of Earth Sciences, Parks Road, Oxford OX1 3PR, U.K., and Centre de Recherches Petrographiques et Geochimiques, BP20, 54501 Vandoeuvre-lesNancy C~dex, France. SIMON M. F. SHEPPARD, Centre de Recherches Petrographiques et Geochimiques, BP20, 54501 Vandoeuvres-les-Nancy C6dex, France. * Present address: Department of Geochemistry, University of Cape Town, Rondebosch 7700, South Africa 9
The Cameroon line, West Africa: a comparison between oceanic and continental alkaline volcanism J. G. Fitton S U M M A RY: The Cameroon line is the only known intra-plate alkaline volcanic province which straddles a continental margin. It consists of a chain of Tertiary to Recent volcanoes stretching from the Atlantic island of Pagalu to the interior of the African continent. It therefore provides a unique area in which to study the differences, if any, between the suboceanic and sub-continental mantle sources for alkali basalt. Although the Cameroon line does not have a graben structure, its origin is closely linked to that of the nearby Benue rift. Geochemical and isotopic data show no significant differences between basaltic rocks in the continental and oceanic sectors. However, the more evolved rocks in the two sectors are quite distinct. The continental magmas evolve towards peralkaline rhyolite, whereas those in the oceanic sector evolve towards phonolite. Progressive crustal contamination of the continental magmas accompanied by crystal fractionation is required to explain this distinction. The striking geochemical similarity between basaltic rocks in the two sectors implies that their parental magmas had very similar mantle sources. This source must lie in the asthenosphere. The old lithosphere mantle beneath Africa will be chemically and isotopically different from the young Atlantic Ocean lithosphere mantle and its involvement would be readily detected in the geochemistry of the basalts. Recent models of intra-plate alkaline magma genesis have stressed the importance of metasomatic enrichment of the mantle in large-ion lithophile elements (LILE) as a precursor to magmatism. Evidence for mantle metasomatism is provided by lithosphere-derived mantle xenoliths. Since the lithosphere is the only place where large domains of enriched mantle are likely to be preserved for long periods, it follows that the Cameroon line magmas could not have had a metasomatically enriched lithosphere source. The Cameroon line alkali basalts are chemically and isotopically similar to most other intra-plate (e.g. ocean island and continental rift) basalts which may also, therefore, have an asthenosphere rather than a metasomatized lithosphere source. The asthenosphere is also the source for mid-ocean ridge basalts (MORB) and must have a bulk composition depleted in LILE. Isotopic differences between MORB and intra-plate basalts require this source to be heterogeneous. The Cameroon line and most other intra-plate alkali basalts can be generated by small-degree melting (less than 1%) of a LILE-depleted MORB mantle source containing LILE-rich streaks.
Introduction The Cameroon line (Fig. 1) is a prominent volcanic lineament, 1600 km long, which comprises both oceanic and continental volcanoes. It includes four large central volcanoes in Cameroon and extensive lava plateaux in Cameroon and northern Nigeria, and extends offshore into the islands and seamounts of the Gulf of Guinea. The volcanic rocks are dominantly alkaline, ranging from transitional basalt to nephelinite and alkali rhyolite to phonolite. Chemical analyses of representative volcanic rocks from the Cameroon line are given in Tables 1-3. General reviews of the geology and physical geography of various parts of the region have been written by G+ze (1943, 1953), Hedberg (1968), MitchellThorn6 (1970), Gouhier et al. (1974) and Dunlop (1983). The occurrence of contemporaneous oceanic and continental alkaline volcanism within a
single province makes the Cameroon line a unique area in which to compare oceanic and continental intra-plate volcanism. The results of a project designed to assess the differences between the sub-oceanic and sub-continental mantle sources for alkaline magmas, the melting processes required to produce them and the effects of crustal contamination upon their subsequent evolution are summarized in this paper.
Age of the Cameroon line The earliest igneous activity on the Cameroon line is represented by a number of small (up to about 10 km) intrusive ring complexes composed of granite and syenite with less abundant gabbro and occasional remnants oftrachyte and rhyolite. These extend along virtually the whole of the continental sector (Fig. 2) and give ages ranging
From: FIT'rON, J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 273-291.
273
J. G. Fitton
274
0
Kilometres
~'~i~l~Biu Plat eau !::;;.,;,~. Mandara
500
;::i;;~
M,s ,0
'
Ngaoundere
--ipJ
........ }enue {rough Bambouto ~1~'Manengouba Mt Cameroon
.....o e , / i
I
i%/o"
FIG. 1. Geological map of part of W Africa showing the outcrop of the Cameroon line volcanic rocks (black) and the Cretaceous sedimentary rocks of the Benue trough (stippled). The boundary between continental and oceanic crust (bold broken line) is taken from Emery & Uchupi (1984), chart XI. The Ngaound6r6 fault system (between Bambouto and the Ngaound6r6 Plateau) and oceanic transform faults are from Gazel (1956) and Sibuet & Mascle (1978) respectively.
from 66 to 30 Ma (Cantagrel et al. 1978; Lasserre 1978 ; Jacquemin et al. 1982). Seismic profiles across the oceanic sector (Grunau et al. 1975) show disturbance of sediments of inferred Upper Cretaceous age. Gorini & Bryan (1976) have suggested that uplift of the ocean floor along the Cameroon line has formed an effective barrier between the Niger Delta and the Douala Basin since the Cretaceous. It seems likely, therefore, that the earliest volcanism in the oceanic and continental sectors was essentially contemporaneous. The oldest dated extrusive rocks are trachyte lavas and rhyolitic ignimbrite flows associated with a small granophyre and microsyenite ring complex at Kirawa on the Nigeria-Cameroon border (Fig. 2). The trachytes and rhyolites give Rb-Sr whole-rock ages of 51.2 Ma and 45.5 Ma respectively (Okeke 1980; Dunlop 1983). The younger extrusive centres (Fig. 1) give K - A r ages ranging from 35 Ma (Mandara Mountains) to Recent with the oldest dated oceanic rocks (Principe) at 31 Ma (Dunlop & Fitton 1979; Dunlop 1983; Fitton & Dunlop 1985, and references cited therein), The available K - A r data on the extrusive centres are summarized in Fig. 3. Signs of geologically recent volcanic activity can be found in most of the volcanic --
I
. L - - dl
Continental sector
I
IB,u i
,
Mandara nmlii
H i
I Ngaoundere
* '
1
m 9
~
i
i
Kirawa (46) Golda Zuelva (66) ~0~
Jos Plateau
~ {jL,/~
/
i Bambouto
J . i l Manengouba
' 9
....
r_~
* oko
1 Mt. Cameroon
~o,,(3~.)- ~,0go, (43)
'*1 Etinde
A: i
"~ ~-~ "5~
,.so~
Oceanic sector
(Namboe ,.~ ~(63) . ] ~ ~ , ~_ r_- Mayo Dar,e (63) ~ |,q~'~v -
, . . ~ ( - b ~ B a n a (30) ~ Nlonako(44) ~ , ~ K Oup6 (46) If~ ~ ?
}
ld0o
Bioko -X- 9
I I
Kiiometres i l ~
.._
Principe
m
I
500
9
115~
FIG. 2. The distribution of early Tertiary intrusive complexes (&) with their ages (in Ma) given in parentheses. Data from Cantagrel et al. (1978), Lasserre (1978) and Jacquemin et al. (1982). The Benue trough and the outcrop of the late Tertiary to Quaternary extrusive rocks of the Cameroon line are shown in outline.
S&o Tome
*__
in
n
Pagalu ")(" a n n e
0
--
10
i,
20 Ma
30
40
FIG. 3. Summary of available K-Ar age data for the Cameroon line volcanic rocks. Asterisks indicate the presence of morphologically recent cinder cones. (After Fitton & Dunlop 1985.)
The Cameroon line
275
TABLE 1. Chemical analyses o f representative Cameroon line basic volcanic rocks Continental sector N31
C147 C174
C89
Cl12
Oceanic sector C65
C5
FP44
P15
P31
P18
ST72 ST96 ST19AN15
Major elements (wt.~) SiO2 A1203 Fe203(t) MgO CaO Na20 K20 TiO2 MnO P2Os LOI
49.18 43.60 45.47 46.62 43.85 13.69 14.50 15.35 16.86 13.65 10.94 12.66 11.55 12.50 12.52 9.78 7.42 5.47 5.56 9.21 7.94 9.78 9.34 8.50 10.55 3.42 3.32 5.08 3.39 3.47 1.57 1.54 2.86 1.37 1.47 2.17 3 . 9 1 2.58 3.18 3.10 0.15 0.16 0.26 0.19 0.19 0.47 0.74 0.91 0.86 1 . 0 3 0.14 1.79 0.93 1.10 0.19
50.14 47.30 14.21 17.20 12.40 10.32 7.94 4.25 7.60 8.87 3.48 5.26 1.35 2.09 2.24 2.93 0.18 0.20 0.40 0.92 0.08 - 0 . 1 2
Total
99.45 99.42 99.79100.13 99.21 100.01 99.21
44.40 46.77 47.71 37.01 43.80 10.80 12.45 14.07 10.88 14.06 13.33 12.66 11.90 14.28 12.46 10.80 9.22 8.15 12.24 8 . 3 1 12.58 9.88 8.00 13.19 10.62 2.48 2.55 3.05 3.10 3 . 6 1 1.19 0.48 1.67 1 . 5 5 1 . 4 1 3.51 2.38 2.94 4.15 3.17 0.19 0.16 0.16 0.21 0.22 0.63 0.41 0.67 1.42 1 . 0 3 0.00 3 . 1 1 1.60 1.62 0.54
49.25 44.25 41.67 17.43 15.30 10.62 9.50 12.95 13.76 4.01 5.63 15.52 7.08 9.50 9.70 3.74 4.76 2.36 2.72 2.15 0.85 2.51 3 . 8 1 3.12 0.19 0.21 0.19 0.99 1.02 0.82 2.22 0.04 0.62
99.90100.07 99.92 99.95 99.23 99.64 99.61 99.23
Trace elements (ppm) Ni Cr V Sc Cu Zn Sr Rb Zr Nb Ba Th La Ce Nd Y
282 376 295 24 44 96 632 35 239 55 527 4 37 77 33 27
88 129 503 19 43 90 1110 34 250 62 485 0 45 87 44 25
54 116 185 15 35 122 1266 74 476 114 1000 12 109 180 65 34
29 148 46 459 182 232 19 23 22 43 98 96 1 0 0 5 1242 22 32 254 190 46 56 517 874 2 1 46 43 94 89 48 46 28 26
197 275 193 23 57 122 429 31 299 47 458 4 41 64 33 29
18 25 177 10 30 103 1202 51 461 118 650 4 97 199 81 38
8.07 29.70 19.36 . -12.95 12.51 7.12 . 5.05 4.29 0.95
12.52 23.30 17.37 . 11.82 17.14 -5.77 . 4.22 5.64 2.21
215 521 294 30 72 107 823 33 336 65 562 7 58 121 59 32
236 351 215 24 60 113 444 13 131 25 174 2 16 40 22 23
167 231 231 21 42 112 625 38 327 49 439 5 43 88 41 29
193 338 307 26 56 108 1297 40 337 118 987 10 103 203 85 35
171 381 237 18 48 114 1130 47 366 92 718 17 79 157 66 33
7 5 124 6 9 128 1056 62 527 105 854 12 89 180 76 37
43 33 234 16 41 133 1113 60 472 99 728 11 83 163 72 36
569 588 268 25 43 118 779 37 253 57 516 6 58 90 43 30
10.13 -26.43 -20.31 11.53 7.40 -14.64 12.77 24.95 8.66 -9.42 19.33 4.63 4.92 5.94 5.73 8.12 1.62 3.46
8.52 15.05 18.39 . 8.76 22.92 -12.58 . 5.14 6.15 2.49
16.59 32.66 23.51 . . -4.94 4.40 6.59 . . 3.96 4.92 2.43
12.86 14.33 14.21 . 14.35 21.67 -7.51 . 5.30 7.33 2.44
5.12 10.74 16.27
CIP W norms (Fez03/(FeO + Fe203) = 0.3) or
ab an
lc ne
di
hy ol cs mt
il ap
9.41 29.38 17.62 . -15.38 2.45 15.97 . 4.48 4.18 1.12
9.42 18.14 20.77 . 5.91 19.50 -11.49 . 5.28 7.67 1.82
17.25 8.23 10.38 29.18 10.81 27.26 . . . 18.13 ~ 24.68 8 . 1 1 -4.37 6.82 9.49 . . . 4.75 5.14 4.99 6.15 2.19 2.07
8.82 14.09 17.66 . 8.56 23.07 -14.19 . 5.14 6.00 2.48
7.08 2.96 10.67 22.47 14.99 21.95 . . 5.70 -35.02 21.11 -14.83 12.85 5.65 . . 5.43 5.32 6.74 4.71 1.51 1.00
5.25 21.87 -27.00 5.68 6.07 1.99
All the analyses in Tables 1, 2 and 3 were carried out in Edinburgh using X-ray fluorescence techniques (see Fitton & Dunlop 1985). Samples were dried at 110~ before analysis. Fe203(t ) is total Fe expressed as Fe203 . LOI, loss on ignition at 1100~ N31, ol-phyric transitional basalt (5 km N N W of Tila Lake, Biu Plateau, Nigeria); C147, aphyric alkali basalt (hillside to S of Kila, Mandara Mountains, Cameroon)" C174, basanite (waterfall on River Vina, 13 km S of Ngaound6r6, Cameroon); C89, plag(-ol-augite)-phyric transitional basalt (4 km SW of Kumbo, E of Mount Oku, Cameroon); C112, ol-augite-phyric basanite (2 km WSW of Lekwu6 Lelo6, near summit of Bambouto, Cameroon); C65, plag-(-ol-augite)-phyric transitional basalt (S wall of Lac de la Femme, Eboga caldera (Manengouba), Cameroon); C5, plag-augite(-ol)-phyric tephrite (summit of Mount Cameroon) (see Fitton et al. (1983) for more analyses of Mount Cameroon lavas); FP44, ankaramite (coast W of Pico Sta Isobel, Bioko (see Piper & Richardson 1972)); P15, olivine tholeiite from blocks in hyaloclastite breccia (River Cambungo. SE side of Principe); P31, ol-augite-phyric transitional basalt (N-S dyke, Praia de Sta. Rita, N coast of Principe); P 18, olivine nephelinite (1 km SE of S. Antonio, Principe) (total includes 0.30~ F); ST72, ol-hbld-augite-plag-phyric basanite (Maria Luisa, N W side of S~o Tom6)" ST96, hbld-phyric hawaiite (NW-SE dyke, Praia Capitango, S coast of S~o Tom6); ST19, aphyric basanite (recent lava flow from Morro Corregado, N coast of S~o Tom6); A N 15, picritic basalt (N side of Pagalu) (see Piper & Richardson 1972).
276
J. G. Fitton
TABLE 2. Chemical analyses o f representative Cameroon line intermediate and evolved volcanic rocks Continental sector C144 C155
C85
Oceanic sector
C87 C104 Cl16
C63
Pll
P12
P16
ST10 ST57 ST73 ST84STl10
Major elements (wt.~) SiO2 A1203 Fe203(t) MgO CaO Na20 K20 TiO2 MnO P205 LOI
74.08 55.75 72.70 71.13 68.56 58.19 53.99 11.40 22.05 10.74 12.56 13.08 19.50 17.03 3.49 2.29 5.69 3 . 3 1 5.74 4.39 8.95 0.17 0.09 0.04 0.04 0.05 0.80 3.30 0.09 1.08 0.12 0.02 0.44 2.88 5.39 5.26 10.93 4.72 5.68 5.87 7.77 5.29 4.05 5.54 4.57 5.26 4.81 4.26 3.02 0.12 0.35 0.37 0.68 0.35 0.91 1.98 0.10 0.19 0.09 0.03 0.13 0.21 0.17 0.01 0.08 0.00 0.09 0.01 0.20 0.65 0.46 1.09 0.22 0.66 0.60 0.84 0.22
54.90 53.98 56.28 57.39 66.69 48.54 61.66 53.64 22.32 19.54 21.47 21.70 18.05 17.42 18.63 18.83 3.13 4.80 2.38 2.20 2.16 9.93 3 . 4 1 6.41 0.30 1.42 0.07 0 . 1 1 0.08 3.51 0.16 1.76 2.00 5.11 2.00 1.24 0.58 8.16 0.94 4.96 9.53 5.39 8.62 9.43 6.29 5.38 7.93 7.28 6.02 4.41 4.86 6.02 4.37 2.45 5.46 3.65 0.48 1.08 0.16 0.24 0.30 2.65 0.26 1.63 0.12 0.13 0.18 0.15 0.11 0.22 0.24 0.16 0.09 0.24 0.02 0.04 0.07 0.93 0.03 0.44 0.85 4.08 3.75 1.19 1.15 0.34 0.99 0.55
Total
99.24 99.43 99.26 99.46 99.64 99.93 99.98
99.75 100.17 99.80 99.72 99.86 99.53 99.70 99.31
Trace elements (ppm) Ni Cr V Sc Cu Zn Sr Rb Zr Nb Ba Th La Ce Nd Y
1 3 7 0 0 307 4 402 1938 576 9 81 41 153 29 78
2 1 1 3 5 4 5 3 14 0 0 0 0 0 2 0 0 0 0 0 131 141 261 300 466 4 15 6 212 172 207 133 1 2 9 1 1 4 0 5 2 0 7 1 1372 167 255 379 251 406 4 80 31 31 38 39 30 60 20 325 249 105 18 659 398 27 19 267 188 22 26 165 112
2 3 2 0 0 75 1549 101 558 152 2501 6 116 183 73 49
24 22 91 9 14 106 856 87 643 107 701 13 76 148 58 32
3 5 45 0 0 79 959 133 431 79 1087 13 65 93 23 15
11 14 56 2 4 87 1115 180 719 119 758 26 68 111 36 20
. . 36.08 21.47 0.34 32.65 . . 6.28 0.78 . -1.28 0.92 0.21
. . 27.19 35.23 16.79 6.73 . . 6.60 . . 2.69 2.02 2.13 0.60
2 4 5 0 0 185 11 314 1745 136 11 76 98 110 17 19
2 4 15 0 0 114 110 223 965 82 154 34 76 95 18 13
2 5 2 0 0 79 231 124 515 141 1361 20 148 254 78 38
29.97 36.19 37.73 29.02 5.74 -20.78 27.26 1.58 . . 3.98 5.26 . . . . 0.43 0.02 1.00 0.11 0.31 0.46 0.04 0.10
11.58 2.12 26.20 54.00 2.45 --. -. 2.03 -0.88 0.58 0.17
19 23 173 9 21 120 1281 64 599 133 870 18 105 197 79 38
1 4 0 0 0 174 19 200 1414 264 70 39 199 338 88 43
6 5 79 4 7 117 1128 114 751 116 845 17 79 148 52 27
s
C IP W norms ( Fe 20 3/ ( FeO + Fe 20 3) = 0.3)
a
C or
ab on
ne ac ns
di WO
hy ol mt
il ap
29.92 . 24.29 36.62 . -2.85 1.24 0.35 . 4.46 -. 0.23 0.03
-29.11 . . 33.33 27.40 25.76 30.18 . . . 31.61 -1 . 8 8 4.64 1.86 1.17 4.31 0.51 . . . -6.26 0.38 -. . . 0.68 0.71 0.20 0.01
23.18 .
16.18 . . . . . . 31.57 28.79 25.47 18.02 35.83 41.10 45.74 43.21 . 5.83 13.90 --11.27 1.05 2.70 4.69 . . . 2.29 0.91 . . . -1.92 6.13 7.18 . . . 3.04 5.72 . . . --1.56 7.65 . 1.79 3.64 1 . 3 1 0.68 1.74 3.80 0.21 0.01 0.48 1.55 .
. .
------14.72 32.76 21.97 26.48 50.12 34.35 16.40 -8.03 10.66 8.86 15.33 -1.47 -. . 15.08 3.98 11.57 . . ---5.24 1.59 1.91 4.07 0.66 2.63 5.10 0.51 3.14 2.25 0.07 1.07
LOI, loss on ignition. C144, rhyolite (Aiguille Mchirgui, 2 km N of Mogode, Mandara Mountains, Cameroon); C155, phonolite (plug near Beka, 20 km N of Ngaound6r6, Cameroon); C85, rhyolite (flow) (Sabga Pass, 16 km E N E of Bamenda; SW of Mount Oku, Cameroon); C87, rhyolitic welded tuff (6 km SW of Jakiri; SE of Mount Oku, Cameroon); C104, trachyte (Pinyin, N slopes of Bambouto, Cameroon); C 116, trachyphonolite (Bambouto summit, Cameroon); C63, benmoreite (NW rim of Eboga caldera (Manengouba), Cameroon); P11, phonolite (large block in River Fria, SE side of Principe); P12, tristanite ( a s Dais Irmfios (plug), SE side of Principe); P16, trachyphonolite (large block in River Cambungo, SE side of Principe); ST10, phonolite (C~.o Grande (plug), S side of S~o Tom~); ST57, quartz trachyte (NE peak of Ilh6u das Cabras, off N E coast of S~,o Tom~); ST73, hbld-plagphyric mugearite (Maria Luisa, N W side of Sao Tom~); ST84, trachyte (S side of Mizambu (plug), E side of S~o Tomb); ST 110, phonolitic tephrite (River Manuel Jorge, 0.5 km N W of Sta. Luisa, E side of S~o Tomb).
The Cameroon line TABLE 3. Chemical analyses o f representative nephelinite lavas f r o m Etinde, Cameroon C150
C24
Major elements (wt.%) SiO2 39.60 40.04 A1203 12.75 14.56 Fe203(t) 13.91 12.34 MgO 8.03 5.28 CaO 14.98 14.00 Na20 2.68 3.47 K20 1.66 2 . 6 1 TiO2 4.49 4.06 MnO 0.20 0.30 P205 0.67 1 . 1 8 SrO 0.12 0.28 BaO 0.09 0.12 SO3 0.00 0.24 C1 0.04 0.10 LOI 1.00 1 . 3 1
Total
C152 C131 C154 39.99 42.65 46.25 17.54 17.39 19.46 9.69 9.94 6.64 3.74 2.70 1.16 10.87 8.79 6.15 6.94 6.41 7.80 3.87 4.48 6.22 2.95 2.54 1.06 0.28 0.38 0.36 0.89 0.68 0.16 0.28 0.33 0.78 0.11 0.14 0.38 1.94 1.41 0.27 0.37 0.49 0.47 0.79 1.95 3.06
100.21 99.87 100.17 100.18 100.11
Tracee&ments(ppm) Ni 46 12 10 6 2 Cr 126 8 5 6 3 V 456 448 342 416 241 Sc 45 15 4 2 0 Cu 113 64 84 38 5 Zn 93 127 134 194 209 Sr 1054 2401 2395 2825 6553 Rb 89 87 114 349 201 Zr 399 656 626 888 914 Nb 96 223 281 320 295 Ba 788 1067 1006 1254 3424 Th 8 25 21 15 4 La 97 238 225 170 131 Ce 195 489 425 272 168 Nd 84 180 142 86 41 Y 28 47 44 52 37 CIP W norms (Fee03/(FeO + Fe 20 3) = 0.3) or ---18.90 23.20 an 18.36 18.12 13.68 12.61 2.60 lc 7.84 12.40 18.20 6.51 11.83 ne 12.36 15.02 23.79 23.03 34.29 hl 0.07 0.17 0.62 0.83 0.81 th -0.44 3.49 2.57 0.50 di 35.37 36.17 25.23 23.75 16.90 wo . . . . 4.59 ol 7.13 1 . 3 8 1.64 1.05 - cs 2.85 0 . 4 1 1.55 - -mt 5.72 5 . 1 1 3.97 4.12 2.79 il 8.69 7 . 9 1 5.69 4.96 2.10 ap 1.62 2.87 2.14 1 . 6 6 0.39
LOI, loss on ignition; C 150, olivine melanephelinite; C24, nephelinite; C152 and C131, haiiyne nephelinite; C154, leucocratic nosean leucite nephelinite. All the samples were collected from loose blocks in stream beds : C24 from Bonenza, and the others from Batoke.
277
centres (Fig. 3). M o u n t C a m e r o o n is still active and last erupted in 1982 (D&uelle et al. 1983; Fitton et al. 1983). It is clear that volcanic activity on both the oceanic and continental sectors of the C a m e r o o n line has been more or less continuous since the end of the Cretaceous.
Geology and petrology Continental sector Intrusive complexes
Relatively little is k n o w n about the m i n e r a l o g y and geochemistry of these complexes. T h a t w h i c h is k n o w n suggests that they are broadly similar to the ring complexes of the Jos Plateau in N i g e r i a (Bowden et al. 1987). Some of the granite bodies in the M a y o Darl~ area contain traces of cassiterite ( C h a p u t et al. 1954; Gazel et al. 1963). Only the M b o u t o u complex (Fig. 2) has been described in detail. Parsons et al. (1986) have published an account of the m i n e r a l o g y and crystallization history of this complex a n d stress the very mildly alkaline nature of its parental m a g m a . J a c q u e m i n et al. (1982) have carried out a g e o c h e m i c a l and isotopic study of the G o l d a Z u e l v a and M b o u t o u complexes and conclude that the m a g m a s , although initially m a n t l e derived, have interacted with crustal rocks as they crystallized. T h e intrusive complexes were not included in the present study because they are only exposed in the continental sector. F u r t h e r m o r e , the effects of crystal a c c u m u l a t i o n and late-stage alteration processes limit the usefulness of coarse-grained rocks in g e o c h e m i c a l studies. T h e following discussion will therefore be confined entirely to the younger extrusive centres shown on Fig. 1. Biu Plateau
The small t o w n of Biu in N E N i g e r i a is built on a plateau c o m p o s e d of basaltic lava flows. These reach a m a x i m u m thickness of 250 m (J. W. du Preez, cited by Carter et al. 1963) and range in composition from basanite to h y - n o r m a t i v e transitional basalt. The most recent volcanism in the area p r o d u c e d a large n u m b e r of c i n d e r cones aligned in a N N W - S S E direction. These usually have well-defined craters w h i c h have often been b r e a c h e d by small lava flows. K - A r ages on the plateau basalts range from 5.3 M a for the basal lavas to about 0.8 M a for a flow high in the succession ( G r a n t et al. 1972; Fitton & D u n l o p 1985). The lava flows and cinder cones often contain a b u n d a n t peridotite xenoliths and clinopyroxene and anorthoclase megacrysts. A large crater to
278
J. G. F i t t o n
the west of the village of Miringa (14 km N N W of Biu) is particularly noteworthy for its abundant and extensive inclusion suite. These include peridotite xenoliths up to 0.5 m in diameter and large megacrysts of clinopyroxene, garnet, anorthoclase, ilmenite and amphibole (Wright 1970; Frisch & Wright 1971). Mandara Mountains
The Mandara Mountains lie along the Cameroon-Nigeria border and are composed of granite and gneiss of Pan-African age. In the area around the small settlements of Roumsiki and Mogode the basement rocks have been intruded by numerous plugs of peralkaline trachyte and rhyolite which form spectacular spines. Occasional remnants of alkali basalt lava flows are also found in this area. K - A r dating (Fitton & Dunlop 1985) gives ages of 30 and 33 Ma for two samples of basalt and 35 Ma for one of the trachyte plugs. Ngaoundbr~ Plateau
The town of Ngaound6rb in northern Cameroon lies on the western edge of an extensive basalt plateau. The earliest lavas were erupted into two broad valleys eroded into the Precambrian basement by the headwaters of the Rivers Bini and Vina. Continued volcanism built a plateau of alkali basalt across the area and culminated in the building of a central volcano, Nganha. This volcano is composed of basanite capped by trachyte and phonolite flows which have yielded K - A r ages of 10-7 Ma (Gouhier et al, 1974). Trachyte flows also occur as small outliers over the whole plateau as do trachyte and phonolite plugs. The most recent volcanism is represented by cinder cones aligned in a W N W - E S E direction. These have sometimes produced small lava flows. All the basic lavas in the area are ne normative and range from alkali basalt to basanite. Some of the flows contain peridotite xenoliths and clinopyroxene, anorthoclase and rare zircon megacrysts.
although basic lava flows also occur higher in the succession. K - A r ages obtained from Oku lavas range from 23 to 17 Ma (Gouhier et al. 1974; Fitton & Dunlop 1985). The basic lavas range in composition from slightly ne normative to slightly hy normative. The volcanic rocks as a whole form a strongly bimodal suite. Numerous cinder cones (some yielding peridotite xenoliths) and explosion craters on and around Mount Oku provide evidence for recent volcanic activity in the area although very few of the vents produced lava flows. The craters reach 2 km in diameter and many form crater lakes. Two of these (Lake Monoun to the south and Lake Nyos to the N W of Oku) recently released large volumes of carbon dioxide with catastrophic results (SEAN 1985, 1986). The cause has not yet been established. Bambouto
Unlike Oku, Bambouto still retains some of its original volcanic form despite extensive erosion. The highest peaks (up to 2740 m) lie around a calder rim 10 km in diameter. Tchoua (1972) identified two episodes of caldera collapse. As with Oku, the lava succession comprises a strongly bimodal basalt-trachyte suite. Basalt lavas dominate the basal part of the succession but are also found on the upper slopes of the volcano. The Bambouto lavas are generally more undersaturated than the Oku suite. All the basic lava samples analysed are slightly ne normative and the salic rocks range from quartz trachyte to trachyphonolite. Rhyolites have not been found by the present author although Tchoua (1973) has reported extensive ignimbrite sheets on and around Bambouto. K - A r dating of lavas from Bambouto gives ages ranging from 23 to 14 Ma (Gouhier et al. 1974; Fitton & Dunlop 1985). Gouhier et al. also report an age of 5.8 Ma for a basalt sample collected about 20 km N of the caldera rim. There is no evidence for recent volcanic activity on Bambouto.
Oku
Manengouba
Mount Oku (3011 m) is a large deeply dissected volcanic massif composed largely of rhyolite and quartz trachyte flows, domes and plugs. Rhyolitic ignimbrite sheets are common in the area and extend as far south as the town of Bamenda (midway between Oku and Bambouto) which sits at the base of a prominent escarpment of welded tuff. The lower part of the volcanic succession is dominated by flows of basalt and hawaiite,
Manengouba is a well-preserved central volcano whose summit region is occupied by two concentric calderas. A poorly defined outer caldera (Elengoum; 6 km in diameter) encloses the younger perfectly preserved Eboga caldera (3 km in diameter). The highest point on the mountain (2411 m) lies on the Elengoum rim. This outer rim was breached on its eastern side during the building of the Eboga volcano and allowed large
T h e C a m e r o o n line
volumes of lava to flow over the north-eastern flanks of the older volcano. The oldest Manengouba lava flow so far dated gives a K - A r age of 1.55 Ma (Gouhier et al. 1974). A recent detailed K - A r and field study of the volcano (C. A. Hirst, unpublished data) has confirmed this date and established a chronology for the evolution of the volcano. The collapse of the Elengoum caldera probably occurred between 0.8 and 0.6 Ma ago. The Eboga lavas give ages ranging from 0.56 to 0.26 Ma, and the collapse of the Eboga caldera is thought to have occurred at about 0.25 Ma. An isolated trachyte spine intruding the eastern rim of the Elengoum caldera gave a K - A r age of 0.26 Ma. More recent volcanism has produced a S W - N E line of crater lakes and cinder cones across the floor of the Eboga caldera. The lavas range in composition from ne- and by-normative basalts to trachyte, quartz trachyte and rare rhyolite. Intermediate lavas are common. In striking contrast with the lavas of Oku and Bambouto, the Manengouba lavas form a complete compositional continuum. Mount Cameroon
At 4095 m, Mount Cameroon is by far the highest mountain in West Africa and one of Africa's largest volcanoes. The base of the lava pile is below sea level. Mount Cameroon is the only currently active centre of the Cameroon line and has had five recorded eruptions this century (Fitton et al. 1983, and references cited therein). It is a composite volcano built of alkali basalt to basanite lava flows interbedded with small amounts of pyroclastic material. Small cinder cones are abundant in several areas on the flanks of the volcano and on the surrounding lowlands. These cones are often aligned along fissures running SW-NE, parallel to the long axis of Mount Cameroon. The Plain of Tombel, between Mount Cameroon and Manengouba, is likewise dotted with numerous cinder cones and explosion craters, some of which are occupied by lakes. The largest of these, Lake Barombi Mbo near the town of Kumbo, is surrounded by a tuff ring containing abundant peridotite xenoliths. The Mount Cameroon lavas are all ne normative and range from almost aphyric to strongly porphyritic (olivine _+plagioclase). Lavas more evolved than hawaiite have not been reported. The few K - A r dates available (Hedberg 1968) suggest that all the exposed lavas were erupted in the last million years although there is field evidence that the volcanic edifice was built upon a basement of older lava flows. These are best exposed along the coast to the SE of Limbe (formerly Victoria) but have not yet been dated.
279
Etinde
Mount Cameroon is not a single volcano but has a large subsidiary peak, Etinde, on its SW flank. Etinde rises from the rain forest a few kilometres from the coast and is densely forested up to its summit (1713 m). Exposures of lava are rare and deeply weathered but the numerous streams contain blocks of fresh volcanic rock unlike anything found on the main volcanic massif. These provide evidence that Etinde is composed entirely of nephelinite lavas. The lavas were first described by Esch (1901). Attempts to date the Etinde lavas have not yet produced consistent results. A sample of nosean leucite nephelinite gave a K - A r age of about 0.1 Ma (Fitton & Dunlop 1985). R. M. MacIntyre (unpublished data) has obtained K - A r ages of 1.1,0.6 and 6.3 Ma from, respectively, a melanephelinite whole-rock sample and nepheline and hafiyne separated from two different nephelinite samples. It is possible that the apparent age of the hatiyne separate is anomalous and is due to excess radiogenic argon. The lavas appear to form a consanguineous series, and it would therefore be unwise to postulate an eruptive history of over 6 Ma on the basis of this one determination. It seems likely that the lavas are less than a million years old and possibly contemporaneous with some of the Mount Cameroon lavas. More work is needed to establish the relative ages of the two volcanic centres. The lavas form a compositional continuum from olivine melanephelinite through nephelinite and haiiyne-nephelinite to a leucocratic nosean leucite-nephelinite (Table 3). Feldspar is absent throughout the whole series and melilite is observed in a few samples. The observed chemical variation among the lavas can be explained by crystal fractionation dominated by aluminous clinopyroxene. This gave rise to residual liquids close in composition to the nepheline-leucitediopside cotectic in the system nepheline-sanidine-diopside (Fig. 4). These highly evolved compositions contain very high concentrations of some incompatible elements. The nosean leucite-nephelinite lavas, for example, contain around 7000 ppm Sr which is present as an unusual strontian melilite (Fitton & Hughes 1981). Oceanic sector Bioko (formerly Fernando P6o)
Despite being only 50 km from the mainland, Bioko is, geologically, the least well known of the Gulf of Guinea islands. From the time the island gained independence from Spain in 1968 until the coup of 1979, the brutal regime of President
J. G. Fitton
280
was formed in the Brunhes epoch and is thus younger than 0.7 Ma. Numerous cinder cones occur on all three volcanic centres although there have been no well-documented reports of eruptions. What may have been a small eruption on the east side of the island was reported by natives around the beginning of this century (Hedberg 1968).
DIOPSIDE
Principe
N E PHE L INE
Weight
-
SANIDINE
FIG. 4. Normative compositions of three leucocratic nosean leucite--nephelinite samples from Etinde projected into the system nepheline-sanidine-diopside (Platt & Edgar 1972). The norms were calculated using Fe_,O3/(FeO + Fe:O0 values of 0.3, 0.5 and 0.7. Chemical analysis ofC154 gives a value of 0.659 for this ratio (Fitton & Hughes 1981). Macias Nguema ensured complete inaccessibility. The current social and economic state of the island has been described by Winchester (1984) who, 4 years after the execution of the President, claimed that Bioko was 'the nastiest place on earth'. Consequently our knowledge of the geology and petrology of the island is based on preindependence field work and rock collections. The most recent geological expedition to the island was undertaken as part of a palaeomagnetic study of the Cameroon line (Piper & Richardson 1972). The cores they collected were used in the present study. The earlier literature on the geology of Bioko and the other Gulf of Guinea islands has been reviewed by MitchellThom~ (1970). Bioko is a youthful volcanic island with three eruptive centres. Pico Santa Isobel, the highest point (2972 m), forms the northern part of the island and the other two volcanoes, San Carlos and Pico Biao, form the southern part. The lavas range from basanite to hy-normative basalt. Unlike the other Gulf of Guinea islands, more evolved rocks have not been reported from Bioko. A single K-Ar age of 1.1 Ma was obtained by Hedberg (1968) from a basanite sample collected near the town of Santa Isabel on the northern coast. Nearly all the palaeomagnetic sites investigated by Piper & Richardson (1972) are normally magnetized which led these workers to conclude that virtually the whole volcanic edifice
Principe is a small and deeply eroded island surrounded by a broad submarine shelf. The prevailing SE trade winds in the Gulf of Guinea veer SW as they approach the mainland. Therefore the south-western parts of the islands receive the highest rainfall and consequently suffer the most erosion. The structure, volcanic stratigraphy and petrology of Principe have been described by Fitton & Hughes (1977), and a K-Ar chronology has been established by Dunlop & Fitton (1979) with an additional determination by Hedberg (1968). The oldest exposed rocks are hyaloclastite breccias which crop out in river valleys in the southeastern part of the island. These represent a submarine phase in the evolution of the island and contain blocks of fresh olivine tholeiite dated at 31 Ma. The oldest sub-aerial lavas (the Older Lava Series) are exposed mostly in the north and range in composition from transitional basalt to hawaiite. They give K-Ar ages ranging from 24 to 19 Ma (Hedberg 1968; Dunlop & Fitton 1979) and are intruded by numerous dykes and small intrusions compositionally similar to the lavas. The Older Lava Series is unconformably overlain by nephelinite and basanite lava flows which have yielded K-Ar ages of 5.6 and 3.5 Ma. This Younger Lava Series is overlain by phonolite lavas in the northern part of the island. The high ground to the south (up to 948 m) is composed of a chaotic assemblage of lava flows and plugs ranging in composition from tristanite to trachyte and phonolite. Four K-Ar determinations on samples collected from these plugs give ages ranging from 7 to 5.3 Ma. There are no signs of recent volcanic activity on Principe. S~o Tom~ S~.o Tom6 is a roughly conical volcanic island rising from about 3000 m below sea level to a height of 2024 m. The ground rises gently in the drier northern and eastern parts of the island but is much more rugged in the S and W. Phonolite and trachyte spines rising vertically out of the rain forest form a spectacular feature of the southern part of the island where rainfall and erosion rates are highest.
The Cameroon line The oldest rocks on the island are conglomerates, sandstones and shales of the Ubabudo Formation (Hedberg 1968). These crop out in stream sections in a small area in the eastern part of the island and are accompanied by oil seepages. They are almost non-fossiliferous and cannot be 15-o o O o00 9
Ca 9 I0~-
5
3.5 P205
o o
3"0
. 9
o o~
. a,. t o %9
dated with certainty, but Hedberg (1968) has suggested a Cretaceous age on the basis of lithological comparisons with Cretaceous sandstones in Gabon and on radiolaria and foraminifera extracted from the shales. The volcanic succession is dominated by basic
el
.~ ",:..,~... ,oO eo
9
281
2"5
9
2"0 9 OO0~
go
k
9
- , o e. e . . .
9
0,5
1"5
0
9
ab
1'0 0-5
OO O
0-0
Total Fe as
Fe2 0 3
o
o o TiO 2
9
. 0.
. ,..O.~~
"0 o
.'~11: o
o
9 .jt.,~
qP
o
9
" _%,-,,T.',l'n_o"
o o
,,g8
n
20 ~
AI2Os
o o
I~
~ .
K20
~o go
9
. ~ 6%,0o ~o
o 9
1o11o 9
~9 o
9 o o
gO 9 ~ D 9 9 9 o9 o ~ QoO o" ~ D , O 0_ - 00 9 9 9 9 9 9 ~o o ~o o4,~jp, oo 9 9 ~ o o Q w c~o 9 9 o ,8,~ 9 o o 9
"
o
oI
Ie~
9
o
Na:O
SiO2 55 I k ~ o
5~f
e 9
45-
o~
9
, 9, . .
40
e 9
o L..P-Ooo
~
5,e"e0.6"-~om o 9 I q' "lid" 9
9 9
9
o i
!
o 9 O--O
9
9...;to.:
.~..~..,. o ~n8
...fie
9
35
o
o i 2
io
1 4
9 io o 6
. I 8
.
"." ~ " " o
~ 10
o o 9 I 12
14
MgO
FIG. 5. 9 Major-element variation (wt.~) among the volcanic rocks of S~o Tom6; 9 analyses of cognate xenoliths collected from intermediate and evolved rocks.
282
J. G. Fitton
lavas, although evolved rocks are common and form part of a continuum from basalt to trachyte and phonolite with no discernible composition gaps (Fig. 5). The chemical variation can be explained by crystal fractionation involving the observed phenocryst phases. Fractionation of the more magnesian magmas (MgO > 6%) was dominated by olivine + augite. The inflection in the variation diagrams marks the onset of plagioclase, hornblende, magnetite and apatite crystallization. Cognate xenoliths composed of these four phases with or without augite are abundant in some of the intermediate and more evolved lavas (open circles in Fig. 5). The basic lavas range in composition from basanite to hy-normative basalt with no clear correlation between degree of undersaturation and age (cf Principe). It may be significant, however, that the rare quartz trachytes are the oldest dated volcanic rocks on $5,o Tom6. Grunau et al. (1975) have obtained an age of 15.7 Ma, but they do not state where the sample was collected. The quartz trachyte plug forming the Ilh~u das Cabras off the NE coast gives a K - A r age of 13 Ma (Fitton & Dunlop 1985). All other dated volcanic rocks from Sgo Tom6 are much younger with ages ranging from 7.6 to 0.1 Ma (Hedberg 1968; Fitton & Dunlop 1985). Evolved rocks were erupted over most of this period, although lavas erupted over the last million years appear to have been entirely basaltic. Recent cinder cones are very common over most of the island. Pagalu (formerly Annobon) Pagalu is the southernmost and smallest of the Gulf of Guinea islands. After independence from Spain it joined Bioko as part of Equatorial Guinea, thereby becoming virtually inaccessible. A suite of rock samples, however, was collected by Piper & Richardson (1972) and these samples were used in the present study. More recently, a visit to the island by a French research vessel resulted in accounts of the geology of the island and the petrology and geochemistry of its volcanic rocks published by Cornen & Maury (1980) and Liotard et al. (1982). The following summary of the geology of Pagalu is based largely on these studies. The oldest rocks on the island are hyaloclastite breccias which are exposed around the coast. They contain large clinopyroxene megacrysts and are intruded by numerous basaltic dykes. Neither the breccias nor the dykes have been dated. This basal unit is overlain by basic lavas ranging in composition from basanite to hy-normative basalt. The lowest flow in this succession has been dated at 18.4 Ma by Piper & Richardson (1972).
The lava pile has been intruded by tristanite (3.9 Ma) and trachyte plugs and by numerous basanite dykes, one of which gave a K - A r age of 5.4 Ma (Cornen & Maury 1980). The most recent basaltic lava flows have been erupted from a wellpreserved crater, now occupied by a lake. These flows contain abundant peridotite xenoliths. A lava sample from one flow gave a K - A r age of 2.6 Ma (Piper & Richardson 1972). The geological history and volcanic rocks of Pagalu are broadly similar to those of Principe. The main difference is that quartz trachyte is the most common salic rock type on Pagalu whereas phonolite is the most abundant on Principe. However, of the three quartz trachyte analyses presented by Cornen & Maury (1980), the two with significant amounts of quartz in their norms are also corundum normative (as is the quartz trachyte from S~.o Tom6; Table 2) which suggests that these rocks have suffered some alkali loss. It is not clear, therefore, whether the original magma was oversaturated or undersaturated. The Pagalu basalts are strikingly magnesian, a feature first noted by Tyrrell (1934). Although not unusually rich in olivine phenocrysts, they contain up to 16.5~ MgO.
Origin of the Cameroon line From the above discussion it is clear that there is no evidence of age progression of volcanic activity on the Cameroon line. Magmatism in both oceanic and continental sectors appears to have commenced in the late Cretaceous, and most of the volcanic centres have shown some activity in the course of the last million years or so. The only currently active volcano (Mount Cameroon) lies half-way along the line. The Cameroon line cannot, therefore, be a hot-spot trail as suggested by Van Houten (1983). Over the past 60 Ma, Africa has rotated anticlockwise and drifted northwards by about 12~ (Smith et al. 1981). If the Cameroon line owes its origin to a deep mantle plume then this must have had the form of a sheet and have been moving so as to keep pace exactly with the lithosphere. Magmatism on the Cameroon line was accompanied by regional uplift. The average height of the Precambrian basement beneath and around the large volcanic centres of Bambouto and Oku, for example, is about 1200 m above sea level with peaks up to 2000 m. The occurrence of inliers of Cretaceous sandstone on S~o Tom~ suggests that the oceanic basement has been similarly uplifted. This uplift is presumably an isostatic response to low-density (? partially melted) mantle beneath the region.
283
The Cameroon line Freeth (1979) has suggested that the Cameroon line is an extensional feature resulting from membrane stresses generated by the northward movement of the African plate away from the equator. This is consistent with regional uplift and alkaline magmatism but direct field evidence for extension is lacking. Freeth (1979) estimates tbat about 1 km of extension has taken place across the Cameroon line. No rift faulting or graben structures have been observed along the continental sector, however, and most of the magmatism has been associated with central rather than fissure volcanoes. Furthermore, intrusive complexes were emplaced at a time when the region should have been under compression (Freeth 1979). For part of its length the Cameroon line runs almost parallel to a major crustal shear zone, the Ngaound6r6 fault (Fig. 1), which was a continuation of the Pernambuco lineament in Brazil before continental separation. Gorini & Bryan (1976) have suggested that the Cameroon line volcanism was caused by reactivation of this ancient lineament. Evidence for this reactivation is provided by the deformation and metamorphism of Cretaceous conglomerates close to the fault zone (Vincent 1968). This explanation is weakened, however, by the lack of evidence that the lineament has exerted any influence over the volcanic activity. For example, the WNW-ESEtrending fissures in the Ngaound~r6 Plateau (marked by lines of cinder cones) cut across the fault zone at an angle of about 70 ~. The oceanic part of the Cameroon line is similarly unaffected by the transform faults it crosses (Fig. 1). An explanation for the Cameroon line may lie in its relationship with the Benue trough (Fitton 1980, 1983). The two features are so remarkably similar in shape and size (Fig. 1) that they can be superimposed exactly by rotating one with respect to the other about a pole in Sudan. Fitton (1980) suggested that this is not a fortuitous coincidence but results from a displacement of the African lithosphere relative to the asthenosphere. In this model the Y-shaped asthenosphere hot zone which would have underlain the Benue trough during its extensional phase in the Cretaceous became displaced relative to the lithosphere so that it now lies beneath Cameroon and the Gulf of Guinea. The Cameroon line and Benue trough are therefore regarded as complementary features. The former is a line of volcanoes with rift valley affinities but lacking rift faulting. The latter is a rift valley containing relatively few volcanic rocks. Magmas originally destined for the Benue rift reached the surface as the Cameroon line instead. The sequence of events postulated in this model is summarized in Fig. 6.
90 Ma
' ~ j
mointAShten~
the 70 Ma
~
~ Present
~
~
/// I
Fla. 6. Postulated sequence of events leading to the development of the Cameroon line. The block diagrams represent segments of crust and upper mantle measuring 1000 km square by 200 km deep. (From Fitton 1983.) Whatever the cause of the Cameroon line volcanism, there can be little doubt that its oceanic and continental components share a common origin. They need not, however, have similar mantle sources. The geochemistry of the Cameroon line basic volcanic rocks has been useful in constraining the composition of the mantle sources for oceanic and continental alkali basalts. These constraints will be discussed in the next section.
Mantle sources and generation of the basic magmas Intra-plate basaltic rocks, both oceanic and continental, are strikingly enriched in large-ion lithophile elements (LILE) by comparison with mid-ocean ridge basalts (MORB). The large
284
J. G. F i t t o n
degree of asthenosphere melting inferred to take place in a relatively shallow melting zone beneath mid-ocean ridges allows the composition of the MORB source to be closely constrained. This source must be depleted in LILE (e.g. K, Rb, Ba and light-rare-earth elements) in comparison with the bulk Earth. Isotope ratios of Sr and Nd show that the MORB source must have had low Rb/Sr and Nd/Sm ratios (and by implication have been depleted in other LILE) for considerable periods of time. Isotope ratios in most intra-plate basic volcanic rocks also imply a long-term LILEdepleted source, although not quite as depleted as the MORB source. This observation has posed a major problem in the understanding of intraplate volcanism. How can such LILE-rich magmas be generated from a LILE-depleted source ? One solution involves very small degrees of partial melting (less than 1~) in order to concentrate LILE (which are incompatible in mantle phases) into the melt. Such small melt fractions have been considered impossible to extract and simple melting models have been rejected as a consequence. Complex melting processes such as zone refining (Harris 1957) and wall-rock reaction (Green & Ringwood 1967) have been proposed to account for the concentration of K and other incompatible elements into a melt. More recent attempts to account for the high LILE concentrations in intra-plate volcanic rocks have invoked metasomatic enrichment of the mantle source shortly before it is melted. The current popularity of mantle metasomatism has led to a proliferation of ad hoc hypotheses involving variable degrees of enrichment through the agency of CO2- and HzO-rich fluids originating in deeper parts of the mantle. Such hypotheses have been applied to both oceanic (Clague & Frey 1982; Wright 1984) and continental (Lloyd & Bailey 1975; Frey et al. 1978; Menzies & Murthy 1980; Bailey 1982, 1987) intra-plate magmas. Di
Direct evidence for the existence of enriched mantle is provided by the occurrence of metasomatized peridotite xenoliths in alkali basalt and kimberlite. These xenoliths almost certainly originate within the continental lithosphere which is the only part of the Earth where large domains of enriched mantle are likely to be preserved for very long periods. The oceanic lithosphere is young and at least as depleted in LILE as the asthenosphere, which accounts for the scarcity of metasomatized mantle xenoliths in oceanic volcanic rocks. The asthenosphere will be well stirred by convection and therefore homogeneous on the scale of individual convection cells. Where the asthenosphere does become locally enriched by metasomatic processes (as it must do above subduction zones, for example), convection will carry the enriched material away from the site of enrichment and mix it with normal depleted asthenosphere. If metasomatically enriched lithosphere mantle plays an important role in the generation of magmas then we would expect significant geochemical differences between oceanic and continental intra-plate basic volcanic rocks. If such differences do not exist then an asthenosphere source is implied and the case for mantle metasomatism as a necessary precursor to intra-plate magmatism will be weakened. In an attempt to detect systematic differences between oceanic and continental intra-plate volcanism Fitton & Dunlop (1985) analysed a large number of volcanic rocks from the oceanic and continental sectors of the Cameroon line for major and trace elements and Sr isotope ratios. The compositional range of the oceanic and continental sector basic ( M g O > 4 ~ ) volcanic rocks is shown in Fig. 7. The highly undersaturated Etinde nephelinites are not included in this diagram. Average incompatible-element concentrations in basic rocks from the main volcanic centres are shown in Fig. 8. Despite minor Q
Di
Q
ooo3-.0
Ne
OI
Ne
OI
FIG. 7. Normative composition of Cameroon line basic volcanic rocks (MgO > 4~) (Etinde nephelinite data not included). Fe203/(FeO + Fe203) is normalized to 0.3 for the norm calculations.
The Cameroon line T
i
i
~
r
i
;
i
;
I
I000
,
"~ o
/' "-~
J"
,""
----
=
'"\'
10
......,, ,,,.,i_7~\ --o
"~
"'O/ .
. ....... . .
N
A "~'~" \... I~ -X~ ~ I \'~,'~*" ",,I','~
9 Manengouba
~ ~t,~
~ ..............
~0
Rib
I Ba
TIh
I K
I Nb
I La cle
I Sr
-~.
N~d
I P
I Zr
I Ti
'~
FIG. 8. Average incompatible-element abundances in basic volcanic rocks (MgO > 4%) from the main volcanic centres of the Cameroon line. Number of analyses: Pagalu, 16; S~o Tom6, 77; Principe, 28; Bioko, 23; Etinde, 15; Mount Cameroon, 24; Manengouba, 15 ; Bambouto and Oku, 22; Ngaound6r6, 27; Biu, 35. Concentrations have been normalized to chondrite and primitive mantle abundances (Sun 1980). The spread in Th data is due to analytical uncertainty. chemical variation between centres it is clear that there are no systematic differences between the oceanic and continental suites. The basic volcanic rocks in each of the centres could have been generated by melting very similar mantle sources. Smaller degrees of melting followed by extensive crystal fractionation can account for the higher incompatible-element concentrations in the Etinde nephelinite lavas. The remarkable similarity between the oceanic and continental basic volcanic rocks is illustrated in Fig. 9 in which the average concentrations of incompatible elements in all the samples (except the Etinde nephelinites) from the two sectors are compared. The two suites of basic rocks are also isotopically indistinguishable (Fitton & Dunlop 1985). These data imply that the oceanic and continental basalts had very similar mantle sources. Since it is most unlikely that the ancient lithosphere mantle beneath the continental sector is chemically and isotopically similar to the young lithosphere mantle beneath the oceanic sector, it follows that lithospheric mantle was not the source of the Cameroon line basalts. The lack of any consistent migration of volcanism with time
285
rules out a source below the 670 km discontinuity. If the Cameroon line were the product of a deepmantle plume it would require that lower-mantle convection has kept pace exactly, in both velocity and direction, with the movement of the African plate over the past 65 Ma. Therefore the convecting upper mantle is the only plausible mantle source. The convecting upper mantle (of which the asthenosphere is the upper part) is also the source of MORB. Fitton & Dunlop (1985) proposed that the LILE-rich Cameroon line basalts could be derived from the LILE-depleted MORB source by small-degree (about 0.2%) melting using bulk partition coefficients consistent with experimentally determined values for mantle silicate phases. They suggested that the convecting upper mantle, although LILE depleted in its bulk composition, contains LILE-rich streaks. The presence of an ancient enriched component in the asthenosphere is necessary to account for the small but consistent isotopic differences between MORB and intraplate basalts. Such heterogeneities could be produced by the migration of magma through the asthenosphere (cf Hawkesworth et al. 1984). They may also be produced by the return of ocean crust (including ocean islands) to the asthenosphere during subduction. Whatever their origin, heterogeneities in the asthenosphere will be drawn out into streaks by convection (Hoffman & McKenzie 1985). Partial melts from these streaks will be selectively incorporated into smalldegree melts, whereas larger-degree melting will tend to homogenize the source and result in magmas (MORB) reflecting the bulk composition of the asthenosphere. T
,
T
CAMEROON
,
LINE
c
n3 lO0
so o
o Oceanic sector mean
lo
I
I
y
I
I
Rib Ba I~ NIb La cle Sir NId P Zr
1
Ti
I
Y
Fie. 9. Normalized abundance patterns (Sun 1980) for
incompatible-elementconcentrations in average continental (134 analyses) and oceanic (144 analyses) basic volcanic rocks from the Cameroon line. Data from the Etinde nephelinites are not included in the averages. (After Fitton & Dunlop 1985).
J. G. F i t t o n
286
from the asthenosphere, the mantle part of the lithosphere and the crust. Radiogenic isotope studies can constrain the possible magma sources in some cases but cannot always distinguish unambiguously between contributions from continental crust and enriched lithosphere mantle. In the case of the Cameroon line we can be confident that the basic magmas have not interacted to any significant extent with either the mantle or crustal parts of the lithosphere. This provides a unique opportunity of studying the effects of environment on the subsequent evolution of two essentially identical sets of basalt magmas during storage in the oceanic and continental lithosphere respectively. The compositional range of the Cameroon line volcanic rock samples collected from all the main centres is illustrated on an alkali-silica diagram in Fig. 10. Despite the similarity of the oceanic and continental basalts, the intermediate and evolved rocks from the two sectors show a clear divergence. With very few exceptions, the oceanic magmas evolve to phonolite and the continental magmas to rhyolite. The two samples of oceanic quartz trachyte (Fig. 10) were both collected from the Ilhbu das Cabras off S~o Tomb; the two continental phonolite samples are from plugs on the Ngaoundbrb Plateau. Apart from these four samples, the separation of oceanic and continental salic rocks is complete. This is further illustrated in Fig. 11 in which the CIPW norms of the salic rocks have been projected into the residua system (quartz-nepheline-kalsilite).
McKenzie (1984, 1985) has recently derived a set of equations describing the compaction of partially molten rock and concluded that melt will begin to flow at very small degrees of melting (less than 0.570) provided that its viscosity is low. The degrees of melting required to produce LILErich intra-plate basalts from the MORB source should no longer be regarded as impossibly small. If the Cameroon line basic magmas were generated by small-degree melting of the MORB source then the same must also be true of other intra-plate basic magmas. This has been discussed by Fitton & James (1986) who compared incompatible-element concentrations in a wide range of intra-plate volcanic rocks with calculated concentrations in liquids generated by variable degrees of equilibrium melting of the MORB source. Partition coefficients required by the Cameroon line data were used in the calculations. The comparisons are encouraging and suggest that ocean island and rift valley magmas may share a common asthenosphere source. Metasomatic enrichment of the source is not necessary.
Origin of the salic rocks The extent to which contamination of magmas by crustal rocks influences their composition and evolution has been debated for many years (see, for example, Moorbath et al. 1984). Magmas in continental provinces may receive contributions
16
L
-s 9
14
9
o
s
:
~ 9
12 0 Z + 10 0 Z
s
eo 9
O ,
0
:3
9
9
9
9O 9
~
t
o
O0
9 O
9 A
8
A 9
%6
0 o ~ 00
0 0
9 9
9
0 " "
0
9
~
9
9
9
9 09
9
9
9
4
A ~-'~3~:
~ 2
0
0
4
,0/
w
~ /oo /
4t5
9 Etinde nephelinites 9 Other continental sector volcanic rocks 9 Oceanic sector volcanic rocks 510
L
55
~ 60
J 65
t 70
J 75
Weight % SiO 2
FIG. 10. Alkali-silicadiagram for the Cameroon line volcanic rocks. The line separating Hawaiian tholeiitic and alkaline rocks (Macdonald & Katsura 1964) is shown for comparison. (From Fitton 1983.)
The Cameroon line
287
Q
9 Continental sector o Oceanic sector s from granulite xenoliths
Ne
Weight %
Ks
FIG. 11. Normative compositions of evolved (more than 80%) normative salic components) volcanic rocks from the Cameroon line projected into the residua system. Additional data from Manengouba (supplied by C. A. Hirst) are included. Phase boundaries (P(H20)= 5 kb) are taken from Hamilton & MacKenzie (1965). The oceanic rocks occupy the thermal valley between the trachyte and phonolite minima, whereas the continental rocks plot in the corresponding valley down to the granite minimum. This separation into undersaturated oceanic and oversaturated continental salic rocks must be related to the environment in which the magmas evolved. The basaltic parental magmas in both sectors of the Cameroon line are transitional to strongly-alkaline in character (Figs 7 and 10). Low-pressure crystal fractionation alone should therefore have produced salic rocks ranging from phonolite to trachyte, with only small amounts of oversaturated magma (as in the oceanic sector). Progressive crustal contamination of the fractionating magmas provides the simplest explanation for the dominance of oversaturated salic rocks in the continental sector. Evidence for crustal contamination is provided by the occurrence of partly digested granulite xenoliths in a basalt lava flow from Bambouto, near to the village of Babadjou. The xenoliths are partially melted and contain patches of fresh glass in a residual mineral assemblage of quartz, plagioclase, orthopyroxene and minor-alkali feldspar. Electron microprobe analyses of the glass
show it to have a composition close to the minimum on the quartz-alkali feldspar cotectic (Fig. 11). Contamination of the continental magmas with crustal rocks or their partial melts could allow the magmas to cross the low-pressure thermal divide and evolve towards rhyolite. Dunlop's (1983) strontium isotope study of the Cameroon line volcanic rocks provides further evidence for crustal contamination. Although the oceanic and continental basalts are generally indistinguishable in their isotope ratios, the distribution of 87Sr/86Sr in the continental basalt samples is skewed slightly towards higher values (Fitton & Dunlop 1985). The contaminated basalt from Bambouto has the highest initial 875r/86Sr ratio (0.70403) of all the continental basic lavas analysed. It is in the evolved rocks, however, that crustal contamination can be most clearly demonstrated. The trachytes and rhyolites from Bambouto, Oku and the Mandara Mountains have initial 87Sr/86Sr values ranging up to 0.715. For comparison, two of the granulite xenoliths from Bambouto have age-corrected 875r/86Sr ratios of 0.71021 and 0.72088 (Dunlop 1983). A possible alternative to crustal contamination as a means of crossing the low-pressure thermal
J. G. F i t t o n
288
divide is provided by fractional crystallization of hornblende-bearing assemblages. Cawthorn et al. (1973), for example, suggested that removal of silica-poor ne-normative hornblende from undersaturated basic magma could lead to the evolution of silica-oversaturated residual liquids. This mechanism cannot apply, however, to the Cameroon line magmas since hornblende is seldom found as a phenocryst phase in the continental volcanic rocks. Curiously, though, hornblende phenocrysts are abundant in the intermediate lavas and even in some of the basalts in the oceanic sector. The abundance of hornblende in the oceanic lavas and its scarcity in the continental rocks is puzzling since the basic magmas in the two sectors were compositionally identical (Fig. 7). The most important factor controlling amphibole stability in basaltic magma is the partial pressure of water vapour (Allen & Boettcher 1978). It is possible that magma stored beneath the islands of the oceanic sector had easier access to water than did the continental magma reservoirs, and that seepage of small amounts of sea-water into the oceanic magma reservoirs stimulated the crystallization of hornblende. The middle and late stages in the evolution of magmas on S~.o Tom6 are dominated by fractionation of hornblende-bearing assemblages (Fig. 5). This is clearly demonstrated by the abundance of apparently cognate hornblende-rich cumulate
blocks. The operation of this process in the oceanic magmas, but not in those of the continent, is well illustrated in a plot of Y against Zr (Fig. 12). Y is a compatible element in amphiboles (Pearce and Norry 1979), and so hornblende removal would cause the concentration of Y to fall. Zr, which is incompatible with all the observed phenocryst phases, is a useful index of fractionation. Figure 11 shows that the Y concentration rises with fractionation in the continental lavas, where hornblende is seldom present, but falls in the intermediate and evolved oceanic sector rocks. The hornblende-bearing cumulate blocks from S~o Tomb are relatively rich in Y. Fractional crystallization involving the removal of this cumulate material from the basic to intermediate magmas (cf. Fig. 5) would increase the Zr/Y ratios in the residual liquids along a line such as that on Fig. 12. This line was constructed by applying the Rayleigh fractionation equation to the average oceanic-sector basic lava composition (from Fitton & Dunlop 1985). Bulk partition coefficients for Zr and Y were taken as the ratio of their average concentrations in the cumulate blocks to the average in the oceanic-sector basic lavas. The degrees of fractionation indicated by the numbers on this line depend heavily upon the concentrations of Zr and Y in the crystal extract and serve only to show that the array of oceanicsector data points could be generated by crystal
/
300
/ 9 Etinde nephelinites
200
/ / /
9 Other continental s e c t o r volcanic r o c k s 9 O c e a n i c s e c t o r volcanic r o c k s
if
/
H o r n b l e n d e - r i c h cumulate b l o c k s
/
..../ ....
/
.//"
loo
j'/
.........,
,/
1.-E
-
-/
.............
9
Q
d'/O ....
~.// /
o
.. .....
o
.......---
o
--b-___.Z~ .@ ~ ./ ~-_:' o o y_.~; oo c~o-_ _--- ~ ~9 / / / -"-~ L /
3
~f
// L,//
10100
........ 9 9 /$~O
LI 9
9 2
20
9
1 " ~'
....
;.%
/ 9 6,1.9 i 9 1 4 .9~ k ~ C z"
_
9
9 9
~,
/1.
-'/
50
,~/~ O ///z~,~. ~
9
/z
from S&o Tome
>-
///"
200
.-
~..// I
I
I
300
400
500
pp.m
I
I
L
I
d
1000
2 O0
Zr
FIG. 12. Y and Zr contents of the Cameroon line volcanic rocks and cognate xenoliths. The fractionalcrystallization path has been calculated by subtracting the average xenolith composition from the average oceanic-sector basic volcanic rock composition. The percentages indicate the amount of crystallization.
The Cameroon line f r a c t i o n a t i o n o f h o r n b l e n d e - b e a r i n g assemblages. T h e r e is no e v i d e n c e t h a t r e m o v a l o f hornb l e n d e has c a u s e d the oceanic-sector m a g m a s to b e c o m e less u n d e r s a t u r a t e d . Virtually all these m a g m a s h a v e e v o l v e d t o w a r d s the p h o n o l i t e m i n i m u m (Figs 10 a n d 11) w h i c h shows clearly t h a t crystal f r a c t i o n a t i o n i n v o l v i n g h o r n b l e n d e is not an efficient m e c h a n i s m for crossing the lowpressure t h e r m a l divide.
289
ACKNOWLEDGMENTS.Field work in Cameroon, Nigeria and the Gulf of Guinea was financed by research grants from the U.K. Natural Environment Research Council and The Royal Society. The author is indebted to D. J. Hughes for his assistance with the field work, to H. M. Dunlop, C. A. Hirst and R. M. Maclntyre for permission to use their unpublished data, and to D. James and G. R. Angell for assistance with the chemical analyses. D. James, B. G. J. Upton and P. Vincent are thanked for their constructive comments on an earlier draft of this paper.
References ALLEN, J. C. & BOETTCHER,A. L. 1978. Amphiboles in andesite and basalt: II. Stability as a function of P - T - f H 2 0 - f O2. Am. Mineral. 63, 1074-87. BAILEY, D. K. 1982. Mantle metasomatism--continuing chemical change within the Earth. Nature, Lond. 296, 525-30. 1987. Mantle metasomatism--perspective and prospect. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, 1-13. BOWDEN, P., BLACK, R., MARTIN, R. F., IKE, E. C., KINNAIRD, J. A. & BATCHELOR,R. A. 1987. NigerNigerian alkaline ring complexes: a classic example of African Phanerozoic anorogenic mid-plate magmatism. In: FITTON, J. G. & UPTON, B. G. J. (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, 357-379. CANTAGREL, J.-M., JAMOND, C. (~; LASSERRE,M. 1978. Le magmatism alkalin de la ligne du Cameroun au Tertiaire inf6rieur: donn6es g6ochronologiques K--Ar. C. R. Somm. Soc. g~ol. Ft. 6, 300-3, CARTER, J. D., BARBER, W. d~; TAIT, E. A. 1963. The geology of parts of Adamawa, Bauchi and Bornu Provinces in northeastern Nigeria. Bull. geol. Surv. Nigeria, 30. CAWTHORN, R. G., CURRAN, E. B. & ARCULUS, R. J. 1973. A petrogenetic model for the origin of the calc-alkaline suite of Grenada, Lesser Antilles. J. Petrol. 14, 327-37. CHAPUT, M., LOMBARD,J., LORMAND,J. & MICHEL, H. 1954. Granites et traces d'btain dans le Nord Cameroun. Bull. Soc. gbol. Ft. 4, 373-93. CLAGUE, D. A. & FREY, F. A. 1982. Petrology and trace element geochemistry of the Honolulu Volcanics, Oahu: implications for the oceanic mantle below Hawaii. J. Petrol. 23, 447-504. CORNEN, G. & MAURY, R. C. 1980. Petrology of the volcanic island of Annobon, Gulf of Guinea. Mar. Geol. 36, 253-67. DI~RUELLE, B., MOREAU, C. & NSIFA, E. N. 1983. Sur la r6cente 6ruption du Mont Cameroun (16 octobre-12 novembre 1982). C. R. Acad. Sci. Paris, 296, 807-12. DUNLOP, H. M. 1983. Strontium isotope geochemistry
and potassium-argon studies on volcanic rocks from the Cameroon line, West Africa. PhD Thesis, University of Edinburgh (unpublished). -& FITTON, J. G. 1979. A K - A r and Sr-isotopic study of the volcanic island of Principe, West Africa--evidence for mantle heterogeneity beneath the Gulf of Guinea. Contrib. Mineral. Petrol. 71, 125-31. EMERY, K. O. & UCHUPI, E. 1984. The Geology of the Atlantic Ocean, Springer, New York. ESCH, E. 1901. Der Vulcan Etinde in Kamerun und seine Gesteine (I). Sitzungsber. Akad. WiNs. Berlin, 277-99. FITTON, J. G. 1980. The Benue trough and Cameroon line--a migrating rift system in West Africa. Earth planet. Sci. Lett. 51, 132-8. -1983. Active versus passive rifting: evidence from the West African rift system. Tectonophysics, 94, 473-81. - - & DUNLOP, H. M. 1985. The Cameroon line, West Africa, and its bearing on the origin of oceanic and continental alkali basalt. Earth planet. Sci. Lett. 72, 23-38. -& HUGHES, D. J. 1977. Petrochemistry of the volcanic rocks of the island of Principe, Gulf of Guinea. Contrib. Mineral. Petrol. 64, 257-72. & -1981. Strontian melilite in a nephelinite lava from Etinde, Cameroon. Mineral. Mag. 44, 261-4. t~ JAMES, D. 1986. Basic volcanism associated with intraplate linear features. Phil. Trans. R. Soc. Lond., Ser. A, 317, 253-66. - - , KILBURN, C. R. J., THIRLWALL,M. F. & HUGHES, D. J. 1983. 1982 eruption of Mt. Cameroon, West Africa. Nature, Lond. 306, 327-32. FREETH, S. J. 1979. Deformation of the African plate as a consequence of membrane stress domains generated by post-Jurassic drift. Earth planet. Sci. Lett. 45, 93-104. FREY, F. A., GREEN, D. H. & ROY, S. D. 1978. Integrated models of basalt petrogenesis: a study of quartz tholeiites to olivine melilitites from southeastern Australia utilising geochemical and experimental petrological data. J. Petrol. 19, 463-513. -
-
-
-
29o
J. G. Fitton
FRISCH, T. & WRIGHT, J. B. 1971. Chemical composition of high-pressure megacrysts from Nigerian Cenozoic lavas. Neues Jb. Mineral. Monatsh. 289304. GAZEL, J. 1956. Carte Gdologique du Cameroun au 1/1,000,000. Direction des Mines et de la G6ologie du Cameroun, Paris. - - , LASSERRE, M., LIMASSET,J.-C. & VACHETTE, M. 1963. Ages absolus des massifs granitiques ultimes et de la min~ralisation en &ain du Cameroun central. C. R. Acad. Sci. Paris, 256, 2875-8. GI~ZE, B. 1943. G6ographie physique et g~ologie du Cameroun occidental. Mem. Mus. natl Hist. nat. Paris, Nouv. Skr. 17, 1-272. -1953. Les volcans du Cameroun occidental. Bull. Volcanol. 13, 63-92. GORINI, M. A. & BRYAN, G. M. 1976. The tectonic fabric of the Equatorial Atlantic and adjoining continental margins: Gulf of Guinea to northeastern Brazil. An. Acad. Bras. Ci~nc. 48 (Suppl.), 10119. GOUHIER, J., NOUGIER, J. & NOUGIER, D. 1974. Contribution ~t l'&ude volcanologique du Cameroun ("Ligne du Cameroun"--Adamaoua). Annls Fac. Sci. Cameroun, 3-48. GRANT, N. K., REX, D. C. & FREETH, S. J. 1972. Potassium-argon ages and strontium isotope ratio measurements from volcanic rocks in northeastern Nigeria. Contrib. Mineral. Petrol. 35, 277-92. GREEN, D. H. & RINGWOOD, A. E. 1967. The genesis of basalt magmas. Contrib. Mineral. Petrol. 15, 103-90. GRUNAU, H. R., LEHNER, P., CLEINTUAR, M. R., ALLENBACH, P. & BAKKER, G. 1975. New radiometric ages and seismic data from Fuerteventura (Canary Islands), Maio (Cape Verde Islands) and S~,o Tom6 (Gulf of Guinea). In: BORRADAILE, G. J., RITSEMA,A. R., RONDEEL, H. E. & SIMON, O. J. (eds) Progress in Geodynamics, pp. 90-118. North-Holland, Amsterdam. HAMILTON, D. L, & MACKENZIE, W. S. 1965. Phase equilibrium studies in the system NaA1SiO,, (nepheline)-KA1SiO4 (kalsilite)-SiO2-H20. Mineral. Mag. 34, 214-31. HARRIS, P. G. 1957. Zone refining and the origin of potassic basalts. Geochim. cosmochim. Acta, 12, 195-208. HAWKESWORTH, C. J., ROGERS, N. W., VAN CALSTEREN, P. W. C. & MENZlES, M. A. 1984. Mantle enrichment processes. Nature, Lond. 311,331-5. HEDBERG, J. D. 1968. A geological analysis of the Cameroun Trend. PhD Thesis, Princeton University (unpublished). HOFFMAN, N. R. A. & MCKENZIE, D. P. 1985. The destruction of geochemical heterogeneities by differential fluid motions during mantle convection. Geophys. J. R. Astron. Soc. 82, 163-206. JACQUEMIN, H., SHEPPARD, S. M. F. & VIDAL, P. 1982. Isotopic geochemistry (O, Sr, Pb) of the Golda Zuelva and Mboutou anDrogenic complexes, North Cameroon: mantle origin with evidence for crustal contamination. Earth planet. Sci. Lett. 61, 97-111.
LASSERRE, M. 1978. Mise au point sur les granit6ides dits "ultimes" du Cameroun: gisement, p6trographie et g+ochronologie. Bull. Bur. Rech. gdol. min., Paris, Ser. 2, Sect IV, no. 2, pp. 143-59. LIOTARD, J. M., DUPUY, C., DOSTAL,J. & CORNEN, G. 1982. Geochemistry of the volcanic island of Annobon, Gulf of Guinea. Chem. Geol. 35, 11528. LLOYD, F. E. & BAILEY, D. K. 1975. Light element metasomatism of the continental mantle: the evidence and the consequences. Phys. Chem. Earth, 9, 389-416. MACDONALD, G. A. & KATSURA, T. 1964. Chemical composition of Hawaiian lavas. J. Petrol. 5, 82133. MCKENZIE, D. 1984. The generation and compaction of partially molten rock. J. Petrol. 25, 713-765. 1985. The extraction of magma from the crust and the mantle. Earth planet. Sci. Lett. 74, 81-9 I. MENZIES, M. A. & MURTHY, V. R. 1980. Nd and Sr isotope geochemistry of hydrous mantle nodules and their host alkali basalts : implications for local heterogeneities in metasomatically veined mantle. Earth planet. Sci. Lett. 46, 323-34. MITCHELL-THOMi~, R. C. 1970. Geology of the South Atlantic Islands. Gebruder Borntraeger, Berlin. MOORBATH, S., THOMPSON, R. N. & OXBURGH, E. R. (eds) 1984. The Relative Contributions of Mantle, Oceanic Crust and Continental Crust to Magma Genesis, 342 pp. The Royal Society, London. OKEKE, P. I. 1980. Petrology of igneous and metamorphic rocks in the area around Gwoza, northeast Nigeria. M. Phil. Thesis, University of Edinburgh (unpublished). PARSONS, I., BROWN, W. L. & JACQUEMIN, H. 1986. Mineral chemistry and crystallization conditions of the Mboutou layered gabbro-syenite-granite complex, North Cameroon. J. Petrol. 27, 13051330. PEARCE, J. A. & NORRY, M. J. 1979. Petrogenetic implications of Ti, Zr, Y, and Nb variations in volcanic rocks. Contrib. Mineral. Petrol. 69, 33-47. PIPER, J. D. A. & RICHARDSON, A. 1972. The palaeomagnetism of the Gulf of Guinea volcanic province, West Africa. Geophys. J. R. Astron. Soc. 29, 147-71. PLATT, R. G. & EDGAR, A. D. 1972. The system nepheline-diopside-sanidine and its significance in the genesis of melilite and olivine bearing alkaline rocks. J. Geol. 80, 224-36. SEAN (Scientific Event Alert Network) 1985. Bulletin, 10(8), 3-4. -1986. Bulletin, 11(8), 2-5. SmUET, J.-C. & MASCLE, J. 1978. Plate kinematic implications of Atlantic equatorial fracture zone trends. J. geophys. Res. 83, 3401-21. SMITH, A. G., HURLEY, A. M. & BRIDEN, J. C. 1981. Phanerozoic Palaeocontinental World Maps, 102 pp. Cambridge University Press, Cambridge. SUN, S.-S. 1980. Lead isotopic study of young volcanic rocks from mid-ocean ridges, ocean islands and island arcs. Phil. Trans. R. Soc. Lond., Ser. A, 297, 409-45. -
-
The Cameroon line TCHOUA, F. 1972. Sur la formation des calderas des Monts Bambouto (Cameroun). C. R. Acad. Sci. Paris, 274, 799-801. 1973. Sur l'existence d'une phase initiale ignimbritique dans le volcanisme des Monts Bambouto (Cameroun). C. R. Acad. Sci. Paris, 276, 2863-6. TYRRELL, G. W. 1934. Petrographical notes on rocks from the Gulf of Guinea. Geol. Mag. 71, 16-23. VAN HOUTEN, F. B. 1983. Sirte Basin, north-central Libya: Cretaceous 'rifting above a fixed mantle hotspot? Geology 11, 115-8. -
-
2 91
VINCENT, P. 1968. Attribution au Cr6tac~ de conglom6rats m6tamorphiques de l'Adamaoua (Cameroun). Annls Fac. Sci. Cameroun, 1, 69-76. WINCHESTER,S. 1984. The nastiest place on Earth? The Sunday Times Magazine, Jan. 29, 1984, pp. 30-37. WRIGHT, J. n. 1970. High pressure phases in Nigerian Cenozoic lavas: distribution and setting. Bull. VolcanoL 34, 833-47. WRIGHT, T. L. 1984. Origin of Hawaiian tholeiite: a metasomatic model. J. geophys. Res. 89, 3233-52.
J. G. FITTON, Grant Institute of Geology, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, U.K.
Outline of the petrology of the Kenya rift alkaline province B. H. Baker S U M M A R Y : The Kenyan and N Tanzanian volcanic province contains sodic alkaline rocks ranging from melilitites and melanephelinites to transitional alkali basalts and their differentiates. Individual volcanoes display three principal magmatic suites: (i) nephelinitic; (ii) alkali basaltic; (iii) transitional basaltic. However, some large volcanoes contain more than one of these suites implying that parental magmas of variable alkalinity were available at certain times and places. A general decrease in alkalinity with time is detectable in the rift zone and for any time period there was a tendency for the least alkaline magmas to be erupted within the central and deepest part of the rift zone. Compositional variation within the suites was largely controlled by low-pressure crystalliquid fractionation. Extended fractionation produced salic differentiates. Liquid fractionation caused upward segregation of phonolitic and trachytic magmas which were erupted in preference to more mafic magmas. Isotopic data suggest that crustal contamination did not occur on a large scale.
Introduction The upper Cenozoic alkaline igneous province associated with the eastern branch of the E African rift system in Kenya and northern Tanzania contains nearly the whole spectrum of sodic alkaline igneous compositions, and includes unusually large proportions of intermediate and silicic rocks. The volcanic and tectonic evolution of the rift valley during the last 30 Ma is now moderately well understood, providing an opportunity to assess interrelations between volcanism and tectonism in time and space, to summarize the petrology and geochemistry of the province and to make inferences concerning petrogenetic processes. This review is an outline of the characteristics of the province, and is biased towards areas and suites familiar to the author. Much of the region is known only from reconnaissance mapping and study of thin sections, supported by a few rock analyses. Few volcanic suites have been analysed systematically for a wide range of geochemical elements. The geochemical data used to prepare variation diagrams include the majority of wellstudied suites of alkaline rocks in Kenya, including some unpublished data. Apart from questions concerning the petrogenesis of individual volcanic series, the major problems of the province concern conditions under which parental magmas were formed, and the characteristics of the mantle source. The fact that there are some systematic variations in the alkalinity of lavas in space and time invites consideration of the tectonic effects of rifting. The bimodality of transitionally alkaline suites and large volumes of salic rocks raise questions concerning low-pressure differentiation processes
and the degree to which crustal contamination is involved. In view of recent progress in understanding physical processes in magma chambers the opportunity is taken to comment on how such processes might influence petrogenesis and eruptive mechanisms.
General characteristics of the province The alkaline volcanic province in Kenya and northern Tanzania is associated with the development of the eastern branch of the E African rift system (Fig. 1). Volcanism began about 30 Ma ago in northern Turkana and extended progressively southward, becoming more voluminous 16 Ma ago and continuing to the present (Baker et al. 1971; Williams 1978; Williams & Truckle 1980). In Kenya and northern Tanzania about 220 000 km 3 of rock were erupted (Williams 1982), of which 68% are mafic rocks (King 1978). This amounts to a dense rock equivalent eruption rate of about 0.006 km 3 per year, but the parental magma production rate must have been at least an order of magnitude greater than this. Tectonic evolution of the rift began with development of a shallow basin in the Turkana region in the N sometime in the early Miocene, in which some 35 000 km 3 of alkali basalt lavas were erupted (Fig. 1). By 15 Ma BP faults had begun to develop on the western side of this depression, and alkali basalts and voluminous phonolites were erupted in a developing halfgraben (Fig. 2, upper part). By 7 Ma BP a major fault extended along the
From: Fir'roy, J. G. & UVTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 293-311.
293
294
B. H. Baker
FXG. 1. Geological map of the alkaline igneous province of Kenya and northern Tanzania (bold lines are major faults): 1, nephelinite-carbonatite suite; 2, Miocene alkali basalts; 3, Miocene P-phonolites; 4, mixed association volcanoes; 5, Pliocene alkali basalts; 6, Pliocene to Recent transitional basalts and trachytes; 7, Pliocene to Recent alkali basalts of the E rift flank (clusters of vents shown as dots). Abbreviations are volcano and formation names referred to in the text. western side of the rift, which was a half-graben bounded by a monoclinal flexure on its eastern side. At this time there was a change to less alkaline volcanism characterized by the building of basalt-trachyte shield volcanoes, although more alkaline central volcanism continued at intervals in the Kavirondo rift, Uganda borderland and N Tanzania. By 4 Ma BP major faults were developing on the eastern side of the rift giving rise to a deeping graben structure in which basalt-trachyte volcanism continued to predominate. It is convenient to divide the tectonic development into pre-rift (3012MaBP), half-graben ( 1 2 - 4 M a BP) and graben stages of development (Fig. 2). The elevation of the flanking plateaux of the
rift valley is due in part to local areas of residual highlands composed of Precambrian rocks, to the accumulation of lavas that periodically filled the rift and overflowed its flanks, and to late uplifts which reached a maximum of 1700 m in the central region (Saggerson & Baker 1965; Williams et al. 1983). The thickness of volcanic rock accumulated in the central section of the rift valley is at least 3 km, and it can be inferred that the sub-volcanic surface is below sea level along much of the rift, being deepest in its central section. Volcanism of the pre-rift stage was dominantly of many small basaltic shields in the Baringo and Turkana regions, with scattered contemporaneous nephelinite-carbonatite central volcanism in the K e n y a - U g a n d a borderland (King et al. 1972) and the Kavirondo rift valley (Le Bas 1977). During the half-graben stage alkali basalts and voluminous phonolites were erupted, but by 7 Ma BP the character of volcanism changed to a transitionally alkaline character marked by the growth of many bimodal basalt-trachyte volcanoes, many of them being calderas with ash-flow tufts (Webb & Weaver 1975 ; Williams 1978). During Plio-Pleistocene times the eastern flank of the rift was the site of widespread fissure volcanism that built chains of many monogenetic alkali basalt cones, such as the Chyulu, Nyambeni and Hurri ranges, and isolated low basaltic shields such as the Marsabit volcano (Fig. 1, Ch, Ny, Hu and Ma). On or near the eastern edge of the rift large central volcanoes composed of a great variety of lavas formed Mount Kenya and Kilimanjaro (Ke and Ki). Within the rift contemporaneous Quaternary volcanism consisted of voluminous eruptions of trachyte and alkali rhyolite, and the building of a chain of basalttrachyte caldera volcanoes (Fig. 1, Su, Lo, Mi, Si and Em).
Associations of alkaline igneous rocks The Kenyan sodic igneous rocks have been divided into suites of contrasting alkalinity, consisting of a strongly alkaline suite of nephelinites, basanites, alkali basalts, tephrites and phonolites, and a weakly alkaline suite of olivine basalts, trachytes, trachyphonolites and alkali rhyolites (Saggerson & Williams 1964; Williams 1969a, 1970, 1972; King 1970; Saggerson 1970; King et al. 1972; Baker et al. 1978; Williams & Truckle 1980). These suites are characterized by normative ne greater and less than 5%, and within the rift zone are broadly represented by Miocene and post-Miocene volcanism (Williams 1972).
Petrology of the Kenya rift A-NORTH My
W flank
o
295
SECTOR
--------- R i f t
,,
2
~
E
flank
=~,~--~~
4 6:
a:
~,~__~_._t_r.__~__=~:~-~'~-~.It i::::::::::::..~.
II i'
10-
12: 14: ]6: 18: 20: 22: 24:
"l
o
2: 4"6:
".... I "'~;b
=====================[
depression
J Ii
My
half graben
B-SOUTH W
=
SECTOR Rift
l
~
~ ~ ~ :
AI
~-~', b --'-,-'_---_"-~_~--7~L.-':~Zb ~ i i ~ i ~ ~ ~:= tr 1
b'~l~l~i~A
A
E
I t,
_tr
~'~..'1"...
igraDell ] half -
j~graben
8: io:
12:
unexposed basanitoid rocks ? . . . . . . . . . . . 7 " . ~ Z Z Z z ~ " ..........
I
Nephelin ite-pho no[ ite-(carbonat ite) Basanite/atk.basalt- phonotite
~
Mixed associations
Basatt - t rachyte-rhyolite/t rachyphonoii te b-basaqite or basalt; ph-phonotite; tr-trachyte; tph-tr&chyphono[ite I periods of faulting FIG. 2. Diagrammatic summary of the ages and extent of lavas of different associations and their relation to phases of rift faulting in the northern and southern sectors of the rift.
Many varied associations can be recognized among the more strongly alkaline rocks (Miyashiro 1978), and it is convenient to divide the strongly alkaline suite further on the character of its differentiated rocks into a nephelinitic suite associated with sphene-bearing phonolites and carbonatites and an alkali olivine basalt-basanite suite associated with sphene-free phonolites. Many volcanoes consist of one or other of these suites, but some large central volcanoes contain
representatives of both and are referred to as 'mixed' association volcanoes. By using association in space and time as an indicator of petrogenetic relationship, the province can be divided into rocks that fall into four types of volcanic association. l Melilitites, melanephelinites, ankaratrites, nephelinites, and G-type phonolites characterized by high Sr and Ba (Table 1 ; Table 2, analyses
296
B. H. Baker
TABLE 1. Nephelinite--phonolite suite 1
2
3
4
5
6
Major elements (wt.~)
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO NazO K20 P2Os
43.0 3 9 . 0 4 9 . 2 5 5 . 4 40.5 43.8 3.1 3.7 1.3 0.5 2.4 2.5 11.2 8.3 18.3 2 0 . 8 15.4 12.9 8.2 6.9 5..0 3.5 6.7 7.1 6.3 7.1 2.9 1.5 3.4 5.2 0.2 0.2 0.2 0.2 0.2 0.3 7.8 9.7 1.7 0.5 3.4 6.0 14.2 19.3 7.2 2.9 16.6 12.7 5.9 3.4 2.4 9.1 9.2 7.2 2.8 1.8 1.9 4.6 5.5 3.1 0.7 0.8 1.2 0.4 0.1 0.9
FeO* Mg~
13.7 50.3
13.3 7.5 4.7 9.5 5 6 . 9 2 9 . 3 14.7 40.1
11.5 48.2
Alkaline norm in we~htpercentvo~ti&~ee~
Ap Mt
Sp Pv Di Geh Akr Wo Ac O1 Or Ab Ne Ks Cg
3.0 7.5 4.8 22.9 10.3
5.5 7.2 19.1 19.7
4.5 1.0 5.9 1.6 2.1 1.3 1.5 3 8 . 4 10.4 5.9
0.3 2.0 0.6
0.5 0.6
0.3
3.8 4.3
4.4
2.5 4.3 1.2 17.0 0.4 0.2 3.7
5.8 13.7 11.6 27.8 36.1 6 3 . 9 50.3 68.4 1.3 0.1 0.1 2.2
2.1 4.8 2.9 23.8 0.3
0.7 1.9 7.7 55.9
Average analyses from Kavirondo rift valley and eastern Uganda (Le Bas 1977). The number of analyses is given in parentheses:l, melanephelinites (43); 2, melilitites (6); 3, nephelinites (14); 4, G-phonolites (55); 5, ijolites (41); 6, estimated parent magma. * Total Fe as FeO t Cation ratio Mg/(Mg + Fe) :~From Le Bas 1973. 1-4). Older volcanoes of this type may have exposed sub-volcanic intrusions composed of ijolite, nepheline syenite and carbonatite. 2 Limburgites, basanites, alkali basalts, tephrites, and P-type phonolites characterized by low Sr and Ba (Table 2, analyses 7 and 8; Table 3, analyses 4 and 7). 3 Transitional basalts, ferrobasalts, scarce mugearites and benmoreites, abundant trachytes, and alkali rhyolites (Table 3, analyses 1-3, 9 and 10; Table 4). Locally the differentiated rocks are trachyphonolites, and in a few suites oversaturated and undersaturated silicic rocks occur on the same volcano.
4 Mixed association central volcanoes which contain usually two of the suites mentioned above, the suites being erupted either sequentially or in alternating phases (Figs 1 and 2). The fields of the first three suites are shown on a Q L M diagram (Fig. 3(a)) in which a continuum of compositions between the transitional (T-ba) and alkali basalt (A-ba) suites, and a gap between the alkali basalt and nephelinitic (M-ne) suites at mafic and intermediate compositions can be seen. The silicic rocks cover the entire range from the oversaturated (Rhy) to the undersaturated (GPh) minima. The trend lines for individual volcanic series are shown in Fig. 3(b). The transitional basalt series exhibits a marked Daly gap, and is associated with voluminous trachytes (TR) and comenditic and pantelleritic rhyolites (RH). Some of the trachytic series contain both rhyolitic and trachyphonolite (T-ph) lavas, suggesting that oversaturated and undersaturated trends can develop from a saturated trachyte stem. Trends within the alkali olivine basalt-basanite suite are more variable, but all trend toward, or end in, the field of P-phonolites (PP). In these suites intermediate rocks such as tephrites and rhomb porphyries are less well represented than the end-member basalts and phonolites. In the strongly-alkaline series, melilitites (ME) and melanephelinites (NE) are usually accompanied by nephelinites and intrusive members of the ijolite series. Strongly-alkaline phonolites (GP) with ne greater than 15% are present in subordinate amounts. There is a tendency for the Daly gap to be most marked in the least alkaline series, becoming absent in the most alkaline series. The phonolitic end-members of nephelinitic and basanitic suites are scarcely distinguishable on the basis of major-element compositions, since both suites converge on the phonolite minimum, but are readily identified by their contrasting Sr, Ba and light rare-earth element (REE) contents (see below).
Distribution and volcanological characteristics of the suites The nephelinite-carbonatite association is represented by large central volcanoes composed largely of pyroclastic rocks and debris flows, exemplified by Mounts Elgon, Napak, Meru, Hanang and O1 Doinyo Lengai. They range in age from 22 Ma to recent, and are found on the N W flank of the rift in the K e n y a - U g a n d a borderland (Davies 1952; King 1965), along the Kavirondo branch of the rift and in southern
Petrology of the Kenya rift
297
TABLE 2. Strongly- and mildly-alkaline suites 1
Major elements (wt.~) SiO2 41.4 TiO2 2.5 Al103 9.2 Fe203 5.9 FeO 6.6 MnO 0.2 MgO 13.6 CaO 13.5 Na20 3.4 K20 1.4 H2O+ 1.8 PzOs 0.5 Trace elements Sc Ni Co Zr Ta Th Rb Sr Ba La Ce Nd Sm Eu Tb Yb
(ppm) 40 240 36 244 6.5 8.2 42 ND ND 63 145 53 10.5 2.5 ND 2.3
2
3
4
5
6
7
8
41.5 2.7 11.9 6.6 5.7 0.2 6.1 14.6 5.7 2.6 1.8 0.9
46.2 1.6 18.6 6.0 3.5 0.2 2.3 7.3 9.3 4.2 1.0 0.5
51.7 0.9 19.3 3.9 2.4 0.2 1.1 4.1 8.9 4.6 2.6 0.3
45.8 2.5 14.0 4.9 8.1 0.3 8.0 10.8 2.9 1.2 1.0 0.5
61.5 0.7 13.7 4.5 4.5 0.3 0.5 1.3 6.8 4.9 1.8 0.3
48.6 3.4 15.1 5.8 5.5 0.2 4.2 8.5 4.1 2.4 ND 0.8
44.3 2.9 12.1 3.3 10.0 0.2 11.3 10.6 2.9 1.1 ND 0.6
16 38 38 263 7.2 8.2 63 1160 915 68 117 52 10.9 2.5 ND 2.5
3.5 17.7 6.0 703 10.0 17.3 116 2260 1650 110 179 53 8.5 2.2 1.2 2.2
3.5 11.5 12.0 840 15.0 24.6 130 1190 1240 93 163 54 10 2.2 ND 2.2
23 135 51 291 3.2 3.8 33 735 440 31 61 30 6.3 2.0 ND 2.0
ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND
14 44 34 396 4.6 9.8 52 1073 882 78 ND 69 13.1 3.7 1.2 2.3
23 ND 66 ND ND ND ND ND ND 52 ND ND 9.8 ND ND ND
3.1
U
ND
Y
29
31
29
3.5 28
0.8 26
ND ND
ND 38
ND ND
FeO* Mg
11.9 67.1
11.6 48.3
8.8 31.8
5.9 24.1
12.6 53.2
7.9 9.3
10.7 41.1
13.0 60.8
1.8
1, Olivine melanephelinites; 2, melanephelinites; 3, nephelinites; 4, phonolites; 5, olivine basalts; 6, trachytes; 7, average Olokisalie basalt (Henage 1976); 8, Average Chyulu basalt (Goles 1975). Nos 1-6, average analyses of samples from Kenya, eastern Uganda and northern Tanzania (Beloussov et al. 1974). ND, no data. Kenya and northern Tanzania (King et al. 1972; Le Bas 1977; Figs 1 and 4). They have been divided into those having olivine-poor melanephelinite and carbonatite (Moroto (Mo) Yelele (Ye), Napak (Na), K a d a m (Ka), Elgon (El), Kisingiri, Wasaki and the early phase of Tinderet (Ti)), and those having olivine-rich nephelinites and basanites without exposed carbonatite (Ketumbeine, Moroto, Yelele and the later phase of Tinderet (Le Bas 1977)). O1 Doinyo Lengai (Fig. 1, O1) is unique in having erupted natrocarbonatite ash and lava in historical times (Dawson 1962). Kerimasi is a
similar but inactive volcano. Evidence of natrocarbonatite pyroclastic eruptions has also been found at Tinderet (Deans & Roberts 1984). The Kenya-Tanzania border region is characterized by many recent explosion craters which erupted carbonate tufts rich in xenoliths (Dawson 1964; Dawson & Powell 1969). Mildly-alkaline basanites and alkali basalts are widely distributed as low small-volume shields in the Turkana depression (Fig. 4), and as discrete series of lavas ranging from 16 to 6 Ma in age both on the flanks and in the northern part of the Gregory rift (Elgeyo basalts). They are associated
B. H. Baker
298
0
0
/ .... ,
~
/
/
'"
-/ ~TI ~ - ~ - - _ 2 ~ - - \ - - - ~
(a) (b) FIG. 3. QLM synoptic diagrams of data from the Kenya alkaline province: (a) fields of all analyses and commonly used rock names; (b) trends of series and volcanoes (ME, melilitites; NE, nephelinites; GP, G-type phonolites; AB, alkali basalts; PP, P-type phonolites; TP, trachyphonolites; TB, transitionally-alkalinebasalts; TR, trachytes: RH, alkali rhyolites). with large volumes of P-type phonolites in the Kamasia region, which reach an aggregate thickness of over 2000 m and a volume of about 4 0 0 0 0 k m 3 (Lippard 1973a and b; Williams 1982). Similar suites are found associated with more strongly-alkaline rocks on Kilimanjaro and other mixed association volcanoes (Figs 2 and 4). In the north-central rift basalts of intermediate alkalinity are found as formations intercalated in a dominantly phonolitic succession, and they show a tendency to become less alkaline with
~-,.~ti:-:i-~:!t-: ,38o ]
t t ~A
A
A
:i
,t
/
9
&A o ~"~ 2- ~ f A A A~
AL
Nephetinite-phonotite central volcanoes ~asanite and
phono[[te shields Quaternary basaLts Mixed ass central volcanoes
2~-
Basalt-trachyte trachyte-rhyotite,& trachyphono[ite shields FauLts
II
200 km l
FIG. 4. Map showing the distribution of major eruptive centres of different associations. Alignment of symbols indicates eruption on fissures.
decreasing age (Truckle 1977, 1980; Koyaguchi 1984). Alkali basalts with a different mode of eruption are widely distributed on the eastern rift flank, and form volcanic ranges composed of many monogenetic cinder cones aligned along fissures. On the SE and E flank they form the Quaternary Chyulu (Saggerson 1963; Goles 1975) and Nyambeni ranges (Mason 1955). On the NE flank similar lavas form the Hurri Hills and Marsabit shield (Fig. 1). Differentiated lavas are scarce or absent, and the magmas have risen rapidly from depth with little low-pressure fractionation. The more extensive basalt lava fields in the NE consist of a thin cover of flows of at least two ages. Those of the Kaisut region (Fig. 1, Ks) are slightly older than those farther NE, but they are all probably of Pleistocene age. Little is known of their composition; they appear to be alkali olivine basalts comparable with the Chyulu lavas. These more strongly-alkaline associations are commonly grouped together on variation diagrams to distinguish them from transitionally-alkaline suites associated with over-saturated differentiates (Saggerson 1970; King & Chapman 1972; Williams 1972; Truckle 1980). Transitionally-alkaline rocks consist of mildlyalkaline olivine basalts, ferrobasalts, very scarce mugearites and benmoreites with abundant feldspar phenocrysts, abundant trachytes and generally minor alkali rhyolites. These e ruptives were erupted within the Gregory rift and within a halfgraben in northern Tanzania, and are everywhere less than 7 Ma in age (Webb & Weaver 1975; Baker & Mitchell 1976; Truckle 1980). The ratio of mafic to silicic rocks varies widely.
Petrology of the Kenya rift
299
TABLE 3. Average analyses o f basalts, phonolites and trachytes 2
3
4
5
6
7
8
9
10
H20PzO 5
45.6 3.8 13.9 2.7 11.2 0.3 5.3 9.9 3.5 1.0 0.2 ND 0.9
46.9 2.7 15.5 4.5 7.5 0.2 5.7 10.8 2.5 0.9 1.8 ND 0.7
59.8 0.7 17.6 5.9 3.1 0.3 0.4 1.7 6.1 4.8 1.6 ND 0.2
55.0 0.5 19.9 2.4 2.3 0.3 0.5 1.4 8.4 5.9 3.2 0.7 0.I
58.2 0.7 16.6 4.0 3.1 0.3 0.6 1.5 6.9 5.2 2.0 1.5 0.I
54.5 0.4 21.2 4.0 0.2 0.2 0.3 1.1 9.4 5.9 2.3 1.0
43.8 2.9 14.2 13.3 ND 0.2 7.3 10.9 3.9 1.5 ND ND 0.6
45.6 2.4 15.6 12.6 ND 0.2 6.9 10.4 3.2 1.3 ND ND 0.6
46.1 2.4 16.0 12.8 ND 0.2 7.1 10.7 3.0 0.8 ND ND 0.4
47.2 3.1 15.1 13.9 ND 0.2 5.7 10.2 3.8 1.0 ND ND 0.7
FeO* Mg
13.7 50.0
11.5 46.9
8.4 8.4
4.4 17.3
6.7 13.0
3.8 13.6
11.9 52.1
11.3 52.0
11.5 52.5
12.5 44.6
1.2 4.1 2.1 9.0 10.0 17.4 27.4
1.2 3.4 2.0 7.6 19.9 25.2 19.2
0.9 3.4 2.0 4.7 23.6 28.5 18.0
1.5 4.3 2.2 5.8 27.7 21.6 20.1
13.4 15.4
15.6 5.8
16.6 2.2
12.8 4.0
1
Major elements (wt.~) SiO2 TiOz A1203 Fe203 FeO MnO MgO CaO Na20 K20
H2O+
Molecular norm (FeO/(Fe203 + FeO) adjusted to 0.85 for all samples) Ap I1 Mt Or Ab An Di Hy O1 Ne Ac
2.0 5.6 3.3 5.9 29.8 20.0 20.0 11.6 1.8
1.5 3.9 2.8 5.7 23.7 29.3 17.4 11.5 4.2
0.4 1.0 1.9 28.4 54.5 6.2 0.8
0.1 0.7 0.5 34.7 26.8
0.1 0.9 1.0 31.2 46.8
0.1 0.6 33.8 22.9
5.2
5.6
4.3
6.9 0.1
2.7 27.7 1.5
4.5 8.2 1.6
2.2 33.7 2.3
1, Average basalt, Emuruangogolak caldera volcano (Weaver 1977); 2, average basalt, Silali caldera volcano (McCall & Hornung 1972); 3, average trachyte, Silali caldera (McCall & Hornung 1972); 4, average plateau Ptype phonolite, alkali basalt suite (Lippard 1973a); 5, average Kenya-type phonolite (Lippard 1973a); 6, average Gwasi G-type phonolite, nephelinite suite (Lippard 1973a); 7, average of 21 Miocene alkali basalts, N Kenya rift (Norry et al. 1980); 8, average of 32 Pliocene alkali basalts, N Kenya rift (Norry et al. 1980); 9, average of 12 Pleistocene basalts, N Kenya rift (Norry et al. 1980); 10, average of 10 Quaternary transitional basalts, N Kenya rift (Norry et al. 1980). ND, no data. Some are basaltic shields with or without extensive flat-lying salic lavas; others are cones and shields c o m p o s e d largely of trachytic lavas and ash-flow tufts on w h i c h late-basaltic satellite eruptions took place from fissures. All the transitionally-alkaline series show m a r k e d D a l y gaps b e t w e e n 4 8 ~ and 5 8 ~ SiO2 (e.g. E m u r u a n g o g o l a k ( W e a v e r 1977)), and, in some, mafic lavas are entirely absent. R a r e occurrences of b e n m o r e i t e lavas are of unusual thickness a n d extent, and are associated with fluid peralkaline trachytes that cover m u c h of the floor of the central and southern part of the rift (Baker & Mitchell 1976). The trachytes range from m e t a l u m i n o u s to peralkaline and are associated with c o m e n d i t i c and pantelleritic rhyolites w h i c h show extreme e n r i c h m e n t of incompatible
trace elements. Large volumes of trachytic ashflow tufts were erupted in the N a i v a s h a - N a k u r u central sector of the rift at about 4 M a BP, and probably c a m e from major caldera structures that were buried u n d e r later lavas. Locally, trachytes and comendites form zones of H o l o c e n e plugdomes in the E b u r r u - N a i v a s h a sector of the central rift and are associated with a geothermal area. T h e c h a i n of Q u a t e r n a r y caldera volcanoes along the axis of the deepest part of the G r e g o r y rift characterizes b a s a l t - t r a c h y t e b i m o d a l volc a n i s m (Scott 1980; Williams et al. 1984). One r e p r e s e n t a t i v e - - M e n e n g a i - - h a s p r o v e d to contain compositionally stratified ash-flow tufts (Leat et al. 1984), and it is likely that m a n y others will be found. Several central volcanoes, including some of
B. H. Baker
300
TABLE 4. Quaternary transitionally-alkaline rocks 1
2
3
4
5
6
7
8
47.5 2.6 14.7 13.9 0.2 5.9 11.6 2.9 1.0 0.5
58.5 1.5 15.9 7.1 0.2 2.2 4.7 5.2 3.9 0.4
61.8 1.0 14.2 7.1 0.3 0.5 1.8 6.2 5.2 0.2
64.8 0.8 13.3 7.3 0.3 0.4 1.2 5.9 5.2 0.1
67.0 0.7 13.7 5.4 0.2 0.4 0.9 5.6 5.2 0.1
60.5 0.7 14.7 8.6 0.4 0.5 1.3 6.9 5.1 0.1
57.1 0.9 16.1 9.0 0.4 0.9 2.6 7.6 5.0 0.2
(ppm) 37 86 45 116 3.0 1.5 2.5 14 453 390 22 20 44 5.4 2.0 0.8 2.3 0.4
33 69 46 140 3.5 1.8 3.1 20 447 472 33 31 39 7.0 2.2 0.9 2.8 0.5
14 ND 17 336 8.9 4.9 9.5 67 282 950 82 126 61 11.2 3.2 1.2 4.3 0.8
7 ND 3 676 15.4 9.6 17.4 116 48 347 135 187 91 16.8 3.3 2.3 8.2 1.4
4 ND 1 1144 28.0 18.4 32.1 159 13 167 185 307 115 21.6 2.6 2.9 10.0 1.8
4 ND 2 1286 36.6 18.7 41.1 181 37 362 203 302 116 24.1 2.7 3.2 12.5 2.0
3 ND 1 1043 26.8 24.5 34.3 166 14 148 218 318 123 24.3 3.5 3.7 ND 2.4
3 ND 3 961 10.2 23.7 35.4 170 48 315 190 293 101 20.6 3.2 2.9 ND 1.9
U Y
ND ND
0.8 32
ND 49
3.3 105
7.5 107
9.8 ND
6.1 131
6.9 111
FeO* Mg
11.4 50.1
12.5 45.8
6.4 38.5
6.4 13.2
6.5 10.5
4.9 12.5
7.7 10.4
8.1 16.7
Major elements (wt.~) SiO2 47.6 TiO2 2.0 A1203 14.8 Fe203 12.7 MnO 0.2 MgO 6.4 CaO 11.5 Na20 2.7 KzO 0.8 P2Os 0.3 Trace elements Sc Ni Co Zr Hf Ta Th Rb Sr Ba La Ce Nd Sm Eu Tb Yb Lu
1, Average of 31 transitionally alkaline Quaternary basalts from the southern rift (Baker, unpublished data); 2, average of 10 O1 Tepesi basalts (Baker et al. 1977); 3, average of 3 benmoreites (Baker et al. 1977); 4, average of 40 trachytes (Plateau and Magadi trachytes) (Baker et al. 1977); 5, average of 6 alkali rhyolites (Baker, unpublished data); 6, average of 39 trachyrhyolites and alkali rhyolites (Limuru trachytes) (Baker, unpublished data); 7, average of 15 trachytes, Suswa pre-caldera shield lavas (Baker, unpublished data); 8, average of 31 trachyphonolites, Suswa post-caldera lavas (Baker, unpublished data). ND, no data. the very high cones such as K i l i m a n j a r o and M o u n t K e n y a , contain at least two of the preceding suites, and are referred to as m i x e d central volcanoes (Fig. 4). K i l i m a n j a r o (Williams 1969b) and Meru (Beloussov et al. 1974) contain nephelinites and alkali basalts and their differentiates, whereas M o u n t K e n y a (Baker 1967), Olorgesailie ( H e n a g e 1976) and the Sattima and Kipipiri centres of the A b e r d a r e R a n g e consist of mildly-alkaline and transitional suites (Shackleton 1945). In the K a v i r o n d o rift and the U g a n d a borderland, M o r o t o (Varne 1968) and T i n d e r e t contain both strongly- and mildly-alkaline suites.
These mixed volcanoes tend to be large steepsided central volcanoes located on the rift flanks, but Olorgesailie is a well-studied example that developed within the Gregory rift ( H e n a g e 1976). Differentiated rocks are subordinate on such volcanoes, but M o u n t K e n y a is an exception; its m a i n cone is constructed of P-type phonolites, succeeded by flank eruptions of alkali basalt a n d trachyphonolite, and later by transitional basalts and trachytes of the I t h a n g u n i centre (Baker 1967). These volcanoes testify to the c o n t e m p o r aneous availability of mafic m a g m a s of a wide range of alkalinity.
P e t r o l o g y o f the K e n y a rift
Petrology and geochemistry Nephelinite--carbonatite suite The compositional fields of several series of the more alkaline rocks is shown on an R1-R2 variation diagram (De la Roche 1980) in Fig. 5. The most strongly alkaline of these is the nephelinite-carbonatite suite (Fig. 5, NE). The suite ranges from melilitites, melanephelinites and ankaratrites to phonolites, and is associated with intrusive complexes of ijolite, nepheline syenite and carbonatite. Nyamaji (NY) in the Kavirondo rift is slightly less alkaline and contains a greater proportion of phonolites (Le Bas 1977). Mount Elgon (EG) on the Uganda border and the Kishalduga shield (KD) on the western side of the southern rift also display trends intermediate between melanephelinite and basanite. Moroto contains distinct nephelinitic and alkali basalt series (Fig. 5, MOa and MOb) (Varne 1968). Representative average analyses of some of these series are given in Table 1 and Table 2, analyses 1-4. The mafic rocks are characterized by nepheline and pyroxene phenocrysts, with or without olivine and/or melilite. Perovskite is present in some lavas, and the phonolites are characterized by accessory sphene, high Sr and Ba, and declining
301
light REE with differentiation. Calcic plagioclase is generally absent in the mode and norm of this suite (Table 1), which accounts for the high Sr and Ba contents of the phonolites. Notable features of the strongly-alkaline suite are the unbroken range of compositions and absence of a Daly gap, subordinate volumes of differentiated rocks, abundance of pyroclastic deposits and a tendency to form large central volcanic piles outside the graben sector of the rift. Le Bas (1977) proposed that the parental magma for the nephelinitic members of the suite is carbonated melanephelinite (Table 1, analysis 6), which gave rise to a carbonatite fraction by liquid immiscibility. The silicate fraction evolved to nephelinite and G-phonolite by fractionation of olivine, pyroxene and sphene. The high contents of incompatible elements are ascribed to a combination of low degrees of partial melting, high pressures and high volatile contents in the source mantle. The phonolitic end-members are distinguished by high Sr and Ba contents owing to the lack of plagioclase fractionation, and are referred to as Gwasi or G-phonolites (Lippard 1973a). Such phonolites are present in subordinate proportions, in contrast with the P-phonolites of the alkali basalt suite (see below). The characteristics of the nephelinite-carbonatite suite are furtherdescribed by Le Bas (1987).
R2
Basanitic and alkali basaltic suite
r --~,
/
,, /
f/~
1--~
NE~-~mel f MO.a I
I
. o/./I.//.
Ii
/ ~:~~
.-2
/
.
" /
. i " .- ? ~ ~ 89
..?A~
(/-. I
-1000
0
I
1000
!
R1
2000
FIG. 5. R1-R2 variation diagram (De La Roche 1980) showing the fields of some strongly-alkaline series : NE, average analyses of the nephelinite-carbonatite suite of the Kavirondo rift (Le Bas 1977) (mel, melilitites; mne, melanephelinites; ij, ijolites; ne, nephelinites; ph, phonolites); NY, Nyamaji suite (Le Bas 1977); MO, Moroto volcano (Varne 1968)(a, nephelinitic series; b, alkali basalt-phonolite series); EG, Elgon volcanics; KD, Kishalduga formation (Crossley & Knight 1981); ab, tb, fields of alkali and transitional basalts; *, points for average compositions given in Table 1.
Basanites and alkali basalts with normative ne greater than 5% are spatially associated with tephrites, ne-mugearites, and exceptionally large volumes of P-phonolite within the Turkana depression and on the flanks and within the northern part of the Gregory rift (Fig. 4). The older mafic series have not been adequately dated but range from 30 Ma in northern Turkana, where they are locally tholeiitic (Bellieni et al. 1981), to 7 Ma, spanning a time range from long before rifting began to the time of the first largescale rift faulting (Baker et al. 1971 ; Williams & Truckle 1980). The mafic lavas frequently carry olivine and augite phenocrysts; plagioclase is less common, and nepheline occurs in the groundmass or is replaced by sodalite. The average Miocene alkali basalt composition is shown in Fig. 6 (Mb), together with the fields of the Miocene Noroyan and Saimo basalts (NO + SO) and associated Pphonolites (Pp). The late Miocene Kaparaina basalts (KA) are less alkaline, and the fields of transitionally-alkaline lavas are shown for comparison (TB, Tr, Tp and Rh). Phonolite lavas cap large areas of the rift
302
B. H. Baker
R2
~
."--.TB~v/
Ch
2000~ .-~ . NO+' " ~ ~ Mo
"9
j,/.y
~#JKn
1000
,~%J-:J ~ I -1000
"
"--"
1000 I
2000 I
R1
FIG. 6. R 1 - R 2 variation diagram of Miocene alkali basalts and phonolites, and of the post-Miocene transitionally-alkalinesuite : NO + SO, Noroyan and Saimo basalts (15 Ma); KA, Kaparaina basalts (5 Ma); Pp, Miocene P-type phonolites; Mb, average Miocene alkali basalt; Pb, average Pliocene alkali basalt; Qb, average Quaternary transitional basalt (Norry et al. 1980) (see Table 3). Average Quaternary basalt of the southern rift (Table 4): TB, Tr, Tp and Rh, fields of transitional basalts, trachyte, trachyphonolites and alkali rhyolites from the southern rift (Baker et al. 1977, and unpublished data) (see Table 4); Kn, late Oligocene tholeiitic and transitional basalts of northern Turkana (Bellieni et al. 1981).
Such occurrences suggest that the basalts and phonolites are genetically related, and imply contrasts in eruptive mechanisms between the rift-floor and rift-flank environments. Other occurrences of series of intermediate alkalinity are in the form of chains of fissurealigned multicentre monogenetic cones and lavas comprising the Chyulu (Fig. 7, CH; Table 2, analysis 8), Nyambeni and Hurri ranges on the eastern rift flank, and large expanses of lavas with maars and small cones in NE Kenya. These fissure-fed lavas show little compositional variation and have only insignificant volumes of differentiated lavas (Saggerson 1963, 1968; Goles 1975). Somewhat more alkaline Miocene lavas in the southern rift are much less voluminous, and consist of localized small limburgite and nephelinite shields (Fig. 5, KD) which are overlain by phonolitic lavas (Crossley 1979; Crossley & Knight 1981).
R2 KI
CH
~SI
2000
NG
~,~
KO
shoulders (Fig. 1), and a greatly expanded succession (up to 2.5 km thick) is found within the north-central part of the Gregory rift, ranging from 16 to 9 Ma in age (Lippard 1973b; Williams & Truckle 1980; Williams e t al. 1983). Phonolite flows up to 270 m thick reach volumes of 300 km 3, and the Yatta phonolite can be traced down the east flank of the rift over a distance of 290 km (Fig. 1, Ya). The phonolites contain alkali feldspar and nepheline phenocrysts, with ferroaugite, apatite and biotite microphenocrysts, and show little variation. They have a similar bulk composition (Table 3, analysis 4) to phonolites of the nephelinite-carbonatite suite (Table 3, analysis 6), but have less Ba and Sr, and light REE abundances which increase with differentiation (Lippard 1973a). Intermediate rocks are scarce among the more voluminous representatives of the suite within the rift, but associations which show gradations from alkali basalt to tephrite, rhomb porphyry and phonolite occur on Mount Kenya and Kilimanjaro on the eastern rift flank, and are described separately below as mixed associations.
...... .... ~ I -1000
KEA .
_ 0
.
0
t
L
.
"
. . . . 1000
2000
. . . .
R1
FIG. 7. R l-R2 variation diagram of Quaternary mixed association volcanoes, Chyulu basalts E of the rift and transitionally-alkalineseries in the southern Kenya rift valley: KI, Kilimanjaro (Williams 1969b); KE, Mount Kenya (Baker 1967, and unpublished data); CH, alkali basalts of the Chyulu range (Goles 1975) (Table 2, analysis 8); SI, NG and KO, Singaraini, Kirikiti and O1 Tepesi transitionally-alkalinebasalts of the southern rift (Baker et al. 1977, and unpublished data) (Table 4, analysis 1); PT, Plateau and Magadi trachytes (Baker et al. 1977; Crossley and Knight 1981) (Table 4, analysis 4); LT, Limuru trachytes (Baker, unpublished data) (Table 4, analysis 6); ME, Menengai (Leat et al. 1984); SU, Suswa volcanics (a, pre-caldera trachytes; b, post-caldera trachyphonolites (Nash et al. 1969; Baker, unpublished data) (Table 4, analyses 7 and 8); HA, Hannington trachyphonolites (Griffiths & Gibson 1980).
P e t r o l o g y o f the K e n y a rift
Transitionally-alkaline series As already noted, the great bulk of transitionallyalkaline basalts, and of the trachytes, trachyphonolites and alkali rhyolites with which they are associated, are less than 7 Ma in age and were erupted within a developing graben. Such rocks are found in strongly bimodal shield volcanoes (Webb & Weaver 1975) as expanses of flood lavas (Baker & Mitchell 1976) and as a chain of Quaternary caldera volcanoes along the axis of the Gregory rift (Macdonald et al. 1970; Leat et al. 1984; Williams et al. 1984). The compositional fields of some of these series are presented in Fig. 7, which shows the pronounced Daly gap between the basalts (SI, NG and KO) and the salic lavas (LT, PT, HA, SU and ME). The trend straddles the critical line of silica saturation, but the salic rocks trend to either the oversaturated (LT) or the undersaturated (SU and HA) side, or extend on both sides (PT and ME). This raises several problems of the transitionally-alkaline suites: the significance of the Daly gap, the occurrence of apparently related rocks of contrasted silica saturation on the same volcano and the occurrence of series that consist largely or entirely of salic rocks. The transitionally-alkaline basalts range from olivine-augite basalts to ferrobasalts with titanomagnetite microphenocrysts (Table 3, analyses 1, 2, 9 and 10; Table 4, analyses 1 and 2). Intermediate lavas are exceedingly rare and include thick benmoreite flows with abundant plagioclase phenocrysts (Baker et al. 1977) (Table 4, analysis 3). Trachyte lavas are well represented and in many series range from mafic trachytes to peralkaline rhyolites (Fig. 7, PT, LT; Table 4, analyses 4-6). Suswa caldera (SU) and the Hannington lavas (HA) exhibit a compositional range from trachyte to trachyphonolite (Nash et al. 1969; Griffiths & Gibson 1980) (Fig. 7; Table 4, analyses 7 and 8). Among the other Quaternary trachytic caldera volcanoes of the axial part of the Gregory graben all but two have pre-caldera shields composed of trachytic and basaltic lavas, overlain by postcaldera trachytic lavas and ash-flow tufts. On several of these volcanoes post-caldera eruptions include basalts erupted from fissures (Williams et al. 1984), testifying to the coexistence of marie and salic magmas throughout much of their eruptive life. The salic rocks on these volcanoes range from metaluminous to peralkaline (Macdonald 1974), and Emuruangogolak (Weaver 1977) shows contrasts in trace-element geochemistry between eruptive phases. Suswa built a pre-caldera shield of trachytic lavas, but post-caldera lavas are
30 3
phonolitic and less differentiated (Fig. 7; Table 4, analyses 7 and 8). A detailed study of Menengai has shown that two major trachytic ash-flow tufts were erupted from a compositionally stratified magma chamber, and the petrogenesis of its trachytic eruptives has been ascribed to a combination of fractional crystallization, magma mixing and unspecified processes of 'liquid-state' differentiation (Leat et al. 1984). The peralkaline salic rocks tend to expel sodium-rich fluid on crystallization, with the result that only obsidians provide true magma compositions. It has been shown that most trace elements are not greatly affected by these processes (Baker & Henage 1977); nevertheless, this effect creates difficulties in mass-balance modelling of differentiation processes.
Mixed volcanic associations Several Quaternary central volcanoes contain rocks of a range of alkalinities; the compositional fields of the Mount Kenya suite (Baker 1967, and unpublished data) and Kilimanjaro (Williams 1978) are shown in Fig. 7 (KI and KE). Both volcanoes were built on the eastern rift flank, and their compositions contrast markedly with those of contemporaneous Quaternary lavas erupted within the rift graben. Mount Kenya consists of a large pile of Pphonolite and rhomb porphyry lavas, with latestage fissure-fed alkali basalts and trachyphonolites erupted on its flanks. Subsequent parasitic volcanism erupted mixed comendite-basalt ashflows and built a trachytic lava cone (Baker 1967; Rock 1976a). The main cone of the volcano contains the alkali basalt-P-phonolite composition range, and satellite eruptions were of basalt, trachyte and rhyolite (not shown in Fig. 7, but equal to the NG and PT fields). Several alkali rhyolite dykes contain xenoliths of quenched basalt representing partial mixing of the two magmas. Kilimanjaro is composed of rocks of a wide range of alkalinity, such as nephelinites, ankaratrites, alkali basalts and phonolites (Fig. 7, KI) (Williams 1969b). Less-well-known volcanoes that also contain mixed associations are Meru, Tinderet and Sattima, and all are large central cones situated outside the Gregory rift (Williams 1969a). One central volcano which contains the whole alkaline spectrum from nephelinites to hypersthene normative basalts is Olorgesailie (Henage 1976), which is a 3.2 Ma old volcano on the floor of the southern Gregory rift. The eruptive sequence shows four stages of contrasting alkalinity: 1, alkali basalt, mugearite and benmoreite;
304
B. H. Baker
2, transitional basalt (hypersthene normative) and mugearite; 3, weakly-alkaline benmoreite, trachyte and trachyphonolite; 4, mixed alkali basalt, nephelinite and G-phonolite. This suggests episodic changes in the nature of the magma being supplied to the volcano. Differentiation within each of the stages can be modelled by fractionation of phenocrysts (Henage 1976). Mixed association volcanoes are of a wide age range and tend to build large cones, and most are situated on the rift flanks. The occurrence of some within the rift, but probably pre-dating its formation, suggests that tapping of magmas of contrasted alkalinities was favoured around the margins of the rift and on its flanks.
Trace-element characterization of volcanic suites and series Few systematic studies of the trace-element geochemistry of volcanic series have been performed. Many of the existing data have been used to address two questions. First, are the mafic and salic members of bimodal transitionallyalkaline series co-magmatic? Second, how were the salic rocks generated (Sceal & Weaver 1971 ; Weaver et al. 1971 ; Weaver 1977)? Contrasts between suites of differing alkalinity are shown by the differing behaviour of Rb, Sr and Ba (Fig. 8). The alkali basalt-phonolite trend (AB-PHO) has a small depletion of Sr and Ba by comparison with the transitionally-alkaline trend
Rb
Sr
Ba
FIG. 8. Ternary variation diagram showing contrasts in the trends of Rb enrichment and the degree of Sr and Ba depletion in rocks of the alkali basaltphonolite ( ; AB, PHO) and transitional basalttrachyte-rhyolite suite ( - - - ; TB, TRA, RHY). The range of the Laacher See phonolitic tephra (LZ) is shown for comparison (W6rner et al. 1983).
Hf
la FIG. 9. Ternary variation diagram showing differences in the Ta/Hf ratio for selected alkaline suites: OS, Olokisalie (Henage 1976); PT, O1 Tepesi basalts and Plateau trachytes (Baker et al. 1977); L, Limuru trachytes and alkali rhyolites (Baker, unpublished data); K, alkali basalts and P-phonolites of Mount Kenya (Baker, unpublished data). A selection of ranges of compositionallyzoned calc-alkaline magmas (MZ, KZ, T, BZ) is shown for comparison (Baker & McBirney 1985) together with the range of the Laacher See tephra (LZ) (W6rner et al. 1983).
in which Sr and Ba were greatly depleted by fractionation of plagioclase and alkali feldspar ( T B - T R A - R H Y ; see Table 4, analyses 2-5). The incompatible trace elements in the alkaline series (Zr, Hf, Nb, Ta, Th, U, Rb and the light REE) are greatly enriched in salic rocks and exhibit linear covariance. This has been ascribed to their having low bulk distribution coefficients owing to the absence of accessory phases that could deplete such elements (Sceal & Weaver 1971; Weaver et al. 1972; Barberi et al. 1975; Weaver 1977; Baker et al. 1977, 1978). Zr/Nb ratios for each volcanic series tend to be constant and have distinctive values. In bimodal series this has been used to argue that the mafic and silicic members are co-magmatic and related by crystal-liquid fractionation (Weaver et al. 1972), but this conclusion is permissive rather than required because the contents of these elements in basaltic lavas have not been determined with sufficient precision. Nevertheless it is necessary to try to demonstrate a co-magmatic relationship in a series of rocks before petrogenetic modelling is performed. The characteristics outlined above extend to Ta, Hf, Th and the light REE except Eu (Baker et al. 1977) and are similar to those of comparable oceanic series (Baker et aL 1978). The proportions of Hf, Ta and Th in several Kenyan suites vary little (Fig. 9, PT, O, L and K); Th is the most
Petrology of the Kenya rift incompatible element owing to its combination of high charge and large ionic radius. The suites show distinctive Hf/Ta ratios, suggesting comagmatic relationships over a wide range of compositions. The comparable behaviour of a selection of calc-alkaline compositionally stratified magma bodies suggests that the petrogenetic processes were similar (Baker & McBirney 1985) (Fig. 9, MZ, KZ, T, L and BZ). In contrast the zoned Laacher See phonolitic tephra shows Ta depletion owing to the crystallization of sphene (W6rner e t al. 1983; Wolff & Storey 1984) 9 A study of the Emuruangogolak Quaternary caldera volcano showed that trachytes erupted between phases of explosive activity and caldera collapse have distinctive Zr/Nb ratios which are thought to be the effects of volatile degassing (Weaver 1977). Much more complex contrasts in trace-element variance were found in the eruptive phases of Menengai caldera volcano (Leat e t al. 1984) and were ascribed to a variety of processes including fractionation, magma mixing and 'liquid-state' processes. Among the more strongly-alkaline phonolites the linear covariance of incompatible trace elements tends to break down owing to the presence of minor phases that incorporate such elements. Lippard (1973a) has shown that the light REE are depleted in Gwasi-type phonolites whereas they are enriched in P-type phonolites and (Kenya-type) trachyphonolites. Similar effects, including depletion of Nb and Ta, have been noted in compositionally stratified phonolitic tufts and ascribed to fractionation of one or another of sphene, apatite, allanite, perovskite or
z z m <
N e,-o
1
g
BaR
Th K
Ta L a C o S r
P SmZrHf
305
salic rocks
2 =1;
,ll,
Z <
~ R!g!F]! $2 (3
q
I l l l l l l Ba RbTh K N b T a L a C e S r
I
I
I
I
P S m Z r Hf Ti Tb Y Yb
FIG. 11. Incompatible-trace-element abundances normalized to primitive-mantle abundances for the transitionally-alkaline O1 Tepesi basalts, benmoreites, plateau trachytes and alkali rhyolites of the southern rift valley (Baker et al. 1977). The vertical broken lines show the compositional range of zoned ash-flow tufts of Menengai (Leat et al. 1984).
melanite (Wolff 1983; W6rner et al. 1983; Wolff & Storey 1984) 9Rb and Th tend to be enriched in all alkaline suites and provide the most reliable indicators of increasing differentiation. The abundances of incompatible trace elements of two representative suites of differing alkalinity normalized to primitive-mantle abundances are shown in Figs 10 and 11. The alkali basalts of Mount Kenya (Fig. 10) are richer in large-ion elements than are the transitionallyalkaline basalts of the rift floor (Fig. 11). The latter show the extent of the Daly gap and t h e great range of enrichments owing to protracted fractionation. The compositional range of one of the zoned ash-flow tufts from Menengai volcano is nearly as great as that of a suite oflavas erupted over a much greater time interval. Similarly, the compositional range of the Laacher See zoned tephra is of the same order as that of all the salic rocks of Mount Kenya. These features suggest that compositional variation in the series and in single compositionally graded eruptives was generated by essentially similar processes of crystal-liquid fractionation.
Ti Tb Y Yb
FIG. 10. Incompatible-element abundances normalized against primitive-mantle abundances for the Mount Kenya suite (Baker 1967, and unpublished data. The compositional range of the Laacher See phonolitic tephra (W6rner et al. 1983) is shown by arrows.
Isotope ratios Isotope ratios have been used to study the comagmatism of Kenyan alkaline series, the degree to which interaction with the crust was involved in the formation or modification of salic rocks
3o6
B. H. Baker
and the nature of the source mantle (Bell & Powell 1970, 1974; Bell et al. 1973; Rock 1976b). Most of the data are for strontium, but some for Sm and Pb isotopes have become available (Norry et al. 1980). Initial Sr isotope ratios for Kenyan stronglyalkaline rocks fall into the range 0.7040-0.7055, comparable with those of oceanic island basalts, and there is no correlation with Rb/Sr ratios. This suggests that crustal involvement has been negligible in the formation of ijolites, nepheline syenites, phonolites and carbonatites (Rock 1976b). A study of Sr, Nd and Pb isotopes of Miocene to Recent mildly-alkaline to transitionally-alkaline lavas of the northern rift shows that they also have ratios similar to those of oceanic islands (Norry et al. 1980). A phonolite and several trachytes of the bimodal caldera volcano Emuruangogolak show the same ratios as those of their associated basaltic lavas. The correlation between enrichment of light REE, the degree of undersaturation and abundances of incompatible elements was explained by metasomatic enrichment of previously depleted mantle by a CO2rich fluid shortly before generation of the parental magmas (Norry et al. 1980). The effects of mantle inhomogeneity, differences in the amount and proportions of water and CO2, degrees of partial melting, pressure differences and zone refining have not been satisfactorily resolved. Unpublished Sr isotope data for several basalttrachyte-rhyolite series of the southern rift also show oceanic island ranges, suggesting that there was little or no crustal involvement in their formation (Baker, unpublished data). Consequently the exceptionally large proportions of salic lavas cannot be due to assimilation of crustal rocks, but the occurrence of oversaturated and undersaturated lavas in close association on the same volcano could be influenced by open-system processes such as the adsorption of water from crustal xenoliths.
Influences of magma chamber processes on differentiation Recent work has suggested that a variety of fluiddynamical processes within sub-volcanic magma chambers can influence the course of differentiation and eruption of magmas. These depend on the development of inhomogeneities of temperature, composition and magma density as a result of cooling and crystallization, and as a result of periodic replenishments by parental magma (McBirney et al. 1985).
In magmas in which the density of the liquid falls with progressive cooling and crystallization there is a possibility for buoyant ascent of residual liquid adjacent to the wall of a magma chamber. Fluid-dynamical calculations suggest that this is possible even if the magma has a high viscosity, and can lead to accumulation of compositionally zoned strongly differentiated liquids at the top of a chamber at rates great enough to account for the volumes and periodicities of successive eruptions of salic magmas (Nilson et al. 1985). Ascent of such boundary-layer liquids should result in shear forces that aid segregation of phenocrysts through the Bagnold effect, and may also result in some mixing with the interior liquids through which the layer ascends (Thompson & McBirney 1985). By such processes differentiated liquid can become segregated as a compositionally stratified gravitationally-stable upper zone, separated from a homogeneous convecting more mafic lowerzone magma. The dominant process in each zone is crystal-liquid fractionation, which takes place by crystal settling in the lower zone and by removal of residual liquid from crystals growing on the walls to generate the upper zone. The course of differentiation may be modified by either periodic eruptions from the upper zone or replenishments of the lower zone (Baker & McBirney 1985). Magmas erupted from a stratified magma chamber would normally be drawn from the top and could give rise to compositionally 'inverted' gradation in ash-flow tufts or frequently erupted lavas. If there is a large downward decrease in viscosity in the erupted volume the lower liquid may overtake the upper during flow through a conduit, resulting in reversal of the usual eruptive sequence (Koyaguchi 1985). The compositions of erupted magmas therefore depend principally on relations between boundary-layer segregation rates and the rates and volumes of individual eruptions. The processes outlined could account for the bimodality and preferential eruption of differentiated magmas that characterize some alkaline volcanic suites. The estimated liquidus densities of representative series of alkaline rocks from Kenya are plotted against their liquidus temperatures in Fig. 12. Uncertainties in the absolute densities are not significant in the context of the following discussion since it is the relative densities along a liquid line of descent that are important for buoyancy effects. The melilitite-melanephelinite-nephelinite-G-phonolite trend shows continuously-falling liquid densities through the whole composition range (Fig. 12(a)). The alkali basalttephrite range shows a much less steep drop in
Petrology of the Kenya rift
3o7
NE
KDw C"
2-8
MO
. . . . ~!.
~'~..~
c.
~_
_ fb
..... -:Sy"c~ . J ~ PA."'~A, 't ~lt" v =..,,-= 9
mne
-~.
2-7
~~\/.
2.6
2"5
-
a
b
( _ ~ - J ph
Z LIJ
r
2-4
=
2"3
I 1200
I
I
I
1000
~C (a)
I 1200
I
I 1000
~C (b)
FIG. 12. Calculated anhydrous liquid densities plotted against estimated liquidus temperatures for suites of varying alkalinity: (a) strongly-alkaline suites (NE, melilitites (me), melanephelinites (mne), nephelinites (ne) and phonolites (ph) of the nephelinite-carbonatite association in the Kavirondo rift valley (Le Bas 1977); KD, limburgites, melanephelinites and nephelinites of the K ishalduga formation, southern rift valley (Crossley & Knight 1981); MO, alkali basalt series of Moroto, Uganda (Varne 1968); (b) mildly- and transitionally-alkaline suites (CH, Chyulu alkali basalts, eastern rift flank (Goles 1975); EM, Emuruangogolak basalt-trachyte suite (Weaver 1977); PA, Paka basalt-trachyte suite; KR, Kirikiti basalts, southern rift (Baker et al. 1977, and unpublished data); OT, O1 Tepesi basalts, benmoreites, trachytes and alkali rhyolites of the southern rift (Baker et al. 1977, and unpublished data); fb, ferrobasalt; be, benmoreite; tr, trachyte ; rh, rhyolite).
densities (Fig. 12(a)), followed by the same density drop to phonolitic compositions as is seen in the nephelinitic series. Alkali basalts show little variation in density (Fig. 12(b)). By contrast transitional basalts have a density minimum, and densities rise to a 'density hump' at ferrobasalt compositions, after which there is a rapid decrelse in density to trachyte and rhyolite (Fig. 12(b)). The appearance of a 'density hump' as alkalinity decreases is due to the increasing role of plagioclase fractionation and iron enrichment as alkalinity decreases. The steepness of the density decrease after the density hump is caused by strong depletion of iron owing to fractionation of titanomagnetite, and this is most marked in the transitional suite (Baker et al. 1977). The negative slope of the density-temperature curve in strongly-alkaline suites should lead to efficient liquid fractionation and to early establishment of a capping zone of differentiated (phonolitic) magma. The fact that bimodality and eruption of disproportionate volumes of salic magma are not features of such suites may be caused by the explosive behaviour of nephelinitic volcanoes, which would disrupt the formation of stratification. Explosive eruptions are uncommon
among alkali basalt-phonolite suites, and phonolitic upper zones could develop and, once formed, would inhibit eruption of underlying basaltic magma. For the transitionally-alkaline suites characterized by a density hump, the processes described would lead to marked bimodality. The density minimum in the basalt composition range represents the composition most readily erupted. More mafic magmas supplied by replenishment or more fractionated magmas formed by crystallization would sink. In the basalt-ferrobasalt range, buoyant ascent of differentiated liquid would be impossible or would be overwhelmed by the down-flow due to thermal contraction (Nilson et al. 1985). As long as the supply of replenishing magma was large enough to offset solidification the magma chamber could not differentiate past the density hump and basalt eruptions would take place. Without replenishment, however, the magma would evolve rapidly to the density maximum without eruptions, and the first eruptions would be trachytic. If replenishment were to slow or cease, the magma would reach the density hump and efficient liquid fractionation could begin. The range of liquidus temperatures of intermediate
3o8
B. H . B a k e r
magmas is small (Fig. 12(b)), and the ascending boundary layer would contain a range of compositions, the most differentiated of which would rise to the top of the chamber by the 'filling-box' mechanism, rapidly establishing a capping layer of differentiated magma which would be available for eruption over the remainder of the cooling time of the chamber. These processes would lead to marked bimodality and preferential eruption of salic magmas. If the uppermost magma reaches volatile saturation it could erupt a compositionally graded ash-flow tuff followed by caldera collapse, and several cycles of segregation, eruption and collapse could take place, drawing on a chamber in which changing distributions and compositions must give rise to contrasts in the products of each eruption (Mahood 1984).
zo"i0................... :ii;il
replenishment . . . . . .
........................................
,~
(a)
t Z 0i
Rh
UZ
I-- Tra Z LU nr I.g I.I. IJ.
Bas
,;-.R
..!,. cc
E E E
JRR TIME
(b) FIG. 13. Schematic representation of the effects of liquid fractionation: (a) a compositionally zoned magma chamber formed by segregation of buoyant residual liquid (replenishment of the chamber is approximately balanced by eruption from a central conduit or, in the later stages, by formation of fissures and flank eruptions; (b) an idealized eruptive sequence from a stratified magma chamber (UZ, differentiated magmas derived from the upper zone; LZ, mafic magmas derived from the lower zone; E, times of eruption; CC, time of caldera collapse; R, times of replenishment of the lower zone by mafic magma). See text for explanation.
Eventually cooling and solidification, aided perhaps by volatile loss and eruptions from the top of the chamber, would inhibit eruptions from the upper zone. The remaining mafic magma in the lower zone would be basaltic and could not penetrate upward unless the upper zone were effectively solid. However, replenishment by mafic magma could result in pressurization of the magma chamber and injection into tensile fissures in its walls, giving rise to basaltic satellite eruptions during the closing stages of volcanic activity. Such a sequence of events has been observed on several Quaternary caldera volcanoes (Williams et al. 1984), e.g. Emuruangogolak (Weaver 1977). A diagram of such a magma chamber is shown in Fig 13(a), accompanied by a schematic representation of its compositional evolution (Fig. 13(b)). During the basaltic stage (Fig. 13(b)) no segregation of a differentiated upper-zone liquid is possible, and compositional variations will be governed by progressive fractionation, offset by periodic replenishments of more primitive magma which are likely to trigger eruptions at the surface. Once the magma has reached the ferrobasalt density hump an upper zone develops rapidly and periodic eruptions will tap the most differentiated liquid. If magma under the roof becomes volatile saturated through crystallization or by adsorption of water from hydrous wallrocks and xenoliths, the magma can erupt explosively and may cause caldera collapse. Each major eruptive phase removes the most differentiated liquid and causes repeated overlaps in the range of compositions erupted. Flank basalt eruptions could take place at any stage if the chamber is pressurized enough to rupture its walls or if the upper zone has solidified. The overall compositional trend would be determined by the balance between the liquid fractionation rate and the eruption rates, and may be towards more or less differentiated magmas. Such processes may be further complicated by back-mixing of ascending compositional boundary-layer magma with the convecting main body of mafic magma or with the deeper parts of a stratified layer of segregated differentiated liquid, and could have generated the mixed trachytehawaiite lavas of Mount Longonot (Jones 1979). Periodic replenishment would cause changes in the composition of lower-zone magma, which would be reflected in minor changes in the compositions of erupted differentiated magmas. On many older basalt-trachyte volcanoes early shield-building eruptions were of basalt followed by trachyte, culminating in pyroclastic eruptions (Truckle 1977). On the Quaternary basalt-trachyte shield volcanoes of the axial zone of the
Petrology of the Kenya rift Kenya rift valley, however, the shield-building stage was dominated by trachytic eruptions, accompanied by some basalt on three of them. Caldera-collapse phases were accompanied by eruption of variable volumes of tufts and ash-flow tufts, followed by post-caldera trachyte lavas and tufts. Late basalts were erupted on four volcanoes, largely from fissures (Williams et al. 1984). The variable behaviour of these volcanoes could be explained by variations in the timing and volumes of magma replenishments and by the magnitudes of extensional tectonic stresses which determine when and where the magma chamber will rupture. Compositionally zoned ash-flow tufts such as those of Menengai (Leat et al. 1984) have been found on many other alkaline volcanoes (Mahood 1980; Wolff & Storey 1980), suggesting that liquid-fractionation processes may play a central role in the formation of salic alkaline magmas.
Summary and conclusions The Kenya and northern Tanzania alkaline province contains 220 000 km 3 of sodic alkaline rocks ranging from melilitites and melanephelinites to transitionally-alkaline basalts and their differentiates. Individual volcanoes display associations of three degrees of undersaturation characterized, in part, by the nature of their differentiated rocks, which are G-phonolites, Pphonolites, and trachyphonolites, trachytes and alkali rhyolites. These can be referred to as the nephelinitic, alkali basaltic and transitional basalt suites. Several large volcanoes contain more than one of these suites, implying that parental magmas of variable alkalinity were available at certain times and places. A general decrease in alkalinity with time is detectable in the rift zone, and for any time period there was a tendency for the least alkaline magmas to be erupted within the central
309
and deepest part of the rift zone. This relationship has been interpreted as being due to progressive ascent of a zone of melting under the rift zone (Wendlandt & Morgan 1982), but the occurrence of mixed-alkalinity volcanoes suggests that other factors such as mantle inhomogeneity and metasomatic processes complicate this simple model. The geochemical variation of the suites is compatible with processes dominated by lowpressure crystal-liquid fractionation. Large enrichments of incompatible trace elements and depletions of Mg, Sr and Ba suggest that extended fractionation gave rise to salic differentiates. The disproportionately large volumes of salic rocks, the strong bimodality of the less alkaline suites and the occurrence of compositionally graded ash-flow tufts suggest that liquid fractionation caused upward segregation of phonolitic and trachytic magmas which were erupted in preference to more mafic magmas. Large-scale contamination by continental crust is not permitted by the isotope data. Much further work remains to be done. The volcanic history and petrology of much of Turkana and of the extensive field of basaltic lavas of NE Kenya is little known. The huge Aberdare volcanic complex, the greater part of which is exposed along a major fault escarpment, is known in bare outline. Few systematic studies have been made of the many varied volcanoes of northern Tanzania. Modern high-precision traceelement and isotopic data are still scanty. The province is associated with a well-defined rift structure which invites examination of relations between its tectonic and volcanic evolution, but variation in space and time of mafic magmas is still little known. Petrogenetic processes are so complex that unsystematic or unrepresentative sampling is not likely to yield useful new insights. The need to trace the evolution of a magma in time requires complete sampling of volcanic series and appreciation of the physical factors that govern the petrogenesis and eruption of magmas.
References BAKER, B. H. 1967. Geology of the Mt. Kenya area.
and geochronology of the Kedong-Olorgesailie area and the evolution of the south Kenya rift valley. J. geol. Soc. Lond. 132, 467-84.
Geol. Surv. Kenya Rep. 42.
--
-
-
--
& HENAGE,L. H. 1977. Compositional changes during crystallization of some peralkaline silicic lavas of the Kenya rift valley. J. Volcanol. geotherm. Res. 2, 17-28. & MCBIRNEY, A. R. 1985. Liquid fractionation: Pt. III--Geochemistry of zoned magmas and the compositional effects of liquid fractionation. J. Volcanol. geotherm. Res. 24, 55-81. & MITCHELL,J. G. 1976. Volcanic stratigraphy
--,
WILLIAMS, L. A. J., MILLER, J. A. & FITCH, F. J.
1971. Sequence and geochronology of the Kenya rift volcanics. Tectonophysics, ll, 191-215. --,
CROSSLEY,R. & GOLES, G. G. 1978. Tectonic and
magmatic evolution of the southern part of the Kenya rift valley. In: NEUMANN, E.-R. & RAMBERG, I. B. (eds) Petrology and Geochemistry of Continental Rifts, pp. 29-50. Reidel, Dordrecht.
3 lO
B . H . Baker
from Kenya and their bearing on the Daly Gap. GOLES, G. G., LEEMAN, W. P. & LINDSTROM, Geol. Mag. 116, 487-9. M. M. 1977. Geochemistry and petrogenesis of a basalt-benmoreite-trachyte suite from the south- KING, B. C. 1965. Petrogenesis of the alkaline igneous rock suites of the volcanic and intrusive centres of ern part of the Gregory Rift, Kenya. Contrib. Eastern Uganda. J. Petrol. 6, 67-100. Mineral. Petrol. 64, 303-32. 1970. Vulcanicity and rift tectonics in East Africa. BARBERI, F., SANTACROCE,R., FERRARA, G., TREUIL, In: CLIFFORD, T. N. & GASS, I. G. (eds) African M. & VARET, J. 1975. A transitional basaltMagmatism and Tectonics, pp. 263-83. Oliver & pantellerite sequence of fractional crystallization, Boyd, Edinburgh. the Boina centre (Afar rift, Ethiopia). J. Petrol. 16, 1978. Structural and volcanic evolution of the 22-56. Gregory Rift Valley. In: W. W. BISHOP (ed.) BELL, K. & POWELL, J. L. 1970. Sr isotopic studies of Geological Background to Fossil Man. Scottish alkalic rocks: the alkalic complexes of East Academic Press, Edinburgh. Uganda. Geol. Soc. Am. Bull. 81, 3481-90. & CHAPMAN,G. R. 1972. Volcanism of the Kenya & 1974. Isotopic composition of strontium in Rift Valley. Phil. Trans. R. Soc. Lond., Ser. A, 271, alkaline rocks. In: SORENSEN,H. (ed.) The Alkaline Rocks, pp. 412-21. Wiley, New York. LE BAS, M. & SUTHERLAND, D. M. 1972. The , DAWSON, J, B. & FARQUHAR, R. M. 1973. - - , history of the alkaline volcanoes and intrusive Strontium isotope studies of aikalic rocks: the complexes of eastern Uganda and western Kenya. active carbonatite volcano Oldoinyo Lengai, TanJ. geol. Soc. Lond. 128, 173-205. zania. Geol. Soc. Am. Bull. 84, 1019-30. BELLIENI, G., JUSTIN VISENTIN, E., ZANETTIN, B. & KOYAGUCHI,T. 1984. Volcanic rocks from the Samburu Hills, northern Kenya. Afr. Stud. Monogr., Suppl. PICCIRILLO, E. M. 1981. Oligocene transitional (2), 147-79. tholeiitic magmatism in northern Turkana, 1985. Magma mixing in a conduit. J. Volcanol. Kenya: comparison with the coeval Ethiopian geotherm. Res. 25, 365-9. volcanism. Bull. Volcanol. 44, 411-27. BELOUSSOV,V. V., GERASIMOVSKY,V. I., GORYATCHEV, LEAT, P. T., MACDONALD, R. & SMITH, R. L. 1984. Geochemical evolution of the Menengai caldera, A. V., DOBROVOLSKY, V. V., KAWTSA, A. P., Kenya. J. geophys. Res. 89, 8571-92. LOGATCHEV, N. A., MILANOVSKY, E. E., POLYLE BAS, M. J. 1973. A norm for feldspathoidal and AKOV, A. I., RYKUNOV,L. N. & SEDOV,V. V. 1974. melilitic igneous rocks. J. Geol. 81, 89-96. The East African Rift System, III--Geochemistry, -1977. Carbonatite-Nephelinite Volcanism, 347 pp. Seismology, Main Results, 288 pp. Nauka, Moscow. Wiley, New York. CROSSLEY, R. 1979. The Cenozoic stratigraphy and 1987. Nephelinites and carbonatites. In: FITTON, structure of the western part of the rift valley in - J. G. & UPTON, B. G. J. (eds) Alkaline Igneous southern Kenya. J. geol. Soc. Lond. 136, 393-405. Rocks, Geol. Soc. Spec. Publ. 30, pp. 53-83. - - & KNIGHT, R. M. 1981. Volcanism in the western LIPPARD, S. J., 1973a. The petrology of phonolites from part of the rift valley in southern Kenya. Bull. the Kenya rift valley. Lithos, 6, 217-34. Volcanol. 44, 117-28. 1973b, Plateau phonolite lava flows, Kenya. Geol. DAVIES, K. A. 1952. The building of Mount Elgon (East Mag. 6, 543-9. Africa). Mere. geol. Surv. Uganda 7. MCBIRNEY, A. R., BAKER, B. H. & NILSON, R. H. 1985. DAWSON, J. B. 1962. The geology of Oldoinyo Lengai. Liquid fractionation. Part I: basic principles and Bull. Volcanol. 24, 349-87. experimental simulations. J. Volcanol. geotherm. -1964. Carbonate tuff cones in northern TanganRes. 24, 1-23. yika. Geol. Mag. 101, 129-37. MCCALL, G. J. H. & HORNUNG, G. 1972. A geochemical -& POWELL, D. G. 1969. The Natron-Engaruka study of Silali volcano, Kenya, with special explosion crater area, Northern Tanzania. Bull. reference to the origin of the intermediate-acid Volcanol. 33, 791-817. eruptives of the central rift valley. Tectonophysics, DEANS, T. & ROBERTS, B. 1984. Carbonatite tufts and 15, 97-113. lava clasts of Tinderet foothills, western Kenya: a MACDONALD, R. 1974. Nomenclature and petrochemstudy of calcified natrocarbonatites. J. geol. Soc. istry of the peralkaline oversaturated extrusive Lond. 141,563-80. rocks. Bull. Volcanol. 38, 498-516. DE LA ROCHE, H., LETERRIER,J., GRANDCLAUDE,P. & BAILEY, D. K. & SUTHERLAND, D. S. 1970. MARCHaL, M. 1980. A classification of volcanic Oversaturated peralkaline glassy trachytes from and plutonic rocks using R1R2-diagram and majorKenya. J. Petrol. 11,507-17. element analyses--its relationships with current MAHOOD, G. 1984. Pyroclastic rocks and calderas nomenclature. Chem. Geol. 29, 183 210. associated with strongly peralkaline magmatism. GOLES, G. G., 1975. Basalts of unusual composition J. geophys. Res. 89, 8540-52. from the Chyulu Hills, Kenya. Lithos, 8, 47-58. MASON, P. 1955. Geology of the Meru-Isiolo area. Geol. GRIFFITHS, P. S. & GmSON, I. S. 1980. The geology and Sure. Kenya Rep. 31. petrology of the Hannington trachyphonolite forMIYASHIRO, A. 1978. Nature of alkalic volcanic rock mation, Kenya Rift Valley. Lithos, 13, 43-53. series. Contrib. Mineral. Petrol. 66, 91-104. HENAGE, F. H. 1976. Geology and compositional NASH, W. P., CARMICHAEL,I. S. E. & JOHNSON, R. W. evolution of Mt. Olokisalie, Kenya. PhD Thesis, 1969. The mineralogy and petrology of Mount University of Oregon (unpublished). Suswa, Kenya. J. Petrol. 20, 268-94. JONES, W. B. 1979. Mixed benmoreite/trachyte flows --,
-
-
-
-
185
-
-
-
-
- 2 0 8 .
Petrology of the Kenya rift NILSON, R. H., MCBIRNEY, A. R. & BAKER, B. H. 1985. Liquid fractionation. Part II: fluid dynamics and quantitative implications for magmatic systems. J. Volcanol. geotherm. Res. 24, 25-54. NORRY, M. J., TRUCKLE, P. H., LIPPARD, S. J., HAWKESWORTH, C. J., WEAVER, S. D . & MARRINER, G. F. 1980. Isotopic and trace element evidence from lavas, bearing on mantle heterogeneity beneath Kenya. Phil. Trans. R. Soc. Lond., Ser. A, 297, 259-71. ROCK, N. M. S. 1976a. Petrogenetic significance of some new xenolithic alkaline rocks from East Africa. Mineral. Mag. 40, 611-25. 1976b. The comparative strontium isotopic composition of alkaline rocks: new data from southern Portugal and East Africa. Contrib. Mineral. Petrol 56, 205-28. SAGGERSON,E. P. 1963. Geology of the Simba-K ibwezi area. Geol. Surv. Kenya Rep. 58. 1968. Eclogites associated with alkali olivine basalts, Kenya. Geol. Rundsch. 57, 890-903. - 1970. The structural control and genesis of alkaline rocks in central Kenya. Bull. Volcanol. 34, 38-76. & BAKER, B. H. 1965. Post-Jurassic erosion surfaces in eastern Kenya and their deformation in relation to rill structure. Q. J. geol. Soc. Lond. 121, 51-72. --• WILLIAMS, U A. J. 1964. Ngurumanite from southern Kenya and its bearing on the origin of rocks in the northern Tanganyika alkaline district. J. Petrol. 5, 40-81. SCEAL, J. S. C. & WEAVER, S. D. 1971. Trace-element data bearing on the origin of salic rocks from the Quaternary volcano Paka, Gregory Rift, Kenya. Earth planet. Sci. Lett. 12, 327-31. ScoTr, S. C. 1980. The geology of Longonot volcano, central Kenya: a question of volumes. Phil. Trans. R. Soc. Lond., Ser. A, 296, 437-65. SHACKLETON, R. M. 1945. Geology of the Nyeri area. Geol. Surv. Kenya Rep. 12. THOMPSON, M. E. & MCBIRNEY, A. R. 1985. Redistribution of phenocrysts by convective flow in a viscous boundary layer. J. Volcano/. geotherm. Res. 24, 83-94. TRUCKLE, P. H. 1977. The geology of the area to the south of Lokori, South Turkana, Kenya. PhD Thesis, Univers!ty of Leicester (unpublished). 1980. Variation of basic lava petrology of the Kenya rift valley. In: Geodynamic Evolution of the A/ro-Arabic Ri/'t System (CARRELLI, A., president), pp. 133-42. Accademia Nazionale dei Lincei, Rome. VARNE, R. 1968. The petrology of Moroto Mountain, Eastern Uganda, and the origin ofnephelinites. J. Petrol. 9, 16990. WEAVER, S. D. 1977. The Quaternary caldera volcano
311
Emuruangogolak, Kenya rift, and the petrology of a bimodal ferrobasalt-pantelleritic trachyte association. Bull. Volcanol. 40, 209-30. - - - , SCEAL, J. S. C. & GIBSON, I. L. 1972. Traceelement data relevant to the origin of trachytic and pantelleritic lava in the East African rift system. Contrib. Mineral. Petrol. 36, 181-94. WEBB, P. K. & WEAVER, S. D. 1975. Trachyte shield volcanoes: a new volcanic form from South Turkana. Bull. Volcanol. 39, 294-312. WENDLANDT, R. F. & MORGAN P. 1982. Lithosphere thinning associated with rifting in East Africa. Nature, Lond. 298, 734-6. WILLIAMS, L. A. J. 1969a. Volcanic associations in the Gregory Rift Valley, East Africa. Nature, Lond. 224, 6 1 - 4 .
1969b. Geochemistry and petrogenesis of the Kilimanjaro volcanic rocks of the Amboseli area, Kenya. Bull. Volcanol. 33, 862-88. -1970. The volcanics of the Gregory Rift Valley, East Africa. Bull. Voleanol. 34, 439-65. ...... 1972. The Kenya rift volcanics: a note on volumes and chemical composition. Tectonophysics, 15, 8396. 1978. The volcanological development of the Kenya rift. In: NEUMANN,E. R. & RAMBERG, I. B. (eds) Petrology" and Geochemistry of Conthlental Rifts, pp. 101-21. Reidel, Dordrecht. . . . . 1982. Physical aspects of magmatism in Continental Rifts. In: PALMASON, G. (ed.) Continental and Oceanic R(fts, pp. 193-222. American Geophysical Union, Washington. - - & TRUCKLE, P. H. 1980. Volcanic sequences in the Kenya Rift. In: Geodynamic Evolution of the Afro-Arabic Rift System (CARRELLI, A., president), pp. 123 32. Accademia Nazionale dei Lincei, Rome. , MACDONALD, R. & CHAPMAN, G. R. 1984. Late Quaternary volcanoes of the Kenya rift valley. J. geophys. Res. 89, 8553-70. - - - , - - - - & LEAT,P. T. 1983. Magmatic and structural evolution of the central part of the Kenya rift. In: Proc. Regional Seminar on Geothermal Energy in E. and S. Africa, Nairobi, pp. 61-7. UNESCO/US AID. WOLFF, J. A. 1983. Variation in Nb/Ta during differentiation of phonolitic magma, Tenerife, Canary Islands. Geochim. cosmochim. Acta, 48, 1345-8. - - - & STOREY, M. 1984. Zoning in highly alkaline magma bodies. Geol. Mag. 121,563-75. WORNER, G., BEUSEN,J.-M., DUCHATEAU,N., GIJBELS, R. & SCHMINCKE, H.-U. 1983. Trace element abundances and mineral/melt distribution coefficients in phonolites from the Laacher See volcano (Germany). Contrib. Mineral. Petrol. 84, 152-73.
B. H. BAKER, Center for Volcanology, University of Oregon, Eugene, OR 97403, U.S.A.
Quaternary peralkaline silicic rocks and caldera volcanoes of Kenya R. Macdonald S U M M A R Y: The late-Quaternary trachytic caldera volcanoes of the Kenya rift provide an unrivalled opportunity for studying the mechanisms of evolution of large peralkaline volcanic complexes and assessing the fundamental problems of magma genesis and chemical differentiation. In a northern set of centres (the Barrier, Emuruangogolak, Silali and Paka) basalt is a major component and mugearites, while scarce, may also be present. Caldera collapse was possibly of Kilauean type. Geochemical variations and the relationships between eruptive rocks and suites of plutonic nodules are consistent with fractional crystallization as the dominant differentiation mechanism in these centres. The southern basalt-absent centres (Menengai, Longonot and Suswa) show Krakatoa-style collapse. The pyroclastic sequences, especially syncaldera ash-flow deposits, indicate the ubiquity of striking compositional zonations within the magma chambers. Chemical variations in these centres have been ascribed to complex interplays of side-wall crystallization, magma mixing and liquid-state differentiation processes, especially involving volatile complexing, The ultimate origin of the trachytes at these centres is still debatable; crystal fractionation and partial melting of basalt are both viable mechanisms. The different development of the two types of centre may be related to the growth of a volatile-rich cap, which profoundly influences the types of differentiation mechanism.
Introduction Recent developments in igneous petrology have focussed attention on the differentiation mechanisms that generate highly evolved volcanic rocks such as peralkaline trachytes and rhyolites. A vigorous debate is in progress on the relative importance of magma mixing, partial melting, assimilation, crystal settling, side-wall crystallization, thermodiffusion, volatile complexing and vapour-phase transport in the development of the strong compositional zoning that characterizes many evolved volcanic systems. Volcanological studies of young centres, geochemical studies of their products, and experimental and theoretical work in geological fluid mechanics have identified new concepts for the evolution of magmatic systems. The most valuable progress will now be made by integrated studies of selected complexes to collect the geological, geochemical and geophysical data specifically designed to test the various competing hypotheses. Caldera volcanoes are particularly useful in such studies. Compositional and mineralogical variations within eruptive units, especially of pyroclastic type, record the common occurrence of zonation within magma chambers at single points of time, while the sequence of eruptive units marks the chemical evolution of the chamber with time. The relationship between the eruptive rocks and associated suites of coarsegrained nodules provides insights into plutonic processes while the volcano is at the magmatic stage.
In Kenya the late-Quaternary caldera volcanoes (Fig. 1) provide an excellent opportunity for placing the high-level stages of peralkaline silicic magmatism into a time-volume-compositional context. The centres are young (less than 1 Ma) and well preserved. Although many problems remain, their stratigraphy is fairly well known. It should be possible to establish rates of magmatic and structural evolution. The presence of zoned phenocrysts and extensive suites of plutonic nodules are evidence of fractional crystallization at various stages during their growth. Magma mixing has been recorded at several complexes and is probably ubiquitous. The importance of the volatile phase in magma genesis and differentiation in this province has been persistently advocated by D. K. Bailey and his colleagues (see below). Despite their potential importance, progress in the study of the caldera volcanoes has been uneven. Much of the large-scale mapping and preliminary geochemical data are in unpublished Ph.D theses, many of which were completed before the impressive developments in physical volcanology of the last 10 years. Although extensive geochemical and isotopic programmes are under way, little has yet been published. In this review geological and geochemical evidence of the origin and evolution of the peralkaline silicic rocks is related to the structural development of the caldera complexes. Attention is focussed on the seven centres of the inner trough of the Kenya rift (Fig. 1) but, where relevant, information from slightly older com-
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 313-333.
313
314
R. Macdonald
~
r'
, ''Q)" i Barrier
~Q~rnuruangogolak
-
KiIombe 9 ~ ~k
~ o "~~
"~ Rift Faults ;
4tl 4
3~ I
0 t
~1
".~.
100kin
C~ downward
and vaults
"~.
_ _ j
1
i\ 35:
36
37":
38<:
39~
FIG. 1.Locality map of the late Quaternary trachytic caldera volcanoes, the Eburru and Naivasha complexes, the Kilombe volcano and the area of the sourthern part of the Kenya rift (SKR) studied by Baker et al. (1977). plexes, e.g. Kilombe (about 2 Ma), and noncaldera complexes, e.g. Eburru and Naivasha, is also included. The Suswa volcano consists of undersaturated trachytes but is an integral part of the province.
The caldera volcanoes The structural development of each of the seven major late-Quaternary trachytic caldera volcanoes has been complex and, to varying degrees,
Silicic rocks of Kenya unique. McCall (1968), Williams (1978a, b, 1982) and Williams et al. (1984) have broadly distinguished two groups. In a northern group comprising the Barrier, Emuruangogolak, Silali and Paka, basalts are important components of the magma series. Intermediate rock types, such as mugearites and trachybasalts, are less common or absent. At Paka, Sceal & Weaver (1971) estimate the proportions of rock types to be 25% basalt, 5% mugearite and 70% trachyte. Weaver (1978) records 20% basalt and 80% trachyte at Emuruangogolak, with a complete absence of rocks in the range 49%-59% SiO 2. Using calculations of 'percentage crystallized' based on incompatibletrace-element (ITE) abundances, Weaver (1978) considers this gap to represent 25% of the crystallization range. A smaller, but real, composition gap is also present at Silali (McCall and Hornung 1972), but no data are presented on the relative proportions of rock types. An important feature of the northern calderas (Williams et al. 1984) is that caldera formation was accompanied by sparse pyroclastic activity, suggesting that caldera collapse was initiated by withdrawal of magma at depth (cfMcCall 1968). In the case of Silali, it is possible that withdrawal of basalt magma from the sub-volcanic reservoir and its eruption as the extensive Katenmening Basalts from fissures and fractures on the shield flanks triggered the collapse. The southern group of calderas (Menengai, Longonot and Suswa) is almost entirely of trachytic composition. Although basalts have
315
erupted between the centres, in the calderas the only mafic magma recorded occurs as mixed trachyte-basalt flows on Longonot (Scott 1980; Scott & Bailey 1984). The occurrence is important, however, in that it suggests that basalt is probably present beneath all the southern centres but normally has been unable to penetrate through lower-density trachytic caps. So far no mugearitic or trachybasaltic rocks have been found at these complexes. This may reflect physical discrimination against such compositions by the eruptive system, or may equally well indicate generation of the mafic and salic members from different source rocks. In contrast with the northern centres, caldera collapse in the southern complexes was of Krakatoan type (Leat 1984; Williams et al. 1984) and was accompanied by the eruption of voluminous air-fall and ash-flow tufts. The differing structural styles and the contrasting proportions of the major rock types in the two caldera groups suggest that the main fractionation mechanisms were different either in style or effect. Relevant evidence is discussed below.
The rock types The range of rock types found in the caldera centres is summarised in Table 1. The salic products of these complexes are trachytes, with maximum SiO 2 contents of about 64%. Rhyolites of Recent age are restricted to the dome field of Naivasha and Eburru, neither of which has evolved to the stage of caldera collapse. The
TABLE 1. Selected features o f the trachytic caldera volcanoes Caldera volcano
Range of rock types
Maximum SiO 2 SiO2gap (eruptives) (eruptives) (%) (%)
Types of nodule
Relative volume of pyroclastics
(%) ?
The Barrier
Basanites, U bas, U haw, U tbas, P phon Emuruangogolak U bas, OP trach, OS trach
64
49--59
Silali
63
50-56
Paka Menengai
O bas, U bas, U haw, O mug, OS trach, OP trach, UP trach, P phon U bas, O mug, OP trach, UP trach OP trach*, OS trach
61 68*
Longonot
OP trach, mixed haw-trach flows
64
Suswa
Up trach, US trach, P phon
59
9 No rock < 60 No rock <61t No rock < 56
Gabbro, syenite Gabbro, 'diorite', syenite Syenite Syenite Syenite
< 10?
53 >20
None recorded
Bas, basalts; haw, hawaiites; mug, mugearite; tbas, trachybasalt; trach, trachyte; phon, phonolite; U, undersaturated; O, oversaturated (hy +_q normative); P, peralkaline; S, subalkaline. * Strictly speaking, some rocks are rhyolitic. ~-Excluding mixed flows.
316
R. Macdonald
trachytes may be entirely oversaturated (Emuruangogolak, Menengai and Longonot), undersaturated (the Barrier?, Suswa) or contain both v a r i e t i e s (Silali, P a k a ) . T h e y a r e o v e r w h e l m i n g l y peralkaline, although subalkaline varieties have b e e n r e c o r d e d a t E m u r u a n g o g o l a k , Silali, M e n e n TABLE 2.
g a i a n d S u s w a a n d a r e i n v a r i a b l y less e v o l v e d than the associated peralkaline trachytes. S e l e c t e d a n a l y s e s a r e g i v e n i n T a b l e 2. T h e Kenyan rocks show the strong enrichments in Zr, Hf, Nb, Ta, Zn, Y, rare-earth elements (REE), F a n d C1 t y p i c a l o f p e r a l k a l i n e r o c k s . T h e e n r i c h -
Representative analyses of Kenyan peralkaline trachytes and rhyolites Trachyte Emuruangogolak $72
Trachyte Paka 4/L
Trachyte Menegai K 1A 1
Pantellerite Eburru K E - 12
Comendite Naivasha 565
Trachyte SKR KLR-54
70.30 0.33 7.62 2.36 6.24 0.26 0.02 0.35 7.28 4.27 0.02 0.10 0.42 0.33
72.5 0.24 10.32 1.93 2.25 0.06 0.02 0.15 5.86 4.39 -0.23 0.95 0.47
63.65 0.94 14.12 2.01 6.03 0.27 0.04 1.31 6.34 5.22 0.07 -------
Major elements ( w t . ~ ) SiO 2 TiO2 A1203 Fe203 FeO MnO MgO CaO NazO K20 P205 H20 + F C1
63.22 0.78 12.54 3.96 4.50 0.38 0.27 1.31 6.55 4.56 0.06 1.28 ---
59.45 0.83 14.89 4.14 3.78 0.30 0.92 2.32 6.00 4.84 0.09 1.41 ---
63.2 0.59 10.53 -9.02* 0.34 0.13 0.83 5.19 4.67 -----
Subtotal O - F, C1 Total
99.41 -99.41
98.97 -98.97
----
99.90 0.26 99.64
99.40 0.51 98.89
Ba Cr Hf Nb Pb Rb Sr Ta Th U Y Zn Zr La Ce Nd Sm Eu Gd Tb Tm Yb Lu
176 4 -167 10 116 18 -14 -112 254 556 99 172 ---------
174 --155 -110 42 ---81 -663 102 202 96 --------
6 < 1 34.5 369 36 216 3 22.3 40.6 7.8 210 416 1356 296 541 208 38.6 2.37 39.4 4.84 2.70 21.0 3.05
--34.2 372 -208 -21.1 32.5 7.3 300 472 1608 246 463 195 44.4 3.71 40.0 8.14 3.82 25.1 3.50
3 -60.5 590 68 637 <2 49.4 124 25.1 296 448 2018 153 310 151 36.8 0.47 42.2 8.22 4.61 34.6 4.33
160 -17.7 207 -115 10 11.1 19.4 -93 -764 152 185 91 17.9 3.17 -2.4 -8.9 1.64
K/Rb Rb/Sr Eu/Eu*
326 6.4 ---
365 2.6 ---
179 72 0.19 6.6
170 -0.27 4.7
57 > 1200 0.04 2.3
547 12 0.58 5.4
Trace elements ( p p m )
Ceu/Yb u
D a t a sources: $72, W e a v e r 1978; 4/L, Sceal & W e a v e r 1971 ; K 1A1, L e a t et al. 1984; K E - 1 2 , Bailey & M a c d o n a l d 1975, a n d u n p u b l i s h e d d a t a ; 565, M a c d o n a l d , u n p u b l i s h e d d a t a ; K L R - 5 4 , B a k e r et al. 1977.
Silicic rocks of Kenya ments are extreme in certain rhyolitic obsidians from Eburru and Naivasha. The F contents are among the highest recorded in silicic obsidians, and the Kenyan province can be regarded as the F-rich end-member of the natural spectrum of suites of ranging F:C1 ratios (Bailey 1980; Hildreth 1981). There is a broad, positive correlation between peralkalinity and concentrations of ITE such as Zr (Fig. 2). The scatter in this diagram is due to several effects: (i) post-emplacement losses or gains of alkalis, affecting the magmatic values of the agpaitic index; (ii) inter-suite variations in Zr abundances; (iii) non-linear relationships between Zr and the agpaitic index in certain suites, e.g. Eburru. It has been noted that differing values of ITE ratios, such as Zr/Nb (Fig. 3), characterize separate volcanoes within the rift (Weaver et al, 1972; Bailey & Macdonald 1975; Baker et al. 1977; Jones 1981). Weaver et al. (1972) thought that the differences reflect changes in the ITE content of the basaltic parents with time, while Bailey & Macdonald (1975) related the traceelement patterns to buffering by vapour phases of variable F/C1 ratio during magmatic evolution. Changes in Zr/Nb ratio during the development of Emuruangogolak coincided with intense pyroclastic activity or caldera collapse. Weaver (1978) has explained this in terms of volatile 9
I
I
|
317
degassing, with partitioning of Zr towards the vapour phase being slightly higher than that of Nb. An important observation is the wide compositional range within rock types. Thus, specimens KLR-54 and K1A1 (Table 2) are peralkaline trachytes with almost identical SiO 2 contents. K 1A 1 has a considerably more evolved character, as indicated by such ratios as K/Rb, Rb/Sr and Eu/Eu* and the chondrite-normalized REE patterns (Fig. 4), in keeping with its more strongly peralkaline character. In contrast, the peralkaline rhyolite 565, by the same criteria is more evolved than rhyolite KE12 despite being considerably less peralkaline. Although mafic rocks are abundant in the northern centres, only 19 major-element analyses, indicating a range from basanite to hy-normative basalt, have been published (Smith 1938; McCall 1970; Brown & Carmichael 1971 ; Sceal & Weaver 1971; McCall & Hornung 1972; Weaver 1978). Therefore it is not yet possible to assess their petrogenesis. One point deserves preliminary comment, however. Norry et al. (1980) note that the Sr, Nd and Pb isotopic compositions of mafic lavas from northern Kenya are comparable with those in oceanic regions, implying not only an insignificant role for contamination by ancient sialic rocks but also that the sources had been impoverished in ITE for much of their history. I
I
l
l
I
3000
2500
o
2000
A
o% oo +
15OO
9
~o
o ~
el Q.
o
x
1000
9x'l" 9 ~x z~ + ~
500
o
[]
9 t&~.
/
~xo
[]
Emuruangogolak
+
Paka
x 9
Kilombe Eburru
o
Naivasha
9
[]
Suswa S . K e ny a R i f t
z~
' 1 "0
'
I
I
I
1" 4
I
1" 8 moh
(Na20
i 2" 2
i
' 2"6
+ K20)/JU203
FIG. 2. Zr-mol. (Na20 + K20)/A1203 (agpaitic index) plot for peralkaline trachytes and rhyolites. Data sources: Emuruangogolak, Weaver (1978); Paka, Sceal & Weaver (1971); Kilombe, Jones (1981); Menengai, Leat et al. (1984); Eburru, Bailey & Macdonald (1975, and unpublished data); Naivasha, Macdonald (unpublished data); southern Kenya Rift, Baker et al. (1977); Suswa, Nash et al. (1969).
318
R. Macdonald
600
|
]
il
I
J
I
500 Kil~ 400
Z
/,,,
i~.
j, .,...,s" ~
.'
(//%j.,.." f / . \". . \Menengai l
300 200 ..
/~, .,,.'"
100
Paka
'J ~
0
I
I
I,
I
I
I
500
1000
1500
2000
2500
3000
3500
Zr ( p p . m . )
FIG. 3. Zr-Nb plot for basalt-trachyte suites of Paka and Emuruangogolak (post-caldera lavas only), the Kilombe trachytes, the first Menengai trachytic ash-flow sheet and the pantelleritic trachyte-pantelleritic obsidians of Eburru. Data sources as ['or Fig. 2. Metasomatism of the sub-Kenya mantle by CO,rich fluids in relatively recent times (less than 100 Ma-ago?) was invoked as a source of ITE in the rift basalts. Nevertheless, the abundances of the ITE remained low for alkali basalts. For example, plots of Ce, Zr and Rb against (F%O3 + FeO)/ (FeO+ Fe:O3 + MgO) (Thompson 1982a, Fig. 34.3) reveal that the ne-normative basalts of Paka (Sceal & Weaver 1971), Emuruangogolak (Weaver 1978), north Naivasha and Silali (University of Lancaster, unpublished data) have levels of ITE much more characteristic of olivine tholeiites than alkali basalts. In this respect the Kenyan rocks compare with the ne-normative members of the Skye Main Lava Series (Thompson et al. 1980; Thompson 1982a, b). This effect is particularly marked for Zr, a feature also noted by Pearce (1983) on the basis of MORBnormalized trace-element patterns (MORB, midocean ridge basalt). It is curious that the basalts associated with some of the most ITE-enriched siticic rocks in the world are relatively ITE impoverished. An important implication is that, if the higher-silica Kenyan rocks are related to the basalts by simple closed-system fractional crystallization, the liquid lines of descent were extremely long. Baker et al. (1977, p. 328) calculate from major-element modelling that the 'commonest' (not the most fractionated) type of Plateau trachyte lava in the southern Gregory rift represents about 3% by
volume of residual liquid derived from the putative parental basalt. That trachyte contains 63% SiO2. For high-silica rhyolites (76% SiO2 or less) of the Naivasha area, which show Rb enrichments of up to 80 relative to possible parental basalts, the calculated volume is around 1%. Such values place constraints on the timespace evolution of the magmatic systems which are as yet unconsidered in most models of fractional crystallization.
Geological evidence for depths of magmatic differentiation On the basis of the approach of Mahood (1984), several lines of evidence combine to suggest that the magma reservoirs beneath the caldera complexes lie at relatively shallow depths. Piutonic nodules There can be little doubt that these nodules are the sub-volcanic, crystallized equivalents of the basalts and trachytes of the Kenyan calderas (McCall 1970; McCall & Hornung 1972; Weaver 1978; Jones 1979a; Leat et al. 1984). The presence of medium-grained varieties of syenite at Emuruangogolak (Weaver 1978) and at the 2 Ma centre Kilombe (Jones 197%) and of miarolitic gabbros with interstitial glass at Silali (McCall 1970) suggests that crystallization took place at shallow levels.
Silicic rocks of Kenya
319
1000
Menengai, KIAI
500
~ ~
.
s s"L&"~
Eburru, KE12" ' -. 9 . . . .
9
100
.
50
SKR, KLR - 54
1
Naivasha |1 565 I
10 -
i
1 |
!
I
|1
II
5 J ka
I
I
I
I
1
1
I
1
Ce
Nd
am
Eu
Gd
Tb
"I'm
Yb
ku
FIG. 4. Chondrite-normalized REE patterns for selected peralkaline trachytes (K 1AI and K LR-54) and rhyolites (KEI2 and 565). Data and localities are given in Table 2. The normalizing values are from Nakamura (1974), except for Tb and Tm which are from Thompson (1982b).
Distribution of recent faults The inner trough of the Gregory rift is cut by swarms of closely spaced normal faults, whose d e v e l o p m e n t has overlapped that of the lateQuaternary calderas. Evidence from Suswa
(Johnson 1969) and from Longonot and Menengai (McCall 1967) shows that these faults virtually die out on the shields of those centres. M a h o o d (1984) has noted that this is a general feature of Quaternary pantelleritic or trachytic shields and suggests that it indicates the existence at shallow
320
R. Macdonald
depths of partially molten zones which cannot support brittle fracture. The situation in the southern centres contrasts with that in the northern group. At both Silali (McCall & Hornung 1972) and Emuruangogolak (Weaver 1978) closely spaced faults and open fissures cut the shields virtually to the caldera rims. This raises the interesting possibility that the northern complexes are underlain by smaller or deeper chambers than those further S.
Experimental studies Phase equilibrium studies on pantellerites and trachytes from central Kenya have provided preliminary information on pressures of equilibration of the magmas. Liquidus data for a halogen-rich (F+C1=0.75 wt. %) water-poor (H20+=0.10%) pantelleritic obsidian from Eburru (Bailey et al. 1974) show that alkali feldspar is the liquidus phase to pressures of about 1 kb, where it is joined by quartz. At higher pressures quartz is on the liquidus. Since quartz phenocrysts are relatively uncommon in pantellerites (Mahood 1984; Sutherland 1974) and virtually always occur in association with alkali feldspar, an upper limit of 1 kb is set for the pressure at which the pantellerites equilibrated. Since liquidi temperatures for pantellerites and trachytes converge at higher pressures (Bailey et al. 1974, Fig. 3), low pressures of equilibration probably also characterize the Kenyan trachytes.
Caldera collapse On the basis of studies of the Sierra La Primavera caldera complex, Mexico, Mahood (1981) has argued that the formation of a caldera 11 km in diameter, on eruption of only 20 km 3 of magma as the caldera-forming unit, requires a magma chamber with a shallow roof (2-6 km?). If eruption had proceeded from a greater depth, collapse would have been accommodated more easily along regional normal faults than along circular ring fractures. Mahood (1984) applied this line of reasoning to peralkaline volcanoes in general, including the Kenyan centres, most of which occur in areas of extensional tectonics. With specific reference to Menengai, the high eruption rates required for the emplacement of two chemically zoned voluminous ash-flow tufts as single flow units also suggest that, prior to eruption, the magma resided in a shallow reservoir (Leat 1984). Although magmatic differentiation involving the salic rocks apparently proceeded at shallow depths, the sizes and shapes of the magma chambers are not known. It is assumed here that
the area of the top of each chamber is no smaller than the area of the caldera above it, and that the chambers have the mushroom shape currently widely accepted (Smith 1979; McBirney 1980; Hildreth 1981) for high-level salic reservoirs.
Basalt-trachyte relationships Fractional crystallization of basaltic magma This has been favoured by S c e a l & Weaver (1971), Weaver et al. (1972), Baker et al. (1977), Weaver (1978), Jones (1979b), Norry et al. (1980) and Baker (1987) as the dominant mechanism producing trachyte in the Pleistocene-Recent sequences of the rift. Prima facie evidence for crystallization processes within the magma chambers of the northern caldera centres is provided by plutonic nodules. McCall (1970) and McCall & Hornung (1972) have described a series of gabbroic and syenite blocks from Silali volcano which are mineralogically and compositionally equivalent to the eruptive rocks and clearly represent the 'plutonic' stage of evolution. A feature of the nodules is their extraordinary abundance: ' . . . the surface of the basalt is littered with a profusion of blocks and bombs, and these are far more conspicuous than outcrops of the basalt' (McCall 1970, p. 257). Their size is also notable: the largest gabbroic block has dimensions in excess of 6 m, thicker than the flow which apparently carried it. McCall suggests that they represent parts of the magma chamber disrupted and detached by gascoring during a terminal phase of ascent of volatile-charged magma which erupted as the pre-caldera Katenmening basalts. The gabbroic nodules contain olivine + titanaugite + plagioclase + Fe-Ti oxides + phonolitic interstitial glass, which in some samples is miarolitic. Some specimens are slightly more evolved, exhibiting a 'dioritic' character, with biotite instead of olivine and plagioclase of andesine composition. Syenite nodules are also mentioned, but few details are given. The Silali nodules are a very important suite, recording conditions in the pre-caldera magma chamber. The simplest interpretation of the relationships is that the interstitial phonolitic liquid in the mafic blocks formed by fractional crystallization and was expelled from the crystallizing wall-rocks to form part of a salic cap to the chamber. Weaver (1978) reported that small gabbroic nodules are occasionally found in the late basaltic fissure eruptions of Emuruangogolak. Some are mineralogically equivalent to the basalts, while
3 21
Silicic r o c k s o f K e n y a
have used linear correlations of TTE in Kenyan basalt-trachyte sequences to suggest that the linearity is best explained by fractional crystallization. The argument is that processes such as volatile transfer and crustal melting are unlikely to add the ITE to the melts in exactly the right proportions and amounts to maintain the linear correlations. In fact it remains an open question whether high ITE correlations can or cannot result from the generation of basalts from the upper mantle and trachytes from the lower crust during the same volatile-induced melting cycle, especially if halogens+COz have a dominant control over trace-element distribution. The question of the linearity of ITE correlations deserves examination, however. Zr-Rb co-variations for two Kenyan suites are shown in Fig. 5. One is Paka, where the data are taken (Sceal & Weaver 1971) to be most consistent with a fractional-crystallizationorigin for the trachytes. The other is a pantelleritic trachyte-pantellerite suite from Eburru used by Bailey & Macdonald (1975) to invoke the necessity of a halogenbearing vapour phase in magma generation. The Eburru rocks show considerably less scatter than the supposedly linear Paka sequence. Further problems relating to the Paka rocks are as follows.
others are cumulate rocks rich in plagioclase, sometimes with well-developed igneous lamination. One flow contains up to 30% nodules in various states of disaggregation. Most are plagioclase-clinopyroxene-magnetite aggregates; others are dunitic. Weaver infers the presence of layered gabbros at high levels in the Emuruangogolak system. As at Silali, the nodules provide direct evidence of fractional crystallization at depth. Another approach to basalt-trachyte relationships has been by geochemical modelling. The most detailed analysis is that by Baker et al. (1977) of a Plio-Pleistocene basalt-benmoreitetrachyte suite from the southern part of the rift which shows compositional similarities to the northern centres. They used mass-balance calculations for major-element compositions and Rayleigh-law modelling of trace-element contents to show that compositional variations within the suite resulted from closed-system fractionation of the observed phenocryst phases or of phases reasonably inferred to have existed at higherpressure stages of evolution. A 10% SiO2 gap which exists between benmoreites and basalts was considered to result from physical discrimination against the relevant magmas by the eruptive system. The eruption of the lavas spanned a considerable period of time (0.8 Ma or less) and took place from several vents. As appreciated by Baker et al. (1977, p. 315), the rocks cannot be placed in a rigorous timecomposition-space context. Sceal & Weaver (1971), Weaver et al. (1972), Ferrara & Treuil (1974) and Baker et al. (1977) L
I
'
1 The complete body of analytical data, and the petrographic and stratigraphic details, have not been published. Judging from the few published analyses (Sceal & Weaver 1971), the Paka suite contains ne- and hy-normative basalts and neand q-normative trachytes. The suite clearly represents more than one lineage, yet these I
]
1
I
J /
4~176
.
t-
300
k-
iQ. 200
o
o
"
o
,%
0
1000
2000
3000
Zr(ppm)
FIG. 5. Zr-Rb co-variation for lavas of the basalt-trachyte series of Paka and the pantelleritic obsidians of Eburru. Data sources are given in Fig. 2. Paka basalts are shown as field. The full line is the computed reduced major-axis line for the Eburru data, and the heavy bar on the axis shows the 95% confidence limits on the intercept.
322
R. Macdonald
complexities are not explored by Sceal & Weaver (1971). 2 In certain graphs, e.g. Z r - R E E (including Y) plots, the more mafic rocks are clearly not collinear with the trachytes. 3 No pyroclastic rocks were included in the analysed sample. The significance of the linearity at Paka cannot therefore be assessed, and these remarks also apply to all the Kenya centres used by Sceal & Weaver (1971) & Weaver et al. (1972). A rather stronger case for linearity is presented by Weaver (1978). On a Z r - N b plot (Fig. 6) the Emuruangogolak trachytes fall on three distinct lines, in accord with stratigraphy. Even here, two of the lines are not truly collinear with the basalt analyses and do not, on extrapolation, pass through the origin. These features may be compatible with different ITE partitioning into the crystallizing phases but a case has not yet been made.
Significance of composition gaps As noted earlier, a scarcity of intermediate rock types is present at all the northern centres, accompanied by a distinct gap over part of the
I
Emuuang
I
I
compositional range. This may reflect the following. 1 Some sort of physical control in which rise to the surface is prevented by high viscosity and/or density (Baker et al. 1977; Jones 1979b). Thus, four flows of the Kilombe volcano contain blebs of anorthoclase-phyric benmoreite within trachyte (Jones 1979b). The blebs range in volume from less than 1% to 10% and in size up to 30 cm in diameter. Benmoreite lavas are not present at the surface, so the blebs bridge the composition gap in an otherwise complete sequence of lavas from basalt to trachyte. Jones argues that the benmoreite magmas had high viscosity, which was perhaps due to the anorthoclase phenocrysts, and were discriminated against during euption. At the Longonot volcano basaltic magma has been erupted only as mixed lavas with peralkaline trachyte (Scott 1980; Scott & Bailey 1984). These lavas were the first products following each of three caldera collapses, and Scott & Bailey (1984) argue that input of basalt magma into a root zone underlying a trachyte magma chamber initiated each pre-caldera pyroclastic event and subsequent caldera formation. Again, the volcano has been an imperfect sampler of its magmatic
I
l
I
I
oak
300
200
!
100
Basalts
Pre-caldera n" pre- tuffs tr~chyte.~
x
Pre-caldera ~
post- ruffs
trachytes
obsidians
Postand
0
and
caldera
rr
trachytes
obsidian
I
I
I
I
I
I
i
200
400
600
800
1000
1200
1400
Zr (p.p.rn.)
FIG. 6. Zr Nb plot for lavas of Emuruangogolak showing, in slightly simplified form, the three groups identified by Weaver (1978, Fig. 6).
Silicic rocks of Kenya components. A requirement for the Kenyan suites is that the compositions being discriminated against are variable: hawaiites and mugearites (5.1%-2.4% MgO) in the Pleistocene sequence of the southern rift studied by Baker et al. (1977), hawaiites to benmoreites (3.9%-0.8% MgO) at Emuruangogolak (Weaver 1978), and benmoreites (about l%MgO) at Kilombe (Jones 1979b). Therefore it cannot be said that a certain compositional range has less potential to erupt. 2 A control related to the crystallization process itself, where the intermediate composition range is seen as one of rapid crystallization over a narrow temperature interval producing low volumes of melt (Weaver 1978). The appropriate compositions are then found, perhaps exclusively, as margins to the major crystallizing phases within the chamber. Direct evidence of this is found in the nodule suite at Silali; the interstitial glass in certain gabbroic cumulates is of phonolitic composition (R. Macdonald, unpublished data). One untested possibility is that the onset of crystallization of Fe-Ti oxides produces rapid changes in major-element composition, especially SIO2. Using calculations of 'percentage crystallized' based on ITE abundances, Weaver (1978) considered that the composition gap between 49% and 59% at Emuruangogolak represented only 25%/0 of the crystallization range between basalt and trachyte. In a similar study Clague (1978) suggested that the transition 50%-57% SiO 2 in the basalt-trachyte sequence of Reunion represented only about 15% of the total range. Both workers related the SiO2 jump to Fe-Ti oxide precipitation. The position of the composition gap in a magmatic series would then be critically dependent on the point at which oxides began to precipitate, which is itself perhaps a function ofpO 2. Such a mechanism might account for a relative scarcity of intermediate lavas but not for their complete absence, as at Emuruangogolak and Silali, unless a particularly efficient combination of 2 and 3 were operating. 3 A real absence of intermediate materials in the erupted and subjacent parts of the volcanoes would cast serious doubt on a fractional-crystallization hypothesis. This plus the relatively high ratio of salic-to-mafic products at even the northern centres has encouraged the idea that the basalts and trachytes have had different sources. Partial melting of mafic deep continental crust The basis of this process (Bailey 1964; Bailey & Schairer 1966; Macdonald et al, 1970; Bailey 1974) is that a mantle-derived halogen+CO2rich volatile phase promotes melting of the lower
323
basaltic parts of the continental crust, trachytes representing greater degrees of melting than rhyolites. The vapour buffers the composition of the melts because it controls the temperature of melting and the proportions of solid phases entering the liquid. It also maintains a buffering action on the melt during its subsequent ascent via crystal ~-~ liquid ~- vapour equilibrium (Bailey 1980). A problem with this model is that it is, as yet, difficult to test quantitatively. Little is known about the way in which elements partitition between crystal, liquid and vapour during melting and subsequent ascent. Similar lack of information has not dissuaded petrologists from a general acceptance of partial melting of metasomatized mantle as a mechanism for generating alkaline basalts. Isotopic data which might point to different sources for the basalts and trachytes are available only for the Emuruangogolak volcano (Norry et al. 1980). The 8VSr/S6Sr ratios of the trachytes (0.70372 + 0.00004-0.70578 + 0.00003) are higher than those of the basalts (0.70338+0.00006 - 0.70343 + 0.00006), which requires either contamination of the trachytes or derivation from a different source. However, Pb and Nd isotope compositions have mantle values and show no consistent differences between the rock groups, and Norry et al. (1980) prefer the explanation that the trachytes were derived from the basalts by fractional crystallization but were contaminated by continental crust. In summary, several features of the northern set of centres are consistent with a fractional crystallization origin of the salic rocks" the presence of plutonic nodules, in some cases spanning the compositional range of the lavas and containing trachyte glass, the relatively complete range of rock types and the high proportion of lavas compared with pyroclastic rocks. It must be noted, however, that the few available isotope data seem to require trachytecrust interactions. There are important composition gaps at each centre and the pyroclastic rocks in these complexes have been analytically neglected.
Trachyte-trachyte and trachyterhyolite relationships One view is to see the crystal-J?actionation mechanisms which produced the trachytes from parental basalts as continuing and controlling compositional variation within the trachytes. One proponent (Weaver 1978, p. 223) acknowledged that at Emuruangogolak, ' . . . in detail the
324
R. Macdonald
validity of the crystal fractionation process is difficult to assess' owing to the cumulophyric nature of many rocks and the loss of alkalis during post-eruptive processes. A favoured mechanism for generating peralkaline from subalkaline melts in the Kenyan sequences (Nash et al. 1969; Baker et al. 1977; Weaver 1978) is the plagioclase effect (Bowen 1945), a process exacerbated by entry of AI into coexisting clinopyroxerie. The transition to peralkalinity within onefeldspar trachytes, lacking discrete plagioclase, is possible only where the feldspar phenocrysts are relatively calcic, i.e. where they have higher normative an than the liquid. The peralkalinity is exacerbated by removal of alkali-feldspardominated assemblages, with minor clinopyroxerie, fayalite and Fe-Ti oxides. If the assemblage is ne normative, it may drive mildly undersaturated trachytic magmas towards silica oversaturation. As noted earlier for the gabbros, there is direct evidence of crystallization processes within the subjacent trachytic magmas in the form of plutonic nodules. At Emuruangogolak, for example, syenite blocks occur in pyroclastic rocks associated with the collapse of caldera II. They are mineralogically similar to the trachyte lavas, although somewhat more leucocratic. The volcano Kilombe (about 2 Ma old) is older than the other caldera centres discussed here but contains an important suite of syenite nodules (Jones 1979a). The nodules are found only in post-caldera flows and range in size up to some metres. Four classes are recognized. 1 Most abundant, and forming all the large blocks, are coarse-grained syenites. These are formed by nearly complete solidification of batches of trachytic liquid, although a small proportion of residual liquid has escaped from these rocks. 2 Fine-grained syenites, which represent essentially solidified trachyte magma. 3 Aplitic syenites, which occur as veins in the other syenites and are again interpreted as essentially crystallized trachyte. 4 Melasyenites which are found in very-latestage post-caldera tufts and are interpreted as Na-Fe-Nb-Y-REE-enriched residual fluid from the crystallization of coarser-grained syenites. Phenocrysts of nepheline occur in some finegrained syenites, and groundmass nepheline and zeolites are common in coarse- and fine-grained varieties. However, all the analysed lavas are slightly SiO2-oversaturated. Furthermore, Jones (1979a) claims that fractional crystallization in the magfiaa chamber has produced more evolved compositions than in the lavas. The general point
is that the nodules provide evidence of compositions at depths either not represented in the extrusive rocks or not yet sampled. They record a stage or stages in the evolution of the magma chamber which is not available in the surface rocks. Syenite blocks or nodules at Menengai were first described by McCall (1967). Their place in the evolutionary history of the centre has been elucidated by Leat et al. (1984). Nodules are found in the second ash-flow sheet, whose eruption accompanied collapse of the Menengai caldera. They are thought to represent crystallization against the roof and sides of the chamber of a trachytic magma formed by rehomogenization of the upper part of the chamber after euption of the first ash-flow sheet. This magma was apparently never tapped by the volcano so, as at Kilombe, an important evolutionary stage is recorded only in the plutonic nodules. Syenites are also found in the Ruplax Tuff, a post-caldera phreatomagmatic deposit. Leat et al. (1984) suggest that these formed by 75~ crystallization of a Ba-rich trachytic magma against the roof; the chemical effects of this crystallization episode can be recognized in the erupted products. The evidence from the nodules is that fractional crystallization was intermittently active at Menengai. Other differentiation processes have also been recognized. Trachyte-trachyte mixing occurred at several stages during the growth of Menengai (Leat et al. 1984). The only direct evidence of mixing is in certain small-volume pre-caldera tufts. These show distinctive compositional variations which are similar to those of the pre-caldera lavas (Fig. 7). It is suggested that the lavas (about 30 km 3 volume) also represent a mixing sequence, although pre-eruptive mixing obliterated petrographic evidence of the end-members. Leat et al. (1984) see this process as being consistent with structural stage: the magma chamber was growing by the addition of two or more trachyte magmas. There is also some evidence that in very recent post-caldera times trachytic magmas from deeper in the magma chamber penetrated into the volatile-rich cap to the chamber, resulting in the formation of mixed magma. Pantellerite-pantellerite and comendite-trachyte-basalt mixing have been noted at Eburru (D. K. Bailey, personal communication) and Naivasha (R. Macdonald, unpublished data) respectively, suggesting that mixing of salic magmas may be ubiquitous in the rift volcanoes. However, other than at Menengai, its effect as a differentiation mechanism has yet to be assessed.
Silicic rocks of Kenya
325
--,
I
l
I
II
|
I
11
I
11
280 .
Menengai
~
240
200
~,,,- / /
160
#
o,
6 9 //\ 9~'("~ , , s o
120
9. 99j ~ J
f
/
/
/
J
/Y J
I
A~.~__Pre-calderatuffs "''
\
o," ,r
80
.."~~ i " I" .s
. . - caldera . . B\\~_-Pre lavas
40
0
I 0
:00
I 400
I
I
I
I
600
800
1000
1200
I 1400
I
|
I
1600
1800
2000
Zr (io.p.m.)
Fie. 7. Zr contents plotted against Y contents in Menengai volcanic rocks (after Leat et al. 1984). The precaldera tuff trend A represents a single zoned air-fall tuff which is megascopically mixed. Trend B is generalized for the pre-caldera lavas. Both are at a high angle to the trends of syn-caldera and post-caldera units, which were interpreted by Leat et al. as resulting mainly from combinations of liquid-state differentiation processes and crystal fractionation. Leat (1983) and Leat et al. (1984) have identified a further process, rehomogenization of magma by convective overturn (Fig. 8) which contributed to magma evolution at the Menengai volcano. Calculations showed that the density of the trachyte melts was controlled mainly by Fe and volatile contents. Prior to eruption of the first ash-flow sheet, a roofward decrease in density of at least the upper parts of the pre-eruptive column was caused by an increase in volatile concentrations, outweighing the contrary effects of an increase in Fe and a decrease in temperature. Following eruption of the ash-flow, and probably accompanying caldera collapse, volatiles were lost to the atmosphere via fractures in the cauldron block. One result was that the Fe-rich magma in the upper zone of the chamber now had a density similar to, or greater than, that of the underlying lower-Fe trachyte. Convective mixing was promoted, producing a homogenized magma. This effect may be common in peralkaline systems, where Fe contents generally increase in
more differentiated, more peralkaline liquids. However, a good understanding of the timecomposition-space relationships at the volcano is necessary to be able to distinguish the homogenized magma chemically. Further complexity in the interplay of processes during the development of Menengai was revealed by detailed studies of the two ash-flow and air-fall tuffsequences which accompanied caldera collapse. Leat et al. (1984) have recorded preeruptive gradients in magmatic composition of the kind well known in calc-alkaline systems (Smith 1979; Hildreth 1981) and increasingly recognized on peralkaline systems (Gibson 1970; Noble et al. 1979; Hildreth 1981 ; Mahood 1981 ; Wolff & Wright 1982). At Menengai there was roofward enrichment in Na, Fe, Mn, F, C1, Rb, Cs, Pb, Th, U, Nb, Ta, Zr, Hf, REE, Y, Zn and Ni, and roofward depletion in Mg, A1, K, Ca, Ti, P, Sc, Ba and possibly Sr (Fig. 9). One important observation at Menengai is the ubiquity of compositional zonation in the pyroclastic units. Even thin (a few metres) post-
326
R. Macdonald
~dl[llllt~llIl!lllTSlllllil]lllll)'~'.]!',]ll]][Y
C REHOMOGEN|ZED
FIG. 8. Diagrammatic representation (much simplified after Leat et al. (1984), Fig. 15)) to show the effects of convective overturn on rehomogenizing trachytic magma. A, Prior to major ash-flow eruption, a stabledensity interface separates volatile-rich and volatilepoor layers within the trachytic cap to the chamber. B, Following caldera collapse, volatiles in upper parts of the chamber are easily lost through the fractured roof and the magma becomes more dense. Convective overturn promotes mixing with underlying trachyte and perhaps also with magma (striped) from deeper in the chamber, rising buoyantly towards the roof zone. C, Convection proceeds within the upper chamber until a volatile-rich cap is re-established. caldera tufts may show substantial gradients; Zr abundances in tuff G, for example, vary from 1215 to 1769 ppm. Another is that the variations may be extreme; enrichment factors exceeding 5 for some elements in the first ash-flow sheet are about the maximum yet recorded in a single zoned eruptive unit. Preliminary study of the Hell's Gate green tuff, a syn-caldera ash-flow of the Longonot volcano, shows that it also is compositionally zoned, although less strongly than the Menengai sheet (enrichment factor 2). It is probable that such zonations are characteristic of all the pyroclastic phases in the southern calderas at least, and further study should document them. As is the case for the calc-alkaline examples (Hildreth 1983; Michael 1983; Miller & Mittle-
fehldt 1984), the origin of the compositional zonations is contentious. One possibility can immediately be ruled out at Menengai. Figure 10 is a section of the SiO2-A1203-(Na20+K20) system (Bailey & Macdonald 1969) on which are shown the compositional range of the second Menengai ash-flow tuff and a feldspar control line constructed by connecting the last-erupted least-differentiated tuff to its coexisting alkali feldspar phenocrysts. Clearly, feldspar fractionation by itself could not have produced the trend towards strong enrichment in peralkalinity shown by the tuff magmas (cf Macdonald et al. 1970; Bailey & Macdonald 1975). Addition of clinopyroxene to the fractionating assemblage is of little help because the hedenbergitic phenocrysts in the sheet are peralkaline, in the sense of having molecular Na20 greater than A1203 (Leat 1983). A further problem regarding fractional crystallization is shown in Fig. 11. Generalized trends for the northern centres Emuruangogolak and Paka show the decrease in K/Rb ratio almost invariably shown with increasing differentiation. The last-erupted parts of the Menengai ash-flows have K/Rb ratios greatly in excess of the trachytes at these centres and, in the case of the first sheet, of any Quaternary basalt in the rift valley. This is quite different to normal geochemical experience and would imply a fractionation mechanism of K and Rb different to that involving crystal --~ liquid processes. Leat et al. (1984) used trace-element data to suggest that poorly specified liquid-state differentiation processes had operated, in combination with minor crystal fractionation, in the upper parts of the Menengai chamber. Their model could not be tested rigorously because of lack of information on mineral-liquid partition coefficients in the Menengai samples and because it is currently impossible to disentangle the relative roles of volatile complexing, vapour-phase transport and, for certain elements, fractional crystallization in the differentiation of peralkaline silicic magmas. A role for volatile complexing in the evolution of the Kenyan salic magmas has been repeatedly championed by D. K. Bailey and his colleagues (Bailey 1986). Using the composition of non-hydrated glassy pantelleritic trachytes from Menengai, Macdonald et al. (1970) suggested that the transition from pantelleritic trachyte to pantellerite in the Kenyan volcanoes was achieved, not by simple crystal fractionation, but by alkali feldspar ~-liquid~,-~- alkali-bearing vapour equilibria. In a study ofpantellerites from Eburru, Bailey & Macdonald (1975) showed that major- and trace-element variations were incompatible with closed-system
3o!
Silicic rocks of Kenya Zr
327
Nb
2.5 4 !
Tm
Zn
2.0 4 I
Hf
I I Th U
La CeNdSm
Cu Mn Fe
1.5 i 1.0
1/1.5 -
Z
1/3.5
2nd ASH-FLOW
TUFF
FIG. 9. Observed enrichment factors in the second Menengai ash-flow tuff (Leat et al. 1984, Fig. 8), calculated by dividing the compositions of the early-erupted by the late-erupted parts of the tuff.
SiO2
A12033 ' SIO2 /
~
~
) I~ FELDSPAR
AI20 3
Mol .~o
~
(Na20+KO
2 )'3SIO2
(Na20+K20) SiO 2
Feldspar control
80
/ 7 ~(= AI203
t I }
,
~ 20
Late Ill i - ~ 1 1 Earl I ,Feldspar ~ I \ O~ AF 10 15
8O
5 (Na;O.K20)
FIG. 10. Enlarged portion of the variation diagram for molecular SiO2, A1203 and (Na20 + K20) (shaded area in inset). AF is the stoichiometric Ca-free alkali feldspar composition. A tie-line connects the compositions of the early- and late-erupted parts of the second Menengai ash-flow tuff (11). The open circle is the composition of feldspar phenocrysts in the late-erupted part of the tuff. Data from Leat (1983).
328
R. Macdonald I
l
I
I
I
1
I
I
I
700 \ M e no n g a i
x
6OO
500
,~
_ ":..2.:: ....
400
elf
Field
-
of b e s a l t s
~
~
"'"'""
~O9o/~ ~
,,,,e
....-....2
...,
300
200
",,?t,
i.
\
i 100
L__ @5
__~
L
..
1 0.6
I
l 0.7
I
1
1
@8
I
I
0.9
(FeO+ F e 2 0 3 ) / ( FeO + Fe203~'MgO ) FIG. 11. K/Rb ratio plotted against (Fee + Fe203)/(FeO + Fe203 + MgO) ratio for the basalt-trachyte suites of Emuruangogolak and Paka, and for the first and second Menengai trachytic ash-flow sheets. Data sources as for Fig. 2. The generalized trend for the oceanic Bouvet Island basalt-comendite sequence is also shown (Imsland et al. 1977). Additional rocks in the field of basalts from Pleistocene-Recent basalts of the Eburru and Naivasha areas (D.K. Bailey and R. Macdonald, unpublished data). crystal fractionation and invoked the role of a halogen-rich vapour phase in magmatic evolution. They suggested that in this phase Zr and Rb complexed preferentially with F, and Nb, Y and Zn complexed preferentially with C1 (Fig. 12). As Mahood (1981) has noted, acceptance of a role for volatile complexing in the evolution of peralkaline silicic magmas does not disguise the fact that the mechanisms are speculative and, as yet, not verified experimentally. We have little information on the nature of the inferred complexes: very high correlations of trace metals with F or C1 in the Eburru obsidians may indicate complexing with halogens (Bailey & Macdonald 1975). However, Scott (1982) presents petrographic evidence that a low-solubility CO2-rich vapour phase coexisted with peralkaline trachyte magma at Longonot, raising the possibility of a role for dissolved carbonate complexes (cfBailey 1980). Bailey (1978) suggests that the vapour coexisting with the silicic magmas was also rich in N and H. However, Kenyan silicic magmatism generally appears to have been very low in H20 (Bailey 1978, 1980). Little is known of the way in which the volatile complexes operate. They possibly migrate along P-Tgradients within the magmas. Alternatively, the complexing of a trace metal with a volatile
phase may lower its chemical potential in the liquid, thereby lowering the likelihood of it entering the crystallizing phases. Continued studies of trace-element patterns in relation to measured halogen, CO2 and H20 contents in the Kenyan centres will provide some insight into this thorny aspect of alkaline rock genesis. The Menengai study has raised several points of general importance in the study of peralkaline volcanic rocks. 1 Whatever the mechanism, it appears that extreme compositional zonations were developed repeatedly through tens of cubic kilometres of magma in times of 102-104 years (Leat et al. 1984). These rates are high when compared with large-volume nonperalkaline systems and must be related to the high Fe and high halogen contents and the relatively low viscosity of the magmas. 2 The most mafic trachytes erupted at Menengai, certain high-Ba post-caldera tufts, are thought to have compositions partly controlled by processes other than crystal fractionation. They have covered their genetic tracks and cannot therefore be used to model likely mafic source rocks or possible liquid lines of descent from basaltic magma. Leat et al. (1984) claim
Silicic r o c k s o f K e n y a
32 9
./
600 E
D ..a
400
z
/
200
~~
/
r=0.955 I
0
I
I
, r=0.943,
I
/ 3000
.J
2000
1000
~ 2000
it
t
.J
/
/
eo"i
o
r=0.876
r=0.994 I
I
1
I
l
I
I
4000
6000
8000
1000
2000
3000
4000
F(ppm)
Cl(ppm)
FIG. 12. Examples of linear and non-linear arrays in the Eburru pantelleritic obsidians (from data in Bailey & Macdonald (1975)). Nb is thought to have formed preferred complexes with C1, and Zr with F, during magmatic evolution.
that parental trachytes have not been recognized at Menengai. The general implication is that indiscriminate use of the chemistry of erupted rocks, without due regard to their place in the evolutionary history of the complex, may lead to spurious petrogenetic conclusions. 3 The inference from the Menengai data (Leat et al. 1984) is that the peralkalinity resulted at least in part from processes involving a volatile phase. The antipathetic behaviour of Na and A1 resulted in those magmas in the upper part of the chamber achieving critical undersaturation in A1203 (cfMahood (1981) for the zoned Tala Tuff, Sierra La Primavera, Mexico). The plagioclase effect cannot be uncritically assumed to have promoted peralkalinity in any of the caldera complexes and especially in the southern set where liquid-state processes may have been important. 4 The evolutionary complexity of volcanoes such as Menengai is unlikely to be recognized unless the pyroclastic rocks are included in the geochemical work. Analyses of lavas only would reveal a relatively limited range of compositions, and genetic models based on these data would be wholly inadequate to explain the total compositional range. The analysed sample automatically biases the genetic conclusions. This has been a weakness of several studies of the Kenyan
peralkaline rocks (Sceal & Weaver 1971 ; McCall & Hornung 1972; Weaver et al. 1972; Bailey & Macdonald 1975; Baker et al. 1977; Weaver 1978).
Generalized models It was accepted earlier that crystal-liquid processes may have dominated the chemical evolution of the northern set of centres. The development of the southern centres was clearly more complex, if Menengai is typical. Here a model for the evolution of both sets is presented. It is highly speculative but it is hoped that it will help to focus future research on the calderas. Leat et al. (1984) suggest that liquid-crystalvolatile processes became dominant over crystalliquid processes at Menengai only when a strong volatile gradient had become established in the magma chamber. The volatile-rich cap provided the environment wherein volatile complexing and possibly vapour-phase transport acted most effectively. Conversely, crystallization was suppressed throughout much of the cap by lowering liquidus temperatures, and magma mixing was inhibited by the inability of more mafic magma to penetrate into the lower-density cap. Therefore it may be no coincidence that those centres
330
R. Macdonald
(Emuruangogolak and Paka) where no role for volatile complexing has been invoked are from the northern group of caldera where caldera collapse was accompanied by only sparse pyroclastic activity. These centres have not (yet) established volatile-rich caps of sufficient size or stability to permit the operation of volatiledominated processes. The reasons for this difference cannot yet be isolated. The size or geometry of the magma chambers, the rate of magma through-put, the local rate of crustal extension (and thus the leakiness of the roof) and the volatile contents of the magmas may all be contributory factors.
NORTHERN 9e mugear~t
layer
Figure 13 attempts to present, in simplified form, the salient features of the evolution of the two sets of caldera complexes as viewed prior to caldera collapse. The general features of the magma chambers follow models of Smith (1979), Hildreth (198 l) and McBirney (1980). In the northern centres a relatively small trachytic cap overlies a dominant volume of basaltic magma, the layers being separated by a mugearitic layer. During periods of eruptive quiescence there are stable density interfaces between the main rock types. Each crystallizes against the side-walls and possibly the roof to form cumulates. The evidence from the intersti-
trachytemagma (compositionallyzoned?) ~ "csyenite umulates" ..~,.,'~;jI=~[F~ ~ PI--I- ~- ~-_FZ~.~ boundary migration uallayer liquids
~
i ~..~ ~'~
\gabbroic'cumulates"
Basalt SOUTHERN convecti?ngtrachyte volatile-rich,zoned;trachytic cap. magma mugearitelayer'~? ~ " / nite brnO;nadtarni~yel Y' "~ trachyte"~
L
~ ~
yelates' ~
~/~ ~l~amS~~ ~
--qt
gabbroic
Basalt -+ trachyte?? FIG. 13. Generalized models to show relationships in northern and southern set calderas prior to caldera collapse.
Silicic rocks of Kenya - - - l - -
I
r
331
-/. _ _ 5
8.0
/~./
-
.
4 0 ,.,
//
3
/
Complex Interl01a y of proceeses
/
7.0 1 ~
/
Crystal .~ l~quk:l processes
m
Menengal
6.0
/
o
'
t
L
200 "-
A
A
J
400 Rb
~
J
600
5.0
~ ' ~ " - ~ ~ _ ~ . -_ ~ _ .................. ~ .....
,..,,
o
4.0
/ ~, Southern -..
3.0
Eburru
Naivasha
Kenya Rift
,:.iD :,
2.0~
/ ;:..:
9
:: -/
Paka
lavas
1.0
k'x Emuruangogolak 0 0
I
I
loo
200
1 300
I _ 400
i
5oo
_
I
eoo
700
Rb(ppm) F]G. 14. K :O-Rb relationships in K enya basalt-trachyte + rhyolite suites. Generalized trends for the Emuruangogolak and southern Kenya rift sequences are shown as dotted lines where composition gaps exist. Menengai I, II and Ill are the trends for the first and second ash-flow sheets and the post-caldera pyroclastic rocks respectively. Data sources are as for Fig. 2. K-Rb space is divided into two overlapping areas where different genetic mechanisms are dominant (inset). tial glasses in the Silali nodules is taken to show that the trachytes (and probably mugearites) formed by fractional crystallization within these cumulate zones. They rose by boundary-layer migration (McBirney 1980; Turner & Gustafson 1981) towards the chamber roof where they were arrested at a level controlled by their density. The cap was probably compositionally zoned. No substantial volatile-rich cap was established. Caldera collapse was achieved mainly by lateral migration of basalts along fissures, disrupting the caldera walls. Small volumes of trachyte were erupted explosively to form a thin pumice mantle. The model of the southern centres is even more hypothetical. The dominant volume of basalt is based on thermal considerations and on the presence of small-volume mixed flows at Longonot. Most of the basalt may actually be stored in interconnecting dykes and sills, rather than as a single body. The gabbroic cumulates are hypothetical, as is the mugearite layer. No attempt is made to prejudge the origin of the trachyte magmas. Some may have come from fractional crystallization of basalt and some from deeper sources, e.g. by partial melting of the mafic lower crust. A critical feature is that the trachyte layer eventually splits into a volatile-rich upper layer and a volatile-poor (convecting?) lower layer. A complex interplay of processes within the cap
produces extremely differentiated residua and may ultimately lead to Krakatoa-type caldera formation. Figure 14 is a preliminary attempt to summarize in terms of K 2 0 and Rb the compositional fields in which each differentiation mechanism plays a dominant, or important, role. The Emuruangogolak, Paka and southern Kenya rift suites are taken to show positive correlations over most of their range, although the very marked K : O gaps in each case must be noted. These suites are used to define a field of crystal~,-~-liquid processes, the main mechanisms being partial melting and crystal fractionation, although magma mixing can produce spread within the field. In the case of fractional crystallization the inflection in the magmatic trends is probably due to a change of plagioclase-dominated to alkalifeldspar-dominated fractionating assemblages. Where partial melting is dominant, the inflection may reflect the differing partitioning of K and Rb into the volatile phase buffering the melting or the disappearance of a potassic phase from the crystalline residue. A second field encompasses the range of the Menengai zoned ash-flow sheets, the Menengai post-caldera tufts, pantelleritic trachytes and pantellerites from Eburru, and the Naivasha comendites. In this compositional range complex interplays of volatile complexing, vapour-phase
332
R. Macdonald
t r a n s p o r t , crystal f r a c t i o n a t i o n , m a g m a m i x i n g a n d crustal c o n t a m i n a t i o n h a v e occurred, probably w i t h i n the volatile-rich caps to the m a g m a c h a m b e r s . This d i a g r a m is a l m o s t certainly also a p p l i c a b l e to the p e r a l k a l i n e silicic rocks o f o t h e r p e t r o g r a p h i c provinces.
ACKNOWLEDGMENTS Thanks are due to Dr. P. T. Leat, Dr. S. C. Scott, Dr. R. S. J. Sparks and Dr. L. A. J. Williams for constructive criticism of the manuscript, and to Professor D. K. Bailey and Dr. L. A. J. Williams for guidance over many years. The Natural Environment Research Council has supported our Kenyan work via research studentships.
References BAILEY, D. K. 1964. Crustal warping--a possible tectonic control of alkaline magmatism. J. geophys. Res. 69, 1103-1 l. - 1974. Melting in the deep crust. In: SORENSEN,H. (ed.) The Alkaline Rocks, pp. 436-42. Wiley, London. --1978. Continental rifting and mantle degassing. In NEUMANN, E. R. & RAMBERG, I. B. (eds) Petrology and Geochemistry of Continental Rifts, pp. 1 13, Reidel, Dordrecht. - 1980. Volcanism, Earth degassing and replenished mantle lithosphere. Phil. Trans. R. Soc. Lond., Ser. A, 297, 309-22. -& MACDONALD,R. 1969. Alkali feldspar fractionation trends and the derivation of peralkaline liquids. Am. J. Sci. 267, 242-8. --- & 1975. Fluorine and chlorine in peralkaline liquids and the need for magma generation in an open system. Mineral. Mag. 40, 405-14. -& SCHAIRER,J. F. 1966. The system Na20-A1203 Fe203-SiO2 at 1 atmosphere, and the petrogenesis of alkaline rocks. J. Petrol. 7, 114-70. - - , COOPER, J. P. & KNIGHT, J. L. 1974. Anhydrous melting and crystallization of peralkaline obsidians. Bull. Volcanol. 38, 653-65. BAKER, B. H. 1987. Outline of the petrology of the Kenya rift alkaline province. In: FITTON, J. G. & UVTON, B. G. J. (eds.) Alkaline Igneous Rocks, Geol. Soc. Publ. 30, pp. 293-312. - - , GOLES, G. G., LEEMAN, W. P. & LINDSTROM,M. M. 1977. Geochemistry and petrogenesis of a basalt-benmoreite-trachyte suite from the southern part of the Gregory Rift, Kenya. Contrib. Mineral. Petrol. 64, 303-32. BOWEN, N. L. 1945. Phase equilibria bearing on the origin of the alkaline rocks. Am. J. Sci. 243A, 7589. BROWN, F. H. & CARMICHAEL,I. S. E. 1971. Quaternary volcanoes of the Lake Rudolf region, II, The lavas of North Island, South Island and the Barrier. Lithos, 4, 305-23. CLAGUE, D. A. 1978. The oceanic basalt-trachyte association: an explanation of the Daly gap. J. Geol. 86, 734 43. FERRARA, G. & TREUIL, M. 1974. Petrological implications of trace element and Sr isotope distributions in basalt-pantellerite series. Bull. Volcanol. 38, 548-74. GIBSON, I. L. 1970. A pantelleritic welded ash-flow from the Ethiopian Rift Valley. Contrib. Mineral. Petrol. 28, 89-111. HILDRETH, W. 1981. Gradients in silicic magma
chambers: implications for lithospheric magmatism. J. geophys. Res. 86, 10153-92. - 1983. Comment on 'Chemical differentiation of the Bishop Tuff and other high-silica magmas through crystallization processes'. Geology 11,6223. I M S L A N D , P . , L A R S E N , J . G . , P R E S T V I K , T . & SIGMOND, E. M. 1977. The geology and petrology of Bouvet~ya, south Atlantic Ocean. Lithos, 10, 213-34. JOHNSON, R. W. 1969. Volcanic geology of Mount Suswa. Phil. Trans. R. Soc. Lond., Ser. A, 265, 383412. JONES, W. B. 1979a. Syenite boulders associated with Kenyan trachyte volcanoes. Lithos, 12, 89-97. --1979b. Mixed benmoreite/trachyte flows from Kenya and their bearing on the Daly Gap. Geol. Mag. 116, 487-9. - 1981. Chemical effects of deuteric alteration in some Kenyan trachyte lavas. Mineral. Mag. 44, 279-85. LEAr, P. T. 1983. The structural and geochemical evolution of Menengai Caldera Volcano, Kenya Rift Valley. PhD thesis, University of Lancaster (Unpublished). - 1984. Geological evolution of the trachytic caldera volcano Menengai, Kenya Rift Valley. J. geol. Soc. Lond. 141, 1057-69. - - , MACDONALD,R. & SMITH, R. L. 1984. Geochemical evolution of the Menengai caldera volcano. J. geophys. Res. 89, 8571 92. MCBIRNEY, A. R. 1980. Mixing and unmixing of magmas. J. Volcanol. geotherm. Res. 7, 357-7l. MCCALL, G. J. H. 1967. Geology of the NakuruThompson's Falls-Lake Hannington area. Rep. geol. Surv. Kenya, 78. - 1968. The five caldera volcanoes of the central rift valley in Kenya. Geol. Soc. London Proc. 1647, 54-9. -1970. Gabbroic and ultramafic nodules: high level intracrustal nodular occurrences in alkalic basalts and associated volcanics from Kenya, described and compared with those of Hawaii. Phys. Earth planet. Inter. 3, 255-72. --& HORNUNG, G. 1972. A geochemical study of Silali volcano, Kenya, with special reference to the origin of intermediate-acid eruptives of the central rift valley. Tectonophysics, 15, 97-113. MACDONALD, R., BAILEY, D. K. & SUTHERLAND,D. S. 1970. Oversaturated peralkaline glassy trachytes from Kenya. J. Petrol. 11,507-17. MAHOOD, G. A. 1981. The chemical evolution of a late Pleistocene rhyolitic centre : The Sierra La Prima-
Silicic rocks of Kenya vera, Jalisco, Mexico. Contrib. Mineral. Petrol. 77, 129-49. 1984. Pyroclastic rocks and calderas associated with strongly peralkaline magmatism. J. geophys. Res. 89, 8540-52. MICHAEL, P. J. 1983. Reply to comment on 'Chemical differentiation of the Bishop Tuff and other highsilica magmas through crystallization processes'. Geology, 11,623-4. MILLER, C. F. &. MITTLEFEHLDT,D. W. 1984. Extreme fractionation in felsic magma chambers : a product of liquid-state diffusion or fractional crystallization? Earth planet. Sct.Lett. 68, 151 8. NAKAMURA, N. 1974. Determination of REE, Ba, Fe, Mg, Na and K in carbonaceous and ordinary chondrites. Geochim. cosmochim Acta, 38, 757-75. NASH, W. P., CARMICHAEL,I. S. E. & JOHNSON, R. W. 1969. The mineralogy and petrology of Mount Suswa, Kenya. J. Petrol. 10, 409-39. NOBLE, D. C., RIGOT, W. L. & BOWMAN, H. R. 1979. Rare-earth-element content of some highly differentiated ash-flow tufts and lavas. Spec. Pap. geol. Soc. Am. 180, 77-85. NORRY, M. J., TRUCKLE, P. J., LIPPARD, S. J., HAWKESWORTH, C. J., WEAVER, S. D. & MARRINER, G. F. 1980. Isotopic and trace element evidence from lavas, bearing on mantle heterogeneity beneath Kenya. Phil. Trans. R. Soc. Lond., Ser. A, 297, 259-71. PEARCE, J. A. 1983. Role of the sub-continental lithosphere in magma genesis at active continental margins. In: HAWKESWORTH, C. J. & NORRY, M. J. (eds) Continental Basalts and Mantle Xenoliths, pp. 230-49. Shiva, Orpington. SCEAL, J. S. C. & WEAVER, S. D. 1971. Trace element data bearing on the origin of salic rocks from the Quaternary volcano Paka, Gregory Rift, Kenya. Earth planet. Sci. Lett. 12, 327-31. SCOTT, S. C. 1980. The geology of Longonot volcano, central Kenya: a question of volumes. Phil. Trans. R. Soc. Lond., Ser. A, 296, 437-65. -1982. Evidence from Longonot volcano, Central Kenya, lending further support to the argument for a coexisting CO_, rich vapour in peralkaline magma. Geol. Mag. 119, 215 7. - & BAILEY, D. K. 1984. Coeruption of contrasting magmas and temporal variations in magma chemistry at Longonot volcano, Central Kenya. Bull. Volcanol. 47, 849-73.
333
SMITH, R. L. 1979. Ash flow magmatism. Spec. Pap. geol. Soc. Am. 180, 5-27. SMITH,W. C. 1938. Petrographic description of volcanic rocks from Turkana, Kenya Colony, with notes on their field occurrence from the manuscript of Mr. A. M. Champion. Q. J. geol. Soc. Lond. 94, 507 50. SUTHERLAND,D. S. 1974. Petrography and mineralogy of the peralkaline silicic rocks. Bull. Volcanol. 38, 517-47. THOMPSON, R. N. 1982a. Geochemistry and magma genesis. In SUTHERLAND,D. S. (ed.) Igneous Rocks of the British Isles, pp 461-77. Wiley, London. - 1982b. Magmatism of the British Tertiary Province. Scott. J. Geol. 18, 49-107. , GIBSON, I. L., MARRINER, G. F., MATTEY,D. P., MORRISON, M. A. 1980. Trace-element evidence of multistage mantle fusion and polybaric fractional crystallization in the Palaeocene lavas of Skye, NW Scotland. J. Petrol. 21,265-93. TURNER, J . S . & Gus'rAFSON, L. B. 1981. Fluid motions and compositional gradients produced by crystallization or melting at vertical boundaries. J. Voleanol. geotherm. Res. 11, 93 125. WEAVER, S. D. 1978. The Quaternary caldera volcano Emuruangogolak, Kenya Rift, and the petrology of a bimodal ferrobasalt-pantelleritic trachyte association. Bull. Volcanol. 40, 209-30. , SCEAL, J. S. C. & GIBSON, I. L. 1972. Trace element data relevant to the origin of trachytic and pantelleritic lavas in the East African Rift System. Contrib. Mineral. Petrol. 36, 181-94. WILLIAMS, L. A. J. 1978a. The volcanological development of the Kenya Rift. In: NEUMANN, E. R. & RAMBERG, I. B. (eds) Petrology and Geochemistry of Continental Rifts pp. 101-21. Reidel, Dordrecht. - 1978b. Character of Quaternary volcanism in the Gregory Rift Valley. In: BISHOV, W. W. (ed.) Geological Background to Fossil Man, pp. 56-69. Scottish Academic Press, Edinburgh. - 1982. Physical aspects ofmagmatism in continental rifts. In." PALMASON, G. (ed.) Continental and Oceanic Rifts, Geodynamics Series 8, pp. 193-222. American Geophysical Union, Washington, DC. --, MACDONALD, R. & CHAPMAN, G. R. 1984. Late Quaternary caldera volcanoes of the Kenya Rift Valley. J. geophys. Res. 89, 8553-70. WOLFF, J. A. & WRIGHT, J. V. 1982. Formation of the Green Tuff, Pantelleria. Bull. Volcanol. 44, 68190.
R. MACDONALD,Department of Environmental Science, University of Lancaster, Bailrigg, Lancaster LA1 4YQ, U.K.
The petrochemistry of the northern part of the Chilwa alkaline province, Malawi A. R. Woolley & G. C. Jones S U M M A R Y: The Chilwa province of alkaline igneous rocks and carbonatites lies at the southern end of the East African rift, and is unique within the rift for its essentially intrusive nature. The province comprises numerous carbonatite centres (some with nepheline syenite and nephelinite), large complexes of nepheline syenite and syenite, and plutons of peralkaline syenite, quartz syenite and granite, together with dykes compositionally equivalent to all the major rock types. 196 rock samples from the northern half of the province have been chemically analysed and the data indicate that three rock series are present: (1) syenitequartz syenite-granite, (2) nepheline syenite-syenite and (3) nephelinite-carbonatitenepheline syenite. It has not proved possible to relate these series to a single parental magma, and it is concluded that three parental magmas of trachytic, phonolitic and nepheliniticcarbonatitic composition were involved. A model is suggested for the generation of these three magmas involving production of a zoned metasomatized lithosphere wedge by lithosphere focussing. Fracturing of the crust above the growing wedge led to pressure release and to the uprise of geotherms and volatiles with consequent partial melting at different levels throughout the metasomatized wedge. The metasomatism may have reached the base of the crust and it was from this region that trachyte magmas were produced by melting of rocks similar to syenitic fenites. Phonolitic magmas were generated directly at intermediate levels, and nephelinites were generated at the base of the metasomatized lithosphere.
Introduction The East African rift system can be traced almost continuously from the Red Sea to just S of Lake Malawi. At its southern extremity it breaks up into a series of faults which merge into those of the approximately E-W-trending Zambesi rift structure. Numerous major faults which merge into monoclinal structures can be traced southwards through Mozambique into South Africa, with the change in structural style corresponding to a decrease in the alkalinity of the associated igneous rocks. The Chilwa alkaline province, originally known as the Chilwa series (Dixey et al. 1937), is the most southerly manifestation of the strongly alkaline volcanism associated with the East African rift. It is the oldest igneous province along the rift, with ages ranging from 135 to 105 Ma (Woolley & Garson 1970, Table 1), and is unique to the rift in being essentially intrusive at the present level of erosion. Although the province is centred on southern Malawi, its rocks are also found in adjacent parts of Mozambique (see Woolley & Garson 1970, Fig. 2); the province as a whole has a diameter of 300-400kin. The major intrusions comprise plutons of peralkaline granite and syenite, the largest of which, Mulanje, covers 640 km 2, and somewhat smaller complexes of nepheline syenite and syenite. There are 17 carbonatite centres in Malawi (Garson 1966), some of which include
nephelinite and nepheline syenite, with a number also in Mozambique; extensive dyke-swarms are also present (Vail 1964). In the Zambesi Valley to the south the Lupata Series includes flows of phonolite, some of which are analcime-phyric (blairmorites), which Garson (1962) considered part of the province. About 30 Ma before the onset of Chilwa magmatism the last manifestations of the Karoo volcanic cycle were extruded in the form of rhyolitic and tholeiitic basalt flows, which lie beneath the Lupata volcanics and extend into southern Malawi (Macdonald et al. 1983), together with an intense tholeiitic dolerite dyke-swarm (Woolley et al. 1979). A general account of the igneous rocks and tectonics of the province is given by Woolley & Garson (1970). Since the pioneering work of Dixey et al. (1937) all of southern Malawi has been mapped geologically and described in detail, principally by the officers of the Geological Survey of Malawi, and numerous maps, memoirs and papers have been devoted specifically to the centres of the Chilwa province. Although these descriptions include a considerable amount of petrography and some chemical analyses, no detailed petrological, mineralogical or geochemical studies of the province have yet been undertaken. The initial results of a geochemical study of the northern part of the province are presented in this paper. The justification for this more detailed work lies in the exceptionally broad range of rock types that
From" FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 335-355.
335
336
A. R. Woolley & G. C. Jones
occur, ranging from granite through nepheline syenite and nephelinite to carbonatite, and in their overwhelmingly felsic character; both these features pose exceptional problems of petrogenesis. Further, the essentially intrusive nature of the province should provide some insight into the deeper levels of the volcanic provinces further north. The northern part of the Chilwa province Within Malawi the Chilwa province can be divided into northern and southern sections, both of which include representatives of all the major rock types of the province. The present account covers only the northern half of the province (Fig. 1). The area is covered on a scale of 1:250 000 by Sheet II of the GeologicalAtlas of Malawi, and on a scale of 1:100 000 by the Zomba sheet which accompanies a detailed account of all the geology of the vicinity (Bloomfield 1965). The largest intrusion of the northern part of the province is the 22 x 23 km complex of Zomba and Malosa (Bloomfield 1965, p. 96). It consists of granite, quartz syenite and syenite, and cuts paragneisses and charnockitic granulites of the Mozambique belt, but to the W it is truncated by the main NE-trending rift valley fault. The complex forms a remarkable massif, rising to
;~'"
X
,,S"
JungunJ
/
., j':' ~.; - - Z,, Mong olowe , 7 ~ j ! i : C ~ a i ; e ~,;+2,+ ,-C hikala ~i~.: ) . .... ,,~ .....~+,~"~17:.-"c7.;,:);-:7;{.,7".+"7hr"~ ..'-,~-?',WI-;'_',t),;7.G:--/-~L~_'~7;,- ;z~'
"'-,..
"
over 2000 m and surrounded on all but its northern sides by precipitous cliffs. Mpyupyu (Bloomfield 1965, p. 100) is a conical hill 1.5-2.0km in diameter standing in the alluvial plain west of Lake Chilwa; this intrusion consists of syenite and quartz syenite and is considered to be consanguineous with Zomba-Malosa. The E - W line of complexes N of Malosa (Fig. 1) form gently-rolling wooded hills. The complexes increase in age from E to W and consist predominantly of nepheline syenite, although a little quartz syenite and syenite are also present, notably on the E side of Chikala and along the SE margin of Mongolowe. All four complexes are multiple with successive intrusions, often in annular form, particularly in Chikala at the E end of the chain. Chikala was mapped by Garson (1960) and Stillman & Cox (1960), the adjacent Chaone Hill by Vail & Monkman (1960) and Mongolowe, which is further W, by Vail & Mallick (1965). Mongolowe makes contact over a short distance with the northern margin of Malosa and the contact was interpreted by Vail & Mallick (1965) as indicating that Mongolowe is the younger of the two. Chinduzi is the least well known, having been mapped by Bloomfield (1965, p. 111) on the basis of only three traverses. There is patchy fenitization of the basement around all four complexes. A number of agglom-
"
"
,"
~
-* - ' - ' " '-- '--4 "~'~1-7"L' r/,,~';
4" %20"".'-t r''-'''st.
~- ,~ ~,'
_ , ~..'-...:~r .\~_'<;: "L7-'7
' ~ , ~ - ~ . "~',/
,,,',,5.: .... ~ ;~ ~\ / ~ ~ ~ ,,-' ',~
",,. / " ,.
p ~ .~,~.~J~-~
----.._
--
..-~
Syenite-granite Granite ring dyke Nepheline syenite. syenite Carbonatite /Solvsbergite - --'-" Phonolite I~" Major faults
:!:!::!:!:7:i :!:!::!:!:!:!:i:!:!:::[: !!!ii!i! i![: # ............. 9"= ::i i[ ii! iii! :: i :: ~ t , ',
A /
t
/
" , t,~
,
/
k a Ke Chilwa
~
~
I
/
( f t;
z
Chilwa ~ " ) Island
'~ / ' - ~ I e..:~.,' ~ ~w~. Ivipyupyu
\
4
/ / /
' ....
X,', , ,' /
,
/ -
>" J"
~ "'75-';-/"
~#"J t
# I /
/ [
0 5 miles O ~ - ~ k m
j /
FIG. 1. The northern part of the Chilwa alkaline igneous province (mainly after Bloomfield 1965, Plate XII).
Petrochemistry of the Chilwa alkaline province erate-choked vents occur (Bloomfield 1965, p. 130; Garson 1965b, p. 91)and areas of metasomatism and alteration of the nepheline syenites are present. Xenoliths of metamorphosed basic volcanic rocks occur in all four complexes and there are a few in-faulted masses up to a kilometre across in Chikala and Chaone. Junguni is an isolated intrusion of nepheline syenite lying to the N of Mongolowe (Bloomfield 1965, p. 111). Contacts with the basement are obscured by Recent sediments, but are exposed in some small hills NE of the main intrusion where the country rocks show the effects of fenitization. Chilwa Island is a multiple carbonatite complex and was the first carbonatite to be recognized in Africa (Dixey et al. 1937). It has been described in detail by Garson & Smith (1958). The present account is not directly concerned with the carbonatites, but rather with the associated plugs and dykes which consist of alkaline silicate rocks including nephelinite, nepheline syenite and aln6ite. The Kangankunde carbonatite (Garson 1965a) lies to the W of the area shown on Fig. 1, but does not have associated silicate rocks although some dykes could be carbonatized nephelinite or aln6ite (Garson 1965a, pp. 45-8). Dykes of microgranite, s61vsbergite, microsyenite and phonolite are widespread (Bloomfield 1965, p. 133 and Plate XII); the s61vsbergites represent the most north-easterly part of a swarm 100 km long extending from Salambidwe on the Mozambique border. Phonolite dykes form a swarm W of Chinduzi (Fig. 1) and are found in the vicinity of, and cutting, all the nepheline syenite complexes.
Petrography Granite, quartz syenite and syenite
337
from the main complex over much of its length by a screen of basement gneiss (Fig. 1). Granites also occur on the western and northern margins but here appear to grade inwards into quartz syenite and syenite. Malosa contains a more heterogeneous suite of syenitic and granitic rocks than Zomba. The syenite of the central plug of Zomba is a coarse rock of perthite, sometimes containing a little quartz, hedenbergitic pyroxene, often zoned to green rims and sometimes mantled by amphibole, variably altered fayalite, a little reddish biotite, opaque oxides and accessory apatite, zircon and pyrite. Malosa rocks are similar but are free from fayalite, while, particularly in the N close to the contact with the nepheline syenites, a variety containing aegirine and arfvedsonite occurs. The rocks of the inner syenitic ring of Zomba consist of variably perthitic feldspar phenocrysts in a groundmass of alkali feldspar, a little quartz, hedenbergitic pyroxene, usually rimmed by amphibole which also forms independent poikilitic crystals, a little altered fayalite and accessories. The granites are generally coarse rocks consisting of tabular perthites, with interstitial quartz, aegirine-(aegirine-augite), riebeckitic and arfvedsonitic amphibole, fayalite or its alteration products and accessories including biotite and pyrochlore. Dyke rocks include aegirine-aenigmatite microgranite and alkali feldspar-quartz porphyries with a felsitic groundmass. Mpyupyu
The rocks of Mpyupyu Hill are hornblendebiotite syenites, sometimes with a little quartz. Perthite forms large very irregular plates. Brown hornblende contains rare pyroxene cores, while biotite mantles amphibole. Pyrochlore-bearing aegirine-riebeckite syenite outcrops just S of the Hill (Garson 1960, p. 32).
Zomba-Malosa
The field relationships and petrography of the Zomba-Malosa complex have been described in some detail by Bloomfield (1965, pp. 96-107). The following brief petrographic account is based on rocks collected by Woolley together with sections made from specimens originally collected by Bloomfield. Bloomfield (1965, p. 96) divided the Zomba part of the complex into a central syenitic plug, an inner syenitic ring and an outer granite ring. The outer ring is a welldefined peripheral ring-dyke some 100 m wide which can be followed for about 25 km along the S and E boundaries of the complex; it is separated
Saturated and oversaturated dykes A regional swarm of trachyte and quartz trachyte dykes, generally referred to in the Malawi geological literature as s61vsbergites, extends approximately from the vicinity of ZombaMalosa to the Salambidwe complex to the SW. They were described by Bloomfield (1965, p. 138) and in references cited therein. They are almost invariably slightly altered with turbid subhedral short prismatic alkali feldspars, often aligned; interstitial quartz varies modally from zero to 15%, and ubiquitous opaque oxides pseudomorph
338
A. R. Woolley & G. C. Jones
mafic minerals. Bloomfield (1965, p. 139) identified riebeckite and aegirine in a few specimens. Some varieties are rich in brown biotite.
Nepheline syenite and syenite Chikala, Chaone, Mongolowe and Chinduzi Chikala, which was mapped by Garson (1960) and Stillman & Cox (1960), Chaone and Mongolowe, which were mapped by Vail & Monkman (1960) and Vail & Mallick (1965), and Chinduzi, which was mapped by Bloomfield (1965), all consist of rocks described as perthosite, pulaskite and foyaite or microfoyaite. There was some slight variation in the definitions of these rock types as used by the above workers but essentially the perthosites are rocks rich in alkali feldspar (perthite) with up to 5% quartz, the pulaskites contain up to 5}/o nepheline and 5%-50~o mafic minerals, and the foyaites contain 5%-50% feldspathoids. The present study has confirmed this range of rock types and indicated a continuous transition from rocks with a little quartz to nepheline syenites with more than 30% nepheline. On Chaone and Mongolowe the field relationships suggest that the perthosites were emplaced first, followed by the pulaskites and then foyaites. Perthosite is the least abundant rock type, occurring principally along the eastern margin of Chikala and the southern margins of Chaone and Mongolowe. It is a coarse rock of large plates of perthite, sometimes a little quartz, a greenish amphibole with more rarely an arfvedsonitic type, often with cores of aegirine-augite, and a little brown biotite. The pulaskites and foyaites vary from coarse rocks with feldspar plates over 1 cm in diameter and patches of mafic minerals of a similar size to finer-grained more homogeneous varieties. The feldspar is invariably perthitic displaying a wide range of complex exsolution and replacement textures. Nepheline is usually fresh with only minor alteration to cancrinite. The nepheline may be interstitial or form short stout subhedral prisms and rather rounded patches enclosed in feldspar. Interstitial sodalite occurs in some of the foyaites, and in Chikala may develop as approximately circular patches, poikilitically including myriads of small feldspars. In Chikala, Chaone and Mongolowe amphibole, usually katophorite (Woolley & Platt 1986), is the principal mafic phase, but sodic pyroxenes and brown biotite are also abundant. Most of the rocks of Chinduzi are aegirinebiotite-nepheline syenites, but amphibole occurs in about one-third of the samples studied. Arfvedsonitic and riebeckitic amphiboles are
found in western Mongolowe, just N of Malosa, but elsewhere the amphibole is a katophorite. Although mica, pyroxene and amphibole are generally clearly primary crystallizing phases, textures commonly indicate some secondary recrystallization, invariably with mica forming at the expense of pyroxene. An opaque oxide phase is ubiquitous, sphene and apatite are abundant, and a little zircon is sometimes present.
Junguni These nepheline syenites contain the most feldspathoid of all the rocks of the area. Perthitic alkali feldspar forms subhedral prisms amongst which nepheline may be interstitial or form stout fresh prisms. Aegirine is abundant, biotite occurs in moderate amounts and amphibole is rare. Sodalite, referred to by Bloomfield (1965, p. 122) as analcite, forms rounded masses, interstitial areas and, in a few rocks, rectangular areas within feldspar prisms which are clearly secondary after nepheline. Sodalite forms between 80 and 90 vol.% of two specimens. Accessories include opaque oxides, sphene, carbonate, apatite and pyrite. Pegmatitic nepheline syenites on Kadongosi Hill, just to the NE of Junguni, include large blades of alkali amphibole, fluorite and several unidentified accessories; Bloomfield (1965, p. 122) described eudialyte from these rocks.
Chilwa Island Nepheline syenite forms a small plug about 100 m in diameter and several sheets cutting s6vite on the NW side of the Chilwa Island carbonatite plug, as well as several microfoyaite dykes in fenitized basement rocks. The nepheline syenites are relatively homogeneous medium-grained rocks comprising alkali feldspar, nepheline, aegirine-(aegirine-augite), a little brown biotite and accessories. The microfoyaites are similar except for the porphyritic nature of the nepheline and the presence of melanite at one locality.
Phonolite dykes The phonolites are texturally and mineralogically rather variable. Alkali feldspar, nepheline, aegirine, alkali amphibole and biotite may occur as phenocrysts. The groundmass is generally a felted mass of prismatic alkali feldspar and nepheline, usually with numerous tiny aegirine needles, while biotite is widespread. Zeolite, probably natrolite, is prominent in some samples and possible sodalite may occur.
Petrochemistry of the Chilwa alkaline province Ijolite, nephelinite, camptonite and alnOite These rock types form dykes on Chilwa Island, usually within fenitized basement rocks but also cutting carbonatite. Ijolite forms one dyke-like mass in K-rich fenites and is characteristically extremely heterogeneous. It consists principally of nepheline, aegirine-augite, melanite and wollastonite, the last sometimes as stellate clusters several centimetres in diameter. K-feldspar is present in some samples and is probably the result of late feldspathization. The nephelinites consist of phenocrysts of olivine and pyroxene in a dense groundmass of pinkish augite prisms, abundant opaque oxides, a little biotite and turbid interstitial material probably composed of zeolite, sericite and possibly alkali feldspar. Only one specimen containing fresh nepheline has been found. Rocks described as camptonites by Garson & Smith (1958, p. 79) occur as concentric dykes in the basement and as a plug within the s6vite. The rock of the plug comprises phenocrysts of olivine, augite and a red-brown amphibole in a groundmass of pyroxene, amphibole, opaque oxides, sodic plagioclase and an unidentified turbid material. A sample from a dyke contains scarce phenocrysts ofaugite and olivine in a groundmass of pyroxene, olivine, reddish-brown amphibole, an opaque phase, probable feldspar and zeolite. Aln6ite, from a dyke 300 m long cutting s6vite, consists of 50% carbonate with numerous phenocrysts of biotite, reddish amphibole, pyroxene, apatite and an opaque phase.
Metavolcanic rocks A collection of metamorphosed basic lavas was made from outcrops on the NW margin of Chaone, which is faulted against basement gneisses but cut by nepheline syenite. These are compact medium-grained melanocratic rocks which are sometimes vaguely foliated and may be cut by a few white veinlets. The dominant minerals are a pale-green pyroxene and a hornblendic amphibole, either of which may be concentrated in monomineralic layers. An opaque phase is abundant and biotite is found in some samples. Nepheline is the dominant felsic mineral with alkali feldspar present in more mesocratic rocks; both may form narrow veinlets.
Petrochemistry 196 rock samples have been analysed for this study. There are 14 published analyses for the area, some of which include a few trace-element
339
data, but these were determined spectrographically. Approximately two-thirds of the newly analysed samples were collected by Woolley, and about 50 were acquired from the collection of the Geological Survey of Malawi. Most of the latter were originally collected by K. Bloomfield over Zomba and Malosa and from widely distributed phonolite dykes, and a few by M. S. Garson in the eastern part of Chikala. Four of the Chikala samples analysed come from the University of Leeds and were originally collected by C. J. Stillman and K. G. Cox. Twelve of the analysed specimens are from the collection of the British Museum (Natural History), the majority of these having been collected by Garson during his mapping of Chilwa Island. Although some of the specimens acquired from these sources were rather small, the material collected specifically for this study consisted of large pieces, generally several kilograms in weight. This material, after initial reduction by screw press and jaw crusher, was subdivided using a sample splitter. A portion was further reduced in a tungsten carbide disc mill to pass 100 gm mesh for analysis. The rocks were analysed as follows. Si, Ti, A1, Fe, Mn, Mg, Ca, Na, K, P, Nb, Zr, Y and La were determined by X-ray fluorescence (XRF) using lithium metaborate discs, and C1 was determined using pressed powder discs. Calibration was with international standard rocks; a few of the samples were analysed gravimetrically and used as standards for the remainder. Li, Be, V, Cr, Ni, Cu, Zn, Rb, Sr, Cs, Ba and Pb were determined by atomic absorption spectroscopy (AAS) in HF-H3BO3 solutions; calibration in this case was performed with synthetic mixures checked against standard rocks. H20 + and CO2 were determined by means of a C H N elemental analyser, FeO was determined by titration after dissolution in HF-H2SO4 and F - was determined by pyrohydrolysis with colorimetric or ionselective electrode finish. A selection of analyses is presented in Tables 1-4, but copies of all the analyses are obtainable from the authors and a full data set has been deposited in the Mineralogy Library of the British Museum (Natural History).
Granite, quartz syenite and syenite Analyses of 43 rock samples from Zomba-Malosa and five from Mpyupyu have been made, and a further six are available in the literature (Dixey et al. 1937; Bloomfield 1965), although all the latter are from the southern end of Zomba Mountain. A selection of new analyses and their normative compositions is given in Table 1. Of
340
A. R. Woolley & G. C. Jones
TABLE 1. A n a l y s e s o f granites, quartz syenites a n d syenites f r o m the Z o m b a - M a l o s a c o m p l e x a n d
M p y u p y u , a n d o f a sblvsbergite d y k e 1
2
10
11
12
13
69.6 0.45 13.0 3.39 1.24 0.08 0.19 0.63 5.4 4.82 0.46 0.32 0.03 0.09 0.72
56.9 1.67 15.1 2.22 7.50 0.25 1.90 4.58 4.6 3.95 0.54 0.06 0.65 0.15 0.44
61.6 0.77 15.8 2.34 4.62 0.25 0.42 1.25 6.1 5.83 0.36 0.08 0.07 0.06 0.36
62.1 0.51 18.4 0.80 2.75 0.11 0.51 2.44 5.9 5.18 0.24 0.11 0.17 0.07 0.43
99.98 100.61 100.46 100.06 100.40 100.02 100.05 100.42 100.51
99.91
99.72 100.28
3
4
5
6
7
8
9
14
Major elements (wt. ~ ) SiO 2 TiO 2 AI203 Fe203 FeO MnO MgO CaO Na20 K.,O H20* H_,OPzO5 CO, Others Total
70.6 0.29 13.7 1.61 2.23 0.13 0.18 1.06 4.4 5.21 0.33 0.10 0.04 0.10 0.52
72.0 0.25 13.2 1.47 1.79 0.09 0.06 0.45 5.0 4.88 0.31 0.14 0.02 0.13 0.42
100.50 100.21
73.3 0.22 13.3 1.02 1.61 0.09 -0.55 4.7 4.50 0.24 0.08 0.04 0.04 0.29
72.0 0.28 10.4 6.39 1.93 0.15 0.10 0.48 3.9 4.33 0.25 0.03 0.02 0.07 0.28
73.3 0.14 12.5 2.35 1.04 0.08 0.07 0.26 4.8 4.74 0.27 0.16 0.02 0.13 0.60
74.4 0.01 14.9 0.10 0.04 0.01 0.02 -4.9 4.67 0.50 0.17 0.02 0.15 0.17
69.0 0.31 14.7 1.21 2.35 0.10 0.11 0.83 5.1 5.56 0.28 0.11 0.03 0.06 0.65
62.0 0.99 15.8 1.77 4.64 0.17 0.94 2.73 4.7 5.07 0.41 -0.30 0.08 0.42
63.3 0.78 15.7 1.32 4.49 0.20 0.58 2.27 4.8 5.46 0.36 0.07 0.24 0.06 0.42
60.1 0.51 18.1 6.33 -0.27 0.24 0.46 5.8 5.86 1.46 0.44 0.31 0.10 0.30
Trace elements (ppm) F CI Be Cr Li Nb Ni Cu Zn V Zr Y Sr Ba Rb
CI P W norms Q c or ab an ac di hy mt hm il ap ce H2 O+ H:O-
3010 1800 5
1290 200 5
1260 220 2
2570 420 10
610 930 6
760 160 4
1240 100 4
1200 610 3
1030 810 3
3860 230 18
1290 570 2
1610 580 8
570 730 2
650 3
14 100
8 100
6 60
13 --
33 100
9 200
12 100
14 70
7 70
58 500
12 65
35 190
7 42
7 215
. 210 . 2200 105 15 370 180
-. 100 50 16 220 580
110 . 3600 65 6 -150
. 115 . 585 75 200 1065 95
145
200
545 80 100 1085 85
2200 220 19 60 350
14 150 55 500 70 310 1260 50
230 -515 100 35 260 170
28.1 1.8 27.6 41.3 -. ---0.1
15.5 . 32.9 43.2 0.8
7.7 . 32.3 40.6 5.2
19.9 . 28.5 40.0 -5.3 1.8 -2,3 . 0.9 0.1 0.2 0.5 0.3
13 95
. 310
--
--
130 . 1000 110 28 160 170
100 . 1800 50 6 -120
850 75 8 30 90
100 50 10 80 210
22.1 . 30.8 37.4 2.2
23.2 27.0 . . . 28.8 26.6 40.7 39.8 -1.9
33.2 . 25.6 29.4 -2.8 0.5 -5.9
--
1.9 2.0 2.3
1.4
1.1 1.8 1.4
.
.
--
0.3 1.8 1.5
--
--
--
0.6 0.1 0.2 0.3 0.1
0.5 0.1 0.3 0.3 0.1
0.4 0.1 0.1 0.2 0.1
.
1.4
0.5 0.1 0.2 0.3 --
27.2 28.0 37.9 -2.7 0.3 0.7 2.1 --
0.3 0.1 0.3 0.3 0.2
.
0.1 0.3 0.5 0.2
.
. 2.5 2.0 1.8 . 0.6 0.1 0.1 0.3 0.1
6.9 . 30.0 39.8 7.1 . 5.1 5.5 2.6 . 1.9 0.7 0.2 0.4 --
.
3.6 5.8 1.9 .
. t.5 0.6 0.1 0.4 0.1
--
2.0 23.4 39.2 8.8
34.5 48.8 --
--
7.4 10.6 3.2 . 3.2 1.5 0.3 0.5 0.1
2.7
--
--
46 -310 27 590 1950 60
98 -580 77 120 720 142
0.2 30.6 49.9 8.4
3.9 2.3 34.6 48.7 --
--
--
4.7 3.3 2.1
1.8 4.1 1.2
1.5 0.2 0.1 0.4 0.1
1.0 0.4 0.2 0.2 0.1
.
-0.5 -6.3 0.6 0.7 0.1 1.5 0.4
1, Amphibole-pyroxenegranite, Zomba (1980, PI9, 16); 2, amphibole-pyroxene granite, Zomba (1937, 232, 10); 3. riebeckiteaegirine granite, Zomba (1980, PI8, 15); 4, riebeckite-aegirine granite, Malosa (1980, PI9, 30); 5, aegirine-riebeckite granite, Malosa (1980, P29, 28); 6, microgranite (dyke), Malosa (1980, P19, 31); 7, amphibole-pyroxene-quartz syenite, Zomba (1980, P18, 13); 8, pyroxene-amphibole-biotite-quartz syenite, Zomba (1980, P19, 1); 9, amphibole-pyroxene-quartz syenite, Zomba (1980, P18, 20); 10, riebeckite-pyroxene-quartz syenite, Malosa (1980, P19, 33); 11, pyroxene-amphibole-fayalite syenite, Malosa (1980, P19, 18); 12, amphibole syenite, Malosa (1980, P19, 11); 13, amphibole-biotite-quartz syenite, Mpyupyu, (1980, P26, 6); 14, s61vsbergite, Chilwa Island (1957, 1056, 172). - - , element below the detection limits, which are as follows: F, 10 ppm; C1, 10 ppm; Be, 1 ppm; Cr, 5 p p m ; Li, 1 ppm; Nb, 10 ppm; Ni, 25 ppm; Cu, 10 ppm; Zn, 1 ppm; V, 50 ppm; Zr, 10 p p m ; Y, 10 ppm; Sr, 5 ppm; Ba, 25 p p m ; Rb, 5 ppm. Analysts, G. C. Jones and V. K. Din. The numbers are British Museum (Natural History) Rock Collection numbers.
Petrochemistry of the Chilwa alkaline province the newly analysed samples 14 have more than 20% normative quartz and are therefore granites, taking the boundary as recommended by Streckeisen (1976) although modal mineralogy is the basis of that system. Of the remaining 34 rocks, 27 are quartz normative, four contain a little nepheline and three are exactly silica saturated. Two of the ne-normative rocks and two of the saturated examples come from the Kasupe area at the extreme northern end of Malosa, close to the contact with the nepheline syenites of Mongolowe, while one of the other ne-normative rocks comes from a small isolated mass of syenite near the NW end of Malosa, again near Mongolowe. There appears to have been some alteration of the rocks at the northern end of Malosa caused by the emplacement of the Mongolowe complex, and this has affected concentrations of several elements; this must be taken into account when considering the significance of the chemical data. There is a distinct bimodality in silica content, with no analyses in the range 64.6%-68.8% SiO2 (Fig. 2). These rocks are also illustrated by the Qz-Ne-Ks diagram of Fig. 3, in which they plot in the 'low-temperature trough' between the granite and feldspar minima at PH2o=I kb (Tuttle & Bowen 1958). 18 of the newly analysed rocks prove to be chemically peralkaline, as indicated by the presence of acmite in the norm. Half the granites fall into this category as well as one-third of the syenites. However, all the latter come from the Kasupe area at the northern end of Malosa which, as mentioned above, is considered to have been affected by the intrusion of Mongolowe. One possible effect of this intrusion would be the introduction of sodium. The variation in the peralkalinity of the suite is illustrated by the SiO2-A1203-(Na20+K20) diagram (Bailey & Macdonald 1969) of Fig. 4. Differentiation at Zomba-Malosa was from syenites to granites and the trend illustrated by Fig. 4 could be generated by fractionation of alkali feldspar, which is invariably slightly peraluminous. The peralkaline granites are chemically very similar to comendites, which lie in the SiO2 oversaturated part of Fig. 4 between the two lines radiating from the Qz apex; the broken line separates comendites from pantellerites according to Macdonald & Bailey (1973, Fig. 6). However, there was certainly some alkali loss from the Zomba-Malosa intrusion during and after emplacement, which is clearly shown at the petrologically similar Mulanje Complex by peripheral fenitization. It is inferred, therefore, that the primary syenite magma probably had an original composition slightly more alkaline than the centre of the syenite cluster in Fig. 4, although it was not quite
341
peralkaline. The syenite differentiated towards more peralkaline compositions, almost certainly by alkali feldspar fractionation (the plagioclase effect) as no other crystallizing phase would enhance the alkali-to-alumina ratio of the liquid. The highly evolved nature of most of the rocks of the province precludes the use of variation diagrams based on Mg- and Fe-dominated differentiation indices. The greatest variation is in SiO2 and this proves to be the most useful plotting index. A range of oxides is plotted against SiO2 in Fig. 2, and most of the Zomba-Malosa and Mpyupyu rocks define simple coherent linear trends which can be interpreted as reflecting a liquid line of descent. Even Ba, which encompasses a great range in the syenites, produces a welldefined trend to low values in the granites, presumably controlled by alkali feldspar and possibly mica fractionation. One group ofsyenites plots markedly off the trends for some elements, e.g. CaO. The majority of these rocks are from the northern end of Malosa and are thought to have been affected by emplacement of the Mongolowe complex. The Ba: Sr ratio is about 8 : 1 and is reasonably linear (Fig. 5). F is generally greater than C1 with some high F:C1 ratios (Fig. 6). For comparison, the shaded area on Fig. 6 is for peralkaline oversaturated glassy lavas from East Africa (Bailey 1977, Fig. 1l). Clearly, there will have been considerable loss of halogens from the Zomba-Malosa rocks but, allowing for this, the patterns are comparable and suggest the early fixing of F and CI, presumably in amphiboles and micas.
Oversaturated dykes Specimens from four s61vsbergite dykes have been analysed, one analysis being given in Table 1. Generally they prove to have similar compositions to the syenite and quartz syenites of ZombaMalosa, but they are rather lower in Mg and much poorer in Ca, plotting in Fig. 2 with the Kasupe-area rocks. It is not clear whether this difference is due to the minor alteration which appears to have affected all the dykes or to separation of mafic phases during emplacement, or whether it is fundamental, indicating a quite separate magma batch.
Nepheline syenite and syenite
ChinduzL Mongolowe, Chaoneand Chikala A selection of analyses and norms of rocks from the E-W-trending nepheline syenite complexes
A. R. Woolley & G. C. Jones
342 f ./~,:
CO2
6
/
4
:: ..9
2
i
:
~ 9
L ',
-
.........~.~. .... ~,~
::~..." x
+~
o
,
/ 10 ~
...."'~'
Bn
,~..." ::..
8
:n 6 MgO
x "~, 9 o
. . . . . .
~
...... - ~
~ Sy
d~,~...P. ~mU'. . _
.,,.+ " ,,,D NS
x
;~':~-&~-"-----" 9 Syenite-granite
~
Oversaturated syenites- C h i k a l a Nepheline syenite - Chikala, C h a o n e , Mongolowe and Chinduzi o Nepheline syenite - J u n g u n i + Phonolite dykes 9 Metavolcanics
.... o
~
: ~ "~N
9
"
.m
~ .
i. Bn
,,
x CaO
......... - Zomba and Matosa
,~ S y e n i t e - Mpyupyu 4, Solvsbergite dykes
9
8
r
Chilwa Island Nepheline syenite e
x
Microfoyaite
t
Average ijolite
-
Nephelinite Camptonite
9 Average analyses (Le Maitre, 1976). N, nephelinite;
~
NS, nepheline syenite; 9
"-~
9. .
"
40
.,. %Sy
50
60 Si02(Wt
Bn, basanite; Sy, Syenite ~
Average sodalite
70
%)
Fio. 2. Plots of a selection of major oxides (wt. %) and barium (ppm) against SiO2. The relatively coherent trends of the Zomba-Malosa-Mpyupyu syenites and granites, the Junguni nepheline syenites and the phonolite dykes are indicated by broken lines. The Junguni trend is extrapolated to an average sodalite composition (Deer et al. 1963, Table 36). The areas occupied by the nephelinites, camptonites and nepheline syenites of Chilwa Island and of the metamorphosed basic volcanics are outlined by dotted lines. Several average compositions are plotted for comparison (Le Maitre 1976). For detailed discussion see text. is given in Table 2. A total of 23 from Chinduzi, 16 from Mongolowe, 14 from Chaone and 26 from Chikala have been analysed. Seven of the analysed rocks have normative quartz, of which five come from the early perthosite sheets of Chikala and one from a thin strip of syenite along the southern contact of Mongolowe. The seventh quartz-normative rock is from the southern contact of Chinduzi and is thought to be a fenite. The five SiO2-oversaturated Chikala rocks have similar compositions to the Zomba syenites as illustrated, for instance, by the plot of SiO2 against CaO (Fig. 2). The fact that Chikala is the
oldest of the E-W-trending nepheline syenite complexes, which are younger than ZombaMalosa, suggests that emplacement of the nepheline syenites was preceded by intrusion of a late pulse of Zomba-Malosa-type syenite. Thus, as illustrated by Fig. 4, the oversaturated perthosites are the only rocks which are common to both the nepheline syenite and syenite-granite complexes. There appears to be a general increase in the degree of undersaturation from Chikala westwards to Chinduzi (Fig. 3). The rocks mapped as pulaskites tend to be less SiO2-undersaturated than the foyaites. However, there is an overlap
Petrochemistry of the Chilwa alkaline province 1
,
1
i
,
,
i
n
i
,
,
,4
u
~l
i
i
f
,
343 ;
,
,
2800 2000 Ba (p.p.m.) 1200
.9
alA
..
9" i :...... . . :.~ ~ ~: ..... -.. ~i..
9
:
drA.
! 9
[ .
'.
. nn!
.......~:~ . . . . . .
400
... .
.. '.
-. ".
x
..
.:'
..
..
o
I I | "..~ l
x
o
. o
~...~ ,~:.~
v..
...
...
* I ~ 9 I 9149 9 9 . ~-,& 9 ~'e 9 ,~*.P'A ~
~ ...
.k ~
..:
~
9
.-~"
9 9 ~,A 9149 9
.r
o
Oli
9
,
o
I
t]
e~l~
~,
A
9
9
"~.~ ,,.
24 o
+
........
20
9
AI203
~NS
~
~e
+ +~s~.~;..
........~ . L
+
" g l ~ { ~ , ~ ". 9
16 : , ,... N, " " m ",,~ ....
~
.
,~-~.~*~
9
~ Bn
' .l""~"
~
-..,
.
x
9
9
, ..
"~
~
9
". ,..
o,,
22
'~..,
",o,
x
14 +,.+ ~, +9
Na20 10
~ .........................+ .....~ , I "
laD
:~:~ .. "'..1 . " ,... 9
J
~
4"O
NS"
,~.-..~,=~a~'~*'~.,
.
~
~'~7,--
-~
9
"~ ' ~:...,
9
: Z ' ~ & ' ~ , ;. o
"~
~ Bn
~
"
9
L
I
I
I
L
I
I
50
I
60
I
I
I
I
t
I
I
I
70
SiO2(Wt %)
and a frequency diagram of normative ne values does not show distinct populations. There is also a marked increase in the degree of peralkalinity from E to W. Only one analysed sample from Chikala has normative ac, but seven from Chaone, 10 from Mongolowe and 21 from Chinduzi contain ac, as illustrated by the peralkaline diagram (Fig. 4, inset). Plots of individual oxides and elements against SiO2 (Fig. 2) show rather vague trends. There is a fall in CaO, for instance, with increasing SiO2, but it is on quite a different trend from that of Zomba-Malosa; the same applies to MgO, NazO and A1203. There are minor variations for particular rock types and between complexes. For example, the more nepheline-rich foyaites and the rocks of Chinduzi with significant sodalite tend to have higher values of Na20 than the other complexes, while Chinduzi rocks are generally lower in TiO2, MgO and P205 and have
higher Fe203 : FeO ratios. The evidence for highlevel fractionation is thus sparse, although there was some variation between magma batches. If a decrease in MgO and C a 9 is taken to indicate the direction of evolution then an average leastevolved rock, at the present level of erosion, has about 58% SiO2, 1% MgO, 2% CaO and 8%Na20, and the main fractionation trend was towards rocks richer in SiO2 and with less modal nepheline. There are some rocks, however, particularly in Chinduzi, which are enriched in soda and are chemically similar to rocks from Junguni. These may have been involved in a different process which concentrated soda, as will be discussed later. There is some coherence in the Ba and Sr values (Fig. 5) and F and C1 values (Fig. 6). Chinduzi has high Ba:Sr ratios and a well-defined trend. Chaone and Mongolowe are somewhat variable with a higher Ba: Sr ratio for pulaskites,
344
A. R. Woolley & G. C. Jones Qz
Qz
Qz
Qz
Qz
Qz
A
Ne
.-.
9 ~o,:
Syenite-granite Syenite Nepheline syenite
: * §
Oversaturated syenite S01vsbergite Phonolite
,A- Nepheline syenite x Microfoyaite 9 Ijolite
Ks
FIG. 3. Rocks of the northern part of the Chilwa province, excluding basic rocks (i.e. metamorphosed basic volcanics, nephelinites and camptonites), plotted in terms of Q z - N e - K s (wt. %). For full key to symbols see Fig. 2. and Chikala seems to follow several trends, although significantly the five quartz-normative samples define a trend with the highest Ba:Sr ratio; this is almost identical with the ZombaMalosa trend, helping to confirm the correlation already suggested. In all the complexes F exceeds CI initially, but they evolve, particularly in Chinduzi, to rocks in which F is less than CI and with C1 values of over 10 000 ppm in two rocks, reflecting the presence of abundant sodalite. In comparison with average nepheline syenites (Le Maitre 1976) the nepheline syenites of Chikala, Chaone, Mongolowe and Chinduzi (Fig. 2) are generally 2 % - 5 ~ higher in SiO, and
somewhat lower in Na20, A1203 and CaO, while Chinduzi also has lower tenors of MgO, TiO2 and P20 5. These differences result in a much lower average normative ne value of 12.2 compared with 21.8 for Le Maitre's average nepheline syenite. This is balanced by higher normative albite in the Malawi rocks. Junguni
18 rock samples from Junguni have been analysed, and a selection is given in Table 3. All are strongly SiO2-undersaturated with normative ne ranging from 14 to 67, and as a whole they
SiO2 Chikala
Chinduzi
1I I !
I AI203.3SIO
/
}
~a i
i
....... xO ~ /~ ~,
,7/
" Zomba-Malosa
v/'
o Junguni
a20 § \K20).
i 9b r
9
1I 1
i [ I
Mongolowe chaone
§ Dykes 9 Metavolcanics
I ~o
FIo. 4. Rocks of the northern part of the Chilwa province plotted in terms of SiO2-AI203 (Na20 + K20) (wt. %). Comendites plot between the full and broken lines radiating from SiO~ and pantellerites to the right of the broken line according to Macdonald & Bailey (1973, Fig. 6).
Petrochemistry o f the Chilwa alkaline province :7omba-I~lalosa ~A~
9
cL 500
~~ #.--
_~_i__~, . . . . . z-- 9 9149 /
9
' 9 -*~*".
9 ,...........
_0 9. . . .
- ,../ ~
Chikala
~
.
., . . . . o-
-o
,#
.
.
345 .
500
.
Chaone - Mongolowe
9
9
~,A
Chinduzi
200 ~ v
T/~/"
9
/
2000
~1000 E d.
x
+
/
500
/
1500
.-->,
E r1 d.
i
v
lOOO -m-.
E .~' 500
o
.~of'~o~
~
dunguni
o
9
500
~
500
i
1000 1500 Ba(p.pm.)
,
J
2000
i
i
500
a
1500 Ba(p.p.m.)
L
Chilwa Island & dykes i
i
2500
FIG. 5. Plots of Ba against Sr (ppm). Clear linear trends are indicated by broken lines. Symbols as for Fig. 2. For discussion see text.
constitute the most undersaturated rocks in the northern part of the Chilwa Province as illustrated by the Qz-Ne-Ks diagram (Fig. 3). Two rocks have no feldspar at all in the norm and leucite occurs in three norms, although this does not reflect high values of K 2 0 - - i n fact these rocks are the poorest in potash of all the nepheline syenites--but demonstrates extreme undersaturation. Some rocks have exceptionally high values of Na20, two containing more than 20%, and the same rocks are highly enriched in C1, with up to 4.5%, reflecting their high modal sodalite. The trend of sodium enrichment clearly correlates with the presence of sodalite, as can be seen from the average sodalite composition (Deer et al. 1963, Table 36) plotted in Fig. 2. This also accounts for the unusual trends of the Junguni rocks in all the SiO2-based plots (Fig. 2). For instance, MgO, CaO, FeO and TiO2 decrease and AI203 and N a 2 0 increase with decreasing SiO2, which is the reverse of the normal situation and reflects a distinctive petrogenesis. All the Junguni rocks are ac normative and Fig. 4 illustrates both the range of SiO2 undersaturation and the peralkalinity of the rocks, which increases as they become more undersaturated because of the higher content of sodalite. The main cluster of Junguni rocks in Fig. 4 corresponds closely to the most undersaturated Chinduzi rocks, so that the increase in SiO2 undersaturation and peralkalinity from Chikala
through Chaone and Mongolowe to Chinduzi continues into Junguni. The average Ba:Sr ratio in the Junguni rocks is only marginally above unity (Fig. 5), which is lower than for the other nepheline syenite complexes. There is therefore some correlation of this ratio with the degree of saturation, ZombaMalosa having the highest and Junguni the lowest ratio. Chinduzi (Fig. 5) does not follow this trend. The amount of F in the Junguni rocks is generally similar to that in the other nepheline syenites but C1 is usually higher, sometimes exceptionally so (Table 3; Fig. 6). There is a sequence, correlating with the degree of saturation and peralkalinity, from Zomba-Malosa through the nepheline syenites: Zomba-Malosa Chikala-ChaoneMongolowe Chinduzi Junguni
F > C1 in nearly all rocks F>C1 in the majority of rocks C1 > F in the rationl4:9 C I > F in the ratio 14:4 but in some rocks in excess of 100:1
Chilwa Island Four nepheline syenites from the small plug on Chilwa Island together with two microfoyaites from dykes have been analysed (Table 3). The microfoyaites, and one in particular, prove to be
346
A. R. Woolley & G. C. Jones Phonolite oot ot
Junguni & Chinduzi
dykes 6000
+
+
d..
//
+
,,._..
/
/
/
12000
/
/
4-
10000
/
/
o
/
/
~' 4000 0..
/
/
+
2000 ~/ //4r +
+
8000 +
/ g"
//
+
++
~
++
+
E d_
Chikala Chaone
6000
6000 E
Mongolowe"
o
/
c c
o
/
,..--.
4000
,/
d. d. .._....
/
// ///
/
(.3
/i
9
..,..
4000
//
0 0
/ /
o
2000
/
9, ;
o~
2000 9
~"o
,,
/ / 9
,/..N:-4 E d..
9 yr
-
2000
....
ZombaMalosa ....
,
d.
---- 1000
9
A*,
e/./
o o
9
,,
o
/
Chilwa Island &
/
metavolcanics
9
2000
4000
F(p.p.m.)
9
2000
E d. 1ooo d
x
o x
2000
p~
0
4000
F(p.p.m.)
FIG. 6. Plots of F against CI. Symbols as for Fig. 2. For discussion see text. rather higher in MgO, CaO and total Fe than the nepheline syenites and may have involved some accumulation of mafic minerals, so that the two groups are sometimes separated on variation diagrams. The discussion is therefore concentrated on the nepheline syenites which, for clarity, are outlined in Fig. 2. The Chilwa Island nepheline syenites are quite distinct from those discussed above, being much poorer in SiO 2 (Fig. 2) with only the more sodalite-rich rocks of Junguni being m o r e S i O 2 depleted. Because of their much lower SiO 2 values they occupy a separate area of Fig. 2 and can also be distinguished by their different concentrations of other oxides and elements, e.g. CaO. Whereas the nepheline syenites of the main complexes are less undersaturated than the
average nepheline syenite of Le Maitre (1976), those from Chilwa Island are more undersaturated with an average normative n e value of 28. The different chemistry of the Chilwa Island nepheline syenites is considered to be of considerable petrogenetic significance, as is discussed later, and is emphasized by the Ba:Sr ratios and CO2 values. The nepheline syenites of all the complexes from Chikala to Junguni contain more Ba than Sr, and this is the normal relationship in such rocks. The Chilwa Island nepheline syenites, however, have more Sr than Ba, usually very much more (Fig. 5), which is a characteristic of carbonatite complexes. Similarly, CO2 is negligible in all the nepheline syenites, syenites and granites so far considered (Fig. 2), but is relatively abundant in all the silicate rocks that have been
Petrochemistry o f the Chilwa alkaline province
347
TABLE 2. Analysesof nepheline syenites, syenitesand quartz syenitesfrom Chinduzi, Mongolowe, Chaone
and Chikala 15
16
17
18
19
20
21
22
23
24
25
26
27
28
Major elements (wt. %) 58.7 SiO 2 0.23 TiO2 20.5 A1203 3.12 Fe203 FeO 1.55 MnO 0.12 0.07 MgO CaO 0.23 9.0 Na20 5.36 K20 0.42 H20 + 0.01 HzOPzOs 0.11 CO2 Others 0.43
60.0 0.76 17.2 2.17 2.84 0.16 0.66 1.88 7.4 5.54 0.44 0.08 0.18 0~12 0.54
60.1 0.31 17.6 2.81 3.15 0.22 0.25 1.27 8.1 5.21 0.38 0.04 0.06 -0.40
56.3 0.20 20.9 1.88 3.02 0.16 0.13 0.92 9.9 5.73 0.43 0.05 0.02 -0.93
59.4 0.25 20.3 2.14 1.91 0.12 0.10 0.66 9.2 5.60 0.26 --0.11 0.47
57.8 0.72 20.1 2.23 1.63 0.20 0.51 1.14 9.3 5.55 0.32 0.09 0.09 0.13 0.31
60.1 1.05 17.6 1.32 2.26 0.16 1.16 1.87 6.2 6.68 0.53 0.06 0.36 -0.57
60.3 0.95 17.9 1.64 1.77 0.24 1.00 1.36 7.0 5.48 0.78 0.14 0.17 0.29 0.53
57.2 0.73 19.7 1.59 1.97 0.20 0.84 1.60 8.8 5.70 0.47 0.14 0.25 0.15 0.55
59.2 0.74 19.4 1.46 2.17 0.13 0.82 1.93 7.2 5.64 0.46 0.06 0.21 0.14 0.26
58.4 0.80 19.0 1.53 2.73 0.17 1.05 2.39 6.8 5.35 0.41 0.02 0.28 0.15 0.42
57.3 0.78 20.1 1.15 2.89 0.17 0.96 2.05 7.8 5.66 0.43 0.06 0.28 0.23 0.50
59.3 1.08 17.3 2.36 2.65 0.19 1.16 2.81 6.0 4.60 0.30 0.11 0.36 0.48 1.22
64.0 0.63 16.5 1.22 3.54 0.19 0.60 1.83 5.2 5.40 0.18 0.10 0.17 0.14 0.28
Total
99.97
99.90 100.57 100.52 100.12
99.92
99.55
99.89
99.82
99.50 100.36
99.92
99.98
1650 270 1 10 15 60
1570 110 9
1690 1640 8
710 260 1
1480 170 5 . 11 90
1340 600 5
13 30
1570 2440 2 10 5 105
- -
99.85
Trace elements (ppm) 410 F C1 3790 Be 5 Cr Li 13 Nb 205 Ni Cu . Zn 105 V Zr 100 Y 60 Sr 25 Ba 95 Rb 230 CIP W norms Q C or
ab an ?le ac
di
hy ol mt hm it ap CC
H2O+ H20-
. 31.7 44.7 0.2 16.5 - -
0.6 . - -
4.5 0.4 - -
- -
0.4 - -
1800 1360 2 . 21 85 .
1580 690 1860 10 780 2 1 . . . 21 7 100 95 .
990 4190 4
- -
37 220
. 101
.
98
71
75
85
10 88
. 145
310 45 160 2070 83
465 45 50 490 131
215 40 20 50 121
165 45 15 75 145
155 45 60 190 87
235 45 280 3100 104
1155 60 160 1480 100
.
.
.
.
32.7 42.7 . 8.1 4.4 6.3 . 1.6 1.0 1.4 0.4 0 . 3
0.4 0 . 1
.
- -
20 275
. 30.8 42.7
.
33.9 24.4 . 27.7 5.4 3.9
. 10.2 6.2 5.2
.
.
.
.
2.3 1.0
2.9 --
0.6 0.1
0.4 0.1
- -
0.4 - -
- -
0.4 0 . 1
. 33.1 39.1 . 18.5 3.0 2.9 . 0.9 1.6 0.5 - -
- -
0.3 - -
32.8 32.6 21.6 5.5 3.6 . 1.0 0.5 1.4 0.2 0 . 3
0.3 0.1
.
39.5 41.4 0.5 6.0
7
.
. 32.4 51.2 1.2 4.4
93
.
. 49
820 55 340 870 138
230 35 380 820 72
.
1.4 1.9
1.6 2.4
33.7 31.1 -20.9 4.3 4.5 . 1.9 0.2
2.0 0.9
1.8 0.4 0.7 0.8 0.1
1.4 0.6 0.3 0.5 0.1
- -
- -
5.2 .
2.1 .
- -
0.5 0.l
.
3 100
1250 360 4 . 13 110 .
1450 990 5 . . 19 135 .
18 90
58
58
62
16 88
485 40 450 1400 132
670 40 340 t250 165
690 60 450 8300 71
200 70 150 900 133
1.1
7.2
27.2 50.8 6.7 --
31.9 44.0 5.7 --
1.5 3.6
1.2 5.7
. 33.3 43.3 4.0 9.6 . 2.8 . 1.8 2.1
31.6 43.2 5.5 7.8 . 3.0
33.5 35.9 3.1 16.3
2.9 2.2
3.1 1.7
3.4
1.8
1.4 0.5 0.3 0.5 0.1
1.5 0.7 0.3 0.4
1.5 0.7 0.5 0.4 0. l
2.1 0.9 1.1 0.3 0.1
1.2 0.4 0.3 0.2 0.1
.
. 3.2
- -
- -
15, Biotite-amphibole-aegirine-nepheline syenite, Chinduzi (1980, P20, 6); 16 aegirine-augite-amphibole-nepheline syenite, Chinduzi (1980, P20, 17); 17, aegirine-biotite amphibole-sodalite syenite, Chinduzi (1980, P20, 21); 18, amphibole biotitenepheline-sodalite syenite, Chinduzi (1980, P20, 38); 19, amphibole-aegirine-biotite-nepheline syenite, Mongolowe (1980,P21, 2); 20, biotite-aegirine-nepheline syenite, Mongolowe (1980, P21, 32); 21, biotite-aegirine-nepheline syenite, Mongolowe (1980, P21, 32); 21, biotite-aegirine-nepheline syenite, Mongolowe (1980, P21, 33); 22, pyroxene biotite-amphibole syenite, Chaone (1980, P21, 27); 23, amphibole-biotite-nepheline syenite, Chaone (1980, P21, 13); 24, amphibole-biotite-nepheline syenite, Chaone (1980, P21, 26); 25, p y r o x e n e - a m p h i b o l e - b i o t i t e - n e p h e l i n e syenite, Chikala (1980, P22, 5); 26, a m p h i b o l e pyroxene-nepheline syenite, Chikala (1980, P22, 16); 27, pyroxene-amphibole-biotite syenite, Chikala (1980, P22, 7) 28, amphibole-quartz syenite, Chikala (1980, P22, 3). - - , as for Table 1. Analysts, G. C. Jones and V. K. Din.
348
A. R. Woolley & G. C. Jones
T A B L E 3. A n a l y s e s o f nepheline a n d sodalite syenites f r o m J u n g u n i a n d nepheline syenites, microfoyaite,
average ijolite, olivine nephelinite, camptonite a n d aln6ite f r o m Chilwa I s l a n d 29
30
31
32
33
34
35
36
37
38
39
40
41
42
13.6 1.98 3.1 4.06 8.20 0.27 6.97 28.88 0.5 1.91 0.39 0.13 5.42 24.28 1.04
Major elements (wt. %) SiO 2 TiO 2 A1203 Fe203 FeO MnO MgO CaO Na20 K20 H20 + H20 P_,O5 CO2 Others
56.4 0.44 21.9 1.4 1.33 0.13 0.22 0.79 9.5 6.19 0.37 0.04 0.06 0.28 0.53
53.7 0.37 21.0 2.4 0.55 0.14 0.19 1.21 11.6 4.71 1.56 0.15 0.04 1.62 0.59
56.7 0.50 21.2 2.7 0.67 0.26 0.33 1.46 9.3 5.24 0.87 -0.07 0.01 0.47
56.1 0.58 20.7 2.50 0.87 0.25 0.43 1.66 10.3 4.96 0.41 0.08 0.11 0.20 0.53
48.8 0.28 24.3 1.3 0.79 0.10 0.01 0.33 17.2 2.71 0.28 0.09 0.06 0.34 2.82
43.1 0.28 26.2 1.9 0.81 0.15 -0.27 20.5 1.95 0.41 0.09 0.03 0.19 3.67
46.0 1.35 20.4 3.36 2.57 0.40 1.47 5.86 7.9 5.75 1.50 0.15 0.62 1.14 0.82
49.8 0.39 19.0 3.79 1.55 0.29 0.12 3.03 8.7 5.78 3.78 0.27 0.04 2.28 0.67
48.0 2.20 12.3 5.32 4.03 0.46 2.52 10.48 5.9 3.55 1.62 0.10 1.27 1.23 0.73
42.3 0.65 14.7 4.90 1.19 0.33 0.83 20.82 5.5 3.95 0.71 0.10 0.82 2.99 0.45
39.0 4.75 10.9 7.11 7.14 0.24 10.14 11.96 1.2 2.79 2.06 0.42 0.95 0.28 0.83
36.6 4.28 10.9 6.69 7.57 0.28 5.20 12.34 4.2 1.40 3.14 0.29 2.57 3.80 0.60
41.5 3.12 12.4 4.23 8.56 0.17 9.35 10.50 2.6 2.04 1.76 0.23 0.84 1.93 0.43
Total
99.58
99.83
99.78
99.68
99.4l
99.55
99.29
99.49
99.71
100.24
99.77
99.86
99.66 100.73
160 840 34730 44060 6 11 . 22 23 136 174 . . . . . . lll 54 77 . . . 982 488 1115 41 23 19 666 161 88 750 100 100 168 100 112
1470 510 4 6 13 187 . . 120 . 250 92 2988 2130 142
310 770 5 -15 270 .
1500 160 6 7 8 367
910 965 3 6 13 75
16 165
12 64 380 420 93 710 610 69
1800 70 2 274 8 40 54 44 125 380 22 84 2983 1150 57
2150 480 3 19 24 68 -17 154 160 135 84 1593 1260 48
1350 10 -247 9 36 65 21 101 200 89 87 1260 690 40
8.3 25.0 6.8
12.1 15.5 16.1 . 3.5
Trace elements (ppm) F C1 Be Cr Li Nb Ni Cu Zn V Zr Y Sr Ba Rb
CIP W norms C or ab an lc ne ac ns di wo ol mt hm il ap cc magn chaly H2 O+ H20-
440 5050 5 . 17 72 . . 52 . 404 17 226 100 119
680 1300 2870 570 6 8 . . . 4 17 215 198 . . . . . . 6 115 . . . 956 1179 41 48 710 558 600 600 102 200
. 35.4 26.7 . . 26.8 3.7
. 27.8 24.8 . . . . 31.7 6.9
- -
1.5 -1.2 0.2 -
-
0.8 0.1 0.6 - -
- -
0.4 - -
. 31.0 33.5
1 . 6
-----
-
-0.1 2.1 0.4 1 . 1
1.6 0 . 2
.
. . 23.5 1.7
1390 1470 11 . 26 165
. 29.3 27.0 . . 28.1 7.2
- -
1.8 1.9 -1.6 1.1 1.0 0.2 -. - -
0.9 -
-
-
-
4.4 0.4 --. 1.1 0.3 0.5 . -
-
0.4 0.1
. 16.0 15.4 .
. 5.5 --
4.8 50.9 66.9 3.8 5.5 7.4 10.2 ----0.6 1.0 --. . . 0.5 0.5 0.1 0.1 0.5 0.4 . . . 0.4 . 0.3 0.4 0.1 0.1
. 34.0 4.1 3.2 -34.0 -. 8.2 1.7 -4.9 . 2.9 1.5 2.6 . . 1.5 0.2
.
124 . 1191 96 2456 480 188
1676 87 1920 730 102
. 34.2 18.3 --25.6 7.1 . 0.3 -1.4 1.9 . 0.7 0.1 5.2 . .
. 21.0 20.0 --12.8 5.7 .
. 17.0 6.1 -4.9 4.2 3.0 2.8 .
. 3.8 0.3
. 10.8 -3.8 9.8 25.2
. 1.6 0.1
. 4.5 29.0 -3.0 2.8 1.2 1.9 6.8 . . 0.7 0.1
. 16.5 0.1 16.1 . 5.4 -. 27.4 -8.8 10.0 0.2 9.0 2.2 0.6 . . 2.0 0.4
.
.
. 5.7 -. 10.8 -6.9 9.7 -
0.1 11.3 4.4 ----
.
-
14.6
--
17.0 6.1
4.4 5.9
- -
8.1 6.1 8.6
-
-
5.9 2.0 4.4
3.8 12.8 38.8 13.8
1.7 0.2
0.2 0.1
. .
2720 290 6 72 40 212 26 33 228 260 125 87 4639 1400 45
. 3.0 2.9
29, Biotite-aegirine-nepheline-cancrinite syenite, Junguni (1980, P25, 2); 30, aegirine-nepheline-cancrinite syenite, Junguni (1980, P25, 8); 31, aegirine-nepheline syenite, Junguni (1980, P25, 17); 32, aegirine-amphibole-nepheline syenite, Junguni (1980, P25, 21); 33, sodalite-biotite-nepheline syenite, Junguni (1980, P25, 15); 34, sodalite-aegirine syenite, Junguni (1980, P25, 18); 35, nepheline syenite, Chilwa Island (1957, 1056, 133); 36, nepheline syenite, Chilwa Island (1957, 1056, 136); 37, microfoyaite, Chilwa Island (1957, 1056, 134); 38, melanite-wollastonite ijolite (average of four analyses), Chilwa Island; 39, olivine nephelinite, Chilwa Island (1957, 1056, 161); 40, olivine nephelinite, Chilwa Island (1957, 1056, 157); 41, camptonite, Chilwa Island (1957, 1056, 182); 42, aln6ite, Chilwa Island (1968, P37, 168). - - , as for Table 1. Analysts, G. C. Jones and V. K. Din.
Petrochemistry of the Chilwa alkaline province analysed from Chilwa Island. It must be stressed that the CO2 is not the result of carbonatite veining or replacement; most of the silicate rocks post-date the carbonatite and textural evidence suggests that the carbonate is a primary crystallizing phase.
Phonolite dykes Samples from 20 phonolite dykes distributed widely over the northern part of the Chilwa province have been analysed and a selection is given in Table 4. Only three specimens were collected by the present authors, 17 being originally collected by K. Bloomfield. Most of the phonolites are strongly SiO2 undersaturated, but are comparable with the main group of Junguni nepheline syenites and those of Chilwa Island (Fig. 3). They are also strongly peralkaline (Fig. 4), which correlates with their Na20 values and Fe 3 + :Fe 2+ ratios, which are higher than for the main nepheline syenites. The dykes have SiO2 values spanning the gap between the main cluster of nepheline syenites and those of Chilwa Island (Fig. 2), and for most elements they define coherent and 'normal' trends, e.g. for CaO, MgO and Ba. Although the dykes, like the main syenites, have exceptionally low concentrations of Cu, Ni and V, they are distinctly enriched in Nb, Be, Rb and Zr, elements that tend also to be enriched in the nepheline syenites of Chilwa Island. The most evolved dykes, those with the lowest MgO, CaO etc., have similar compositions to some of the main group of Junguni rocks and certain of the foyaites of the main complexes, while the least evolved dyke rocks overlap with the nepheline syenites of Chilwa Island for some elements although not for others. However, the CO2 values and Ba:Sr ratios clearly point to a correlation of the phonolites with the Chilwa Island nepheline syenites. They contain a moderate amount of CO2 (Fig. 2) which is otherwise a unique feature of the Chilwa Island rocks, and Sr is greater than Ba for many of the dykes which, as already discussed, is also characteristic of Chilwa Island rocks. Therefore, although the phonolites are widely distributed over the whole province and some of the analysed samples come from more than 50 km west of Chilwa Island, their compositions strongly suggest that they are consanguineous with the Chilwa Island centre rather than with the large nepheline syenite complexes.
Nephelinite, camptonite, ijolite and alnOite Four olivine nephelinites, two camptonites, four ijolites and an aln6ite from Chilwa Island have
349
been analysed and some results are given in Table 3. All four analysed specimes of ijolite come from the same intrusion to the N W of the carbonatite plug. One specimen was relatively homogeneous but the others are extremely heterogeneous and in part pegmatitic, so that very large samples were crushed and quartered. However, the analyses are still rather variable so an average of the four is given in Table 3, and this value is plotted on the variation diagrams although all four are shown in Fig. 3. The very high CaO in the ijolites, one of which contains 27% CaO, is not in carbonate but in wollastonite, melanite and pyroxene. The first two minerals appear to have crystallized very late, and some metasomatism may have been involved. The suite of rocks as a whole is chemically distinct from all the other rocks of the province, with the sole exception of the metavolcanics. The clearest distinction is the lower silica, where the nephelinites all have less than 40% SiOz and the camptonites only slightly more. These rocks are also lower in alkalis and A1203, Zr, Rb and Nb but contain more CaO, MgO, FeO, TiO2, V, Ni and Cu; this is clearly shown in Fig. 2. Sr exceeds Ba for all Chilwa Island rocks (Fig. 5). The camptonites have compositions similar to the average nephelinite of Le Maitre (1976) (Fig. 2), while the olivine nephelinites are similar to the melanephelinites of Homa Bay, Kenya (Le Bas 1977, Appendix 2). Compared with Homa Bay, the Malawi rocks are somewhat poorer in Na20; in three samples K20 is greater than Na20, and two samples contain substantial CO2. Only three of the nephelinites contain normative ne and this in small amounts (5.4-5.7). Normative alkali feldspar is abundant, but this value is probably inflated by the high potassium held in modal biotite and possibly in groundmass zeolite. A thorough understanding of the chemistry of the nephelinites and camptonites must await an electron microprobe study of their mineralogy, but they are the most basic silicate rocks of the province and, together with the metavolcanics, clearly represent potential parental magmas. In this regard it is noteworthy that feasible fractionation trends for the nephelinites, perhaps by separation of olivine and later pyroxene, extrapolate towards the Chilwa Island nepheline syenites rather than the nepheline syenites of the large complexes; this is illustrated, for instance, by CaO and N a 2 0 in Fig. 2. The analysis of aln6ite (Table 3) confirms the trend towards carbonatite, the norm indicating more than 50% carbonate and over 12% apatite. The very high concentrations of Sr and Ba and moderately high Nb are consistent with this tendency.
350
A. R. Woolley & G. C. Jones
T A B L E 4. A n a l y s e s o f p h o n o l i t e d y k e s a n d m e t a m o r p h o s e d 43
nephelinites from Chaone
44
45
46
47
Majorelements(wt.%) SiO 2 57.8 TiO: 0.13 AlzO3 17.6 Fe203 4.86 FeO 1.68 MnO 0.36 MgO 0.11 CaO 1.06 Na20 9.3 K20 4.89 H20 + 1.03 H20-P.,Os 0.02 CO, 0.25 Others 0.85
55.6 -20.0 4.61 0.13 0.27 0.06 0.56 11.6 4.37 0.70 0.01 0.05 1.04 0.51
51.4 0.37 21.8 0.77 3.00 0.26 0.34 1.61 7.6 8.57 2.07 0.04 0.04 1.38 0.89
59.4 0.14 17.9 4.11 2.20 0.29 0.10 1.15 8.1 4.93 1.09 0.04 0.04 0.14 0.56
52.4 1.34 17.5 5.47 1.60 0.50 1.13 1.31 10.2 5.42 0.86 0.08 0.44 0.98 1.38
40.2 3.62 14.3 5.49 9.38 0.27 6.63 10.72 4.7 1.45 1.09 0.02 1.07 0.35 0.51
41.5 2.60 13.7 4.21 7.71 0.19 7.76 14.09 4.8 1.72 0.67 0.02 0.51 0.33 0.53
42.4 3.31 14.8 4.67 9.05 0.20 5.86 9.95 4.9 2.49 1.01 0.04 0.66 0.32 0.66
Total
99.51
100.14
100.19
100.61
99.80
100.34
100.32
2060 1360 9 . 19 100 . . 130 . 600 20 2242 2230 270
2890 230 10
3770 4830 19
41 300
71 600
2660 40 1 20 21 85
4040 130 7 385 4 45 85 45 109 250 245 30 910 130 40
99.94
Trace e&ments ( p p m ) F 4180 C1 262." Be 13 Cr . Li 50 Nb 325 Ni . Cu . Zn 170 V . Zr 1190 Y 165 Sr 8 Ba 80 Rb 260 CIPWnorms C or ab an lc ne nc ac di wo ol cs mt hm il ap cc H2 O+
H20-
-28.9 36.6 -. 14.5 0.6 10.6 4.5 -0.3 . 1.8 -0.3 0.1 . 1.0 --
940 390 20 .
. 13 200 .
.
.
. 160
.
. 2400 30 16 50 150 -25.8 36.7 --
.
. 22.7 2.5 6.6 0.3 0.9 --
.
. 1.3 1.4 -0.1
.
. 0.7 --
0.4 50.7 6.4 7.7 . 22.5 3.3 ---4.3 . 1.1 . 0.7 0.1 . 2.1 --
48
.
.
49
.
. 200
200
1700 60 65 -240
2800 75 1065 670 250
30 149 350 310 35 1160 830 13
.
. 29.1 49.3 --
.
.
. 8.6 7.1 13.7
32.0 19.8 --
.
21.7 2.4 12.8 2.9 . 1.6
--10.8 8.0 22.0 --37.3
.
17.6 --24.4 .
.
4.9 0.3 0.1
1.5 . 2.5 1.0
1.1 --
0.8 0.1
.
.
8.0 .
71 136 300 385 35 940 1490 55
14.7 4.5 10.9 20.1 --26.1
.
8.6
.
.
3830 300 3 25 2 75
.
. 8.3 0.3 2.1 3.5 0.6 --
50
5.9 2.3 6.1
7.0 6.8
4.9 1.2 0.8 0.7 --
6.3 1.6 0.7 1.0 --
. 6.9 2.5 0.8 1.0 --
43, F e l d s p a r p h y r i c p y r o x e n e - b i o t i t e - a m p h i b o l e p h o n o l i t e , N E side o f C h i n d u z i (1980, P20, 1); 44, a e g i r i n e - c a n c r i n i t e p h o n o l i t e , S E o f Z o m b a (1980, P29, 1); 45, c a n c r i n i t e - b i o t i t e p h o n o l i t e , 4 k m N o f C h i n d u z i (1980, P I 9 , 24); 46, a e g i r i n e - a m p h i b o l e - b i o t i t e p h o n o l i t e , S W o f C h i n d u z i (1980, P29, 16); 47, n a t r o l i t e p h o n o l i t e , E o f M a l o s a (1980, P29, 27); 48, a m p h i b o l e - p y r o x e n e - a l k a l i f e l d s p a r - n e p h e l i n e rock, N W side o f C h a o n e (1980, P21, 40); 49, p y r o x e n e a m p h i b o l e - a l k a l i f e l d s p a r - n e p h e l i n e rock, N W side o f C h a o n e (1980, P21, 42); 50, p y r o x e n e - a m p h i b o l e - b i o t i t e - a l k a l i f e l d s p a r rock, N W side o f C h a o n e (1980, P21, 44). - - , as for T a b l e 1. Analysts, G . C. Jones a n d V. K . D i n .
Petrochemistry of the Chilwa alkal&e province Metavolcanic rocks
Eight specimens of metavolcanic rocks from the Chaone locality have been analysed (Table 4), care being taken to select samples with a minimum of veining by nepheline syenite; only two samples have very small veinlets. The resulting analyses lie very close to the average nephelinite of Le Maitre (1976) (Fig. 2) with normative n e plus leucite values of 14-30 and feldspar from 11 to 31. An analysed sample of metamorphosed basic rock from Chikala (Stillman & Cox 1960) was considered by them to be close to an average nepheline tephrite but their sample, with 6.5% normative ne, was much more felsic (normative feldspar 59%) than those analysed here. It is not certain how closely the present analyses reflect the original rock compositions, but on most variation diagrams they cluster relatively tightly (Figs. 2 and 6). On the SiOz plots (Fig. 2) they trend towards the Chilwa Island nepheline syenites, but such trends could equally well represent primary differences. However, if the metavolcanics had been substantially altered then their original compositions would have been even more basic than nephelinite, which seems unlikely. The metavolcanics are close in composition to the olivine nephelinites and camptonites of Chilwa Island, differing only a little in CO2, K20, F and Zr. The chemical evidence indicates strongly that the metavolcanics were originally nephelinite lavas, and thus also represent potential parental magma in any petrogenetic model for the province.
Discussion The Chilwa alkaline igneous province is remarkable for the extreme range of rocks it encompasses, from carbonatite to granite with many intermediate rock types. Individually, some of these rock types pose petrogenetic problems that have engaged petrologists for decades, but when such relatively exotic rocks are concentrated in one province understanding their origin becomes even more intractable. The wide range of rock types of the Tundulu carbonatite complex was considered by Garson (1962) to have evolved from kimberlite or mica-peridotite primary magmas by combinations of crystal fractionation, fenitization and rheomorphism, including the production of trachytes and phonolites by the mobilization of feldspathic and nephelinized fenites. It has also been suggested that the large intrusions of nepheline syenite represent mobilized nepheline fenites (Garson & Walshaw 1969), and these workers speculate that the syenite-
351
granite intrusions were generated by the rheomorphism of fenites formed around deep-seated carbonatite and ijolite bodies. Woolley & Garson (1970) considered that fractionation of primary alkali olivine basalt magma generated at intermediate depths, together with nephelinite-carbonatite magmas at greater depths within the mantle, could account for the province as a whole. Again, rheomorphism of fenites and nepheline fenites was suggested to have produced certain rock types. A similar scheme was erected by Woolley (1969) but called on primary nephelinitic, phonolitic and carbonatitic magmas. The syenites, quartz syenites and granites of Zomba-Malosa and Mpyupyu define coherent trends on most geochemical plots, which strongly suggest that crystal-liquid fractionation controlled a descent from slightly SiO2 oversaturated syenite to granite, with a concomitant increase in peralkalinity. The granite plots define simpler trends than the syenites but the scatter in the syenites seems to be mainly attributable to rocks at the northern end of Malosa which have undergone some metasomatic alteration associated with the emplacement of the Mongolowe nepheline syenite. Fractionation of alkali feldspar, fayalite, pyroxene, amphibole and an opaque phase is consistent with the chemistry of the series. I n contrast, the nepheline syenites and syenites of Chikala, Chaone, Mongolowe and Chinduzi are chemically similar and show only moderate alkali enrichment, particularly in Chinduzi, and a rather muted trend towards more SiO2-rich syenites. Although the nepheline syenites could be produced by fractionation of alkali feldspar from a slightly undersaturated syenitic magma driving the liquid towards the phonolite residuum, the great preponderance of nepheline syenites and the correlation of higher CaO, MgO etc. with greater SiO2 undersaturation suggests that the nepheline syenites are primary and that the few syenites are differentiates. There are slight differences between each complex, and an overall increase in SiO2 undersaturation and peralkalinity from E to W, but the indications are that these highly evolved rocks were emplaced as such and that little further evolution took place at their present level. Junguni is an exception to this and certainly underwent SiO2 impoverishment and Na and C1 enrichment leading to sodalite-rich rocks of unusual composition. The textural evidence of replacement indicates that this was a high-level effect which probably took place at the present level of intrusion, and probably involved Na concentration at the top of a magma column, perhaps by C1 streaming with the Na transported as gaseous C1 complexes.
352
A. R. Woolley & G. C. Jones
Experimental work on the system A b - N e - N aC1H20 has shown that NaCl is strongly partitioned into the vapour phase (Binsted 1981). Syenites are present in both the nepheline syenite and granite complexes, but they are chemically distinct. The only syenites that occur out of context, from the point of view of field relationships, are the quartz syenites of Chikala whose compositions match those of the syenites associated with the granites. Although there is an overlap on some chemical diagrams between the syenites of Zomba-Malosa and the syenites and nepheline syenites of the nepheline syenite complexes, there are other chemical indications that they are distinct, and this distinction forms a barrier which it is not easy to circumvent. On the SiOz-CaO plot (Fig. 2), for instance, not only is there a gap between the two rock populations but the Zomba-Malosa syenites are actually higher in CaO as well as SiO2, which poses considerable difficulties in deriving one group of rocks from the other by the crystal fractionation of an acceptable mineral phase or phases. The chemical differences between these two groups of rocks are therefore, such, as to require two parental magmas. It has been shown that the nepheline syenites of Chilwa Island, as well as being in a rather different geological context, have very different compositions from the nepheline syenites of the large complexes. They do not lie on common trends with the latter but show affinities with the spatially associated nephelinites and camptonites. Again, therefore, separation in the field correlates with chemical differences which appear to be fundamental and necessitate the postulation of a third 'primary' magma. The correlation of the phonolite dykes with the Chilwa Island centre rather than with the large nepheline syenite complexes was unexpected, but suggests that the centre played a somewhat larger role than suggested by its present small area. Although they are volumetrically insignificant, the nephelinites and camptonites of Chilwa Island must inevitably have an important place in any petrogenetic scheme because of their potential parental role. This is supported by the complete absence of basalts from the province. Karoo dolerites are widespread and the Karoo lavas which are now confined to the extreme south of Malawi probably formerly extended further north, but these rocks are 20-30 Ma older than the earliest Chilwa rocks (Woolley & Garson 1970, Table 1) and are tholeiitic (Woolley et al. 1979; Macdonald et al. 1983). The remnants of metamorphosed lavas found on Chikala, Chaone and Mongolowe must also occupy an important place in any scheme of petrogenesis, especially
now that those of Chaone have been shown to be nephelinitic and rather similar to the nephelinites of Chilwa Island. It seems very probable that the nephelinitic lavas were extruded from the Chilwa Island centre which, together with the correlation of the phonolite dykes with Chilwa Island, suggests that this was originally a volcano of some considerable magnitude. The chemistry and field associations of the rocks of the northern part of the Chilwa province indicate that three suites of rocks are present: (1) syenite-quartz syenite-granite, (2) nepheline syenite-syenite and (3) nephelinite-carbonatitenepheline syenite. If undoubted high-level fractionation is taken into account, the presence of three parental magmas is indicated, with compositions which are essentially trachytic (slightly oversaturated), phonolitic and nephelinitic-carbonatitic. The rocks representing these three magma series are shown in Fig. 7. Only number three of the postulated magmas is basic, and there is no indication within the province of the former presence of other less evolved basic rocks from which the nepheline syenites, syenites and granites could reasonably be expected to have derived. There are no basic plugs, or basic areas within the main complexes and, most significantly, there are no basic dykes apart from the Karoo dykes which, as already explained, cannot play a direct role in petrogenesis. If alkali basalt or some more alkaline basic rock had been available some evidence of it would be expected. There may, of course, be large basic bodies at depth beneath the province, but if such bodies were present they would need to be of enormous size to generate, by crystal fractionation, the large nepheline syenite and granite-syenite complexes of the northern part of the province, and the problem would be compounded in the south by the huge peralkaline granite-syenite pluton of Mulanje which covers nearly 640 km 2. Unfortunately, no geophysical work has yet been undertaken which might indicate the likely presence or absence of such bodies, but the lack of any surface manifestation of alkali basaltic rocks is thought to indicate that they are very unlikely to be present. The southern part of the Chilwa province includes the same range of rock types as the northern, and will be the subject of a future petrochemical study. There are a few published analyses of rocks of the southern area, notably of the syenites and granites of Mulanje Mountain (Garson & Walshaw 1969), which are variably peralkaline rocks similar to those of ZombaMalosa. Of more significance, perhaps, are the two nephelinites, foyaite and phonolite from the Tundulu carbonatite complex (Garson 1962),
Petrochemistry of the Chilwa alkaline province plotted in Fig. 7, which are clearly similar to comparable rocks from Chilwa Island (Fig. 2), as is also a foyaite from the small Nkalonje carbonatite centre. A nepheline syenite from the large Mauze intrusion, in contrast, is richer in SiO 2 and more comparable with the nepheline syenites from the large undersaturated plutons of the northern part of the province. Although the data are few it seems that the important petrogenetic distinction between a nephelinite (carbonatite)-nepheline syenite series and a syenitenepheline syenite series can also be made in the southern part of the province, while the oversaturated series represented by Zomba-Malosa is to be found in Mulanje. Since the 1960s the concept of mantle metasomatism has become more generally accepted and, combined with the notion of lithosphere focussing as developed in particular by Bailey (1964, 1980), is particularly well suited to explaining the genesis of the Chilwa rocks, bearing in mind the exceptionally broad range of rock types involved, their overwhelming felsic character and the paucity of basic rocks. Further, the Chilwa province is situated on a large dome, as are other alkaline provinces in Africa, and there is a close
353
temporal association of doming, rifting and igneous activity. It is considered that the dome lies above an area of low-density mantle produced by metasomatism over a long period of time, and probably initiated by the Karoo igneous event at least 50 Ma, and probably much longer, before the earliest manifestation of alkaline igneous activity in the Chilwa province. There is evidence of faulting in the area long before Chilwa times (Dixey et al. 1937; Dixey 1956) and this may be related to the deep structures that initially helped to localize the metasomatism. The wedge of anomalous mantle produced in this way is thought to have been enriched in phlogopite, carbonate and probably apatite, passing upwards into more amphibole-rich rocks. The metasomatism progressed upwards with time until even the lower crust was affected and rocks probably barely distinguishable from fenites were generated. The wedge ofmetasomatized mantle is thus envisaged as being zoned mineralogically and extending vertically at least from the base of the lithosphere to the lower crust. The dome eventually fractured (rifted) with consequent release of volatiles and upward
+
14
+
§
§
le §
~.1.§
12
§
+
10
CaO
+§ §
8
6
§
e2 4 +
§
§
I
e3
~1%
o
2 o i
4()
o§~ § ~|5
o o,
~
ee
o~;.o.-~,~j.
"-' "
§ § ~%'og~+o o . o <~
jo
z.o+~
50
,+
60
",
9 .
:~
9
9
70
S i O 2 (Wt %)
FIG. 7. Plot of SiO2 against CaO (same data as in Fig. 2) but showing the three postulated magma series: 9 syenite-quartz syenite-granite; O, nepheline syenite-syenite; + , nephelinite-nepheline syenite. Also shown are plots of six published analyses from the southern part of the Chilwa province which appear to indicate that the distinction between nepheline syenites associated with carbonatite centres, and probably fractionated from nephelinite, and nepheline syenites of large complexes also applies in this part of the province. Numbered analyses are as follows: (1) nephelinites, (3) foyaite and (5) phonolite--all from the Tundulu carbonatite centre (Garson 1962) (2) foyaite from the Nkalonje carbonatite centre and (4) foyaite from Mauze Hill (Dixey et al.
1937).
354
A. R. Woolley& G. C. Jones
migration of geotherms initiating the igneous cycle. It is envisaged that partial melting took place t h r o u g h o u t the anomalous m a n t l e wedge and that the m a g m a batches that r e a c h e d the surface originated at a range of depths with nephelinite and carbonatite deriving from the greatest depths (towards the base of or below the lithosphere), phonolite at i n t e r m e d i a t e depths and t r a c h y t e - q u a r t z trachyte at the shallowest depths in the vicinity of the base of the crust. It is i m p o r t a n t to stress that the phonolitic and trachytic liquids are considered not to have been derived by fractionation from more basic liquids but to be partial melts. Some fractionation of these melts took place to produce the s y e n i t e quartz syenite-granite, n e p h e l i n e syenite-syenite and n e p h e l i n i t e - n e p h e l i n e syenite suites, as well as the phonolite dyke spectrum, but the already highly evolved nature of the p r i m a r y m a g m a s is a product of the partial melting of mantle particularly enriched in volatiles, alkalis, carbonate and the range of incompatible elements
characteristic of alkaline m a g m a s , and is not a product of a long line of liquid descent. It is intended to elaborate this model in detail elsewhere, but it must be stressed that it has been developed to explain the rather special case of the C h i l w a province, although it is believed that modified versions can be applied to other intraplate alkaline provinces, especially those in w h i c h felsic rocks are p r e d o m i n a n t a n d basic rocks are scarce. ACKNOWLEDGMENTS: We should like to thank Dr A. C. Bishop, Dr K. Bloomfield, Dr M. S. Garson, Dr P. Henderson and Dr D. R. C. Kempe for reading the manuscript and for making suggestions for its improvement. We are also grateful to the Director of the Geological Survey of Malawi for giving permission for the removal of samples from the Survey Collection, Dr K. G. Cox for providing additional material and Mr T. F. Johnston for arranging access to the collection of the University of Leeds. Mr V. K. Din determined some of the trace elements and Miss V. Jones drafted the figures.
References BAILEY, D. K. 1964. Crustal warping--a possible tectonic control of alkaline magmatism. J. geophys. Res. 69, 1103-11. -1977. Igneous Rocks and the Degassing o f the Earth, Ninth Tomkeieff Memorial Lecture, pp. 1-41. Geology Department, University of Newcastle upon Tyne. -1980. Volcanism, Earth degassing and replenished lithosphere mantle. Phil. Trans. R. Soc. Lond., Ser. A, 297, 309 22. - - & MACDONALD,R. 1969. Alkali-feldspar fractionation trends and the derivation of peralkaline liquids. Am. J. Sci. 267, 242-8. BINSTED, N. 1981. The system Ab-Ne-NaC1-H20. Prog. Exp. Petrol. NERC Set. D, 5, 34-6. BLOOMFIELD,K. 1965. The geology of the Zomba area. Bull. geol. Surv. Malawi, 16. DEER, W. A., HOWIE, R. A. & ZUSSMAN,J. 1963. Rock.forming minerals, Vol. 4, Framework Silicates, Longmans, London. DIXEV, F. 1956. The East African Rift System. Bull. Suppl. colon, geol. miner. Res. 1. - - S M I T H , W. C. & BISSET, C. B 1937, The Chilwa Series of southern Nyasaland ; a group of alkaline and other intrusive and extrusive rocks and associated limestones. Bull. geol. Surv. Nyasaland, 5. GARSON, M. S. 1960. The geology of the Lake Chilwa area. Bull. geol. Surv. Nyasaland, 12. -1962. The Tundulu carbonatite ring-complex in southern Nyasaland. Mere. geol. Surv. Nyasaland, 2. -1965a. Carbonatite and agglomerate vents in the western Shire Valley. Mem. geol. Surv, MalawL 3.
Carbonatites in southern Malawi. Bull. geol. Surv. Malawi, 15. -1966. Carbonatites in Malawi. In: TUTTLE O. F. & GITTINS,J. (eds.) Carbonatites, pp. 33-71. WileyInterscience, New York. - & SMITH, W. C. 1958. Chilwa Island. Mem. geol. Surv. Nyasaland, 1. --& WALSrtaW, R. D. 1969. The geology of the Mlanje area. Bull. geol. Surv. Malawi, 21. LE BAS, M. J. 1977. Carbonatite-Nephelinite Volcanism. Wiley-Interscience, New York. LE MAtTRE, R. W. 1976. The chemical variability of some common igneous rocks. J. Petrol. 17, 589637. MACDONALD,R. & BAILEY,D. K. 1973. The chemistry of the peralkaline oversaturated obsidians. Prof Pap. U.S. Geol. Surv. 440-N-1, pp. 1-37. - - , CROSSLEY,R. & WATERHOUSE,K. S. 1983. Karroo basalts of southern Malawi and their regional petrogenetic significance. Mineral. Mag. 47, 3819. STILLMAN, C. J. & COX, K. G. 1960. The Chikala Hill syenite-complex of southern Nyasaland. Trans. Proc. geol. Soc. Aft. 63, 99 117. STRECKEISEN,A. 1976. To each plutonic rock its proper name. Earth Sci. Rev. 12, 1-33. TURTLE, O. F. & BOWEN, N. L. 1958. Origin of granite in the light of experimental studies in the system NaA1Si308-SiOz-H20. Mem. geol. Soe. Am. 74. VAIL, J. R. 1964. Mesozoic igneous activity in central Africa. Rep. Int. Geol. Cong. 16, 212-25. - & MALLICK,D. I. J. 1965. The Mongolowe Hills nepheline-syenite ring-complex, southern Malawi. Rec. geol. Surv. Malawi, 3, 49-60. - - 1 9 6 5 b .
Petrochemistry of the Chilwa alkaline province & MONKMAN, L. J. 1960. A geological reconnaissance survey of the Chaone Hill ring complex, southern Nyasaland. Trans. Proc. geol. Soc. S. Afr. 63, 119-35. WOOLLEV, A. R. 1969. Some aspects of fenitization with particular reference to Chilwa Island and Kangankunde, Malawi. Bull. Br. Mus. nat. Hist. (Mineral.) 2, 189-219. - - & GARSON, M. S. 1970. Petrochemical and tectonic relationship of the Malawi carbonatitealkaline province and the Lupata-Lebombo vol-
-
-
355
canics. In: CLIFFORD, T. N. & GASS, I. G. (eds) African Magmatism and Tectonics, pp. 237-62. Oliver & Boyd, Edinburgh. - - - - & PLATT,R. G. 1986. The mineralogy of nepheline syenite complexes from the northern part of the Chilwa Province, Malawi. Mineral. Mag. 50, 597610. , BEVAN, J. C. & ELLIOTT, C. J. 1979. The Karroo dolerites of southern Malawi and their regional geochemical implications. Mineral. Mag, 43, 48795.
A. R. WOOLLEY & G. C. JONES, Department of Mineralogy, British Museum (Natural History), Cromwell Road, London SW7 5BD, U.K.
Niger-Nigerian alkaline ring complexes: a classic example of African Phanerozoic anorogenic mid-plate magmatism P. Bowden, R. Black, R. F. Martin, E. C. Ike, J. A. Kinnaird & R. A. Batchelor SUMMARY: The Pan-African orogeny, regarded as a major event in the construction of Gondwanaland, played a dominant role in providing a variety of source materials for mineralization, melting or assimilation as components of the Benin-Nigerian shield. Although the parent magmas of the silica-oversaturated alkaline rocks which constitute the Niger-Nigerian anorogenic province were initiated in the asthenosphere, it was the exploitation and reactivation of the Pan-African shear zones and transcurrent faults during the fragmentation of Gondwanaland that controlled the locations of Phanerozoic intra-plate magmatism in W Africa. The Niger-Nigerian anorogenic province represents one of those regions where progressive uplift has been accompanied by periodic sequential development of chains of volcanoes now exposed as ring complexes with Palaeozoic and Mesozoic ages. The magmatic lineage comparable with other alkaline provinces can be established from preserved volcanic sequences. Associated andesite compositions can be attributed to magma mixing. In Niger, an important petrogenetic parameter is the occurrence of leucogabbros, anorthosites and monzoanorthosites as part of the Palaeozoic sub-volcanic association with syenites, peralkaline granites and biotite granites. The partially eroded volcanic cover of rhyolitic ignimbrites has provided the source for substantial uranium deposits. The southern centres of upper-Silurian-lower-Devonian age are mineralized with columbite and cassiterite. The general geological and geochemical features of the Nigerian Triassic-Jurassic anorogenic centres are well known. While their magmatic derivation from mantle and crustal sources can be conclusively demonstrated, petrological and geochemical research has shown that many of the Nigerian anorogenic centres demonstrate substantial evidence for post-magmatic metasomatism linked to mineralization. There are certain parallels between alkali metasomatism in alkaline granite ring complexes and fenitization associated with carbonatites.
Introduction Of necessity this review has been restricted to a limited range of topics on research completed over the past decade in W Africa, which has particular relevance to the anorogenic alkaline ring complex province in Niger and Nigeria. For example, the Pan-African orogeny has played a dominant role in providing a variety of source materials for mineralization, melting or assimilation as components of the continental lithosphere. Recently it has been accepted that the Pan-African event represents a major period in the construction of Gondwanaland as well as defining the boundary between the end of the Precambrian and the beginning of the Phanerozoic in Africa. Particularly relevant in the application of plate tectonics to the late Precambrian in W Africa are the discoveries of volcanoclastic sequences as island-arc and marginal trough material in the Hoggar (Caby 1970; Caby et al. 1981), associated with ophiolites in Morocco (Leblanc 1981), in Saudi Arabia (Bakor et al. 1976; Greenwood et al. 1976) and in Egypt and Sudan (Garson & Shalaby 1976). For W Africa
in general, and the Niger-Nigerian Precambrian basement in particular, Pan-African reactivation may represent the deep structural level of the same processes produced by the collision of India with N E Asia (Black 1984). Although the Phanerozoic parental magmas of African silica-oversaturated alkaline rocks were generated in the asthenosphere, it was the exploitation and reactivation of the Pan-African shear zones and transcurrent faults that controlled the locations of these mid-plate A-type ring complexes in Niger, Nigeria and elsewhere. The sub-volcanic ring complexes of Niger and Nigeria form three separate sub-provinces of anorogenic mid-plate magmatism spaced out in a generally southerly direction from Adrar Bous to Tarrouadji (Air) (Fig. 1), Zinder to the Mounio massif (southern Niger) (Fig. 2) and Dutse to Afu (northern Nigeria) (Fig. 3). Over 80 ring structures in these sub-provinces exhibit migratory magmatic centres spread over 340 Ma, which were episodically concentrated into three 6070 Ma periods: upper Ordovician to lower Devonian, Carboniferous to Permian and upper Triassic to upper Jurassic. Thus the Air (480-
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 357-379.
357
358
P. Bowden et al.
Southern.,0er
]
45~
|
~
"-:.
:
o
FIG. 1. Geological sketch map of the Air (Niger) massif illustrating the distribution of ring complexes and their Palaeozoic Ordovician-Silurian ages (corrected to Rb-Sr decay constant 1.42 x 101 ~~- 1): B, Adrar Bous; TK, Tamgak; O, Ofoud; A, Aroyan; Ta, Tarrouadji. Additional ages of 435 Ma (Agalak) and 426 Ma (Iskou) taken from Karche & Vachette (1976) are identified by broken underlining. The age of Ofoud (Agarageur) is taken from Bowden et al. (1976) and identified by parentheses, l, Meugueur-Meugueur basic cone sheet; 2, syenitic alkaline granite complexes associated with gabbros and anorthositic rocks; 3, syenite alkaline granite complexes; 4, dominantly volcanic succession of rhyolitic ignimbrite with occasionally intercalated basalt, hawaiite and trachyte; 5, Recent volcanic rocks; 6, Precambrian basement; 7, edges of Cretaceous to Recent sedimentary cover; 8, Palaeozoic sedimentary basins; 9, structural trends in the Precambrian basement.
~
G
,
:
o
,
r
e
,oo
, " M258 ,-"N I G E R I A
9~ "T'"
u
..........
I
FIG. 2. Geological sketch map of the southern Niger and Nigerian border region illustrating the distribution of ring complexes, and their PermoCarboniferous ages (corrected to Rb-Sr decay constant 1.42 x 1011 0~- 1): Z, Zinder; G, Gour6; D, Daura; M, Matsena. Daura and Matsena belong geographically to the Nigerian centres but chronologically to the southern Niger group. Additional ages are identified by broken underlining. The age in parentheses represents an average of values reported by Karche & Vachette (1976) for TchouniZarniski ; the other age of 289 Ma (Adoumchi) is taken from Bowden et al. (1976). 1 edge of Cretaceous to Recent sedimentary cover; 2. dominantly volcanic successions of rhyolitic ignimbrite; 3 syenite alkaline granite complexes;4 Precambrian basement; 5 structural trends in the Precambrian basement. Geological information from Black (1963) and Mignon (1970). magmatism. This aspect is considered in the first part of this review using examples from Niger. The aspect emphasized in the second part of this review is concerned with the post-magmatic crystallization history within the ring complexes. The reactions of residual fluids in the Nigerian biotite granites and the associated U - N b , Z n - S n mineralization are used as examples.
Tectonic and structural controls of anorogenic magmatism 400 Ma), S Niger (320-290 Ma) and Nigerian (215-140 Ma) alkaline sub-provinces are a unique feature in the world of Palaeozoic to Mesozoic within-plate anorogenic volcanism and plutonism with progressive southern shift of centres of magmatic activity (Bowden e t al. 1976; Karche & Vachette 1976; Karche & Moreau 1978). Similar rock types are developed in all three subprovinces. Thus both the processes in the asthenosphere and the structure of the continental lithosphere exert an important tectonic and chemical control on the products of anorogenic
Tectonic setting: influence of the Pan-African orogeny The evolution of the alkaline igneous ring complexes known as the Niger-Nigerian anorogenie province occupied the period between the construction of Gondwanaland and its immediate fragmentation. The magmatism presents a record of intra-plate activity of asthenospheric magma interacting with continental lithosphere. Such Phanerozoic alkaline magmatism in the Palaeo-
dS~
~
~
,oow
Kano
'~,5:
Kano .......
(.f " ~[~]-~?
O
WA
tXgGODU
Zaria
fir,/
IEN 173 175 FIIR1WAIGURAKA 170 ZUKU@~ SAIYA ~ N G O L O
~SH~WA
3ERE-SANGA
\
OBa~.,
@ ZARANDA186
tt ~
~KERKU ....
i
:.-
% .....
% !',iiIf
!: :Jb u~uRu
!
o..........
~
Eoqeol sedimenlarvCOVe,
[]
Volcanltrocks
[]
.....................
[]
........
~
Syeqlle$
0
'~I~
lo
:,,,
30
~,l,:s
JAK
FIG. 3. Sketch map of the distribution of anorogenic ring complexes in Nigeria with their Triassic-Jurassic ages in relation to the direction of major crustal lineaments shown by the black arrows. Ages from Rahaman et al. (1984). Based on geological map compiled by Kinnaird (1981).
360
P . B o w d e n e t al.
zoic and Mesozoic eras is not unique to Niger and Nigeria but occurs elsewhere in Africa confined to domains of mobile belts generated during the Pan-African orogeny. When considering contemporaneous orogenic cycles in other continents, the Pan-African corresponds to a major event in the Earth's history with worldwide distribution of ramified mobile belts marking the limit between the Precambrian and the Phanerozoic. In effect the Pan-African is recognized as part of the sequence of events which welded Gondwanaland together. Studies completed in the last 10 years along the eastern margin of the W African craton have shown that the Pan-African orogeny result from the collision between the passive margin of the W African craton and the active margin of a continent to the E. The resultant rock types generated by orogeny, subduction and obduction form the Tuareg shield to the N (Fig. 4) and the Benin-Nigerian shield to the S (Caby et al. 1981). The alkaline igneous rocks which constitute the Niger (Air) group of Palaeozoic age were emplaced into the southern extremity of the Tuareg shield, and the Nigerian alkaline ring complexes of Mesozoic age were intruded into the northern segment of the Benin-Nigerian shield. Such Pan-African belts are often characterized by thick continental crust but thin subcrustal lithosphere. The Tuareg shield (Fig. 4) is divided from W to E by major N-S shear zones into three tectonic
Tuareg Shield
-~ Pan AfrLcan m o b i l e belt West A f r i c a n c r a t o n
0
500kin
Fro. 4. Schematic map showing the location of the Pan-African mobile belt in W Africa. (From Black 1984.)
domains: a western domain known as the Pharusian belt, characterized by the abundance of Upper Proterozoic volcano-clastic sequences; a central polycyclic Hoggar-A~r domain E of the main shear zone, largely composed of ancient gneisses, reactivated and injected by abundant calc-alkaline to subalkaline granitoids during the Pan-African orogeny; an eastern Hoggar-Tenere domain which was apparently stabilized at an early stage of the Pan-African around 730 Ma ago. The late Pan-African event marking the EW collision affected the entire Tuareg shield overprinting earlier tectonic patterns. Prior to collision, the palaeogeographic picture along the western edge of the shield was that of an island arc to the W and a cordilleran-type continental margin to the E. Ocean closure is marked by a suture outlined by a string of positive-gravity anomalies corresponding to basic and ultrabasic complexes, and perhaps ophiolites, which can be traced over a distance of 2000 km along the eastern margin of the W African craton (Bayer & Lesquer 1978). The Benin-Nigerian shield (Fig. 4) and its north-eastward prolongation to the Damagaram of S Niger represents the southern continuation of the Pan-African belt. Ideally on the platetectonic model it should correspond to the same divisions as those proposed for the Hoggar-Iforas segment by Caby et al. 1981). Considering the Pan-African belt E of the W African craton as a whole, the Pharusian cordilleran-type assemblages are preserved in a re-entrant of the craton. To the S, the Benin-Nigerian shield is directly thrust onto the passive margin of the craton. Here, the ocean opening was either very small or the ocean crust and active margin were entirely subducted. In SW Nigeria the basement consists of charnockitic rocks of unknown age and a gneissic complex comprising banded gneisses, migmatites, quartzites, schists, biotite__amphibole gneisses, amphibolites and marbles metamorphosed to the almandine-amphibolite facies. They are cut by foliated granites which have yielded Lower Proterozic ages (Rb-Sr wholerock isochrons of 2267_+30 Ma and 2283_+ 70 Ma) and by porphyritic granites, subordinate quartz-diorites and two-mica granites which are considered to be Pan-African. The central part of Nigeria, locus of the Triassic-Jurassic alkaline ring complexes, is occupied by abundant PanAfrican granitoids which have been dated at 700-500 Ma (van Breemen et al. 1977). Although it is difficult to relate Pan-African granitoids to subduction zones there is an overall zonation. Diorites and granodiorites with low initial strontium ratios predominate in the Pharusian do-
Niger-Nigerian alkaline ring complexes mains (Liegeois & Black 1984), whereas monzogranites and quartz monzonites dominate in the polycyclic Hoggar-Air domain. Similar contrasts are noted between the western part of the Benin shield and the central Nigerian shield. The location of the Palaeozoic and Mesozoic oversaturated alkaline provinces of Air, S Niger and Nigeria in the former Pan-African mobile belts appears to be related to a regime of distension and reactivation of Pan-African shear zones and transcurrent faults. The loci of anorogenie magmatism are often focussed at the intersection of the N-S megashears and the NNW-SSE, N N E - S S W transcurrent fault system. Elsewhere in Africa, Phanerozoic silicaoversaturated alkaline magmatism is likewise confined to Pan-African domains. By Tertiary-Quaternary times mixed silicasaturated and silica-undersaturated alkali basaltphonolite-trachyte-rhyolite volcanism developed widely in the Pan-African domains and was associated with domal uplift and reactivation of pre-existing Pan-African fault patterns, often on sites of former silica-saturated alkaline provinces in Air and Nigeria as well as elsewhere in Africa. To summarize, it is important to note the variety of rock types from mantle and crustal sources found in the Pan-African mobile belt, the dominant and influential megashear zones, the continued progressive uplift in the Palaeozoic, and the critical structural control of the location of the ring centres in Niger and Nigeria defined by Pan-African lineaments, megashears and the intersecting transcurrent wrench fault system. On fragmentation of Gondwanaland the PanAfrican crustal lineaments, which may fracture through the continental lithosphere, control the directions of rifting and continued separation. Perhaps it is not surprising that large-scale lineaments on the African continent can be traced, sometimes with a change in direction, as transform faults into the oceanic lithosphere (Sykes 1978; Guiraudetal. 1985). Tectonic constraints on alkaline magmatism From the plate-tectonic viewpoint there is a correlation between alkaline magmatism and changes in the direction of plate movements provoking reactivation of lithospheric shear zones and rifting within plates. If this reactivation caused a reversal in the sense of movement along the N-S megashear zones in the Pan-African domains, the associated oblique sets of transcurrent faults originally under compression would open and propagate as tensional faults. This would allow fracturing through the continental lithosphere causing pressure release, channelling
361
of volatiles, partial melting and generation of magma from the asthenosphere. The correspondence between within-plate magmatic activity in the Niger-Nigerian province and orogens at the African plate margins has been known for some time (Karche & Vachette 1976). Thus the overall plate-tectonic approach suggests that it is the within-plate stress fields and fault reactivation which controls the sites of alkaline magmatism in the continental lithosphere and provides the triggering mechanism for diapiric processes of magma generation in the asthenosphere. On this model the episodic partly mobile thermal anomaly in the mantle (Bowden & Karche 1984) is relegated from a causal plume to a passive hotspot whose location is determined by the exploitation of pre-existing zones of weakness in the African plate. Anorogenic magmatic evolution The magmatic evolution of the anorogenic complexes is comparable with that of other alkaline provinces. It can be established from preserved volcanic sequences of comenditic ignimbrites and lavas with minor amounts of intercalated intermediate and basic lavas (Turner & Bowden 1979). Associated andesite compositions can be attributed to mixing of acid and basic endmembers (Bowden & Karche 1984). The subvolcanic assemblages include minor gabbros, monzogabbros, nordmarkites, syenites, albiterich and albite-poor aegirine-arfvedsonite-granites, fayalite-hedenbergite granites, amphibolebiotite granites and biotite granites. Their relative proportions change from N to S: peralkaline granites and quartz syenites predominate in the Air, where they may be associated with anorthosites, leucogabbros and lenses of Fe-Ti oxides, often as layered funnel-shaped intrusions (Black 1965; Black et al. 1967; Husch & Moreau 1982; Leger 1985), and S Niger (Black 1963; Mignon 1970), whereas minor proportions of peralkaline granites and major proportions of biotite granites are the most prevalent rock types in Nigeria (Bowden & Turner 1974; Hossain & Turaki 1983). The three provinces lie between longitudes 8~ and 10~ in a N-S zone 1300 km long of the former Pan-African orogenic belt. In Air extensive transcurrent faulting occurred prior to the emplacement of the ring complexes (Fig. 5) and was accompanied by crustal doming as indicated by the southerly tapering out of the Palaeozoic sedimentary rocks along the western border (Karche & Vachette 1976). To the S in Nigeria, Rahaman et al. (1984) have shown that the age migration of centres occurred along E N E - W S W and N N W - S S E lineaments (Fig. 3); the E N E -
362
P. B o w d e n et al. i
AGALAK
9
9
J
ORAYAN
BAGUEZANS
"
\
\
~TARAOUADJI
a ' a n ' ~~ ' "
9'OO'E
FIG. 5. Geological sketch map of the southern section of the Air massif showing the dominant structural control of the sets of transcurrent faults on the location of Palaeozoic ring complexes and subsequent Tertiary to Recent alkaline basaltic magmatism: 1, Palaeozoic ring complexes; 2, Tertiary alkaline magmatism; 3, QuaternaryRecent alkaline magmatism. (From Black et al. 1985.)
WSW lineaments correspond to the direction of late-Pan-African dextral transcurrent faults in the basement and are parallel to the marginal faults of the Benue trough which has recently been interpreted as a pull-apart basin determined by sinistral shear (Benkhelil 1982), and which contains transitional basalts and alkali rhyolite with an Rb-Sr age of 113_+3 Ma (Umedji & Caen-Vachette 1983).
The majority of the Niger massifs were emplaced into the southern extension of the Precambrian Tuareg shield as roots of volcanoes which erupted minor basalts, trachytes and mainly peralkaline ignimbritic rhyolites. From work completed at Bilete (Fig. 6), where volcanic rocks are still preserved, it is evident that two magmas, one basaltic and the other acid, were freely available during the early eruptive history
Niger-Nigerian alkaline ring complexes
363
FIG. 6. View of the northern end of the Bilete volcanic massif, Niger, with abundant ignimbrite units. of each centre. However, the acid magma was dominant, producing extensive ignimbrite flows enriched in uranium. Part of this uranium has been deposited in the coeval adjacent sedimentary basin to the W (Fig. 7) and is currently mined at Arlit and elsewhere in Niger (Fig. 8). Although basaltic magmatism is of limited occurrence, erosion at the sub-volcanic level in Air has revealed anorthosites and gabbroic rocks associated with the sub-volcanic alkaline ring complexes outcropping over more than 500 km 2 at seven centres. Between N and S over a distance of 250 km similar associations are found at Adrar Bous, Enfoud, Taguei and Abontorok (Karche & Moreau 1977; Moreau et al. 1978; Morel & Moreau 1979; Husch & Moreau 1982; Moreau 1982), as well as at Iskou (Leger 1980, 1985) and Meugueur-Meugueur (Black 1965; Black et al. 1967).
The Phanerozoic anorthosite association, Air, Niger Although anorthosite massifs are exceptionally abundant in the Precambrian worldwide, there are few well-documented examples in the Phanerozoic apart from the Air occurrences. However, they provide an important critical appraisal of the possible petrogenetic link between the two associations of gabbro-anorthosite and syenitegranite in the same centres. The key to this link
lies in the interpretation of monzoanorthosites (Husch & Moreau 1982). This aspect has been well researched by Moreau and his co-workers. Of particular relevance are the associations at Ofoud, Taguei and Abontorok. An alternative aspect of the fundamental petrogenetic evolution of sub-volcanic anorthosites and related rocks is provided by Leger (1980, 1985) based upon his studies at Iskou. These and other aspects of the gabbro-anorthosite suite are briefly reviewed here. For further details the reader is recommended to consult the original publications. A summary of the geological features of the AYr (Niger) anorthosites is provided in Table 1. Ofoud (Figs 1 and 9) is one of the largest leucogabbro-anorthosite syenite-granite ring complexes in Air with an areal extent of some 900 km2.Essentially half the massif in the S is composed of basic rocks, but unlike Adrar Bous (Fig. 9) these are mainly anorthosites and leucogabbros with only subsidiary amounts of gabbro and microgabbro. The high central zone of the Ofoud complex (Fig. 10) consists of a circular syenite-alkaline granite massif which cuts the anorthosites to the S and E. The basic rocks are bordered to the S by the high granite ridge of Agueraguer, and to the N and E by syenites and microsyenites. The border zones of leucogabbro have partially aligned plagioclase crystals (An6o) varying from 5 to 20 mm in length imparting a foliation that dips concentrically
364
P. Bowden et al. 1200 Mersoi
. ....
U w i t h Cu I I O 0 at A z e l i k - -
Trhazar series
Lower Cretaceous
I00(
9O( U at irnouraren~:=:> , Agadez series 800 ~
Continental i ntercalaire
700 600 5OO A'I" r
massif
U at Arlit
-
9 i i9. . -:::.:.:::
: -i.
I
-
0
Permian
400
200 U at Madaouela~
k
Izegouanda series
~176
Tagora series
Lower Marine,Carbonifepicontinental erous
~Talak series
FIG. 8. Schematic stratigraphic column for sedimentary deposits adjacent to the Air massif, showing the location of uranium deposits : horizontal shading, shales/siltstones; fine stipple, sandstones; diagonal etching, sandstones with current bedding;., conglomerates: v, volcanogenic sandstones with analcime.
In G a l l basin AGADES.="
FIG. 7. Geological sketch map of the basins to the west of the Air massif. Tin Mersoi basin, Upper Palaeozoic sedimentary rocks; Arlit (Izegouda series), Permian; Agades series, Trias to Jurassic; In Gall basin, Cretaceous 9The thick black lines represent postdepositional faults, related to continued movement of the Air block, providing 'traps' for the retention of uranium in carbonaceous sedimentary horizons. (From Bowden et al. 1981.) 200-60 ~ towards the centre of the complex giving the impression of a funnel-shaped array. The structure is rather like a lopolith with the feeder zone offset towards the S. Away from the border zone the texture coarsens with modally less olivine and clinopyroxene. Towards the centre anorthosite contains plagioclase crystals up to 48 cm long. Some rhythmic layering is developed near the margins. Interstitial minerals in the main anorthosite zone are brown to green amphibole, titanomagnetite with coronas of secondary biotite, minor olivine and, rarely, primary clinopyroxene (Table 2). In the coarsertextured anorthosites, titanomagnetite and ilmenite constitute about 10~ of the whole rock. However, one distinctive feature of the Ofoud
basic series is the presence of numerous layers and lenses of titanomagnetite (Karche & Moreau 1978) up to 1 0 m thick intercalated with the anorthosites. These ores contain up to 50~ modal euhedral cumulus olivine. The attitude of these mafic titanomagnetite-enriched layers confirms the overall inward dip of the layered leucogabbroanorthosite units (Fig. 10). Major-element compositions for both Ofoud and Adrar Bous are given in Table 3. TABLE 1. Anorthosites a n d associated rocks (ATr, Niger) Sub-volcanic centres
Adrar Bous, Tamgak, Ofoud, Taguei, Meugueur-Meugueur, Abontorok, Iskou Palaeozoic age 470 412 Ma Funnel-shaped With oriented and inclined intrusions plagioclase crystals near border zones; dips of 20~ ~towards centre Petrographic Gabbros, leucogabbros, association anorthosites, monzogabbros, monzoanorthosites, ferrosyenites, syenites, alkaline and peralkaline granites Enclaves of leucotroctolite; leuconorite in leucogabbro Titanomagnetite lenses in anorthosites and leucogabbros Data from Bowden et al. (1976), Karche & Vachette (1976), Karche & Moreau (1977), Moreau et al. (1978), Leger (1980) and Husch & Moreau (1982).
Niger-Nigerian alkaline ring complexes
FIG. 9. Geological sketch maps of four principal anorthosite centres Air, Niger: 1, basic rocks (gabbros, leucogabbros); 2, syenite-granite series; 3, monzoanorthosite. (From Husch & Moreau 1982.)
365
A similar concentric array of aligned platy plagioclase crystals with an inward-dipping foliation (about 60 ~ can be seen at Abontorok, a circular plug about 2.5 km in diameter (Fig. 11). Within the basal series there is an obvious variation in modal content from the outside towards the centre (Fig. 12). With increasing distance from the centre, the amount of plagioclase diminishes while olivine and opaque minerals increase. The plagioclase crystals become more equidimensional and alignment less pronounced. At about 100 m from the contact the rock contains approximately 65~-70% modal plagioclase. Within 5 m of the contact the grain size decreases dramatically. Cross-cutting the leucogabbro-anorthosite unit are dykes of syenite and alkaline granite, while the central part of the small massif is invaded by syenite breccia. In the leucogabbro margins and anorthositic centre, plagioclase crystals are compositionally zoned from a core of An60 to a rim of An40. Furthermore, the proportions of olivine, clinopyroxene and titanomagnetite vary considerably. This tiny massif of Abontorok convincingly demonstrates the evidence for low-pressure fractional crystallization of anorthositic liquids in the Air sub-volcanic massifs. The apparent centripetal fractionation witnessed at Abontorok and Ofoud suggests that flow differentiation may have been a dominant factor in their development.
FIG. 10. Distance view of Ofoud massif with syenite-granite peak. The lower slopes and middle ground contain inwardly-dipping anorthosite-titanomagnetite units of the Ofoud complex. The foreground shows coarse-grained anorthosite at the southern edge of the Ofoud complex.
366
P. Bowden et al. TABLE 2. Modal analyses of anorthosites and related rocks from Adrar Bous and
Ofoud, Agr, Niger vol. %
1
Quartz Alkali feldspar Plagioclase Clinopyroxene Olivine Amphibole Opaques Biotite Others
2
--62.5 7.2 29.4 -0.9 ---
--59.9 32.2 6.5 -1.4 ---
3
4
5
6
--68.3 14.9 14.3 -1.2 --
0.7 1.2 93.5 -0.3 3.0 1.0 0.3
-0.7 94.4 0.4 -3.1 0.4 --
1.0 5.4 81.8 --8.6 3.2 --
1.3
--
1.0
--
1, 2, 3, Adrar Bous; 4, 5, 6, Ofoud. Details from Husch & Moreau (1982). Major-element analyses in Table 3. TABLE 3. Major-element analyses of anorthosites
and related rocks from Adrar Bous and Ofoud, A~r, Niger wt. %
SiO2
1
2
3
4
5
6
Fe203 FeO MgO CaO Na20 K20 TiO2 P205 LOI
46.33 49.05 47.81 52.32 52.75 49.87 17.73 15.74 20.29 24.61 25.82 23.14 0.89 2.19 0.97 1.85 0.33 0.86 10.32 7.09 6.65 2.99 2.53 5.03 0.16 0.15 0.11 0.08 0.04 0.06 9.75 13.55 11.46 9.30 9.62 9.04 2.59 2.47 2.53 4.51 4.95 4.14 0.15 0.16 0.11 0.80 0.69 0.50 0.61 0.86 0.26 1.03 0.56 1.16 0.17 0.09 0.01 0.07 0.15 0.24 0.60 -0.53 0.97 1.66 2.38
Total
98.90 98.92 98.92 99.63 99.76 99.30
AI203
c o m p o s i t i o n s are p l o t t e d in Fig. 13 as Niggli p a r a m e t e r s . Syenite a n d g r a n i t e c o m p o s i t i o n s for N i g e r are i n c l u d e d in this d i a g r a m as well as t h e t r e n d for the N i g e r i a n ring c o m p l e x e s .
1,2,3, Adrar Bous; 4,5,6, Ofoud. LOI, loss on ignition. Details in Husch & Moreau (1982). Modal analyses in Table 2. F l o w sorting o f crystals at the initial e m p l a c e m e n t stage m u s t h a v e b e e n a i d e d by o t h e r m e c h a n i s m s o f m i n e r a l layering on cooling. M a j o r - e l e m e n t c o m p o s i t i o n s o f the various rock types f o u n d at A b o n t o r o k are p r e s e n t e d in Table 4, a n d c a n be c o m p a r e d w i t h c o r r e s p o n d i n g d a t a for A d r a r Bous a n d O f o u d in Table 3. T h e s e m a j o r - e l e m e n t
Petrogenesis
T h e r e is little d o u b t t h a t t h e e v o l u t i o n o f a n o r t h o s i t i c c o m p l e x e s is the result o f a n o r o g e n i c m a g m a t i s m (Emslie 1978) in b o t h the P r e c a m b r i a n a n d the P h a n e r o z o i c . H o w e v e r , the n a t u r e o f the p a r e n t m a g m a a n d t h e m e c h a n i s m s by w h i c h c o n c e n t r a t i o n s o f plagioclase c a n o c c u r are the subject of c o n t r o v e r s y (Emslie 1973, 1978; D e W a a r d et al. 1974; R y d e r 1974). V a r i o u s c o m p o n e n t s h a v e b e e n p r o p o s e d for the c o m p o sition of the p a r e n t a l m a g m a o f a n o r t h o s i t i c rocks a n d the possible link w i t h the associated acid rock types (Fig. 13). Possible p a r e n t a l m a g m a c o m p o s i t i o n s include l e u c o g a b b r o (Budd i n g t o n 1969; S i m m o n s & H a n s o n 1978), g a b b r o (Emslie 1978), m o n z o g a b b r o ( D e W a a r d et al. 1974; D u c h e s n e et al. 1974) a n d diorite ( P h i l p o t t s 1966; G r e e n 1969). A c c o r d i n g to Leger (1980, 1985) the p e t r o g e n e t i c e v o l u t i o n at the I s k o u m a s s i f (Air, Niger) s u p p o r t s the ideas of M o r s e (1979a, b) w h o suggests t h a t a basaltic m a g m a could evolve t o w a r d s a l u m i n a - r i c h c o m p o s i t i o n by f r a c t i o n a t i o n o f m e t a s t a b l e olivine. E
W
0 3 km I
I
FiG. 11. Interpretative cross-section of Abontorok : 1, Pan-African domain of Tuareg shield ; 2, gabbro border facies; 3, leucogabbro; 4, gabbroic anorthosite; 5, anorthosite ; 6, syenite; 7, granite and granite porphyry; 8, central syenitic breccia. (From Moreau 1982.)
367
Niger-Nigerian alkaline ring complexes
Major-element analyses of the gabbroanorthosite series, Abontorok, Agr, Niger
TABLE 4.
ABONTOROK
"4" 15
[\
I ~,
~,ol -\\
',,,, cpx
01 .5
centre
I 05
I 1
km
Fza. 12. Modal distribution of major mineral phases across the Abontorok complex, Air, Niger: plg, plagioclase; ol, olivine ; cpx, clinopyroxene; op, opaque minerals. (From Husch & Moreau 1982.)
wt. ~
Abl
Ab4
Ab9
Abl0
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO N%O K20 P205 LOI
42.74 5.32 16.26 2.18 14.21 0.21 5.92 7.23 3.37 0.91 0.79 1.09
45.96 3.62 16.88 3.23 10.53 0.20 3.75 8.91 3.82 1.31 1.14 0.23
53.05 0.79 25.24 0.68 2.33 0.04 0.54 9.74 4.32 0.81 0.42 0.99
54.08 0.72 25.25 0.64 2.64 0.05 0.64 9.64 4.30 1.32 0.26 0.77
Total
99.63
100.29
99.21
100.60
Abl, ferrogabbro; Ab4, anorthositic gabbro; Ab9, Abl0, anorthosite. Taken from Moreau (1982).
c+fm
t
III FII I" t 0 I I OIBm,,II II I I I IB~A I II
i
11211
tI i i
/
9
~,-A n 50 Ab50 /
1(13
w a
CO Oi 0
I o
, t n
alc
t
i i
al
FIG. 13. Niggli diagram describing the major-element variations in the Air ring complexes: squares and triangles, gabbro-anorthosite series; circles and diamonds, syenite-granite series; broken lines, trend of Nigerian anorogenic complexes. Analyses compiled from many sources, and summarized by Husch & Moreau (1982) and Leger (1985).
368
P. B o w d e n et al.
At Iskou and the other Niger anorthositic centres (Table 1) Leger (1985) has identified three distinct stages of magmatic evolution from the variations recorded in the clinopyroxene compositions (Fig. 14). The first stage is indicated by the occurrence of leuconorite cumulates and orthopyroxene relics included in olivine in the leucotroctolites. The leuconorites probably developed during crystallization at high pressure (around 10-11 kb (Green & Ringwood 1976)). Such cumulates were generated during a shortlived residence time at the crust-mantle boundary. Such sub-crustal magma chambers may have been the source which fed the important Meugueur-Meugueur ring dyke. The Meugueur-Meugueur structure is a large ring dyke 65 km in diameter which is somewhat eccentric in its relation to the Ofoud massif. The width of the dyke varies between 200 and 400 m. Its occurrence is noted topographically by erosional hollows bordered by the Precambrian basement or the sub-volcanic centres which it dissects. Compositionally the ring dyke is variable with an abundance of enclaves of leucotroctolite, leuconorite, peridotite, leucogabbro and rare ferros_yenite set in a matrix of olivine monzogabbro (30-40 vol. ~ olivine, plagioclase An 6~176 and rare clinopyroxene) with olivinerich concentrates towards the centre and more plagioclase-rich zones near the margins. Even along its perimeter the Meugueur-Meugueur ring shows further textural and compositional contrasts as well as a variable-enclave suite. For example, magmatic layering with an almost
/ / ~
Fo"
k
/ --'-.
Ar
, "~------
k'~-. ~
vertical dip is apparent in the southern part. In contrast, the N E quadrant contains a dark picritic microgabbro, anorthositic enclaves, plagioclase xenocrysts and a border containing syenitic to granitic margins. The overall impression of this remarkable structure is a steep-sided composite ring dyke (Table 5) confined to a deeplypenetrating ring fracture through the continental lithosphere with enclaves providing a remarkable record of periods of crystallization in sub-crustal and crustal magma chambers. There is some comparison between the feeder-dyke system which fed the Muskox layered intrusion in Labrador and the Meugueur-Meugueur ring dyke, although the layered gabbro-anorthosite suites noted at the present erosional level in Air are of much smaller dimensions than their Precambrian counterparts elsewhere. The first stage of deep-level mantle crystallization was followed by the formation of anorthositic cumulates in a crustal magma chamber. A complicated fractionation process is envisaged with sedimentation of ferromagnesian minerals at the base of the magma chamber and plagioclase near the roof. This model follows that proposed by Morse (1979a, b) for the Kiglapait layered intrusion from Labrador, and at the same time helps to explain the enrichment of Fe in the magma and the presence of anorthositic enclaves in the Iskou complex and the Meugueur-Meugueur ring dyke. Leger (1980, 1985) then finally considered the formation of the sequence of leucogabbro-monzogabbros and ferrosyenites as the third stage occurring close to the level of .1
\
Ac
:: ~4
.,'----o---*----I\
....___-,--\
______ ...i,,~., "\""~ ~"5.
~
,7
og
=
-- = ~
- ~ F:a
FIG. 14. Compositional variations in olivines and pyroxenes from the Iskou complex, Niger: -/t, leuconorite; V, leucotroetolite; A, anorthosite; II, basic and [-], intermediate rocks; 0, olivine-pyroxene syenite; <~, amphibole-aegirine syenite and granite; O amphibole-biotite syenite and granosyenite; Sk, Skaergaard trend ; Bu, Bushveld trend; K, Klokken trend (Parsons 1979). (After Leger 1985.)
369
Niger-Nigerian alkaline ring complexes
TABLE 6. Chemical analyses and CIP W norms of
TABLE 5. Analyses and CIP W norms o f the
monzoanorthosite and leucogabbro from Taguei, A~r, Niger
composite Meugueur-Meugueur ring dyke, A~r, Niger wt. ~
Picrite
Gabbro
Syenite Alkaline granite
wt.~
SiO2 TiO2 A1203 Fe203 FeO MnO MgO CaO
37.87 2.50 6.34 1.79 21.33 0.28 22.40 3.11 1.44 0.44 0.32 0.91
45.89 1.16 20.45 0.82 9.03 0.12 8.27 8.76 3.24 0.31 0.23 0.04
58.18 1.34 15.15 3.62 6.53 0.16 1.54 4.55 4.21 2.15 0.36 0.71
73.21 0.17 13.42 0.45 2.76 0.06 0.42 0.87 2.99 5.17 0.03 0.49
-2.60 7.83 9.54 2.36 -3.10 -64.32 2.60 4.75 0.75
-1.83 24.16 40.34 1.76 -1.32 -24.95 1.17 2.20 0.54
12.10 12.70 35.62 16.09 --3.49 9.16 -5.25 2.54 0.85
31.57 30.55 25.30 4.12 -1.40 -5.58 -0.65 0.32 0.07
Na20 K20 P205 LOI Q
or ab an ne C
di hy ol mt il ap
e m p l a c e m e n t a n d a i d e d by m e c h a n i c a l segregation due to the B a g n o l d effect in the c o n d u i t o f rising m a g m a f r o m the crustal m a g m a c h a m b e r to its h i g h level of e m p l a c e m e n t . T h e r e r e m a i n s , h o w e v e r , the d i l e m m a of t h e link b e t w e e n the g a b b r o - a n o r t h o s i t e suite a n d the s y e n i t e - g r a n i t e c o m p o n e n t s o f all the ring c o m p l e x e s in the Air. A possible critical rock type d e s c r i b e d by M o r e a u (1982) as m o n z o a n o r thosite f r o m T a g u e i a n d O f o u d m a y hold the key to the m a g m a genesis. T a g u e i (Fig. 9) is the smallest of the Air a n o r t h o s i t i c c o m p l e x e s ; it is a l m o s t perfectly circular in p l a n a n d is a p p r o x i m a t e l y 700 m in d i a m e t e r . It consists o f a central p l a i n o f l e u c o g a b b r o , a n o r t h o s i t e a n d a central zone o f m o n z o a n o r t h o s i t e 550 m in d i a m e t e r enclosed by an external ridge of alkaline g r a n i t e in c o n t a c t w i t h the P r e c a m b r i a n T u a r e g shield. Majore l e m e n t analyses o f t h e a n o r t h o s i t i c varieties at T a g u e i are r e c o r d e d in T a b l e 6. T h e m o n z o a n o r thosite at T a g u e i c o n t a i n s platelets o f plagioclase o f t e n strongly z o n e d f r o m An60 to rims of An~o s e p a r a t e d by interstitial s y m p l e c t i c crystallization o f q u a r t z - o r t h o c l a s e i n t e r g r o w t h s . T h e plagioclase grains in direct c o n t a c t w i t h interstitial felsic i n t e r g r o w t h s h a v e c o r r o d e d a n d e m b a y e d
78 T a g l
76 T a g l
77 Tag 18
SiO2 TiO2 A120 3 Fe203 FeO MnO MgO CaO Na20 K20 P205 LOI
54.72 1.13 21.77 0.95 2.96 0.05 1.11 8.53 4.05 1.55 0.29 1.42
54.48 1.26 21.83 1.40 3.69 0.07 1.36 7.85 4.00 1.75 0.28 1.52
47.49 1.67 21.14 1.33 8.07 0.13 5.73 9.49 3.37 0.46 0.41 0.12
Qz or ab an
5.62 9.16 34.27 36.64
4.93 10.34 33.85 36.44
-2.72 28.52 41.20
?/e
-
-
-
-
-
-
C
-
-
-
-
-
-
di hy ol mt il ap
3.13 4.09
0.56 6.77
1.38 2.15 0.68
2.03 1.39 0.66
2.63 1.67 16.41 1.93 3.17 0.97
78 Tag l, 76 Tag 1, monzoanorthosite' 77 Tag 18, leucogabbro. TABLE 7. Modal analyses o f monzoanorthosite and
leucogabbro from Taguei, A~r, Niger vol.~ Quartz Alkali feldspar Plagioclase Clinopyroxene Olivine Amphibole Biotite Opaques Others An (%)
78 T a g l
76 T a g l
77 T a g l 8
4.4 13.3 75.6 2.0 Trace 1.3 0.4 0.9 2.1 Zoned 60-20
3.5 16.3 73.3 --3.4 -2 2.1 Zoned 56-20
--87.8 3.1 7.0 -0.6 1 0.5 Unzoned 60
78 Tag l, 76 Tag 1, monzoanorthosite; 77 Tag 18, leucogabbro. borders w i t h A n c o m p o s i t i o n s t o w a r d s oligoclase. H o r n b l e n d e as well as s e c o n d a r y a n d accessory m i n e r a l s such as epidote, chlorite, s p h e n e a n d several g e n e r a t i o n s o f zircon are f o u n d as m i n o r a c c o m p a n y i n g constituents, b u t olivine is conspicuously a b s e n t as a m a j o r m i n e r a l (Table 7). F u r t h e r m o r e , M o r e a u (1982) has n o t e d t h a t there
370
P. Bowden et al.
is an increase in the amount of quartz-orthoclase intergrowth in the monzoanorthosite towards the centre of Taguei. Various alternatives have been proposed to account for the acidic residuum including liquid immiscibility, granite liquid permeation and residual products of gabbroanorthosite crystallization. The significance of monzoanorthosites in the petrogenetic evolution of the Air alkaline province must await further research.
The effect of late-stage fluids on anorogenic magmatism Introduction In the previous section the tectonic and structural controls of anorogenic magmatism were discussed, utilizing examples from Niger. However, the mineralogical assemblage of the complexes is often a result of reactions with residual fluids. Fluids affect the late-magmatic and particularly the post-magmatic (sub-solidus) crystallization history of the cooling sub-volcanic pluton and to some extent the overlying volcanic pile. With continued convection and scavenging from the surrounding country rocks, some of the alkaline ring complexes have become extensively mineralized in a wide spectrum of ore minerals which bear some similarities to economic concentrations in carbonatites. Such fluid reactions, whilst recorded in the Niger complexes (Perez & Rocci 1985), have been well documented in the biotite granites of the Nigerian province (Kinnaird 1985; Kinnaird et al. 1985). Biotite granites and their mineralization Biotite granite (Table 8) and its sub-solidus textural variants (Table 9) are the most abundant rock types in the Nigerian anorogenic province. Structurally, biotite granite can occupy ring-dyke fractures but is generally found as circular plutons, cupolas and sheets. The majority of the biotite granites are mildly corundum normative, indicating alumina oversaturation. The peraluruinous character was not inherited from magmatic processes of crystal fractionation but is a post-magmatic effect of substantial rock-fluid interaction (Martin & Bowden 1981). The excess A 1 is partly linked with Li to form a series of trioctahedral micas varying from annite through siderophyllite to zinnwaldite, and in the latestage development of topaz. Trace-element signatures clearly show the strong hydrothermal overprinting which has taken place in the
constitution and formation of biotite granites (Fig. 15). Not only does the fluid autometasomatize its host but some fluid escapes, partly as a vapour phase, to react with and modify the adjacent and overlying consanguineous rock series as well as the surrounding Precambrian basement. Recently Ike et al. (1985) have clearly demonstrated that hydrothermal alteration of porphyries in the Tibchi anorogenic ring complex (Ike 1983; Ike et al. 1984) is directly related to emplacement and subsequent hydrothermal evolution of the co-genetic biotite granite. Subsolidus reactions are recorded in a plug of quartz porphyry by the transformation of the primary assemblage of hedenbergite, olivine and fayalite to sodi-calcic amphibole compositions. These transformations are progressive and related to the relative distance the fluids have penetrated from the biotite granite cupola. Within the biotite granite itself, the compositions of the Fe-Ti oxides can be correlated with various stages of metasomatism. In most samples from the metasomatic aureole the mafic mineralogy in the quartz porphyry can be correlated with distance from the biotite granite contact. The detailed chemical and mineralogical changes of the sub-solidus variants of biotite granite in the Nigerian province can be considered in terms of sodic metasomatism, potassic metasomatism and acid metasomatism, and can be displayed diagrammatically by distinctive trends in the cationic Q-F diagram (Fig. 16). Table 10 defines some of the mineralogical criteria used to identify the metasomatic overprints on an original magmatic assemblage. Sodic metasomatism The actual mineral assemblages which form in response to sodic metasomatism depend on the intensity of rock-fluid interaction. Such reactions can be variously described as Na-silicate alteration, albitization or Na-for-K exchange. In the cationic diagram of Fig. 16 sodic metasomatism deflects the trends towards the AF join. This is well demonstrated by trend V and initially by trend X. Minerals characteristic of sodic metasomatism include albite, and mica compositions ranging from lepidolite and cryophyllite to zinnwaldite. When sodic metasomatism is intense and the dominant process, the trend of data points is directed towards the albite pole A. This is particularly well demonstrated for the albite granites (Bowden & Kinnaird 1984). The mineralogy changes from a biotite perthite granite to a zinnwaldite albite granite with columbite or an arfvedsonite albite granite with pyrochlore, where fluid alkalinity is higher.
Niger-Niger&n alkaline ring complexes
37
5.Z 0
,.lzl
.~
Z Z d
"13
~,.~,
r
z9 ~
< ,,~. ~. o. ~. e.:. o. ~. o. ~. t--:. o. ~. -.z. o " ,,~
z od rll
.<
o ~ o
~ 1 7 6 1 7 6o ~
~ -~ z <~
372
P. Bowden et al.
E
9-..~
"~
~
~
.2. .~
~
E
g.
#
.
#
i!
"6
,,
II
>-
,_ cr~
"I-
o
._
~
.
~3
.
3
I
I
,,,
E
>-
.-:
f/.-
i
I
Z
ell
"~,
I
~
,I
l
I
D
N
|
o
!
~
o],!Jpuoqo
I
~ool::J
~2
Niger-Nigerian alkaline ring complexes
373
TABLE 9. Major- and trace-element analyses of some Nigerian greisens (gr) and their mieroeline-rieh
borders (wr) B36gr
B64gr
B91gr
B95gr DUT32wr
F G l l g r RN58Awr
RN58Bgr
Major elements (wt.%) SiO2 TiO2 Al203 FezO3 FeO MnO MgO CaO Na20 K20 F 205 HzOt F O~F
76.30 0.10 ll.82 8.77 0.00 0.10 0.02 0.43 0.09 2.43 0.00 0.60 0.00 0.00
81.00 0.30 10.59 5.85 0.00 0.13 0.01 0.33 0.05 1.30 0.00 0.90 0.00 0.00
74.00 0.30 16.98 6.80 0.00 0.17 0.11 0.33 0.05 1.36 0.00 1.34 0.00 0.00
61.00 0.10 17.3l 1,26 8.85 0.13 0.17 2.60 0.05 2.78 0.12 0.78 0.00 0.00
77.80 0.20 9.30 7.89 0.00 0.06 0.01 0.02 2.43 3.78 0.00 0.02 0.00 0.00
77.80 0.10 12.18 8.75 0.00 0.1l 0.01 0.21 0.03 2.13 0.01 0.58 0.00 0.00
76.20 0.08 10.89 1.96 0.00 0.06 0.0l 0.05 0.01 9.76 0.06 0.37 0.17 -0.07
84.40 0.17 4.28 7.71 0.00 0.18 0.01 0.18 0.01 1.58 0.08 1.20 0.77 -0.32
Total
100.66
100.46
101.44
95.15
101.51
101.91
99.55
100.25
A/CNK Q F
3.32 363.15 41.02
5.07 415.66 20.11
7.89 375.61 21.38
2.22 246.44 1t.06
1.15 272.19 1.51
4.45 382.38 40.51
1.02 214.07 206.00
2.09 431.63 30.01
624 8 0
316 2 3
624 3 16
10 15 0
309 1 0
524 16 164 103 116 850 5 41 97 4 311 93
648 62 24 72 14 81 315 20 34 2 92 19
1244 114 18 37 6 550 329 0 7 0 208 35
425 7 491 6820 359
670 1 74 84 134 62 1 91 109 6 64 39
340 3 0 21 292 1203 0 43 238 170 160 1 50 98 11 182 66
1512 5 0 49 2436 682 1 125 171 87 350 0 85 159 11 717 158
Trace elements (pprn) Li Be V Cu Zn Rb Sr Y Zr Nb Sn Ba La Ce Hf Pb Th
532 2 0 2 194 905 18 126 169 93 210 30 35 84 8 87 47
0 255 644 162 84 88
B, Banke; RN, Ririwai; DUT, Dutse; FG Fagarn.
In thin section zinnwaldite albite granite d e m o n s t r a t e s the characteristic 'snowball' texture of albite aplogranites with a b u n d a n t laths of ordered albite surrounding and enclosed by larger a n h e d r a l crystals of u n t w i n n e d turbid intermediate microcline and u n s t r a i n e d quartz. T o p a z is a c o m m o n accessory m i n e r a l with small localized areas of z i n n w a l d i t e - t o p a z - q u a r t z greisen developing at the expense of albite and microcline. C o l u m b i t e and Th-rich monazite associated with m i c a aggregates are occasionally observed in thin section. R e m n a n t cores of more annitic and siderophyllitic m i c a can be o v e r g r o w n or replaced by zinnwaldite.
Potassic metasomatism Biotite granites particularly from Ririwai, have been considerably potash m e t a s o m a t i z e d in the cupola roof zones by boiling of residual fluids. Minerals characteristic ofpotassic m e t a s o m a t i s m include m i c a compositions ranging from original annite to siderophyllite and the g e n e r a t i o n of i n t e r m e d i a t e to o r d e r e d microcline. This ordering is a c c o m p a n i e d by the release and deposition of microscopic plates of h e m a t i t e giving the alkali feldspar a distinctive p i n k to red tinge. Some a m p h i b o l e compositions like ferroactinolite m a y also be valuable petrological criteria. Such min-
P. B o w d e n et al.
374
TABLE 10. Mineralogical criteria for post-magmatic metasomation in alkaline granite ring complexes Sodic metasomatism Na-for-K exchange 1 Albitization 2 Sodium silicate alteration 3 Ordered albite 4 Protolithionite to trilithionite 'cryophyllite '5 Riebeckite, lithian arfvedsonite 6 Aegirine-neptunite ~ Cryolite, villiaumite s
Potassic metasomatism
Acid metasomatism
K-for-Na exchange H§ exchange Microclinization ) Greisenization Potassium silicate alteration Intermediate to ordered microcline Feldspar breakdown to topaz/sericite Ferrous siderophyllite Lithian siderophyllite Ferroedenite, ferroactinolite Chlorite Fluorite/topaz
1, Cationic exchange reactions recorded in perthitic feldspar; 2, 3, generalized terminology; 4, structural state of alkali feldspar; 5, tri-octahedral mica compositions; 6, amphibole compositions or equivalent; 7, pyroxene compositional range; 8, fluorine-rich mineral assemblages.
eral reactions can be described as K-silicate alteration (Rose & Burt 1979), microclinization and K-for-Na exchange. More than one period of potassic metasomatism has been recognized, depending on the mineralogical variants observed in the sub-solidus assemblages. This aspect has been discussed more fully by Kinnaird et al. (1985). Acid metasomatism
Acid metasomatism in Nigeria has affected three distinctive mineralogical assemblages; perthitic granite, albitized variants and microclinized variants. The chemical composition of these variations, superimposed on either sodic metasomatism or potash metasomatism, are shown in the Q-F diagram (Fig. 16) by trends X, Y and Z towards the silica pole. Acid metasomatism reactions are concerned with the instability of granitic minerals in the presence of HF-rich fluids and the growth of new sub-solidus assemblages (Kinnaird 1985). Mineralogical and chemical reactions include the breakdown of perthitic alkali feldspar to sericite and topaz, the destabilization of albite to form fluorite, cryolite and topaz with some montmorillonite, and the destruction of microcline to form micaceous aggregates, zones of chlorite and, more rarely, kaolinite. Mica compositions generated by acid metasomatism of perthitic granite show compositional variations from siderophyllite, a characteristic green mica, which appears initially as overgrowths to more annitic compositions and ultimately as independent crystals of more uniform optical properties. Other mica compositions include protolithionite, zinnwaldite and lepidolite associated with acid metasomatism of albite roof rocks, and lithian siderophyllite generated from microcline-rich facies. Some of these reac-
tions released silica to the formation of quartz veins. All the petrological changes discussed above have been referred to as 'greisenization' in the literature and are regarded as a series of complex pneumatolytic-hydrothermal processes (Shcherba 1970). Greisenization can be found superimposed on various earlier mineral assemblages. The geochemical characteristics and the extent of greisen development depend upon the intensity and retention of earlier stages of albitization and/or microclinization and the timing of fracturing and fissuring in the roof zone. All the major-and trace-element data collated under greisens (Table 9 ) a n d wall-rock assemblages show many distinctive geochemical features (Kinnaird 1985). Some of these features are indicated as trace-element signatures in Fig. 15, and are displayed as major-element variations in Fig. 16. Trends X, Y and Z show clearly the differences in compositional variations of bulkrock chemistry and hydrothermal fluid related to ore deposition. These variations are attributed to varying degrees of rock-fluid interaction whose chemistry was controlled by the parental granite from which the fluid phase exsolved and by the chemical composition of the enclosing rocks with which exchange took place. The implications of this diagram specifically for metasomatic reactions recorded in the Ririwai complex are discussed in detail by Kinnaird et al. (1985), with the ore minerals generated during greisenization tabulated in Fig. 17. There are suites of greisens, particularly those from Banke, Fagam and Jos-Bukuru in Nigeria, and Tarouadji, Guissat and El Meki in Niger, where greisenization is pervasive. This occurs associated with sericite, topaz and lithian siderophyllite, and is defined as a trend to the left centre on the Q-F diagram in terms of the cationic
N i g e r - N i g e r i a n alkaline ring c o m p l e x e s
375
Qu
Oredeposition \ in marginalveins\ BANKE:FAGAM:/ JOS: KUDARU//
Major ore deposition of sphalerite- cassiterite in lodes stockworks RIRIWAT: RISHI: AFU: TIBCHI (2)
/
Oredeposit~ TIBCHI{I) granite
O//~redeposition / GINDIAKWAT
% %
~.__", A
F
I
MICA
M
FIG. 16, Summary of major trends of metasomatism, greisenization and ore deposition in the Nigerian anorogenic ring complexes. The triangular apices Qu (quartz), A (albite) and M (microcline) are delineated as a sketch diagram based on Q-F parameters (Charoy 1979). Trends V, X, Y and Z are described in the text. The rectangular closed box labelled mica is lithian siderophyllite. The line connecting the mica and quartz poles represents the zone along which greisens develop in the Nigerian anorogenic province. The broken line connecting the quartz pole and the open square is the zone of greisen formation for calc-alkaline granitoids. (After Charoy 1979.)
proportions of quartz, albite and microcline. The trend is similar to that presented by Charoy (1979) for St. Michael's Mount and Cligga Head, Cornwall, and the Blue Tier batholith, Australia. In the Nigerian samples showing pervasive greisenization, boiling of the fluid phase was retarded until a late stage. Such features can be established by field evidence of brecciated greisens at Banke (Bennett et al. 1984), and by fluidinclusion studies (Kinnaird 1985). Furthermore, substantial ore deposition only occurs on boiling of mineralizing fluids. A completely separate trend is defined for the samples from Ririwai (Kinnaird et al. 1985). They lie close to and along the quartz-mica join after following an initial path towards the microcline pole. The greisen veins and wall-rock assemblages occur as stockworks or lodes with central-textured quartz bounded by limited greisen zones and microcline-richborders containing abundant sphalerite and cassiterite. Fluid-inclusion studies (Kinnaird et al. 1985) indicate a dominantly boiling assemblage for the quartz centres of the lode system. There are therefore
two distinct metasomatic trends related to mineralization which can be expressed on the Q-F diagram. The first is pervasive greisenization with late retarded boiling in the roof zones of granite cupolas forming only marginal veins. Little development of potash metasomatism has occurred but the dominant processes are sodic metasomatism followed by acid metasomatism. In contrast, the second trend is shown by major lode systems and stockworks, often extensively mineralized, in which boiling of the mineralizing fluid has had a major influence on the metasomatic sequences, and by the development of important economic sources of ore metals. It is only in biotite granite intrusions with shallow outward-dipping contacts with the centre's own volcanic cover or the Precambrian basement that the fluids can be retained in the roof zone and intensively interact. The endocontact zone of the granite cupola and the exogenic region within the volcanic cover or Precambrian country rocks become potential sites for substantial rock-fluid interaction. Initially the reactions are widespread and pervasive but, as the solidifying granite cools
376
P . B o w d e n e t al. Greisen
Monazite Zircon Ilmenite Cassiterite
valuable petrological criteria for assessing the ore-bearing potential.
Quartz Velh
.
Conclusions
Wolframite Columbite TiO2(Rutile) Molybdenite Sphalerite Stannite Pyrite Marcasite
I
.
m
Chalcopyrite Cubanite Pyrrhotite Mackinawite Bismuth Bismuthinite Galena Hematite Chalcocite Covellite Blaubleibender Covellite
m
m
n
FIG. 17. Ore mineral assemblages identified at the Ririwai complex in the lode system of the biotite granite. (Compiled by Ixer in Kinnaird et al. 1985.)
and contracts, a series of microfractures and later joints provide more local access and channelways within the intrusive body for circulating fluids. According to Taylor et al. (1980) both metaluminous and peralkaline granites exsolve a peralkaline fluid phase from a residual silicate melt at shallow depth. It is the behaviour of this fluid phase in a low-pressure environment that dictates the formation of the new sub-solidus assemblages and the ore deposits in Nigeria associated with the younger granites. From the mineralization viewpoint there are many parallels with alkalineundersaturated rocks and carbonatites, particularly with the abundance of sphalerite, rare-earth minerals, zircon and complex titanium silicates, uranium, thorium, columbite and pyrochlore. In addition to the distinctive alkaline mineralization suite, there is a more normal group of ore minerals such as cassiterite, wolframite, chalcopyrite and galena. In general all granitic structures developed at high levels in the continental crust, as roots of volcanoes, are potential hosts for metasomatic reactions by residual fluids. The reactions are recorded as new sub-solidus mineral assemblages with trapped fluid inclusions, providing
Although the source region for the parental magma of the Niger-Nigerian anorogenic province lies in the asthenosphere, it was the exploitation and reactivation of Pan-African shear zones and transcurrent faults during the fragmentation of Gondwanaland that controlled the locations of Phanerozoic intra-plate magmatism in W Africa. The occurrence of leucogabbros, anorthosites and monzoanorthosites as part of the Palaeozoic sub-volcanic association with syenites, peralkaline granites and biotite granites in Niger provides critical evidence for assessing the magmatic evolution of the Niger-Nigerian anorogenic province. Distinctive mineralogical and chemical changes induced by post-magmatic processes are superimposed on the magmatic paragenesis. Such processes are essentially the result of fluid interaction with the original magmatic assemblage of minerals which re-equilibrated with their residual fluids. These fluids are magmatichydrothermal in origin but their composition is modified by exchange reactions, boiling and selective loss of gaseous components. Therefore the distinctive petrology and geochemistry of the Niger-Nigerian ring complexes reflect not only their magmatic evolution but also the importance of post-magmatic metasomatism. ACKNOWLEDGEMENTS: Part of the work summarized in
this review was financed by the Overseas Development Administration, Grant R2679. The authors (PB, ECI, JAK and RAB) would like to acknowledge the longterm support provided by the British and Nigerian governments for continued research at St. Andrews, and to their colleagues and fellow members of the academic and technical staff for assistance. ECI is also grateful to the Commonwealth Foundation for a research scholarship for the period 1975-1978 and to Professor C. A. Kogbe for leave of absence from teaching duties. PB acknowledges the continued interest by French colleagues in the Air anorthosites. Particular thanks are due to J.-P. Karche, C. Moreau, (with J. Husch), J.-M. Leger and B. Mai-Manga for allowing us to incorporate some of the results of their research into this review. RFM is grateful to the National Science Foundation of Canada for research funds to visit Niger and Nigeria, and for continued support in Canada for monitoring fluid reactions by mineralogical observations on sub-solidus granite assemblages. Many research workers on the Niger-Nigerian
Niger-Nigerian alkaline ring complexes anorogenic province have contributed to the data and interpretations discussed in the paper. We would particularly like to thank J. Lameyre, A. Giret,
377
B. Bonin, D. C. Turner, J. N. Bennett, C. A. Abernethy, A. B. Moyes, D. A. Ajakaiye, C. A. Kogbe, S. I. Abaa and U. M. Turaki.
References BAKOR, A. R., GASS, I. G. & NEARY, C. R. 1976. Jabel al Wask; northwest Saudi Arabia: an Eocambrian back arc ophiolite. Earth planet. Sci. Lett. 300, 19. BAYER, R. & LESQUER, A. 1978. Interpr&ation des anomalies gravim&riques de la bordure orientale du craton ouest-africain: geom&rie d'une suture. Bull. Soc. geol. Fr. 20, 863-76. BENKHELIL, J. 1982. Benue trough and Benue chain. Geol. Mag. 119, 155-68. BENNETT, J. N., TURNER, n . C., IKE, E. C. & BOWDEN, P. 1984. Geology of the northern Nigerian ring complexes. Overseas Geol. miner. Resources 61. BLACK, R. 1963. Note sur les complexes annulaires de Tchouni-Zarniski et de Goure. Bull. Bur. Rech. gdol min. (Fr) 1, 31-45. 1965. Sur la signification p6trog6n6tique de la d6couverte d'anorthosites associ6es aus complexes annulaires sub-volcaniques du Niger. C.R. Acad. Sci. Paris, 260, 5829-32. 1984. The Pan-African event in the geological framework of Africa. Pangea, 2, 6-16. , BONIN, B. & LAMEYRE, J. 1985. The structural setting of alkaline ring complexes. J. Aft. Earth Sci. 3, 5-16. , JAUJOU, M. • PELLATON, C. 1967. Notice Explicative de la Carte Gdologique de l'Air ~ l'Echelle 1/500,000. Directorate de Geologie et Mines, Niger. BOWDEN, KARCHE, J-P. 1984. Mid-plate A-type magmatism in the Niger-Nigeria anorogenic province: age variations and implications. In: KLERKX, J. & MICHOT, J. (eds) African Geology, pp. 167-77. Mus6e Royal de l'Afrique Centrale, Tervuren. & KINNAIRD, J. A. 1984. Petrological criteria for the identification of potential ore-bearing Nigerian alkaline granitoids, Proc. Int. Geological Congr., Moscow, Vol. 1, pp. 85-119. V N U Science Press, Utrecht. & TURNER, D. C. 1974. Peralkaline and associated ring complexes in the Nigeria-Niger province West Africa. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp. 350-51. Wiley, New York. , BENNETT, J. N., KINNAIRD, J. A., WHITLEY, J. E., ABAA, S. I. & HADZIGEIRGIOU-STAVRAKIS,P. K. 1981. Uranium in the Niger-Nigerian younger granite province. Mineral. Mag. 44, 379-89. , VAN BREEMEN, O., HUTCHINSON, J. & TURNER, D. C. 1976. Palaeozoic and Mesozoic age trends -
-
-
-
P.
-
-
&
for some ring complexes in Niger and Nigeria. Nature, Lond. 259, 297-9. BUDDINGTON, A. F. 1969. Adirondack anorthositic series. In : ISACI-ISEN,I. W. (ed.) Origin of Anorthosite and Related Rocks, N Y State Mus. Sci. Serv. Mem. 18, 215-31. CABY, R. 1970. La Chaine pharusienne dans le NordOuest de l'Ahaggar (Sahara central, Alg6rie); Sa place dans l'orogen6se du Pr6cambrian sup6rieur en Afrique. Thbse d'Etat, Universit6 de Montpellier (unpublished). - - , BERTRAND,J. L. & BLACK, R. 1981. Pan-African closure and continental collision in the HoggarIforas segment, central Sahara, In: KRONER, A. (ed.) Precambrian Plate Tectonics, pp. 407-51. Elsevier, Amsterdam. CHAROY, B. 1979. D6finition et importance des ph6nom+nes deut6riques et des fluides associ6s dans les granites. Cons6quences m&allog+niques. Mem. Sci. Terre, 37. DEWAARD, D., DUCHESNE, J. C. & MICHOT, J. 1974. Anorthosites and their environment. In: Centenaire Soc. Geol. Belg., Gdologie des Domaines Cristallins, Li6ge, pp. 323-46. DUCHESNE, J. C., ROELANDTS, I., DEMAIFFE, D., HERTOGEN, J., GIJBELS, R. & DE WINTER, J. 1974. Rare-earth data on monzonoritic rocks related to anorthosites and their bearing on the nature of the parental magma of the anorthositic series. Earth planet. Sci. Lett. 24, 325-35. EMSLIE, R. F. 1973. Some chemical characteristics of anorthositic suites and their significance. Can. J. Earth Sci. 10, 54-71. 1978. Anorthosite massifs, rapakivi granites and late Proterozoic rifting of north America. Precambrian Res. 7, 61-98. GARSON, M. S. & SHALABY,Ii M. 1976. PrecambrianLower Palaeozoic plate tectonics and metallogenesis in the Red Sea region. Geol. Assoc. Can. spec. Pap. 14, 573-96. GREEN, D. H. & RINGWOOD, A. E. 1976. The genesis of basaltic magma. Contrib. Mineral. Petrol. 15, I 03-90. GREEN, T. H. 1969. High-pressure experimental studies on the origin of anorthosite. Can. J. Earth Sci. 6, 427-40. GREENWOOD, W. R., ANDERSON, R. E., FLECK, R. J. & SCHMIDT, D. L. 1976. Late Proterozoic cratonization in southwestern Saudi Arabia. Phil. Trans. R. Soc. Lond., Ser. A, 280, 517-27. -
-
378
P. B o w d e n et al.
GUIRAUD, R., ISSAWI, B. & BELLION, Y. 1985. Les lin6aments guin6o-nubiens: un trait structural majeur ~i l'6chelle de la plaque africaine. C.R. Acad. Sci. Paris, 300, 17-20. HOSSAIN, M. T. & TURAKI, U. M. 1983. Bibliography on the younger granite ring complexes and tin mineralization with emphasis on Nigeria. J. Afr. Earth Sci. 1, 73-81. HUSCH, J. M. & Moreau, C. 1982. Geology and major element geochemistry of anorthositic rocks associated with Palaeozoic hypabyssal ring complexes, Air Massif, Niger, West Africa. J. Volcanol. Geotherm. Res. 14, 47-66. IKE, E. C. 1983. The structural evolution of the Tibchi ring-complex: a case study for the Nigerian younger granite province. J. geol. Soc. Lond. 140, 781-8. , BOWDEN, P. & MARTIN, R. F. 1984. Fayalite and clinopyroxene in the porphyries of the Tibchi anorogenic ring-complex, Nigeria: postmagmatic initiation of a peralkaline trend. Can. Mineral. 22, 401-9. -- & -1985. Amphiboles in the porphyries 'of the Tibchi anorogenic ring-complex, Nigeria: product of deuteric adjustments. Can. Mineral. 23, 447-56. KARCHE, J.-P- & MOREAU, C. 1977. Note pr~liminaire sur le massif subvolcanique ~ structure annulaire d'Abontorok (Air-Niger) C.R. Acad. Sci. Paris, Sdr D, 248, 1259-62. & -1978. Sur quelques gites de fer titane associ6s aux anorthosites de l'Ofoud (Air centralNiger). Annls Univ. Niamey, 1, 121-5. -& VACHETTE, M. 1976. Migration des complexes subvolcaniques ~t structure annulaire du Niger: cons6quences. C.R. Acad. Sci. Paris, 282, 2033-6. KINNAIRD,J. A. 1981. Geology of the Nigerian Anorogenic Ring Complexes. Map Scale 1:500,000. John Bartholomew, Edinburgh. 1985. Hydrothermal alteration and mineralization of the alkaline anorogenic ring complexes of Nigeria. J. Afr. Earth Sci. 3, 229-51. , BOWDEN, P., IXER, R. A. & ODLING, N. 1985. Mineralogy, geochemistry and mineralization of the Ririwai complex, Northern Nigeria. J. Aft. Earth Sci. 3, 185-222. LEBLANC, M. 1981. The late Proterozoic ophiolites of Bou Azzer (Morocco): evidence for Pan-African plate tectonics. In: KRONER, A. (ed.) Precambrian Plate Tectonics, pp. 435-51. Elsevier, Amsterdam. LEGER, J-M. 1980. Evolution p~trologique des magmas basiques et alkalins dans le complexe anorog~nique d'Iskou (Air-Niger). Thbse 3brae Cycle, Universit~ de Paris 6 (unpublished). 1985. G6ologie et bvolution magmatique du complexe plutonique d' Iskou (Air, Niger). J. Aft. Earth Sci. 3, 89-96. LIEGEOIS, J-P. & BLACK, R. 1984. P6trographie et g6ochronologie Rb-Sr de la transition calcoalcaline fini-panafricaine dans l'Adrar des Iforas (Mali): accr6tion crustale au Pr6cambrien sup6rieur. In: KLERKX, J. & MICHOT, J. (eds) African Geology, pp. 115-45. Mus6e Royal de l'Afrique Centrale, Turvuren.
-
-
-
-
-
-
MARTIN, R. F. & BOWDEN, P. 1981. Peraluminous granites produced by rock-fluid interaction in the Ririwai non-orogenic ring complex, Nigeria: mineralogical evidence. Can. Miner. 19, 65• MIGNON, R. 1970. Etude g6ologique et prospection du Damagaram Mounio et du Sud Maradi. Rap. Bur. Rech. Geol. Miner. Mines Geol., Niamey. MOREAU, C. 1982. Les complexes annulaires anorog6niques h suite anorthosique de FAir Central et septentrional (Niger). Thbse d'Etat, Universit6 de Nancy I (unpublished). - - ~ , KARCHE,J-P & TRICHET,J. 1978. Remarques sur les anorthosites des complexes sub-volcaniques de l'Air (Niger). C. R. Somm. Sdanc. Soc. Geol. Fr. 1, 21-3. MOREL, A. & MOREAU, C. 1979. Le mod/~le et la p6trologie des anorthosites et roches associ6es de l'Ofoud (Air-Niger). Rev. Geogr. phys. Geol. dynam. 21,247-55. MORSE, S. A. 1979a. Kiglapait geochemistry I: systematics sampling and density. J. Petrol. 20, 555-90. 1979b. Kiglapait geochemistry II: petrography. J. Petrol. 20, 591-624. PARSONS, I. 1979. The Klokken gabbro-syenite complex, South Greenland: cryptic variation and origin of inversely graded layering. J. Petrol. 20, 653-94. PEREZ, J.-B. & ROCCI, G. 1985. Fluid interaction during the crystallisation of granites in the Taghouaji ring complex (Air, Niger): textures, paragenesis of rock-forming minerals and accessory minerals. Abstr. 13th Colloq. on African Geology, St. Andrews (CIFEG), pp. 304-5. Geol. Dept. St. Andrews Univ. PHILPOTTS, A. R. 1966. Origin of the anorthositemangerite rocks in southern Quebec. J. Petrol. 7, -
-
1 - 6 4 .
RAHAMAN, M. A., VAN BREEMEN, O., BOWDEN, P. & BENNETT,J. N. 1984. Age migrationsofanorogenic ring complexes in northern Nigeria. J. Geol. 92, 173-84. ROSE, A. W. & BURT, D. M. 1979. Hydrothermal alteration. In: BARNES,H. L. (ed.) Geochemistry of Hydrothermal Ore Deposits, pp. 173-235. Wiley, New York. RYDER, G. 1974. A rationale for the origins of massif anorthosites. Lithos, 7, 139-46. SHCHERBA, G. N. 1970. Greisens. Int. Geol. Rev. 12, 114-50, 239-50. SIMMONS,E. C. & HANSON,G. H. 1978. Geochemistry and origin of massif-type anorthosites. Contrib. Mineral. Petrol. 66, 119-35. SYKES, L. 1978. Intraplate seismicity, reactivation of pre-existing zones of weakness, alkaline magmatism, and other tectonism post-dating continental fragmentation. Rev. Geophys. space Phys. 16, 62188. TAYLOR, R. P., STRONG,D. F. & KLAN, B. F. 1980. The Topsails igneous complex: Silurian-Devonian peralkaline magmatism in western Newfoundland. Can. J. Earth Sci. 17, 425-39. TURNER, D. C. & BOWDEN, P. 1979. The Ningi-Burra complex, Nigeria: dissected calderas and migrating magmatic centres. J. geol. Soc. Lond. 136, 105-19.
Niger-Nigerian alkaline ring complexes UMEDJI, A. C. & CAEN-VACHETTE, M. 1983. Rb-Sr isochron from Gboko and Kyuen rhyolites and its implication of the age and evolution of the Benue trough, Nigeria. Geol. Mag. 20, 529-33.
379
VANBREEMEN,O., PIDGEON, R. T. & BOWDEN, P. 1977. Age and isotopic studies of some Pan African granites and related rocks, north-central Nigeria. Precambrian Res. 4, 307-19.
PETER BOWDEN, RICHARDA. BATCHELOR& JUDITH A. KINNAIRD,Department of Geology, University of St. Andrews, Fife KY16 9ST, U.K. RUSSELLBLACK,Laboratoire de P&rologie, Universitb de Paris 6, 75230 Paris, France ROBERT F. MARTIN, Department of Geological Sciences, McGill University, Montreal, Canada ECHEFU C. IKE, Department of Geology, Ahmadu Bello University, Zaria, Nigeria
Alkaline magmatism subsequent to collision in the Pan-African belt of the Adrar des Iforas (Mali) J. P. Li geois and R. Black SUM MARY: The Pan-African Trans-Saharan belt in the Iforas displays a rapid switch from subduction and collision-related calc-alkaline to typical A-type magmatism, which is accompanied by transcurrent movements along major shear zones inducing weak distension. Detailed Rb-Sr geochronology and geochemical data point to different mantle sources for orogenic (lithospheric depleted mantle + oceanic crust) and within-plate magmatism (more primitive asthenospheric mantle). Both groups suffer lower-crustal contamination. A model is proposed whereby asthenospheric mantle originally underlying the subducted plate has risen to shallow depth beneath the continental lithosphere after the rupture of the cold plunging plate. This source, which is often proposed for alkaline rocks, explains the great similarity of oversaturated alkaline ring-complexes whatever their environment. The peculiarities of the alkaline province, for example the lack of Sn mineralization when compared with the Niger-Nigerian province, may be related to the nature and composition of the basement.
Introduction The Cambrian saturated alkaline ring complexes of the Iforas (Ba et al. 1985) mark the end of the Pan-African, characterized here by oceanic closure around 600 Ma ago and oblique collision between the passive margin of the W African craton and the active margin of the Tuareg shield (Black et al. 1979; Caby et al. 1981 ; Fabre et al. 1982; Ball & Caby 1984). This post-tectonic setting contrasts with that of the anorogenic within-plate Niger-Nigeria province where there is a considerable lapse of time (several hundred million years) between the age of the country rocks and that of the alkaline intrusions (Black et al. 1985). The aim of this paper is to outline very briefly the main structural and petrological characteristics of the Iforas alkaline province, to trace chronologically and geochemically the transition from calc-alkaline to alkaline magmatism and finally to propose a geodynamic model which may also fit other collisional terrains displaying similar magmatic sequences, e.g. Pan-African of Saudi Arabia (Duyvermann et al. 1982; Harris 1982), Permian of Corsica (Bonin 1980), Tertiary volcanism of the Turkish-Soviet Armenia-W Iranian Alpine segment (Innocenti et al. 1982) and the Basin and Range province (Eaton 1982).
General geology Whereas the 2000 Ma old W African craton is simply covered by Upper Proterozoic and Phanerozoic fiat-lying sediments, the adjacent Tuareg shield emerging from beneath Ordovician strata
has a complex geological history with diversified and abundant igneous activity between 850 and 530 Ma ago. These features are best explained by a complete Wilson cycle in Pan-African times. After rifting around an RRR triple junction in the Gourma area (Moussine-Pouchkine & Bertrand-Sarfati 1978) followed by oceanic spreading to the E of the W African craton, a long period of subduction (720-620 Ma) is recorded in the Iforas along the SW margin of the Tuareg shield (Ducrot et al. 1979; Caby et al., 1986. The subduction-related magmatism over an easterlydipping Benioff plane appears in two environments: an oceanic trench island arc to the W of the Iforas (Caby 1970, 1981) and a cordilleran assemblage with andesites in the Iforas (Chikhaoui 1981). Oceanic closure and collision occurred 620-590 Ma ago, the suture between the W African craton and the Tuareg shield being marked by an array of strong positive gravity anomalies (Bayer & Lesquer 1978). Collision was not frontal as in the Himalaya but oblique (Ball & Caby 1984), inducing N - S mega-shear zones which sliced the Iforas and absorbed part of the collision. Nappes, comprising elements of the passive continental margin, oceanic basalts, possible ophiolites and an inner unit of eclogitic mica schists, were translated westward onto the W African craton (Caby 1980). This craton behaved as a relatively rigid mass devoid of autochtonous Pan-African magmatism, in contrast with the active margin of the Tuareg shield which was strongly mobilized and invaded by abundant granitoids which form a huge composite late-tectonic calc-alkaline batholith 100150 km to the E of the suture. The calc-alkaline manifestations continued in post-tectonic condi-
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 381-401.
38I
382
J. P. LiOgeois and R. Black
FIG 1. Structural map of the Iforas-Ahnet area (1-5, passive margin of the W African craton; 6-12; active margin of the Iforas) : 1, Eburnean unreactivated basement; 2, autochthonous passive margin (Gourma aulacogen); 3, Timetrine-Gourma nappes; 4, Permian Tessofi graben; 5, Permian Tadhak undersaturated ringcomplexes province; 6, Eburnean granulites (Archaean substratum?); 7, reactivated basement; 8, island arc; 9, cordilleran volcano-sedimentary unit; 10, composite calc-alkaline batholith; 11, oversaturated alkaline ringcomplexes and lavas; 12, Cambrian molasse; 13, transcurrent faults; 14, thrusts; 15, suture zone as indicated by a string of positive gravity anomalies; F2, area of Fig. 2. (After Fabre et al. 1982.) tions with W N W - E S E dyke-swarms and highlevel circular plutons which indicate uplift of the continent subsequent to collision. This uplift led to unroofing the batholith just after the emplacement of the first granite displaying some alkaline affinities (Tahrmert). Renewed intermittent movements along the N-S shear zones and a change in the stress field was accompanied by the injection of spectacular N-S dyke-swarms in the axis of the batholith, feeding extensive plateaux of rhyolites and ignimbrites beneath which were emplaced the alkaline ring-complexes.
The Iforas alkaline province The Cambrian alkaline province is superimposed on the Pan-African composite calc-alkaline batholith of Western Iforas. It comprises remnants of an extrusive cover of rhyolites and ignimbrites of fissural origin (Tiralrar, Ichoualen plateaux), dense N-S-oriented acid dyke-swarms and over 15 plutons displaying a variety of intrusion forms including huge ring-complexes, crescentic sheeted intrusions and sub-circular stocks. Contacts with the country rocks and between succes-
Alkaline magmatism in the lforas province
[]13 F ~ tl ~]~]] ~o
~-~ 7
[]4 [] 3
~2
FIG. 2. Schematic map of the studied Kidal-Tiralrar area: 1, island arc; 2, reactivated basement; 3, volcano-sedimentary sequences; 4, pre-tectonic tonalite; 5, late-tectonic granodiorite; 6, late-tectonic porphyritic monzogranite; 7, late-tectonic fine-grained monzogranite; 8, post-tectonic E-W dyke-swarms; 9, post-tectonic syenogranite; 10, post-tectonic alkaline granite pluton; 11, extrusive alkaline rhyolites and ignimbrites; 12, syenitic ring-dykes; 13, alkaline oversaturated ring-complexes. For clarity, N-S dykes have not been drawn. (See Fig. 3.)
sive phases are sharp and typically show a chilled marginal facies with frequent development of miarolitic cavities. All the complexes have been emplaced at shallow depth in a rigid environment beneath a thick volcanic cover by a process of subterranean cauldron subsidence and major stoping. They are composed of typical A-type quartz syenites and granites (Loiselle & Wones 1979) embracing metatuminous, peralkaline and peraluminous compositions. The province is exclusively oversaturated and quite distinct from the much younger Permian Tadhak province of undersaturated alkaline rocks and carbonatites located on the eastern margin of the W African
383
craton (Fig. 1) (Li~geois et al. 1983; Sauvage & Savard 1985). A general description of the Iforas alkaline province has recently been published (Ba et al. 1985). Only a brief description of the units studied in this paper (Fig. 2) is given here, followed by a summary of the main geochemical and petrological features of the province. The plateau lavas occurring as thick flows of devitrified and often altered rhyolites and ignimbrites have not been studied in detail. Inside the Ichoualen ring structure they are seen to overlie the eroded Tahrmert granite, a large irregularshaped massif characterized by pronounced horizontaljointing which invades the Telabit and Dohendal WNW-ESE dyke-swarms to the N and W of the Timedjelalen ring-complex (Fig. 3). This coarse-grained hypersolvus granite already displays distinct alkaline affinities and, with its clustered rounded quartzes, perthites and abundant accessory minerals (zircon, Fe-Ti oxides, sphene and fluorite), it closely resembles some of the metaluminous biotite granites of the alkaline ring-complexes. The mafic minerals (brown phlogopitic biotite and occasional pale-blue richterite) are magnesian, however, and suggest that this alkaline precursor has a mixed source (Ba et al. 1985). The N-S dykes are mainly composed of devitrified and often altered rhyolites, felsites and quartz-feldspar porphyries (believed to be feeders to the plateau rhyolites), quartz microsyenites often containing basic xenoliths, porphyry granites and granophyres. Both metaluminous and peralkaline varieties are present. Basic dykes are rare and have only been encountered to the N of Tiralrar. Among the ring-complexes, the Kidal massif (Figs 3 and 4) with a diameter of 30 km is the largest and most intricate. It has been segmented and offset by a N N W sinistral transcurrent fault. The early arcuate ramified and polygonal ringdykes of quartz syenite (K 1) are partly obliterated by the plutonic centre which comprises 13 granitic phases whose order of intrusions is indicated in the legend of Fig. 4. A coarse hypersolvus granite (K3), displaying both metaluminous and peralkaline facies with related granite porphyries (K2K3'), occurs as a fiat tabular sheet covering twothirds of the area and constitutes the upper structural level of the complex. It is thought to have intruded permissively a long and subhorizontal tensional roof fracture fed by a peripheral ring-dyke. Granite porphyry (K4) marks the fragmentation of the underlying block and the start of an aluminous cycle with the emplacement of metaluminous hypersolvus granite (K5-K5') followed by subsolvus granite (K6-
J. P. Libgeois and R. B l a c k
384 I
1
IoO0'E
2oO0'E
Tessalit: -20oO0'N
Boughessq
Bezzeg
( ~ Dourit~ I" ~"~Tinebrorhen
',• // ',3
~,Alimamas -19~
I
Timedjelolen ~.t~-H
Sendemen
l,
'~'~\
Dykeswarms
t::::::::l .o,o o.,o.,oe roni,e
Djounhane~-g
Yenchichi
Mylonites, faults
Microsyenites Rhyolites
Tibeljel
i!i!!~:. Tahrmert alkaline granite Takellout 0 I
50 km I
FIG. 3. Geological map of the Iforas alkaline province (Ba et al. 1985). K8) and the late granite porphyry (K9) which defines several centres of subsidence. Magmatic activity is then confined to central and eastern parts of the complex with the intrusion beneath K3 and K7 of a peralkaline medium-grained granite (K10) which in turn is underlain by a
metaluminous coarse-grained hypersolvus granite ( K l l ) and by albitized peralkaline granite (K12) mineralized in Th minerals. The general shape of the massif has been confirmed by a gravity study (Ly et al. 1984) which indicates that the alkaline granites in the western part of the
385
Alkaline magmatism in the Iforas province I
I
i~
~
1~
/ jjt8~
-
--...,d"
KIDAL
0
I~--_1
5krn
AKKISE DI
DE
D3
AI
A2A3
KIO
KII
KIDAL-TIBELJELJ ELINE KI
K2
K3
K3'
K4
K5
K5'
K6
K7
K8
K9
KI2
F[o. 4. Geological map of the Kidal-Tibeljelejeline,Djounhane and Akkise ring-complexes: K 1, syenite porphyry and quartz syenite (Fe-augite, amphibole, biotite, _ fayalite); K2, amphibole-biotite granite porphyry; K3, metaluminous coarse-grained granite (perthite, hedenbergite, amphibole, biotite, +_fayalite) to peralkaline (perthite, aegirine-augite, Ca-Na- and Na-amphiboles, + aenigmatite); K3', peralkaline granite porphyry; K4, metaluminous granite porphyry (K-feldspar, hedenbergite, amphibole, biotite and two feldspars, amphibole, biotite); K5, fine-grained granite (perthite, biotite, chlorite, _+amphibole); K5', medium-grained granite (perthite, amphibole, biotite, +_fayalite); K6, fine-grained granite (perthite, oligoclase, amphibole, biotite); K7, fine- to coarse-grained granite (perthite, oligoclase, amphibole, biotite); K8, fine-grained granite (perthite, oligoclase, biotite, chlorite); K9, granite porphyry (perthite, oligoclase, biotite, chlorite); K 10, metaluminous medium-grained granite (hedenbergite, aegirine-augite, Fe-richterite, arvfedsonite, biotite) to peralkaline (aegirine, arvfedsonite, astrophyllite); K 11, coarse-grained granite (perthite, amphibole, biotite); K 12, finegrained granite (microcline, albite, aegirine, Ca-Na- amphibole and microcline, albite, arfvedsonite); A1 = K2; A2 = K8; A3, granite porphyry (perthite, oligoclase, amphibole, biotite); D1 ---K3; D2 - K7; D3 - K8. (Ba et al. 1985). complex are thin compared with those in the E which have been estimated to have a thickness of 5 km. In contrast, the Timedjelalen ring-complex (Figs 3 and 5), with six concentrically disposed units younging towards the centre, is relatively simple despite its large size (32 km x 22 km). The sequence (see legend to Fig. 5) starts with mildlyperalkaline quartz-poor granite porphyry (T1) and medium-grained hypersolvus granite (T2)
which form the steeply-dipping external ringdyke injected in a multiple-fracture zone. Foundering of a central block permitted the emplacement of a peralkaline coarse-grained hypersolvus granite (T3) to form a tabular sheet corresponding to the upper level of the complex. Renewed subsidence marks a change in the chemistry with the intrusion beneath T3 of a metaluminous hypersolvus granite (T4) followed by a subsolvus granite (T5) displaying low outward-dipping
J. P. LiOgeois and R. Black
386
I
IoO0'E A TIME
L
.::.:-)-!i-:7. I
18o50'N -
l
A'
0 5 km P--re.---roB---1
At
A
T2
T5
T4
T5
T6
m FIG. 5. Geological map of the Timedjelalen ring-complex: T 1, granite porphyry; T2, medium-grained granite (perthite, aegirine-augite, Ca-Na- amphibole, biotite); T3, coarse-grained granite (perthite, aegirine-augite, CaNa- and Na-amphiboles); T4, heterogranular granite (perthite, oligoclase, amphibole, biotite); T5, fine-grained granite (perthite, oligoclase, biotite); T6, fine- to medium-grained granite (aegirine, arfvedsonite). (Ba et al. 1985.) contacts. Lastly, a peralkaline medium-grained hypersolvus granite (T6) occurs, piercing T5 as a shallow dome. Despite the mineralogical diversity, the compositional range of the Iforas alkaline province is very narrow. All the rocks are rich in silica with values of 64%-68% for the quartz syenites and 72~-79% for the granites. CaO and MgO are always low, and are reflected by sodic plagioclase (maximum Anls) and by mafic minerals with high FeO/(FeO + MgO): fayalite, pyroxenes (Feaugite, hedenbergite and aegirine), amphiboles (Fe-hornblende, Fe-richterite and arfvedsonite) and Fe-biotite. K20 (4.2~-5.6~) and Na20
(3.4%-4.1%) are not particularly high for granites. In molecular proportion, Na20 is generally slightly in excess of K20. The rocks are mainly peralkaline or metaluminous with rare peraluminous varieties containing less than 1~ of normative corundum. Fluorine is always abundant but, characteristically, boron is very low and tourmaline is absent. Contents of Th, Zr, Y, Rb and the rare-earth elements (REE) are high, particularly in the peralkaline facies. Petrologically, the most striking features of the quartz syenites and granites is the abundance of perthites displaying a wide range of textures and the rich assemblage of ferromagnesian minerals
A l k a l i n e m a g m a t i ~ m in the Iforas province AI
often displaying beautiful reactions and mantling relationships. Ba et al. (1985) have shown the presence of two distinct trends. 1 A sodic peralkaline trend extends as an evolving fractionation sequence from a metaluruinous quartz syenite, characterized by a miaskitic sequence in which the mafic minerals (fayalite, Fe-augite, Ca-amphibole and biotite) crystallize earlier than the K-feldspar and quartz, to peralkaline granite where the Ca-Na- and Naamphiboles and aegirine are late and mould the feldspars. Relict fayalite is only present in the less-differentiated metaluminous syenite and granite; it is frequently altered and mantled by Ca-amphibole. Clinopyroxenes are nearly always present and form an apparently continuous suite: Fe-augite-hedenbergite-Na-hedenbergite-aegirine. Within a single intrusion, Ca- and Ca-Napyroxenes are early and separated from late Napyroxenes by the crystallization of amphiboles. Amphiboles show a very wide range in composition. The green Ca-amphiboles have compositions between hornblende, edenite and barroisite and mantle the fayalite and Ca-pyroxene. The blue-green amphiboles mantling the Ca-amphiboles in some of the metaluminous granite and occurring as the main amphiboles in the peraIkaline granite range from between barroisite and richterite to arfvedsonite and may be rimmed by riebeckite. Micas are not frequent and are represented by Mg-biotite in the syenites, biotiteannite in the metaluminous granites and lepidomelane in the peralkaline granites. Other typical peralkaline minerals present in the province are aenigmatite and astrophyllite. The accessory minerals include coarse zircon, zoned allanite, Fe-Ti oxides, apatite, fluorite, stilpnomelane and tchevkinite. Sphene is rare in the Kidal complex but is abundant in the Timedjelalen. The albitemicrocline-arvfedsonite granite present in the Kidal complex (K12) has undergone strong albitization by sub-solidus alteration of the preexisting magmatic feldspars and is similar to those described in Nigeria (Jacobson et al. 1958). 2 A potassic-aluminous trend is represented by granite porphyries, hypersolvus amphibole-biotite granite, subsolvus amphibole-biotite and biotite-chlorite granite. Anhydrous early-formed mafic minerals (fayalite, clinopyroxene) are rare and the ferromagnesian reaction series is much more limited. The calcic amphiboles belong to the hastingsite-actinolite series and are generally euhedral. They are mantled by biotite (annitelepidomelane) which contains inclusions of zircon, allanite, apatite, Fe-Ti oxides and sphene. Late alteration plays an increasing role on passing from the hypersolvus granite to the subsolvus
387
/ ~
Na
V 35
.
,-
V 30
,30
V -25
K
FIG. 6. Na-K-AI (atomic) diagram for the Kidal massif; 0, syenites; I , peralkaline hypersolvus granites; 9 peralkaline albitized granites; A, metaluminous granites. (Ba et al. 1985.)
granite with the appearance of fluorite-albitechlorite-epidote + muscovite. The trends appear clearly on the Na-K-AI diagram (Fig. 6). The evolution from metaluminous quartz syenite to peralkaline granite, passing through transitional facies with fayalite, hedenbergite and Ca-Na-amphiboles, occurs with an increasing (Na + K)/A1 ratio and a constant K/ Na ratio; after moving into the peralkaline field the K/Na ratio decreases. This path can be explained by a plagioclase effect accompanied by fractionation of early mafic minerals (fayalite, Ca-pyroxene and amphiboles), which is followed by intense fractionation of K-feldspar (orthoclase effect) as it crosses into the peralkaline field and descends the thermal valley (Bailey & Macdonald 1969). This stage may be followed by deuteric recrystallization in the presence of hydrothermal F-bearing fluids giving rise to a sub-trend characterized by a high (K + Na)/A1 ratio and a fall in the K/Na ratio corresponding to the albitization of the K-feldspars which produced the albite-microcline-arfvedsonite granite. The aluminous trend from granite porphyries to subsolvus biotite-chlorite granite is marked by decrease of ( N a + K ) and Na/K ratios. This reflects the appearance of biotite and of perthites with K greater than Na. A possible explanation may be the fractionation of basic oligoclase, Fehornblende and clinopyroxene if the dark inclusions sometimes encountered are cumulates. The independence of the two trends is suggested by the Sr and Ba concentrations. The metaluminous granite porphyries have more Sr (100 ppm) and less Ba (500-800 ppm) than the
388
J. P. LiOgeois and R. Black
syenites (50 ppm Sr, 1000 ppm Ba). This suggests a divergent evolution before the syenite stage, with the common source perhaps being of monzonitic composition. These trends are very similar to those described by various workers in Nigeria, Niger, Corsica, Sudan and New Hampshire, and also occur in mixed oversaturated and undersaturated provinces which may be in continental (Gardar, Quebec, Oslo) or oceanic (Kerguelen) areas. These were first reviewed over 40 years ago by Barth (1944). To conclude, the Iforas alkaline province displays all the main petrological and geochemical characteristics of typical anorogenic oversaturated alkaline provinces. When compared with the Nigerian province (Jacobson et al. 1958; Bowden & Turner 1974) and the Niger province (Black 1963; Black et al. 1967), the main differences can be summarized as follows. 1 The volcanism is fissural and the ringcomplexes do not appear to have acted as central volcanoes. 2 Basic rocks (gabbros, anorthosites) are absent. It should be noted, however, that gravity data suggests that dense rocks may be present at depth (Ly et al. 1984) and some microgabbros have been observed as dykes. 3 Economic mineralization is absent. Among the metaluminous granites, subsolvus varieties are abundant whereas the hypersolvus facies predominate in Niger-Nigeria where the albitized granites are associated with Sn-bearing greisens and contain columbite.
The Pan-African orogenic context The Iforas alkaline ring-complexes and associated dyke-swarms lie within the Pan-African belt 100-150 km to the E of the suture between the two palaeo-continents, the W African craton and the Tuareg shield. They are aligned N-S along the axis of the composite calc-alkaline batholith of western Iforas, which is a major structural unit bound by N-S mega-shears (Fig. 1). Clearly, an understanding of the alkaline rocks and of the rapid switch from calc-alkaline to alkaline magmatism must take into account Pan-African orogenic evolution. Detailed mapping of the segment of the batholith between the Kidal ring-complex (latitude 18~ ') and the Tiralrar plateau (latitude 19~ ') (see Figs 2 and 3) has established the relative chronology of magmatic events in this area. The granitoids fall into four groups: (1) lowK calc-alkaline pre-tectonic; (2) high-K calcalkaline late-tectonic; (3) high-K calc-alkaline post-tectonic; (4) alkaline post-tectonic.
All groups cut the tillite-bearing Tafeliant volcano-sedimentary group whose age can be bracketed between 693_ 7 Ma (U-Pb age from the granodiorite beneath the base of the Tafeliant group (Caby & Andreopoulos-Renaud 1985)) and 615+5 Ma (U-Pb on zircon from the crosscutting Adma granodiorite (Ducrot et al. 1979)). Detailed petrographic descriptions of the different massifs have recently been published (Li6geois & Black 1984), and they can be summarized as follows. (1) The pre-tectonic group is represented by the Erecher tonalite composed essentially of zoned andesine, quartz, green biotite and hornblende, and accessory microcline, sphene, Fe-Ti oxides and zircons. The massif has been mylonitized to variable degrees, locally producing a pronounced foliation and partial recrystallization in high greenschist facies. Some granite plutons (Yenchichi 1 type) may also belong to the pre-tectonic group. (2) The late-tectonic group is by far the more important in volume and is composed of three major rock types. (a) The granodiorites, the Adma pluton being typical, display two crystallization steps" the first (quartz, andesine, microcline, green biotite, hornblende, __+clinopyroxene, __+orthopyroxene) is locally affected by the E-W compression producing N-S foliation, whereas the second (amphibole, biotite, quartz-feldspar symplektites) is not affected and is posttectonic. (b) The porphyritic monzogranites are composed of slightly perthitic microcline phenocrysts, quartz, oligoclase, green biotite and amphibole, abundant accessory sphene and apatite, Fe-Ti oxides, zircon and secondary chlorite and epidote. They also often display a planar fabric with strong gradients of deformation developed in proximity to shear zones. (c) The fine-grained monzogranites are composed of millimetre-sized quartz, oligoclase, slightly perthitic microcline, chloritized biotite and accessory sphene, Fe-Ti oxides, allanite, apatite and zircons, and present only a weak planar fabric. Intrusion has followed the porphyritic monzogranite closely, with contact zones often being hybrid and characterized by the presence of large K-feldspar xenocrysts. (3) The post-tectonic calc-alkaline group is represented by dense W N W - E S E dyke-swarms (Yenchichi and Telabit) and circular sharpcutting syenogranite plutons. The dykes comprise intermediate to acid rocks (55%-77% SiO2) and
Alkaline magmatism in the lforas province some of them have alkaline affinities (Dohendal dyke-swarm). The plutons are undeformed, homogeneous in composition and more acid than the preceding groups (Yenchichi 2 pluton). (4) The late post-tectonic alkaline group includes the Tahrmert granite, spectacular N-S dyke-swarms (quartz syenite porphyries, granophyres, quartz-feldspar porphyries and rhyolites), extrusive rhyolites and ignimbrites (Tiralrar), and high-level ring-complexes (Kidal, Timedjelalen) composed of quartz microsyenites and metaluminous and peralkaline granites.
Ages, isotopic and main geochemical characteristics of the calc-alkalinealkaline transition The Rb-Sr system gives for the pre-tectonic family a rehomogenization age which corresponds to the collision event (602__ 13 Ma, 0.70590+0.0008, 9 whole-rock samples (WR), MSWD = 1.0) (Fig. 7). The two Rb-Sr isochrons on the Adma granodiorite and on the Aoukenek fine-grained monzogranite (late-tectonic family) give similar ages around 595 Ma with relatively low initial ratios (Adma: 595+24 Ma, 0.70482__+0.00026, 9 WR, MSWD = 0.7; Aoukenek' 591 • 18 Ma, 0.7035+ 0.0005, MSWD = 0.9) (Fig. 7). As a U-Pb zircon age of 615• Ma has been obtained for the Adma pluton (Ducrot et al. 1979), the slightly younger Rb-Sr ages are believed to date final consolidation marked by the second stage of crystallization and to correspond to the end of the collision. The porphyritic monzogranite only yields an errorchron, probably owing to the wide scatter of sampling. As its age, on the field evidence can be assumed to be almost contemporaneous with the fine-grained monzogranite, its initial ratio can be bracketed between 0.7042 and 0.7053. Three dyke-swarms (Yenchichi, Dohendal and Telabit) and one granite pluton (Yenchichi 2) of the post-tectonic calc-alkaline family have been dated using the Rb-Sr method. This group seems to show an age trend from S (Yenchichi swarm" 565+ 14 Ma, 0.7048___0.0005, 7 WR, MSWD=4.6; Yenchichi 2 pluton: 577+ 14 Ma, 0.7038 + 0.0010, 7 WR, MSWD = 1.6) to N (Dohendal swarm" 558+ 10 Ma, 0.705 11 +0.00012, 7 WR, MSWD =0.7;Telabitswarm: 544+ 12 Ma, 0.70505+0.00010, 13WR, M S W D = 4 . 6 ) ( F i g . 7). All the representatives of this group also have low initial 878r/86 Sr ratios with values between 0.7038 and 0.7051.
389
Representatives of the alkaline group, on field evidence, are all younger than the calc-alkaline intrusions. The truncation by the Timedjelalen ring-complex of the transcurrent fault cutting the Takellout and Kidal complexes indicates S-N migration (Figs 2 and 3). The geochronological data confirm the post-tectonic age trend from S to N: S, Kidal ring-complex ( 5 6 1 _ 7 M a , 0.7061 +0.0007, 25 WR MSWD =2.1); N, Tahrinert pluton (542 • 7 Ma, 0.7061 _ 0.0004, 12 WR, MSWD=3.6), N-S dykes (543-1-9 Ma, 0.7050+0.0003, 14 WR, MSWD=2.0) and Timedjelalen ring-complex (546__+7 Ma, 0.7058 + 0.0003, 21 WR, MSWD=3.3) (Fig. 7). In the N alI the intrusions are contemporaneous, within the • Ma limit of error, but the relative chronology established by the field relationships is as follows: 1, Tarhmert; 2, N-S dykes and lavas; 3, Timedjelalen. All the alkaline units have Sr initial ratios between 0.7050 and 0.7060. The NW-SE sinistral transcurrent faulting has been dated. Such a fault cutting the Yenchichi 2 pluton also affects a part of the earlier Yenchichi 1 pluton. An isochron based on Yenchichi 1 mylonites gives an age around 545 Ma and can be interpreted as the rehomogenization of the Rb-Sr system during the shearing process (544 ___ 16Ma, 0.7063-1-0.0005, M S W D = I . 6 , 9 W R ) (Fig. 7). A geochemical evolution, related to age and tectonic setting, can be traced from the calcalkaline pre-tectonic group through the latetectonic and post-tectonic calc-alkaline groups to the alkaline post-tectonic magmatism. This appears on the simple SiO2 versus N a 2 0 + K 2 0 diagram (Fig. 8(a)). Three trends are clearly distinguished' low-K calc-alkaline represented by the pre-tectonic granitoids, a high-K calcalkaline trend by late- and post-tectonic groups and an alkaline trend for late post-tectonic granites. The distinction between the three trends also appears in other geochemical diagrams, for example in the Rb versus K20 diagram (Fig. 8(b)). In this diagram, the alkaline group is dispersed perpendicular to Shaw's main trend (Shaw 1968) as in the case of the Rallier du Baty ring-complex in the Kerguelen archipelago (Lameyre et al. 1976; Vidal et al. 1979). However, the pre- and late-tectonic calc-alkaline families follow the main trend, although with some higher KzO/Rb ratios. It is interesting to note that the calc-alkaline post-tectonic group shows some alkaline affinities. Several samples of the peralkaline phases of the Timedjelalen ring-complex follow the now classical 'pegmatitic-hydrotherreal' trend (Shaw 1968), probably in response to post-magmatic auto-metamorphic fluids (Vidal et al. 1979).
390 0.75
J. P. LiOgeois and R. Black Adma
Erecher
_
f
0.715
-
0.72
o/01 01 0.71
-
o
/, 0.701 0 (a)
0.710 602 +_ 13 Ma O. 70 60 + 0 . 0 0 0 8 MSWD = 1.0
9
o . -~"
1 0.5
/o/ 9
I
I 1.5
I
I 2
I 2.5
595_+ 24 Ma 0.70482+0.00026 MSWD=0.7
0.705 0.700
I
0
0.2
I
0.4
I
0.6
I
I
0.8
1.2
I
I
1.4
1.6
1.8
(b)
/
Yenchichi E-W dykes
Aoukenek 0.76
~
I
1.0
0.74
03
9149 9
/ 0.70 L 0
MSWD
I
I
I
2
565+14 Ma 0.7048 + 0.0005 MSWD = 4.6
591_+18 Ma 0.7055+ O. 0005
0.I 9
0.72
I
5
= 0.96 I
1
4
5
0.7 0
(c)
1.0
0.9
_
I0
I 20
L 50
I
50
(dl
!
Dohendol E-W d y k e s /
0.75~-
/
_
Telabit E-W dykes
.I
O. 72 ~
/ o .,.,,. i
558+10Ma 0.70510 + 0.00012 MSWO = 0.7
0.8 -
0.7 L/ 9 0 (e)
I
40
I I0
] 20
I 50 87Rb/8%r
I 40
l 50
II
0.7f~
0.70 0 (f )
1, ./ oId,/ 9"~
I 0.5
/ I 1.0
/
~
9
544 + 12 Ma 0.70505 + 0.00010 MSWD = 4 .6
l 1.5
1 2.0
I 2.5
87Rb/8%r
FIG. 7. Rb-Sr isochrons : (a) rehomogenized pre-tectonic Erecher tonalite; (b) late-tectonic A d m a granodiorite; (c) late-tectonic Aoukenek monzogranite; (d) post-tectonic Yenchichi E - W dyke-swarm; (e) post-tectonic Dohendal E - W dyke-swarm; (f) post-tectonic Telabit E - W dyke-swarm; (g) post-tectonic Yenchichi 2 circular syenogranite; (h) post-tectonic Tahrmert alkaline granite; (i) post-tectonic alkaline N - S dyke-swarm; (j) posttectonic Timedjelalen ring-complex; (k) post-tectonic Kidal ring-complex (some samples with very high Rb/Sr ratios are outside the figure (see Li6geois & Black 1984); (1) pre-tectonic (?) Yenchichi 1 granite. Analysed mylonitized samples indicate rehomogenization during late transcurrent faulting events. The calculations were made following Williamson (1968) with k = 1.42 x 10-11 a - 1. The errors are at the 20 level. (From Li6geois & Black 1984.)
Alkaline magmatism in the Iforas province
[
Yehchichi 2
1.0
391
Tohrmert
./~
1.0
<
J
0.9
0.9
olO.~176 0.8 .f"
.1
0.7
I 5
0 (g)
I I0
I 15
0.85
0.80 Z
.Y
./
/4
l I0
I
5
0
0.7
I 25
e/
I 15
0 (h}
I
I
I
I0
20
50
1 40
oL ~ ..//'"
0.8
I 20
0.7
I 25
60
./
/546+_7Mo
ring-complex
545_+. 9Mo 0,7050+0.0005 MSWD=2.0
I 50
./"
Timedjelalen
(2o-)
" 0.7058+0.0005
(2o-)
MSWD = :5.5
--
/~ 0
[
I
I
I
I
I
I0
20
30
40
50
60
(j)
(i)
Kidol
0.90
542+_7Ma 0.7061 + 0 . 0 0 0 4 MSWD =5.6
,/"
I,/"
0.9
/~o/~ 0.70
I 20
0/ 9
~
~ .o
-
0.8
/
N - S dykes
0.75
577+ 14Ma 0.7058 + 0.0010 MSWD: 1.6
. /
/.
ring- complex
Yenchichi
I
0o
0.85
Z
0.80
/
./
0.77 -
Y
0.75
/ / / / "
~176
--
/~
so"
..,," / 0-
0,75
_
3.7~ 5 6 1 + 7 M o (2o-) 0.7061 + 0 . 0 0 0 7 (2o-) MSWD : 2.1 0 71
0/~
./ 0o~
0 70 0 (k)
I 5
/.ofI./~
I I0
I 15
I 20
I 25
87Rb186Sr
In the alkaline ring-complexes the associated subsolvus metaluminous granites are mineralogically practically indistinguishable from the posttectonic acid members of the high-K calc-alkaline group. They may be discriminated, however, in the SiOz versus agpaitic index (AI) diagram (Fig.
i 0 ( L)
5 4 4 + 16 Ma 0.7065 + 0 . 0 0 0 5 MSWD = 1.6
,./I I I
I 2
I 3
1 4
I 5
8"tRb/86Sr
9), the limit between the two groups correspondi n g to an AI of 0.88. In a REE chondritenormalized diagram (Fig. 10) (Li6geois & Hertogen, unpublished data) the alkaline group including the metaluminous granites displays a general enrichment in REE, particularly in heavy REE,
I 6
J. P. Li#geois and R. Black
392
X
X
X
x o
4o
x ~ Xxx ~
9
~ ~o
9
!
x x x
9,,,
.
o**
x
x ~ ,,~z~x x ~ , f ~ x _
x
8
.. 9149
Z
9
oOoo # 9 ~149~
o~
x
o O
6 0
4
1 55
I 60
O~
oo
o
0
I 65
I 70
(a)
I 75
I 80
% Si0z
MT
6
x
x
9 x
I
X
b-
x
X
o 4v 04
xx x ,.y.X X~
x
X
x
~ ,/A ~ ' X ~ ,'l~l. ^ x ,',
x X x O0
9
,
i
I00
I
PH x
. x
~.zx. ,.x./X ~. o ",;,":,~ 9 g . 9 ~ / 9
3-
(b)
/x
x xx xxtx • x
0
^^x / x, XA j X
~
n
I
200
300
J
I
400
J
I
t
500
ppm Rb
FIG. 8. (a) SiO 2 versus Na20 + K20 diagram showing the different trends of the Iforas magmatic groups from the pre-tectonic low-K calc-alkaline (O) to alkaline ( • ) through late (O) and post-tectonic high-K calc-alkaline (&); (b) Rb versus K20 diagram (same symbols as in Fig 8(a)). The lines labelled MT enclose the 'main trend' and PH represents the 'pegmatitic-hydrothermal trend' of Shaw (1968). with a pronounced Eu anomaly, whereas the high-K calc-alkaline rocks, as shown by the latetectonic group, yield steep curves with low contents of heavy REE. This distinction between granitic rocks affiliated to the alkaline or calcalkaline group based on R E E patterns has been successfully used by Harris (1985) in a study of mixed anorogenic complexes in Saudi Arabia. The four groups plot distinctly on the Rb against Yb + Ta discrimination diagram (Fig. 11) (Pearce et al. 1985). Whilst the pre-tectonic group clearly falls in the V A G field, the late-tectonic granitoids lie astride the V A G and syn-COLG join; the main alkaline representatives plot in the W P G
field and the earlier transitional E - W dykes and Tahrmert granite lie to the left in an intermediate position. The geochemical data will be published in a separate paper.
Interpretation and the petrogenetic model Three distinct trends, low-K calc-alkaline, highK calc-alkaline and alkaline, are easily distinguishable in the field and in the laboratory and can be related to successive palaeo-environments.
Alkaline magmatism in the Iforas province
393
1.2 x
+
x
x
x x
x
§
1.0
~
x
•
"~)x • 'i
X X
xx ,j,o, ~
0.8
" J o , ,*
9 9 9 9
9149
00 9 ~ 9 9 ~
O~"
0.6 -
9 9
;.
X9 ~' x ~ ,. oo
8 o
9
o 0
{~
000
0
#o 9
9
1
I
I
J
I
J
55
60
65
70
75
80
% SiO2
FIG. 9. SiO., versus AI (molecular (Na20 + KzO)/AI203)" O, pre-tectonic low-K calc-alkaline- O, late-tectonic high-K calc-alkaline; A, post-tectonic high-K calc-alkaline; +, post-tectonic alkaline N-S dykes; x, posttectonic alkaline ring-complexes.
Ioo
W
to
La CIe
Nd
Sm Eu Gd Tb
Ho
Yb Lu
FtG. 10. Range of chondrite-normalized REE for (1) the late-tectonic high-K calc-alkaline group (six samples) and (2) the post-tectonic alkaline group including metaluminous and peralkaline samples (six samples). Data from Li6geois & Hertogen, unpublished.)
Allthe groups have relatively low STSr/S6Sr initial ratios (0.7035-0.7061) (Fig. 12), which preclude any important participation of an old upper continental crust even during the collision (also shown for eastern Iforas by Bertrand & Davison (1981)). The two remaining possible sources are the lower crust and the upper mantle. The granulites of the Iforas (Boullier et al. 1980) offer the possibility of estimating the 878r/S6 Sr ratio of the lower crust. Preliminary results show a minimum mean value 600 Ma ago of 0.7066 for the STSr/S6Sr ratio (Li6geois, unpublished data). As this value is much higher than that of the magmas produced, these granulites cannot represent the source (Fig. 12). If they are not representative of the Iforas lower crust, the fact that the generation of magmas with 87Sr/S6Sr ratios as low as 0.7035 (Aoukenek granite) requires strong depletion of the source in lithophile elements, implying a highly refractory lower crust, is incompatible with abundant magma production. This non-crustal origin is confirmed by the lack of crustal-derived inclusions in the studied plutons. In contrast, the island arc situated west of the batholith and composed essentially by basic and intermediate rocks reflects the composition of the depleted mantle under the Iforas in Pan-African times and provides homogeneous 87Sr/86Sr initial ratios between 0.7025 and 0.7030 (Caby et al., 1986). It would therefore appear that the Iforas
394
J. P. L i @ e o i s a n d R . B l a c k
I000 Syn-COLG
9 -~ ~ W P G x 410
.~
~
x
x x
I00 x
VAS o
I
I0 Yb + To ppm
FIG. 1l. Discrimination diagram Yb + Ta against Rb (field boundaries from Pearce et al. 1985): O, pretectonic low-K calc-alkaline; O, late-tectonic high-K calc-alkaline; A, post-tectonic high-K E-W dykes; ,, post-tectonic Tahrmert pluton; x, post-tectonic alkaline N-S dykes and ring-complexes; VAG, volcanic-arc granite; syn-COLG, syn-collision granite; WPG, within-plate granite; ORG, ocean ridge granite. See text for discussion. (Data from Li~geois & Hertogen, unpublished)
calc-alkaline granitoids have been generated in the mantle with crustal contamination during their ascent. The generation of important amounts of granitic rocks in the mantle has been discussed for many years, the main problem being the need to have at least an equal quantity of basic rocks at depth. In this connection it must not be forgotten that the great crustal thickness in the Andean cordillera (up to 70 km (James 1971) is due not to compression but essentially to vertical accretion from the underlying mantle (Brown 1977). In Western Iforas E - W crustal shortening across the volcano-sedimentary sequences related to the 600 Ma collision has been estimated at only about 65% (Ball & Caby 1984) and probably much less than in the granitoids (Li~geois & Black 1984). It is likely that the mountain building in the Iforas has been due, at least in part, to vertical accretion under the continent in both the subduction and the collision epochs. If we accept this origin for the Iforas granitoids at mantle depth, several classical models for andesite generation are available (Best 1975; Fyfe & McBirney 1975 ; Thorpe et al. 1976). The most plausible model is probably partial melting in the upper mantle above the subduction zone
with addition of H20-rich fluids coming from the dehydration of the down-going oceanic lithosphere, which would provide alkalis and favour melting by lowering the solidus temperature (Thorpe et al. 1976). This phenomenon can introduce radiogenic Sr in the mantle because of the alteration by seawater (Pan-African 8VSr/S6Sr around 0.708; extrapolation from Peterman et al. (1970)) of the oceanic basalt crust (Hawkesworth et al. 1979). However, in the Iforas the Sr isotopic composition of uncontaminated mantle melts can be approached in the island arc (0.7025-0.7030) (Caby et al., 1986). The slightly higher values which have been measured in the cordillera for the calc-alkaline groups (0.7035-0.7051) must then be attributed to crustal contamination (Fig. 12). This participation of the crust in the genesis of the mantle Iforas granitoids can also explain their relatively high silica contents. In conclusion, we propose for the subductionand collision-related calc-alkaline rocks a common origin in the mantle with subducted oceanic crust participation and some crustal contamination during the magma ascent (Fig. 13). A main difference between the pre- and late-tectonic magmas is their K20 (and Rb) contents. This can be explained by a greater depth for the mantle mobilization during the collision (K-h relation) (Dickinson & Hatherton 1967; Arculus & Johnson 1978 ; Dupuy et al. 1978). In fact the collision between the Tuareg shield and the W African craton in the Iforas has consisted of 'docking' rather than a Himalayan-type confrontation as would appear to have been the case further S in Benin (Burke & Dewey 1972; Trompette 1980). This would explain the weak crustal mobilization and the total absence of S-type granites (in contrast with the Himalayan leucogranites (Le Fort 1981)) and weak syn-COLG affinity of the late-tectonic granitoids in the Rb against Yb + Ta diagram (Fig. 11). The last calc-alkaline group is poorly represented and is clearly post-tectonic. Since its geochemistry is very similar to that of the latetectonic group, it is taken as the last manifestation of the source mobilized during the collision. As some samples display alkaline affinities (Figs 8 and 9), a beginning of participation of the alkaline source is likely, particularly in the N where the two post-tectonic families are approximately contemporaneous. The emplacement of this group occurred during the rapid uplift of the belt (Fig. 13). The Iforas alkaline rocks follow the calcalkaline post-tectonic group without a significant time break towards the end of the uplift. Indeed, only the early alkaline Tahrmert pluton was eroded during the unroofing of the batholith. This
Alkaline magmatism in the Iforas province
395
0.707 Lower limit of Iforas granulitic
0.706
i i
~_~ Yenchichi I
I
unit (Iforas Dower crust) _A I Erecher
I--4 P---I t ~ Kidal __ --
E-WT Dohendal F 0.705
I
oo
"~ Adma T
0.704 Aoukenek I
I-'
0 . 7 0 3 - Depleted upper mantle
0.702
I 650
I
--
T ~
I
Tahrmert
Timedjelalen w Telabit
',,~ I- ~NS' I E - W Yenchichi
Yenchichi Tf -'I
I ',
--
I 600
I 550
I 500
Age (Ma)
FIG. 12. 8~Sr/86Sr initial ratios of the intrusions as a function of their isochron age (the error bars represent 2 level isochron errors): A, pre-tectonic low-K calc-alkaline plutons (rehomogenized values); I , late- and posttectonic high-K calc-alkaline plutons and dykes (end-of-crystallization values); O, post-tectonic high-level alkaline plutons and dykes (intrusion values). The 87Sr/86Srratio of the Iforas lower crust is inferred from the Iforas granulitic unit and that of the depleted mantle from the island arc (see text).
implies that the lava plateaux, fed by the N-S dyke-swarms, represent relics of a thick volcanic cover which extended over most of the Iforas and under which the ring-complexes were intruded. Volumetrically, the alkaline magmatism was very important, as is confirmed by the rhyolitic and ignimbritic composition of the molasse associated with the belt (Fabre 1982). The onset of alkaline magmatism was accompanied by reactivation of N-S mega-shear zones related to a change in the direction of major constraints shown by the switch in strike in the dyke-swarms from W N W ESE (calc-alkaline) to N-S (alkaline). Movements along sinistral N N W - S S E transcurrent faults has been dated around 545 Ma (Yenchichi I mylonites). A 560-535 Ma age range based on U - P b zircon and 39Ar/4~ Ar ages has been obtained from deformed granite within the nearby mega-shear zone defining the western limit of the Iforas granulite unit (Fig. 1) (Lancelot et al 1983). Distension, produced essentially by horizontal movements, was weak and marked by
narrow grabens filled with molasse located along major shear zones Fabre (1982). The geochronological data seem to indicate the diachronous emplacement of the two post-tectonic families (from S, 580-560 Ma, to N, 550540 Ma). This could be related to differences in the rate of uplift which may possibly be linked to the obliquity of the collision (Ball & Caby 1984). The sudden and radical change in post-tectonic conditions from typical subduction and collisionrelated calc-alkaline magmatism to alkaline magmatism, displaying all the geochemical and petrological characteristics of within-plate anorogenic complexes, implies the tapping of a new source. Trace-element and REE patterns indicate a less depleted water-poor mantle origin in contrast with a subduction-related source. Pb isotope data on feldspars (Li6geois & Lancelot, unpublished data) show that the calc-alkaline group had a source distinct from the more radiogenic source for the alkaline rocks. The Iforas alkaline ring-complexes have slightly
396
J. P. LiOgeois and R. Black "t. Subducti0n ( 7 2 0 - 6 2 0 Mo)
Trr.uplift
South 5 9 0 - 5 7 0 Ma North 5 9 0 - 5 5 0 Mo
FIG. 13. Schematic representation of the proposed model for the Pan-African magmatic evolution of the Iforas.
Alkaline magmatism in the Iforas province
397
398
J. P. LiOgeois and R. Black
higher 87Sr/86Sr initial ratios (0.7050-0.7061) than those of the components of the surrounding calc-alkaline batholith. This precludes crustal contamination at shallow depth as a means of raising the initial ratios. Moreover, oxygen isotope data have shown interaction with meteoric water to have been only very local (Weis et al.). Some participation of lower crust is likely but it is difficult to see why it should have been greater than in the case of the calc-alkaline group. We suggest that the new source is less depleted asthenospheric mantle, originally underlying the subducted plate, which has risen to shallow depth after rupture of the cold plunging plate (Fig. 13). This source would then be the same as that often proposed for within-plate alkaline complexes emplaced in a strictly anorogenic environment. In the case of the nearby Tadhak alkaline undersaturated and carbonatite province of Permian age (Li6geois et al. 1983), the main ringcomplex (Tadhak) can be interpreted as a pure mantle product on the basis of Sr and Pb isotopic data (Weis & Li6geois 1983) with a source of the ocean island basalt type. An estimation of the Sr isotopic composition of this deep mantle in PanAfrican times can be obtained by calculating back the isochron-deduced 87Sr/86 Sr initial ratio to 550 Ma. This gives a ratio of about 0.7043, which is clearly distinct from that of the depleted lithospheric mantle represented by the island-arc lavas (0.7025-0.7030).
Conclusions The Cambrian oversaturated alkaline province, with its extrusive plateau rhyolites, spectacular acid dyke-swarms and typical alkaline 'youngergranite' ring-complexes, displays all the major petrographical and geochemical characteristics of anorogenic within-plate magmatism, despite its appearance only a few million years after a major orogeny (Pan-African) and its location along a sutured convergent plate boundary. As the magmatic products are essentially the same in contrasting settings, the source must be identical and an explanation for peculiarities must be sought in the basement geology (Black et al. 1985). There is now considerable evidence for interpreting the Pan-African Trans-Saharan orogenic belt (Cahen et al. 1984) in terms of a Wilson cycle ending with collision between the passive continental margin of the W African craton and the active margin of the Tuareg shield around 620 Ma (Black et al. 1979; Caby et al. 1981). In the Iforas, subduction-related low-K calc-alkaline pre-tectonic magmatism is recorded in the interval 725-
620 Ma and is found in both an island-arc and a cordilleran environment. During and after collision high-K calc-alkaline late-tectonic granitoids forming the main component of the Western Iforas batholith were emplaced between 620 and 590 Ma and were followed by affiliated posttectonic plutons (580-550 Ma) during the rapid uplift which led to unroofing of the batholith. The switch to alkaline magmatism occurred around 560 Ma ago in the S and 545 Ma ago in the N indicating a northward migration. All the calc-alkaline magmatism related to subduction or to collision present low 87Sr/S6Sr initial ratios between 0.7035 and 0.7051 and may be related to a lithospheric depleted mantle source with participation of subducted oceanic crust at depth and lower-crust Eburnean granulites during ascent. The island-arc material suggests that this mantle source has had an 87Sr/ 86Sr initial ratio around 0.7025--0.7030 (Caby et al., 1986). The absence of typical crustal granites (S-type) is probably due to the oblique nature of the collision which in the Western Iforas did not lead to a doubling of the crust as in the Himalayan situation. These results, in contrast with W. Q. Kennedy's original view of the Pan-African as an episode of basement reactivation (Kennedy 1964), provide strong evidence of important crustal accretion during this event. The switch to alkaline magmatism with its distinctive geochemical characteristics is attributed to a different more primitive mantle source corresponding to the asthenosphere. The alkaline rocks have 87Sr/86Sr initial ratios ranging from 0.7050 to 0.7061 compared with an estimated mean value for this source of 0.7043, based on Rb-Sr and Pb-Pb data from the nearby Permian Tadhak undersaturated complex which is a pure mantle product (Weis & Li6geois 1983). Structurally, the onset of alkaline magmatism in the Iforas occurred at the end of major uplift and was accompanied by a change in the stress field with renewed and intermittent transcurrent movements along lithospheric mega-shear zones and related wrench faults producing distension, as indicated by the presence of dyke-swarms and narrow molasse-filled grabens. In the model proposed, access to a new source is explained by rise of the asthenosphere to shallow depths beneath continental lithosphere after rupture of the cold subducted plate. Asthenospheric mantle, which has often been considered as a source for alkaline magmatism, takes into account the petrological and geochemical features common to all A-type granites. It can be argued that differences such as the presence or absence of associated tin mineralization are due to the nature of the country rocks. The
Alkaline magmatism in the Iforas province chances of finding economic Sn deposits are considerably enhanced when the intrusions cut crustally derived granitoids which have already redistributed and concentrated the metal, e.g. Nigerian Pan-African granites and tin-bearing pegmatites. In the Iforas, where the basement is composed of depleted granulitic rocks and essentially mantle-derived granitoids, there is little possibility of finding such mineralization (Black 1984).
Lastly, we would like to stress that whilst the source of alkaline magmatism is to be sought in the deep mantle, its location and nature are largely controlled by the structure, composition
399
and dynamics of the overlying continental lithosphere (Black et al. 1985).
ACKNOWLEDGEMENTS: This work is a result of collaboration between the Direction Nationale de la G~ologie et des Mines, Mali, the Centre G6ologique et G6ophysique, Montpellier, the Laboratoire de P&rologie, Paris VI, and the Mus~e Royal de l'Afrique Centrale, Tervuren. We acknowledge the financial support of the FNRS and Minist6re de l'Education Nationale of Belgium and of the CNRS and Fond d'Aide et de Cooperation of the French Republic.
References ARCULUS, R. J. & JOHNSON, R. W. 1978. Criticism of generalised models for the magmatic evolution of arc-trench systems. Earth planet. Sci. Lett. 39, 11826. BA, H., BLACK, R., BENZIANE, B., DIOMBANA, D., HASCOET-FENDER, J., BONIN, B., FABRE, J. & LI~GEOIS, J. P. 1985. La province des complexes annulaires sursatur& de l'Adrar des Iforas, Mali. J. Afr. Earth Sci. 3, 123-42. BAILEY,D. K. & MACDONALD,R. 1969. Alkali-feldspar fractionation trends and the derivation of peralkaline liquids. Am. J. Sci. 267, 242-48. BALL, E. & CABY, R. 1984. Open folding and constriction synchronous with nappe tectronics along a mega-shear zone of Pan-African age. In : KLERKX, J. & MICHOT, J. (eds) Gbologie ajHcaine--Ajkican Geology, pp. 75-90. Mus6e Royal de l'Afrique Centrale, Tervuren. BARTH, T. F. W. 1944. Studies on the igneous rock complex of the Oslo region. II. Systematic, petrography of the plutonic rocks. Skr. Norsk. Vitensk.Akad. Oslo I, 9, 1-104. BAYER, R. & LESQUER, A. 1978. Anomalies gravim&riques de la bordure orientale du craton ouestafricain: g6om&rie d'une suture pan-africaine. Bull. Soc. gkol. Fr. 20, 863-76. BERTRAND, J. M. L. & DAVISON, I. 1981. Pan-African granitoids emplacement in the Adrar des Iforas mobile belt (Mali). A Rb-Sr isotope study. Precambr. Res. 14, 333-62. BEST, M. G. 1975. Migration of hydrous fluids in the upper mantle and potassium variation in calcalkalic rocks. Geology, 3, 429-32. BLACK, R. 1963. Note sur les complexes annulaires de Tchouni-Zarniski et de Gour6 (Niger). Bull. Bur. Rech. geol. minikre, 1, 31-45. -1984. The Pan-African event in the geological framework of Africa. Pangea, 2, 8-16. , CABY, R., MOUSSINE-POUCHKINE,A., BAYER, R., BERTRAND, J. M. L., BOULLIER, A. M., FABRE, J. & LESQUER, A. 1979. Evidence for Precambrian plate tectonics in West Africa. Nature, Lond. 278, 223-7.
---,
JAUJOU, M. & PELLATON, C. 1967. Notice explicative de la carte g6ologique de FAir ~tl'6chelle 1/500 000. Dir. Mines Gkol., Niamey, Niger --, LAMEYRE, J. & BONIN, B. 1985. The structural setting of alkaline complexes J. Afr. Earth Sci. 3, 5-16. BONIN, B. 1980. Les complexes acides alcalins anorog~niques continentaux: l'exemple de la Corse. ThOse Etat, Universit~ de Paris VI (unpublished). BOULLIER, A. M., DAVISON, I., BERTRAND, J. M. L. & COWARD, M. 1980. L'unit~ granulitique des Iforas: une nappe de socle d'fige Pan-Africain pr6coce. Bull. Soc. gbol. Ft. 20, 877-82. BOWDEN, P. & TURNER, D. C. 1974. Peralkaline and associated ring-complexes in the Nigeria-Niger province, West Africa. In: SORENSEN,H. (ed.) The alkaline rocks, pp. 330-5 I. Wiley, London. BROWN, G. C. 1977, Mantle origin of cordilleran granites. Nature, Lond. 265, 21-4. BURKE, K. & DEWEY, J. F. 1972. Orogeny in Africa. In: DESSAUVAGIE, T. F. J. & WHITEMAN, A. J. (eds.) African Geology, pp. 583-608. Ibadan University Press, Ibadan. CABY, R. 1970. La chaine pharusienne dans le nordouest de l'Ahaggar (Sahara central, Alg6rie); sa place dans l'orogen~se du Prbcambrien sup~rieur en Afrique. Thbse Etat, Universit6 de Montpellier. 1980. Les nappes pr6cambriennes du Gourma dans la cha~ne pan-africaine du Mali. Comparaison avec les Alpes occidentales. Rev. G~ogr.phys. Gkol. dyn. 21,365-76. 1981. Associations volcaniques et plutoniques pr&ectoniques de la bordure de la chaine panafricaine en Adrar des Iforas (Mali): un site de type arc-cordill~re au Protbrozoique sup6rieur. llth Colloq. on African Geology, Milton Keynes, p. 30. & ANDREOPOULOS-RENAUD, U. 1985. Etude p&rostructurale et g6ochronologique d'une m&adiorite quartzique de la chaine pan-africaine de l'Adrar des Iforas (Mall). Bull. Soc. Gbol. Fr., Sbr 8, l, 899-903.
400
J. P. Likgeois and R. Black
, BERTRAND, J. M. L. & BLACK, R. 1981. PanAfrican ocean closure and continental collision in the Hoggar-Iforas segment, central Sahara. In: KRONER, A. (ed.) Precambrian Plate Tectonics, pp. 407-34. Elsevier, Amsterdam. CABY, R., LII~GEOIS,J. P., DOSTAL, C., DUPUY, C. & ANDREOPOULOS-RENAUD, U. 1986. The Tilemsi arc and the Pan-African suture zone in Northern Mali. Int. Field Conf. Proterozoic Geology and Geochemistry (IGCP 215-17), Colorado, USA, p. 88. CAHEN, L., SNELLING, N. J., DELHAL, J. & VAIL, J. R. 1984. The Geochronology and Evolution of Africa, 512 pp. Clarendon Press, Oxford. CHIKHAOUI, M. 1981. Les roches volcaniques du Prot6rozoique sup6rieur de la chaine pan-africaine du NW de l'Afrique (Hoggar, Anti-Atlas, Adrar des Iforas). Caract~risation g6ochimique et min6ralogique--implications g6odynamiques. Thkse Etat, Universit~ de Montpellier. DICKINSON, W. R. & HATHERTON, T. 1967. Andesitic volcanism and seismicity around the Pacific. Science, 157, 801-3. DUCROT, J., DE LABOISSE,H., RENAUD,U. & LANCELOT, J. 1979. Synth~se g6ochronologique sur la succession des 6v~nements magmatiques pan-africains au Maroc, dans l'Adrar des Iforas et dans l'est du Hoggar. 10~ Colloq. de Gdologie d'Afrique, Montpellier, p. 40. nupuY, C., DOSTAL,J. & VERNIERES,J. 1978. Genesis of volcanic rocks related to subduction zones, geochemical point of view. Bull. Soc. gbol. Fr. 19, 1233-44. DUYVERMANN, H. J., HARRIS, N . B. W. & HAWKESWORTH, C. J. 1982. Crustal accretion in the PanAfrican: Nd and Sr evidence from the Arabian shield. Earth planet. Sci. Lett. 54, 315-26. EATON, G. P. 1982. The Basin and Range province, origin and tectonic significance. Annu. Rev. Earth planet. Sci. 10, 409-40. FABRE, J. 1982. Pan-African volcano-sedimentary formations in the Adrar des Iforas. Precambr. Res. 19, 201-14. , BA, H., BLACK, R., CABY, R., LEBLANC, M. & LESQUER, A. 1982. La chaine pan-africaine, son avant-pays et la zone de suture au Mali. Notice explicative de la carte gdologique et gravimdtrique de l'Adrar des Iforas au 1/500 000, Bamako, Mali. FYFE, W. S. & MCBIRNEY, A. R. 1975. Subduction and the structure of andesitic volcanic belts. Am. J. Sci. 275A, 285-97. HARRIS, N. B. W. 1982. The petrogenesis of alkaline intrusives from Arabia and northeast Africa and their implications for within-plate magmatism. Tectonophysics, 83, 243-58. -1985. Alkaline complexes from the Arabian shield. J. Aft. Earth Sci. 3, 83-8. HAWKESWORTH, C. J., NORRY, M. J. RODDICK, J. C. & BAKER, P. E. 1979. 143Nd/14*Nd, 87Sr/86Sr, and incompatible element variations in calc-alkaline andesites and plateau lavas from South America. Earth planet. Sci. Lett. 42, 45-57. ]NNOCENTI,F., MAZZUOLI,R., PASQUARE,G., RADICATI DI BROZOLO F. & VILLARI, L. 1982. Tertiary and
Quaternary volcanism of the Erzurum kars area (Eastern Turkey). Geochronological and geodynamic evolution. J. Volcanol. geotherm Res. 13, 22340. JACOBSON, R. R. E., MACLEOD, W. N. & BLACK, RI 1958. Ring-complexes in the Younger Granite province of northern Nigeria. Mem. geol. Soc. Lond. 1, 1-72. JAMES,D. E. 1971. Plate tectonic model for the evolution of the Central Andes. Geol. Soc. Am. Bull. 82, 3325-46. KENNEDY, W. Q. 1964. The structural differentiation of Africa in the Pan-African (_+ 500 m.y.) tectonic episode. Res. Inst. Afr. Geol. Univ. Leeds, 8th Annu. Rep., pp. 48-9. LANEYRE, J., MAROT, A., ZININE, S., CANTAGREL, J. M., DOSSO, L. & VIDAL, P. 1976. Chronological evolution of the Kerguelen islands syenite-granite ring-complex. Nature, Lond. 263, 306-7. LANCELOT, J. R., BOULLIER, A. M., MALUSKI, H. & DUCROT, J. 1983. Deformation and related radiochronology in a late Pan-African mylonite bearing shear zone, Adrar des Iforas, Mali. Contrib. Mineral. Petrol. 82, 312-26. LE FORT, P. 1981. Manaslu leucogranite: a collision signature of the Hymalaya. A model for its genesis and emplacement. J. geophys. Res. 16, 10 545-68. LIf/GEOIS, J. P. & BLACK, R. 1984. P6trographie et g6ochronologie Rb-Sr de la transition calcoalcaline--alcaline fini-pan-africaine dans l'Adrar des Iforas (Mali): accretion crustale au Pr6cambrien sup6rieur. In: KLERKX, J. & MICHOT, J. (eds.) Gbologie Africaine--African Geology, pp. 115-45. Mus6e Royal de l'Affique Centrale, Tervuren. - - , BERTRAND,H., BLACK, R., CABY, R. & FABRE, J. 1983. Permian alkaline undersaturated and carbonatite province, and rifting along the West African craton. Nature, Lond. 305, 42-3. LOISELLI~,M. C. & WONES, n . R. 1979. Characteristics and origin of anDrogenic granites. Abstracts 92nd G.S.A. and Annu. Meet. 11,468. LY, S., LESQUER,A., BA, H. & BLACK,R. 1984. Structure profonde du batholite occidental de l'Adrar des Iforas (Mali) : une synth~se des donn~es gravim~triques et g6ologiques. Rev. Gbogr. phys. Gbol. dyn. 25, 33-44. MOUSSINE-POUCHKINE, A. & BERTRAND-SARFATI, J. 1978. Le Gourma: un aulacog~ne du Pr~cambrien sup&ieur? Bull. Soc. gbol. Ft. 20, 851-6. PEARCE, J. A. HARRIS, N. B. W. & TINDLE, A. G. 1985. Trace element discrimination diagrams for the tectonic interpretation of granitic rocks. J. Petrol. 25, 956-83. PETERMAN, Z. E., HEDGE, C. E. & TOURTELOT, H. A. 1970. Isotopic composition of strontium in seawater throughout Phaneorozoic time. Geochim. cosmochim Acta, 34, 105-20. SAUVAGE, J. F. & SAVARD, R. 1985. Les complexes alcalins sous-satur6s ~ carbonatites de la r6gion d'In Imanal (Sahara malien): une presentation. J. Afr. Earth Sci. 3, 83-8. SHAW, D. M. 1968. A review of K - R b fractionation trends by covariance analysis. Geochim. cosmochim Acta, 32, 573-601.
Alkaline magmatism in the IJbras province THORPE, R. S., POT'I'S, P. J. & FRANCIS, P. W. 1976. Rare earth data and petrogenesis of andesite from the north Chilean Andes. Contrib. Mineral. Petrol. 54, 65-78. TROMPETTE, R. 1980. La chaine pan-africaine des Dahomeyides et le bassin des Volta (bordure SE du craton ouest-africain). In: BESSOLES, B. (ed.) Gbologie de l'Afrique ."la Chaine Pan-Africaine, Zone Mobile d'Afrique Centrale (Partie Sud) et Zone Mobile Soudanaise, Mem. Bur. Rech. G~ol. Min. Fr. 92, 11. VIDAL, P., DOSSO, L., BOWDEN,P. & LAMEYRE,J. 1979. Strontium isotope geochemistry in syenite-alka-
401
line granite complexes. /n: AHRENS, L. H. (ed.) Origin and Distribution of the Elements, pp. 223-31. Pergamon Press, Oxford. WEIS, D. & LII'GEOIS, J. P. 1983. U-Pb whole rock isochron in the Tadhak ring-complex (Mali). Petrogenetic implications. Int. Conf. on Alkaline Ring-complexes, p. 36. Zaria and Jos, Nigeria. WEIS, D., LI~GEOIS,J. P. & BLACK,R. Tadhak alkaline ring complex (Mali) : existence of U-Pb isochrons and 'Dupal' signature 270 Ma ago. Earth planet. Sci. Lett. in press. WILLIAMSON, J. H. 1968. Least-square fitting of a straight line. Can. J. Phys. 46, 1845-7.
J. P. LII~GEOIS, Service de G6ochronologie, Mus+e Royal de l'Afrique Centrale, 1980 Tervuren, Belgium. R. BLACK,Laboratoire de P&rologie, CNRS-UA728, Universit6 P. et M. Curie, 75230 Paris Cedex 05, France.
The petrology, chemistry and crystallization history of the Velasco alkaline province, eastern Bolivia C. J. N. Fletcher & B. Beddoe-Stephens S U M M A R Y: The Velasco alkaline province consists of an early volcanic suite, a series of 14 interfering circular and elliptical alkaline plutons, numerous dykes and a silicified carbonatitic complex. The alkaline rocks, excluding the dykes, are restricted to a narrow NE-trending belt that has been traced for nearly 80 km. The plutons can be divided into two groups defined by the silica-oversaturated or silicaundersaturated character of the rock units within each pluton. It has been possible to deduce from textural relationships and whole-rock and mineral chemistry that the two groups can be derived from a single pulaskitic magma type, in turn producible by minor fractionation of a mantle-derived trachytic parent. Sr isotopic data indicate that the trend from pulaskite through quartz syenite to granite was developed by progressive crustal contamination due to slow rates of emplacement. In contrast, the undersaturated trend towards foyaite shows no evidence of crustal influence, the plutons probably being emplaced at a greater rate. These plutons exhibit marked sub-vertical crystal layering parallel to the intrusion walls suggesting that crystal-fluid processes were responsible for the fractionation trend. A separate strongly undersaturated magma batch is represented in the Velasco province and occurs as a melasyenitic dyke-swarm. Some dykes are strongly enriched in rare-earth elements (REE), concentrated in the mineral britholite, and indicate interaction with a hydrous phase enriched in REE.
Introduction The Velasco alkaline province was first discovered in 1977 during a regional mapping and mineral exploration programme undertaken over that part of the Brazilian Precambrian shield lying within eastern Bolivia (Fig. 1). The initial identification of the province was made on Landsat imagery which showed a number of circular structures and a large isolated hill (Cerro Manomo (Fig. 1)) aligned in a north-easterly direction. Subsequent field mapping revealed that the circular structures were formed by a series of alkaline plutons and that Cerro Manomo was a silicified carbonatitic complex enriched in rare-earth elements (REE) (Fletcher & Litherland 1981; Fletcher et al. 1981). R b - S r whole-rock isochron ages (140___6 Ma and 143__+4 Ma) and K - A r mineral ages (134-142 Ma) have shown that the alkaline plutons were emplaced during Late Jurassic or very early Cretaceous time into Precambrian gneisses which have yielded an age of 1366 Ma (Darbyshire & Fletcher 1979). The province lies close to the geographic centre of S America and over 1500 km from the eastern continental margin. It has been suggested that the alkaline magmatism was generated along a deep crustal fracture related to a possible triple junction which was initiated by the rifting of the S America-Africa plate (Fletcher & Litherland 1981.) The petrography and chemistry of rocks within
the main south-western part of the Velasco complex are described in this paper and postulates regarding their genesis are discussed.
Geological setting The Velasco alkaline province consists of an early volcanic suite, a series of 14 interfering ring plutons, numerous dykes and a silicified carbonatitic complex. The alkaline rocks, excluding t h e dykes, crop out over 460 km 2 but are restricted to a narrow belt which has been traced for nearly 80 km (Fig. 1). Little evidence is left for what was presumably a fairly extensive early volcanic episode which heralded the intrusion of the alkaline plutons. Volcanics are only found in situ along the Rio Paragua, although small xenoliths are found within the later oversaturated plutonic rock types. The volcanics consist of agglomerates, brecciated and welded tufts and porphyritic trachytes. The majority of the plutons have circular outcrop patterns with diameters from 3 to 8 kin. Over half display a concentric distribution of the major rock units defining ring structures having the form of truncated upright cones or cylinders. It has been possible to show from their intersecting relationships and the distribution of faulting that the plutons were intruded along a major fault which was active intermittently throughout the plutonic activity (Fletcher & Litherland
1981).
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 403-413.
4o3
C. J. N. Fletcher & B. Beddoe-Stephens
404
60~
I CERRO iiiiiSHIELD,iiiiiiii
1s~
!!i B R A Z I L I A N
MANOMO
,
.~
SHIELD'!!!!
+ + +
+i|+++~1
LEGEND Laterite
+
,;,.,,.._.
jj/
+ + + + ~/-~ -
ManomOcomplecarbonatitic x Oversaturated alkaline plutons Undersaturated alkaline plutons
1.++++ + +
+
+
+
~
t
VELASCO PROVINCE
TE RTIARY
t CRETACEOUS TO LATE JURASSIC
Foyaite Volcanics Gneiss and granite
PRECAMBRIAN
Fault Scale
5 ' ....
0 '
5 '
10 kilometres i
Geological contact ~,~
Tren~ lines of gneissic foliation
FIG. 1. Location and generalized geology of the Velasco alkaline province. The plutons and surrounding gneisses are intruded by numerous alkaline dykes, the majority of which range in width from 1 to 3 m. They have been observed up to 30 km from the nearest pluton.
Petrology Early volcanic suite The agglomerates are regarded as being the remnants of a volcanic neck and contain angular
blocks of hydrothermally altered tufts in a siliceous matrix. The surrounding tufts contain lapilli-sized clasts of trachytic lava, spherulitic glass and microbreccias. Welded tufts exhibit a parataxitic texture that dips away from the summit of the hill. The porphyritic trachyte is the most important unit within the volcanic suite as it could provide an indication of parental magma compositions relevant to the genesis of the plutonic rocks. These lavas contain phenocrysts of perthite, plagioclase (An12_60) and augite which is often
Petrology of the Velasco alkaline province replaced by amphibole and biotite. The matrix of these rocks consists of fine- to medium-grained feldspar with some hornblende and occasional biotite. Minor quartz may be present.
Plutonic rocks These can be divided into two groups: silicaundersaturated plutons containing pulaskite and foyaite, and silica-oversaturated plutons with quartz syenite and granite. A slightly oversaturated syenite, nordmarkite, can occur in both suites although it has a strong spatial coherence with undersaturated rock types. Neither within individual plutons nor between plutons is there any indication of a compositional development between the undersaturated and oversaturated rock types. The order of intrusion of the oversaturated and undersaturated plutons is random with respect to both time and space, suggesting distinct modes of genesis prior to intrusion of the various melts. The undersaturated rocks show textures that can be described using the cumulus, adcumulus and post-cumulus terminology of Wager & Brown (1968), although there is no evidence for crystal settling in the Velasco plutons. The non-genetic use of these terms (Irvine 1982) is implied. This suite is composed predominantly of pulaskite (i.e. 0-5% nepheline) which occupies about 24% of the total exposure of the plutonic complex. Foyaites, with up to 35% nepheline, only occupy about 5% and occur either as pods within pulaskite or as the cores to ring plutons containing pulaskite (Fig. 1). These rocks both show cumulus tabular perthite, whereas cumulus nepheline is only seen in the foyaites. In contrast, cumulus augite and aegirine-augite is restricted to the pulaskites compared with inter-cumulus aegirines characteristic of the foyaites. Amphibole is common in the pulaskites, coexisting with or in place ofpyroxene, but rare in foyaite. Biotite is common throughout. Sphene, titanomagnetite, zircon and apatite occur as accessory phases with melanite and sodalite evident in some foyaites. Nepheline is frequently replaced by cancrinite, gieseckite or analcime. The rock type nordmarkite, whilst oversaturated and containing minor quartz, is very similar to pulaskite except that igneous lamination of the feldspars is less apparent. Amphibole is the main mafic phase. This rock occupies all of one pluton or the outer zones of others. The oversaturated plutons are composed of quartz syenite and granite, the difference being the quartz content. Mafic phases are hornblende and biotite, although in some cases the granites
4o5
show development of riebeckite and aegirine, occasionally completely replacing hornblende and biotite. These riebeckite granites form the whole of the most northerly saturated pluton and occur as selvedges within biotite-hornblende granites adjacent to undersaturated intrusions. The riebeckite granites also exhibit the rare development of astrophyllite, baddeleyite, molybdenite and cassiterite along with eudialyte in aegirine-rich veins. These features suggest that the riebeckite granites formed by a process akin to fenitization. A leucocratic granite with essentially no mafic phases, but with accessory zircon and magnetite, occurs as the matrix to breccias forming at the contacts of some plutons with the Precambrian country rocks. This rock is termed an aplogranite.
Dykes Over 100 dykes associated with the Velasco province have been recorded, ranging in composition from microfoyaites to aplites. However, a distinct melasyenitic suite of dykes has been studied in more detail, along with a variety enriched in REE. The normal melasyenites contain zoned augites/aegirine-augites, alkali feldspar and hornblende phenocrysts. They are nepheline bearing. The REE-enriched dyke-rock contains abundant aegirine with feldspar and biotite. Extensive replacement of pyroxene by the C a - R E E silicate britholite has occurred with interstitial development of mesolite.
Rock chemistry 22 samples from the Velasco province were analysed for major and trace elements (Table 1). All the samples, except C5-6, formed part of a radiometric dating programme (Darbyshire & Fletcher 1979) and analyses represent splits from 10 kg of homogenized bulk sample.
Major-element variations The majority of the rocks from the Velasco province are peralkaline ( ( N a + K ) / A I > I ) in contrast with the surrounding Precambrian granites which are calc-alkaline. It should also be noted at this stage that, despite the field and petrographical distinction between the biotitehornblende granites and the riebeckite granites, there is no significant chemical difference between them although the latter exhibit minor acmite in the norm. The large-scale desilication or sodium enrichment that characterizes other fenitized complexes (McKie 1966) is absent.
C. J. N. Fletcher & B. Beddoe-Stephens
406
0 O< O0 O0
~
0
m
oo "~t" O0 , ~
c',l~ er~
on
o~'V= r~
. . . .
~
.
-
-
.
~
-~o-
--N
o o
K
~-~
c~
1"4
~N m ~
.
.~
.~
.~
.
.m
.
. ~7 ~
9
. ~
. ~
. ~
.
. ~
9
~N
t'N
..
.
~)
.'~ ~
~.~ N
m
~
kJ .
.
~
.
~
.
.
~
.
.
.
r
r
O~ I ~
r
I~ 0
~
r
--
on ,-- m
.
x O t'xl A
~
~
--~ o n
--
.-~ t"q
~o
~
c~ <
. ~ ~
o
o
~
Petrology of the Velasco alkaline province Traditional Harker diagrams (Fig. 2) illustrate variations amongst the plutonic, volcanic and dyke rocks. The plutonic rocks cover a wide range of SiO2 content reflecting the spread from extremely silica saturated to undersaturated magma types. These two suites are spatially separate, as stated above, and two trends can be drawn which meet at about 62%-64~ SiO2, which corresponds to a pulaskitic-nordmarkitic composition sitting on the division between quartz- and nepheline-normative compositions. Thus a magma of this composition can be seen as parental to both trends.
Trachyte 9< RE-Melasyenite dyke 9 Melesyenite dykes 9 Foyeite 9 Pulaskite
~7
2 TiO2
AI203
9
1
20
Nordmarkite Q u a r t z syenite 9 Biotite granite | Riebeckite granite 9 Aplogranite [] i~
~ 9.
9
15
~'~'~oo~o~. *
101
=
Fo=O3 Total
Fe
4
Ii-----1.1._~__~Ai L][~_D
0
u
21- I,
The trend from quartz syenite to aplogranite is well defined and linear, and could be viewed simply as a trend towards the thermal minimum in the granite system (Fig. 3). However, isotopic evidence discussed below, together with the occurrence of aplogranite as a probable countryrock partial melt, indicates progressive crustal contamination as a more likely process. In contrast, the pulaskite to foyaite trend involves a decrease in silica and a strong increase in Na20 that could not be due to crustal-rock incorporation. The fact that the foyaites cluster around a thermal minimum in the undersaturated portion of the residua system (Fig. 3) can be ascribed to crystal-liquid differentiation. The trachytic volcanic rocks plot in a distinct field but with a similar range of silica to the pulaskitic-nordmarkitic field and similarly straddle the plane of critical silica undersaturation. They are slightly more primitive than the pulaskites, in view of the higher Mg, Fe and Ca, and could represent magmas parental to the pulaskites. The normal melasyenites form a separate group with low silica values, but enhanced Mg, Fe and lower Na compared with the plutonic rocks. They exhibit a trend towards an evolved undersaturated composition similar to that of foyaite, but this is distinct from the plutonic trend and suggests that the melasyenites derive from a separate strongly undersaturated magma batch produced at depth. The major-element composition of the rare-earth (RE) melasyenite has clearly been affected by hydrothermal alteration, particularly evidenced by its K20 depletion.
~7
MgO 0
~
407
II
9
Minor-element variations
== m
Recognition of systematic minor-element trends has proved difficult, especially for those elements Foyaite- Pulaskite with concentrations near detection limits. It is rTrend Pulaskite -Aplogranite CaO therefore proposed to restrict discussion to K - R b and Nb-Ti relationships for which comparative data are readily available. I 9 i~(-~ " ~ - ) ~ ~i~ I Since there is an overall geochemical coherence 12between K and Rb variations, the K/Rb ratio 9Na20 can prove useful in formulating petrogenetic 6models. The K/Rb ratio shows a wide variation 3in the rocks of the province as shown in Fig. 4. 08 i i The oversaturated trend towards aplogranite defines a trend of decreasing K/Rb with higher K20 silica values, which is consistent with a mixing 4~ ~ model between a pulaskitic magma and a country2 9 rock partial melt. As has been suggested by Shaw , , 0-~ 50 60 70 8b (1970), the breakdown of biotite gneiss can lead Si02% to the formation of an aqueous fluid phase with FIG. 2. Harker variation diagrams for the Velasco rocks. low K/Rb. 8-
C. J. N. Fletcher & B. Beddoe-Stephens
408
Qz
QUARTZ
ALKALI FELDSPARss
\
Or LEUCITE
Le
N E PH ELI N E ss
Ne
'
KALSI LITE Ss
~
Ks
FIG. 3 Plutonic rocks plotted in the system Ne-Ks-Qz (Petrogeny's Residua System). Phase boundaries are for PH2o= 1 kb. m' and m" are the thermal minima in the two portions of the diagram. They are linked by thermal valleys (broken lines) to the minimum on the Ab-Or join. The arrows give generalized crystallization paths. Symbols as in Fig. 2. (From Hamilton & MacKenzie 1965).
400-
,~.ob
v 9
9
K/.b ?~/I(
9
A;:~'*'"
l.m~. 50
- "o.',,
60
7'0 SiOz%
FIG. 4. K/Rb versus SiO, for the Velasco rocks. Symbols as in Fig. 2. The trend from pulaskite to foyaite involves on average a decrease in K/Rb. Since this trend is largely due to alkali feldspar fractionation (Fig. 3) together with some accumulation of nepheline, this is expected as Heier (1966) has demonstrated that K/Rb in alkali feldspar is higher than in the coexisting melt. Complementary to this is the probability that K/Rb in nepheline is lower than in alkali feldspar as indicated by data from the Kangerdlugssuaq intrusion (Kempe & Deer 1970) and Stjern6y syenites (Heier 1966).
K/Rb ratios for the trachytes overlap those of the pulaskites, whereas K/Rb ratios of the normal and RE melasyenites overlap with, or are lower than, those of the foyaitic rocks. In undersaturated alkaline residual liquids Nb forms large complex ions which restrict the crystallization of Nb-rich phases, and this can be used as an indicator of differentiation. In a NbTi plot (Fig. 5) the foyaites are enriched in Nb relative to the pulaskites with Ti relatively constant, indicating the removal of sphene or ilmenite during fractionation. The oversaturated rocks define a rather scattered trend, but the quartz syenites and granites fall between the pulaskite-trachyte fields and the aplogranite and thus do not invalidate the progressive crustal contamination hypothesis suggested above. Similarly, other trace elements, with the exception of Y, fulfil this criterion.
Minor elements #7 the rare-earth melasyenite dyke The REE content of the analysed RE melasyenite is 9.5~ and it is also enriched in Be, V, Sr, Y, Sn, Ba, Th and U compared with nearly all other Velasco rocks. Beryllium is concentrated in the residual liquids of magmatic rocks and can be incorporated in minerals with high-valency cations (e.g. Ti, Zr, Nb, REE) and low-valency
Petrology of the Velasco alkaline province Volcanic field ~ / "
~'~ \ 9
Undersaturated Trend
~--. []y)
Ti
9
9
(ppm)
[]
1ooo -
loo
.
.
i
.
o
~
.
.
.
9
.
~
~o
Nb (ppm)
FIG. 5. Semi-logarithmic plot of Ti versus Nb for the Velasco rocks. Symbols as in Fig. 2. anions (e.g. F, OH) characteristic of alkaline rocks (Taylor 1965). Thus the high Be content in these dykes suggests the hypothesis that they originated from residual liquids, either from the crystallizing undersaturated plutons or from the fractionation of normal melasyenitic dyke magmas, although it appears that the REE were introduced by later hydrous fluids. The concentration of V can largely be accounted for by its incorporation in aegirine. Electron microprobe data indicate values from 600 to 1200 ppm. Gomes et al. (1970) and Kempe & Deer (1970) also note the concentration of V into alkali pyroxene in preference to coexisting phases. The anomalous concentrations of other elements are largely a result of extreme fractionation
[]
INITIAL
87Sr/86Sr
I
I =,'+
I II
I
~'6
do
66
The initial 87Sr/86Sr ratios for many of the samples analysed were determined during the course of radiometric dating (Darbyshire & Fletcher 1979). Figure 6 illustrates these data plotted against SiO2. Two groupings are immediately evident: the first has an essentially constant initial ratio in the range 0.704--0.705, and the second shows a spread from 0.706 to 0.709. In the latter case the errors on the determinations make it impossible to infer correlation with SiO2. The two groups correspond to the suites already defined on chemical and field considerations, i.e. the undersaturated and oversaturated associations. These data are strongly suggestive of an undersaturated suite of rocks that have been mantle derived and suffered very little or no crustal contamination during fractionation and emplacement. In contrast, the oversaturated suite, including some pulaskite, must have suffered considerable interaction with crustal material since its close spatial and geochemical coherence with the undersaturated rocks precludes derivation from a different mantle source. There is a tentative increase in the STSr/86Sr ratio with SiO2 in the oversaturated suite which could indicate that simple mixing between pulaskite and aplogranite can explain the spectrum of granite compositions, but this is inconclusive.
Pyroxene and amphibole relations in the undersaturated rocks
9 0.707-
i I
combined with the ease with which some elements form complexes in hydrous fluids. It is probable that most of these elements are contained within britholite and apatite. Strontium isotopes
9
i
409
"6
SiOz
FIG. 6. Initial Sr87/Sr86 ratios (with error bars) versus SiO2 for the Velasco plutonic rocks. Symbols as in Fig. 2.
Pyroxenes and amphiboles from the undersaturated plutonic rocks, melasyenites and volcanics were analysed by microprobe in order to compare with other suites and to complement whole-rock chemical data in developing a relationship between these rocks. Fe 3 + in pyroxene was estimated by assuming that exactly four cations per six oxygen atoms are present (cf Larsen 1976), and the assumption of 13 cations in the Y + Z sites per 23 oxygens allowed a similar estimate to be made for amphibole (cf Leake 1978). Analyses are available from the authors on request. Pyroxenes have been plotted on a conventional N a - M g - ( F e 2+, Mn) diagram in Fig. 7. Overall there is a variation from augite through aegirineaugite to aegirine, as defined by Deer et al. (1978),
C. J. N. Fletcher & B. Beddoe-Stephens
4IO
Na
9 Foyaite 9 Pulaskite
v
9 e"
Trachyte Melasyenite R E-Melasyenite
I I I
C16-2 HJ41-3
L
r" =ms
C37-1 ---~~ H J S - - 1
Mg
Fa 2+ + M n
FIG. 7. Na-Mg-(Fe 2+, Mn) variation diagram for the pyroxenes. Solid tie-lines link analyses within one crystal and the arrows indicate core-to-rim variation. The inset shows other alkaline pyroxene trends for comparison: a) Itapirapua, Brazil (Gomes et al. 1970); b) Uganda (Tyler & King 1967); c) South Q6roq (Stephenson 1972); d) Igdlerfigssalik, Greenland (PoweU 1978); e) Ilimaussaq, Greenland (Larsen 1976).
in common with many other suites. The Velasco data define a trend intermediate to the S Q6roq and Itapirapua trends (see Fig. 7, inset) in showing only moderate enrichment towards hedenbergite. The least evolved pyroxene compositions, as indicated by the least Na/Ca and Fe 2 +/Mg, are those occurring as phenocrysts in the volcanic trachytes. These fall into the augite category and only exhibit a limited trend parallel to the diopside-hedenbergite join. The total extent of zoning in these pyroxenes may be disguised by marginal alteration to amphibole. Within the pulaskites and foyaites, pyroxenes fall into three groups that can conveniently be termed augite, aegirine-augite and aegirine, although these slightly overlap the fields defined by Deer e t al. (1978). Three of the pulaskites analysed (Fig. 7) exhibit pale-green to colourless sodian augites, slightly more evolved than those in the trachytes. Single-crystal zoning is limited but shows enrichment in Na. These augites also show a reaction relationship with a brown-green ferroan pargasite or pargasitic hornblende. This is illustrated in Fig. 8, which shows that chemical evolution across this reaction is continuous in terms of the Na/Ca ratio, a fact noted by Larsen
(1976) in the Ilimaussaq intrusion. The stage at which pyroxene can be replaced by amphibole will depend on the H20 activity. Alternatively, the low to zero Fe 3 + estimated for these amphi-
2.
"1" 9
amphibole pyroxene
Fe2§
/////!// 1
lI i/
I
II I I ii ~ HJ5-1fj, //i//
~tlL~lo"
-
4
GJ21-
~GJ209 o
I 0
1 Na/Ca
FIG. 8. Fe z § versus Na/Ca for coexisting pyroxenes and amphiboles in the pulaskites.
Petrology of the Velasco alkaline province boles indicates lowfO2, consistent with Powell's (1978) interpretation that limited pyroxene zoning and the appearance of amphiboles with low Na/Ca is due to sharply decreasing fO2 as the temperature falls. In contrast, pulaskite H J40--4 contains pyroxenes of aegirine-augite composition but no amphibole. Foyaitic pyroxenes are predominantly interand post-cumulus aegirines with limited zoning, though occasional subhedral grains are aegirineaugite. This indicates a degree of disequilibrium preserved from a pulaskitic precursor. There are no amphiboles coexisting with these pyroxenes, and hence sodic amphiboles such as kataphorite and arfvedsonite, characteristic of evolved syenites in other provinces (e.g. Larsen 1976; Powell 1978), do not appear to have been stabilized. The pyroxene data discussed so far indicate that an evolving magma could precipitate first trachytic-type augites followed by pulaskitic and then finally foyaitic aegirine-augites and aegirines. This is consistent with suggestions made above based on whole-rock data. With regard to the dykes there is also a range of pyroxene compositions that parallel the pulaskite-foyaite trend, except that within the normal melasyenites the phenocrysts show zoning toward aegirineaugite with a higher hedenbergite component than that seen in the plutonics. Aegirines in the RE melasyenite are acmitic and overlap with aegirines in foyaites. However, the former are
411
distinct in other ways and contain significantly more Ti, Mn and Zr. Thus, again in accord with whole-rock data, the melasyenites, including possibly the RE varieties, appear to have been derived from a significantly different magma batch.
Petrogenesis In developing a coherent petrogenetic scheme for the formation of the Velasco province, in addition to considering whole-rock and mineral chemistry, isotopic data and petrography, the following points must also be included. 1 The trachytic volcanics were the earliest expression of magmatism. 2 There is a division into oversaturated and undersaturated plutons which, however, were intruded randomly in space and time. 3 With regard to the undersaturated plutonic suite (a) foyaite is much less abundant than pulaskite with which it is always associated, (b) sub-vertical crystal layering and cumulus textures are restricted to this suite and (c) cumulus nepheline occurs only in the foyaites. Figure 9 shows all the rock compositions plotted on an alkali-silica diagram along with analysed
F P N M G T 9 -I-
Nepheline
[]
Or 16
Foyaite Pulaskite Nordmorkite M elasyenite
Granite / Quartz Syenite Trachyte
Feldspar in volcanics Feldspar in melasyenite dykes
Ne-norm
+B~/
Wt. % NazO + K20
~Qz ..... Ab
gl i Pargasites
~<~/
/
/S
. ~y
'0 Wt. % SiO 2
FIG. 9. Total alkalis versus SiO2 plot showing whole-rock fields and mineral compositions. For further discussion see text.
4IZ
C. J. N. Fletcher & B. Beddoe-Stephens
and other relevant mineral compositions. As noted earlier the pulaskite-nordmarkite and trachytic fields both straddle the plane of critical silica undersaturation which can form an effective thermal divide. This gives the potential for trachytic magmas to differentiate towards thermal minima on either side of the divide, shown in Fig. 3. All features point to a trachytic magma as being parental to the plutonic series; Fig. 9 indicates that minor augite plus plagioclase fractionation will produce a range of compositions in the pulaskite-nordmarkite field. Various evidence considered previously indicates foyaite to be derived by differentiation of pulaskite and, as Figs 3 and 8 show, this is largely controlled by alkali feldspar fractionation. The initial part of the trend may be influenced by pyroxene, but in the foyaites pyroxene is entirely post-cumulus and reflects rather than controls residual liquid compositions. Some foyaites with abundant cumulus nepheline are unlikely to have been produced solely by alkali feldspar separation of a pulaskite magma. This is apparent in Fig. 3 where foyaite C5 plots well within the primaryphase field of nepheline and must reflect nepheline accumulation. The sub-vertical layering in the plutons precludes crystal settling as a process and must reflect the geometry of the cooling surface and a slow cooling history (McBirney & Noyes 1979). Foyaitic pods within pulaskite may reflect local fluid-dynamical accumulation of suspended nepheline crystals in excess of that predicted by equilibrium crystallization of a fluid on the feldspar-nepheline cotectic. The oversaturated portion of the Velasco plutonic trend also follows a thermally decreasing path towards a eutectic point (Fig. 3); however, other evidence, in particular isotopic, indicates that crustal contamination played a large part in the production of this trend. Similar isotopic evidence in the Gardar province of Greenland indicates that silica-oversaturated rocks were also derived by contamination of mantle-derived magmas (Blaxland et al. 1978). This seemingly selective contamination of some magmas, but not others, may be related to the rate of ascent. At high rates of ascent trachytic-pulaskitic magmas could penetrate quickly to the upper crust, whereas at lower rates partially melted country rock may have time to intermix thoroughly. Another factor that may aid the sharp division between the two suites is the presence of the thermal divide referred to above. Once a magma slips down the undersaturated side of the divide it will prove impossible to mix homogeneously with a low-temperature component on the far side, unless a lot of thermal energy is applied. No such constraints will apply to mixing of, say, a
nordmarkitic magma with a lower-temperature aplogranitic component. During the development of the oversaturated plutonic series and under the influence of closely related undersaturated alkaline intrusive activity, the riebeckite granites formed by mild fenitization of the biotite granites. The textures and mineralogy in these rocks are similar to those in the outer zones of fenite complexes associated with other alkaline-carbonatite bodies, e.g. Chilwa Island (Garson & Smith 1958), Kangankunde (Woolley 1969) and Aln6 Island (von Eckermann 1948). However, broad chemical changes such as desilication and alkali metasomatism are minor at Velasco. The melasyenites have been shown from wholerock chemical and mineralogical data to be distinct from both the volcanics and the plutonics. They are strongly undersaturated and cannot be regarded as deriving from pulaskitic-, foyaitic- or trachytic-type magmas. However, foyaitic-type compositions could be produced by fractionation from the normal melasyenites of the observed phenocryst assemblage (pyroxene, alkali feldspar and pargasite) (Figs 3 and 8). The RE melasyenite could represent such a residual composition subsequently affected by low-temperature reaction with hydrous REE-enriched fluids. Although isotopic evidence clearly shows that mantle-derived magmas must have been parental to the development of the province, there is little evidence of the primary magmas, either as basaltic extrusives or as gabbroic intrusions. In the Gardar province, for example, early and parental basic magmas are well represented as dykes and volcanics (Upton 1974; Blaxland et al. 1978). However, it must be remembered that this paper represents a first look at an inaccessible and poorly exposed province and future work will almost certainly change the picture.
ACKNOWLEDGMENTS:The authors would like to thank N. C. W. Anderson, D. Appleton, L. Ault, A. E. Davies, J. G. Fitton, S. Hobbs, J. E. Robbins, T. K. Smith and B. A. R. Tait for their analytical work on the samples. J. Bennett, J. G. Fitton, N. M. S. Rock and A. R. Woolley read various versions of the manuscript, and the paper has greatly benefited from their comments and suggestions. D. P. F. Darbyshire supplied additional unpublished strontium isotope data. This paper is published with the permission of the Directors of the British Geological Survey (Natural Environmental Research Council) and the Servicio Geologico de Bolivia. It resulted from work carried out whilst one of the authors (CJNF) was on a technical cooperation assignment in Bolivia funded by the Overseas Development Administration, Foreign and Commonwealth Office.
Petrology of the Velasco alkaline province
413
References BLAXLAND,A. B., VAN BREEMEN,O., EMELEUS,C. H. ~,z mineralogy of the Kangerdlugssuaq alkaline intruANDERSON,J.G. 1978. Age and origin of the major sion, east Greenland. Meddr. Gronland, 190(2), 1syenitic centres in the Gardar Province of south 49. Greenland: Rb-Sr studies. Geol. Soc. Am. Bull. 89, LARSEN, L. M. 1976. Ciinopyroxenes and coexisting 231-44. mafic minerals from the alkaline Ilimaussaq DARBYSHIRE, D. P. F. & FLETCHER, C. J. N. 1979. A Intrusion, South Greenland. J. Petrol. 17, 258-90. Mesozoic alkaline province in eastern Bolivia. LEAKE, B. E. 1978. Nomenclature of amphiboles. Am. Geology 7, 545-8. Mineral. 63, 1023-52. DEER, W. A. HOWlE, R. A. & ZUSSMAN,J. 1978. Rock MCBIRNEY,A. R. & NOYES,R. M. 1979. Crystallisation Forming Minerals, Vol. 2A, Single Chain Silicates. and layering of the Skaergaard Intrusion. J. Petrol Longman, London. 20, 487-554. ECKERMANN,H. VON 1948. The alkaline district of Aln6 MCKIE, D. 1966. Fenitisation. In: TUTTLE, O. F. & Island. Sver. geol. Unders. AJh. Ca. 36. GITTINS, J (eds) The Carbonatites, pp. 261-94. FLETCHER, C. J. N. & LITHERLAND, M. 1981. The Wiley, New York. geology and tectonic setting of the Velasco Alkaline POWELL, M. 1978. The crystallisation history of the Province, eastern Bolivia. J. geol. Soc. Lond. 138, Igdlerfigssalik nepheline syenite intrusion, Green541-8. land. Lithos, I1, 99-120. , APPLETON, J. D. WEBB, B. C. & BASHAM,I. R. SHAW, D. M. 1970. Trace element fractionation during 1981. Mineralisation in the Cerro Manomo Caranatexis. Geochim. cosmochim. Acta, 34, 237-43. bonatite Complex, eastern Bolivia. Trans. lnstn STEPHENSON, D. 1972. Alkali clinopyroxenes from Min. Metall., Sect. B, 90, 37-50. nepheline syenites of the South Q6roq centre, GARSON, M. S. & SMITH, W. C. 1958. Chilwa Island. south Greenland. Lithos, 5, 187-201. Mem. geol. Surv. Dep. Nyasaland, 1. TAYLOR, S. R. 1965. The application of trace element GOMES, C. D., MORO, S. L. & DUTRA, C. V. 1970. data to problems in petrology. Phys. Chem. Earth,, Pyroxenes from the alkaline rocks of Itapirapua, 6, 133-213. S~o Paulo, Brazil. Am. Mineral. 55, 224-30. TYLER, R. C. & KING, B. C. 1967. The pyroxenes of the HAMILTON, D. L. & MACKENZIE, W. S. 1965. Phase alkaline igneous complexes of eastern Uganda. equilibrium studies in the system NaA1SiO4 Mineral. Mag. 36, 5-21. (nepheline)-KA1SiO4 (kalsilite)-SiO2-H20. Min- UPTON, B. G. J. 1974. The Alkaline Province of southeral. Mag. 34, 214-31. west Greenland. In: SORENSON, H. (ed.) The HEIER, K. S. 1966. Some crystallo-chemical relations of Alkaline Rocks, pp. 221-38. Wiley, London. nephelines and feldspars on Stjern6y, North WAGER, L. R. & BROWN, G. M. 1968. Layered Igneous Norway. J. Petrol. 7, 95-113. Rocks. Oliver & Boyd, Edinburgh and London. IRVINE, T. N. 1982. Terminology for layered intrusions. WOOLLEY, A. R. 1969. Some aspects of fenitisation J. Petrol. 23, 127-62. with particular reference to Chilwa Island and KEMPE, D. R. C. & DEER, W. A. 1970. Geological Kangankunde, Malawi. Bull. Brit. Mus. Nat. Hist. investigations in east Greenland, part 9. The Mineral. 2, 191-221. C. J. N. FLETCHER*,Overseas Directorate, British Geological Survey, Keyworth, Nottingham NG12 5GG, U.K. B. BEDDOE-STEPHENS, British Geological Survey, Murchison House, West Mains Road, Edinburgh EH9 3LA, U.K. * Present address: British Geological Survey, Bryn Eithyn Hall, Llanfarian, Aberystwyth, Dyfed, U.K.
Tertiary alkaline magmatism in Trans-Pecos Texas Daniel S. Barker S U M M A R Y : Alkaline magmatism in far W Texas extended from 48 to 16 Ma ago, peaking in volume during the interval 37-26 Ma when all the felsic rocks were emplaced. Primitive magmas are rare, if not entirely lacking, but ne- and ol+hy-normative hawaiite and mugearite liquids (Ni < 120 ppm; 100Mg/(Mg + Fe) < 66) erupted through the entire time span. These evolved through ne-normative and q-normative trachytes to phonolite and rhyolite, including peralkaline varieties of both. Mafic q-normative rocks are distinctly subordinate. Transition from a contractional to an extensional regime occurred between 32 and 30 Ma ago, after which Trans-Pecos magmatism was almost exclusively mafic. Major- and traceelement compositions show gradation south-westward to the calc-alkaline rocks of the Sierra Madre Occidental in western Mexico, which were erupted from 35 to 28 Ma ago. This compositional gradient suggests that the alkaline and calc-alkaline rocks were all related to subduction of the Farallon plate. The Trans-Pecos province is not a continental rift, and probably is not a product of back-arc spreading. Peralkaline phonolites and rhyolites in Texas and calc-alkaline rhyolites in Mexico were erupted at the same time and probably in the same stress regime, and their compositional differences must reflect differences in source rocks and conditions of melting.
Location and extent
Early work
The Trans-Pecos province contains all the Cenozoic igneous rocks within an area bounded on the N by an E - W line 12 km N of the New MexicoTexas border, on the E by the Pecos River, and on the W and S by the international border defined by the Rio Grande (Fig. 1). The southwestern and south-eastern limits of the province are arbitrary; compositions of igneous rocks grade to the SW into less alkaline but coeval rocks, and to the SE alkaline rocks like those of the Trans-Pecos province extend for an unknown distance into Mexico (Bloomfield & CepedaDavila 1973; Robin 1974). The northern boundary is also arbitrary, because similar rocks of the same ages do occur farther N in New Mexico (Foord et al. 1983). The north-eastern limit of the province, coinciding with the inland edge of the Basin and Range structural province in Texas, is a natural boundary, because no Cenozoic igneous rocks are exposed within 50 km W of the Pecos River. As discussed later, Basin and Range faulting occurred later than most Trans-Pecos magmatism. The Cenozoic alkaline rocks of the TransPecos province make up one segment of a belt extending along the eastern flank of the N American Cordillera through its entire length from Canada into Mexico (Barker 1974). Several hundred intrusive bodies, as well as less widespread tracts of lavas and pyroclastic rocks, occur within the 63 000 km 2 area of the Trans-Pecos province. Erosion has removed much of the original volcanic cover.
Petrographic descriptions and rock analyses from the Trans-Pecos province were provided by Osann (1896), Clarke (1904) and Meigen & Nachreiner (1925). Lonsdale (1940) presented descriptions and analyses of many intrusive bodies in the Terlingua-Solitario region W of Big Bend National Park (Fig. 1). Meanwhile, Udden (1907) and Baker (1934) had pioneered structural and geomorphic studies of the region. Geological mapping and volcanic stratigraphy began with the work of Eifler (1943, 1951), Goldich & Elms (1949), Erickson (1953), Moon (1953) and McAnulty (1955). From 1948 to 1970 R. K. DeFord and his many graduate students built an excellent data base through their detailed mapping and stratigraphic studies. Albritton & Smith (1965) and King (1965) published classic reports for the U.S. Geological Survey. R. A. Maxwell's monumental map of Big Bend National Park was accompanied by petrological data by Lonsdale and stratigraphic and palaeontological data by Hazzard and Wilson in a comprehensive report (Maxwell et al. 1967). Most of these early efforts were hindered by lack of adequate topographic maps; quadrangle base maps on a scale of 1 : 24 000 began to appear in the 1960s, although some still await publication. The University of Texas Bureau of Economic Geology has completed a series of four geological maps on a scale 1:250 000 covering the province as part of the Geologic Atlas o f Texas (Barnes 1968, 1979a, b, 1982). Wilson (1971) summarized biostratigraphic
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 415--431.
415
4 [6
D . S . Barker
Bc
_
~
Fo
Rio Gronde~kRR
I
I00
Rm
~
N~ IN
PP
Fro. 1. Index map of the Trans-Pecos province, showing the principal localities cited: BB, Big Bend National Park; BC, Buckhorn caldera; BG, Black Gap; BO, Bofecillos Volcano; CA, Cerro Alto; CH, Chinati Mountains caldera; DP, Diablo Plateau; EM, Eagle Mountains caldera; FD, Fort Davis; IN, Infiernito caldera; MA, Mariscal Mountain; MC, Marble Canyon; NP, Nine Point Mesa; PC, Pine Canyon caldera; PP, Paisano Volcano : QM, Quitman Mountains caldera; RM, Rattlesnake Mountain, RR, Rim Rock dyke-swarm; SB, Sierra Blanca; SO, Solitario; SQ, Sierra Quemada caldera; TE, Terlingua; XM, Christmas Mountains.
research on the vertebrate fauna of the TransPecos province. K - A r dating has been published by Maxwell et al. (1967), Dasch et al. (1969), Barker et al. (1977), Daily (1979), McDowell (1979), Parker & McDowell (1979) and Laughlin et al. (1982). Petrological reviews relating magmatism to tectonic setting were attempted by Barker (1977, 1979b). The guidebooks edited by Walton & Henry (1979) and Price et al (1986b) also contain overviews as well as detailed accounts of specific areas and problems. Henry & Price (1984) provide a comprehensive review of Trans-Pecos calderas and associated mineralization.
Igneous rock types Magmatic products in the Trans-Pecos province are conveniently classified in five groups: (1) alkali basalt, hawaiite, mugearite and benmoreite, and their coarse-grained equivalents; (2) phonolite, nepheline syenite and nepheline-normative trachyte; (3) trachybasalt and trachyan-
desite, and their coarse-grained equivalents; (4) quartz trachyte and quartz syenite; (5) rhyolite and granite. In the classification scheme used for TransPecos rocks by Barker (1979a), these types are discriminated (Fig. 2) according to the ThorntonTuttle (1960)differentiation index (sum of normative ab, or, and q or ne), normative plagioclase composition (100 an/(an+ab)) and silica saturation (wt.% q or he). If the differentiation index is less than 75, the rock is classified according to silica saturation and normative plagioclase composition. If the differentiation index equals or exceeds 75, the normative percentage of quartz or nepheline is used without consideration of normative plagioclase (because many of the felsic rocks, both ne- and q-normative, are peralkaline). The discrimination of the mafic silica-oversaturated rocks (trachybasalt and trachyandesite) is difficult, because many analysed samples are so close to silica saturation that a small change in ferric to ferrous iron ratio can change the CIPW norms from ne to q or o l + h y bearing. If all iron in the analyses were treated as ferrous, many but not all of the samples plotting as trachybasalt and
Tertiary magmatism in Trans-Pecos Texas Tholeiite
50
Trochyondesite
o Q
30
I
0o
..__~_o~~
o
Trachybasalt
(~:00~
Ezo
~,,
Quartz trachyte
---to
0 Benmoreite
+
]
Rhyolite
0
417
o
15 c~
Mugearite
30
8
o
oq~ Qo
o~o
~~ 30
o Phonolite
Nbo ~:~ Hawaiite
50
O0
Alkali basalt
I
20
I
40
[
I
~
60
L
75
[
90
I00
Differentiation i n d e x
FIG. 2. Classification of Trans-Pecos igneous rocks (Barker 1979a) using the Thornton-Tuttle (1960) differentiation index, the normative plagioclase composition and, for rocks with differentiation index exceeding 75, the weight per cent of normative quartz or nepheline. For rocks with differentiation index less than 75, compositions above the horizontal zero line are q-normative, and those below are ne- or ol + hy-normative. (O) Alkali belt, (O) metaluminous belt.
trachyandesite would be moved into the hawaiite and mugearite fields. Figure 3 presents generalized maps showing the distribution of intrusive bodies of each of four groups, excluding the mafic silica-oversaturated rocks. Table 1 lists representative major- and trace-element analyses and Table 2 shows mineral composition ranges, as determined by electron probe microanalysis.
Alkali basalt, hawaiite, mugearite and benmoreite Mafic silica-undersaturated rocks generally have 100Mg/(Mg + Fe) ratios less than or equal to 66, and nickel rarely exceeds 120 ppm, indicating that few, if any, of these rocks represent primitive magmas although a few do carry ultramafic xenoliths. Schieffer & Nelson (1981) reported 'basanite'in an abstract but did not give analytical data. Alkali basalt and benmoreite are very rare in the Trans-Pecos province, but hawaiite and mugearite are widespread in the southern part (Fig. 3(a)). Phenocryst phases are olivine, titanaugite, kaersutitic amphibole, titanomagnetite, apatite and plagioclase (Table 2) in all four types.
Phonolite, nepheline syenite and nephelinenormative traehyte Silica-undersaturated rocks of low colour index occur in a belt along the eastern margin of the province (Fig. 3(b)). The most silica-undersaturated felsic rocks occur at the northern limit of the province, in the Diablo Plateau (Barker & Hodges 1977; Barker et al. 1977), and are distinctly peralkaline, carrying eudialyte and catapleiite. Phenocrysts are anorthoclase, clinopyroxene (sodian augite through hedenbergite to acmite), manganoan fayalite, edenitic to arfvedsonitic amphiboles, and aenigmatite. Nepheline and sodalite form phenocrysts in a few occurrences of phonolite, but generally these phases are interstitial in the groundmass and partly or entirely converted to analcime. Most phonolites and trachytes show well-developed flow alignment of tabular alkali feldspars in the groundmass, with interstitial acmite, arfvedsonite and aenigmatite. Plagioclase is a phenocryst phase in a few examples of trachytes toward the southern limit of their occurrence. The compositional gap between mafic and felsic silica-undersaturated rocks apparently reflects the derivation of nepheline-normative trachytic and phonolitic liquids by efficient filter-
418
D. S. Barker
TABLE 1. Representative analyses o f Trans-Pecos igneous rocks 1
10
11
12
3
4
5
6
7
8
9
Major elements (wt. %) SiO2 44.99 43.24 TiO2 3.55 3.69 A120 3 16.68 15.56 Fe203 2.73 4.49 FeO 9.56 7.87 MnO 0.17 0.16 MgO 4.70 6.22 CaO 9.59 8.31 Na20 3.69 4.64 K20 1.29 1.62 H2 O§ 1.26 2.65 H200.29 0.38 P205 0.74 0.86 CO2 0.15 0.10
44.70 2.61 14.89 3.85 7.96 0.19 7.07 9.42 3.21 1.47 2.40 0.24 1.20 0.01
50.15 2.45 17.55 4.20 5.58 0.18 3.49 6.61 4.63 2.51 0.92 0.10 0.93 0.05
50.42 2.09 16.98 3.05 7.00 0.12 2.70 5.50 5.07 3.20 0.52 0.32 1.38 0.70
55.15 1.28 18.41 3.24 3.28 0.17 1.78 3.75 5.86 4.71 0.76 0.11 0.62 0.22
56.10 0.18 19.57 2.87 1.86 0.29 0.13 0.90 10.77 4.90 2.42 0.22 0.04 0.08
58.60 0.11 18.54 3.70 2.08 0.31 0.30 1.62 7.65 5.47 1.90 0.17 0.11 0.26
59.70 0.11 17.80 3.84 1.71 0.26 0.12 1.13 7.82 5.21 1.55 0.21 0.07 0.12
62.20 0.37 17.08 3.67 1.79 0.23 0.30 1.17 7.04 5.50 0.66 0.12 0.15 0.00
63.00 0.42 16.97 2.02 2.92 0.35 0.41 0.92 7.36 4.98 0.35 0.11 0.17 0.05
49.83 1.62 16.35 2.53 6.79 0.11 5.52 9.19 3.37 0.82 1.87 1.03 0.42 0.25
Sum
99.22
99.95
99.05
99.34
100.33
100.82
99.65
100.28
100.03
99.70
147 55 55 920 171 <3 98
93 8 161 742 190 <3 104
99.39
2
99.79
Trace elements (ppm) Rb 45 38 Sr 1054 1100 Y 16 29 Zr 226 373 Nb 46 59 Ni <3 50 Zn ND 105 Diff. index 35.46 an (~o) 51.3 ne(wt. ~o) 4.01
39.99 45.8 10.48
42 1094 60 150 85 78 100 34.33 47.8 1.79
63 1000 32 280 41 3 91 53.65 33.9 0.43
44 915 33 390 50 <3 111 60.88 25.3 1.10
85 970 29 520 59 <3 90 73.43 19.7 4.71
260 85 56 2549 260 4 154 81.84 0 24.23
142 123 52 1640 154 <3 118 88.00 0.2 1.07
254 68 72 1700 218 4 137 86.78 0 7.76
88.42 0 1.01
15 470 38 160 < 10 40 96
88.44 34.72 0 48.7 0.66 q(wt. ~) 0.50
All major-element oxide analyses by G. K. Hoops and all trace-element analyses by D. S. Barker, both of the Department of Geological Sciences, University of Texas at Austin. ND, no data. Silica-undersaturated rocks" 1, Black Hill, alkali basalt, 103~176 W of Nine Point Mesa (Bobeck 1985); 2, Cob-2, hawaiite, 103~176 3, Tib-l, hawaiite, 104~176 intruding upper Rawls Formation, whole-rock K - A r age 17.6+0.3 Ma (McDowell 1979); 4, XM-12-3, hawaiite, 103~176 dyke in Christmas Mountains (Lewis 1978); 5, Viuda, mugearite, 103~176 La Viuda, Tascotal Mesa quadrangle' 6, AS-5, mugearite, 105~176 Cornudas Mountains group, Diablo Plateau, biotite K-Ar age 34.9__+1.1 Ma (Barker et al., 1977); 7, CH-1, phonolite, 105~176 Chattfield Mountain, Cornudas Mountains group, Diablo Plateau (Barker et al. 1977); 8, WN-11, nepheline syenite, 105~176 Wind Mountain, Cornudas Mountains group, Diablo Plateau (Barker et al. 1977); 9, SP-28, trachyte, 105~176 chilled selvedge of Sierra Prieta pluton, Diablo Plateau (Barker et al. 1977); 10, Ht-1, trachyte, 103~176 Heart Mountain; ll, PP-3, trachyte, 103~176 Paisano Mountain, Paisano Pass area (Parker 1983).
pressing from largely crystallized mafic m a g m a s . N e t - v e i n i n g and diapiric segregations of felsic feldspathoidal rocks occur within mafic hosts at Mariscal M o u n t a i n ( B u m g a r d n e r 1976), Rattlesnake M o u n t a i n ( C a r m a n et al. 1975), the Christmas M o u n t a i n s (Jungyusuk 1977; Lewis 1978), N i n e Point Mesa (Bobeck 1985) and Marble C a n y o n (Price et al., 1986a). M a n y other mafic bodies also show small segregations of felsic rock. In contrast, the felsic intrusive masses are h o m o g e n e o u s and rarely carry inclusions that could be interpreted as mafic autoliths (Barker et al. 1977).
C o m p a r i s o n of the distribution m a p s (Figs 3(a) and 3(b)) shows that large bodies of felsic and mafic silica-undersaturated rocks are mutually exclusive, except in the Diablo Plateau. This lack of areal overlap m a y be due to varying ease of separation and ascent of felsic m a g m a s , and to varying a m o u n t s of uplift a n d erosion. N o n e of the felsic silica-undersaturated rock bodies is k n o w n with certainty to have b r e a c h e d the surface, but m a n y show chilled contacts and a b u n d a n t miarolitic cavities and must have been e m p l a c e d at depths of less t h a n 1 km. In the S, w h e r e mafic bodies are the rule, shallower felsic
Tertiary magmatism in Trans-Pecos Texas
13
14
15
16
17
18
19
58.93 1.34 15.78 4.60 2.61 0.15 1.33 3.85 5.56 4.21 0.79 0.43 0.71 0.25
62.23 0.94 16.64 4.09 0.96 0.04 0.75 2.41 5.21 4.67 1.04 0.23 0.39 0.30
62.80 0.40 17.50 5.85 0.44 0.12 0.05 0.36 6.00 5.73 1.26 0.13 0.18 0.00
65.49 0.71 15.09 4.48 0.40 0.12 0.28 0.74 5.06 5.90 0.56 0.26 0.23 0.02
100.54
99.24
100.82
99.22
419
20
21
22
23
66.60 0.34 16.65 1.62 0.52 0.15 0.26 0.73 6.28 5.31 0.33 0.11 0.11 0.05
67.60 0.49 14.30 2.23 1.92 0.18 0.10 0.82 5.17 5.65 0.61 0.23 0.06 0.04
70.23 0.30 12.42 3.29 1.97 0.08 0.01 0.48 5.36 4.65 0.01 0.25 0.11 0.05
73.37 0.17 11.53 2.27 1.80 0.09 0.02 0.14 5.06 4.37 0.45 0.13 0.11 0.02
99.06
99.40
99.21
99.53
24
Major elements (wt. %)
SiOe TiO2 A1203 Fe203 FeO MnO MgO CaO Na20 K20 H2 O+ H20P205 COz
49.34 1.68 15.27 6.78 4.01 0.16 6.81 8.15 3.38 0.87 1.52 0.50 0.46 0.20
53.00 1.87 15.19 5.09 4.30 0.14 4.73 7.23 4.10 1.99 0.67 0.35 0.86 0.08
54.91 1.89 16.23 4.52 3.06 0.16 1.87 3.66 5.05 3.85 1.76 0.98 0.82 0.28
Sum
99.13
99.60
99.04
75.60 0.09 14.00 0.49 0.30 0.06 0.02 0.34 4.86 3.40 0.24 0.08 0.00 0.02 100.00
Tracee&men~ (ppm)
Rb Sr Y Zr Nb Ni Zn
13 520 23 97 20 102 101
Diff. index an (%)
q (wt. %)
34.95 45.6 0.99
22 1055 27 300 20 40 100 50.41 33.1 3.96
70 470 62 620 50 <3 93 70.56 19.4 2.14
95 520 59 640 43 <3 101 75.11 10.8 3.17
106 295 97 638 31 ND 83 81.24 14.6 9.56
198 20 41 1150 109 <3 89 90.09 1.2 5.46
92 22 85 381 42 <3 91 90.05 2.4 12.37
150 129 29 547 118 <3 72 92.98 2.9 8.19
115 7 58 534 102 <3 98 90.14 0 14.65
397 12 132 1777 195 13 198 87.02 0 21.54
205 10 119 2200 166 9 108 88.52 0 27.72
1071 0 258 41 63 14 102 95.82 3.7 31.65
Silica-oversaturated rocks: 12, TPC-435, trachybasalt, 103~176 Bee Mountain Basalt, Big Bend National Park; 13, 77042, trachybasalt, 102~176 flow at Black Gap, whole-rock K-Ar age 22.7 Ma (F. W. McDowell, personal communication, 1984); 14, QB-1, trachybasalt, 105~176 intrusive plug W of Quitman Mountains, whole-rock K-Ar age 33.3 _+0.6 Ma (McDowell 1979; Barker 1980) (in metaluminous belt); 15, 77036, trachyandesite, 103~176 Alamo Creek Basalt, Big Bend National Park; 16, 76001, quartz trachyte, 103~176 flow NW of Puertacitas Hills; 17, DP142, quartz trachyte, 103~176 flow in Fox Canyon Formation (Parker & McDowell 1979); 18, RH-3, quartz trachyte, Star 105~176 Red Hills pluton, Diablo Plateau (Barker et al. 1977); 19, DP-140, quartz trachyte, 103~176 Mountain Rhyolite, alkali feldspar K - A r age 37.2_+0.7 Ma (Parker & McDowell 1979); 20, Alto-l, quartz syenite, 105~176 Cerro Alto; 21, Aguja-1, quartz syenite, 103~176 Big Aguja pluton, Davis Mountains (Parker 1986); 22, TPC-376, peralkaline rhyolite, 103~176 basal vitrophyre, Pine Canyon Rhyolite, Big Bend National Park (Ogley 1979); 23, PP-254, peralkaline rhyolite, 103~176 Paisano Rhyolite, Paisano Pass area (Parker 1983); 24, Blanca-1, metaluminous rhyolite, 105~176 Sierra Blanca (Barker 1980).
intrusions or extrusions m a y have been eroded. Conversely, in the N erosion has rarely been deep e n o u g h to expose the mafic parents.
Trachybasalt and trachyandesite These rocks are sparsely r e p r e s e n t e d by analyses (Fig. 2) primarily because m a n y show strong oxidation (Table 1), and high Fe3+ :Fe 2§ ratios are suspected to be the cause of slight silica oversaturation in the n o r m s of the majority of examples. So few m i c r o p r o b e data are yet available that these mafic q-normative rocks are
not included in Table 2; mineralogically there is no a p p a r e n t difference b e t w e e n these rocks and their silica-undersaturated counterparts. Alt h o u g h h y p e r s t h e n e appears in the norms, no orthopyroxene has been identified in thin sections.
Quartz trachyte and quartz syenite The most voluminous a n d w i d e s p r e a d a m o n g intrusive rocks, and probably a m o n g volcanic rocks as well, in the Trans-Pecos p r o v i n c e are felsic rocks c o n t a i n i n g up to 20% interstitial
420
D . S. B a r k e r
~ \!1
I00 km
~. Alkeli Basalt, Hawoiite, Mugeerite, Ben'moreite '
I00 km
M "'
I
X
/
\\
I
I "
9
. %')
"~,q~,)}
P h o ~ Nephelinesyenite, Trachyte
" (
a)
-'~
/,/
b)
\
~ \\ " 9 L
I0'0km
I
i
\ Q~hyte,
I~ km ~ \
\\/
"~'f
~ '..
Quartzsyenite
(c)
Rhyolite,Granite
.. ~-,,~.
d)
FIG. 3. Distributions of four groups of Trans-Pecos intrusive rocks. The broken line is the boundary between the alkalic belt on the E and the metaluminous belt on the W. (Revised from Barker 1977, Fig. 5.)
quartz. Other prominent phases are cryptoperthitic to microperthitic alkali feldspar, plagioclase as calcic as An46 (largely as cores mantled by anorthoclase, and not present in all occurrences because many samples are peralkaline), fayalitic olivine, clinopyroxene, a wide range of amphiboles and biotite (Table 2). Fluorite and zircon are common accessory phases. Examples are described by Barker et al. (1977), Barker & Hodges (1977), Indest & Carman (1979), Parker (1983) and Bobeck (1985). Barker (1977) divided the Trans-Pecos province into two belts, alkalic to the E and metaluruinous to the W, with the interpolated boundary approximately parallel to the south-western margin of the N American plate in early Cenozoic time. The significance of this boundary will be discussed later; its existence must be mentioned
now because both belts contain silica-oversaturated rocks (Figs 3(c) and 3(d)). Peralkaline varieties do occur in the metaluminous belt (Cepeda & Henry 1983; Cameron & Cameron, personal communication, 1984) but are clearly subordinate to the associated metaluminous rocks. In the Diablo Plateau (Barker et al. 1977) small bodies of slightly oversaturated quartz syenite occur at the margins of large intrusive masses of ne-normative syenite. Because ratios among Rb, Y, Zr and Nb of the two rock types in each composite pluton are so similar, the quartz syenite selvedges are now interpreted as deuteric products formed during near-solidus loss of alkalis from the peralkaline ne-normative syenites. Other large quartz syenite bodies are definitely magmatic.
Tertiary magmatism in Trans-Pecos Texas
421
TABLE 2. Mineral compositions in Trans-Pecos rocks Alkali basalt, hawaiite, mugearite, benmoreite
Phonolite, nepheline syenite, trachyte
Quartz trachyte, quartz syenite
Olivine Fo (mol. %) Fa (tool. %) Tp (tool. %) Ca-ol (mol. %)
P 42-44 53-55 2 1
P 1-23 72-87 4-11 1
P 10 85 4 1
P 0-1.5 95-98 2-5 0-1
Clinopyroxene Ca (at. %) Mg (at. %) Fe (at. %) TiO2 (wt. ~o) A1203 (wt. ~) NazO (wt. %)
G 45-48 34-41 13-20 1-4 4-8 0.3-1.0
G 6-47 1-34 19-93 0.2-0.6 0.7-1.8 0.8-13
G 42-44 23-28 30-33 0.6-1.0 1.0-1.2 0.2-0.5
P 41 1 58 0.5-0.7 0.4 1.1-1.2
Amphibole TiO 2 (wt. ~) 100Mg/(Mg + Fe) (atomic)
G 1.7-6.2 36-64
G 0.2-2.6 1-63
G 1.5-2.8 3-35
G 0.2-1.5 0.2-2.3
Biotite TiO 2 (wt. %) 100Mg/(Mg + Fe) (atomic)
G 6.9-9.0 43-61
G 0.9-7.2 3-40
G 2.3-6.2 7-63
G 2.6-4.1 1.3-1.7
Magnetite X' usp Ilmenite X' ilm
G 17-92 G 75-99
G 7-71 --
G 23-45 G 92-100
Plagioclase An (tool. %) Ab (mol. %) Or (mol. %)
P 12-75 26-85 0-7
P 28-45 53-67 2-5
P 1-46 51-84 3-15
Alkali feldspar An (mol. %) Ab (mol. %) Or (mol. %)
G 8-21 33-76 12-65
P 0-5 38-91 4-61
P 0-15 45-74 11-65
Rhyolite, granite
G 1-10 0 90-99 2.6-4.6 0.3-0.8 13-14
-G 99
P 0-4 52-69 31-48
P, phenocrysts' G, groundmass. Sources: Barker, unpublished data; Barker & Hodges 1977; Becker 1976; Bobeck 1985 Carman et al. 1975' Indest & Carman 1979" Jungyusuk 1977; Lewis 1978; Parker 1983.
Unfortunately, igneous rock terms were incorporated into some Trans-Pecos stratigraphic names before adequate analytical data were available. As a result, for example, the widespread Crossen Trachyte is a rhyolite but some samples of the Star Mountain Rhyolite are quartz trachyte (Table 1). Rhyolite and granite
In the alkalic belt many examples of rhyolite, but fewer of granite, are peralkaline. However, the persistence of groundmass arfvedsonitic amphi-
bole and aenigmatite in many samples in which the atomic ratio of alkalis to A1 is now less than unity suggests that a substantial portion of these rocks was originally peralkaline but has lost Na and K. Phenocrysts are cryptoperthitic anorthoclase (Christoffersen & Schedl 1980; Yund & Chapple 1980), augite and rare fayalite. Quartz is a phenocryst phase only in the most highly evolved rhyolites in both the alkalic and the metaluminous belts, e.g. in the youngest lavas, ash flows and dykes of the alkalic Pine Canyon caldera (Ogley 1979; Barker et al, 1986) and in laccoliths of the metaluminous Sierra Blanca
422
D. S. Barker
group (Barker, 1980). Orthopyroxene phenocrysts occur in some metaluminous rhyolites (M. Cameron et al. 1982; C. D. Henry, personal communication, 1985), and pigeonite occurs in at least one (J. G. Price, personal communication, 1985). Other alkalic rhyolites are described by Gibbon (1969), Gibbon & Wyllie (1969), McKay & Rogers (1970), Becker (1976), Sharp (1979) and Parker (1983). Burt (1970), K. L. Cameron et al. (1979) and Cepeda & Henry (1983) described rhyolites of the metaluminous belt.
Chronology Figure 4 shows Thornton-Tuttle differentiation indices for samples dated by the K - A r method. Rocks with differentiation indices less than 75 were emplaced over the entire span from 48 to 16 Ma ago, and were joined by felsic rocks from 37 to 26 Ma ago. Implications of the timecomposition relations will be discussed at the end of this paper.
Field relations Intrusive bodies Most, if not all, Trans-Pecos intrusions crystallized within 2 km of the surface. Support for this statement comes from reconstruction of the local stratigraphic sections at the time of emplacement,
9O
% ._
70
--Shortening
~- - - - 4 -
Extension
~6 q-
.~_ 50 d3
9
O0
9
50 50
40
50
20
K - A r Qge, Mo
FIG. 4. Thornton-Tuttle differentiation index versus K-Ar age. Each point represents a sample upon which both a chemical analysis and a K-Ar age determination were performed. Samples of Trans-Pecos volcanic rocks are omitted from this and other figures if their vents are known to be in Mexico, which is outside the Trans-Pecos province as defined in this paper. The only two felsic samples with ages less than 30 Ma are both from the BofeciUosVolcano. Times of crustal shortening and extension (Price et al., in press) are indicated.
from the characteristically fine grain-size and vesicular or miarolitic textures of the igneous rocks, and from structural evidence (lifting of roofs, lateral injection along bedding surfaces). Laccoliths and gently discordant sheets ('trapdoor intrusions') are common, especially in the Diablo Plateau (Barker et al. 1977), and were preferentially injected along specific stratigraphic levels, especially at unconformities and other inter-formational contacts. In the S, quartz syenite forms large tabular bodies up to 300 m thick and with outcrop areas exceeding 65 km2; floors of these masses are concordant on a large scale but abruptly step through the stratigraphic section from one preferred level to another, and at margins the intrusions finger out into a stack of sills, each about 10 m thick (e.g. Nine Point Mesa (Bobeck 1985)). In the S, mafic sills from 6 to 50 m thick tend to be more concordant than the felsic bodies and crop out for distances of up to 12 km (Bumgardner 1976; McCulloh 1977; DeCamp 1981; Bobeck 1985). Thicker lessregular mafic sills were described by Lonsdale (1940), Maxwell et al. (1967) and Carman et al. (1975). Ring-dykes occur in the Quitman Mountains, Pine Canyon and Sierra Quemada calderas (Henry & Price 1984). Radial dyke-swarms are well developed around Dominguez Mountain in Big Bend National Park (Maxwell et al. 1967) and the Christmas Mountains complex (Jungyusuk 1977; Lewis 1978). Other dyke arrays tend to show higher degrees of preferred orientation related to regional stress, as around the Quitman Mountains, Eagle Mountains, Paisano Volcano, Chinati Mountains and Pine Canyon calderas (Price & Henry 1984). A set of young mafic dykes strikes N N W - W in the Rim Rock country (Dasch et al. 1969). Large irregular plutons of quartz syenite and granite occur in the Infiernito, Chinati Mountains and Eagle Mountains calderas of the metaluminous belt; Henry & Price (1984) interpret these as products of resurgent doming. Similar plutons occur on the flanks of the Pine Canyon, Buckhorn and Paisano calderas of the alkalic belt, which are smaller than those of the metaluminous belt and show no signs of resurgence (Henry & Price 1984). Still others, unrelated to calderas, occur throughout the Trans-Pecos province S of the latitude of Fort Davis. Discrete clusters of mafic or felsic, silicaoversaturated or silica-undersaturated, intrusive bodies suggest the presence of subjacent plutons (Barker 1979b, Fig. 3) with a mean diameter of 40 km. These clusters differ in ratios among Si, Ti, K, Rb, Sr, Y, Zr and Nb, indicating a separate magma reservoir for each cluster. Rb and Sr data
Tertiary magmatb~m in Trans-Pecos Texas
423
Henry & Price (1984) summarize Trans-Pecos calderas and their associated ash-flows, and readers are urged to consult their paper, of which only a few points are mentioned here. In the metaluminous belt six calderas are the sources of major ash-flows, including the widespread Mitchell Mesa Rhyolite (Burt 1970; Cepeda & Henry 1983) consisting of six flow units making one cooling unit with an estimated volume exceeding 1000 km 3, from the Chinati Mountains caldera, and the peralkaline Buckshot Ignimbrite (Anderson 1976) with an estimated volume of 3040 km 3, from the Infiernito caldera. In the alkalic belt four calderas are certain, and two more are probable, sources of ash-flows; the most voluminous single unit is the peralkaline Gomez Tuff which is a densely welded pantellerite with an estimated volume of 220 km 3 (Parker & McDowell 1979; Parker 1986). In both belts, the earliest ash-flow was peralkaline. The Pine Canyon caldera erupted peralkaline rhyolites in two episodes separated by an interval in which quartz trachyte was emitted; the magmatic reservoir system under this volcano was complex in its geometry and compositional variation (Barker et al. 1986).
(Bramson 1984) favour fractional crystallization over fractional fusion as the single most satisfactory explanation of compositional diversity within each reservoir, but magma mixing was locally important, as in the Marble Canyon pluton (Price et al. 1986a). Volcanic rocks
Extrusive products in the Trans-Pecos province include flood lavas, accumulations of lava and pyroclastic debris in shield volcanoes, and ashflow tuff, air-fall tuff and minor flows and domes genetically linked to calderas. Flood lavas include the Crossen Trachyte with an area of 1500 km 2 and an estimated volume of 90km 3 (McAnulty 1955; Parker & McDowell 1979), the Star Mountain Rhyolite with an area of 2500 km 2 and a volume of 205 km 3 (Gibbon 1969; Parker & McDowell 1979), the Tule Mountain Trachyandesite (Maxwell et al. 1967) and the Alamo Creek Basalt (Stewart 1982). Gibbon (1969) identified three flow units in the Star Mountain Rhyolite; the other bodies called flood lavas here demand additional field study to find the locations and natures of their source vents. Shield volcanoes include that at Paisano Pass (Parker 1983) which is surmounted by a small caldera, the younger and more mafic Bofecillos Volcano (McKnight 1970), and the broad accumulations called Sheep Canyon Basalt and Cottonwood Springs Basalt (McAnulty 1955) and the lower and upper mafic units of Parker (1983).
Trace-element geochemistry Trace-element data (for Rb, Sr, Y, Zr, Nb, Ni and Zn) have now been obtained by X-ray fluorescence spectrometry for 166 samples for which major-element analyses have also been
9
700 -
9
oo
oo
oo
9
o
Oo
9
o o o
500
9
300
I,
9 9
'o ~
9 m
,-.,00 0
~
Jo 0 ~176 o
9
0
0 0
I00
0 I 5
0 (a)
I I0
[ 15
[ 20
J 25
0 (b
[ 5
I I0
I 15
Zr/Nb
I 20
I 25
~0 ~" 0
I 5
I I0
I 15
[ 20
(c
FIG. 5. K/Rb versus Zr/Nb' (a) 0 , alkali basalt, hawaiite, mugearite, benmoreite; (3, phonolite, nepheline syenite, trachyte; (b) O, trachybasalt, trachyandesite; O, quartz trachyte, quartz syenite; (c) Q, rhyolite and granite of the metaluminous belt; (3, rhyolite and granite of the alkalic belt.
]
25
424
D. S. Barker I'0
normative and the ne-normative series, increasing differentiation index is reflected by increasing Zr/Ti and Nb, and by decreasing K/Rb. Zircon is a microphenocryst phase in many samples of quartz trachyte and quartz syenite (Parker 1983) and in peralkaline rhyolites. In peralkaline phonolites and nepheline syenites, Zr occurs in eudialyte, catapleiite, rosenbuschite and baddeleyite (Price et al. 1986a), and acmitic clinopyroxenes contain up to 2 wt. ~ ZrO 2. Rhyolites from the metaluminous belt differ from those of the alkalic belt in having higher Rb/Zr and lower Nb (Fig. 6).
0.6 0,4
O0 I
I0O
o
(Q) I'0 t
500
0
0"6 ~0 . 4 ~ .
~o,o /
o ,~0,-~o
I
o~ ~ o
ooO
IO0
o
o o 1
o
0
-2 L
c~oOoo o o o 00
-5
50O
9
(b)
o
I-0 -4f~ 0 (Q)
0.6 0"4
o
o
I 40
I
I 60
I
0
0
I 80
-2
0 J
IOO (c)
I
oo ~
o I
OI 20
I
500
O0
0
ppm Nb
FIG. 6. Rb/Zr versus Nb (ppm): (a) 0 , alkali basalt, hawaiite, mugearite, benmoreite ; O, phonolite, nepheline syenite, trachyte; (b) O, trachybasalt, trachyandesite; 9 quartz trachyte, quartz syenite; (c) Q, rhyolite and granite of the metaluminous belt; O, rhyolite and granite of the alkalic belt.
made on dissolved aliquots of the same powders. Earlier data for the Trans-Pecos province were published by Carman et al. (1975), Barker et al. (1977) and Parker (1983). Figures 5, 6 and 7 show K/Rb versus Zr/Nb, Rb/Zr versus Nb (ppm) and log (Zr/TiO2) versus Zr/Y. Within any specific volcanic or intrusive complex, samples show nearly constant values of Zr/Nb and Rb/Zr, as expected for a co-magmatic alkaline suite (Weaver et al. 1972). Variations in these ratios shown in Figs 5 and 6 reflect the diversity of magma batches feeding individual centres. Compared with the mafic silica-undersaturated rocks, the few analysed samples of mafic qnormative rocks (trachybasalt and trachyandesite) have higher Zr/Nb ratios (Fig. 5), but there are no other obvious differences among the available trace-element data. In both the q-
b4
-4
oo -o i
o (b)
l
I
2o
40
L
I
I
I
60
80
-f#o
-3
bo
-4 L
0 (C)
L
20
L
I
40 Zr/Y
L
I
60
L
I
80
FIG. 7. Log (Zr/TiO2) versus Zr/Y: (a) Q, alkali basalt, hawaiite, mugearite, benrnoreite; O, phonolite, nepheline syenite, trachyte; (b) 0 , trachybasalt, trachyandesite; O, quartz trachyte, quartz syenite; (c) O, rhyolite and granite of the metaluminous belt; 9 rhyolite and granite of the alkalic belt.
Tertiary magmatism
Contrasting behaviour between 'large-ion lithophile' (LIL) elements (including K and Rb) and 'high field strength' (HFS) elements (including Zr, Nb and Y) has recently been recognized (see, for example, Brown et al. 1984). The Trans-Pecos metaluminous belt has higher LIL/HFS ratios than the alkalic belt, suggesting greater involvement of crust in magmatic processes (partial fusion, assimilation) under the metaluminous belt and generation of alkaline magmas from subcrustal lithosphere or asthenosphere under the alkalic belt. Rare-earth data are still very sparse for TransPecos rocks. Parker (1983) gives data for one mugearite and one peralkaline rhyolite from the alkalic belt. Irving & Frey (1984) list rare-earth concentrations in one sample of alkali basalt from the alkalic belt and one of hawaiite from the metaluminous belt. K. L. Cameron & M. Cameron (personal communication, 1984) report data for four metaluminous rhyolites and two peralkaline rhyolites, all from one caldera in the metaluminous belt. K. L. Cameron et al. (1980) give a plot for one rhyolite of the alkalic belt. In Fig. 8 these data are summarized in a conventional chondrite-normalized plot. All rhyolites show pronounced negative Eu anomalies. Peralkaline rhyolites show nearly identical profiles, even though one sample represents early and voluminous lava from the Paisano volcano, another represents a lava dome in the Pine Canyon caldera (both in the alkalic belt) and the other two are late and volumetrically-minor products (a dome and a dyke) of the Chinati Mountains caldera in the metaluminous belt.
I000
-
-
I00
2
in T r a n s - P e c o s T e x a s
425
Metaluminous rhyolites are less enriched in all rare-earth elements (REE) compared with peralkaline rhyolites. Mafic silica-undersaturated rocks from both belts show little difference and no Eu anomaly. Clearly many more data are needed. REE data are given in abstracts by Nelson et al. (1981), McDonough et al. (1982) and Schieffer et al. (1982).
Isotope geochemistry No Pb, Nd, H or C isotopic analyses are known to the writer for Trans-Pecos igneous rocks, and Sr and O data are meagre. STSrff%r initial ratios are provided by Hedge (1966), Dasch (1969), Barker et al. (1977), Lewis (1978) and K . L. Cameron & M. Cameron (personal communication, 1984). Ranges are as follows: mafic silicaundersaturated rocks (eight samples), 0.70260.7044; mafic silica-oversaturated rocks (two samples), 0.7040-0.7043; phonolites, nepheline syenites and trachytes (23 samples), 0.70410.7052 and 0.7078-0.7120; quartz trachytes and quartz syenites (eight samples), 0.7030-0.7091; metaluminous rhyolite (one sample), 0.7140; alkalic rhyolites (two samples), 0.7067-0.7093. Samples showing the higher initial ratios are very low in Sr and therefore were most susceptible to contamination by radiogenic Sr from Precambrian basement and from groundwater that has percolated through Permian and Cretaceous carbonate wall-rocks. Xenoliths of Precambrian rocks have been recognized in the Diablo Plateau (Barker et al. 1977) and in the Paisano Volcano (Parker 1983). For all rock types except the rhyolites, for which data are obviously inadequate, initial Sr ratios are low enough to permit magma generation in the upper mantle or lower crust. Schwarzer & Johnson (1976) give r values of 6.8-8.4, increasing with higher silica content, for unspecified volcanic rocks in the Davis Mountains, but 10.2-10.8 for Star Mountain Rhyolite, implying crustal involvement through assimilation or partial fusion.
to
Megacrysts and mantle-derived xenoliths t
I
Ce
I I
[
[
I
I
1
Sm Gd Dy Er Yb Eu Yb FIG. 8. Chondrite-normalized REE plot for 11 TransPecos rocks. Sources of data are given in the text. Horizontal shading, mafic silica-undersaturated rocks (three samples); vertical shading, peralkaline rhyolites (four samples); black, metaluminous rhyolites (four samples).
Ultramafic xenoliths occur in mafic silica-undersaturated rocks N of Black Gap (St. John 1966) and the Terlingua area (Schieffer & Nelson 1981 ; Schieffer & Mattison 1982). All are spinel-bearing lherzolite, dunite, harzburgite and websterite. In the Terlingua locality megacrysts are also found. Olivine, spinel and titanomagnetite are interpreted as fragments of xenoliths, but orthopyrox-
426
D. S. Barker
ene, clinopyroxene and sodic plagioclase appear to have precipitated from the host magma. Megacrysts of Al-rich orthopyroxene, spinel and plagioclase (Merrill & Irving 1977; Irving & Frey 1984) occur in Sheep Canyon Basalt (alkalic belt) and megacrysts of apatite, titanomagnetite, kaersutite, biotite, plagioclase and anorthoclase (Irving 1977; Irving & Frey 1984) were found in a hawaiite dyke of the Rim Rock swarm (metaluminous belt). The Sheep Canyon and Rim Rock megacrysts may all have precipitated from the liquids that now form their hosts, or from earlier more highly evolved liquids. It is unlikely that any megacryst phase at these two localities is a remnant of the mantle parent from which the transporting magma was derived. Parker (1972) reported xenoliths containing olivine and chromian diopside in mafic dykes in the north-eastern Davis Mountains. Presumably other occurrences of mantle-derived xenoliths and megacrysts, and of megacrysts precipitated from ascending magmas, await recognition, but they are probably rare because even the most mafic magmas appear to have fractionated during ascent and therefore any suspended xenoliths and megacrysts are likely to have settled out.
Mineralization Mercury deposits in the Terlingua district are associated with quartz syenite and rhyolite intrusions (Yates & Thompson 1959) and were worked for many years. Mineralization related to caldera complexes (Henry & Price 1984, Tables 2 and 3), nearly all in the metaluminous belt, has yielded significant production of Ag, Pb, Zn and fluorite; Cu, W, Be and Sn are locally concentrated but have not supported mines. Uranium has been prospected in the Buckshot Ignimbrite, and radiometric anomalies identified by airborne surveys occur in many intrusive bodies of felsic silica-undersaturated rocks. A porphyry molybdenum deposit (Sharp 1979) occurs in rhyolite of the Cave Peak intrusion near the Marble Canyon pluton. Price et al. (1983) have compiled a bibliography of Trans-Pecos mineral deposits, including the many that are not associated with Cenozoic magmatism.
Structure Basin and Range faulting occurs throughout the Trans-Pecos magmatic province and, over a large part in the W and S, was superimposed on Laramide folds (Cobb & Poth 1980; Muehlberger 1980; DeCamp 1981). Mafic sills in the southern part of Big Bend National Park (Bumgardner 1976) occur within folded Cretaceous limestones
and, because felsic layers and diapiric bodies (representing late-stage segregations of liquid within the nearly solidified mafic host) tend to be either parallel or perpendicular to bedding surfaces in the folded wall-rock, the sills are interpreted as having been emplaced before folding started. Orientations of dykes and hydrothermal veins within plutons further suggest that E - N E shortening persisted at least until 32 Ma ago (Price & Henry 1984). Widespread ash-flow tufts, such as the 32 Ma old Mitchell Mesa Rhyolite (Burt 1970), predate normal faulting, except for that forming caldera margins. E - N E extension probably began between 32 and 30 Ma ago (Price & Henry 1984), but Basin and Range faulting with large displacement began as late as 24 Ma ago; N-NW-striking dyke-swarms 24-18 Ma old formed in the Rim Rock country (Dasch et al. 1969) and in the flanks of the Bofecillos shield volcano (McKnight 1970; McDowell 1979). Mafic flows in the Black Gap area are 23 Ma old (McDowell, personal communication, 1984) and are displaced by normal faults (St. John 1966; Moustafa 1983). 23 Ma old mafic flows in the Bofecillos Volcano interfinger with debris from fault scarps (McDowell 1981). The youngest volcanic rock yet recognized in the Trans-Pecos province is an isolated remnant of a hawaiite flow on Cox Mountain in the southern Diablo Plateau (King 1965; Barker 1980), dated at 16.7 ___0.3 Ma (McDowell 1979). Muehlberger et al. (1978) demonstrated that faulting still continues, long after the cessation of Trans-Pecos magmatism. The overall pattern (Moustafa 1983) is one of Laramide contraction parallel to a N70~ direction, resulting in an average of 3.4% shortening, followed by Basin and Range extension (an average of 6.5% elongation). The contraction reactivated W - N W striking basement faults with left-lateral strikeslip displacement, and produced some folds, high-angle reverse faults and low-angle thrusts. Extension also reactivated basement faults, producing pull-apart grabens terminating against W-NW-striking right-lateral strike-slip faults and bounded by N-NW-striking normal faults. With the possible exception of two samples from the Bofecillos Volcano, all dated and analysed Trans-Pecos igneous rocks with differentiation index greater than 50 were emplaced during the shortening episode (Fig. 4).
Relations between tectonic setting and magmatism Igneous rocks of the alkalic belt in the TransPecos province have long been compared with
Tertiary magmatism
the hawaiites, mugearites, trachytes, phonolites and peralkaline rhyolites of the Kenya (Gregory) rift (Prior 1903; Smith 1931; Barker 1977). Although the analogy extends further than rock compositions and mineralogy to include eruptive styles (flood lavas and shield volcanoes topped by non-resurgent calderas similar to Longonot and Silali (Williams et al. 1984)) and the associated normal faults, the Trans-Pecos magmatic province does not mark a continental rift. There is no evidence of crustal thinning or of an underlying axial body of mafic or ultramafic rock close to the surface. The Rio Grande rift adjoining the TransPecos province on the NW shares many structural similarities (ages, orientations and styles of faults (Muehlberger et al. 1978)), because both the Rio Grande rift and the Trans-Pecos region are parts of the Basin and Range province; however, Trans-Pecos magmatism began much earlier and felsic magmatism occurred in a contractional regime and ended before igneous activity in the Rio Grande rift began (Barker 1979b); igneous rock compositions in the two provinces show little overlap. The Trans-Pecos province is a relatively small outlier of the extensive Cenozoic volcanic field dominated by the calc-alkaline calderas of the Sierra Madre Occidental in Mexico (K. L. Cameron et al. 1979, 1980; McDowell & Clabaugh 1979; Swanson & McDowell 1984). Analogous alkaline tracts occur E of the Andes (references in Brown et al. 1984), as well as in N America (Barker 1974). Accumulating traceelement data (Cameron & Cameron 1976; K. L. Cameron et al. 1980; Bramson 1984) support the tentative conclusion of Barker (1975 ; 1979b) and McDowell & Clabaugh (1979), based on majorelement data, that there is gradation in felsic magma compositions emplaced in the 37-26 Ma interval north-eastward from the Sierra Madre Occidental through eastern Chihuahua to the Trans-Pecos province. Among the compositional parameters that progressively change north-eastward are the normative colour index at a given normative plagioclase composition, and the Mg/Fe ratio, the atomic ratio (Na + K)/A1 and total alkalis at a given silica content (Barker 1979b, Figs 5, 6 and 7; McDowell & Clabaugh 1979, Fig. 8). From the Sierra Madre Occidental north-eastward into the Trans-Pecos province, at a given silica content, decreases in Ti, Ca, Sr, Mg/Fe and Y/Nb are coupled with increases in Mn, Na, K, Rb, Y, Zr, Nb, Zn and Zr/Rb (Bramson, McDowell & Barker, in preparation). The discrimination between alkalic and metaluminous belts within the Trans-Pecos province apparently reflects this gradation, which is much more
in Trans-Pecos Texas
427
pronounced in felsic than in mafic rocks (on the basis of major or trace elements in rocks with less than 55 wt.% SiO2, it is difficult to distinguish Trans-Pecos from eastern Chihuahua and Sierra Madre Occidental samples). Calc-alkaline magmatism in the Sierra Madre Occidental from 35 to 28 M a ago was linked to subduction of oceanic lithosphere of the Farallon Plate under the south-western margin of the N American plate (McDowell & Clabaugh 1979; Swanson & McDowell 1984). Felsic magmatism in the Trans-Pecos province occurred mostly in the interval 38-32 Ma ago (Fig. 4), and has also been interpreted as subduction related (Lipman et al. 1972; Keith 1978; Henry & Price 1984) or as a product of back-arc spreading over an asthenospheric diapir (Barker 1979b). If all the coeval magmas in the 500 km span between the Gulf of California and the Pecos were fed, directly or indirectly, from a subducting plate dipping to the E-NE, compositional gradation would be expected. Greater depth of melting, thicker and more thoroughly metasomatized lithosphere, and higher CO2/H20 ratios (Barker 1979b) above the older, initially colder, thicker and more dehydrated oceanic slab at greater depth could explain the increased alkalinity and higher HFS/LIL ratios of Trans-Pecos magmas. If, however, the calc-alkaline and alkaline magmas were generated by different mechanisms (one related to subduction and the other to backarc spreading), no true gradation in compositions would be expected, other than the small-scale effects of magma mixing in the region where magmas of both affinities were ascending. The data so far available favour gradation, and therefore suggest that all the Cenozoic felsic magmas were subduction related. The alkaline and calc-alkaline felsic rocks, with the possible exception of those in the Bofecillos Volcano, were all emplaced during crustal shortening, with maximum horizontal stress oriented approximately parallel to the direction of subduction. Price et al. (in press) discuss the compositional differences between Trans-Pecos rocks that formed during shortening and extension. Mafic rocks were emplaced in both episodes; those contemporaneous with extension generally contain more Mg and less Ti, but there are no marked differences in silica saturation between the two stress regimes.
Suggestions for future research The Trans-Pecos province offers abundant opportunities for additional work. Detailed mapping and volcanic stratigraphy, combined with
428
D. S. Barker
geochronology, are n e e d e d to identify vent areas and clusters of c o - m a g m a t i c bodies, and to date more precisely the c h a n g e from a compressional to an extensional stress regime. M o r e and better microprobe, trace-element and isotopic data are n e e d e d to answer the following questions. 1 Is the compositional c h a n g e with distance from the Oligocene t r e n c h truly gradational, or are there steps or inflections ? 2 D o the mafic silica-oversaturated rocks represent an i n d e p e n d e n t m a g m a lineage, or are they merely altered or c o n t a m i n a t e d deviants from the silica-undersaturated mafic m a g m a s ? 3 Felsic silica-undersaturated m a g m a s apparently formed by low-pressure fractionation from mafic silica-undersaturated m a g m a s , but are the differences in their areal distributions adequately explained by vagaries of uplift and erosion ? 4 Did the quartz syenites and quartz trachytes form by fractionation from the mafic silicau n d e r s a t u r a t e d m a g m a s , and if so h o w ? 5 A p p a r e n t l y the peralkaline rhyolites formed by crystal-liquid fractionation from quartz syenite m a g m a s , but did the m e t a l u m i n o u s rhyolites
also, or did their g e n e r a t i o n involve substantial assimilation or partial fusion of crustal rocks ? 6 By w h a t m e c h a n i s m s did both peralkaline and m e t a l u m i n o u s rhyolites form u n d e r calderas of the m e t a l u m i n o u s belt ? 7 W h a t limits can be placed on the compositions of m a g m a sources in the m a n t l e and crust u n d e r the Trans-Pecos province, and how do these limits differ from those of m a g m a sources u n d e r the Sierra M a d r e Occidental and u n d e r eastern Chihuahua? Certainly other petrologic questions will arise if and w h e n these are answered.
ACKNOWLEDGMENTS:My research in the Trans-Pecos Province, starting in 1970, has been supported by the U.S. National Science Foundation, Grants GA-11154, GA-32089 and EAR 75-22201, and by the University of Texas Geology Foundation. Fred McDowell, Don Parker, Steve Clabaugh, Bill Muehlberger, Virgil Barnes, Chris Henry, Jon Price, Max Carman, Ken Cameron and Maryellen Cameron have generously shared their unpublished data but do not necessarily agree with the interpretations in this review. For 15 years, G. Karl Hoops has enthusiastically performed high-quality wet chemical analyses for major-element oxides in more than 200 Trans-Pecos rock samples.
References ALBRITTON, C. C. 8z SMITH, J. F. 1965. Geology of the Sierra Blanca area, Hudspeth County, Texas. U.S. Geol. Surv. Prof. Pap. 479. ANDERSON, W. B. 1976. Folds, ramp structures, and tumuli of the Buckshot Ignimbrite, Presidio and Jeff Davis Counties, Texas. Geol. Soc. Am. Abstr. Progm. 8, 2. BAKER, C. L. 1934. Major structural features of TransPecos Texas. Texas Univ. Bull. 3401, 137-214. BARKER,D. S. 1974. Alkaline rocks of North America. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp. 160-71. Wiley, London. 1975. Tectonomagmatic relations between the Trans-Pecos and Sierra Madre Occidental magmatic provinces. Geol. Soc. Am. Abstr. Progm. 7, 144-5.
---
--
--
1977. Northern Trans-Pecos magmatic province: Introduction and comparison with the Kenya Rift. Geol. Soc. Am. Bull. 88, 1421-7. 1979a. Magmatic evolution in the Trans-Pecos province. In: WALTON,A. W. & HENRY, C. D. (eds) Cenozoic Geology of the Trans-Pecos Volcanic Field of Texas, Guidebook 19, pp. 4-9. Bureau of Economic Geology, University of Texas at Austin. 1979b. Cenozoic magmatism in the Trans-Pecos province: Relation to the Rio Grande Rift. In: RIECKER, R. E. (ed.) Rio Grande Rift: Tectonics and Magmatism. pp. 382-92. American Geophysical Union, Washington, DC. 1980. Cenozoic igneous rocks, Sierra Blanca area,
Texas. New Mexico Geol. Soc. Guidebook, 31st Field Conf., Trans-Pecos Region, pp. 219-23. New Mexico Geol Soc. - & HODGES, F. N. 1977. Mineralogy of intrusions in the Diablo Plateau, northern Trans-Pecos magmatic province, Texas and New Mexico. Geol. Soc. Am. Bull. 88, 1428-36. - - - , HENRY, C. O. & McDOWELL, F. W. 1986. Pine Canyon Caldera, Big Bend National Park: a mildly peralkaline magmatic system. In: PRICE, J. G., HENRY, C. D., PARKER, D. F. & BARKER, D. S. (eds) Igneous Geology of Trans-Pecos Texas. Field Trip Guide and Research Articles, Guidebook 23, 266-85, Bureau of Economic Geology, University of Texas at Austin. , LONG, L. E., HooPs, G. K. & HODGES, F. N. 1977. Petrology and Rb-Sr isotope geochemistry of intrusions in the Diablo Plateau, northern Trans-Pecos magmatic province, Texas and New Mexico. Geol. Soc. Am. Bull. 88, 1437-46. BARNES, V. E. 1968. Van Horn-E1 Paso sheet. In: Geologic Atlas of Texas, Bureau of Economic Geology, University of Texas at Austin (scale 1 : 250 000, one sheet with explanatory text). -1979a. Marfa sheet. In: Geologic Atlas of Texas, Bureau of Economic Geology, University of Texas at Austin (scale 1 : 250 000, one sheet with explanatory text). -1979b. Emory Peak-Presidio sheet. In: Geologic Atlas of Texas, Bureau of Economic Geology,
Tertiary magmatism in Trans-Pecos Texas University of Texas at Austin (scale 1:250 000, one sheet with explanatory text). - - - 1982. Fort Stockton sheet. In: Geologic Atlas of Texas, Bureau of Economic Geology, University of Texas at Austin (scale 1:250 000, one sheet with explanatory text). BECKER, S. 1976. Field relations and petrology of the Burro Mesa 'Riebeckite' Rhyolite, Big Bend National Park, Texas. M.S. Thesis, University of California at Santa Cruz (unpublished). BLOOMFIELD,K. & CEPEDA-DAVILA,L. 1973. Oligocene alkaline igneous activity in NE Mexico. Geol. Mag. ll0, 551-5. BOBECK, P. A. 1985. Igneous petrology and structural geology of Nine Point Mesa, Brewster County, Texas. M.A. Thesis, University of Texas at Austin (unpublished). BRAMSON, E. 1984. Trace-element study of Tertiary volcanic rocks from the Sierra Madre Occidental, Mexico to Trans-Pecos Texas. M.A. Thesis, University of Texas at Austin (unpublished). BROWN, G. C., THORPE, R. S. & WEBB, P. C. 1984. The geochemical characteristics of granitoids in contrasting arcs and comments on magma sources. J. geol. Soc. Lond. 141,413 26. BUMGARDNER, J. E. 1976. Geology of the gabbro sill surrounding Mariscal Mountain, Big Bend National Park, Texas. M.A. Thesis, University of Texas at Austin (unpublished). BURT, E. R., III 1970. Petrology of the Mitchell Mesa Rhyolite, Trans-Pecos Texas. PhD Dissertation, University of Texas at Austin (unpublished). CAMERON,K. L. • CAMERON,M. 1976. Reconnaissance investigation of volcanic rocks from southeastern Chihuahua, Mexico: a comparison with the Sierra Madre Occidental and Trans-Pecos Texas. Geol. Soc. Am. Abstr. Progm. 8, 10. , --, BAGBY, W. C. & MOLL, E. J. 1979. Petrologic characteristics of mid-Tertiary igneous suites in Chihuahua, Mexico, and adjacent west Texas. In: WALTON, A. W. & HENRY, C. D. (eds) Cenozoic Geology of the Trans-Pecos Volcanic Field of Texas, Guidebook 19, pp. 1-3. Bureau of Economic Geology, University of Texas at Austin. , --, MOLL, E. J. & DRAKE, R. E. 1980. Petrologic characteristics of mid-Tertiary volcanic suites, Chihuahua, Mexico. Geology, 8, 87-91. CAMERON, M., CAMERON, K. L. & CEPEDA, J. C. 1982. Geochemistry of Oligocene volcanic rocks from the Chinati Mountains, west Texas. Geol. Soc. Am. Abstr. Progm. 14, 107. CARMAN, M. F., JR, CAMERON, M., GUNN, B., CAMERON, K. L. & BUTLER, J. C. 1975. Petrology of the Rattlesnake Mountain sill, Big Bend National Park, Texas. Geol. Soc. Am. Bull. 86, 177-93. CEPEDA, J. C. & HENRY, C. D. 1983. Oligocene volcanism and multiple caldera formation in the Chinati Mountains Caldera, Presidio County, Texas. Rep. Invest. 135, Bureau of Economic Geology, University of Texas at Austin. CHRISTOFFERSEN,R. & SCHEDL,A. 1980. Microstructure and thermal history of cryptoperthites in a dike from Big Bend, Texas. Am. Mineral. 65, 444-8. CLARKE, F. W. 1904. Analyses of rocks from the
-
429
laboratory of the U.S. Geological Survey, 1880 to 1903. U.S. Geol. Survey Bull. 228. COBB, R. C. & POTH, S. 1980. Superposed deformation in the Santiago and northern Del Carmen Mountains, Trans-Pecos Texas. New Mexico Geol. Soc. Guidebook, 31st Field Conf., Trans-Pecos Region, pp. 71-5. New Mexico Geol. Soc. DAILY, M. 1979. Age relations in alkaline rocks from the Big Bend region, Texas. In: WALTON, A. W. & HENRY, C. D. (eds) Cenozoic Geology of the TransPecos Volcanic Field of Texas, Guidebook 19, pp. 92-96. Bureau of Economic Geology, University of Texas at Austin. DASCI-I,E. J. 1969. Strontium isotope disequilibrium in a porphyritic alkali basalt and its bearing on magmatic processes. J. geophys. Res. 74, 560-5. , ARMSTRONG,R. L. & CLABAUGH,S. E. 1969. Age of Rim Rock dike swarm, Trans-Pecos Texas. Geol. Soc. Am. Bull. 80, 1819-24. DECAMP, D. W. 1981. Structural geology of Mesa de Anguila, Big Bend National Park, Trans-Pecos Texas. M.A. Thesis, University of Texas at Austin (unpublished). EIFLER, G. K., JR. 1943. Geology of the Santiago Peak quadrangle, Texas. Geol. Soc. Am. Bull. 54, 161344. 1951. Geology of the Barrilla Mountains, Texas. Geol. Soc. Am. Bull. 62, 339-53. ERICKSON, R. L. 1953. Stratigraphy and petrology of the Tascotal Mesa quadrangle, Texas. Geol. Soc. Am. Bull. 64, 1353-86. FOORD, E. E., MOORE, S. L., MARVIN, R. F. & TAGGART, J. E., JR. 1983. Petrology of the Proterozoic syenites at Pajarito Mountain and the Tertiary igneous complex of the Sierra Blanca region, Otero County, New Mexico. Abstracts, Mineralogical Society of America, Symp. on Alkaline Complexes, Wausau, WI. GIBBON, D. L. 1969. Origin and development of the Star Mountain Rhyolite. Bull Volcanol. 33, 438-74. & WYLLIE, P. J. 1969. Experimental studies of igneous rock series : the Farrington complex, North Carolina, and the Star Mountain Rhyolite, Texas. J. Geol. 77, 221-39. GOLDICH, S. S. & ELMS, M. A. 1949. Stratigraphy and petrology of the Buck Hill quadrangle, Texas. Geol. Soc. Am. Bull. 60, 1133-82. HEDGE, C. E. 1966. Variations in radiogenic strontium found in volcanic rocks. J. geophys. Res. 71, 611926. HENRY, C. D. & PRICE, J. G. 1984. Variations in caldera development in the Tertiary volcanic field of Trans-Pecos Texas. J. geophys. Res. 89, 876586. INDEST, S. & CARMAN, M. F., JR. 1979. Crystallization history of the Wildhorse Mountain quartz syenite intrusion and its relation to some other Big Bend intrusions. In: WALTON, A. W. & HENRY, C. D. (eds) Cenozoic' Geology of the Trans-Pecos Volcanic Field of Texas, Guidebook 19, pp. 72-82. Bureau of Economic Geology, University of Texas at Austin. IRVING, A. J. 1977. Origin of megacryst suites in basaltic dikes near 96 Ranch, west Texas and Hoover Dam, Arizona. Los, 58, 526. -
-
-
430 --
D. S. Barker
& FREY, F. A. 1984. Trace element abundances in megacrysts and their host basalts: constraints on partition coefficients and megacryst genesis. Geochim. cosmochim. Acta, 48, 1201-21. JUNGYUSUK, N. 1977. Petrology of the Christmas Mountains igneous rocks, Trans-Pecos Texas. M.A. Thesis, University of Texas at Austin (unpublished). KEITH, S. B. 1978. Paleosubduction geometries inferred from Cretaceous and Tertiary magmatic patterns in southwestern North America. Geology', 6, 51621. KING, P. B. 1965. Geology of the Sierra Diablo region, Texas. U.S. Geol. Surv. Prof. Pap. 480. LAUGHLIN, A. W., KRESS, V. C. & ALORICH, M. J. 1982. K-Ar ages of dike rocks, Big Bend National Park, Texas. Isoehron West, 35, 17-18. LEWIS, P. S. 1978. Igneous petrology and strontium isotope geochemistry of the Christmas Mountains, Brewster County, Texas. M.A. Thesis, University of Texas at Austin (unpublished). LIPMAN, P. W., PROSTKA,H. J. & CHRISTIANSEN,R. L. 1972. Cenozoic volcanism and plate-tectonic evolution of the western United States. I. Early and middle Cenozoic. Phil. Trans. R. Soc., Ser. A. 271, 217-48. LONSDALE, J. T. 1940. Igneous rocks of the TerlinguaSolitario region, Texas. Geol. Soc. Am. Bull. 51, 1539-1626. MCANULTY, W. N. 1955. Geology of Cathedral Mountain quadrangle, Brewster County, Texas. Geol. Soc. Am. Bull. 66, 531-78. MCCULLOH, R. P. 1977. Geology of Yellow Hill Quadrangle, Brewster County, Texas. M.A. Thesis, University of Texas at Austin (unpublished). MCDONOUGH, W. F., NELSON, D. O. & MATTISON, G. D. 1982. Major and trace element variation in a dynamically evolving silicic magma chamber. Geol. Soc. Am. Abstr. Progm. 14, 116. McDOWELL, F. W. 1979. Potassium-argon dating in the Trans-Pecos Texas volcanic field. In: WALTON, A. W. & HENRY, C. D. (eds) Cenozoic Geology of the Trans-Pecos Volcanic FieM of Texas, Guidebook 19, pp. 10-18. Bureau of Economic Geology, University of Texas at Austin. -1981. Chronology of Tertiary volcanism in TransPecos Texas. Geol. Soc. Am. Abstr. Progm. 13, 242. --& CLABAUGH,S, E. 1979. Ignimbrites of the Sierra Madre Occidental and their relation to the tectonic history of western Mexico. Geol. Soc. Am. Spec. Pap. 180, 113-24. MCKAY, G. A. & ROGERS, J. J. W. 1970. Ignimbrites of the Seven Springs Formation, Barrilla Mountains, western Texas. Geol. Soc. Am. Bull. 81, 33518. MCKNIGHT, J. F. 1970. Geology of Bofecillos Mountains area, Trans-Pecos Texas. Geol. Quadrangle Map 37, Bureau of Economic Geology, University of Texas at Austin. MAXWELL, R. A., LONSDALE,J. T., HAZZARD, R. T. & WILSON, J. A. 1967. Geology of Big Bend National Park, Brewster County, Texas. Publ. 6711, University of Texas at Austin. MEIGEN, W. & NACHREINER,F. 1925. Analysen einiger
Gesteine von Alter Pedroso und aus den ApacheMts. Zentbl. Mineral. Geol. Palaeontol., Abt. A, 331-3. MERRILL, R. B. & IRVING, A. J. 1977. Chemistry and phase relations of an orthopyroxene-bearing transitional alkalic basalt. Eos, 58, 526. MooN, C. G. 1953. Geology of Agua Fria quadrangle, Brewster County, Texas. Geol. Soe. Am. Bull. 64, 151-95.
MOUSTAFA, A. R. 1983. Analysis of Laramide and younger deformation of a segment of the Big Bend region, Texas. PhD Dissertation, University of Texas at Austin (unpublished). MUEHLBERGER,W. R. 1980. Texas Lineament revisited. New Mexico Geol. Soc. Guidebook, 31st Field Conf., Trans-Pecos Region, pp. 113-21. New Mexico Geol. Soc. - - , BELCHER,R. C. & GOETZ, L. K. 1978. Quaternary faulting in Trans-Pecos Texas. Geology, 6, 337-40. NELSON, D. O., McDoNOUGH, W. F. & MATTISON, G. D. 1981. Trace element geochemistry of the Sawtooth Mountain syenites, Trans-Pecos magmatic belt of west Texas. Geol. Soc. Am. Abstr. Progm. 13, 259. OGLEY, D. S. 1979. Eruptive history of the Pine Canyon caldera, Big Bend Park. In: WALTON, A. W. & HENRY, C. D. (eds) Cenozoic Geology of the TramPecos Volcanic Field of Texas, Guidebook 19, pp. 67-71. Bureau of Economic Geology, University of Texas at Austin. OSANN, C. A. 1896. Beitrage zur Geologie und Petrographie der Apache (Davis) Mountains, Westtexas, Tschermaks Mitt. N.F. 15, 394-456. PARKER, D. F. 1972. Stratigraphy, petrography and K Ar geochronology of volcanic rocks, northeastern Davis Mountains, Trans-Pecos Texas. M.A. Thesis, University of Texas at Austin (unpublished). 1983. Origin of the trachyte-quartz trachyteperalkalic rhyolite suite of the Oligocene Paisano volcano, Trans-Pecos Texas. Geol. Soc. Am. Bull. 94, 614-29. 1986. Stratigraphic, structural, and petrologic development of the Buckhorn caldera, northern Davis Mountains, Trans-Pecos Texas. In: PRICE, J. G., HENRY, C. D., PARKER, D. F. & BARKER, D.
S. (eds)Igneous Geology of Trans-Pecos Texas. Field Trip Guide and Research Articles, Guidebook 23, pp. 286-302. Bureau of Economic Geology, University of Texas at Austin. - - & McDOWELL, F. W. 1979. K - A t geochronology of Oligocene volcanic rocks, Davis and Barrilla Mountains, Texas. Geol. Soc. Am. Bull. 90, 110010. PRICE, J. G. & HENRY, C. D. 1984. Stress orientations during Oligocene volcanism in Trans-Pecos Texas: Timing the transition from Laramide compression to Basin and Range tension. Geology, 12, 238-41. --, BARKER, D. S. & PARKER, D. F. Alkalic rocks of contrasting tectonic settings in TransPecos Texas. In : MORRIS, E. M. & PASTERIS,J. D. (eds) Mantle Metasomatism and Alkaline Magmatism. Geol. Soc. Am. Spec. Paper 215, in press. -& RUBIN, J. N. 1986a. Petrology of the Marble Canyon stock, Culberson County,
Tertiary magmatism in Trans-Pecos Texas Texas. In: PRICE, J. G., HENRY, C. D., PARKER, D. F. & BARKER, D. S. (eds) Igneous Geology of Trans-Pecos Texas. FieM Trip Guide and Research Articles, Guidebook 23, pp. 303-19. Bureau of Economic Geology, University of Texas at Austin. --,--, PARKER,D. F. & BARKER,D. S. (eds) 1986b. Igneous Geology of Trans-Pecos Texas. Field Trip Guide and Research Articles, Guidebook 23, Bureau of Economic Geology, University of Texas at Austin. d~ STANDEN, A. R. 1983. Annotated bibliography of mineral deposits in Trans,Pecos Texas. Min. Resources Circ. 73, Bureau of Economic Geology, University of Texas at Austin. PRIOR, G. T. 1903. Contributions to the petrology of British East Africa. Mineral. Nag. 13, 228-63. ROBIN, C. 1974. Premieres donn~es sur les sbries magmatiques alcalines de la Sierra de Tamaulipas (Est mexicain). C.R. Acad. Sci. Paris, Ser. D, 279, 1741-4. SCHIEFFER, J. H. & MATTISON,G. D. 1982. Nature and origin of alkalic and calcic veinlets in xenoliths from the Terlingua district, west Texas. Geol. Soc. Am. Abstr. Progm. 14, 609-10. & NELSON,D. O. 1981. Petrology and geochemistry of megacrysts, xenoliths, and their host basalts from the Terlingua mercury district of west Texas. Geol. Soc. Am. Abstr. Progm. 13, 547. , MATTISON, G. D. & NELSON, D. O. 1982. The mineralogy and geochemistry of the igneous rocks of the Terlingua district, Brewster County, Texas. Geol. Soc. Am. Abstr. Progm. 14, 135. SCHWARZER, R. R. & JOHNSON, M. 1976. Oxygen isotope geochemistry of a continental alkali olivine basalt rock series. Geol. Soc. Am. Abstr. Progm. 8, 64-5. SHARP, J. E. 1979. Cave Peak, a molybdenummineralized breccia pipe complex in Culberson County, Texas. Econ. Geol. 74, 517-34. SMITH, W. C. 1931. A classification of some rhyolites, trachytes and phonolites from part of Kenya Colony, with a note on some associated basaltic rocks. Q. J. geol. Soc. Lond. 87, 212-56.
-
-
431
STEWART, R. M. 1982. A stratigraphic and petrologic characterization of the Alamo Creek Basalt, Big Bend National Park, Texas. Geol. Soc. Am. Abstr. Progm. 14, 137. ST. JOHN, B. E. 1966. Geology of Black Gap area, Brewster County, Texas. Geol. Quadrangle Map 30, Bureau of Economic Geology, University of Texas at Austin. SWANSON, E. R. & McDOWELL, F. W. 1984. Calderas of the Sierra Madre Occidental volcanic field, western Mexico. J. geophys. Res. 89, 8787-99. THORNTON, C. P. & TUTTLE, O. F. 1960. Chemistry of igneous rocks: I. Differentiation index. Am. J. Sci. 258, 664-84. UDDEN, J. m. 1907. A sketch of the geology of the Chisos country, Brewster County, Texas. Texas Univ. Bull. 93, 101 pp. WALTON, A. W. & HENRY, C. D. (eds) 1979. Cenozoic Geology of the Trans-Pecos Volcanic Field of Texas, Guidebook 19, Bureau of Economic Geology, University of Texas at Austin. WEAVER, S. D., SCEAL, J. S. C. & GIBSON, I. L. 1972. Trace-element data relevant to the origin of trachytic and pantelleritic lavas in the East African rift system. Contrib. Mineral. Petrol. 36, 181-94. WILLIAMS, L. A. J., MACDONALD, R. & CHAPMAN, G. R. 1984. Late Quaternary caldera volcanoes of the Kenya Rift Valley. J. geophys. Res. 89, 855370. WILSON, J. A. 1971. Vertebrate biostratigraphy of Trans-Pecos Texas. In: SEEWALD,K. & SUNDEEN, D. (eds) The Geologic Framework of the Chihuahua Tectonic Belt, West Texas Geol. Soc. Publ. 71-59, pp. 156-66. YATES, R. G. & THOMPSON, G. A. 1959. Geology and quicksilver deposits of the Terlingua district, Texas. U.S. Geol. Survey Prof. Pap. 312. YUND, R. A. & CHAPPLE, W. M. 1980. Thermal histories of two lava flows estimated from cryptoperthite lamellar spacings. Am. Mineral. 65, 43843.
DANIEL S. BARKER, Department of Geological Sciences, The University of Texas at Austin, Austin, TX 78713-7909, U.S.A.
The Monteregian Hills and White Mountain alkaline igneous provinces, eastern North America G. Nelson Eby SUMMARY: The Monteregian Hills and White Mountain provinces consist of stocks, plugs, ring-dyke complexes and several large granite bodies emplaced into Precambrian gneisses, flat-lying Cambro-Ordovician sediments and the deformed Lower Palaeozoic section of the Appalachian fold belt. Felsic rocks dominate in the Appalachian fold belt, while elsewhere mafic and ultramafic rocks are significant components of the plutons. Igneous activity extended from 240 to 90 Ma ago with two major periods of magmatism, correlated with events in the opening of the N Atlantic Ocean, occurring between 200165 Ma and 140-110 Ma ago. Five major rock series have been identified: (1) undersaturated CO2-rich rocks, carbonatite and aln6ite; (2) moderately to strongly undersaturated diorites-nepheline syenites; (3) slightly undersaturated to slightly oversaturated pyroxenites-gabbros-diorites-syenites; (4)alkali syenite-quartz syenite-granite; (5)metaluminous biotite granite. Series (1), (2) and (3) magmas were drawn from an isotopicallydepleted mantle which was enriched in incompatible elements shortly before or synchronous with melting. These magmas were produced by variable degrees of melting of garnet or spinel lherzolite. Series (4) and (5) magmas represent partial melts of a heterogeneous crustal section consisting of both meta-sedimentary and meta-igneous rocks of either Grenville (Precambrian) or Lower Palaeozoic age.
Introduction The Monteregian Hills and White Mountain alkaline igneous provinces consist of a group of plutons and associated dykes which were emplaced 240-90 Ma ago (Foland & Faul 1977; Eby 1984a) into the crust of southern Quebec and northern New England (Fig. 1). The New England Seamount chain, emplaced between 103 and 82 Ma ago, is considered to be an oceanic extension of the continental alkaline magmatism (Duncan 1982). The plutons were intruded into a variety of geological settings: the Precambrian Grenville province of the Canadian shield, the flat-lying Cambro-Ordovician sediments of the St. Lawrence Lowland and the deformed Lower Palaeozoic section of the Appalachian orogen. The plutons vary in lithology from pyroxenites to peralkaline granites, and both silica-undersaturated and silica-oversaturated rocks are found, often in spatial association. In regions of thick continental crust felsic silica-saturated rocks predominate, while in regions of thinner continental crust ultramafic and mafic rocks predominate. The igneous activity is episodic, with most of the intrusions emplaced between 200-165 and 140-110 Ma ago (Foland & Faul 1977; Eby & Creasy 1983; Eby 1984a). McHone & Butler (1984) have related the two major periods of igneous activity to events in the opening of the N Atlantic Ocean. The major emphasis of this review will be on the younger period (140-110 Ma) of igneous activity, but the older period
(200-165 Ma) of igneous activity will be briefly considered in the following section.
Older White Mountains The igneous activity of the Older White Mountain period is essentially confined to central and northern New Hampshire (Fig. 1). This is the area of greatest current crustal thickness as deduced from the seismic studies of Taylor & Toks6z (1982). Projections of the bulk compositions of the granitic units associated with the various plutons onto the Qz + Ab + Or system suggest that these rocks crystallized at a Pn2o of about 2 kb (Czamanske et al. 1977; Eby, unpublished data), which indicates a depth of emplacement of about 6 km. Essentially all the rocks are silica saturated, with the notable exception of Red Hill, and felsic rocks greatly predominate over mafic rocks. Three distinct petrological associations can be recognized: (1) gabbro-diorite-monzonite; (2) alkali syenite-quartz syenitegranite; (3) metaluminous biotite granite. The geology and geochemistry of the White Mountain batholith, Pliny Range, Belknap Mountain complex and Red Hill complex will be briefly considered as representative examples of this older period of igneous activity. The White Mountain batholith is volumetrically the most important member of the older White Mountain period (see Fig. 1 for locations). Creasy (1974) described the geology of the
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 433-447.
433
G. N. Eby
434 ~'Tk.,'~Pre_C/'~ ft.
/
~ ' / ~ Ac-7/7] - " [ e r e ~ ~ '# n ' o n
,#.oo
,'2.
~.,
i
ml i
/
/
I
,Maine-
--'-- --~
/
_ _
~ --
(
~
7'o.
. ~
. , ~ - ~ ' - ~ T///////I//////~ i-- 9 -
~ F--.--------
t--!~
fi' _ _
.
LI~". . . . . . . / ~ ~ : - -
II
a~ / /_
-
/
FIG. 1. Geological setting and locations of Monteregian Hills and White Mountain plutons : horizontally shaded areas, plutons emplaced between 240 and 200 Ma ago; diagonally shaded areas, plutons emplaced between 200 and 165 Ma ago; black areas, plutons emplaced between 140 and I00 Ma ago. Specific plutons are indicated by number : 1, Oka; 2, Mount Royal; 3, Mount St. Bruno; 4, Mount St. Hilaire; 5, Mount Rougemont; 6, Mount Johnson; 7, Mount Yamaska; 8, Mount Shefford; 9, Mount Brome; 10, Mount Megantic; 11, Pliny Range; 12, White Mountain batholith; 13, Red Hill complex; 14, Ossipee ring-complex; 15, Cuttingsville; 16, Belknap Mountain complex; 17, Merrymeeting Lake; 18, Mount Pawtuckaway. LGB, limit of Grenville basement. batholith in some detail. It consists of several overlapping caldera complexes. The two major units are an alkali granite and a metaluminous biotite granite, which have been intruded by ring dykes of alkali syenite-quartz syenite composition. Peralkaline rhyolite and trachyte volcanic sections are locally preserved as down-dropped blocks and attest to the prior existence of a substantial volcanic edifice. For example, at Moat Mountain the volcanic section is approximately 3000 m thick. The various lithological units of the batholith were emplaced between 196 and 168 Ma ago (Eby & Creasy 1983). A wide range of initial Sr isotope ratios has been determined for the various units from relatively non-radiogenic ratios of 0.7039 to relatively radiogenic ratios of 0.7165, while the Pb isotope ratios have a much more restricted range (Eby & Creasy 1983). The isotopic data suggest that the source regions for the melts had relatively uniform
U/Pb and Th/Pb ratios but highly variable Rb/ Sr ratios. The isotopically and chemically primitive samples have small to moderate negative Eu anomalies and intermediate La/Yb ratios (1020), consistent with the generation of these magmas from lower-crustal material containing residual plagioclase and pyroxene. The subsequent fractionation of these magmas was controlled by the removal of alkali feldspar+ pyroxene +__fayalite (Creasy & Eby 1983). The geology of the Pliny Range has been described in some detail by Czamanske et al. (1977). Four phases of igneous activity were recognized. The rocks constituting each phase, in chronological order from oldest to youngest, are as follows: (1)coarse syenite and mediumgrained syenite; (2)diorite and quartz monzodiorite; (3)porphyritic quartz monzonite, pink biotite granite, granite prophyry and hastingsite quartz syenite; (4) hastingsite biotite granite and Conway granite. The diorite forms a central plug and is surrounded by ring dykes composed of quartz monzodiorite and the group (3) rocks. The group (4) rocks occur as small stocks which cut the other units. Czamanske et al. (1977) concluded that the rock and mineral chemistry precluded the generation of the various rock types by differentiation of a single parent magma, and suggested instead that multiple fusion events at different levels in the crust best explained the chemical data. Randall et al. (1983) found that the rocks of the Pliny range were intruded in a short time interval centred around an age of 183 Ma. Initial Sr isotope ratios varied between 0.7035 and 0.707 and initial Nd isotope ratios varied between 0.5126 and 0.5121, and the ratios were inversely correlated indicating extensive interaction of a mantle-derived magma with the crust. The initial magma had an eNa of approximately + 4 and an initial Sr isotope ratio of 0.7035, values which are appropriate for a source depleted in large-ion lithophile (LIL) elements. Modell (1936) provided the first detailed description of the geology of the Belknap Mountain complex. A substantial amount of additional petrographic and geochemical work was subsequently carried out by Loiselle (1978). The complex consists of a diorite stock which is surrounded by a series of partial ring dykes consisting of monzodiorite and quartz syenites. Two small syenite stocks outcrop within the complex and were apparently emplaced prior to the ring dykes. The last event is the intrusion of a central mass of Conway granite. Loiselle (1978) obtained a Rb-Sr whole-rock isochron age of 168 Ma for the complex. The diorite, a monzodiorite, some of the syenites and the granite had low initial Sr isotope ratios of about 0.7035, while
The Monteregian Hills another monzodiorite and the quartz syenites had higher initial Sr isotope ratios in the range 0.7040-0.7045. Trace-element data indicated that the mafic rocks were derived from an alkali basalt magma. The syenites and granites could not be derived from this mafic melt by simple crystal fractionation, and more complex models were invoked involving assimilation of metamorphosed tholeiitic basalt at the base of the crust and partial melting of an alkaline gabbro. The Red Hill complex is unusual in that it contains nepheline-normative rocks, in sharp contrast with the other plutons emplaced within the Appalachian fold belt which are silica saturated. Size (1972) described the geology of this complex in some detail. The complex is circular in outline and encompasses an area of some 20 km 2. The apparent order of emplacement is an outer coarse-grained syenite, inward of which is a nepheline-sodalite syenite. The central portion consists of a syenite unit into which were subsequently emplaced three pluglike intrusions of syenite and fine-grained granite. Foland & Friedman (1977) found that three of the syenite units gave a Rb-Sr whole-rock isochron age of 198 Ma and an initial Sr isotope ratio of 0.70330. The interior fine-grained granite yielded a mineral Rb-Sr isochron age of 188 Ma and an initial Sr isotope ratio of 0.70573. Initial ratios for the nepheline-sodalite syenite ranged between 0.7035 and 0.7065. On the basis of their Sr and O isotopic data, Foland & Friedman (1977) concluded that the range of rock types was the result of interactions of mantle-derived magmas with various crustal components. Currie et al. (1986) have recently suggested that the interaction of mafic magmas with old brines stored in crustal reservoirs could lead to the production of sodalite-rich rocks. A similar explanation may be applicable for the nephelinesodalite syenite at Red Hill. The data currently available for the older White Mountain magma series are consistent with a model involving the generation of magmas in the mantle and the subsequent interaction or thermal exchange of these magmas with crustal materials, leading to felsic residua. The initial magmas are most probably alkali olivine basalts which were extracted from a depleted mantle source region. These magmas ascended various distances into the overlying crust and were either contaminated by crustal material or served as heat sources for local crustal melting. In this area the crustal section is quite complex and consists of mafic and felsic meta-volcanics and a variety of pelitic and feldspathic meta-sedimentary rocks. Previous melting episodes during the Ordovician and Devonian extracted granitic
435
components. Thus, depending upon the degree of interaction between the mafic magmas and the crust, the depth of melting and the lithology of the crust in the zone of melting, a variety of distinct magma types can be generated. In general, with the notable exceptions of the Pliny Range and the Belknap Mountains complex, mafic rocks are not exposed at the current level of erosion. The felsic rocks show a wide range in initial Sr isotopic ratios, but a number of the felsic units have low initial ratios which suggest derivation from mantle-like source regions. Whether these rocks are the products of differentiation of a mantle melt or are largely of crustal origin is still an open question. Within the White Mountain batholith, the bulk of the rocks have Pb isotope ratios appropriate for the crust at their time of emplacement, and the Pb isotopes do not reveal any interaction or derivation of these melts from Precambrian material. The older White Mountain plutons define a relatively localized thermal anomaly which persisted for approximately 30 Ma, a constraint which must be considered when developing a dynamo-thermal model, i.e. hot-spot versus decompression melting versus leaky transform, for this period of igneous activity.
Younger White Mountains and Monteregian Hills The plutons forming the younger White Mountain and Monteregian Hills provinces are distributed over a much greater area than the older White Mountain plutons and are intruded into several geological provinces (Fig. 1). Those plutons which occur within the Appalachian fold belt have a significant felsic component, while those outside the fold belt have a significant mafic component. The locations of the plutons described in the following sections can be found on Fig. 1.
Younger White Mountains Included in this group are the plutons located in the Appalachian fold belt and plutons of the same age found in Vermont. Volumetrically these plutons tend to be smaller than those belonging to the older White Mountain series. The geology of the Ossipee ring-complex, Merrymeeting Lake, Mount Pawtuckaway and Cuttingsville will briefly be considered as representative examples of this group of plutons. The Ossipee ring-complex was originally mapped by Kingsley (1931) and the geology was
436
G. N. Eby
subsequently modified and petrochemical data obtained by Carr (1980). The complex is circular in outline, 14 km in diameter, and consists of an outer ring-dyke of medium- to coarse-grained porphyritic quartz syenite and an interior section of older basalts, intrusive rhyolite and younger coarse-grained biotite granite. Carr (1980) found relations within the basalt sequence to be quite complex. Two pre-rhyolite basalts occur, an earlier coarsely-porphyritic variety and a later dense massive variety. A third basalt, which is post-rhyolite, is also dense, massive and sparsely porphyritic. The biotite granite occurs as a circular mass which makes up the centre of the intrusion. Carr (1980) concluded that the rhyolite was emplaced explosively at a shallow level, and that the plutonic rocks crystallized at a depth of about 3 km. Gravity and aeromagnetic data (Sharp & Simmons 1978) indicate that a vertical dense mafic plug underlies the central portion of the complex and extends to depths of 4-7 km. The geology of Merrymeeting Lake has been described by Quinn & Stewart (1941). The bulk of the pluton consists of biotite granite, similar to that found at the Ossipee ring-complex and elsewhere in the White Mountains. Leucorhyolite is found at high elevations and occurs as inclusions in the biotite granite. Flow structures and other evidence of volcanic origin have not been observed for this unit. A discontinuous body of granodiorite occurs along the NW margin of the biotite granite, and this unit has apparently been intruded by the biotite granite. This group of rocks forms a circular structure with a diameter of approximately 10 km. A small plug of dioritegabbro is found along the eastern margin, and this body has been intruded by a granite porphyry. An aeromagnetic maximum centred on the diorite-gabbro body indicates the presence of mafic rocks at depth (Weston Geophysical Research Inc. 1976). The geology of Mount Pawtuckaway was originally described by Roy & Freedman (1944) and was subsequently modifed by Eby (1984b). The pluton is a small circular structure approximately 3 k m in diameter. A coarse-grained monzonite-syenite ring-dyke rims about 70% of the intrusion. A central plug of fine-grained monzonite has been intruded by coarse-grained monzonite. Between the two monzonite-syenite units is a body of mafic rock ranging in composition from gabbro to monzodiorite. A foliated subfacies of the diorite can be distinguished and the foliation dips steeply inward. Gravity and aeromagnetic data (Griscom 1962; Joyner 1963) indicate that the plug-like structure of the complex extends to depth. Field relations and petrography establish that several separate
mafic magmas were intruded up the conduit, followed by emplacement of the fine-grained central monzonite plug, and then subsidence of this block and the crystallization of the coarsegrained monzonite-syenite unit of the peripheral ring-dyke. The geology of the Cuttingsville complex was originally described by Eggleston (1918) and subsequently modified by Pierson (1970). The intrusion is elliptical in outline (2.3 km • 1.7 km) and is a pipe-like body. The intrusion is composite and consists of a nearly-circular core of foyaite which has been intruded by an essexite-plagiofoyaite unit. Diorites and porphyritic diorite cut the foyaite and are in turn intruded by the essexite-plagiofoyaite unit. A syenite-quartz syenite unit intrudes all the previous lithologies. This complex is distinctly different from those previously discussed in that feldspathoid-bearing and quartz-bearing rocks exist in close spatial association.
Monteregian Hills The Monteregian Hills consist of alkaline stocks and plugs intruded into a variety of geological settings (Fig. 1). At their present level of exposure, the cover must have been less than 3 km as evidenced by the concordance of apatite fissiontrack ages (reset at temperatures above 100~ and Rb-Sr whole-rock isochron ages (Eby 1984a). Explosive igneous activity, represented by diatremes and breccia pipes, is largely confined to the western end of the province. Elsewhere the emplacement seems to have been relatively passive. Where developed, primary igneous foliation is steeply dipping towards the centre of the intrusions, suggesting crystallization from the walls inward. Most of the intrusions are surrounded by high-grade metamorphic aureoles, and near the contacts the country rocks dip steeply inward, suggesting that local melting caused them to founder into the intruding magma (Philpotts 1969). The Oka carbonatite complex (Gold 1963, 1972) consists of two ring-like structures emplaced into Grenville-age gneisses. The cores of the two rings are occupied by s6vite while the outer portions of the structures consist of intercalated arcuate dykes of okaite (melilite-rich rock), jacupirangite, ijolite and s6vite. An irregular zone of fenitization surrounds the complex. Aln6ites and intrusive breccias cut the other rocks and represent the last stage of igneous activity. Mount Royal (Woussen 1969; Eby 1984c) largely consists of a heterogeneous mixture of pyroxenite and gabbro. The contact between the
The Monteregian Hills two lithologies is gradational over short distances. Cumulus olivine (Fo74_68)and titanaugite are the essential minerals and amphibole (kaersutite) is locally an important constituent. Plagioclase (An83_44) varies from an interstitial phase to an essential component. Leucogabbro intrudes the gabbro-pyroxenite body and apparently represents a differentiate of the magma which crystallized the gabbros and pyroxenites. None of these rocks contain feldspathoid minerals, but they are slightly silica undersaturated (nepheline normative). The gabbro--pyroxenite body is intruded by nepheline-bearing diorite and several generations ofnepheline- and sodalite-bearing monzonite and syenite dykes. Mount St. Bruno (Philpotts 1976; Eby 1984c) essentially consists of a gabbro-pyroxenite body similar to that at Mount Royal. The olivine (Fo84_74) and plagioclase (An84-55) compositions are somewhat more restricted, and kaersutite is relatively less abundant while olivine is relatively more abundant. Locally, quartz is found in the gabbro facies, presumably in response to the contamination of the magma by siliceous country rock. A slightly oversaturated quartz-bearing syenite intrudes the mafic rocks and represents the last stage of igneous activity. Mount St. Hilaire (Currie 1983) is a composite intrusion consisting of a western half which is largely gabbroic in character and an eastern half composed of peralkaline strongly undersaturated syenites. Igneous breccias occur between the two main lithologies. The gabbroic rocks are locally pyroxene rich and appear to grade into an amphibole-rich pyroxene-poor facies. Olivine is rarely found and feldspathoids are absent. These rocks are intruded by nepheline-bearing diorite and dykes of peralkaline syenite similar to the rocks which form the eastern half of the intrusion. Sodalite is locally abundant in the syenites (up to 50 vol.%), and a number of rare alkaline minerals have been identified at Mount St. Hilaire. Mount Rougemont (Philpotts 1976; Jacobson 1981) is essentially composed of pyroxenite and gabbro. The two lithologies are generally intercalated, although a mappable gabbro body occurs along the western side of the intrusion. Olivine (Fo81_67) is locally abundant and plagioclase (An95_36) varies from an interstitial phase (pyroxenite) to a major component (gabbro). Feldspathoids are absent. Mount Johnson (Pajari 1967) is a small circular intrusion consisting of a fine-grained essexite core surrounded in sequence by medium- to coarsegrained essexite, pulaskite and nepheline syenite. Nepheline is found in all the units. Mount Yamaska (Gandhi 1966) is largely composed of pyroxenite (yamaskite) and gabbro.
437
Two separate pyroxenite intrusions have been mapped. The pluton is partly rimmed by akerite (augite monzonite) which is apparently a hybrid rock formed by interaction of the marie magma with the country rock (Eby, unpublished data). None of these rocks carry feldspathoids, and quartz is found locally. The pyroxenites and gabbros were subsequently intruded by essexite and nepheline-bearing syenites. Mount Shefford (Valiquette & Pouliot 1977; Eby 1985a) appears to consist of two overlapping circular structures. The core of the larger structure is diorite-gabbro, locally amphibole rich, while the smaller structure has a core of breccias and sub-volcanic porphyries and a partial rim of diorite. The two cores are surrounded by arcuate units of micropulaskite, pulaskite and nordmarkite. Several small bodies of nepheline-bearing diorite are found in the pulaskite. Mount Brome (Valiquette & Pouliot 1977; Eby 1985a), the second largest of the Monteregian plutons, consists of an arcuate gabbro body which partly encloses a large mass of pulaskite. The gabbro body consists of several cyclical units which represent cumulates from separate magma pulses. Nordmarkite, which appears to be a facies of the pulaskite that has undergone crustal contamination, occurs adjacent to the pulaskite and locally along the outer edge of the gabbro. Nepheline-bearing diorites and monzonites, foyaites, tinguaites and laurdalite have been intruded into the earlier syenite units. Mount Megantic (Reid 1976; Eby 1985a) is the largest of the Monteregian plutons. It consists of a central granite plug which is almost completely surrounded by a nordmarkite ring-dyke. Gabbros and diorites occupy the area between the ringdyke and the central granite plug. The intrusion of the gabbros, diorites and nordmarkites apparently occurred almost simultaneously, since a mixed zone is found between the mafic and felsic rocks. All the rocks of the pluton are silica saturated to oversaturated, and the central biotite granite is similar to the granites found in the White Mountains. From the previous discussion of the individual plutons, and reference to Fig. 1, it is apparent that the plutons to the W of the folded Appalachians are largely composed of ultramafic and mafic rocks. Mount Yamaska, which is just E of Logan's Line, is also largely ultramafic to mafic in character. Within the Appalachian province, felsic rocks become an important component of the plutons, ultramafie rocks are essentially absent and the areal extent of the plutons increases. Two distinct petrographic groups can be identified, one consisting of rocks that do not contain feldspathoids but may contain quartz
438
G. N. Eby
and the other consisting of feldspathoid-bearing rocks. Where intrusive relations can be observed, the feldspathoid-bearing units intrude the nonfeldspathoid-bearing units.
4
I
Dykes The dykes of the Monteregian Hills and younger White Mountains can be divided into three groups: (1) K20-rich aln6ites; (2) moderately to strongly silica-undersaturated monchiquites, basanites, camptonites and feldspathoid-bearing felsic dykes; (3) slightly silica-undersaturated to slightly silica-oversaturated camptonites, alkali olivine basalts and quartz-bearing felsic dykes. The group (1) dykes are confined to the western end of the Monteregian Hills, group (2) dykes predominate in the Montreal area and the Lake Champlain valley, and the group (3) dykes predominate in the eastern portion of the province (Hodgson 1968; McHone 1978; McHone & Corneille 1980). Hodgson (1968) distinguished two major differentiation trends for the dykes of the Monteregian Hills, one a silica-saturated trend leading to bostonites and the other a silicaundersaturated trend leading to nepheline syenires and tinguaites.
Geochronology All the currently available age dates for the Monteregian Hills and younger White Mountains are plotted in Fig. 2. When necessary these ages have been recalculated using the decay constants recommended by Steiger & Jaeger (1977). In the cases where multiple dates have been obtained for a rock series, only the average age is plotted. As has been previously noted (Eby 1984a), the Monteregian ages fall into two distinct clusters, one between 118 and 122 Ma which includes all the strongly silica-undersaturated units and the other between 130 and 140 Ma which contains all the slightly undersaturated to saturated units. A quartz-bearing syenite unit at Mount Shefford falls in the same group as the strongly undersaturated rocks, but the melt from which these rocks crystallized is inferred to have a crustal source (Eby 1985a). In the case of the composite intrusions Mounts Royal, St. Hilaire, Shefford and Brome, which contain both petrographic groups, a hiatus of 10-15 Ma occurred between the intrusive events. Few plutons which correspond to the older intrusive period of the Monteregian Hills are found within the White Mountain province. However, a number of plutons were emplaced during the time corresponding to the younger event in the Monteregian Hills. Significantly, none of these plutons, which were emplaced
I
1
I
I
m
White Mountoins I
2
3
[
2
3
E
g2 o
m
,
Dykes
4t o
~ N I00
,
hl,N 108
I 116
I 124
,1~
I
NI
1.32
140
Age (Me)
F]G. 2. Frequency of ages for Monteregian Hills and younger White Mountain plutons and dykes. Where multiple dates were determined for a rock series, the average age is plotted. Letters identify the ornamentation for each rock type or rock series. Monteregian Hills : U, undersaturated series; N, Mount Shefford nordmarkite; G, Mount Megantic granite; S, slightly undersaturated to saturated series. White Mountains : same ornamentation as Monteregian Hills except G, granite. Dykes: M, monchiquites; A, aln6ites; C, camptonites; B, basanites; O, alkali olivine basalt. Data sources: Foland & Fau11977; Eby 1984a, 1984b, 1985b; Krueger, personal communication, 1984; McHone 1984; Currie et al. 1986. within the folded Appalachians, contain strongly silica-undersaturated rocks. A third and younger period of igneous activity is found in the White Mountains and a younger intrusion of strongly undersaturated rocks (Cuttingsville) is found in south-western Vermont. The dykes encompass the total time span represented by the plutons of the Monteregian Hills and younger White Mountains. Both silicaundersaturated and silica-oversaturated dykes are found in all the age groups. As previously noted (Eby 1985b), there is no apparent geographical pattern to the ages of the dykes.
Chemistry and isotope geology The oldest period of igneous activity in the Monteregian Hills is largely represented by ultramafic to mafic rocks, most of which show cumulate textures. Felsic rocks become an important component of this sequence at the eastern end of the province. These rocks rarely contain feldspathoids, but are often nepheline normative
The Monteregian Hills owing to the presence of abundant amphibole. On a plot of total alkalis versus silica (Fig. 3) this sequence is represented by the field labelled mafic cumulates and the syenites at Mounts Brome and Shefford. The trend of magmatic evolution is apparently from slightly undersaturated to slightly oversaturated liquids, and Sr isotopic data indicate that crustal contamination played a role in producing the oversaturated rocks (Eby 1984c, 1985a). In all these rocks N a 2 0 is significantly enriched with respect to K20, rareearth elements (REE) and incompatible elements are enriched and MgO, CaO and TiO2 are relatively high, as expected for subalkaline to alkaline igneous rocks. Similar sequences to those described above within the Appalachian fold belt (including the eastern-most Monteregian pluton, Mount Megantic) differ in some important respects. Ultramafic rocks are generally absent and the mafic sequences consist of gabbros and diorites, most of which have cumulate textures. N a 2 0 is still significantly enriched with respect to K20, but
MgO and TiO2 concentrations are relatively lower and the absolute concentrations of R E E and incompatible elements tend to be lower. These sequences apparently evolve from slightly undersaturated or just saturated melts towards silica-oversaturated residua, as shown by the nordmarkites at Mount Megantic. In the case of Mount Megantic at least, Sr isotopic data indicate that crustal contamination played a role in developing the silica-oversaturated residua (Eby 1985a). Mount Pawtuckaway, Ossipee and Cuttingsville (in the CS field) are examples of this magmatic trend (Fig. 3). It is concluded that there are two distinct sequences of igneous rocks which represent the crystallization products of slightly silica-undersaturated magmas which were evolving towards silica-saturated or silicaoversaturated residua. The one sequence is represented by the largely ultramafic to mafic plutons of the Monteregian Hills, whilst the other sequence is represented by the mafic plutons of the White Mountain province. The younger period of igneous activity in the
IGF
//
/
1-
1
,zJ-
/
439
/
8 L~
/
/ ~-~ ~ ~ " ~
.
.
.
.
/ / ! / D ' " P,+i"s')+r',,++'Ls t:
,-"~'--J
~~
1 S
V
/
,tl +STl ",~..L/ /,~/"
/
.
~S~
4
|
1 L
-I
~:--:-.-":'~:u
I
--/7"
J
+
SiO2 FIG. 3. Total alkalis versus silica for Monteregian Hills and the younger White Mountain plutons: ST, Mount St. Hilaire; BR, Mount Brome; SH, Mount Shefford; SHN, Mount Shefford nordmarkites; OS, Ossipee syenites. Lines with arrowheads indicate trends for the various rock series. Fields are shown where no trends are found. RY I, RY II and RY III are dyke sets from Mount Royal. The field labelled 'Mafic cumulates' includes all the medium- to coarse-grained ultramafic and mafic rocks from the various plutons. The field labelled 'Granites' includes the granites at Mount Megantic, Ossipee and Merrymeeting Lake. The two solid circles connected by the broken line labelled Johnson represent the possible immiscible pair at Mount Johnson (Eby 1979). The bold broken line divides those magmas which evolve towards more silica-undersaturated compositions (US) from those which evolve towards silica-oversaturated compositions (CS). The alkaline-subalkaline boundary is from Miyashiro (1978). Data sources: Gold 1972; Laurent & Pierson 1973; Carr 1980; Jacobson 1981 ; Currie 1983; Eby 1984c, 1985a, unpublished data.
G. N. Eby
440
Monteregian Hills and some of the units exposed at Cuttingsville define another sequence of igneous rocks which consists of feldspathoidbearing varieties. N a 2 0 is significantly enriched with respect to K 2 0 and these rocks have very high alkali contents relative to silica. REE and incompatible elements are highly enriched in this sequence. The initial magmas are silica undersaturated and evolve towards more strongly undersaturated residua. The silicate rocks in the Oka complex also belong to this sequence. The alkalis versus silica trends for this sequence are shown in Fig. 3 (in the US field). Two of the felsic dyke sets at Mount Royal ( R Y I and R Y I I ) are apparent exceptions to the trend of increasing silica undersaturation, but Sr isotopic data (Eby 1984c) support the conclusion that crustal contamination is responsible for the decrease in silica undersaturation. Currie et al. (1986) have recently suggested that the strongly nepheline-normative character, and abundant sodalite, of the nepheline syenites at Mount St. Hilaire (Fig. 3) is due to the interaction of an alkaline magma with an NaC1 brine stored in the crust. For Mount Johnson (Fig. 3) the two full circles represent the compositions of the two immiscible liquids
i
i
208pb 39.2~ 204pb 38.4
204pb
|
!
i
!
~ -~
- - ~ e~ s
J
2o7pb ~~'l
believed to have played a role in the formation of this pluton (Eby 1979). On the basis of trace-element and isotopic data the granites and some of the syenites are inferred to have a crustal origin (Eby 1985a). These rocks tend to be lower in total alkalis, relative to silica, and in the case of the granites they have a subalkaline character (Fig. 3). In many of the granite samples K 2 0 exceeds Na20. For these rocks, REE patterns show pronounced negative Eu anomalies indicating substantial feldspar fractionation and/or a feldspar-rich source material. The Sr and Pb isotopic data are shown in Fig. 4 and detailed information can be found in Eby (1985c). In most instances the strongly undersaturated magmas were intruded into older rocks belonging to the slightly undersaturated series and thus were isolated from crustal rocks. Rocks of the strongly undersaturated series have low initial Sr isotope ratios and in terms of Pb and Sr isotopes fall along the Atlantic mid-ocean ridge basalt (MORB) trend. The average Pb and Sr isotope ratios for the Oka carbonatite complex (Griinenfelder et al. 1985) and the isotopicallyprimitive fine-grained monzonites from Mount
~
:
~ ~6\
~
'
'
,rxsl-~ , 7 ~ i ~ \ ~ ~ \ \ ( ~ s ~ - L
Lx~-~-.... ~
~r~5~ gl~-
,
~
/--/0 t ~ /
--
'
'
....
'
~e~,
- ~"
:,---,_ c r~s~ a~,_ ~.~- ~,
19.6
206pb le.e
2o4pb
I 18"0().7050
I 0.7038
I
I 0.7046
I
I 0.7054
I
87Sr/86Sr FIG. 4. Age-corrected Pb isotopic ratios versus age-corrected 87Sr/86Sr ratios: SUS, strongly undersaturated series; CS, undersaturated to saturated series; MM, Merrymeeting Lake diorites. Symbols represent average values for isotopic ratios: O, Oka carbonatite complex (Griinenfelder et al. 1985); I , Mount Pawtuckaway finegrained monzonite; I--],Mount Pawtuckaway coarse-grained monzonite; A, Mount Shefford nordmarkite; 0, Merrymeeting Lake granite; 0, Mount Megantic granite. The MORB trend ( ) is from Cohen & O'Nions (1982), and the Pb and Sr isotopic data are from Eby (1985c).
The Monteregian Hills Pawtuckaway also plot with the strongly undersaturated series. The isotopic ratios for the slightly undersaturated series fall in a broader area but do overlap with the ratios of the strongly undersaturated series. These observations support the inference that the magmas which gave rise to the strongly undersaturated and slightly undersaturated series and the Oka carbonatite complex were drawn from a depleted mantle similar to that which gives rise to ocean island basalts and possibly some MORBs. Foland et al. (1983) report Sr and Nd isotopic data for the Ascutney Mountain complex of eastern Vermont which also indicate a depleted mantle source region. On the basis of trace-element data it has been inferred that the Mount Megantic and Merrymeeting Lake granites were derived from crustal material, most probably of Lower Palaeozoic age. The Mount Shefford nordmarkite was inferred to have been derived by partial melting of granulitic material, most probably of Grenville (Precambrian) age (Eby 1985a). The generally lower Pb isotope ratios for the Mount Shefford nordmarkite (Fig. 4) are consistent with its derivation from Precambrian material. These felsic units have been used to delimit a broad band (Fig. 4) of isotopic ratios which may be appropriate for crustal material in this region. If the initial mantle-derived magmas had Pb and Sr ratios similar to those of the strongly undersaturated series, then it would be possible to interpret the broad range in ratios of the slightly undersaturated series as representing mixtures of mantlederived magmas and crustal material. The slightly
441
undersaturated magmas were emplaced first and a number of the plutons are surrounded by rheomorphic breccias indicating substantial heat transfer to the country rocks and the possible ingestion of significant amounts of crustal material. The isotopic ratios for the coarse-grained monzonite at Mount Pawtuckaway and the diorite at Merrymeeting Lake can also be interpreted as evidence of crustal contamination of mantle-derived melts.
Petrogenesis Most of the mafic rocks in the Monteregian Hills and White Mountain provinces are at least of partly cumulate origin, so that it is not usually possible to define liquid lines of descent. However, if a rock can be found which represents the composition of the magma at some stage in its evolution, it may be possible to calculate the initial composition of the magma. If only ferromagnesian silicates are fractionated from the magma, and if the chemistry of these minerals and the relative proportions in which they are removed is known, they can be added to the liquid composition until an Mg/Fe ratio appropriate for a liquid in equilibrium with mantle olivine (Fo90) is obtained. This type of calculation has been carried out to determine the initial composition of the magmas which formed the slightly undersaturated to saturated series at Mounts Royal, St. Bruno and Brome and the undersaturated series at Mount Royal (Eby 1984c, 1985a), and the results are given in Table 1. The samples selected for the calculations have low
TABLE 1. Compositions o f initial melts (wt.~)
SiO 2 TiO2 A1203 Fe203 FeO MnO MgO CaO Na20 K20 P205 mg number Ne Ol
Slightly undersaturated
Undersaturated
Royal 1 St. Bruno 1 Brome 2
Royal 1
Aln6ite 3
43.95 2.21 10.45 1.84 11.51 0.11 17.89 8.74 2.29 0.70 0.30 73.5 5.9 36.2
43.96 2.72 11.67 3.74 8.33 0.19 12.96 10.97 3.31 1.26 0.91 73.5 10.1 19.5
33.58 2.27 6.93 1.98 10.10 -20.08 12.89 0.18 0.77 1.16 78.0 0.9 47.3
44.90 2.18 8.11 1.57 12.72 0.11 19.65 7.96 1.75 0.75 0.27 73.4 1.0 41.5
47.98 1.97 14.63 1.79 7.85 0.15 12.42 7.61 3.55 1.42 0.64 73.8 2.8 25.7
Dykes Monchiquite 4 Camptonite s 45.13 2.04 11.15 4.88 7.12 0.22 13.16 11.04 3.62 1.07 0.56 76.7 10.2 16.5
46.96 2.63 10.51 3.24 7.96 0.16 12.57 11.77 2.53 1.04 0.63 73.8 2.6 17.1
1, From Eby 1984c; 2, from Eby 1985a; 3, xenolith-free matrix composition from Marchand (1970, p. 17), (Fe z+ =0.85 Fea-); 4, sample P8802 from Hodgson (1968, p. 155); 5, sample P8776 from Hodgson (1968, p. 153)+ 0.11Fo84.
G. N. Eby
442
initial Sr ratios and have apparently not been contaminated by crustal material. Also included in Table 1 are the measured compositions of xenolith-free matrix material from the Ile Bizard aln6ite and a monchiquite dyke from the Montreal area, both of which have M g / ( M g + F e 2+) ratios appropriate for liquids in equilibrium with mantle olivine (Fo92 and Fo91 respectively). The initial composition of a camptonite liquid has also been calculated (Table 1) by adding olivine, which is a phenocryst phase, until an appropriate Mg/(Mg + Fe z +) ratio is obtained. The original liquids which gave rise to the slightly undersaturated to saturated series, particularly in the case of Mounts Royal and St. Bruno, are quite similar to alkali picrites. The initial magma at Mount Brome is less magnesian and has a lower normative olivine content. It should also be noted that the Mount Brome m a g m a has a significantly higher P205 and K 2 0 content (both P and K are incompatible elements during the melting of anhydrous mantle rocks), which implies a smaller degree of partial melting. All these initial melts are slightly nepheline normative. The camptonite is slightly nepheline normative and is chemically similar to an alkali olivine basalt. Such a magma could give rise to the mafic cumulate sequences found at Mounts Megantic and Pawtuckaway and to the basalts at Ossipee (Carr 1980). The camptonites, therefore, may be
a reasonable analogue for the magmatic evolution of the mafic series in the White Mountain plutons. At our present level of understanding it seems reasonable to conclude that two distinct magma types were involved in the formation of the slightly undersaturated to saturated series; alkali picrite magmas in the case of the Monteregian Hills and alkali basalt magmas in the case of the White Mountains. The original liquid which gave rise to the strongly undersaturated series at Mount Royal is moderately nepheline normative. The initial melt resembles a basanite in composition. The calculated composition is also quite similar to the measured composition of the monchiquite, indicating a similar origin. The aln6ite is distinctly different, having a much lower silica content, K 2 0 > N a 2 0 and high P205, suggesting a small degree of melting, and a high normative olivine content. The aln6ites of the Oka area have similar chemistries and along with the carbonatites collectively constitute a chemical group very different from the rest of the rocks in the Monteregian Hills and White Mountains. The trace-element contents of the initial melts have been calculated (Table 2) by adding olivine and pyroxene in the amounts used to determine the major-element compositions of the initial melts. Stepwise Rayleigh fractionation has been assumed and the partition coefficients are taken
TABLE 2. Trace-element abundancesfor initial melts (ppm)* Slightly undersaturated
Sc Co Rb Sr Ba La Ce Nd Sm Eu Tb Yb Lu Zr Hf Ta Th Rb/Ba • La/Yb Zr/Hf
Royal
St. Bruno
25 61 16 700 280 35 72 29 6.2 2.1 0.70 1.3 0.21 180 4.1 2.9 3.3 5.7 16.3 44
29 87 19 800 240 30 65 31 6.4 2.1 0.86 1.7 0.27 200 5.3 2.4 3.2 7.9 10.7 38
Brome 13.7 23 36 1130 450 48 117 55 10.8 3.5 1.3 2.1 0.31 360 9.7 8.3 7.0 8.0 13.8 37
Undersaturated Royal 13.5 25 26 1120 540 65 132 67 13.1 3.9 1.4 2.2 0.35 300 7.0 3.6 5.7 4.8 18.8 43
Dykes Aln6ite 27 64 84 -1020 91 187 78 14.7 4.0 1.3 1.5 0.22 400 11.9 9.4 7.8 8.2 37 34
Monchiquite 11.8 26 27 -93 200 84 15.0 4.4 1.7 2.9 0.41 290 8.0 8.3 7.7 -19.4 36
* Calculated from data in Eby (1984c, 1985a, 1985b); aln6ite abundances are measured values.
Camptonite 23 55 32 --40 83 33 7.4 2.4 1.1 1.8 0.28 200 5.8 4.2 5.0 -13.5 34
The Monteregian Hills from Frey et al. (1978) and Eby (1984c). The Rb/Ba • 102 ratios of 4.8-8.2 and the Zr/Hfratios of 34-44 are within the range of those found for primary basalts (Frey et al. 1978). Within the group of slightly undersaturated to saturated magmas (alkali picrite sub-group) incompatibleelement concentrations vary inversely with compatible-element concentrations, giving a set of relationships consistent with variable degrees of partial melting of a similar source region. The strongly undersaturated magma is enriched in incompatible elements and relatively depleted in compatible elements, suggesting a small degree of melting. The monchiquitic magma has similar trace-element concentrations to the strongly undersaturated magma, whilst the camptonitic magma (alkali olivine basalt) has lower incompatible-element abundances and a lower La/Yb ratio. The aln6itic magma has relatively high transition metal concentrations, suggesting either a relatively large degree of melting or derivation from mantle material from which a melt phase had previously been extracted. Incompatibleelement concentrations for the aln6ite are also quite high, suggesting either a small degree of melting and/or metasomatic enrichment of the
443
source region. This apparent contradiction can be resolved if the source of the aln6itic magma was a depleted mantle (high compatible-element abundances) which was subsequently metasomatized. Mysen & Kushiro (1977) suggested that alkali picrite magmas could be produced by moderate degrees of melting of a garnet lherzolite. Wyllie (1979) has suggested that a carbonated mantle could yield 'carbonatitic' liquids if melted at pressures of about 27 kb (in the garnet stability field). Isotopic data described previously indicate that these magmas were extracted from a depleted mantle (a mantle which had undergone an earlier Precambrian melting event). However, incompatible-element abundances are high, so that for any reasonable degree of melting a metasomatized mantle source would seem to be required. Such a model is illustrated in Fig. 5 (a) in which a metasomatized mantle similar to that inferred by Frey et al. (1978) is melted in the garnet lherzolite stability field. The melting equation for a garnet lherzolite deduced by Mysen & Kushiro (1977) is used in this illustration. It is seen that alkali picrites and aln6ites could be derived from a similar source by variable degrees of partial
""-.... I00
I00
50
~
50
o
....
_ 55 OJ:50 Opx:lO Cpx: 5 Sp 10
io
5
5
Lo Ce (a)
Nd
I I Sm Eu
I Tb
REE At. no.
I I Yb Lu
I La Ce (b)
I Nd
I I Sm Eu
I Tb
I L Yb Lu
REE At. No.
FIG. 5. Models for the generation of the various mantle-derived melts.(a) Generation of alkali picrites and aln6ites from metasomatized garnet lherzolite mantle : - - - , 6% partial melt; 9 , 2% partial melt. Fields represent range of compositions for the initial magmas. Data from this paper and Eby (1975). (b) Generation of basanites and alkali olivine basalts lu metasomatized spinel lherzolite mantle. Field represents various degrees of partial melting from 0.5% (top of field) to 4% (bottom of field). Compositions of initial magmas: SD5, monchiquite; RY33, basanite; PAB, camptonite; MP5, alkali olivine basalt; IB 1, aln6ite. Note that with increasing percentage of partial melt the REE patterns evolve from those of the strongly undersaturated magmas to those of the slightly undersaturated to saturated magmas. The aln6ite is plotted for comparison and has a much higher La/Yb ratio. Data from this paper and Eby (1985b).
444
G.N.
melting. Superimposed on this simple model, in the case of the CO2-rich aln6ites and the carbonatites, must be the effect of CO2 enrichment in the source region which can lead to substantial enrichment in the light REE (Wendlandt & Harrison 1979). Mysen & Kushiro (1977) suggested that very small degrees of partial melting of a spinel lherzolite would produce basanitic liquids while larger-percentage melts would be alkali olivine basalts. Such a model relating the strongly undersaturated (basanitic) magmas and the slightly undersaturated (alkali olivine basalt) magmas is illustrated in Fig. 5 (b). The melting equation for a spinel lherzolite deduced by Mysen (1979) and the same metasomatized mantle from Fig. 5 (a) are used in this illustration. The REE data do permit these two magma types to be related through variable degrees of partial melting of a metasomatized spinel lherzolite mantle. The melting models described above are relatively simplistic. In particular they do not consider the effect of the prior mantle-melting event on the amount of residual garnet nor the role of minor minerals introduced during the postulated metasomatic event. The trace-element composition of the metasomatized mantle is not well constrained, although it should be noted that the relative REE patterns inferred for a metasomatized mantle (Frey et al. 1978) are quite similar. The relatively limited amount of information on the chemistry of the initial melts, however, does not seem to warrant more complex models. At our present level of understanding, chemical and experimental data support the inference that the alkali picrite and aln6ite magmas were drawn from deeper depths (garnet lherzolite stability field) than the basanite and alkali olivine basalt magmas (spinel lherzolite stability field), and the last two magma types can be related by variable degrees of melting of a spinel lherzolite mantle.
Tectonic setting The mechanism(s) leading to melting in the Monteregian Hills-White Mountain provinces have been a subject of debate for some years. Among the possible mechanisms are leaky transform faults (Foland & Faul 1977), hot-spots (Crough 1981) and south-eastward propagating tensional events (Bedard 1985). At present, no single model seems to have overwhelming support. Mesozoic magmatism in this portion of North America is represented by tholeiitic basalts (both flows and dykes and sills) associated with the
Eby
Triassic basins and the subalkaline to alkaline magmatism of the Monteregian Hills and White Mountains. The older period of White Mountain magmatism (200-165 Ma) is time correlative with the tholeiitic basalt magmatism. The alkaline magmas were emplaced into a relatively restricted geographic area, and the mafic magmas (alkali olivine basalts) apparently originated in a spinel lherzolite mantle implying a relatively shallow depth of melting. The magmatic activity persisted for about 35 Ma, and there is no regularity to the ages which would suggest a hot-spot trace. It is therefore difficult to envisage this period of magmatism as being due to a mantle plume, and pressure-release melting accompanied by the influx of volatiles would seem to be the most appropriate mechanism. The younger period of alkaline magmatism (140-110 Ma ago) is much more complex in terms of age, distribution, magma types and geographic extent. In the case of the Monteregian Hills the igneous activity appears to be localized by preexisting fracture zones (Kumarapeli 1978), and a similar control is likely for the White Mountains. In the Monteregian Hills there is a well-defined periodicity to the igneous activity which is related to magma type, in that alkali picrite magmas precede basanite magmas. A similar periodicity has not been observed for the White Mountains, and the bulk of the plutons were emplaced during the younger pulse of Monteregian Hills activity (Fig. 2). The mafic magmas involved in the formation of the White Mountain plutons are apparently derived from alkali olivine basalts. The continental magmatic activity has an oceanic extension represented by the New England Seamounts chain, alkali olivine basalts in the Orpheus graben off Nova Scotia, and alkali basalts and trachytes of the Newfoundland Seamounts. The New England Seamounts chain shows a regular decrease in age eastward from 103 to 82 Ma (Duncan 1982). The alkali olivine basalts of the Orpheus graben have an age of about 116 Ma (Jansa & Pe-Piper 1985), whilst the eastward trachytes of the Newfoundland Seamounts have an age of 98 Ma (Keen et al. 1977). In most, but not all, cases the volcanism appears to be related to pre-existing fracture zones. In most cases the location of the alkaline magmatism was controlled by pre-existing zones of crustal weakness, and there is an apparent trend towards younger ages as the igneous activity moved from the continental to oceanic setting. Isotopic data suggest that the magmas emplaced into the continental crust were extracted from the sub-continental lithosphere, whilst the isotopic signatures of the magmas emplaced into the
The Monteregian Hills oceanic crust indicate a different mantle source (Eby 1985c). A possible scenario for this period of alkaline magmatism involves the translation of the North American plate across a rising mantle diapir. The diapir impinges into the base of the thick continental lithosphere and provides fluids and heat energy which migrate upwards causing metasomatism and melting. The impingement of the diapir causes some stretching of the overlying crust and the reactivation of preexisting zones of weakness which serve as channelways for the fluids and magmas. In the case of the Monteregian Hills, at least, the depth and percentage of melting decrease with time which is consistent with this scenario. As the plate moves westward the diapir penetrates farther into the continental lithosphere, resulting in melting at shallower depths and greater percentages of melting (White Mountain province). The translation of thin oceanic lithosphere over the diapir results in melting at even shallower depths and the direct involvement of material from the diapir in the formation of the magmas (as suggested by the isotopic data). Pre-existing zones of structural weakness serve to localize the magmatism in the oceanic setting. The mantle
445
diapir is envisaged to cover a broad area, unlike the more restricted geographic extent usually implied by the term mantle plume. While this model for the younger period of alkaline magmatism is still rather speculative, it does fit the currently available data.
ACKNOWLEDGMENTS: Over the years a number of people have assisted with various aspects of the Monteregian Hills-White Mountain project. D. Gold, A. R. Philpotts and G. Pouliot introduced the author to the fascinating problems of the Monteregian Hills. J. Creasy introduced the author to the White Mountain batholith, has been a stimulating collaborator and whetted the author's appetite for White Mountain problems. S. Moorbath, Oxford, provided laboratory facilities for isotopic work and N. Charnley, Oxford, has graciously provided microprobe time. The University of Lowell Nuclear Center has lurched from crisis to crisis during the past 9 years, but has somehow kept things going so that INAA trace-element analysis has continued. Lastly, J. Bruce, a former Dean of the College of Science, provided financial support at a crucial point which permitted the establishment of a major-element rockanalysis laboratory. To all I extend my thanks. The National Science Foundation provided the funds to establish a fission-track dating laboratory.
References BEDARD, J. H. 1985. The opening of the Atlantic, the Mesozoic New England igneous province, and mechanisms of continental breakup. Tectonophysics, 113, 209-32. CARR,R. S. 1980. Geology and petrology of the Ossipee ring-complex, Carroll County, New Hampshire. M.A. Thesis, Dartmouth College, Hanover, NH (unpublished). COHEN, R. S. & O'NIONS, R. K. 1982. The lead, neodymium and strontium isotopic structure of ocean ridge basalts. J. Petrol. 23, 299-324. CREASY,J. W. 1974. Mineralogy and petrology of the White Mountain batholith, Franconia and Crawford Notch Quadrangles, New Hampshire. PhD Thesis, Harvard University, Cambridge, MA (unpublished). - - & EBY,G. N. 1983. The White Mountain batholith as a model of Mesozoic felsic magmatism in New England. Geol. Soc. Am. Abstr. Progm. 15, 549. CROUGH, S. T. 1981. Mesozoic hotspot epeirogeny in eastern North America. Geology 9, 2-6. CURRIE, K. L. 1983. An interim report on the geology and petrology of the Mount Saint Hilaire pluton, Quebec. Curr. Res. geol. Surv. Can. 83 (1B), 39-46. - - , EBY, G. N. & GITTINS, J. 1986. The petrology of the Mount Saint Hilaire pluton, southern Quebec: an alkaline gabbro-peralkaline syenite association. Lithos, 19, 65-81. CZAMANSKE,G. K., WONES, D. R. & EICHELBERGER,J. C. 1977. Mineralogy and petrology of the intrusive
-
-
-
-
-
complex of the Pliny Range, New Hampshire. Am. J. Sci. 277, 1073-123. DUNCAN, R. A. 1982. The New England Seamounts and the absolute motion of North America since mid-Cretaceous time. Eos 63, 1103-4. EBv, G. N. 1975. Abundance and distribution of the rare-earth elements and yttrium in the rocks and minerals of the Oka carbonatite complex, Quebec. Geochim. cosmochim. Acta, 39, 597-620. -1979. Mount Johnson, Quebec--An example of silicate liquid immiscibility? Geology, 7, 491-4. 1984a. Geochronology of the Monteregian Hills alkaline igneous province, Quebec. Geology, 12, 468-70. 1984b. Mount Pawtuckaway ring-dike complex. In: HANSON, L. (ed) Geology of the Coastal Lowlands, Boston, MA to Kennebunk, ME, New England Intercollegiate Geological Conf. Salem, pp. 240-8. Salem State College, Salem. 1984c. Monteregian Hills I. Petrography, major and trace element geochemistry, and strontium isotopic chemistry of the western intrusions: Mounts Royal, St. Bruno, and Johnson. J. Petrol. 25, 421-52. 1985a. Monteregian Hills II. Petrography, major and trace element geochemistry, and strontium isotopic chemistry of the eastern intrusions: Mounts Shefford, Brome, and Megantic. J. Petrol. 26, 418-48. 1985b. Age relations, chemistry, and petrogenesis -
-
-
-
-
446
-
G. N. Eby
of mafic alkaline dikes from the Monteregian Hills and younger White Mountain igneous provinces. Can. J. Earth Sci. 22, 1103-1111. 1985c. Sr and Pb isotopes, U and Th chemistry of the alkaline Monteregian and White Mountain provinces, eastern North America. Geochim. cosmochim. Acta, 49, 1143-54. & CREASY,J. W. 1983. Strontium and lead isotope geology of the Jurassic White Mountain batholith, New Hampshire. Geol. Soc. Am. Abstr. Progm. 15, 188. EGGLESTON,J. W. 1918. Eruptive rocks at Cuttingsville, Vermont. Am. J. Sci. 45, 377-410. FOLAND, K. A. & FAUL, F. 1977. Ages of the White Mountain intrusives--New Hampshire, Vermont, and Maine, U.S.A. Am. J. Sci. 277, 888-904. -& FRIEDMAN, I. 1977. Application of Sr and O isotope relations to the petrogenesis of the alkaline rocks of the Red Hill complex, New Hampshire, USA. Contrib. Mineral. Petrol. 65, 213-25. , RACZEK, I. & HOFFMANN, A. W. 1983. Nd isotopic evidence for crustal contamination during evolution of the mafic rock suite at the Ascutney Mountain complex, Vermont. Geol. Soc. Am. Abstr. Progm. 15, 575. FREY, F. A., GREEN, D. H. & ROY, S. D. 1978. Integrated models of basalt petrogenesis: a study of quartz tholeiite to olivine melilitites from south eastern Australia utilizing geochemical and experimental petrological data. J. Petrol. 19, 463-513. GANDHI, S. S. 1966. Igneous petrology at Mount Yamaska. PhD Thesis McGill University, Montreal, Quebec (unpublished). GOLD, D. P. 1963. The relationship between the limestones and the alkaline igneous rocks of Oka and St. Hilaire, Quebec. PhD Thesis, McGill University, Montreal, Quebec (unpublished). 1972. The Monteregian Hills: Ultra-alkaline rocks and the Oka carbonatite complex. Int. Geol. Congr., 24th Session, Excursion B-11, Ottawa. GRISCOM, A. 1962. Aeromagnetic evidence for the structure of stocks and ring dikes in New Hampshire and Vermont. Geol. Soc. Am. Spec. Pap. 68, 186. GRUNENFELDER, M. H., TILTON, G. R., BELL, K. & BLENKINSOP, J. 1985. Lead and strontium isotope relationships in the Oka carbonatite complex, Quebec. Geochim. cosmochim. Acta, 50, 461-68. HODGSON, C. J. 1968. Monteregian dike rocks. PhD Thesis, McGill University, Montreal, Quebec (unpublished). JACOaSON, C. A. C. 1981. Major and rare-earth geochemistry of Mount Rougemont, Quebec. M.A. Thesis, Boston University, Boston, MA (unpublished). JANSA, L. F. & PE-PtPER, G. 1985. Early Cretaceous volcanism on the northeastern American margin and implications for plate tectonics. Geol. Soc. Am. Bull. 96, 83-91. JOYNER, W. B. 1963. Gravity in north-central New England. Geol. Soc. Am. Bull. 74, 831-58. KEEN, C. E., HALL, B. R. & SULLIVAN, K. D. 1977. Mesozoic evolution of the Newfoundland Basin. Earth planet. Sci. Lett. 37, 307-20. -
-
-
-
KINGSLEY, L. 1931. Cauldron-subsidence of the Ossipee Mountains. Am. J. Sci. 22, 139-68. KUMARAPELI, P. S. 1978. The St. Lawrence Paleorift system: a comparative study. In: RAMBERG, I. B. & NEUMANN, E.-R. (eds) Tectonics and Geophysics of Continental Rifts, pp. 367-84, Reidel, Dordrecht. LAURENT, R. & PIERSON, T. C. 1973. Petrology of alkaline rocks from Cuttingsville and the Shelburne Peninsula, Vermont. Can. J. Earth Sci. 10, 1244-56. LOISELLE, M. C. 1978. Geochemistry and petrogenesis of the Belknap Mountains complex and Pliny Range, White Mountain series, New Hampshire. PhD Thesis, Massachusetts Institute of Technology, Cambridge, MA (unpublished). MCHONE, J. G. 1978. Distribution, orientations, and ages of mafic dikes in central New England. Geol. Soc. Am. Bull. 89, 1645-55. 1984. Mesozoic igneous rocks of northern New England and adjacent Quebec: Summary, description of map, and bibliography of data sources. Geol. Soc. Am. Map Chart Ser. MC-49. & BUTLER,J. R. 1984. Mesozoic igneous provinces of New England and the opening of the North Atlantic Ocean. Geol. Soc. Am. Bull. 95, 757-65. -& CORNEILLE, E. S. 1980. Alkalic dikes of the Lake Champlain Valley. Vermont Geol. 1, 16-21. MARCHAND, M. 1970. Ultramafic nodules from Ile Bizard, Quebec. M.Sc. Thesis, McGill University, Montreal, Quebec (unpublished). MIYASHIRO, A. 1978. Nature of alkalic volcanic rock series. Contrib. Mineral. Petrol. 66, 91-104. MODELL, D. 1936. Ring-dike complex of the Belknap Mountains, New Hampshire. Geol. Soc. Am. Bull. 47, 1885-932. MYSEN, B. O. 1979. Trace-element partitioning between garnet peridotite minerals and water-rich vapor: experimental data from 5 to 30 kbar. Am. Mineral. 64, 274-87. -& KUSHIRO, I. 1977. Compositional variations of coexisting phases with degree of melting of peridotite in the upper mantle. Am. Mineral. 62, 843-56. PAJARI, G. E. 1967. The mineralogy and petrochemistry of Mount Johnson. PhD Thesis, University of Cambridge, Cambridge (unpublished). PHILPOTTS,A. R. 1969. Geology of Mount Rougemont. In: POULIOT, G. (ed.) Guidebook for the Geology of Monteregian Hills, pp. 77-84. Mineralogical Association of Canada, Montreal. -1976. Petrography of Mounts Saint-Bruno and Rougemont. Rep. ES-16, Quebec Minist6re des Richesses Naturelles. PIERSON, T. C. 1970. Petrogenesis and emplacement of the alkaline intrusion at Cuttingsville, Vermont: a reevaluation. BSc. Thesis, Middlebury College, Middlebury, VT (unpublished). QUINN, A. & STEWART, G. W. 1941. Igneous rocks of the Merrymeeting Lake area of New Hampshire. Am. Mineral. 11,633-45. RANDALL, K. A., FOLAND, K. A., RACZEK, I. & HENDERSON, C. M. B. 1983. Evolution of the Pliny complex, northern New Hampshire: Isotopic and -
The Monteregian Hills chemical relationships. Geol. Soc. Am. Abstr. Progm. 15, 666. REID, A. M. 1976. Geology of Mount Megantic. Rep. ES-25, Quebec Minist+re des Richesses Naturelles. RoY, C. J. ~,~FREEDMAN, J. 1944. The petrology of the Pawtuckaway Mountains, New Hampshire. Geol. Soc. Am. Bull. $5, 905-20. SHARP, J. A. & SIMMONS,G. 1978. Geologic/geophysical models ofintrusives of the White Mountain magma series (WMMS). Geol. Soc. Am. Abstr. Progm. 10, 85. SIZE, W. B. 1972. Petrology of the Red Hill syenitic complex, New Hampshire. Geol. Soc. Am. Bull. 83, 3747-60. STEIGER, R. H. & JAEGER, E. 1977. Subcommission on geochronology: Convention on the use of decay constants in geo- and cosmochronology. Earth planet. Sci. Lett. 36, 359-62. TAYLOR, S. R. & TOKS6Z, M. N. 1982. Crust and uppermantle velocity structure in the Appalachian
447
orogenic belt: Implications for tectonic evolution. Geol. Soc. Am. Bull. 93, 315-29. VALIQUETTE, G. & POULIOT, G. 1977. Geology of Mounts Brome and Shefford. Rep. ES-28, Quebec Minist~re des Richesses Naturelles. WENDLANDT, R. F. & HARRISON, W. J. 1979. Rare earth partitioning between immiscible carbonate and silicate liquids and CO2 vapor: results and implications for the formation of light rare earthenriched rocks. Contrib. Mineral. Petrol. 69, 40919. WESTON GEOPHYSICALRESEARCHINC. 1976. Geological Investigations for the Pilgrim H Station, BE-SG 7701, Weston Geophysical Research Inc., Weston, MA. WOUSSEN, G. 1969. Notes on the geology of Mount Royal. In: POULIOT, G. (ed.) Guidebook for the Geology of Monteregian Hills, pp. 63-74. Mineralogical Association of Canada, Montreal. WYLLIE, P. J. 1979. Magmas and volatile components. Am. Mineral. 64, 469-500.
G. NELSON EBY, Department of Earth Sciences, University of Lowell, Lowell, MA 01854, U.S.A.
Mid-Proterozoic alkaline magmatism in southern Greenland" the Gardar province B. G. J. Upton & C. H. Emeleus S U M M A R Y : The Gardar province (c. 1320-1120 Ma old) relates to repetitive rifting within the Mid-Proterozoic continent and the attendant large-scale generation of alkali and transitional olivine basaltic and hawaiitic magmas. Lavas from the earliest episodes are preserved in down-faulted units (the Eriksfjord formation), whereas younger basic magmatism is recorded only as intrusions, mainly dykes. The province shows a compositional continuum from basaltic to salic alkaline compositions. The latter are represented among the early lavas, as dykes and, most prominently, as central-type complexes commonly involving layered (cumulate) syenitic intrusions. Most of the central complexes display either a silica-oversaturated trend (augite syenite-quartz syenite-(A-type) alkali granite) or a silica-undersaturated trend (augite syenite-pulaskite-foyaite). Strongly peralkaline products are present in each association with agpaitic rocks occurring as differentiates at Ilimaussaq and Motzfeldt. The oversaturated and undersaturated complexes, irrespective of age, define southerly and northerly zones respectively in the province. 'Giant-dyke complexes', produced in late-Gardar rifting events (contemporaneous with Keweenawan activity in N America) are important in providing a link between simple dilatational dykes and the central-type complexes. An ultramafic lamprophyre--carbonatite association plays a subordinate role in the province, and carbonatites are also associated with some of the salic complexes. The carbonatites at Gmnnedal-Ika are thought to be residual from CO2-rich phonolitic magmas, whereas carbonate (siderite)-fluoride rocks at Ivigtut crystallized from volatile-rich residues associated with an alkali granite. The salic alkaline magmas were produced from alkali or transitional olivine basaltic parent magmas, with crystal fractionation being the dominant mechanism. Generally low 87Sr/86Sr(i) ratios in the alkaline complexes suggest little assimilation of old radiogenic crustal materials. The basaltic magmas were characterized by high AI203/CaO and FeO/MgO ratios, and it is inferred that they themselves were residues from more magnesian parent magmas that had undergone extended olivine and clinopyroxene fractionation at sub-crustal levels. The primary magmas may have been derived from garnet lherzolite source rocks with high garnet to clinopyroxene ratios. Late-Gardar basic magmas of the Tugtut6q-IlimaussaqNunataq zone are geochemically distinct from all other Gardar basic magmas in being enriched in P, Ba, Nb and light rare earth elements. This feature may be due to metasomatic enrichment of sotirce rocks along a lineament sub-parallel to the Archaean craton margin. Gardar alkaline magmas were rich in F, C1and C compounds (oxides and/or hydrocarbons) while being relatively anhydrous. Many crystallized under low fO2 conditions. The fundamental cause of the repetitive alkaline magma generation may have been the ascent of F-, C1- and C-rich fluids from the deep mantle, leading to partial melting of the lithospheric mantle close to the boundary between Archaean and Proterozoic crustal units.
Introduction Intra-plate alkaline magmatism in southern Greenland in the period 1320-1120 Ma gave rise to the Gardar igneous province. Since general reviews of the province have already been published (e.g. Upton 1974; Emeleus & Upton 1976), an attempt is made in this paper to highlight new data and concepts that have accrued during the past decade. The Gardar province forms an integral part of the great tract traversing the supercontinent which for much of Proterozoic time was composed of North America, Greenland and Europe (Piper 1982), whose tectono-magmatic evolution
involved the emplacement of massif-type anorthosites (with associated adamellites, quartz mangerites, ferrodiorites etc.) and Neohelikian (post 1400 Ma) intra-contintental rifting (Emslie 1977). Large-scale extrusion of basalt (flood basalts) attended the younger rifting stages and is exemplified by the Keweenawan, Seal Lake and Coppermine Bay lavas. These were typically tholeiitic, although relatively enriched in largeion lithophile (LIL) elements and sometimes verging towards alkalic compositions (Baragar 1977). The magmas parental to the North American anorthosites were themselves tholeiitic in character and may have evolved from olivine tholeiites (Emslie 1977).
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks, Geological Society Special Publication No. 30, pp. 449-471.
449
450
B. G. J. Upton & C. H. Emeleus
In the southern Greenland area of the Proterozoic supercontinent it is suggested that successive melting events in the mantle for around 200 Ma following c. 1320 Ma were generally less extensive and produced more alkalic magmas. Rather than ferrodiorites forming as a result of closed-system crystal fractionation of the high-A1 magmas parental to the anorthosites, ferrosyenites developed in southern Greenland from the crystallization of benmoreitic residual magmas. Of the geochemical features related to the enhanced alkalinity of these Gardar magmas there is evidence suggesting that volatile contents (F, C1 and CO2) were notably high. This may have contributed significantly to lowering viscosities, facilitating convection, crystal fractionation and liquid-state differentiation, as well as promoting late-stage processes of mineralization, pegmatite formation and wall-rock metasomatism. Whereas the great majority of Gardar igneous rocks are relatively siliceous (more than 45% SiO2) , their genesis was associated with small-volume, but persistent, production of subsilicic (feldspar-free) magmas ranging from lamprophyres to carbonatites.
stones and (ii) intrusives emplaced in the southern part of the Archaean craton and the early Proterozoic rocks to the S of the craton. The Gardar rocks, which lie N of the Grenville Front, are unaffected by any younger orogenic episodes and are consequently generally undeformed and unmetamorphosed save for those effects due to thermal metamorphism, metasomatism and hydrothermal activity associated with the Gardar magmatism itself. The intrusions include alkaline and peralkaline central complexes (Fig. 1). Salic intrusions constitute 87% of the total exposed outcrop of Gardar 'plutonic' intrusions, the remaining 13% being gabbroic (Watt 1966). However, there is abundant evidence for the widespread eruption of basaltic and hawaiitic magmas, principally through dykes, whose areal distribution was underestimated in Watt's planimetric analysis. Consequently the 'plutonic' rocks almost certainly give a misleading impression of the actual volumetric relationship among the Gardar magmas. About 10 central complexes (the precise number depending upon definition !) occur in a belt some 70 km by 200 km (Fig. 1). These vary in size from a few hundred metres across (e.g. Ivigtut) to about 45 km in diameter (Nunarssuit). They are typically steep-sided bodies, commonly composed of several separate intrusions. The
Central complexes and giant dykes Gardar igneous rocks occur as (i) relict lava successions interbedded with continental sand-
~
CC.cvc.=,:~t~,_,"~.t ~; GREE
NLAND
. . .~.!,. . ,~,Jj"
/
g - - N u n 02"., atc]q ,
~,'
Far
Klokken
Isortoq
~.\~176 \
Margin of gJaoers Undersaturated J'~'~ Over sf turated E~
oo
17
~
i..
t Oardar Maior )ntrus~ons
Enksfiord Formohon Iovas
0.
.I0
. 20
. 30
40
50 K m
Q
FIG. 1. Sketch map of the Gardar igneous province showing the relationship between part of the Archaean craton and the early Proterozoic (Ketilidian) terrain to the S and the principal Gardar intrusive complexes.
M i d - P r o t e r o z o i c m a g m a t & m & southern G r e e n l a n d latter may be stock-like, emplaced by piecemeal stoping and/or by repeated foundering of slablike masses of roofing rocks (e.g. Kfingngtt and Gronnedal-Ika (Emeleus & Upton 1976)), or may have the form of ring dykes. Localization of these central-type complexes appears to have been controlled in several cases (e.g. Gronnedal-Ika, Dyrnaes-Narssaq, Ilimaussaq, the Q6roq complexes and Motzfeldt) by intersection of dykeswarms with (approximately) E-W left-lateral faults that were active throughout the Gardar period (Henriksen 1960; Stephenson 1976a). Descriptions of the individual complexes have been given by Emeleus & Upton (1976) and Blaxland et al. (1978). Whereas typical Gardar dykes consist of transitional or mildly alkaline olivine basalt and hawaiite (and their coarser-grained equivalents), the central complexes principally comprise salic alkaline rocks. These grade from alkali granite via quartz syenite, syenite and pulaskite to foyaite. The saturated syenites are commonly composed of cryptoperthitic to microperthitic alkali feldspar, fayalitic olivine, salitic pyroxene, amphibole (ferro-pargasite to hastingsite), biotite, apatite and Ti-magnetite. These augite syenites are common to most of the central complexes and are believed to have played the role of parental magmas. The central complexes are characterized either by a SiO2-oversaturated or a SiO2-undersaturated trend. However, the adjacent (and intersecting) Dyrnaes-Narssaq and Ilimaussaq complexes are remarkable in displaying an intrusive sequence showing the following alternation: oversaturated, undersaturated, oversaturated, undersaturated. In the province as a whole, however, undersaturated complexes predominate in the north and oversaturated complexes predominate in the southern half (Fig. 1). The larger intrusive units of the central complexes are commonly composed of welllayered cumulates. Since the review of Emeleus & Upton (1976), detailed studies have been made of the layering processes that were involved at Ilimaussaq (Larsen & Sorensen 1987) and Klokken (Parsons 1979, 1981). Klokken displays unusual layering consisting of alternating granular and laminated syenites in which the granular syenites are characterized by an inverted cryptic layering whereas the (more fractionated) laminated syenites lack cryptic variation. Strongly disturbed pseudo-sedimentary structures at the bases of the granular syenite layers in contact with laminated syenite indicate that the granular syenite layers successively spalled off from the roof zone of the intrusion and settled onto crystal mushes that subsequently formed the laminated syenites. Compositional considerations suggest
451
that the uppermost (least fractionated) granular syenite layer originated very close to the roof of the intrusion. Transmission electron microscope studies of the structure of cryptoperthitic feldspars from the top of the granular syenite sequence indicate much more {apid rates of cooling than those from deeper in the pluton (Brown et al. 1983). One of the interesting features of the province is the occurrence of giant dykes. These exceptionally broad dykes with widths of up to 800 m are restricted to the late-Gardar (less than 1200 Ma ago) dyke swarms. As with the majority of Gardar dykes, these were initiated by ascent of basaltic or hawaiitic magma. The tectonics of emplacement were complex but appear to have involved dilation accompanied in some cases by lateral shear and vertical displacement of one side-wall relative to the other. Slow cooling of the basic magmas produced coarse gabbroic and syenogabbroic cumulates. However, augite syenites and occasionally still more differentiated rocks locally occupy the axial regions of these giant dykes. In some cases these have been emplaced by later intrusion of salic magma from deep sources whereas in other instances they appear to have developed by in situ differentiation, largely resulting from side-wall crystallization. Evidence from giant dykes in the Isortoq and Tugtut6q areas and in the nunataqs N of Motzfeldt indicates that the low-density salic magmas accumulated at (or migrated to) the upper parts of the giantdyke complexes, and that continued ascent by stoping progressively superseded the earlier stages of emplacement in which the denser basic magmas occupied potential fissures produced by extensional tectonics (Bridgwater & Coe 1970; Upton & Thomas 1980; Becker 1984; Upton & Fitton 1985). In giant dykes of the Nunarssuit area (at Bangs Havn) and the younger giant dyke on Tugtut6q there is gradation from augite syenite through quartz syenite to alkali granite. Conversely, in the Tugtut6q older giant dyke and in a giant dyke north of Motzfeldt, augite syenites grade with increasing silica undersaturation into peralkaline nepheline syenites (foyaites). There is consequently a close similarity in the range of lithologies (as well as in the nature of cumulate layering) between the giant-dyke complexes and the central-type complexes. The principal differences between the two situations lie (i) in their geometric form and mode of emplacement and (ii) in the fact that basic rocks predominate in the giant-dyke complexes but have subordinate status within the central-type complexes. The giant-dyke complexes thus play a crucial role in linking simple dilational dykes (typically basic)
452
B. G. J. Upton & C. H. Emeleus
on the one hand to quasi-cylindrical central-type complexes (dominantly salic) on the other (Upton 1974). Numerous detailed reports on the mineralogy of the various intrusive complexes have appeared over the past 15 years. Because of analytical difficulties associated with exsolution phenomena the mineral chemistry of the alkali feldspars has been relatively neglected and the studies have largely concerned the ferromagnesian silicate and opaque oxides. Reports relating to the mineralogy of the intrusions include (a) Igdlerfigssalik (Powell 1978), (b) Ilimaussaq (Larsen 1976, 1977, 1979; Karup-Moller 1978), (c) Isortoq giant dykes (Becker 1984), (d) Klokken (Parsons 1979, 1981 ; Parsons & Butterfield 1981), (e) Kfingn~t (Stephenson & Upton 1982),(f) Motzfeldt (Jones & Peckett 1980; Jones 1984), (g) N Q6roq (Chambers 1976), (h) Nunarssuit (Parsons & Butterfield 1981), (i) S Q6roq (Stephenson 1972,1974, 1976b), (j) Tugtut6q older giant-dyke complex (Upton et al. 1985) and (k) Tugtut6q younger giant-dyke complex (Upton & Thomas 1980). The peralkaline nature of many of the salic intrusions is manifest in the occurrence of such minerals as aegirine, arfvedsonite-riebeckite~, aenigmatite, astrophyllite, eudialyte etc. Agpaitic rocks are developed at Motzfeldt (Jones 1980) and particularly at Ilimaussaq (Larsen & Sorensen 1987). Detailed petrological studies of individual complexes (e.g. Motzfeldt (Jones 1980), Kfingn~t (Stephenson & Upton 1982), Tugtut6q younger giant dyke (Upton & Thomas 1980) and the Isortoq giant dykes (Becker 1984)) indicate that the salic rocks are differentiates of the basic magmas through crystal fractionation processes. However, in view of the field evidence clearly demonstrating that stoping was important in the emplacement of the central complexes and the syenitic components of some giant dykes (Bridgwater & Coe 1970; Becker 1984), the compositions of many of the large salic intrusions must have been substantially modified by crustal assimilation. Whereas the initial 87Sr/86Sr ratios of most of the major Gardar intrusions fall in the range 0.702-0.704, some lie in the range 0.7040.707. Thus the biotite granite in the Nunarssuit complex, with an initial ratio (at 1137 Ma) of 0.7068+0.0013, can be presumed to contain a large proportion of older crustal components (Blaxland et al. 1978). Exceptionally high initial Sr isotopic ratios recorded from Ivigtut (0.7125 + 0.0048) and from the latest component of the Ilimaussaq complex (agpaitic syenites, 0.7096+ 0.0022) were attributed by Blaxland et al. (1978) to selective enrichment in radiogenic Sr rather than to bulk assimilation.
Stratified magma chambers Intrusive sequences at some of the complexes indicate that stratified magma bodies were not uncommonly developed at depth, with their uppermost parts composed of siliceous alkali-rich compositions and grading downwards into more basic compositions. The successions at S Q6roq (Stephenson 1976b) and Motzfeldt (Jones 1980), where the most evolved foyaitic components were emplaced early, with more primitive compositions (lower Fe/Mg and (Na + K)/Ca) emplaced later after relatively short cooling intervals, have been explained as being the result of sequential tapping of deeper levels in such stratified chambers. Kfingn~t (Stephenson & Upton 1982) provides an analogous example among the oversaturated complexes. Supplementary evidence for magma stratification prior to high-level emplacement in serial batches has been adduced from compositional changes across dykes (Bridgwater 1967), and variation from augite syenite to foyaite in the Tugtut6q older giant dyke has been attributed to the crystallization in place of a compositionally stratified magma body (Upton et al. 1985). Whereas continuity from basalt through to trachyte (and thence to either phonolite or comendite) can be demonstrated within fastchilled dyke-swarms (Martin 1985; Upton & Fitton 1985), evidence from some of the giantdyke and central-type complexes containing both syenitic and gabbroic components suggests rather abrupt changes from basic magmas (less than 50~ SiOz; more than 3% MgO) to salic magmas (more than 54~ SiO2; less than 2~ MgO). Consequently, stratification inferred for the 'parental' magma chambers may have involved at least one relatively sharply defined compositional break between overlying salic magmas and underlying basic magmas. The deep-lying basic levels were only tapped during ring-dyke formation during late-stage cauldron subsidence at Kfingn~t, Igdlerfigssalik and Motzfeldt.
Cycles of Gardar magmatism It has been proposed that Gardar activity involved three principal cycles of magmatism (early, mid and late), each of which (i) commenced with a large-scale uprise of olivine basalt or hawaiite magmas during phases of crustal attenuation and (ii) terminated with emplacement of central-type complexes at, or towards the end of, the extensional phases (Upton & Blundell 1978). Most of the central complexes were
Mid-Proterozoic magmatism in southern Greenland emplaced at the end of the first and the last of these three cycles (i.e. early and late Gardar respectively), while only the small complexes at Kfingn~t and Ivigtut marked the end of the second (mid-Gardar) cycle. The distributions of the larger intrusions and major dyke swarms (and, in the case of the early Gardar, the lavas) for these three cycles are shown in Figs 2, 3 and 4. There are clear exceptions, however, to this simple cyclical pattern and, in the late Gardar (less than 1200 Ma ago), some of the major (central-type) intrusions occur at an early stage. Whereas those at Nunarssuit, central Tugtut6q, Ilimaussaq and Klokken, together with the late Igdlerfigssalik syenites, are clearly among the latest manifestations ofGardar magmatism, postdating the great majority of dykes in their respective areas, this is not so for the S Q6roq and early-Igdlerfigssalik intrusions. The S Q6roq syenites, emplaced 1160_ 8 Ma* ago, could be contemporaneous with the Tugtut6q older giantdyke complex (1154+ 16 Ma) with which (after allowing for younger faulting) it is on-strike. Whereas the S Q6roq (and early Igdlerfigssalik) syenites could be regarded as culminating midGardar intrusions, they are considered here as among the oldest rocks of the late-Gardar magmatism.
453
The Dyrnaes-Narssaq complex is also anomalous in terms of a simple three-cycle hypothesis. It is younger than the Tugtut6q giant dykes (see below) but older than the Ilimaussaq complex and some of the late-Gardar alkaline dykes (as well as being affected, like S Q6roq, by left-lateral transcurrent faulting) and was emplaced in the middle stages of the late-Gardar 'cycle'. Collerson (1982) noted close parallels between Proterozoic alkaline activity in Central Labrador and that of southern Greenland. The Labrador occurrences (including peralkaline granites and agpaites) date between 1460 and 1260 Ma, suggesting a closer age relationship to the early Gardar than to the mid- or late-Gardar magmatic events.
Distribution and chronology of Gardar lavas and dyke swarms Early Gardar
Lavas compose about half the total thickness (c. 3000 m) of the preserved sequences of the Eriksfjord formation and represent the earliest
'5, '..
" "v" :;, 9 ,x
,/,
,
,
:7
('
c.
?
~ ; -J -'" '". .
,.. :
'" v'-..-
""\
......
-
,;, .-,'"-".
j ..s - : j
co
D~
FIG. 2. Distribution of early-Gardar igneous rocks" dotted areas, Eriksfjord formation lavas; black areas, Gronnedal-Ika, N Q6roq and Motzfeldt nepheline syenite complexes. * Ages quoted are from Blaxland et al. (1978), recalculated by Martin (1985) using constants recommended by Steiger & Jager (1977).
B. G. J. Upton & C. H. Emeleus
454
", ".- ;'-" - -G:
9 !:,",,\,,'
jJr / ,4.-.
-'-P I
,
~-l .-, ,'.__.
")
.
.
.
.
--2
?'
,-~"
:'--'
--
c."
t
k
0~
FIG. 3. Distribution of mid-Gardar igneous intrusions. The KfingnfitFjeld and Ivigtut complexes lie to the W (the Ivigtut locality is indicated with an arrow). WNW ESE-trending BD0 dykes traverse the region. Younger mid-Gardar dykes are mainly concentrated in the Ivigtut region. The Fox Bay swarm is indicated diagrammatically in the eastern part of the map. record of Gardar magmatism (Fig. 2). Most have basaltic to hawaiitic compositions. Three thick sub-aerial volcanic units occur" from oldest to youngest these are the Mussartfit, Ulukasik and Ilimaussaq volcanic members, and they are separated by sandstone units (Poulsen 1964). Well-preserved pahoehoe surfaces point to high eruption rates with little erosion occurring between one eruption and the next. Whereas the conduits through which these lavas were supplied have not been identified, the lavas are interpreted as low-viscosity products of volcanism in a subsiding basin bounded by roughly E-W boundary faults. The Eriksfjord formation remains poorly dated; it rests unconformably on granites dated at 1850-1600 Ma and is cut by the Motzfeldt and N Q6roq complexes dated at 1310_+31 Ma and 1291 _+61 Ma respectively (Blaxland et al. 1978). No unconformities are known within the formation and the youngest lavas of the Ilimaussaq member may have been erupted from the N Q6roq or Motzfeldt centres (Upton & Blundell 1978; Jones 1980). Consequently the whole formation may be no older than about 1320 Ma. Despite the dating uncertainty' it appears likely that the Eriksfjord lavas are distinctly older than North American and Greenlandic Neohelikian flood basalts such as the Keweenawan (11501100 Ma), Seal Lake (1280 Ma), Coppermine Bay
(1215 Ma) and Zig-Zag Dal (1230 Ma) sequences (Baragar 1977; Kalsbeek & Jepsen 1984) and must rank among the earliest-known continental flood basalts. The mean composition of the lavas (those with MgO in the range 4~-8~o) in each of the three members (Table 1, analyses 1-3) becomes more evolved upwards; whereas the Mussartfit and Ulukasik volcanic members consist only of basalts and hawaiites, the uppermost (Ilimaussaq) volcanic member also includes trachybasalts, trachyandesites, trachytes and phonolites (J. G. Larsen 1977; Jones 1980). Thus, eruptions through the Eriksfjord formation display a generalized tendency to become more evolved with time. Mid-Gardar
Following the emplacement of the large earlyGardar undersaturated alkaline complexes (Motzfeldt, N Q6roq and Gronnedal-Ika), swarms of basic dykes (individually up to 200 m wide) were intruded (Berthelsen & Henriksen 1975). The earliest of these were lamprophyric (spessartites, kersantites and aln6ite-monchiquites) but subsequent, and more abundant, dykes were dominantly doleritic or trachydoleritic. These younger 'brown dykes' (so-called on account of their weathering properties) have been
Mid-Proterozoic magmatism
455
in s o u t h e r n G r e e n l a n d
: ,,::,,y,,,",,,--,, ( ', , / " / ;
(-, .z - 2 -..
:q i
- ?" ~ .
"
I~"~ ........ -,,
) -,i , .....
.'
i
.,"..s.-.~
,;..?
,:"..
9
FIG. 4. Distribution of late-Gardar igneous intrusions. The principal WSW ENE swarms are indicated diagrammatically: the Nunarssuit Isort6q swarm lies to the N and W of the Tugtut6q-Illmaussaq-Nunataq swarm. sub-divided into four generations 0, 1, 2 and 3. The generation 0 brown dykes (BD0) trend W N W - E S E , while younger generations tended to be intruded with rotation of the trend direction anticlockwise towards a S W - N E orientation (Berthelsen & Henriksen 1975). The distribution of the larger mid-Gardar dykes is shown in Fig. 3. The main concentration of these dykes is on and around the boundary between the Archaean craton to the N and the early Proterozoic (Ketilidian) rocks which compose much of southern Greenland. Their emplacement age is bracketed by upper and lower limits of 1219 + 16 Ma (Kfingngtt) and 1299_+ 17 Ma (Gronnedal-Ika) (Patchett et al. 1978). The mid-Gardar activity can be regarded as a local manifestation of the large-scale magmatic event that affected much of the preGrenville cratonic region of Canada and Fennoscandia (Patchett et al. 1978). In the Gardar province its termination was marked by the intrusions of Ivigtut (1222_+ 25 Ma (Blaxland et al. 1978)) and Kfingngtt. An important ENEtrending swarm oftrachytic and phonolitic dykes, the Fox Bay swarm, occurs in the eastern part of the province in the vicinity of Igaliko (Fig. 3). The swarm, which is about 10km wide and traceable for 20 kin, includes individual dykes up to 20 m across. The swarm is cut by a small intrusion, the Ostfjordsdal syenite, which may be contemporaneous with the S Q6roq complex.
While the upper age limit is defined by the Ostfjordsdal syenite, the lower limit is not well constrained and the swarm can be regarded as of mid- or possibly early-Gardar age. Late Gardar
Resumption of activity saw migration of the principal magmatic foci southwards away from the craton. Dyke swarms developed along W S W E N E to S W - N E trends, sub-parallel to the craton margin, and were concentrated along two main zones. Those of the more northerly NunarssuitIsortoq zone (Fig. 4) are demonstrably younger than the mid-Gardar BD0 dykes (about 1250 Ma (Patchett et al. 1978)). Similarly, those of the more southerly Tugtut6q-Ilimaussaq zone (Fig. 4) are post-1250 Ma and earlier than the Ilimaussaq and central Tugtut6q complexes (at 1143 _+21 and 1124_+ 20 Ma respectively (Blaxland et al. 1978; Martin 1985)). They are more precisely defined, however, than those of the Nunarssuit-Isortoq zone since what may be the earliest dyke intrusion in the Tugtut6q Ilimaussaq zone (the Hviddal or Tugtut6q older giant dyke) is itself dated at 1154 -I- 16 Ma (Upton et al. 1985). It could be that the dyke-swarms in these two zones were broadly contemporaneous. However, in the absence of dates from the earliest NE and E N E dykes of the Nunarssuit-Isortoq zone (some of which could correlate with the mid-
B. G. J. Upton & C. H. Emeleus
456
TABLE 1. Average compositionsof basaltic rocks (4-8 wt. %MgO) from the Gardar province 1
2
3
4
5
6
7
8
9
10
11
SiO2 A1203 FezO3* MgO CaO NaEO K20 TiO, MnO P_,O5
45.87 16.15 13.42 5.96 7.38 2.82 1.42 2.61 0.18 0.66
46.26 16.36 13.46 5.64 7.77 3.59 0.93 2.52 0.16 0.49
47.66 15.08 13.77 5.18 7.80 4.11 1.19 2.40 0.14 0.43
48.90 15.17 14.13 5.41 8.41 3.30 1.04 2.62 0.19 0.49
46.65 15.92 14.74 6.38 5.85 3.22 1.15 2.40 0.19 0.39
47.00 17.03 13.69 6.14 8.60 3.26 1.06 2.09 0.18 0.42
43.47 15.65 15.57 4.76 7.70 3.45 1.81 4.40 0.19 1.95
46.00 16.71 14.91 5.93 7.78 3.55 1.45 2.63 0.19 0.86
46.06 17.13 14.21 6.12 7.96 3.43 1.41 2.51 0.18 0.83
45.71 t5.48 14.07 5.25 7.63 3.49 2.25 3.04 0.19 1.06
44.09 14.73 14.93 5.54 9.18 3.20 1.71 2.80 0.27 1.38
Total
96.47
97.18
97.62
99.66
99.62
99.47
98.95
100.01
99.84
98.17
97.83
Major elements (wt. ~ )
Trace elements (ppm) Ni Cr V Sc Cu Zn Sr Rb Zr Nb Ba La Ce Nd Y
ND ND 198 28 31 96 521 37 137 10 762 17 41 24 26
ND ND 160 22 38 106 516 31 102 8 387 9 23 16 25
ND ND 161 23 8 103 310 29 185 13 424 13 35 22 32
62 85 232 26 41 112 379 30 223 11 536 24 53 28 34
78 108 219 25 52 106 538 31 225 27 663 22 49 27 30
77 57 179 22 49 92 533 22 124 8 576 15 30 19 27
FeO*/(FeO* + MgO) (wt.) A1203/CaO K/Rb Ba/Sr Zr/Nb La/Y
0.67 2.19 319 1.46 13.7 0.65
0.68 2.11 248 0.75 12.8 0.36
0.71 1.93 341 1.37 14.2 0.41
0.70 1.80 287 1.41 20.3 0.71
0.68 1.86 306 1.23 8.33 0.73
0.67 1.98 400 1.08 15.5 0.56
22 5 138 16 40 81 1039 39 162 31 1669 47 103 57 35
52 30 160 17 39 91 901 23 150 22 1120 44 64 34 27
0.75 0.69 2.03 12.15 385 522 1.651 1.24 5.2 6.8 1.34 1.63
63 42 168 17 44 90 921 25 141 21 1052 27 63 32 26 0.68 2.15 468 1.14 6.7 1.04
47 72 181 21 38 109 854 79 192 40 1251 40 87 45 31 0.71 2.03 237 1.46 4.8 1.29
52 40 170 19 59 133 1099 44 178 59 1284 118 253 110 43 0.71 1.60 323 1.17 3.0 2.74
1, Mussartfit volcanic member, Eriksfjord formation (20); 2, Ulukasik volcanic member, Eriksfjord formation (9); 3, Ilimaussaq volcanic member lavas (6); 4, early W N W - E S E (BD0) dykes (12); 5, N E and E N E dykes of the Ivigtut region (24); 6, NE and E N E dykes of the Nunarsuit-Isort6q region (23); 7, older giant dyke, Tugtut6q (chilled facies) (3); 8, younger giant dyke, Tugtut6q (chilled facies) (9); 9, giant dykes of the Nunataq region N and NE of Motzfeldt (9); 10, other basic E N E dykes, Tugtut6q region (46); 11, other basic ENE dykes, Nunataq region (27). (Numbers in parentheses are the numbers of analysed samples).
* total iron as Fe203 or FeO.
ND, not determined.
Gardar NE and ENE dykes of the Ivigtut area), this dyke swarm may have commenced substantially before that of the Tugtut6q-Ilimaussaq zone. Despite this reservation, the dykes, and associated major intrusions, of these two zones will be collectively treated as late Gardar. The late-Gardar dykes form a broad swathe, some 70 km across, within which the two zones are separated by a tract in which Gardar dykes are notably scarce. The restriction of giant dykes to the late phases of Gardar activity, and their occurrence within both these zones, may denote response of the lithosphere to exceptionally high
strain rates. If so, these may reflect particularly vigorous mantle diapirism. Despite their common possession of giantdyke components there are significant differences between the two zones. 1 Despite the presence of salic differentiates associated with the giant dykes of the NunarssuitIsortoq zone (Harry & Pulvertaft 1963; Bridgwater & Coe 1970; Becker 1984), dykes more evolved than hawaiite are scarce in the Nunarssuit-Isortoq zone but extremely abundant in the Tugtut6q-Illmaussaq zone.
457
Mid-Proterozoic magmat&m in southern Greenland
to the three volcanic members of the Eriksfjord formation, groups 4 and 5 comprise dyke samples from the mid Gardar, group 4 being the BD0 dykes and group 5 the NE and ENE dykes of the Ivigtut region, and group 6 comprises the lateGardar Nunarssuit-Isort6q dykes. The largest body of data (groups 7-9) pertains to the lateGardar intrusions of the Tugtut6q-Ilimaussaq zone and their ENE continuation (Upton & Fitton 1985). As stated above, the late-Gardar dykes of the Nunarssuit-Isort6q zone and the Tugtut6q-Ilimaussaq zone may be roughly contemporaneous. One outstanding feature of the basic magmas erupted in this continental intra-plate province was their relative compositional uniformity in space and time. Alkali-to-silica ratios (Fig. 5) and the absence of low-Ca pyroxenes denote their alkaline affinities (despite the fact that many are hy-normative at reasonable Fe oxidation ratios). All groups have KzO > 0.9% and TiO2 > 2%: The single most distinctive and unifying feature is their high mean AlzO3/CaO ratios
2 As will be shown in the following section, compositions of the basic dykes in the two zones are significantly different.
Compositions of Gardar basic rocks Since the olivine basalt and hawaiite magmas are considered to be parental to the salic rocks it is necessary to consider their compositions in some detail. Gardar basic rocks with the assemblage plagioclase-olivine-clinopyroxene rarely contain over 8% MgO, and it appears that basaltic magmas (as opposed to ultramafic lamprophyres) more magnesian than this were rarely, if ever, erupted during the Gardar activity. In this discussion the mean composition of 11 groups of fast-cooled phenocryst-poor basic rocks with between 4 and 8 wt.% MgO are compared (Table 1). The data are from analyses of over 200 samples of lavas or fne-grained dyke-rocks. Groups 1-3 correspond
%
+Na20
I
/
/
",\\
/ "
..~"
--e. + o.
9 9 "/o~ /~" o
",,
9
N
+
II "~
\\\
"(
o
9
9
O/ /_ /
I
I
m u g e a ri te s
I
\\/
t- "~0. ~' 9.Odl " e
o/
7 ///
/
.fL____~.., 'iIbasaltlc " i'+
9
9 / " +landesites
40 g + o ~ ~" ~ ./ 9 q l ~ " ~A e./
:, ;.'+
O, Oe)
\\
Ii
9
~r
+ /~/~-~.o ~"~Ko"~--~ . . . . . . . . 9. / +-~LQS' """ 9
,,{.+,
-
I
.-
4~ + A-I- + . /
,, 9
/
o
, 0 ~ / i/0
",, 9
/F
9 /'0 / O /'b ~ // % g hawaiites //o~ § OoO o o 6:, o ,6 q ,o I (30 0 / 0 ~o't" / " zO A jr" 0
/
%
\ / /
o/
I~"
/ /
I
/
I~O
lea
t I I
~,\
0
,/.~
/
/
I
\ t/
// /
! //
/
\
b
\
9
++
/"
~ ". \
a/~,
a
I t s
.f
./
_j
J
I
32
I
I
36
i
9
,
40
i
l
44
i
I
48
I
I
52
*
% SiO 2
FIG. 5. Alkali versus silica diagram for basic (MgO > 4 wt. %) lavas and hypabyssal intrusions of the Gardar province: Eriksfjord formation lavas; A, mid-Gardar dykes; O, late-Gardar dykes of Nunarssuit-Isortoq; O, late-Gardar dykes of the Tugtut6q-Ilimaussaq-Nunataq zone; m, ultramafic lamprophyres (early to late Gardar). The composition fields are those suggested by Cox et al. (1979). The Hawaiian alkali-tholeiitic basalt divide is indicated (--.--) for visual reference.
-
458
B. G. J. Upton & C. H. Emeleus
(1.75-2.40) which set these Mid-Proterozoic Gardar magmas apart from the earlier Proterozoic (Ketilidian) basalts of southern Greenland and from the Phanerozoic basalts exemplified by the Mesozoic 'coast-marginal' dykes and the early Tertiary flood basalts of both the E and W Greenland provinces. The A12Og/CaO ratios confer high contents of relatively sodic plagioclase (both normatively and modally) and concomitantly low contents of clinopyroxene. Typically, the chilled rocks are plagioclase phyric or plagioclase-olivine phyric, and the coarsePb
Rb
Ba
K
ND
grained products are troctolitic. Dykes in the Nain area of Labrador intruded in the interval 1450-1040 Ma have Fe-rich troctolitic compositions while possessing alkaline to transitional normative characteristics (Wiebe 1985). Clearly the broad characteristics of the Gardar basic magmas are shared by their time equivalents in Labrador. The absence of highly magnesian basalts has already been noted. Each of the group means has a Mg number (Mg = 100 x Mg/(Mg + Fe") atomic where Fe" is calculated assuming that Fe203/ La
Ce
,
400 •"
~ /A\ ...f"ll'~.\ ,," ,/ \ ~
Sr
T
Nd
,
P
r
Zr
T~
'
i
T
i
i
i
Y
'~ O G ; C "[ug'utoq 9 u 9 O t h e r E N E d y k e s T u g t u t 6 q zone 9 ENE dykes N u n a r s s u i t - I s o r t 6 q
2OO
100
5O
2O
Late Gardar i
i
i
i
20(
i
9
I
i
i
i
'i
d y k e s BDOs v
10(
Rock Chonc]~rite 5O
2O
Mid Gardar I
i
i
20i /
J 9 x x
I
I
i
I
9 Ilimaussaq v o l c a n i c 9 Ulukasik .....
I
I
I
I
I
I
I Nd
I P
I Zr
I Ti
I Y
member
lO(
5o
Early Ga rclar i Pb
I Rb
i Ba
I K
I Nb
I La
I Ce
I Sr
FIG. 6. Chondrite-normalized incompatible-element plots for averaged groups of Gardar basaltic-hawaiitic rocks. Early Gardar: group 1 (Mussartfit member), group 2 (Ulukasik member) and group 3 (Ilimaussaq member). Mid-Gardar : group 4 (BD0 dykes) and group 5 (younger ENE and NE dykes of the Ivigtut area). Late Gardar: group 6 (Nunarssuit-Isort6q dykes), group 7 (chilled margins, older giant-dyke complex, Tugtut6q), group 8 (chilled margins, younger giant-dyke complex, Tugtut6q) and group 10 (other ENE basic dykes of the Tugtut6q swarm). Group numbers correspond to columns in Table 1.
Mid-Proterozoic magmatism in southern Greenland FeO=0.15) of less than 50. Compositions of olivine phenocrysts are more Fe rich than Foso and typically more so than Fovo. Whereas the Gardar basalt-hawaiite magmas were characterized, as a whole, by a combination of high A1203/CaO and low Mg numbers, these features are most accentuated in (i) the early~ Gardar lavas and (ii) the late-Gardar dykes of the Tugtut6q-Ilimaussaq zone (and their ENE extensions). A general geochemical affinity between the various groups is brought out in chondritenormalized incompatible-element plots (Fig. 6; chondrite values from Sun (1980)). All tend to show high values for the most highly incompatible elements Pb, Rb and especially Ba. Almost all show low values of Nb (relative to neighbouring elements in the plot), the mid-Gardar Ivigtut dykes providing the only exception. A P peak is an obvious feature in the early-Gardar lavas and the late-Gardar Tugtut6q-Ilimaussaq dykes, and is weakly developed in the mean of the Nunarssuit-Isort6q dykes. The incompatible-element patterns of the early- and late-Gardar basic magmas, with Ba and K peaks and Nb troughs, are of interest. Thompson et al. (1983) have noted that Nb troughs are commonly developed in continental flood basalts, whereas strongly-alkaline basalts more generally display relative Nb enrichment. Ti is generally low (e.g. in comparison with average oceanic alkaline basalt (Sun 1980)) throughout the Gardar basalts. Whereas these
%
7
6
9
o
o
5
o
o
~
Oo
9
o
o
TiO 2 o A
~
o ,-..J'~§ ~-~ §
cP ~ ~~ o
31 , , ~ _ A , ~
1 -o Iio
o
o 2
o
o I
I 1
I
I 2
1
I 3
P~O s
I
I 4
I
I 5
[
%
FIG. 7. TiO 2 versus PzOsplot for Gardar basic dykes and lavas: +, Eriksfjord formation lavas; A, midGardar dykes; O, late-Gardar dykes of the Nunarssuit-]sort6q zone; 9 late-Gardar dykes of the Tugtut6q-Ilimaussaq-Nunataq zone.
459
ppm
1500 o
o
Sr
o
o
o
o
o
8
100o
+
~ o%~ 80
o
~ ~
o8
o oo 0 o o~oOOOOo % ~
o
"~ +%o o ~ 9 ~o~oo ~o oo o .++e + o+11oO +i' + t ,
+Ao
o
A
o
o
+
+
+
+
+.;
,,if+
+ i
,
10100
l
20100
I
3000
Bo
FIG. 8. Sr versus Ba plot for Gardar basic dykes and lavas. Symbols as for Fig. 7. geochemical features could have resulted from assimilation of crustal rocks, it is more likely that they reflect primitive characteristics. Despite overall similarities, the basic magmas of the Tugtut6q-Ilimaussaq zone dykes and their ENE continuations into the Nunataq region are geochemically distinct from all the earlier-Gardar basaltic rocks as well as from those of the Nunarssuit-Isort6q zone. Firstly, they tend to be more alkaline (Fig. 5), a feature largely due to comparatively high K20 and low SiO2 contents. They also tend to be richer in P, Ba, St, light rareearth elements (light REE) and Nb as is shown in Figs 7-10. Figure 7 shows the higher P2Os/ TiO2 ratios and higher P2Os absolute values. Figure 8 shows that, despite broad data scatter, the Tugtut6q-Ilimaussaq-Nunataq basic dykes tend to be richer in Ba and St, whereas Fig. 9 demonstrates the tendency for these dykes to be enriched in light REE (exemplified by La) and to have higher La/Y ratios. The most striking discriminant, however, is the Zr versus Nb plot (Fig. 10). Early- and mid-Gardar basalts and hawaiites and those of Nunarssuit-Isort6q show strong Z r - N b correlation with an average Zr/Nb ratio of about 18, whereas the Tugtut6q-Ilimaussaq-Nunataq dykes are distinctly richer in Nb with Zr/Nb ratios generally between 3 and 7. The fact that the major-element compositions of the Tugtut6q-Ilimaussaq-Nunataq basalts and hawaiites are otherwise similar to those of the other groups (e.g. similar Mg numbers) and that such relatively-incompatible elements as Ti and Zr are not notably concentrated suggests that the enrichment in P etc. did not stem from advanced crystal fractionation but indicates more
B. G. J. Upton & C. H. Emeleus
460 ppm
o
o o Oo
9 A
o
9 o
+~
+9
o
oo
Ii 9 c §
A+ c
~,~+ ~
+
+
~
~ 9 o<~Ao o
o +o
~
o
o~
06) @
c
~
A+ 9 + ~ o " A++ w~
oo
o
o
o
~-~'
o
o o
l++o o
I
210
I
I 40
I
6
/ 0
I
' 80
'
Lo
FIG. 9. Y versus La plot for Gardar basic dykes and lavas. Symbols as for Fig. 7.
fundamental differences, e.g. in source-rock compositions, depth or degree of the melting processes and/or wall-rock reaction processes. One conclusion from this study is that the lateGardar dykes of the Nunarssuit-Isort6q zone differ from their roughly contemporary counterparts in the Tugtut6q-Ilimaussaq zone in having both lower Mg numbers and distinctly lower contents of incompatible elements. Their comparatively primitive character is emphasized by high Ni contents. The few REE data for the Gardar basic rocks suggest that NunarssuitIsort6q dykes, as well as being less alkaline, have flatter chondrite-normalized patterns with distinctly less light-REE enrichment than those of the Tugtut6q-Ilimaussaq swarm (Fig. 11). Whereas the possibility of mantle heterogeneity cannot be ruled out, the existing data are compatible with the hypothesis that the Nunarssuit-Isort6q basic magmas resulted from more extensive (and shallower ?) mantle partial melting than did those of the Tugtut6q-IllmaussaqNunataq zone. Average compositions for the Keweenawan, Seal Lake and Coppermine Bay basalts were presented by Baragar (1977). The various groups of Gardar basalts (Table 1) show lower SiO2, Cr, Ni, V and Cu, but higher TiO2, P205, Rb and Sr, emphasizing their more alkaline character. The widespread nature of generally similar basaltic and hawaiitic magmas repeatedly erupted over a continental tract of up to 20 000 km 2 for some 200 Ma raises the possibility that they themselves may represent primary compositions. If so, they must have been gener-
ated from an abnormally Fe-rich mantle source and, in view of their high AlzO3/CaO ratios, one deficient in diopside relative to aluminous phases (e.g. garnet or spinel). It is more probable that the erupted magmas were themselves residues after extensive crystal fractionation; the high AlzO3/CaO ratios and low Mg numbers may reflect extended fractionation of olivine and clinopyroxene giving rise to dunite, wehrlite and/ or clinopyroxenite cumulates within the lower crust or lithospheric mantle. Beyond this stage it is suggested that the magmas were held at deep crustal or sub-crustal levels until fractionation of olivine and clinopyroxene had reduced their densities sufficiently for the relatively Fe-rich and diopside-poor troctolitic residual magmas to ascend through the crust during intermittent tensional episodes (Cox 1980; Morse 1982). Olivine (and clinopyroxene) separation during 'dry' fractionation of basaltic magmas leads to a density minimum typically in the 7-10 wt.% MgO range (Sparks & Huppert 1984). The high contents of normative plagioclase coupled with high Sr and lack of negative Eu anomalies suggests that many of the more basic magmas had undergone no significant plagioclase fractionation. Skeletal plagioclase phenocrysts are beautifully developed, often as radiating clusters giving rise to 'snow-flake' or 'star-basalt' textures, not only in the early-Gardar lavas but also in the late-Gardar dykes of the Tugtut6qIlimaussaq zone. Such textures imply that the magmas were erupted supersaturated with respect to plagioclase (Berg 1980).
ppm o o
t o I
+
o
300 9 9
o o o
+A
o o
o
2OO
~_ §
Zr 9
113o oo
100
1-§ , 20
,
| 40
| '
6%
'
80
| 1~0
, 120|
, 140
Nb FIG. 10. Zr versus Nb plot for Gardar basic dykes and lavas. Symbols as for Fig. 7.
Mid-Proterozoic magmatism in southern Greenland
461
O Older giant dyke, 86120 Tugtutoq
200
9 Younger ,,
40452
"9 Bongs Havn giant dyke 101386 Nunarssuit
100
"O~
9 Nunagorsuk
,,
,,
101338 Isort6q
5O Rock Chondrite 20
~'a C'e /r N'd -'
S'~ E'o dd
-'
0', H" E'r -'
~'b L'u
FIG. 11. Chondrite-normalized REE plot for chilled margins of two giant dykes from Tugtut6q and two giant dykes from the Nunarssuit and Isort6q areas. (Analyses by J. N. Walsh, University of London.)
Anorthosite cumulates and benmoreite residues The abundant occurrence of anorthositic xenoliths and plagioclase megacrysts in basic to intermediate Gardar intrusions has been taken as evidence for the large-scale occurrence of granular labradorite-andesine anorthosites at depth (Bridgwater 1967; Bridgwater & Harry 1968). Bridgwater (1967) concluded that (i) these formed as a continuation of the belt of Canadian anorthosites, (ii) they were produced throughout Gardar times as flotation cumulates, (iii) they tend to be more alkaline than their Canadian counterparts and (iv) their origin was intimately associated with the genesis of the Gardar syenites. The concentration of such feldspathic xenoliths and megacrysts in magmas of hawaiite to mugearite composition (Bridgwater & Harry 1968; Upton & Fitton 1985) is probably the result of density control. The anorthosites are inferred to underlie much of the province, probably as stratiform bodies in the lower or middle crust. It is concluded that batches of the high-alumina (troctolitic) magmas, withheld in the lower crust, crystallized slowly to yield anorthositic and troctolitic (and probably dunitic) cumulates. The residues from this relatively low-pressure fractionation, involving copious separation of plagioclase (and lesser olivine), evolved to the composition of benmoreite. During formation of the anorthositic cumulates, liquid densities are likely to have risen (Sparks & Huppert 1984). However, following the onset of Ti-magnetite crystallization (typically in the mugearitic stages of magma evolution), the benmoreitic magmas may have attained densities low enough to permit
further ascent by stoping to produce the large augite syenite intrusions. Fine-grained apparently uncontaminated chilled marginal facies of the Ilimaussaq augite syenite occur around its southern contact zone. An analysis of this chili, approximating the composition of the Ilimaussaq parental magma, is given in Table 2. Benmoreitic compositions similar to this were abundantly represented in the Tugtut6q-Ilimaussaq-Nunataq dyke swarm prior to the emplacment of the Ilimaussaq intrusion (Table 2). The further evolution of these magmas appears to have involved both crystal fractionation and liquid fractionation, the latter possibly occurring by thermodiffusion or volatile transfer or both. Anorthoclase was the dominant phase in the crystallizing assemblages, resulting in increasing impoverishment of residual magmas in Ba and Sr and giving rise to strong negative Eu anomalies.
Highly differentiated salic magmas Peralkaline (typically persodic) compositions with high concentrations of Zr, Nb, U, Th, Zn, Pb, Sn, Ga, Rb, Li, Y and REE occur in both the silica-poor (phonolitic agpaitic) and silica-rich (comenditic) salic associations. Eudialyte and aenigmatite are prominent among the wide range ofpersodic minerals crystallized from the agpaitic magmas. Zr-rich aegirines (up to 7% ZrO2) have been described from peralkaline rocks at Motzfeldt (Jones & Peckett 1980). Table 3 shows four analyses from agpaitic dykes in and close to the Ilimaussaq intrusion, which were presented by Martin (1985) and Larsen & Steenfelt (1974). Inasmuch as these are analyses of relatively fast-
462
B. G. J. Upton & C. H. Emeleus
TABLE 2. Composition of the chilled margin of the Ilimaussaq augite syenite and some benmoreite dykes from the Tugtut6qIlimaussaq-Nunataq swarm 1
2
3
4
5
Major elements (wt. %) SiO 2 54.99 5 4 . 0 6 Al:O3 15.48 1 4 . 8 7 Fe203* 11.53 1 2 . 0 0 MgO 1.60 1.51 CaO 4.39 4.15 Na20 4.79 4.59 K20 4.64 4.89 TiO: 1.95 1.82 MnO 0.22 0.24 P20s 0.76 0.73
55.84 15.21 10.76 1.33 3.71 4.70 5.30 1.68 0.20 0.61
5 4 . 6 6 55.59 1 7 . 1 9 16.44 8.99 9.81 1.64 1.43 4.24 3.23 4.92 4.89 4.77 5.12 1.58 1.75 0.24 0.23 0.65 0.65
Total
99.33
98.88
100.35 98.85
Trace elements (ppm) Ni 3 4 Cr 5 6 V --Sc 29 17 Cu 23 21 Zn 127 168 Sr 312 223 Rb 82 103 Zr 268 493 Nb 48 66 Ba 2438 1539 Pb 6 17 Th -4 La 69 99 Ce 142 212 Nd 71 94 Y 45 57 La/Y K/Rb Zr/Nb
1.53 467 5.58
1.73 370 7.47
4 5 -19 14 158 201 99 398 55 1570 14 4 76 170 78 49 1.55 444 7.24
5 6 27 15 6 142 577 97 577 130 2113 18 2 96 206 86 46 2.09 421 4.44
99.14 3 4 -21 8 139 439 82 380 75 2542 7 -94 191 88 50 1.88 518 5.07
1, GGU 40528, Ilimaussaq augite syenite marginal zone, N side of Kangerdluarsuq; 2, GGU 212117, 15 m ENE-trending dyke, Mellemlandet; 3, GGU 212141, 12 m ENE-trending dyke, Mellemlandet; 4, ARM 281/80, 20 m ENE-trending dyke, N shore of Tunugdliarfik fjord; 5, ARM 285/80, 3 m ENEtrending dyke, N shore of Tunugdliarfik fjord. GGU numbers relate to collections of the Geological Survey of Greenland, ARM numbers to Martin (1985). Fe203* ,
total Fe as
Fe203.
chilled dykes, they can be taken as rough guides to approximate agpaite magma compositions. Table 3 also includes analyses of two (devitrifled) comendite dykes from Tugtut6q (Martin 1985) and of two more-slowly-cooled granitic intrusive sheets from Kfingn~t (Macdonald et al. 1973). The Kfingn~tt granites, like those of
Malenefjeld (in the Nunarssuit area), Ilimaussaq, Dyrnaes-Narssaq, Central Tugtut6q and others, are typical A-type granites ('alkaline, anorogenic and anhydrous') (Loiselle & Wones 1979; Collins et al. 1982). The characteristic geochemical (highGa/A1, high-Nb, large-Eu depletions) and mineralogical (alkali feldspar, annite-rich biotites, fluorite) features are well displayed (Upton 1960; Stephenson & Upton 1982; Martin 1985). The Ilimaussaq 'green granite' is of interest for its content of Be-bearing minerals (e.g. epididymite NaBe(OH) (Si307)) (Hamilton 1964). One of the most extreme silicic magma compositions from the Gardar province is represented by a devitrifled pantelleritic trachyte dyke in the Tugtut6q area which contains a range of persodic minerals including narsarsukite (Na2(Ti,Fe)Si4010(F)), Zn-nordite (Na3Ce(Sr,Ca)ZnzSi6018), emeleusite (LizNa4Fe~'Si12030) and N a - R E E - r i c h apatite (Upton et al. 1976, 1978). Such features of the alkali granites, comendites etc. as low K / R b ratios (less than 100; Table 3) and high Ga/AI ratios (Ga x 104/A1 values up to 6 (Upton 1960)), combined with low initial Sr isotopic ratios and the field evidence, indicate that they formed by differentiation of trachytic (syenitic) magmas and not by partial melting of an anhydrous halogen-rich granulitic source (Collins et al. 1982).
Carbonatites and ultramafic lamprophyres The largest carbonatite body (of early-Gardar age) cuts the foyaite complex at Gr~nnedal-Ika. It is composed mainly of calcite, siderite, magnetite and lesser amounts of apatite, sphalerite, pyrite and Sr-bearing barite (Emeleus 1964). Since there appears to be a continuum from cancrinite foyaite to carbonatite in this complex, the carbonatite probably represents a differentiate from a CO2-rich phonolitic magma. Other carbonatitic bodies in the province, however, appear to be intimately related to a suite of silica-deficient ultramafic lamprophyres (UML). This association is seen in the eastern part of the province in the vicinities of the Motzfeldt, Q6roq and Igdlerfigssalik complexes and is well developed at Qagssiarssuk, a few kilometres W of S Q6roq (Stewart 1970). Comparable U M L also occur in the Tugtut6qIlimaussaq area and in the vicinity of Kfingnfit and Ivigtut. Inasmuch as these may represent primitive (or even primary) small-degree melt extracts from the mantle they hold an intrinsic interest quite disproportionate to their actual
463
M i d - P r o t e r o z o i c m a g m a t i s m in southern Greenland TABLE 3. Compositions of some highly fractionated salic hypabyssal rocks from the Gardar province Silica,oversaturated rocks
Silica-undersaturated rocks
1
2
3
4
5
6
7
8
71.67 10.11 3.99 0.02 2.94 3.87 4.35 0.18 0.04 -ND ND ND
75.69 10.40 4.40 0.01 0.01 3.43 4.40 0.15 0.03 -ND ND ND
63.9 14.79 7.12 0.02 0.85 7.93 4.23 0.24 0.14 0.04 0.28 0.84 0.006
--
0.36
0.04
--
--
0.51
0.41
97.17
98.52
100.03
100.25
94.54
92.44
99.72
99.40
ND ND ND 490 30 470 180 2440 ND 45 60 55 30 805 1550 ND 220
ND ND ND 590 80 95 65 3140 ND 75 770 360 110 420 940 ND 480
4 11 -3495 9i 471 ND 1647 254 39 545 1154 ND 5361 7863 1859 851
5 --2613 68 6 ND 1374 276 46 188 735 ND 3200 6638 2279 1366
ND ND 10 ND 47 ND ND 4900 1100 < 10 ND ND 19 437 868 ND 350
ND ND 10 ND 81 ND ND 4800 1100 < 10 ND ND 20 453 900 ND 365
3.66 75 --
0.88 22 --
6.30 193 6.48
2.34 83 4.98
1.25 -4.45
1.24 -4.36
Major elements (wt. ~ ) SiOz A1203 Fe203* MgO CaO Na:O K20 TiO2 MnO
PzOs H2 O+ F C1 O - F , CI Total
75.3 11.10 4.20 0.03 0.15 7.95 0,26 0.30 0.16 0.08 0.21 0.07 0.48
50.64 11.29 16,91 0.12 0.37 11.24 2.12 0.27 0.71 0.87 ND ND ND
52.04 12.53 12.60 0.25 1.04 t2.30 0.06 0.21 0.80 0.62 ND ND ND
51.71 13.89 12.96 0.21 2.58 10.00 4.87 0.53 0.42 0.12 1.54 1.54 0.97
51.88 13.81 13.08 0.15 2.73 8,63 4.95 0.53 0.45 0.15 2.42 2.42 0~10
Trace elements (ppm) Ni Cr V Zn Sr Rb Li Zr Nb Ba Pb Th U La Ce Nd Y
3 2 -277 78 694 ND 5736 631 7 137 92 ND 495 1168 487 515
4 4 5 48 15 420 ND 3918 545 30 27 118 ND 189 474 193 346
La/Y K/Rb Zr/Nb
0.96 52 9.10
0.55 87 7.19
1, ARM 117/80, 2 m ENE-trending comendite dyke, Tugtut6q (Martin 1985); 2, ARM 219/80, 1.5 m ENEtrending comendite dyke, Igdlutalik (Martin 1985); 3, GGU 86157, aplitic sheet, W Kfingnfit (Macdonald et al. 1973); 4, G G U 86190, aplitic sheet, W Kfingn~t (Macdonald et al. 1973); 5, ARM 6/81A, 7.5 cm ENEtrending lujavrite dyke chill, Kvanefjeld plateau (Martin 1985); 6, ARM 8/81, 1.5 m ENE-trending lujavrite dyke chill, Kvanefjeld plateau (Martin 1985); 7, G G U 22093, 10-30 m micro-kakortokite dyke (high-alkali facies) SE of Ilimaussaq (Larsen & Steenfelt 1974); 8, G G U 42474, 10-30 m micro-kakortokite dyke (high-alkali facies) SE of Ilimaussaq (Larsen & Steenfelt 1974). Fe203*, total iron as Fe203. ND, not determined.
quantity. T h e U M L are m a i n l y r e p r e s e n t e d by dykes and o t h e r m i n o r intrusions but also occur as lavas within the Eriksfjord f o r m a t i o n (Stewart 1970; J. G. L a r s e n 1977, and u n p u b l i s h e d data).
T h e Qagssiarssuk o c c u r r e n c e s display an association of c a r b o n a t i t e (s6vite) and silicate rocks (monchiquite, aln6ite, mica-peridotite, melilitite and m i c a pyroxenite) a n d m a r k the site o f an
464
B. G. J. Upton & C. H. Emeleus
TABLE 4. Compositions of some ultramafic lamprophyres from the Gardar province 1
2
3
4
5
6
7
21.46
Major elements (wt. ~) SiO2 A1203 Fe203 MgO CaO Na20 K20 TiO2 MnO PzOs LOI
38.73 8.23 17.02 12.97 10.47 1.43 1.85 3.69 0.21 0.82 4.11
35.07
9.79 17.38 9.33 11.94 1.93 3.22 5.90 0.23 0.79 3.64
42.57 9.48 15.65 16.10 6.21 0.63 2.99 3.29 1.10 0.39 2.03
36.65 6.95 12.25 27.99 4.45 0.03 0.10 4.06 0.41 0.67 7.76
36.20 9.07 20.55 15.90 5.55 0.36 3.13 5.47 0.23 0.67 1.83
3.46 16.44 7.85 20.14 -1.13 2.24 0.58 0.98 24.16
28.94 2.53 13.81 8.52 16.97 -0.33 1.87 0.52 1.02 24.17
Total
99.60
99.22
99.44
100.32
98.97
98.42
98.62
Trace elements (ppm) Ni Cr V Sc Cu Zn Sr Rb Zr Nb Ba Pb Th La Ce Nd Y
438 252 277 20 77 86 1018 52 224 5 1044 3 1 38 81 42 226
1449 170 406 23 81 125 1164 109 422 107 1213 1 5 57 122 56 34
528 731 367 27 4 121 536 100 208 43 1175 --33 56 26 20
893 945 315 23 -91 140 14 293 78 3 6 3 11 26 19 20
717 712 285 30 199 175 639 160 329 81 1626 --56 118 661 29
129 103 172 10 44 296 2936 36 265 274 2535 23 32 437 905 379 82
176 197 152 8 76 152 2219 16 148 271 655 20 14 343 692 272 65
La/Y K/Rb Zr/Nb
1.46 298 4.07
1.68 245 3.94
1.65 248 4.83
0.55 59.3 3.76
1.93 162 4.06
5.33 261 0.97
5.28 169 0.55
1, GGU 81116, 1.5 m ENE-trending UML dyke, Kajartalik (Kfingngtt area); 2, GGU 81119, 3 m ENE-trending UML dyke, Kajartalik (Kfingn~,t area); 3, GGU 101221, c. 10 m thick lava unit, Nunarssarnaussaq (Ilirnaussaq area); 4, GGU 181917, c. 10 m thick lava unit, Nunarssarnaussaq (Ilimaussaq area); 5, GGU 101258, UML, Karra, Igdlutalik (Tugtut6q area); 6, GGU 216638, ultramafic sheet in diatreme, Qagssiarssuk; 7, GGU 216639, ultramafic sheet in gneiss, Qagssiarssuk. Fe203*, total Fe as Fe203. e a r l y - G a r d a r volcano (Stewart 1970). N o fresh melilite, however, has been preserved, the identification being based on the m o r p h o l o g y of c a r b o n a t e pseudomorphs. H o w e v e r , D e a n s & Roberts (1984) have suggested that c a r b o n a t e p s e u d o m o r p h s at Qagssiarssuk m a y be after nyerereite (Na2Ca(CO3)2) r a t h e r t h a n melilite. Biotite-rich U M L dykes were intruded early in the m i d - G a r d a r swarms of the Ivigtut area (Berthelsen & H e n r i k s e n 1975; P a t c h e t t et al. 1978). Small intrusions of U M L (often rich in c a r b o n a t e and sulphide) are relatively a b u n d a n t
b e t w e e n Qagssiarssuk and the n u n a t a q s N of Motzfeldt and intrude the N Q6roq a n d Motzfeldt centres (Walton 1965; E m e l e u s & H a r r y 1970; Jones 1980). Some at least are of l a t e - G a r d a r age ( U p t o n & Fitton 1985). Hypabyssal intrusions of l a t e - G a r d a r U M L also occur in the Tugtut6q area w h e r e they have been described as fine-grained m i c a pyroxenites and m i c a peridotites ( U p t o n & T h o m a s 1973). W h e r e a s the n a m e s aln6ite, m o n c h i q u i t e , biotite pyroxenite, biotite peridotite a n d ultramafic lava have been variously used to describe the m o r e silicate-rich m e m b e r s of the
Mid-Proterozoic magmatism in southern Greenland association (Stewart 1970; Upton & Thomas 1973; J. G. Larsen 1977), the terms aln6ite and aillikite appear most appropriate according to whether or not melilite appears to have been an original component (Rock 1986). Carbonated aln6itic rocks have, so far, been described only from the Qagssiarssuk and Motzfeldt areas (Stewart 1970; Jones 1980). It appears likely that these silica-poor rocks form a coherent suite ranging from silicate-FeTi oxide-rich (UML) variants to carbonate-rich members, with silica contents from about 38% to less than 5%. They are feldspar-free, with the UML members dominated by olivine, clinopyroxene, biotite and FeTi oxides. With increasing modal carbonate the suite appears to grade towards beforsites and s6vites. While it is often difficult to determine whether carbonate is secondary or primary, particularly in the UML, there can be no doubt that the UML magmas were CO2-rich. Much of the mineralogical alteration (carbonatization, serpentinization etc.) that is common in this suite is probably deuteric. The mineralogy of some of the UML is known to be highly complex: melanite, monticellite, wollastonite and perovskite are among the many minerals occurring in these silica-poor rocks (Craven 1986). Apatite is ubiquitous throughout the UML-carbonatite association and phases rich in Ba, Sr and light REE occur in the carbonatitic rocks. Some representative UML analyses are presented in Table 4. Compositionally the UML are distinct in combining low SiO 2 and A1203 with high CaO, FezO3 (total Fe), MgO, TiO2 and, commonly, CO2. Alkali contents vary widely (Fig. 5), but the U M L are typically potassic with K20 > Na/O. They are often leucite normative and sometimes larnite normative. However, despite the extreme nature of some of the compositions it is probable that the UML grade, with increasing silica, via feldspathic lamprophyres, into more normal alkali olivine basaltic compositions. Among the trace elements, high Ni and Cr contents (several hundred parts per million) are seen together with high concentrations of Ba, Sr and light REE. Th, Zn and Nb are abundant relative to their concentrations in the basaltic rocks. Ba/Sr, Zr/Nb and K/Rb ratios tend to be lower than in the Gardar basalts and hawaiites while La/Y ratios are generally higher. Zr/Nb ratios are typically less than 4 (and commonly less than unity) and K/Rb ratios are typically less than 300 (and commonly less than 200). LightREE enrichment is pronounced, with La/Y ratios frequently greater than 5. The high Mg, Ni and Cr contents suggest that
465
the UML represent unfractionated primitive compositions. The UML dykes often contain small carbonated inclusions that may be altered xenoliths of mantle peridotite. A xenolith of s6vite in one such dyke provides further evidence for a close genetic link between carbonatites and UML. A UML intrusion on Igdlutalik (near Tugtut6q) is host to abundant ultramafic xenoliths. These are severely altered but include relict peridotite (with olivine (Fo81)) and glimmerite xenoliths of presumed mantle origin. Chondritenormalized incompatible-element patterns for three aillikitic intrusions are shown in Fig. 12. The most obvious features are troughs at Ba, K and Sr. Ba (and P) peaks and Nb troughs noted for the basaltic compositions are conspicuously absent. Little detailed geochemical or mineralogical study of the carbonatites has yet been made. A reconnaissance study of a set of UML and carbonatite intrusions from the E of the province showed that the carbonatites were enriched in Sr, Ba, REE and Y relative to the UML, but much poorer in Ni, Cr and V (Upton & Fitton 1985). The Sr contents show an inverse correlation with SiO/, rising to over 10 000 ppm in the least siliceous carbonatites; Ba/Sr ratios tend to be even lower than those of the UML. Zr/Nb ratios are variable but appear to be generally as low as, or lower than, those of the UML while the high La/Y ratios are similar to those of the UML.
Alkali metasomatism in the vicinity of Gardar intrusions Many clear examples of late-stage loss of alkalibearing fluids during crystallization of the alkaline intrusions have been recorded. The most striking of these include the fenitization of the early-Gardar lavas in the roof zone of the Ilimaussaq intrusion (Sorensen et al. 1974; Kunzendorf et al. 1982) and the extensive hydrothermal alteration and mineralization of the trachyte and phonolite lavas composing the roof of the Motzfeldt complex (Jones 1980; Tukiainen et al. 1983). Apart from alkalis, Th, U, Zr, Nb, REE and F were also transferred to the country rocks. Bailey et al. (1981) have drawn attention to the high Th/U ratios of the late fluids generated at Ilimaussaq. Metasomatism of the older granites around the N Q6roq syenite affected rocks up to 200 m from the contact. Metasomatizing fluids conveyed Na, K, Fe, Mg, Ca, CO2 and F, and resulted in secondary formation of albite, NaFe amphiboles
466
B. G. J. Upton & C. H. Emeleus
1000
Roc k Chondrite 100
f
9 Aillikite sill:Mellem Lander 2121s6 Aillikite dyke : Narssarssuaq:1012 56 9 Mela-aillikite: mean of three samples
10
from plug on lugtutbq l
Rb
Ia
B
I
Th
I
K
N
L
I
La
I
Ce
I
Sr
I
Nd
I
P
I
Zr
I
Ti
'
Y
FIG. 12. Chondrite-normalized incompatible-element plots for some UML (aillikite) compositions.
and pyroxenes, carbonates and fluorite (Chambers 1976). Gabbros adjacent to quartz syenites and alkali granites at K~ngn~.t and Narssaq underwent enrichment in K, F and H20. Of the trace elements involved in the metasomatism at Kfingnfit, Zn, Th, U, Cs, Pb, Li and Rb were prominent (Macdonald et al. 1973).
Volatile component of Gardar magmas The loss of water from slowly cooled Gardar syenites and granites has been alluded to in the previous section. However, crystallization sequences in the basic rocks suggest that the basaltic magmas were relatively anhydrous (Upton & Thomas 1980). Nevertheless the ubiquitous presence of F-bearing minerals qualitatively suggests that this element was not only concentrated (as'complex fluoride ions?) in residual fluids, but was also relatively abundant in the primitive basaltic magmas. The high concentration (more than 1900 ppm) of F in the chilled marginal facies of one of the giant dykes is thought to imply high F contents in the relatively unfractionated basaltic magmas. There is clear evidence in some of the differentiated intrusions that F was fractionated into early-crystallizing apatite (Stephenson & Upton 1982; Upton et al. 1985).
Whereas F-bearing biotites, amphiboles and apatites are common in the Gardar intrusions, fluorides occur in the late-stage differentiates of the intrusive complexes. Thus fluorite is a common accessory in several of the quartz syenite and alkaline granite intrusions and in some of the foyaites. It is also widespread as a vein, jointcoating and amygdale mineral (frequently in association with carbonates) in the vicinity of the main complexes and is conspicuous among the metasomatized rocks referred to in the previous section. It is one of many fluorides present in the famous Ivigtut pegmatite, formed from the latest volatile-rich differentiates of the alkaline granite magma (Berthelsen & Henriksen 1975). The Ivigtut assemblage however, is, dominated by cryolite (Na3A1F6). Villiaumite (NaF) is prominent in some facies of the Ilimaussaq agpaitic lujavrites (Bondam & Ferguson 1962). The large size (more than 50 cm) of crystals seen in anorthositic xenoliths and plagioclase megacrysts, the remarkable development of cumulate layering in rocks ranging from anorthosite to syenite to foyaite and even granitic compositions (Emeleus & Upton 1976), and the strong in situ fractionation even within small intrusions all provide evidence for unusually low magma viscosities that may well be related to high F contents. The role of C1 is emphasized by the common development of sodalite in the nepheline syenites.
Mid-Proterozoic magmatism in southern Greenland This is particularly so with respect to the remarkable sodalite cumulates (naujaites) of the Ilimaussaq agpaitic suite. Primary magmatic carbonates are also widely developed in the Gardar alkaline rocks. Siderite and calcite (sometimes in association with fluorite) occur interstitially and as pegmatite minerals in the syenites at Kfingnfit and the Tugtut6q giant dykes. Fluid inclusions in the granites and quartz syenites vary from CO2-CH4(-H20) mixtures to aqueous saline fluids (Konnerup-Madsen 1984). Konnerup-Madsen has observed up to five different solid phases in these inclusions including two occasionally rhombohedral phases that are probably carbonates and another that may be nahcolite (NaHCO3), as well as halite and sylvine. The Ivigtut cryolite deposit contained some 20% siderite (Berthelsen & Henriksen 1975), possibly resulting from a reaction, in a CO2-F-rich alkali granite magma, of the type : NaA1Si308 + 2NaFe"Si206 + 3F 2 -k-2CO2 --~ Na3A1F6 + 2Fe"CO3 + 7SIO2 + 202 postulated by Stormer & Carmichael (1970). Interstitial carbonate sometimes occurs in the foyaites (Emeleus 1964; Upton & Fitton 1985), but more commonly CO2 in the undersaturated magmas became incorporated in cancrinite (Emeleus 1964; Emeleus & Harry 1970). At Ilimaussaq, however, CO3-bearing minerals are rare or absent and the highly reduced agpaitic magmas crystallized graphite (Sorensen et al. 1981), while fluid inclusions contain a variety of hydrocarbons (Petersilie & Sorensen 1970; Konnerup-Madsen et al. 1981 ; Sorensen et al. 1981). CO2-rich magmas clearly played a major role in the evolution of the UML-carbonatite suite. Fluorine also appears to have been an important constituent and fluorite occurs in, or in association with, some of the carbonatites (Emeleus & Harry 1970; Stewart 1970). Carbonatite dykes in the Motzfeldt area contain up to 40% modal fluorite. Apart from evidence that the dissolved volatile components in the Gardar magmas were comparatively H20 poor but rich in halogen and C compounds, the significant modal contribution of sulphides in many of the UML and carbonatitic rocks suggests that S-bearing phases were also present in significant amounts.
Petrogenetic summary and conclusions The model outlined above for the genesis of the alkaline and transitional olivine basalt magmas with high A1203 and low Mg number that were
467
typical of so much of the Gardar activity and which are thought to be parental to the major alkaline complexes is similar to views expressed by Thompson et al. (1983) for the evolution ofdipoor hawaiite magmas at or near the base of the crust, as well as to conclusions reached by Emslie (1971, 1977) with respect to the Proterozoic troctolitic magmas of Labrador. Morse (1981, 1982), considering the genesis of the high-A1 magmas in Labrador, concluded that the explanation lay in the partial-melting process in the mantle rather than in subsequent crystal-fractionation histories. He inferred a depleted spinel-rich lherzolite or harzburgite source beneath an incipient continental rift. Magma genesis involved exhaustion of clinopyroxene in the source rock; spinel was largely, but not wholly, consumed, thus presumably leaving refractory spinel harzburgite. Morse's (1982) model appeals to a mantle source which was abnormally rich in Fe, ab and STSr but generally depleted in LIL elements. Taylor et al. (1983) agreed that the magmas responsible for the Proterozoic massif anorthosites (and, by implication, for the Gardar basic magmas) were related to incipient or failed rifting and concluded that the source rocks must have been rich in AI, Sr and Eu, but 'dry', poor in LIL elements and with low Rb/Sr ratios. While these features could relate to abnormal mantle, these workers proposed that lower crustal granulites constituted the magma source. However, in the Gardar province, a crustal source for the aluminous (troctolitic) magmas appears untenable. Even if these magmas are near primary in composition, almost total (dry) melting of the granulite source rocks would be called for, requiring quite exceptional thermal gradients. We conclude that the Gardar basic magmas were derived from the mantle. This may have been anomalous to the extent of having had unusually low cpx/gt (or cpx/sp) ratios. The geochemistry of the erupted magmas was largely controlled by crystal fractionation at all levels, but particularly at or close to the crust-mantle boundary. The combination of low 87Sr/86Sr(i) ratios with high contents of Rb, Ba, K and P may denote mantle metasomatism (with secondary growth of phlogopite and apatite ?) shortly prior to melting. The pervasive (if qualitative) evidence in the province for the importance of halogen and Crich volatile components prompts the speculation that the cycles of Gardar magmatism might have been initiated by infiltration of F-, CI- and C-rich fluids from the deep mantle, promoting metasomatism and partial melting in the deep subcontinental lithosphere, in response to extensional tectonic stresses.
468
B. G. J. Upton & C. H. Emeleus
Very small-scale partial melting (mainly involving carbonates, phlogopite and apatite, with lesser olivine and clinopyroxene (Wyllie 1977)) is inferred to have produced highly-mobile magmas which ascended as near-primary U M L and carbonatites. (Some of the Gardar carbonatites, however, appear to be crystallization residues from salic magmas.) Larger-scale melting, resulting not only in consumption of mantle clinopyroxene but also involving a major contribution from an aluminous phase (garnet?), produced the primary magmas which, after fractionation, gave rise to the typical plagioclase-saturated basaltic and hawaiitic magmas. Further fractionation of these within the crust generated the anorthosite cumulates and the benmoreite (or augite syenite) magmas parental to the major complexes. Differentiation of these benmoreitic magmas is thought to have involved both crystal fractionation and liquid fractionation leading, in extreme instances, to the generation of either phonolites of agpaitic composition or to peralkaline rhyolites. The model favoured here thus differs from those proposed for the generally similar MidProterozoic alkaline associations of Central Labrador and Pike's Peak, Colorado, where crustal anatexis caused by the ascent of basalt has been considered the principal genetic process giving rise to the salic magmas (Barker et al. 1975; Collerson 1982). The incompatible-element patterns (Fig. 6) show closest similarity between (i) the earlyGardar lavas and (ii) the late-Gardar basic rocks, particularly those of the Tugtut6q-Ilimaussaq (-Nunataq) zone. There is some resemblance between these patterns and those for the midGardar BD0 dykes. The least similarity is shown by the mid-Gardar dykes of the Ivigtut region. This suggests the possibility of lithospheric control on the geochemistry, with the Ivigtut dykes being controlled by the Archaean craton, the BD0 dykes traversing both craton and adjacent Ketilidian rocks, and both the early-
Gardar lavas and the late-Gardar intrusions lying well away from the craton margin. The degree of crustal contamination in the basic magmas requires evaluation. Their low 87Sr/S6Sr ratios indicate that any such contamination is likely to have been small scale or to have involved relatively non-radiogenic lower crustal (granulite-facies) rocks. The Nb(-Ta) troughs common in chondrite-normalized trace-element patterns of continental flood basalts are ascribed by Thompson et al. (1983) to contamination by readily-fusible crustal rocks which themselves have marked Nb(-Ta) troughs. While this process might be appealed to for the Gardar basalts, it is implausible to attribute the Ba and P peaks to such a process; the sub-craton source rocks may have been more depleted in K, Ba and P, while having higher Nb (or lesser ability to retain it during melting) than the source rocks beneath the adjacent Proterozoic terrain. The pronounced geochemical difference between the basaltic-hawaiitic magmas of the lateGardar Tugtut6q-Ilimaussaq-Nunataq dykeswarm and those produced at other times and places in the Gardar province may denote that the source rocks beneath this lineament had been affected by a metasomatic event by which the mantle source rocks were differentially enriched in some incompatible elements, particularly K, Ba, P and light REE.
ACKNOWLEDGMENTS:The manuscript benefited from comment and criticism by J. A. Craven, J. G. Fitton, R. Macdonald, A. R. Martin, I. Parsons, H. Sorensen, L. M. Larsen and W. S. Watt, to whom we extend our thanks. We are also grateful to A. R. Martin for permission to reproduce some of his analyses. Our thanks go also to D. James for X-ray fluorescence analyses, to J. N. Walsh for ICPS analyses of REE and to P. Stewart for preparation of the manuscript. Additionally, we are grateful to the Director of the Geological Survey of Greenland for permission to publish.
References BAILEY, J. C., ROSE-HANSEN, J., LOVBORG, L. & SORENSEN,H. 1981. Evolution of Th and U wholerock contents in the Ilimaussaq intrusion. In." J. C. BAILEY, L. M. LARSEN& H. SORENSEN(eds) The Ilimaussaq Intrusion, South Greenland. A Progress Report on Geology, Mineralogy, Geochemistry and Economic Geology, Rapp. Gronlands geol. Unders,
103, 87-98. BARAGAR,W. R. A. 1977. Volcanismofthe stable crust. In." Volcanic Regimes in Canada, Geol. Ass. Can. Spec. Pap. 16, pp. 377-405.
BARKER,F., WONES,D. R., SHARP, W. N. & DESBOROUGH, G. A. 1975. The Pike's Peak Batholith, Colorado Front Range, and a model for the origin of the gabbro-anorthosite-syenite-potassic granite suite. Precambrian Res. 2, 97-160. BECKER S. M. 1984. Petrology of the giant dykes of Isortoq and feldspar variation in the Klokken complex, South Greenland. PhD Thesis, University of Aberdeen (unpublished). BERG, J. H. 1980. Snowflake troctolite in the Hettasch intrusion, Labrador: evidence for magma mixing
Mid-Proterozoic magmatism in southern Greenland and supercooling in a plutonic environment. Contrib. Mineral. Petrol. 72, 339-51. BERTHELSEN, A. & HENRIKSEN, N. 1975. Geological Map of Greenland. 1 .100,00, Ivigtut 61 v. 1 Syd. The Orogenic and Cratogenic Geology of a Precambrian Shield Area, 169 pp. Geological Survey of Greenland, Copenhagen. BLAXLAND,A. B., VAN BREEMEN,O., EMELEUS,C. H. & ANDERSON, J. G. 1978. Age and origin of the major syenite centers in the Gardar province of South Greenland: Rb-Sr studies. Geol. Soc. Am. Bull. 89, 231-44. BONDAM, J. 8z FERGUSON, J. 1962. An occurrence of villiaumite in the Ilimaussaq Intrusion, South Greenland. Meddr. Gronland, 172 (2). BRIDGWATER, D. 1967. Feldspathic inclusions in the Gardar igneous rocks of South Greenland and their relevance to the formation of major anorthosites in the Canadian Shield. Can. J. Earth Sci. 4, 995-1014. & COE, K. 1970. The role of stoping in the emplacement of the giant dykes of Isortoq, South Greenland. Geol. J. Spec. Iss. 2, 67-78. & HARRY, W. T. 1968. Anorthosite xenoliths and plagioclase megacrysts in Precambrian intrusions in South Greenland. Bull. Gronlands geol. Unders. 77 (also Meddr. Gronland, 185 (2)). BROWN, W. L., BECKER, S. M. & PARSONS, I. 1983. Cryptoperthites and cooling rate in a layered syenite pluton. Contrib. Mineral. Petrol. 82, 13-25. CHAMBERS,A. D. 1976. The petrology and geochemistry of the North Q6roq Centre, Igaliko complex, South Greenland. PhD Thesis, University of Durham (unpublished). COLLERSON, K. D. 1982. Geochemistry and Rb-Sr geochronology of associated Proterozoic peralkaline and subalkaline anorogenic granites from Labrador. Contrib. Mineral. Petrol. 81, 126-47. COLLINS, W. J., BEAMS, S. D., WHITE, A. J. R. & CHAPPELL, B. W. 1982. Nature and origin of Atype granites with particular reference to southeastern Australia. Contrib. Mineral. Petrol. 80, 189200. Cox, K. G. 1980. A model for flood-basalt vulcanism. J. Petrol. 21,629-50. --, BELL, J. D. & PANKHURST, R. J. 1979. The Interpretation of Igneous Rocks, Allen & Unwin, London. CRAVEN, J. A. 1986. The petrogenesis of some ultramafic rocks from the Gardar Province, SW Greenland. PhD Thesis, University of Edinburgh (unpublished). DEANS, T. & ROBERTS, B. 1984. Carbonatite tufts and lava clasts of the Tinderet foothills, western Kenya: a study of calcified natrocarbortatites. J. geol. Soc. Lond. 141,563-80. EMELEUS, C. H. i964. The Gronnedal-[ka alkaline complex, South Greenland. The structure and geological history of the complex. Bull. Gronlands geol. Unders. 45 (also Meddr. Groland, 172 (3)). & HARRY, W. T. 1970. The Igaliko nepheline syenite complex. General description. Bull. Gronlands geol. Unders. 85 (also Meddr. Gronland, 186 (3)). -
-
--
-
8,~ UPTON, B. G. J. 1976. The Gardar period in Southern Greenland. In: A. ESCHER• W. S. WATT (eds) The Geology of Greenland, pp. 153-81. The Geological Survey of Greenland, Copenhagen. EMSLIE, R. F. 1971. Liquidus relations and subsolidus reactions in some plagioclase-bearing systems. Yb. Carnegie Instn. Wash. 69, 148-55. 1977. Anorthosite massifs, Rapakivi granites, and late Proterozoic rifting of North America. Precambrian Res. 7, 61-98. HAMILTON,E. I. 1964. The geochemistryofthe northern part of the Ilimaussaq intrusion, S. W. Greenland. Bull. Gronlands geol. Unders. 42 (also Meddr. Gronland, 162 (10)). HARRY, W. T. & PULVERTAFT, C. T. R. 1963. The Nunarssuit intrusive complex, South Greenland. Bull. Gronlands geol. Unders. 36 (also Meddr. Gronland, 169 (1)). HENRIKSEN, N. 1960. Structural analysis of a fault in South-West Greenland. Bull. Gronlands geol. Unders. 26 (also Meddr. Gronland, 162). JONES, A. P. 1980. The petrology and structure of the Motzfeldt Centre, Igaliko, South Greenland. PhD Thesis, University of Durham (unpublished). 1984. Mafic silicates from the nepheline syenites of the Motzfeldt centre, south Greenland. Mineral. Mag. 48, 1-12. & PECKETT,A. 1980. Zirconium-bearing aegirines from Motzfeldt, South Greenland. Contrib. Mineral. Petrol. 75, 251-5. KALSBEEK,F. & JEPSEN, H. J. 1984. The late Proterozoic Zig-Zag Dal basalt formation of eastern North Greenland. J. Petrol. 25, 644-64. KARUP-MOLLER, S. 1978. The ore minerals of the Ilimaussaq intrusion: their mode of occurrence and their conditions of formation. Bull. Gronlands geol. Unders. 127. KONNERUP-MADSEN, J. 1984. Compositions of fluid inclusions in granites and quartz syenites from the Gardar continental rift province (South Greenland). Bull. Mineral. 107, 327-40. - - , ROSE-HANSEN, J. & LARSEN, E. 1981. Hydrocarbon gases associated with alkaline igneous activity: evidence from compositions of fluid inclusions. In : J. C. BAILEY, L. M. LARSEN& H. SORENSEN(eds) The Ilimaussaq Intrusion, South Greenland. A Progress Report on Geology, Mineralogy, Geochemistry and Economic Geology, Rapp. Gronlands geol. Unders, 103, 99-108. KUNZENDORF, H., NYEGAARD, P. & NIELSEN, B. L. 1982. Distribution of characteristic elements in the radioactive rocks of the northern part of Kvanefjeld, Ilimaussaq intrusion, South Greenland. Rapp. Gronlands geol. Unders. 109. LARSEN, J. G. 1977. Petrology of the late lavas of the Eriksfjord Formation, Gardar Province, South Greenland. Bull Gronlands geol. Unders. 125. LARSEN, L. M. 1976. Clinopyroxenes and coexisting mafic minerals from the alkaline Ilimaussaq Intrusion, South Greenland. J. Petrol. 17, 258-90. 1977. Aenigmatites from the Ilimaussaq intrusion, south Greenland: Chemistry and petrological implications. Lithos, 10, 257-70. 1979. Distribution of REE and other trace -
-
-
-
-
469
-
-
470
-
-
B. G. J. Upton & C. H. Emeleus
elements between phenocrysts and peralkaline undersaturated magmas, exemplified by rocks from the Gardar igneous province, South Greenland. Lithos, 12, 303-15. & SORENSEN,H. 1987. The Ilimaussaq intrusion-progressive crystallization and formation of layering in an agpaitic magma. In: J. G. FITTON & B. G. J. UPTON (eds) Alkaline Igneous Rocks, Geol. Soc. Spec. Publ. 30, pp. 473-88. & STEENFELT, A. 1974. Alkali loss and retention in an iron-rich peralkaline phonolite dyke from the Gardar province, South Greenland. Lithos, 7, 81-90. LOISELLE, M. C. & WONES, D. R. 1979. Characteristics and origin of anorogenic granites. Geol. Soc. Am. Abstr. Progm. 11,468. MACDONALD, R., UPTON, B. G. J. & THOMAS, J. E. 1973. Potassium- and fluorine-rich hydrous phase coexisting with peralkaline granite in South Greenland. Earth planet. Sci. Lett. 18, 217-22. MARTIN, A. R. 1985. The evolution of the Tugtut6qIlimaussaq dyke swarm, southwest Greenland. PhD Thesis', University of Edinburgh (unpublished). MORSE, S. A. 1981. Kiglapait geochemistry IV: the major elements. Geochim. cosmochim. Acta, 45, 461-79. 1982. A partisan review of Proterozoic anorthosites. Am. Mineral. 67, 1087-100. PARSONS, I. 1979. The Klokken gabbro-syenite complex, South Greenland: cryptic variation and origin of inversely graded layering. J. Petrol. 20, 653-94. 1981. The Klokken gabbro-syenite complex, South Greenland: quantitative interpretation of mineral chemistry. J. Petrol. 22, 233-60. & BUTTERFIELD,A. W. 1981. Sedimentary features of the Nunarssuit and Klokken syenites, South Greenland. J. geol. Soc. Lond. 138, 289-306. PATCHETT, P. J., BYLUND, G. & UPTON, B. G. J. 1978. Palaeomagnetism and the Grenville orogeny: new Rb-Sr ages from dolerites in Canada and Greenland. Earth planet. Sei. Lett. 40, 349-64. PETERSILIE, I. A. & SORENSEN, H. 1970. Hydrocarbon gases and bituminous substances in rocks from the Ilimaussaq alkaline intrusion, South Greenland. Lithos, 3, 59-76. PIPER, J. D. A. 1982. The Precambrian palaeomagnetic record : the case for the Proterozoic supercontinent. Earth planet. Sci. Lett. 59, 61-89. POULSEN, V. 1964. The sandstones of the Precambrian Eriksfjord Formation in South Greenland. Rapp. Gronlands geol. Unders. 2. POWELL, M. 1978. The crystallization history of the Igdlerfigssalik nepheline syenite intrusion, Greenland. Lithos, l l , 99-120. ROCK, N. M. S. 1986. The nature and origin of ultramafic lamprophyres : aln6ites and allied rocks. J. Petrol. 27, 155-96. SORENSEN, H., ROSE-HANSEN, J., NEILSEN, B. L., LOvBORG, L., SORENSEN, E. & LUNDGAARD, T. 1974. The uranium deposit at Kvanefield, the Ilimaussaq intrusion, South Greenland. Geology, reserves and beneficiation. Rapp. Gronlands geol. Unders. 60. -
-
-
-
-
-
& PETERSEN,O. V. 1981. The mineralogy of the llimaussaq intrusion. In." J. C. BAILEY, L. M. LARSEN & H. SORENSEN (eds) The Ilimaussaq Intrusion, South Greenland. A Progress Report on Geology, Mineralogy, Geochemistry and Economic Geology, Rapp. Gronlands geol. Unders, 103, 19-24. SPARKS, R. S. J. & HUPPERT, H. E. 1984. Density changes during fractional crystallisation of basaltic magmas: fluid dynamic implications. Contrib. Mineral. Petrol. 85, 300-9. STEIGER, R. H. & JAGER, E. 1977. Subcommission on geochronology: Convention on the use of decay constants in geo- and cosmochronology. Earth planet. Sci. Lett. 36, 359. STEPHENSON, D. 1972. Alkali clinopyroxenes from nepheline syenites of the South Q6roq Centre, South Greenland. Lithos, 5, 187-201. -1974. Mn and Ca enriched olivines from nepheline syenites of the South Q6roq Centre, South Greenland. Mineral. Mag. 46, 283-300. 1976a. A simple-shear model for the ductile deformation of high-level intrusions in South Greenland. J. geol. Soc. Lond. 132, 307-18. -1976b. The South Q6roq nepheline syenites, South Greenland : petrology, felsic mineralogy and petrogenesis. Bull. Gronlands geol. Unders. 118. & UPTON, B. G. J. 1982. Ferromagnesian silicates in a differentiated alkaline complex: Kfingn~t Fjeld, South Greenland. Mineral. Mag. 46, 283300. STEWART,J. W. 1970. Precambrian alkaline-ultramafic/ carbonatite volcanism at Qagssiarssuk, South Greenland. Bull. Gronlands geol. Unders. 84 (also Meddr. Gronland, 186 (4)), STORMER,J. C. & CARMICHAEL,I. S. E. 1970. Villiaumite and the occurrence of fluoride minerals in igneous rocks. Am. Mineral. 55, 126-34. SUN, S. S. 1980. Lead isotope study of young volcanic rocks from mid-ocean ridges, ocean islands and island arcs. Phil. Trans. R. Soc. Lond., Ser. A, 297, 409-45. TAYLOR, S. R. CAMPBELL,I. H., MCCULLOCH,M. T. & MCLENAN, S. M. 1983. A lower crustal origin for massif-type anorthosites. Nature, Lond. 311, 3724. THOMPSON, R. N., MORRISON, M. A., DICKIN, A. P. & HENDRY, G. L. 1983. Continental flood basalts... arachnids rule OK? In." C. J. HAWKESWORTH & M. J. NORRY (eds) Continental Basalts and Mantle Xenoliths, pp. 158-85. Shiva, Orpington. TUKIAINEN,T., BRADSHAW,C. & EMELEUS,C. H. 1983. Geological and radiometric mapping of the Motzfeldt Centre of the Igaliko Complex, South Greenland. Rapp. Gronlands geol. Unders, 120, 78-83. UPTON, B. G. J. 1960. The alkaline complex of Kfingnfit Fjeld, South Greenland. Bull. Gronlands geol. Unders. 27 (also Meddr. Gronland, 123). 197 I. Melting experiments on chilled gabbros and syenogabbros. Yb. Carnegie Instn. Wash. 70, 1128. 1974. The alkaline province of South-West Greenland. In: H. SORENSEN (ed.) The Alkaline Rocks, pp. 221-38. Wiley, New York. -& BLUNDELL, D. J. 1978. The Gardar igneous -
-
-
-
Mid-Proterozoic magmatism in southern Greenland
--
-
-
- - - - ,
province: evidence for Proterozoic continental rifting. In. E. R. NEUMANN& I. B. RAMBERG(eds) Petrology and Geochemistry qf Continental Rt'fis, pp. 163 -72. Reidel, Dordrecht. & FIT'rON, J. G. 1985. Gardar dykes north of the Igaliko Syenite Complex, southern Greenland. Rapp. Gronlands geol. Unders. 127. & THOMAS, J. E. 1973. Precambrian potassic ultramafic rocks: South Greenland. J. Petrol. 14, 509-34. & -1980. The Tugtut6q younger giant dyke complex, South Greenland: fractional crystallisation of transitional olivine basalt magma. J. Petrol. 21,167-98. , HILL, P. G., JOHNSEN, O. & PETERSEN, O. V. 1978. Emeleusite: a new LiNaFe'" silicate from South Greenland. Mineral. Mag. 42, 31-4. MACDONALD, R., HILL, P. G., JEFFRIES, B. & FORD, C. E. 1976. Narsarsukite: a new occurrence w
471
in peralkaline trachyte, South Greenland. Mineral. Mag. 40, 737-46. ---, STEPHENSON, D. & MARTIN, A. R. 1985. The Tugtut6q older giant dyke complex: mineralogy and geochemistry of an alkali-gabbro-augitesyenite-foyaite association in the Gardar province of South Greenland. Mineral. Mag. 49, 623 -42. WALTON, B. J. 1965. Sanerutian appinitic rocks and Gardar dykes and diatremes north of Narssarssuaq, South Greenland. Bull. Gronlands geol. Unders. 57 (also Meddr. Gronland, 179 (9)). WATT, W. S. 1966. Chemical analyses for the Gardar igneous province, South Greenland. Rapp. Gronlands geol. Unders. 6. WIEBE, R. A. 1985. Proterozoic basalt dikes in the Nain anorthosite complex, Labrador. Can. J. Earth Sei. 22, 1149-57. WYLLIE, P. J. 1977. Effects of H20 and CO: on magma generation in the crust and mantle. J. geol. Soe. Lond. 134, 215-34.
B. G. J. UPTON, Grant Institute of Geology, University of Edinburgh, Edinburgh EH9 3JW, U.K. C. H. EMELEUS, Department of Geological Sciences, University of Durham, Durham DH1 3LE, U.K.
The llimaussaq intrusion progressive crystallization and formation of layering in an agpaitic magma* Lotte Melchior Larsen & Henning Sorensen S U M M A R Y : Agpaitic rocks form the major part of the Ilimaussaq intrusion in S Greenland. The agpaitic magma developed as the narrow top zone in a large stratified basalt-syenite magma chamber at depth. The extreme composition of the agpaitic magma is related to an unusually high crustal position of the cupola. After emplacement, the agpaitic magma developed in an essentially closed system. The magma chamber was shallow and the volatile-rich alkaline magma was light and fluid. Heat loss was mainly through the roof, and the earliest agpaitic rocks crystallized successively downwards from the roof. The magma was probably well mixed in the early stage, but there is evidence for accumulation of residual components in a layer below the roof. This accumulation of low-melting components eventually suppressed the downwards crystallization of the roof rocks. The exposed floor rocks, kakortokites and lujavrites, are younger than the roof rocks, and at this stage the magma had probably developed repeated layering. The layering in the kakortokites, with density-graded units 7 m thick repeated continuously over the whole exposed floor, can be simply explained if they formed from a layered magma by successive upwards crystallization of individual layers. The magma at this stage was nearly volatile saturated, and each layer crystallized in response to the upward loss of a certain amount of volatiles. The lujavrites conformably overlie the kakortokites and formed after a roof collapse which caused the rate of heat loss from the remaining magma to increase. The upward crystallization became faster than the upward transport of residual components, and the successive lujavrites contain more and more of these components which finally gave rise to potentially economic concentrations of U, Be and other rare elements. Finally, a hydrothermal phase was lost from the s,'stem.
Introduction The Ilimaussaq intrusion belongs to the latePrecambrian Gardar magmatic province in S Greenland (Upton 1974; Emeleus & Upton 1976) and has a radiometric age of 1168 Ma (Blaxland et al. 1976). It is in many ways unique: it consists of peralkaline nepheline syenites (agpaites) unusually enriched in rare and generally incompatible elements such as Zr, Nb, rare-earth elements, U, Th and halogens, and it is depleted to an extreme degree in such compatible elements as Mg, Ni, Cr, Sc and V. This unusual geochemistry has led to the development of a wide diversity of minerals, many of which are very rare or only found in Ilimaussaq; Sorensen et al. (1981) have listed 194 minerals and a few unnamed compounds. Rock types which have few or no counterparts in other parts of the world are present, and many rocks show spectacular layering and intrusive features. The intrusion was first mapped and described at length by Ussing (1912). Renewed activity after 1955 has resulted in the publication of many papers on the mineralogy, petrology, geochemistry and economic potential of the intrusion (Sorensen 1962, 1969; Ferguson 1964, 1970a; Petersilie & Sorensen 1970; Bohse et al. 1971 ; Engell * Contribution to the Mineralogy of Ilimaussaq No. 81.
1973; Sorensen et al. 1974; Steenfelt & Bohse 1975; Larsen 1976, 1977; Karup-Moller 1978; Konnerup-Madsen et al. 1979, 1981 ; Andersen et al. 1981a; Bailey et al. 1981c; Bohse & Andersen 1981). The series Contributions to the Mineralogy o f l l i m a u s s a q (Sorensen et al. 1981) includes more than 80 papers. An interim petrological synthesis of the intrusion has been given by Bailey et al. (1981b). The spectacular layering exhibited by many of the rocks of Ilimaussaq has led to much speculation as to its origin and the processes in the magma chamber (Upton (1961) and Ferguson & Pulvertaft (1963) in addition to the references given above). Considerable development in ideas on the processes in crystallizing magma chambers has taken place in recent years (e.g. Sparks et al. 1984), and in this paper the crystallization of the agpaitic magma is reviewed in the light of these proposals.
Geological setting Figure 1 shows a map of the intrusion which is elongated and measures some 8 km • 17 km. It is emplaced at a high level in the crust and cuts both the basement granite and the overlying Gardar sandstones and lavas. Remnants of the lava roof are preserved, but the original thickness
From: FITTON,J. G. & UPTON,B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 473-488.
473
474
L. M. Larsen & H. Sorensen 46-[ .~.':I'.IL./M.AUSSA.Q.I~II~III : : I : : : : : I : : : : : : I : I : : I I I I I : : : : : I 1.61.~
~ : : :: :: :: :: :: :: :: ::W~o: : ,-< : G E O L O G I C A L M A P OF T H E E L D " " - ~ o ~ - - % - ' , ~ \ i%o~ , ~ , o " , " " , . . . . . . . . . . . . . . . .' ::~:ii=:i::::::iiii~.' ' / 9 .~.)~=~ I L IMAUSSAQ INTRUSION 9. ~ - ~ ' K V A N E F J
9 ........
Superficial "K,L i"/'~" SOn SIOR x, deposits Fault Lujavrite M - C Lujavrite, black Lujavrite, green Kakortokite Naujaite ~Z,-~ Sodalite foyaite ~ Pulaskite, foyaite Alkali granite, quartz syenite Augite syenite Narssaq intrusion I , ~ Gardar supracrustals Basement granite
9
,
"NUNASARNAUSSAQ
I
0
2
4
km
FIG. 1. Generalized map of the Ilimaussaq intrusion. (After Ferguson 1964.) of the roofing lavas is uncertain; estimates are 25 km (Serensen 1969) and 3-4 km (Larsen 1974, p. 110). In all cases the lid on the magma chamber was thin in relation to the horizontal extent of the chamber. It would therefore be expected that material from the magma chamber had easy access to the surface where a phonolitic volcano like those in the E African rift system could have formed, but apparently this did n o t happen. The Ilimaussaq rocks bear evidence of retention to the very last stages of huge amounts of alkalis and volatiles which would have been released if there had been easy access to the surface. The agpaitic magma, which was of extreme composition from the beginning, developed in an essentially closed system which allowed its unique character to develop.
Structure of the Ilimaussaq intrusion The intrusion consists of a discontinuous marginal shell of augite syenite and an inner series of conformably layered agpaitic nepheline syenites
among which there are both roof cumulates (foyaite and naujaite) and floor cumulates (kakortokite). A 'sandwich horizon' (lujavrite) in the middle has intruded and brecciated the roof cumulates. The structure is shown in Fig. 2. The thickness of the exposed rocks is only 1500 m, so that the part of the magma chamber which we know, and which was the last to crystallize, was very flat and disc shaped. We do not know which rock types are present beneath the exposed rocks nor the position of the bottom of the intrusion. Indirect evidence of the character of the substratum in Ilimaussaq comes from geophysical data. Heat-flow measurements (Sass et al. 1972) have shown that the agpaitic rocks constitute a surface layer with a thickness not much greater than that of the exposed agpaitic rocks. A large positive gravity anomaly is centred in the Ilimaussaq area (Blundell 1978; Forsberg & Rasmussen 1978) where there is also a large magnetic anomaly (L. Thorning, personal communication). Forsberg & Rasmussen (1978) showed that threedimensional prism models gave the best fit to the gravity data when a heavy body of density at
The Ilimaussaq intrusion viscosity calculations. F and OH are also quite similar in size and mass, and it is possible that they may also be combined in density calculations. The calculations in the following sections were made with water as the only volatile though in reality this will comprise water+fluorine. Chlorine is too different to be included. Agpaitic magmas have a large capacity for dissolving volatile components which are probably present in the melt as alkali complexes (Kogarko 1974). Therefore, even though the concentrations of these components are high, their activities (fugacities) will be low. Thus the initial agpaitic magma of Ilimaussaq was H20 undersaturated and had a low oxygen fugacity buffered by a QFM-type equilibrium (Larsen 1976). Two separate fluid phases exsolved during the crystallization: one was rich in H20 and the other was rich in hydrocarbons (KonnerupMadsen et al. 1979, 1981), while the content of carbon oxide species was very low, as also was the sulphur fugacity. Native tin and lead formed in the late stages of crystallization (Karup-Moller 1978). Kogarko et al. (1977) and Kogarko & Romanchev (1983) noted that agpaitic magmas displaying the 'agpaitic order of crystallization' have partial water pressures of less than 200 bars and, according to Kogarko et al.'s (1977) experiments, therefore have maximum water contents of 1.5% at 1000~ However, the Ilimaussaq agpaitic magma was multiply saturated (anchieutectic), and all major mineral phases appeared more or less simultaneou,:ly on the liquidus. This allows higher water vapour pressures (Kogarko et al. 1977). The same experiments showed a maximum water content of 4.3% in an agpaitic magma at 1 kbar and 1000~ and this is not in conflict with our estimate of maximum 4% water + fluorine for the Ilimaussaq magma which was under a confining pressure of around 1 kbar.
477 500
I000
Ilirnaussaq
roof
rocks sodalite nepheline
s o I i d
feldspar
U S
~
--
~....
Arf--
-Ac ......... Astr ~lNept
i
~ eudialyte / - Fa olivine --Kat
Aeg
? --
amphibole
Y'
\~As Hed
Aen .
.
.
.
Mgt--Ti-
I
i
J
1
I000 bars
-20
pyroxene
Hed
phase
estimated trapping "# of hydrocarbons
......... -I0
~.
i
j
I
QFM
........
._# g~ _d
30
/~" /
/~calculated
log fo ,T
/" late hydrothermal veins (b)
300
500 Temperature.
IOOO ~
FIG. 4. (a) Sequence of crystallization of minerals in the Ilimaussaq intrusion compared with (b) the calculated (logfO2, T) conditions for possible equilibration of Ilimaussaq gases: Ac, acmite; Aeg, aegirine; Aen, aenigmatite; Arf, arfvedsonite; Astr, astrophyllite; Fa, fayalite; Hed, hedenbergite; Kat, kataphorite; Mgt, magnetite; Nept, neptunite. Data compiled from Piotrowski & Edgar (1970), Larsen (1976, 1977), Karup-Moller (1978) and KonnerupMadsen et al. (1979, 198 I). (From Konnerup-Madsen et al. 1981.)
Temperatures The temperature of the initial agpaitic magma was estimated to be 850-900~ by Serensen (1969) and Larsen (1976). The olivine equilibration method of Ford et al. (1983) gives a temperature of 950~ for the phonolite dyke (Table 1); this is a maximum estimate if any volatiles and alkalis have been lost. A temperature of 900~ was used for the density and viscosity calculations discussed below. Solidus temperatures are around 450~ according to melting experiments by Piotrowski & Edgar (1970) and Sood & Edgar (1970). The agpaitic rocks thus had unusually large crystallization intervals (Fig. 4).
Density Magma densities are crucially dependent on volatile content. The density of the composition in Table 1, except for the halogens, was calculated using the method of Bottinga et al. (1982), with H20 incorporated as in the method of Bottinga & Weill (1970). At 900~ and a pressure of 1 kb the density is 2.41 g cm -3. Addition of water up to 4.20 wt.% decreases the density to 2.29 g cm - 3, i.e. the density of sodalite. When naujaite formed, sodalite usually--but not always-floated in the magma which must have had a density close to 2.29gcm -3. Although the
476
L. M. Larsen & H. Sorensen
amounts of crystallization because the residual magmas formed at any time tend to keep physically separate and collect upwards (Turner & Gustafson 1981 ; Sparks et al. 1984). The deepseated magma chamber where the development took place must have been very large and also of appreciable vertical extent, and thus must have been situated in a considerable temperature gradient. It is feasible that this magma became compositionally zoned and fractionated by sidewall crystallization and upward migration of the residual liquids. If the magma body were broad based and tapered upwards, the narrow cupola at the top would be a very efficient trap for the light residual liquids. The agpaitic magma could thus develop concomitantly with, and as an effect of, the formation of the heavy magnetic rocks which underlie the intrusion. The composition of the magma at a given height would be a function of the temperature there, and the higher in the crust the top of the magma chamber was situated the more differentiated would be the magma. When the top zone was not too high in the crust, augite syenitic magma would form, and could be intruded into the higher crust during tectonic disturbance. Possibly as a result of the same event, the cupola on the deep magma chamber was extended further upwards into lower-temperature regions in the crust where more extensively differentiated magma started to develop. The extreme composition of the agpaitic magma is due to an unusually high position of the cupola in the crust. The possibility that the intrusion itself is this cupola and that the intrusion is 'bottomless' is unlikely as the agpaites are clearly intrusive into the earlier rocks (Fig. 3). However, because the dense mafic rocks appear to be situated at depths of only 2-5 km (Forsberg & Rasmussen 1978), the subsided central block or fragments cannot be thicker than this. The known agpaitic rocks have a volume of about 200 km 3 which would involve some 10 000 km 3 of basalt. This volume can reasonably be accommodated beneath the Ilimaussaq intrusion and its surroundings.
The properties of the agpaitic magma Composition The earliest peralkaline syenites of Ilimaussaq have such variable compositions that no quantitative estimates of the composition of the initial agpaitic magma can be obtained from them. However, the country rocks around the southern
part of the intrusion contain an extensive swarm of trachytic to phonolitic dykes (Allaart 1969), and each dyke has a well-defined composition. One of these dykes is a eudialyte-bearing agpaitic phonolite with many petrographical and chemical similarities to the Ilimaussaq rocks (Larsen & Steenfelt 1974; Larsen 1979). Its composition is given in Table 1, and we consider that the composition of the initial agpaitic magma of Ilimaussaq was not very different from this. The halogen contents of the rocks in the intrusion (Gerassimovsky & Kuznetsova 1967; Ferguson 1970b) are, however, very different from those of the dyke (Table 1).
Volatile contents and fugacities The dyke represented in Table 1 has in all probability lost volatiles and perhaps also alkalis, and the values for these elements in Table 1 are consequently minimum values. Density considerations given later indicate that the total of water+fluorine content was around 4wt.%. Recent experiments by Dingwell et al. (1985) have shown that fluorine is analogous to water in its effect on magma polymerization state, and that fluorine and water may be combined in
TABLE 1. Chemical composition of a dyke believed to be similar to the initial agpaitic magma of the Ilimaussaq intrusion wt.%
1
2
SiO2 TiO2 ZrOz A1203 Fe203 FeO MnO MgO CaO Na2 K20 P_,Os H_,O F C1
51.83 0.55 0.57 14.57 7.56 4.61 0.48 0.14 2.54 8.81 4.87 0.08 2.03 0.84 O.33
49.79 0.33 0.72 18.72 5.02 3.58 0.23 0.22 1.40 12.99 3.67 0.07 1.97 0.25 1.38
Sum
99.81
100.34
0.43
0.42
99.38
99.92
O = F, C1 Total
1, GGU 42475, Gardar dyke SE of the Illmaussaq intrusion (Larsen 1979); 2, average composition of agpaitic nepheline syenites of the Ilimaussaq intrusion (Ferguson 1970b).
The Ilimaussaq intrusion viscosity calculations. F and OH are also quite similar in size and mass, and it is possible that they may also be combined in density calculations. The calculations in the following sections were made with water as the only volatile though in reality this will comprise water+fluorine. Chlorine is too different to be included. Agpaitic magmas have a large capacity for dissolving volatile components which are probably present in the melt as alkali complexes (Kogarko 1974). Therefore, even though the concentrations of these components are high, their activities (fugacities) will be low. Thus the initial agpaitic magma of Ilimaussaq was H20 undersaturated and had a low oxygen fugacity buffered by a QFM-type equilibrium (Larsen 1976). Two separate fluid phases exsolved during the crystallization: one was rich in H20 and the other was rich in hydrocarbons (KonnerupMadsen et al. 1979, 1981), while the content of carbon oxide species was very low, as also was the sulphur fugacity. Native tin and lead formed in the late stages of crystallization (Karup-Moller 1978). Kogarko et al. (1977) and Kogarko & Romanchev (1983) noted that agpaitic magmas displaying the 'agpaitic order of crystallization' have partial water pressures of less than 200 bars and, according to Kogarko et al.'s (1977) experiments, therefore have maximum water contents of 1.5% at 1000~ However, the Ilimaussaq agpaitic magma was multiply saturated (anchieutectic), and all major mineral phases appeared more or less simultaneou,:ly on the liquidus. This allows higher water vapour pressures (Kogarko et al. 1977). The same experiments showed a maximum water content of 4.3% in an agpaitic magma at 1 kbar and 1000~ and this is not in conflict with our estimate of maximum 4% water + fluorine for the Ilimaussaq magma which was under a confining pressure of around 1 kbar.
477 500
I000
Ilirnaussaq
roof
rocks sodalite nepheline
s o I i d
feldspar
U S
~
--
~....
Arf--
-Ac ......... Astr ~lNept
i
~ eudialyte / - Fa olivine --Kat
Aeg
? --
amphibole
Y'
\~As Hed
Aen .
.
.
.
Mgt--Ti-
I
i
J
1
I000 bars
-20
pyroxene
Hed
phase
estimated trapping "# of hydrocarbons
......... -I0
~.
i
j
I
QFM
........
._# g~ _d
30
/~" /
/~calculated
log fo ,T
/" late hydrothermal veins (b)
300
500 Temperature.
IOOO ~
FIG. 4. (a) Sequence of crystallization of minerals in the Ilimaussaq intrusion compared with (b) the calculated (logfO2, T) conditions for possible equilibration of Ilimaussaq gases: Ac, acmite; Aeg, aegirine; Aen, aenigmatite; Arf, arfvedsonite; Astr, astrophyllite; Fa, fayalite; Hed, hedenbergite; Kat, kataphorite; Mgt, magnetite; Nept, neptunite. Data compiled from Piotrowski & Edgar (1970), Larsen (1976, 1977), Karup-Moller (1978) and KonnerupMadsen et al. (1979, 198 I). (From Konnerup-Madsen et al. 1981.)
Temperatures The temperature of the initial agpaitic magma was estimated to be 850-900~ by Serensen (1969) and Larsen (1976). The olivine equilibration method of Ford et al. (1983) gives a temperature of 950~ for the phonolite dyke (Table 1); this is a maximum estimate if any volatiles and alkalis have been lost. A temperature of 900~ was used for the density and viscosity calculations discussed below. Solidus temperatures are around 450~ according to melting experiments by Piotrowski & Edgar (1970) and Sood & Edgar (1970). The agpaitic rocks thus had unusually large crystallization intervals (Fig. 4).
Density Magma densities are crucially dependent on volatile content. The density of the composition in Table 1, except for the halogens, was calculated using the method of Bottinga et al. (1982), with H20 incorporated as in the method of Bottinga & Weill (1970). At 900~ and a pressure of 1 kb the density is 2.41 g cm -3. Addition of water up to 4.20 wt.% decreases the density to 2.29 g cm - 3, i.e. the density of sodalite. When naujaite formed, sodalite usually--but not always-floated in the magma which must have had a density close to 2.29gcm -3. Although the
478
L. M. Larsen & H. Sorensen
magma composition at that time was modified somewhat relative to the initial composition, the content of water and fluorine could not have deviated much from the 4 wt.% level. The naujaite magma was also rich in C1, but this is much denser than H20 and F.
Viscosity Like the density, the viscosity is strongly volatile dependent. Viscosity calculations after Shaw (1972) yield a viscosity of 8980 poise at 900~ for the composition in Table 1, and this is decreased to 1390 poise when the water content is increased to 4 wt.% (chlorine is not considered, while fluorine is included with the water (Dingwell et al. 1985)). The ratio of non-bridging oxygens to tetrahedrally co-ordinated cations (NBO/T) is decisive for the viscosity of a magma (e.g. Mysen et al. 1982). For the dry composition in Table 1, NBO/T= 0.51, and for the volatile-rich Ilimaussaq magma NBO/T was probably at least 0.6. These figures for viscosity and liquid structure are in the range of those for basalt (Shaw 1969; Basaltic Volcanism Study Project 1981, p. 703; Mysen et al. 1982), and the agpaitic magma was at least as fluid as a basalt despite the low temperature. This is also indicated by dykes of lujavrite which, though centimetre thin, can be followed for hundreds of metres.
Crystallization of the agpaitic magma The classical picture of a crystallizing magma chamber involves a convecting homogeneous magma in which crystallization takes place along the roof and walls of the chamber, and crystals are sedimented on the floor. Layering is due to differential crystal settling or sorting by convection currents (Wager & Brown 1967; Kogarko et al. 1982). This simple picture has been considerably modified in recent years because it has become increasingly evident that magmas are not homogeneous (see the review by Sparks et al. 1984). Laboratory experiments have shown that liquids in a temperature gradient, like a magma in a chamber, will develop concentration gradients and become compositionally stratified (e.g. Chen & Turner 1980; Turner 1980). Natural examples of compositionally stratified magma bodies have been identified by their eruption products, both in rhyolitic systems (Hildreth 1979) and in undersaturated phonolitic systems (Wolff & Storey 1984; W6rner & Schmincke 1984). In the magma chambers themselves, now
represented by plutonic rock complexes, various workers have interpreted layered rocks as formed from layered magmas (Bussen & Sakharov 1967; McBirney & Noyes 1979; Irvine et al. 1983; Wilson & Larsen 1985). As yet, there are no criteria which can be used as infallible proof that a layered rock formed from a layered magma (Sparks et al. 1984, p. 533). However, as stressed by these workers, layering seems to be inevitable in magma chambers. The application of these ideas to the agpaitic magma chamber of Ilimaussaq immediately helps to resolve an old controversy, namely whether the magma was convecting and thereby homogeneous (Bohse et al. 1971; Engell 1973) or stagnant and with compositional gradients (Sorensen 1962, 1969; Ferguson 1964, 1970a). There is evidence in the intrusion for both current action and upward concentration of volatiles and alkalis, and this is conceivable if the magma was layered. There would then be compositional gradients between the layers, and each layer would have convected separately by doublediffusive convection which would have been driven by the material and heat flux through the system and served to keep the composition of each layer constant. If the agpaitic magma was layered at some stage, the number and thickness of the layers are open to speculation. On the one hand it is tempting to visualize each unit in a layered rock as formed from one magma layer, in which case there were many layers with a thickness of 10 m or less. On the other hand there may have been only a few thick layers; Irvine (1983) suggested that the Skaergaard magma chamber had three layers of which the middle one comprised most of the chamber volume. All intermediate situations are possible, and the configuration would probably also change with time. What is certain is that the major part of the heat loss during the crystallization of the exposed rocks of Ilimaussaq took place through the roof and to a lesser extent through the walls of the flat chamber (Fig. 2). In experiments both these methods of cooling lead to the development of repeated double-diffusive layering in the magma (Chen & Turner 1980), and it is very probable that the agpaitic Ilimaussaq magma also developed layering. It was light and fluid with a high volatile content and a very long crystallization interval, and had all possibilities of developing density and temperature gradients. However, the situation is not so simple that the conformably layered Ilimaussaq rocks can be interpreted as having formed in a correspondingly layered magma chamber. This would necessitate a regular sequence from higher-temperature less evolved
The Ilimaussaq intrusion rocks at the bottom to lower-temperature more evolved rocks at the top of the intrusion, and this does not occur. In the following sections the crystallization of the agpaitic magma of Ilimaussaq is developed with special emphasis on evidence for gradients and layering in the magma and on any connection between this and the layering in the rocks.
The roof rocks The roof rocks are the earliest agpaitic rocks and have crystallized downwards from the roof as a result of the high rate of heat loss under the thin roof. The sequence consists of comformable layers of, in descending order, pulaskite, foyaite, sodalite foyaite and naujaite (Fig. 2).
Mineralogy All the roof rocks have a water-free high temperature liquidus mineral assemblage of alkali feldspar, fayalite, hedenbergite, titanomagnetite and apatite, with additional nepheline and sodalite joining in the sodalite foyaite and naujaite. Their estimated liquidus temperatures range from 900~ to perhaps 800~ (Fig. 4). The mafic liquidus minerals are common just below the roof and become less abundant downwards in the sequence, while their Fe/Mg ratios increase (Larsen 1976). At a certain stage in the crystalli-
FIG. 5. Naujaite texture: grey sodalite crystals poikilitically enclosed in white alkali feldspar, dark grey eudialyte and black arfvedsonite and aegirine.
479
zation of a rock the high-temperature mafic minerals reacted out and were replaced by an alkaline hydrous assemblage of alkali amphibole, aegirine, aenigmatite and eudialyte (Fig. 4). This stage may be equivalent to that at which a separate fluid phase exsolved around 700~ (Fig. 4). This happened late in the crystallization history of the earliest rocks, and consecutively earlier during the formation of the roof sequence. The naujaite is considered to be a flotation cumulate (Ussing 1912; Ferguson 1964, 1970a; Sorensen 1969; Engell 1973). It consists of around 40% of sodalite dodecahedra or slender hexagonal prisms of dimensions 2-3 mm poikilitically enclosed in centimetre- to decimetre-sized crystals of feldspar, nepheline and the low-temperature mafic minerals mentioned above (Fig. 5). The amount of high-temperature mafic minerals in the naujaite is very small. The height of the magma column which contributed sodalite to the naujaite must have been considerable, and perhaps the whole magma was involved.
Layering The most extensive layering in the roof rocks is the four-member sequence of different rock types itself. However, the possibility that these crystallized from separate magma layers can be discounted because, as described in the preceding section, the roof rocks formed from successively more differentiated, more volatile-rich and slightly cooler magmas which would not have been stable on top of each other. Rather, the development of the roof rocks suggests a situation with a large well-mixed reservoir which was slowly being modified and cooled. Evidence for convection in such a reservoir should be sought in the contemporaneous bottom cumulates which are not exposed. The roof rocks present many features of modal layering, and more or less distinct horizons rich in sodalite, mafic minerals and eudialyte may be followed for short distances. The roof rocks contain evidence for concentration of residual components, mainly volatiles, immediately under the crystallizing roof in a manner similar to experimental results reported by Chen & Turner (1980). This evidence comes from extensive conformable pegmatite layers. The foyaite is the earliest rock with such pegmatite layers. It has developed a sequence 20 m thick of layers 1 m thick with alternating coarse and pegmatitic grain-size (Fig. 6). This can be explained by repetitive accumulation of volatiles under the roof (Ferguson & Pulvertaft 1963; Ferguson 1964). The high-temperature mineralogy of the layers is the same irrespective of the grain-size, and this shows that the volatile
480
L. M. Larsen & H. Sorensen
FIG. 6. Metre-scale repeated layering in foyaite, southern Ilimaussaq. Each layer is pegmatitic in the upper part, with crystals grown perpendicularly to the contact, and grades downwards into 'normal' coarse-grained foyaite. (Photograph by J. Ferguson.) accumulations were not sufficiently large to saturate the magma. Conformable pegmatite layers are particularly well developed in the naujaite (Fig. 7). Each pegmatite layer is around 0.5 m thick and the spacing between the layers is 10-30 m. They were described by Ussing (1912) and Sorensen (1962). In contrast with the surrounding rock, the layers contain no cumulus sodalite crystals. Sodalite nucleated on the sharp upper margin of the pegmatite layer and grew perpendicularly downwards in large prismatic crystals, and the other
FIG. 7. Naujaite cliffs S of Tunugdliarfik fjord. The near-horizontal massive naujaite benches are separated by conformable pegmatite layers 0.5 m thick that weather out as light gravel. (Photograph by J. Ferguson.)
minerals (feldspar, nepheline, arfvedsonite, eudialyte, aenigmatite) are thus not poikilitic. This structure can be explained if the density of the uppermost magma layer had dropped below that of sodalite, i.e. around 2.29 g c m - 3 The upwardmigrating sodalite crystals would then stop at this density level and leave the uppermost magma layer free of cumulus sodalite except for the perpendicular crystals grown there. The density decrease of the uppermost magma layer was most probably caused by accumulation of volatiles under the roof, as in the foyaite. The volatiles would also depress the liquidus temperature, and the layer would stay liquid while crystallization proceeded underneath it at higher temperatures. As the downward crystallization continued, the rate of heat loss through the roof diminished and the crystallization rate decreased. Volatile accumulation upwards continued, and consequently the amount of pegmatites increases downwards in the naujaite. Eventually the liquidus temperature of the magma under the roof was constantly depressed below the actual temperature there, and the downwards crystallization stopped. The bottom of the naujaite was left exposed to a very reactive magma for a long time, which explains the heavy alteration of some of the lowest naujaites. In conclusion, the situation that is envisaged for the upper part of the agpaitic magma chamber is a competition between two directionally opposed processes: upward accumulation of residual components like volatiles and volatiledependent elements, and downward-proceeding crystallization. Because the processes were opposed, the development in compositional gradient was only moderate and the most developed
The Ilimaussaq intrusion material was continuously caught up in crystallization or cut off in isolated layers. There were probably gradients in a relatively-thin upper layer, and a large convecting, slowly-cooling magma reservoir below. This magma would also have been involved in crystallization along the walls and bottom of the chamber, and the residual magma and sodalite crystals produced thereby would have moved upwards and joined the residual magma expelled from the crystallizing roof. Discontinuous screens of naujaite in the border pegmatite along the marginal contacts of the bottom rocks may represent remnants of such marginally crystallized naujaite. Whether any chemical diffusion was involved in the process (the 'thermogravitational diffusion' of Hildreth (1979)) or not (the 'convective fractionation' of Rice (1981) and Sparks et al. (1984)) is undetermined.
The bottom rocks As noted before, the rocks which presumably formed at the bottom of the magma chamber while the roof rocks were forming at the top are not exposed. They are inferred to be hightemperature hedenbergite-fayalite syenites. The lowermost exposed rocks, the kakortokites, are younger than the naujaites. They contain no hightemperature mafic minerals and formed from a magma that was cooler than that of the roof rocks (see below). Further, it has been shown that a large inclusion of naujaite in the middle of the kakortokite series fell from the lowermost part of the naujaite (Steenfelt & Bohse 1975). The naujaite series is therefore older than at least the
481
upper half of the kakortokite series, and it must be imagined that the kakortokites formed at the bottom while volatiles were continuously concentrating in the upper stagnant part of the magma chamber where no crystallization took place.
Mineralogy and layering The kakortokite magma behaved as an anchieutectic system having feldspar, nepheline, eudialyte and arfvedsonite on the liquidus (Sorensen 1969; Kogarko & Romanchev 1983). Liquidus temperatures are estimated at 700-800~ on the basis of stability data for arfvedsonite (Bailey 1969) and eudialyte (Kogarko et al. 1982). Some of the minerals contain fluid inclusions of both aqueous and hydrocarbon-rich types, similar to those found in the roof rocks (J. KonnerupMadsen, personal communication), and the kakortokite magma must have been nearly volatile saturated (Fig. 4). The visible part of the magma chamber was 300-600 m high at the kakortokite stage and, owing to the volatile enrichment, crystallization was probably depressed in the whole main magma body except in a bottom layer, as discussed below. The kakortokite series is about 280 m thick. The lowest 210 m section is spectacularly layered (Fig. 8). The rocks and the layering have been described by Ussing (1912), Upton (1961), Ferguson & Pulvertaft (1963), Ferguson (1964, 1970a), Bohse et al. ( 1971) and Bohse & Andersen (1981). The sequence comprises 29 exposed layered units numbered - 11 to + 17 (Bohse et al. 1971) which are 7 m thick on average. Each unit consists of a lower black layer rich in arfvedsonite, a red layer rich in eudialyte and an upper white layer--by far the thickest--rich in feldspar and
FIG. 8. Layered kakortokites, S coast of Kangerdluarssuk. Note the large roof inclusion wrapped by the layering in the centre of the picture. The jagged mountain range of Kidtlav~t shaped from basement granite along the contact to the intrusion can be seen in the background.
482
L. M . L a r s e n & H. Sorensen
The numerous roof blocks in unit +3 are interpreted as originating from one roof-collapse event, and each layered unit thus must have formed simultaneously over the whole floor.
Theories of formation of kakortokite layering
FIG. 9. Close-up of kakortokite layering, with large inclusion to the left.
nepheline (Fig. 9). The contacts between the black layers and the underlying white layers are sharp; other contacts are gradational. The red layers may be suppressed and in a few cases are absent. Each unit is thus density stratified and is usually well laminated; lineation was described by Ferguson (1970a). Individual units extend over the whole exposed floor of the intrusion with constant thickness and structure. The layers dip around 10~ towards the centre of the intrusion; close to the margins the layering steepens considerably and becomes indistinct and more fine scale. This 'saucer-shaped' structure is believed to be primary, although accentuated by late-stage sagging and compaction. The fact that one of the units ( - 4) has slumped (Bohse et al. 1971) shows that the original floor was indeed sloping. Apparent signs of current activity are found in the lower part of the kakortokite pile where trough banding and current bedding features occur in the steepened fine-scale layered rocks near the margins (S~rensen 1969; Bohse et al. 1971). These features die out upwards and only reappear in unit + 4 where they occur both in the steepened marginal rocks and in the steepened layers draped over the numerous large roof inclusions bulging up from the underlying unit + 3 (Bohse et al. 1971). Irvine (1983) attributed similar current structures near inclusions in the Skaergaard intrusion to eddies set up in larger currents when they flow past obstacles. It thus appears that the kakortokite magma was first convecting and later stagnant, except after the fall of the roof blocks into unit +3. It is theoretically possible, however, that there was convection all the time but that tracks were only left in special circumstances.
Several theories and discussions of the origin of the layering have been published (S~rensen 1969). A single black-red-white triple unit can be interpreted as formed by gravitational separation of crystals in the very fluid magma, and this mechanism has been proposed by almost all students of the kakortokites. Arguments against crystal settling have been advanced in recent years (McBirney & Noyes 1979) but, as stated by Sparks et al. (1984, p. 519), 'crystal settling is only likely to become a dominant factor in thin fluid layers of low viscosity magma'. The kakortokite magma would serve well as such a case (S~rensen 1969). The real interpretational problem in the kakortokite series has been the mechanism causing the recurrence of the layered units. Ussing (1912), Ferguson (1964, 1970a) and Sorensen (1969) favoured intermittent crystallization in a largely stagnant magma, while Upton (1961) and Bohse et aL (1971) suggested that the layering was produced by periodic convective overturns of the magma. However, if the magma did not convect during the formation of the upper part of the kakortokite pile, then convection cannot have played any role in the formation of the layering. This can also be argued from the flatness of the magma chamber: with an aspect ratio of 8-50 (Bailey et al. 1981b) for the chamber, even topto-bottom convection must have been broken up into several individual convection cells, and these are unlikely to produce floor deposits which are continuous over more than 4 km, as is the case for the kakortokite units. The researchers who have favoured intermittent crystallization have referred to a mechanism whereby a release in volatile pressure has shifted the position of a eutectic or changed the nucleation rates. However, the experiments by Yoder (1954), referred to by Ferguson (1964, 1970a), and also those of Edgar (1964) and Yoder (1965) indicate that volatile pressure release should stabilize feldspar relative to pyroxene and thus produce inverted layering. Moreover, release in volatile pressure was invoked by Parsons (1979) and Parsons & Butterfield (1981) to account for the inversely graded layering in the Klokken syenite by changing the nucleation rates. In any case pressure release appears to lead to inverted layering.
48 3
The Ilimaussaq intrusion
interface between the two layers would thereby be obliterated, and the rocks produced from a magma layer would be thinner than the magma layer. The next magma layer would start crystallizing when the appropriate amount of volatiles had been carried away from it into the overlying layer. The composition of the crystalline material was thus buffered by the volatiles, and this explains the almost constant composition of the kakortokites. The phase assemblage and majorelement phase compositions are constant up through the sequence, and even many traceelement contents and ratios are constant (Fig. 10) (H. Bohse, personal communication). If the succeeding magma layers had decreasing temperatures, systematic changes in the trace-element composition of the phases should be detectable. A systematic increase in the uranium content of the eudialyte with height in the kakortokite sequence has been described by Steenfelt & Bohse (1975), and the Mg content (a minor element) in the arfvedsonite appears to decrease with height (Larsen 1976). The concept of a layered magma, with each layer crystallizing successively as the necessary amount of heat and volatiles leaves it, provides a simple explanation of the recurrence of the layering and the almost-constant composition of the sequence. There are difficulties, however, in reconciling the sloping layers on the floor with a horizontally layered magma, and also in explaining the draping of the kakortokite units over the large
In the light of recent experiments with layered liquids it is tempting to view the kakortokites as products of a layered magma. If the magma was divided into double-diffusive layers, temperature and composition (density) profiles would be stepped. Heat would be transported upwards in the system by convection and diffusion across layer boundaries, and so probably would volatile components, especially if they were exsolved from some stage in one or two separate phases. The bottom layer, however, lost heat and material to the layer above without receiving anything from below, and consequently its liquidus temperature rose to meet the actual magma temperature. The liquidus minerals arfvedsonite, eudialyte, feldspar and nepheline nucleated throughout the layer and settled to the floor. Mafic minerals nucleate at smaller degrees of undercooling than feldspar does (Carmichael et al. 1974), and arfvedsonite and eudialyte could grow and start settling before feldspar and nepheline. This would increase the density sorting. The same differences in nucleation rate were used by Parsons (1979) to explain the inversely graded layers in the Klokken syenite as a result of gradual volatile increase. In Ilimaussaq the normal grading within a unit is due to gradual volatile loss, while the recurrence is due to the sudden volatile increase represented by the switch of crystallization to the next magma layer. The residual magma in the crystallizing layer would be volatile rich and light and would mix with the overlying magma layer. The original
M-C
lujavrite
9
~
'n?-T
n n
9
9
.
__
.800
-
-700 :-
orfvedsonite lujavrite n
with nauiakosite
.
600 0 0-
aegirine lujavrite II'~ a--Tt%eds~] av r~ fe j aegirine
lujavrite
transition
zone
-500 o
i "
._
-400 bo"
oegirine
lujavrite
layered
9 black
~,red
~
T
,r
iij
~
"transit~l layered,, k a k o r t o k i t e s ~ g h t 17~_layered k o k o r t o k l t e
lower
_
II
~ ~ .
kakortokite
c
.
| .
9
~white
9
300 ~
/ .
"0
.
-200 o
9
J
.100
~""
~,* v
I
l
~ ] I I I [[ 10
I
I
Z r//U
I
I
I
I Ill
100
I
|
~
I
l
,ell
1000
I
,
!
I 10
9J
Z r/Y
L
i
0
m
50
FIG. 10. Zr/U and Zr/Y stratigraphy in the kakortokite-lujavrite sequence, southern Ilimaussaq : Q, black kakortokites; ~, red kakortokites; O, white kakortokites; I , lujavrites (vertical bars indicate uncertainty in stratigraphic height); -~, analyses of separated eudialyte (e) and steenstrupine (s). (From Andersen et al. 1981a.)
484
L. M. Larsen & H. Sorensen
inclusions (Fig. 8). This would involve a good deal of layer sagging and compaction. The unconsolidated crystal mush was about 20 m thick (Bohse et al. 1971) and must have contained a fair amount of residual liquid which could later have been squeezed out by the weight of the overlying rocks. This liquid could have formed the pegmatitic veins in the marginal pegmatite which is thickest in its lower parts (Fig. 2) (Bohse & Andersen 1981). It is also possible to argue that the successive kakortokite units formed from one or a few thick magma layers by the process described above. In this case the residual liquid would stay in the thick crystallizing magma layer and only gradually be transferred to the overlying layer. An oscillation of the liquidus temperature around the magma temperature as a function of the volatile content of the magma and the heat released during crystallization will have to be postulated, and the explanation of the recurrence of the layering loses some of its simplicity. This traditional and slightly more complicated process was favoured by S~rensen (1969).
The final stages: the lujavrites The transition from the kakortokites to the lujavrites was continuous, and faintly but distinctly layered lujavrites still accumulated on the bottom, on top of the kakortokites (Bohse &
Andersen 1981). Although the upper lujavrites appear to be intrusive there was no new intrusion of magma. The lujavrite magma developed in situ under the naujaite roof by accumulation of the volatiles and associated low-melting components that were transported upwards through the underlying magma as described in the preceding sections. The lujavrites are relatively fine-grained laminated rocks. They consist of microcline and albite (in separate crystals, showing that the temperature had now decreased below that of the alkali feldspar solvus around 660~ nepheline, eudialyte, aegirine and arfvedsonite in a matrix of natrolite, analcime and sodalite. The lower lujavrites are dominated by aegirine (green lujavrite (Figs 1 and 2)) while the upper lujavrites are dominated by arfvedsonite (black lujavrite (Figs 1 and 2)). Later, intrusive lujavrites contain increasing amounts of alkaline minerals like naujakasite (Na6FeAI4SisO26), steenstrupine (a complex N a - R E E - T h , U - S i phosphate) and villiaumite (NaF), while numerous veins in the lujavrites contain analcime, ussingite (Na2 OH A1Si308), chkalovite (Na2BeSi206) and other rare minerals. The lujavrites and the veins obviously formed from material very rich in Na, H20 and F, and also enriched in a number of rare elements such as Th, U, Be and Sn which form complexes with alkalis and volatiles. At the beginning of the lujavrite stage there was apparently a physical change: the roof collapsed. The magma chamber at that time was
FIG. 11. Dark lujavrite has intruded into overlying solid naujaite (light) and spalled offlarge blocks which have sunk into the lujavrite (N coast of Tunugdliarfik; height of cliff section, 600 m.)
The Ilimaussaq intrusion very flat, only around 200 m high, and two or more sub-chambers were established by the collapse. The magma brecciated the roof and was forced into fractures there. It is possible that there was surface activity associated with the event and that some material was lost from the system. The transition from arfvedsonite-dominated kakortokite to aegirine-dominated green lujavrite could indicate volatile loss and oxidation due to this loss. However, the system was rapidly closed again, and the lujavrites show no signs of decreased contents of alkalis and volatiles. Figure 11 shows the lujavrite spalling off large blocks from the naujaite roof. By the brecciation of the roof, fresh naujaite was exposed to the magma, and thus both very fresh and very altered naujaite inclusions are found together in the lujavrite. The altered naujaite blocks are the remains of the lower surface of the naujaite roof which was exposed to the magma for a long time. The roof collapse resulted in an increase in the rate of heat loss from the system, and the balance between the rate of heat loss and the upward migration of volatiles that had pertained during the formation of the kakortokites was disrupted. The almost constant composition of the kakortokites gave way to a rapid compositional development up through the lujavrite sequence which shows systematic changes with height of several element ratios (Fig. 10). While the crystallization was still taking place from the bottom upwards, the crystallization rate increased, with the result that more and more of the low-melting components were incorporated into the crystallizing rock (Andersen et al. 1981 a; Bailey et al. 198 la, c). The last pools of magma crystallized to form the latest lujavrites and veins, of extreme composition and exotic mineralogy. Some of this magma was mobilized and intruded towards the N W to form the uranium-rich deposit at Kvanefjeld (Fig. 3). While there is both geological and geochemical evidence for crystal-fractionation processes in the agpaitic magma up to the late lujavrite stage, smooth geochemical profiles like those in Fig. 10 become blurred in the latest arfvedsonite lujavrites which possess a very large innate geochemical scatter (e.g. Andersen et al. 1981; Bailey et al. 1981 a, c). This is attributed to the influence of processes other than crystal fractionation in the final stages, and is taken to indicate the dominance of liquid-fluid distribution processes. The average lujavrite contains about 3 wt.~ volatiles, mainly H20 (Gerassimovsky & Kuznetsova 1967). This is considerably less than the contents in the magma must have been, and it must be inferred that a hydrous phase has left the system. The Th/U evolution of the intrusion shows that material with high Th/U ratios has
485
been lost from the system (Bailey et al. 1981c), and radioactive veins with high Th/U ratios occur outside the intrusion (Hansen 1968; Nielsen 1981). Metasomatism in the overlying lavas has been described from Kvanefjeld (Sorensen et al. 1974; Kunzendorf et al. 1982), and the closing stage in the life of the intrusion, with hydrothermal activity in the overlying lavas, may have been the chemically most open part of it.
Conclusions The layered agpaitic rocks of the Ilimaussaq intrusion present a very well exposed and complete section through the products of crystallization in the upper part of a magma chamber of highly differentiated magma. The initial agpaitic magma had a temperature around 900~ and was as fluid as a basaltic melt. With an estimated H20 + F content of around 4 wt.~ it was volatile undersaturated; volatile saturation was achieved around 700~ The early crystallization was downwards from the roof, and the magma at this stage was well mixed and cooled slowly, perhaps down to around 800~ Later crystallization took place from the bottom upwards, and the magma at this stage is suggested to have been repetitively layered, with temperatures lower than 800~ Residual liquid rich in volatiles and volatileassociated elements continually accumulated at the top of the magma. In the early stage the accumulation was counteracted by the downward crystallization, but when this stopped the upwardcrystallizing floor rocks contributed a steady stream of residual liquid and most probably also a separate fluid phase to the top part of the magma, where the fluid may have redissolved (Kogarko et al. 1977). The last 200 m of magma, which was extremely enriched in residual components, crystallized as a sandwich horizon, whilst incorporating successively more of these components, at temperatures below 700~ There may have been a continuous transition into the hydrothermal stage at temperatures below 500~ The extreme composition of the residual magma which collected at the top of the magma column, and from which the lujavrites crystallized, is an indication that the system was very effectively closed. Only after a roof collapse and at the very end of the crystallization is there indication of material loss from the system. The magma chamber apparently had very little associated surface activity, and this was mainly of hydrothermal character. ACKNOWLEDGMENT: Publication of this paper was
authorized by the Director of the Geological Survey of Greenland.
486
L. M. Larsen & H. Sorensen
References ALLAART, J. H. 1969. The chronology and petrography of the Gardar dykes between Igaliko Fjord and Redekammen, South Greenland. Rapp. Gronlands geol. Unders. 25. ANDERSEN, S., BAILEY, J. C. & BOHSE, H. 1981a. ZrY - U stratigraphy of the kakortokite-lujavrite sequence, southern Ilimaussaq intrusion. Rapp. Gronlands geol. Unders. 103, 69-76. - - , BOHSE, H. & STEENFELT, A. 1981b. A geological section through the southern part of the Ilimaussaq intrusion. Rapp. Gronlands geol. Unders. 103, 3942. BAILEY, D. K. 1969. The stability of acmite in the presence of water. Am. J. Sci. 267A, 1-16. BAILEY, J. C., GWOZDZ, R., ROSE-HANSEN, J., ANDERSEN, L. S. & SORENSEN, H. 1981a. Distribution of lithium and fluorine in lujavrites from the Ilimaussaq intrusion. Rapp. Gronlands geol. Unders. 103, 77-86. , LARSEN, L. M. & SORENSEN,H. 1981b. Introduction to the Ilimaussaq intrusion with a summary of the reported investigations. Rapp. Gronlands geol. Unders. 103, 5-17. --, ROSE-HANSEN, J., LOVBORG, L. & SORENSEN, H. 1981 c. Evolution of Th and U whole-rock contents in the Ilimaussaq intrusion. Rapp. Gronlands geol. Unders. 103, 87-98. BASALTIC VOLCANISM STUDY PROJECT 1981. Basaltic Volcanism on the Terrestrial Planets. Pergamon Press, New York. BLAXLAND, A., VAN BREEMEN, O. & STEENFELT, A. 1976. Age and origin of agpaitic magmatism at Ilimaussaq, South Greenland : Rb-Sr study. Lithos, 9, 31-8. BLUNDELL, D. J. 1978. A gravity survey across the Gardar Igneous Province, SW Greenland. J. geol. Soc. Lond. 135, 545-54. BOHSE, H. & ANDERSEN, S. 1981. Review of the stratigraphic divisions of the kakortokite and lujavrite in southern Ilimaussaq. Rapp. Gronlands geol. Unders. 103, 53-62. --, BROOKS, C. K. & KUNZENDORF, H. 1971. Field observations on the kakortokites of the Ilimaussaq intrusion, South Greenland, including mapping and analyses by portable X-ray fluorescence equipment for zirconium and niobium. Rapp. Gronlands geol. Unders. 38. BOTTINGA, Y. & WEILL, D. 1970. Densities of liquid silicate systems calculated from partial molar volumes of oxide components. Am. J. Sci. 269, 169-82. --, -& RICHET, P. 1982. Density calculations for silicate liquids. I. Revised method for aluminosilicate compositions. Geochim. cosmochim. Acta, 46, 909-19. BUSSEN, I. V. & SAKHAROV, A. S. 1967. Geology of Lovozero Tundra, pp. 1-125. Nauka, Moscow (in Russian). CARMICHAELI. S. E., TURNER, F. J. & VERHOOGEN, J. 1974. Igneous Petrology. McGraw-Hill, New York. CHEN, C. F. & TURNER, J. S. 1980. Crystallization in a
double-diffusive system. J. geophys. Res. 85 (B5), 2573-93. DINGWELL, D. B., SCARFE, C. M. & CRONIN, D. J. 1985. The effect of fluorine on viscosities in the system Na20-A1203-SiO2 : implications for phonolites, trachytes and rhyolites. Am. Mineral. 70, 807. EDGAR, A. D. 1964. Phase-equilibrium relations in the system nepheline-albite-water at 1000 kg/cm 2. J. Geol. 72, 448-60. EMELEUS, C. H. & UPTON, B. G. J. 1976. The Gardar period in southern Greenland. In: ESCHER, A. & WATT, W. S. (eds) Geology of Greenland, pp. 15281. Gr~nlands Geologiske Undersogelse, Copenhagen. ENGELL, J. 1973. A closed system crystal-fractionation model for the agpaitic Ilimaussaq intrusion, South Greenland, with special reference to the lujavrites. Bull. geol. Soc. Denmark, 22, 334-62. FERGUSON J. 1964. Geology of the Ilimaussaq alkaline intrusion, South Greenland. Description of map and structure. Bull. Gronlands geol. Unders. 39 (also Meddr. Gronland, 172 (4)). -1970a. The significance of the kakortokite in the evolution of the Ilimaussaq intrusion, South Greenland. Bull. Gronlands geol. Unders. 89 (also Meddr. Gronland, 190(1)). -1970b. The differentiation of agpaitic magmas: the Ilimaussaq intrusion, South Greenland. Can. Mineral. 10, 335-49. -& PULVERTAFT,T. C. R. 1963. Contrasted styles of igneous layering in the Gardar province of South Greenland. Mineral. Soc. Am. Spec. Pap. 1, 10-21. FORD, C. E., RUSSELL, D. G., CRAVEN, J. A. & FISK, M. R. 1983. Olivine-liquid equilibria: temperature, pressure and composition dependence of the crystal/liquid cation partition coefficients for Mg, Fe 2 +, Ca and Mn. J. Petrol. 24, 256-65. FORSBERG, R. & RASMUSSEN,K. L. 1978. Gravity and rock densities in the Ilimaussaq area, South Greenland. Rapp. Gronlands geol. Unders. 90, 814. GERASSIMOVSKY,V. I. & KUZNETSOVA,S. YA. 1967. On the petrochemistry of the Ilimaussaq intrusion, South Greenland. Geochem. Int. 4(2), 236-46. HANSEN, J. 1968. A study of radioactive veins containing rare-earth minerals in the area surrounding the Ilimaussaq alkaline intrusion in South Greenland. Bull. Gronlands geol. Unders. 76 (also Meddr. Gronland, 181(8)). HILDRETH, W. 1979. The Bishop Tuff: Evidence for the origin of compositional zonation in silicic magma chambers. In: CHAPIN, C. E. & ELSTON, W. E. (eds) Ash-flow Tufts, Geol. Soc. Am. Spec. Pap. 180, 43-75. IRVINE, T. N. 1983. Skaergaard trough-layering structures. Yb. Carnegie Instn. Wash. 82, 289-95. --, KEITH, D. W. & TODD, S. G. 1983. The J-M platinum-palladium reef of the Stillwater complex, Montana: II. Origin by double-diffusive convec-
The Ilimaussaq intrusion tive magma mixing and implications for the Bushveld complex. Econ. Geol. 78, 1287-334. KARUP-MOLLER, S. 1978. The ore minerals of the Ilimaussaq intrusion: their mode of occurrence and their conditions of formation. Bull. Gronlands geol. Umlers. 127. KOGARKO, L. N. 1974. Role of volatiles. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp. 474-87. Wiley, London. & ROMANCHEV, B. P. 1983. Phase equilibria in alkaline melts. Int. Geol. Rev. 25, 534-46. BURNHAM,C. & SHETTLE,D. 1977. Water regime in alkalic magmas. Geochem. Int. 14(3), 1-8. . . . . , LAZUTKINA, L. N. & ROMANCHEV, B. P. 1982. The origin of eudialyte mineralization. Geochem. Int. 19(5), 128-45. KONNERUP-MADSEN, J. & ROSE-HANSEN, J. 1984. Composition and significance of fluid inclusions in the Ilimaussaq peralkaline granite, South Greenland. Bull. Mineral. 107, 317-26. --, LARSEN, E. & ROSE-HANSEN, J. 1979. Hydrocarbon-rich fluid inclusions in minerals from the alkaline Ilimaussaq intrusion, South Greenland. Bull. Mineral. 102, 642-53. --- -, ROSE-HANSEN, J. & LARSEN, E. 1981. Hydrocarbon gases associated with alkaline igneous activity : evidence from compositions of fluid inclusions. Rapp. Gronlands geol. UDders. 103, 99- 108. KUNZENDORF, H., NYEGAARD, P. & NIELSEN, B. L. 1982. Distribution of characteristic elements in the radioactive rocks of the northern part of Kvanefjeld, Ilimaussaq Intrusion, South Greenland. Rapp. Gronlands geol. UDders. 109. LARSEN, J. G. 1974. Stratigrafi, geokemi og petrologi i den ovre vulkanske del af Eriksfjord Formationen, Gardar-provinsen, Sydgronland, Thesis, University of Copenhagen (unpublished). LARSEN, L. M. 1976. Clinopyroxenes and coexisting mafic minerals from the alkaline Ilimaussaq intrusion, South Greenland. J. Petrol. 17, 258-90. . . . . 1977. Aenigmatites from the Ilimaussaq intrusion, South Greenland: Chemistry and petrological implications. Lithos, 10, 257-70. 1979. Distribution of REE and other trace elements between phenocrysts and peralkaline undersaturated magmas, exemplified by rocks from the Gardar igneous province, South Greenland. Lithos, 12, 303-15. . . . . & STEENFELT, A. 1974. Alkali loss and retention in an iron-rich peralkaline phonolite dyke from the Gardar province, South Greenland. Lithos, 7, 81-90. MCBIRNEY, A. R. & NOYES, R. M. 1979. Crystallization and layering of the Skaergaard intrusion. J. Petrol. 20, 487-554. MYSEN, B. O., VIRGO, n. & SEIFERT, F. A. 1982. The structure of silicate melts: Implications for chemical and physical properties of natural magma. Rev. Geophys. space Phys. 20, 353- 83. NIELSEN, B. L. 1981. Radioactive albitites bordering the Ilimaussaq complex: Agpat and Sondre Siorassuit. Rapp. Gronlands geol. UDders. 103, 11923. -
-
-
-
-
-
,
487
& STEENFELT, A. 1979. Intrusive events at Kvanel~jeld in the Ilimaussaq igneous complex. Bull. geol. Soc. Denmark 27, 143-55. PARSONS, I. 1979. The Klokken gabbro-syenite complex, South Greenland: cryptic variation and origin of inversely graded layering. J. Petrol. 20, 65 3 94. ..... & BUTTERFIELD,A. W. 1981. Sedimentary features of the Nunarssuit and Klokken syenites, S. Greenland. J. geol. Soc. Lond. 138, 289-306. PETERSILIE, I. A. & SORENSEN, H. 1970. Hydrocarbon gases and bituminous substances in rocks from the Ilimaussaq alkaline intrusion, South Greenland. Lithos, 3, 59-76. PIOTROWSKI, J. M. & EDGAR, A. n. 1970. Melting relations of undersaturated alkaline rocks from South Greenland compared to those of Africa and Canada. Meddr. Gronland, 181(9). RICE, A. 1981. Convective fractionation. A mechanism to provide cryptic zoning (macrosegregation), layering, crescumulates, banded tufts and explosive volcanism in igneous processes. J. geophys. Res. 86(B1), 405-17. SASS, J. H., NIELSEN, B. L., WOLLENBERG, H. A. & MUNROE, R. J. 1972. Heat flow and surface radioactivity at two sites in southern Greenland. J. geophys. Res. 77, 6435 -44. SHAW, H. R. 1969. Rheology of basalt in the melting range. J. Petrol. 10, 520-35. - - - 1972. Viscosities of magmatic silicate liquids: an empirical method of prediction. Am. J. Sei. 272, 870-93. SOOD, M. K. & EDGAR, A. D. 1970. Melting relations of undersaturated alkaline rocks from the Ilimaussaq intrusion and Gronnedal-ika complex South Greenland, under water vapour and controlled partial oxygen pressure. Meddr. Gronland, 181(12). SORENSEN, H. 1962. On the occurrence of steenstrupine in the Ilimaussaq massif, Southwest Greenland. Bull. Gronlands geol. UDders. 32 (also Meddr. Gronland, 167 (1)). 1966. On the magmatic evolution of the alkaline igneous province of South Greenland. Rapp. Gronlands geol. UDders. 7. 1969. Rhythmic igneous layering in peralkaline intrusions. An essay review on Ilimaussaq (Greenland) and Lovozero (Kola, USSR). Lithos, 2, 26183. , ROSE-HANSEN, J., NIELSEN, B. L., LOVBORG, L., SORENSEN, E. & LUNDGAARD, T. 1974. The uranium deposit at Kvanefjeld, the Ilimaussaq intrusion, South Greenland. Geology, reserves and beneficiation. Rapp. Gronlands geol. UDders. 60. ..... & PETERSEN,O. V. 1981. The mineralogy of the Ilimaussaq intrusion. Rapp. Gronlands geol. UDders. 103, 19-24. SPARKS, R. S'. J., HUPPERT, H. E. & TURNER, J. S. 1984. The fluid dynamics of evolving magma chambers. Phil. Trans. R. Soc. Lond., Ser. A, 310, 511-34. STEENFELr, A. & BOHSE, H. 1975. Variations in the content of uranium in eudialyte from the differentiated alkaline Ilimaussaq intrusion, South Greenland. Lithos, 8, 39-45. -
-
-
-
-
-
488
L. M. Larsen & H. Sorensen
TURNER, J. S. 1980. A fluid-dynamic model of differentiation and layering in magma chambers. Nature, Lond. 285, 213-5. & GUSTAFSON, L. B. 1981. Fluid motions and compositional gradients produced by crystallization or melting at vertical boundaries. J. Volcanol. geotherm. Res. 11, 93-125. UPTON, B. G. J. 1961. Textural features of some contrasted igneous cumulates from South Greenland. Bull. Gronlandsgeol. Unders. 29 (also Meddr. Gronland, 123(6)). -1974. The alkaline province of south-west Greenland. In: SORENSEN,H. (ed.) The Alkaline Rocks, pp. 221-38. Wiley, London. USSING, N. V. 1912. Geology of the country around Julianehaab, Greenland. Meddr. Gronland, 38.
-
-
-
WAGER, L. R. & BROWN, G. M. 1967. Layered Igneous Rocks, Oliver & Boyd, Edinburgh. WILSON, J. R. & LARSEN,S. B. 1985. Two-dimensional study of a layered intrusion--the Hyllingen Series, Norway. Geol. Mag. 122, 97-124. WOLFF, J. A. • STOREY, i . 1984. Zoning in highly alkaline magma bodies. Geol. Mag. 121,563-75. WORMER, G. 8z SCHMINCKE,H.-U. 1984. Mineralogical and chemical zonation of the Laacher See tephra sequence (East Eifel, W. Germany). J. Petrol. 25, 805-35. YODER, H. S. 1954, Synthetic basalt. Yb. Carnegielnstn. Wash. 53, 106-7. 1965. Diopside-anorthite-water at five and ten kilobars and its bearing on explosive volcanism. Yb. Carnegie Instn. Wash. 64, 82-9. -
LOTTE MELCHIOR LARSEN, Gronlands Geologiske Undersogelse, Oster Voldgade 10, DK1350 Copenhagen K, Denmark. HENNING SORENSEN,Geologisk Centralinstitut, Oster Voldgade 10, DK-1350 Copenhagen K, Denmark.
Tertiary alkaline magmatism in East Greenland: a review T. F. D. Nielsen S U M M A R Y : The Tertiary alkaline magmatism in E Greenland is the result of the processes related to the continental break-up of the N Atlantic in the Lower Tertiary. The alkaline rocks are divided into three groups: A, inland alkaline lavas, dykes and intrusive complexes of nephelinitic parentage; B, mildly-alkaline to nephelinitic mafic lavas and dykes mostly in areas that previously experienced tholeiitic magmatism; C, salic subalkaline to alkaline dykes and intrusive complexes related to group B. All three groups are exposed over about 1000 km along the E coast of Greenland, and close to 100 intrusions have been identified, excluding the very intense dyke-swarms. The rocks of nephelinitic parentage occur inland about 100 km W of the coastal zone of most intense magmatism and are believed to constitute an alkaline 'flank' activity to the tholeiitic magmatism of the spreading ridge. Early tholeiites and the nephelinites are thought to be derived from the same mantle source. Both group B and group C rocks occur along the coast and along what has been called the 'initial magmatic lineament', which in some areas evolved to a spreading zone of intense tholeiitic magmatism. The group B rocks may have been generated in the tholeiite feeder system after the end of the tholeiitic magmatism by processes in replenished magma chambers. Interaction between the magmas in the feeder system and the Archaean crust generated the subalkaline to alkaline salic liquids of group C. The occurrences of the alkaline rocks and their relationships with one another, with the preceding tholeiitic magmatism and with the process of continental break-up are described, and petrogenetic models are presented.
Introduction The Tertiary alkaline magmatism in E Greenland is dominated by salic intrusions and there are only minor volumes of extrusive and intrusive mafic alkaline rocks. The magmatism is spatially and chronologically closely related to the preceding voluminous tholeiitic rift magmatism, and both the alkaline and the tholeiitic magmatism are understood within the framework of the continental break-up process in the N Atlantic in the Lower Tertiary. General reviews of the Tertiary of E Greenland have been given by Noe-Nygaard (1974, 1976), Deer (1976) and Brooks & Nielsen (1982a, b). This review focusses on the distribution of the major types of alkaline magmatism and their possible relationship with each other, with the tholeiitic magmatism and with the tectonic development in the E Greenland area in Lower Tertiary time. Many of the geophysical studies of the structure of the shelf areas and studies of the petrology and geochemistry of the igneous rocks made over the last decade are unpublished. Some of the information in this review is from the author's unpublished data, and some is from unpublished data provided by numerous colleagues. There are still large gaps in our knowledge, and the models for the Tertiary alkaline magmatism presented
below are to be regarded as a reference flame for further studies. The alkaline magmatism of the Kangerdlugssuaq area is described in somewhat greater detail because it is believed to exemplify the alkaline magmatism along the coast of E Greenland.
Tectonic framework Since the reconstruction of the N Atlantic, for example by Bullard et al. (1965), workers have been puzzled by the lack of fit of shelf margins along E Greenland, and several researchers have suggested that large parts of the E Greenland shelf formerly regarded as mainly continental might in fact be oceanic. Brooks (1973) and Burke & Dewey (1973) suggested that the initial continental break-up followed the present-day coastline S of the Kangerdlugssuaq area (Fig. 1). Myers (1980), Nielsen & Brooks (1981) and Faller & Soper (1981) contributed to the discussion. Geophysical investigations were presented by Johnson et al. (1975) and Larsen (1978). The most recent contributions are by Larsen (1984), Larsen & Watt (1985) and H. C. Larsen (personal communication). Larsen & Watt (1985) showed that the initial rifting followed the present E Greenland coast S of Kangerdlugssuaq but was displaced to the E
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 489-515.
489
490
T. F. D. Nielsen
FIo. 1. The distribution of Tertiary alkaline rocks in E Greenland (lavas, dyke-swarms and complexes are shown schematically): - - -, initial spreading line (H. C. Larsen, personal communication; Larsen & Watt 1985); . 9 o, initial magmatic lineament (IML); A.24, ocean-floor anomaly 24; B, Borgtinderne syenite; B~, Bontekoe tO; F J, Kcjser Franz Joseph Fjord : GC, Gardiner complex : GH, Gauss Halvo, GHC, Kap Gustav Holm complex; HWH, Hold with Hope: KC, Kialineq centre: K EH, Kap Edvard Holm complex; KI, Kangerdlugssuaq intrusion: KOF, Kong Oscar Fjord : K PC, Kap Parry complex : KSC, Kap Simpson complex; KV, K~erven syenite; JS, I. C. Jacobsen Fjord syenite: JL, Jameson Land: L, Lilloise intrusion; NC, Nualik centre; P-A, P/ttfil~jivit-Nordre Aputiteq centres (tholeiitic): PWB, Prinsen af Wales Bjerge, RS, Ryberg Fjord syenite: SH, Shannon: SK, Skaergaard intrusion; T, Tugtilik : TA, Tasiilaq (fjord); TN, Trekant Nunatakker; TO, Traill O; WBC, Werner Bjerge complex; WB, Watkins Bjerge. Complexes 2 and 3 on Hold with Hope represent the Myggbukta and Kap Broer Ruys intrusions respectively.
Tertiary magmatism in E Greenland along the Blosseville Kyst and Liverpool Land, and almost rejoined the initial trend (seen S of Kangerdlugssuaq) N of Kong Oscar Fjord (Fig. 1). Later sea-floor spreading events occurred in the Blosseville Kyst shelf so that the transition from the continental to oceanic crust along the coast of E Greenland S of Kong Oscar Fjord is situated close to the present-day coastline (Fig. 1).
The Alkaline Magmatism As will be seen from the more detailed descriptions and the chronological information below, three groups of alkaline rocks can be identified. A Strongly alkaline flank magmatismofcentral type, pene-contemporaneous with tholeiitic magmatism along the initial rift in the N Atlantic along the coast of E Greenland. Age about 5550 Ma. B Prolonged and increasingly potassic and alkaline fissure-type magmatism in areas that previously experienced voluminous tholeiitic magmatism, mainly along the 'initial magmatic lineament' (IML) as defined below. Age about 55-33 Ma. C Basic to salic plutonism in areas of alkaline dyke,swarms of group B. The basic components are equivalent to alkaline dykes, whereas the salic component is partly derived from the basic component and is partly of crustal origin. Age about 55-28 Ma. It can be seen in Fig. 1 that the northwards extrapolation of the initial rift trend from S of Kangerdlugssuaq cuts through the basalt areas of the Blosseville Kyst and Jameson Land before it almost rejoins the trend of the initial rift N of Hold with Hope. The following features are located along this IML. 1 Early tholeiitic and later alkaline dykeswarms in Miki Fjord (see Fig. 5). 2 A sodalite syenite occurrence in I.C. Jakobsen Fjord. 3 A nepheline syenite in Ryberg Fjord (N. J. Soper & C. K. Brooks, personal communication). 4 A hidden intrusion in Watkins Bjerge (Matthews 1979). 5 Dyke-swarms in the inner part of Scoresby Sund (Larsen & Watt 1985; Watt et al., in press). 6 A hidden intrusion in western Jameson Land (H.C. Larsen, personal communication). 7 The Werner Bjerge granite-syenite-alkali gabbro complex (Bearth 1959; Brooks et al. 1982). 8 Minor granitic to syenitic intrusions between Werner Bjerge and Antarctic Havn (Kapp 1960).
491
9 Major granitic centres on Traill t0 (Kap Simpson and Kap Parry) (Tyrrell 1932; Schaub 1938, 1942; Engell 1975). 10 Intrusive centres (Myggbukta and Kap Broer Ruys) on Hold with Hope (Upton et al. 1984b). Even though these occurrences are not all contemporaneous, they all appear to be related to the IML described above. Exceptions are as follows. 1 A nephelinite tuffin the inner part of Scoresby Sund (Fig. 1) (Larsen & Watt 1985) which lies off the IML and is provisionally grouped with the off-lineament alkaline magmatism of group A. 2 The nepheline and sodalite syenites of Borgtinderne (Brown et al. 1978), which are referred to group C. 3 The Lilloise alkaline gabbro (Brown 1973), which is referred to group B on the basis of its petrochemical affinities.
Group A: inland alkaline lavas and intrusions Prinsen af Wales Bjerge The Prinsen af Wales Bjerge lavas (PWB in Fig. 1) are poorly known. Anwar (1955) investigated a few profiles, and their isotopic compositions have recently been studied by Evans & Brown (1981). Additional information on the chemistry of the lavas comes from erratic blocks in the Kangerdlugssuaq region and on the Kangerdlugssuaq Gletscher (Brooks & Rucklidge 1974; Fawcett et al. 1982) and unpublished results collected by J. Gittins and M. P. Gorton. Access to the Prinsen af Wales Bjerge is difficult, and the distribution of the lavas can only be judged from aerial photographs. They may be exposed over some 500 km 2 (Bengaard & Henriksen 1984). Wager (1947) estimated the total thickness of the alkaline lavas to be more than 700 m, but it is at present believed that no more than 400 m are preserved. The original thickness is unknown, but observations made by J. Gittins & M. P. Gorton (personal communication) suggest that the alkaline lavas were never much thicker than presently observed. The often thin lava flows have high dips and, as suggested by Wager (1947), the alkaline lavas appear to consist of a large number of overlapping volcanic cones. The alkaline magmatism was of central type as opposed to the fissure-type tholeiitic magmatism. The alkaline lavas are mostly plagioclase, clinopyroxene and/or olivine phyric and include
492
T. F. D. Nielsen above sea level. It lies at the contact between the Archaean basement and the overlying tholeiitic plateau basalts. An age of about 50 Ma is given by Gleadow & Brooks (1979). The ultramafic suite is composed of rings of dunite, wehrlite and pyroxenite cumulates with no systematic petrographic variation (Nielsen 1981) (Fig. 2). They formed in an open magma chamber 1-2 km below a volcano that could have fed lavas similar to those of the Prinsen af Wales Bjerge (Fig. 5(a)). A late ring-dyke system in the centre of the complex includes all the very alkaline evolved rock types mentioned above. They formed from magmas similar to those responsible for the olivine and clinopyroxene cumulates, which, however, evolved by fractionation in closed
hawaiites, nepheline-hawaiites, alkali basalts, basanites and olivine-nephelinites. Sparse alkaline lavas also cap the Trekant Nunatakker about 40 km SW of Prinsen af Wales Bjerge.
The Gardiner complex The Gardiner complex (Fig. 2) about 50 km S of the Prinsen af Wales Bjerge is largely ultramafic but includes a late suite of melilitolites (plutonic rocks composed of about 80% melilite), agpaitic syenites, carbonatites and apatite and magnetite rocks. The complex was described by Frisch & Keusen (1977) and Nielsen (1979, 1980, 1981). The complex, which is at the head of the Kangerdlugssuaq, is approximately 5 km in diameter and is exposed between 800 and 2000 m
~
KANGERDLUGSSUAQ
"%
~
/
" ""
*.
**.
" -:
''~
BASEMENT
MARGINAL SHEETS
CPX- DUNITES
VOLCANICS
BANDED PERIDOTITES
OL- PYROXENITES
ALLUVIUM
ULTRAMP,FICS
MT- PYROXENITES
DUNITES
IJOLITES
CONTACT ROCK
I
I
MELILITOLITES
FIG. 2. The geologyof the Gardiner complex. (After Nielsen 1981.)
Tertiary magmat&m in E Greenland systems. Mafic alkaline dykes are abundant. The complex may contain as much as 200 million tons of apatite in ores with grades of 5%-50% apatite. Alkaline lavas on the nunataq region at 74~ Alkaline lavas occur in the inland areas around 74~ (Katz 1952; Hailer 1956; Brooks et al. 1979) at high elevation as isolated caps on the metamorphic basement. Their original extent is not known. They include nephelinites, basanites and nepheline-hawaiites not unlike those of the Prinsen af Wales Bjerge and dyke-rocks of the Gardiner complex (Fig. 5(a)). Nephelinite tuff in Scoresby Sund The nephelinite tuff in the inner part of Scorcsby Sund erupted after the tholeiitic activity had moved E (Larsen & Watt 1985) and thus at some distance W of the active rift, as did the alkaline lavas of the nunataq zone and the Prinsen af Wales Bjerge lavas.
Group B: alkaline dyke-swarms and associated lavas Major dyke-swarms are exposed along the coasts of E Greenland between 66~ and 70~ (Wager & Deer 1938; Nielsen 1978; Myers 1980) (Fig. 1). The dyke-swarms comprise (1) early tholeiitic dykes affected by coast-parallel flexuring and (2) younger near-vertical dykes of saturated to undersaturated potassic composition (Nielsen 1978). The tholeiitic dyke injection was by far the most intense and the alkaline dykes only account for 1%-25% of the total number of dykes. The alkaline dykes are relatively thin (1-5 m as opposed to an average of above 10 m for the tholeiites), and the total volume in the alkaline dykes represents only a few per cent of the total dyke volume. S of Kangerdlugssuaq the dykes have been studied at Tugtilik and at Tasiilaq (Williams 1974; Rucklidge et al. 1980). At Tugtilik the dykes vary from mildly alkaline to nephelinitic. Similar dykes occur at Tasiilaq. Further dyke-swarms are observed at Kangerdlugssuaq (Nielsen 1978). The tholeiitic part of the dyke-swarm (Thol-1) at Kangerdlugssuaq is deflected by 80 ~ to the E (Figs 1 and 3). A few dykes continue the N N E trend from S of Kangerdlugssuaq into the basement, sediment and basalt areas NE of Kangerdlugssuaq. A second slightlyyounger tholeiitic dyke-swarm (Thol-2) radiates from the mouth of Kangerdlugssuaq (Nielsen
493
1978). Thol-1 formed between 55 and 58 Ma ago, and Thol-2 formed about 55 Ma ago. The tholeiitic swarms are succeeded by three groups of alkaline dykes. The first group follows the main tholeiitic swarm (Thol-1) through the 80 ~ deflection at Kangerdlugssuaq. A second group continues into the areas to the NE. The third group (the late dyke-swarm of Wager (1947)) parallels the Kangerdlugssuaq fjord to the NW. The intensity of these dyke-swarms is shown in Fig. 4. Brooks (1973) and Burke & Dewey (1973) suggested, on the basis of the dominant dyke directions, that a triple junction formed at Kangerdlugssuaq during the intense magmatism, doming and continental break-up. In all the three alkaline dyke-swarms mentioned above there is an increase in alkalinity (mainly K20) with decreasing age (Fig. 5(b)). Independently of their orientation they can be divided into two groups on the basis of their age relative to the syenitic magmatism at Kangerdlugssuaq. The syenites are generally assumed to be about 50 Ma old (Nielsen 1978; Brooks & Nielsen 1982b) and the alkaline dykes pre-dating the syenites are potassic tholeiites, alkali basalts and hawaiites, whereas those post-dating the syenites are more 'lamprophyric' and include camptonites but also a few tholeiites. Nielsen (1978) termed the two generations Alk-1 and Alk2. A description of the dykes (both Alk-1 and Alk-2) cutting the Skaergaard intrusion was given by Vincent (1953). On the basis of fission-track dating and the age of the tholeiitic and syenitic magmatism, it has been proposed that the Alk-1 dykes were emplaced between 55 and 50 Ma ago and the Alk-2 dykes between about 50 and 33 Ma ago (Gleadow & Brooks 1979). The youngest dykes in the Kangerdlugssuaq area compose an E-W group of tholeiitic to transitional dykes (Trans-1 of Nielsen (1978)) along the southern part of the Blosseville Kyst. Their age is not known, but they have provisionally been regarded as about 35 Ma old (Nielsen 1978). The only information on alkaline dykes from the Blosseville Kyst comes from the Lilloise area in Wiedemann Fjord (Brooks & Rucklidge 1973). Here, camptonite dykes are divided into a NE coast-parallel swarm and an E-W swarm. The dykes are related to the Lilloise alkali gabbro complex at the head of Wiedemann Fjord (Brown 1973). One of the dykes has been dated to 52.2_ 1.2 Ma (Gleadow & Brooks 1979). A feature of interest is the occurrence of an upper-mantle harzburgite nodule in one of the dykes (Brooks & Rucklidge 1973). The dykes are not described in detail, but Brooks & Rucklidge (1973) compare them with the camptonites of the Skaergaard
494
T. F. D. Nielsen
t ' x ~ ..-
Y4.~ ?
d
) .......
;~:~i'~~'~ .~, ~ 7 ,~"..# ~ "
6--8 15N-----~
',
"
TL
~
|
..... i
K ANGERDLUGSSUAQ COMPLEX ...~,
~: ~.~
'i
*,,,, ... KONTAKTBJERG
-
~it
~//I
KAP
EOV
KAP
BOSWFLL
LDV ~NDR
APUTITEQ
HOLM
I-K[)LM
O()MPLEX
IN 1 I--I
I,
!0 i,m
FIG. 3. Simplified map of the Tertiary intrusive complexes around Kangerdlugssuaq (after Gleadow & Brooks 1979) cross-hatched, gabbros; black, syenites (s. 1.) and Gardiner complex; SK, Skaergaard intrusion. Dykeswarms are indicated schematically, after Wager & Deer (1938), Wager (1947) and Nielsen (1978).
1
,,,~ --v., "", f
,
Ww ~
~
",
I
- ~
:5-//"~
,
I
~
........
'-..~ ......
FIG. 4. The western part of Kraemers O (after Brooks & Nielsen 1978). Kraemers O lies inside the coastparallel dyke-swarm and only the intensity of the tholeiitic Thol-2 and inland alkaline dyke-swarms is indicated. The inset shows schematically the dominant trends of these dyke generations : l KraemersO syenite; 2 contact breccia; 3 basaltic inclusions;4 tholeiitic grabbros; 5 Archaean basement;6 faults; 7 alluvium; 8 lake;9 dykes.
Tertiary magmatism in E Greenland intrusion described by Vincent (1953) referred to the Alk-1 and Alk-2 generations of Nielsen (1978). Dykes from the inner parts of Scoresby Sund have not yet been described, but they include early tholeiitic basalts and younger monchiquites and alkaline trachytes (Larsen & Watt 1985). More dyke-swarms are known N of Scoresby Sund but not in any detail. The reader is referred to Noe-Nygaard (1976) for information. A dyke-swarm sharing many similarities with those of the Lilloise and Kangerdlugssuaq areas is related to the Werner Bjerge complex (Fig. 1). The swarm trends NE (fewer dykes in the NW). The dykes are believed to be about 30 Ma old (Rex et al. 1979b) and contemporaneous with the youngest alkaline dykes in the Kangerdlugssuaq area about 33 Ma old) (Gleadow & Brooks 1979). Kapp (1960) described the alkaline dykes, which include tholeiites and alkaline 'lamprophyre' dykes very similar to those described from the other plutonic centres in E Greenland. Few occurrences from between Kong Oscar Fjord and Kejser Franz Joseph Fjord are known. Noe-Nygaard (1976) shows dykes trending N-S and NE-SW, but no details are given. Some are basaltic and others are alkaline. More information is available from the northern parts of the E Greenland coast (Hald 1978; Upton et al. 1980, 1984 a, b). Tholeiitic plateau basalts are exposed between Bontekoe ~ and Shannon (Fig. 1). On Gauss Halvo and Hold with Hope the lower-plateau lava series (LPLS) of Kpoor tholeiites with MORB affinities is followed by up to 350 m of upper-plateau lava series (UPLS) which is much more variable than the LPLS. Even though Upton et al. (1984a) have classified the UPLS as K-rich tholeiites, many of them, including the alkaline 'anomalous' UPLS, fall on the alkaline side of the Hawaiian divide. The UPLS include aphyric as well as stronglyporphyritic picrites, ankaramites and hawaiites. Some contain cognate xenoliths. Upton et al. (1984a) suggested that the UPLS were fed from a central-type shield volcano resting on the LPLS. Chemically and chronologically the UPLS correspond to the Alk-1 dykes and probably some of the Alk-2 dykes from the Kangerdlugssuaq area (Fig. 5(a)). Alk-1 dykes at Kangerdlugssuaq are 55-50 Ma old. The UPLS of Hold with Hope have given a K/Ar age of about 51 Ma (Upton et al. 1984b). The UPLS are cut by NE-SW and N-S alkaline dyke-swarms with ages of 48-29 Ma which recall the Alk-1 and Alk-2 dyke-swarms in the Kangerdlugssuaq area (Upton et al. 1984b). The characteristics of dyke-swarms of group B can be summarized as follows.
495
1 They generally follow a N E - S W trend parallel to those of preceding tholeiitic dyke-swarms, except when related to intrusive complexes. Nearly all the occurrences of these dykes and lavas seem to be related to the IML (Fig. 1). 2 The dyke-swarms and related lavas post-date the K-poor tholeiites common along most of the E Greenland coast. 3 The alkaline dykes and lavas formed over a period of more than 25 Ma. The lack of systematic chemical evolution, although increasingly alkaline rock types occur, suggests that the lavas and dykes are not the result of some uniform melting event in the mantle but represent individual magma batches which followed individual evolutionary trends. 4 Many of the dykes are related to the gabbrosyenite-granite complexes of alkaline affinities described below.
Group C: syenitic to granitic complexes and dykes related to group B There are nearly 100 intrusions of Tertiary age along the E Greenland coast and the IML. The group C intrusions include all the complexes composed of quartz porphyries, granites, quartz syenites, saturated syenites and undersaturated syenites. The complexes, going northwards from the most southerly complex at Kap Gustav Holm, are briefly characterized below. The location of the complexes is shown in Fig. 1.
Kap Gustav Holm complex (66035'N) A preliminary description was given by Wager (1934), and a more recent study has been reported by Myers (1980). The complex (Fig. 6) was emplaced at the basement-sediment interface and includes (1) an early tholeiitic gabbro, (2) hornblende monzonite, (3) a major layered monzonite, (4) a syenite ring-dyke intrusion and (5) a granitic plug. Younger swarms of mafic to leucocratic dykes, both radial and coast parallel, cut the intrusions. The tholeiitic gabbros are regarded as contemporaneous with, for example, the Skaergaard intrusion at about 55 Ma. An apatite fiission-track age of 47+ 2.9 Ma was obtained from an alkali rhyolite dyke just N of the complex (Gleadow & Brooks 1979) and is taken as the age of the late salic magmatism. The granitic plug and lamprophyre dykes may be even younger.
496
T. F. D. Nielsen
FIG. 5. (a) The variations in the inland alkaline rock occurrences (group A) compared with the alkaline suites in (b) (excluding the Tugtulik dykes) and the tholeiitic basalts (data from quoted references and unpublished sources). (b) the variability of the group B alkaline dyke-swarms and group C syenites (sensu lato) and granites (data from quoted references and unpublished sources). The field of the tholeiites is shown for comparison. The minimum melt at 2 kb H20 in the Qz-Ab-Or system (40~ Qz, 35~ Ab and 25~ Or) is indicated by the encircled cross (references in the text).
Tertiary m a g m a t & m in E Greenland
497
FIG. 6. K ap Gustav Holm and Kialineq district (after Myers 1980); Brown et al. 1977); cross-hatched, gabbros; grey, acid-basic complexes; black, syenites (sensu lato). The dyke-swarms are indicated schematically.
The Kialineq centre (67~ The Kialineq centre is a very large intrusive centre at Kial~q, which includes more than 10 individual intrusive bodies (Fig. 6). As at Kap Gustav Holm the earliest intrusion is tholeiitic gabbro-the Imilik gabbro-described by Brown & Farmer (1972) and Myers (1980) who called attention to the multiple nature of the gabbro intrusion. The gabbros were followed by acid-
basic net-veined complexes and plugs (Brooks 1977). These were followed by intrusions of saturated to oversaturated syenites, with granitic intrusions as the latest event (Brown et al. 1977; Brooks, personal communication). The Imilik gabbro is assumed to be about 55 Ma old. Samples from a number of the younger intrusions gave a Rb/Sr isochron of 35+2 Ma (Brown et al. 1977). However, the isochron could be a mixing line between salic and mafic
498
T. F. D. Nielsen
compositions of different Sr i giving only an apparent age. However, a fission-track age of about 35 Ma was obtained from a syenite of the centre (Gleadow & Brooks 1979). As at Kap Gustav Holm the most evolved rocks have alkaline affinities.
The Nualik centre (67~ The principal complex is the very large Kruuse Fjord or Agtertia gabbro (Myers et al. 1980). Other gabbros occur on Kap Louise Ussing, and three gabbro bodies occur on Sondre Aputiteq (Rex et al. 1979a; T.S. Petersen, personal communication). A sphene from one of these gabbros gave a fission-track age of 52.5 Ma (Gleadow & Brooks 1979). Dodemandspynten and parts of Ersingerseq are composed of breccias comprising an andesitic matrix with inclusions of basalts and Archaean gneisses. On Ersingerseq the breccia is intruded by a diorite plug. Finally, the earlier rocks are invaded by granitic veins. A Rb/Sr isochron on 'contaminated' or hybridized grano-
phyres has given an 'apparent' age of about 55 Ma (Rex et al. 1979a). A sketch map of the Nualik centre is given by Rex et al. (1979a).
The Kangerdlugssuaq district (68~ More than 20 syenite-alkali granite complexes are known in the Kangerdlugssuaq district (Fig. 3). They share many geochemical characteristics (see next section) and are generally assumed to have comparable origins (Deer et al. 1984). Plutonism in the Kangerdlugssuaq district was initiated by large tholeiitic gabbro intrusions: the Kap Edvard Holm intrusion, the Kap Edvard Holm complex, the Skaergaard intrusion, the gabbroic macrodykes of Miki Fjord and Kr~emer O, the Ka~rven gabbro intrusion and marie plugs in Watkins Fjord, Courtauld Fjord and the inner parts of Kangerdlugssuaq Fjord (Brooks & Nielsen 1982b). Where contact relations can be observed, the syenites and the granites are seen to intrude the gabbros and the age relationships are similar to those observed in the plutonic
FIG. 7. Geological reconnaissance map of the Kap Boswell complex based partly on field work carried out in 1979 and partly on Abbott & Deer (1972). The position is indicated in Fig. 3.
Tert&ry magmatism
districts to the S. This is also confirmed by studies of dyke-swarms (Nielsen 1978) and by extensive radiometric dating (Gleadow & Brooks 1979). The Kap Boswell complex (Fig. 7) is composed mainly of oversaturated syenites and granites (seven individual intrusive phases) surrounded by a zone up to 1000 m wide intruded by so-called pillow dykes. They are composed of acid and basic material, where the basic material forms pillows with chilled or gradational margins. The intensity of the veining is variable, and the zone may occasionally be termed a breccia and can be compared with breccias in the complexes further to the S. The youngest rocks of the complex are oversaturated tinguaites and peralkaline monazite-bearing granites. Strongly altered silicic volcanics on Barberkniven have been preserved by cauldron subsidence. Petrographic and geochemical details are given by Deer et al. (1984). A K/Ar age of 51.8 +__3.9 Ma was obtained from an amphibole of a late peralkaline granite (Beckinsale et al. 1970). Four individual syenite intrusions occur in the Kap Deichman area (Fig. 8). The syenites are saturated to oversaturated and mostly aenigmatite-fayalite-hedenbergite bearing (Brooks & Nielsen 1982b; Deer et al. 1984) As at Kap Boswell the intrusions are surrounded by zones of pillow dykes and breccias (Brooks & Nielsen 1982b). The Kap Deichmann syenites are believed to be contemporaneous with the Kap Boswell syenites. The last major intrusion in the Kap Edvard Holm complex is a breccia in Kontaktbjerg (Fig. 3). Most of the intrusion is a granite-syenite breccia, which in the centre is intruded by a 'hybrid' compositionally equivalent to an oversaturated trachyandesite (Deer et al. 1984). The Kangerdlugssuaq intrusion of nordmarkites, pulaskites and foyaites may more appropriately be termed the 'Kangerdlugssuaq complex'. It includes the Kangerdlugssuaq intrusion proper and a number of both older and younger satellite intrusions (Fig. 9). The exact genetic relationship between the satellites and the Kangerdlugssuaq intrusion is not known. The Kangerdlugssuaq intrusion is circular (33 km in diameter) and is composed of an outer ring of nordmarkites (more than 90~ of the complex), a pulaskite ring (5.3%) and a core of foyaites (0.6%) (Kempe et al. 1970). The geology, petrography and geochemistry of the complex is described by Wager (1965), Kempe et al. (1970), Kempe & Deer (1970, 1976), Pankhurst et al. (1976) and Brooks & Gill (1982). Wager (1965) visualized the intrusion as saucer shaped with the undersaturated foyaites in the centre. No internal intrusive contacts were described (Wager 1965), even though pulaskites
in E G r e e n l a n d
499
form intrusive sheets in the nordmarkites in Amdrup Fjord (C.K. Brooks, personal communication). Wager (1965) argued that the oversaturated liquids of the nordmarkites crossed the thermal divide to foyaitic compositions by crystal fractionation. Pankhurst et al. (1976) suggested that the phonolite liquids of the foyaites were oversaturated liquids depleted in SiO2 by extraction of SiO2 through circulating meteoric water. Brooks & Gill (1982), however, showed that the alkali clinopyroxenes of the foyaites are more primitive than those of the oversaturated rocks. In conjunction with the isotopic differences between the oversaturated and undersaturated syenites they suggested that the foyaites are fractionates of an undersaturated 'basic' parent and that the oversaturated nordmarkites and pulaskites are basement-contaminated foyaites. The basic parent, according to Deer et al. (1984), would be similar to an Azores alkali basalt. Pankhurst et al. (1976) found an age of about 50 Ma for the complex. The satellite intrusions (Fig. 9) include the younger Snout series nordmarkites at the E margin, the nordmarkitic Bagnaesset syenite complex to the SE, a younger biotite granite plug to the S, an older augite syenite to the W and the older Kaerven syenite at the NE margin. An earlier syenite (peak 2005 m syenite) in the northern part of the Kangerdlugssuaq intrusion is engulfed by the nordmarkites (Deer & Kempe 1976). The syenites on Bagnaesset form domes and sheets in the often strongly metasomatized Archaean gneisses (Larsen 1982). Small 'diorite' domes and pillows intrude alongside the syenites in the 'Astrophyllite Bay' complex (Fig. 10) and form irregular intrusive masses with crenulated and chilled margins in the Bagnaesset syenites (Brooks & Nielsen 1982b; Larsen 1982). The 'diorite' domes and pillows have rheomorphic basement contacts that back-vein the diorites. As observed in the complexes described previously the salic liquids coexisted with more mafic liquids. Lamprophyre dykes cutting the nordmarkites in the SE sector of the Kangerdlugssuaq intrusion often carry kaersutite gabbro inclusions (Brooks & Platt 1975), and a major gravimetric anomaly (Blank & Gettins, unpublished data) suggests that a major layered kaersutite gabbro underlies this part of the Kangerdlugssuaq intrusion. The Kaerven syenite in the NE contact is assumed to be older than the Kangerdlugssuaq intrusion. The youngest (?) intrusive rocks in the 'Kangerdlugssuaq complex' occur in the sub-volcanic 'Flammefjeld complex' as mineralized quartz porphyry breccias and microgranite sheets in the
500
T. F. D. Nielsen
FIG. 8. Kap Deichmann and Hutchinson Gletscher syenites based partly on Abbott & Deer (1972) and partly on field work carried out in 1979. A major deviation from the map of Abbott & Deer is the identification of the Kap Deichmann syenite II.
T e r t i a r y m a g m a t i s m in E G r e e n l a n d
501
FIG. 9. Simplified geological map of the Kangerdlugssuaq intrusion and satellites (after Kempe et al. (1970), with additional observations from Bagnaesset (Larsen 1982) and 'Flammefjeld' (C. K. Brooks, personal communication; Geyti & Thomassen 1985). The 'Astrophyllite Bay' complex (ABC) is indicated in black together with an alkali gabbro-diorite plug at the contact to the Snout series. SE part of the 'complex'. Interesting mineralizations of Mo, Pb, Zn, Ag and Au, recalling the classic quartz porphyry molybdenum deposits of the Climax type, have been observed (Geyti & Thomassen 1985). The last major syenite body in the Kangerdlugssuaq district is the Kraemers ~ syenite of which only a minor part is exposed (Fig. 3). The syenites are layered and rich in basalt inclusions. They are separated from the Archaean basement by a syenite breccia, which suggests explosive emplacement (Brooks & Nielsen 1982b). Parts of the syenites are inclusion free and may indicate multiple intrusion. Of special interest are associated peralkaline rhyolite dykes (Brooks & Rucklidge 1976). The granites are not alkaline, but the very alkaline rhyolite dykes and pegmatites relate the intrusion to the other alkaline complexes of the Kangerdlugssuaq district (Brooks & Rucklidge 1976; Brooks & Nielsen 1982b; Layne et al. 1982).
The Blosseville Kyst (68~176
FIG. 10. Trachyandesite 'pillows' in the 'Astrophyllite Bay' complex emplaced in a mobile basement. Rheomorphic basement veins have intruded the surrounding basement and back-veined the trachyandesite bodies. Basement structures are preserved in the foreground. (Photograph by C. K. Brooks.)
Few intrusive complexes have been reported from this part of E Greenland and only little is known about them. In I.C. Jakobsen Fjord (Fig. 1) a large number of sphene-rich sodalite syenite blocks have been collected in a restricted area. The intrusion has not been found but their distribution suggests that they are of local origin. They are best compared with the syenites of the Borgtinderne intrusion (see below). Soper et al.
502
T. F. D. Nielsen Bearth (1959), Kapp (1960) and Brooks et al. (1982). The largest complex, the Werner Bjerge intrusion, includes three individual centres: an early south-eastern 'basic' alkaline centre, a northern alkali syenite and granite centre and a south-western nepheline syenite centre, which is believed to be the youngest. The 'basic' centre is sub-volcanic and includes agglomerates, breccias and plutonic rocks. The plutonic rocks range from alkaline pyroxenites and gabbros, syenogabbros and syenites to alkali granites. The northern centre is dominated by undersaturated syenites and pulaskites but includes a late nordmarkite and alkali granite body. The late oversaturated rocks of both the 'basic' and the northern centres are mineralized and the molybdenum deposit at Malmbjerg (Kirchner 1964) is a stockwork mineralization related to the late granites of the northern centre. Numerous dykes were emplaced during and after the emplacement of the plutonic rocks. An isochron based on samples from several of the intrusive bodies, excluding the basic centre, gave an age of 30 4- 2 Ma (Rex et al. 1979a). Four intrusive complexes of granites, syenites and minor alkaline mafic units are exposed
(1976) reported an undersaturated very sodic syenitic body in Ryberg Fjord. The Lilloise alkaline gabbro complex (Brown 1973" Sheppard et al. 1977) and the Borgtinderne nepheline syenite (Brown et al. 1978) are emplaced into the Tertiary plateau lavas and fall outside the IML (Fig. 1). Brown (1973) described the Lilloise gabbro as a mafic layered pluton intruded by later syenite veins. The parental magma was alkaline. Lamprophyre dykes intruded during and after the emplacement of the complex are described above. The Borgtinderne complex is composed of leucocratic nepheline syenites and'dark' syenites. The 'dark' syenites are developed from the leucocratic syenites through contamination with metamorphosed and metasomatized tholeiitic plateau basalts (Brown et al. 1978). Gleadow & Brooks (1979) found a weighted mean age of 47.4_ 0.9 Ma for the complex. The Mestersvig area (72~ The alkaline complexes of the Mestersvig area (Fig. 11) are described by Noe-Nygaard (1941), I
24W/
'C'JO,90 BJERGE
KAP
,9 NORTH
COMPLEX
BJERG
COMPLEX
WERNER
r
BJERGE
k
i
5km
COMPLEX FIG. 11. Simplified geological map of the intrusive complexes S of Kong Oscar Fjord (after Bearth (1959) and Kapp (1960)): cross-hatched, the basic complex of the Werner Bjerge complex and the basic part of Theresabjerg complex; black, undersaturated and oversaturated syenites; dotted, alkali granites; Mo, Malmbjerg Mo stockwork deposit.
Tertiary magmatism in E Greenland between the Werner Bjerge complex and Antarctic Havn in Kong Oscar Fjord (Kapp 1960) (Fig. 11). The Pictet Bjerge complex is a large alkali granite body that formed late in the regional magmatic development (Noe-Nygaard 1941, 1976). Large breccia plugs are related to the granite and to the neighbouring Theresabjerg complex, which includes an early alkaline gabbro-diorite followed by syenites and monzonites. Although the Pictet Bjerge alkali granite lies outside the Theresabjerg complex, it probably marks the end of the magmatic evolution in this complex (Noe-Nygaard 1976). Syenites and nordmarkites of the Oksehorn complex and the Kap Syenite intrusion are exposed E and W of Theresabjerg (Fig. 11). A large number of dykes (see previous section) and sills are found in the host rocks. They include a large variety of alkaline mafic rock types resembling the lamprophyric Alk-2 dykes of the Kangerdlugssuaq district. The leucocratic dyke-rocks mimic the spectrum observed in the salic plutonic rocks.
503
72
15'-
Traiil O (72~ The alkali granites and syenites of the Kap Simpson complex or Forchhammer pluton are exposed on the SE peninsula of the island (Tyrrell 1932; Schaub 1938; Parr & Punzengruber 1969; Noe-Nygaard 1976) (Fig. 12). Only the roof of the complex is exposed and large sediment blocks within the perimeter of the complex indicate cauldron subsidence (Noe-Nygaard 1976). Hydrothermal alteration is extensive and parts of the complex are mineralized. The structure of the complex is not known, but ring-dykes are present along the eastern and western margins. Sills and dykes of granites and syenites are common in the neighbouring Mesozoic sediments. It has been dated at about 38 Ma (Gleadow & Brooks 1979; Rex et al. 1979b). The syenites and granites of the Kap Parry complex are exposed on the NE peninsula (Tyrrell 1932; Schaub 1942; Engell 1975). The complex is composed of breccias and plutonic and volcanic rocks from three volcanic centres, including acid volcanic breccias, quartz syenites (nordmarkites) and at least two generations of alkali granites. The roof over the complex is partly preserved. No details have been published. The age of the complex is reported to be about 40 Ma (Rex et al. 1979b).
Hold with Hope (74~ The most northerly exposure of intrusive Tertiary rocks in E Greenland is found in this area. The sub-volcanic mainly basaltic Myggbukta complex
FIG. 12. Outline of the Kap Simpson and Kap Parry syenite complexes (after Schaub (1938), Engell (1975) and Parr & Punzengruber (1969)): black, syenites (s. 1.); shaded volcanic breccias; dotted, older sediments and contemporaneous acid volcanics downfaulted owing to cauldron subsidence (Noe-Nygaard 1976). is only partly exposed at Mackenzie Bugt S of Hold with Hope. The Kap Broer Ruys complex 35 km to the E is badly exposed and only sparse felsites and granophyres are observed (Upton et al. 1984b). Large positive magnetic anomalies are related to both of the complexes, suggesting large basic masses at depth (H.C. Larsen, personal communication). Upton et al. (1984b) have given a detailed description of the Myggbukta complex. The basic rocks of the complex are genetically related to the UPLS lavas of group B. Upton et al. found K/Ar ages of 3428 Ma for the Myggbukta complex and ages of 48-46 Ma for the Kap Broer Ruys complex.
Geochemistry The geochemistry of the Tertiary alkaline igneous rocks of E Greenland cannot be described in detail here, partly because of the very large geochemical variations and partly because many of the occurrences of alkaline rocks in E Greenland are only superficially described. There are
504
T. F. D. Nielsen
some general geochemical distributions, however, which may throw light on some of the fundamental processes operating during and after the continental break-up along the E coast of Greenland. The alkaline rock occurrences have been divided into three groups in the previous sections: A, inland alkaline intrusives and extrusives; B, dyke-swarms and lavas along the coast and the IML; C, the intrusive complexes. Alkalis versus MgO plots for the inland alkaline intrusives and extrusives of group A are shown in Fig. 5(a). The large grey field encompasses all the alkaline compositions of groups B and C (Fig. 5(b), excluding the anomalous Tugtilik dykes). It can be seen that the Prinsen af Wales Bjerge lavas and the nephelinites and nepheline-hawaiites of the nunataq zone at 74~ overlap, and that the whole group is generally more alkaline, at a given MgO concentration, than the rocks of group B. Distinctly less alkaline rocks from the inland areas are basanites of the nunataq zone. The mafic dyke-rocks of the Gardiner complex (15% MgO or more) overlap entirely with the inland lavas, but as they become more and more evolved they become increasingly alkaline. The representative analyses in Tables 1 and 2 show that the alkaline dykes and lavas of group A are sodic compared with the more potassic dykes and lavas of group B. Large amounts of trace-element data exist for the Tertiary alkaline rocks of E Greenland, but it is beyond the scope of this review to go into a more detailed discussion of, for example, the distribution of rare-earth elements. The composition of the alkaline rocks of groups B and C, which are closely related spatially and chronologically (see earlier section) is summarised in Fig. 5(b). A number of important observations can be made in this Figure. 1 Alk-1 dykes from Kangerdlugssuaq overlap with the UPLS of the Hold with Hope area. 2 There is a gradual and general increase in total alkalis with time in the alkaline dykeswarms and lavas of group B, independent of the degree of differentiation expressed as percentage MgO. 3 The basic parts of the pillow dykes and complexes and related hybrids of syenite complexes (trachyandesites) overlap compositionally with the most evolved Alk-2 dykes and the large syenite field, and they may form the continuation of the field of the Alk-2 dykes. 4 The rheomorphic basement at the contacts of trachyandesite bodies of the 'Astrophyllite Bay' complex plot within the syenite field and are more alkaline than expected for a minimum melt from the metamorphic-basement complex.
Isotopes A summary of the (87Sr/S6Sr)o = Sri data from the various groups and sub-groups is shown in Fig. 13. The only sub-groups with comparatively low Sr i are the rocks of the Gardiner complex, some of the inland nephelinite lavas, the tholeiites of the UPLS and some of the tholeiitic plateau lavas. All other groups and sub-groups record higher and much more variable ratios. The lowest ratios of the nephelinites, Gardiner rocks and tholeiites approach those of the most enriched tholeiites of Iceland, which suggests that all other groups of alkaline rocks have been contaminated by an 8VSr-rich source (e.g. the continental crust) or are derived from an 87Sr-enriched mantle. Carter et al. (1979) suggested that all the E Greenland tholeiites were derived from a source similar to that below Iceland. Pb isotope compositions of galena deposits throughout the Tertiary of E Greenland have been described by P. G. Coomer, C. K. Brooks & B. Thomassen (unpublished data). Their interpretation suggests that the Pb is mostly recycled pre-Tertiary in age. Only in the Malmbjerg Mo deposit in the Werner Bjerge complex does the Pb seem to be of Tertiary age. The galena deposits are mainly related to the oversaturated syenites and late granites (see earlier section).
Discussion and petrological models As mentioned in the Introduction there is a close relationship between the earlier tholeiitic magmatism and the alkaline magmatism of group B, and the understanding of the location, chronology and petrology of the alkaline magmatism is based on an understanding of the location and chronology of the tholeiitic magmatism.
The tholeiitic magmatism The tholeiitic magmatism was initially located along the IML. Basalt areas formed at Kangerdlugssuaq (Nielsen et al. 1981) and in the inner part of Scoresby Sund (Larsen & Watt 1985). Plateau basalts may have been erupted along other parts of the IML, but have subsequently been removed by erosion. In the Kangerdlugssuaq area the early basalts include a group of picrites and genetically related basalts and andesites with high levels of large-ion lithophile elements (LILE) (Nielsen et al. 1981) and a second group of less LILE-rich Fe-Ti tholeiites (Brooks et al. 1976) relating to the dominant tholeiitic plateau basalts in E Greenland (Larsen
Tertiary magmatism in E Greenland
5o5
TABLE 1. Inland alkaline rocks of group A 1
2
3
4
5
6
7
8
9
10
11
12
13
14
MgO CaO Na20 K20 MnO TiO2 P205 LOI
(wt.~) 38.92 39.27 4.48 7.09 4.89 5.67 11.12 9.89 18.09 16.08 10.96 10.57 2.43 2.37 1.19 0.96 0.20 0.26 4.10 5.14 0.58 0.66 1.86 2.44
42.85 6.30 5.63 7.69 20.32 8.71 1.62 1.37 0.19 2.12 0.22 2.37
44.58 6.83 4.41 9.02 10.93 10.66 2.24 2.17 0.16 4.60 0.60 2.83
43.36 7.55 5.15 10.74 10.10 13.48 2.25 1.22 0.22 4.94 0.52 1.48
42.19 6.84 3.99 10.85 11.24 11.18 2.51 1.03 0.19 5.29 0.64 1.72
46.34 10.59 4.63 9.22 5.71 8.99 3.98 1.69 0.18 5.24 0.63 2.17
51.14 13.59 5.21 6.13 4.01 7.12 4.65 1.99 0.15 3.45 0.81 1.02
44.74 13.98 6.87 5.23 6.28 8.94 3.67 2.01 0.26 3.07 1.14 4.04
49.51 12.97 4.43 7.34 4.42 7.23 6.81 1.48 0.14 3.17 0.68 1.10
47.34 16.06 3.81 4.37 2.96 6.58 6.59 2.97 0.19 2.94 0.59 0.85
47.79 15.32 6.14 2.27 2.77 5.83 10.09 2.84 0.21 2.87 0.61 0.53
46.49 25.02 3.59 0.52 0.95 0.91 15.25 3.25 0.07 0.39 0.02 1.60
1.52 0.04 0.17 0.48 0.37 50.40 0.27 0.03 0.06 0.10 1.66 44.42
Total
98.82
99.39
99.03
100.94
97.67
99.37
99.27
100.07
99.28
95.07
97.06
96.46
99.52
. 8.10 8.11 5.87 . 3.03 .
. 12.82 13.98 2.17 . . 2.69 . .
6.09 13.18 4.36 . 4.36 .
9.09 28.45 6.04 . 2.83 .
1.08 11.76 39.35 10.33 . -.
. . . . . 12.34 8.75 17.55 16.78 26.30 34.94 24.82 8.15 16.36 0.45 5.47 -. . . . 3.24 12,29 16.77 29.71 . . 17.76 . 0.52 17.14 25.00 15.94 14.88 1.19 1.93 . . . . . 5.83 2.74 --. . . . 9.16 6.42 5.52 -6.06 6.02 5.58 4.79 . 0.59 2.71 1.58 1.37 1.41
Major elements
SiO2 AlzO 3 FezO3
FeO
100.36
C I P W norms
O or ab an lc ne ac ns
. ---5.51 8.89 3.65
.
.
.
di wo
32.33 .
.
hy ol In mt il pf ap
.
.
29.09 3.09 5.26 7.79 .
. 5.79 3.44 5.99 . 9.23 . .
34.04
.
28.30 . .
.
1.34
30.91 . 8.16 4.03 .
1.56
.
0.51
. 5.93 . .
45.48
. .
. .
10.86 . . 6.39 8.74
8.10
.
.
37.51
. .
1.39
.
.
37.15
. .
21.59 . 8.39 9.97
.
. 7.24 8.18 6.93
.
7.50 9.43 .
1.21
.
13.14 . 5.79 10.05
.
.
.
27.61
.
4.15 . 6.71 9.95 .
1.48
.
15.57
.
1.46
4.18 -. 7.55 6.55 .
1.88
19.21 0.68 -59.55 10.39 1.55 3.47 -0.83 -0.74 -0.05
1, Mela-nephelinite, volcanic, Kangerdlugssuaq (Brooks & Rucklidge 1974, Table 3, column 1); 2, nephelinite, volcanic, nunataq zone at 74~ (Brooks et al. 1979, Table 10, column 3); 3, nephelinite, dyke, Gardiner complex, G M 55201D; 4, limburgite, volcanic, Kangerdlugssuaq (Brooks & Rucklidge 1974, Table 3, column 2); 5, basanite, volcanic, nunataq zone at 74~ (Brooks et al. 1979, Table 10, column 6); 6, basanite, lava, Prinsen af Wales Bjerge, JG22 (Gittins, personal communication); 7, N e hawaiite, lava, Prinsen af Wales Bjerge, JG58 (Gittins, personal communication); 8, hawaiite, lava, Prinsen af Wales Bjerge, JG63 (Gittins, personal communication); 9, Ne-hawaiite, volcanic, nunataq zone at 74~ (Brooks et al. 1979, Table 10, column 8); 10, nephelinite, dyke, Gardiner complex, GM29959;
11, nephelinite, dyke, Gardiner complex, GM29912; 12, mafic
phonolite, dyke, Gardiner complex, GM29902; 13, urtite, dyke, Gardiner complex, GM29938/2; 14, carbonatite, dyke, Gardiner complex (Frisch & Keusen 1977, Table 19, column 18). LOI, loss on ignition. * About 95% CaCO 3.
& Watt 1985). Following Brooks & Nielsen (1982a), the picrites are assumed to be derived from greater depth in the mantle and to have provoked the melting of more depleted upper mantle which gave rise to the less enriched Fe-Ti tholeiites (Fig. 14). The tholeiite magmas intruded the continental crust along the IML and in the areas S of Kangerdlugssuaq and N of Hold with Hope where spreading and ocean-floor formation took place. Between Kangerdlugssuaq and Hold with Hope the tholeiites occasionally penetrated the continental crust along the IML. For some reason the crust between Kangerdlugssuaq and Hold with Hope resisted spreading, but large amounts of tholeiitic liquid were probably
present at depth along the IML. As spreading continued both S of Kangerdlugssuaq and N of Hold with Hope the slowly east-moving spreading axis attempted to take a short cut between Kangerdlugssuaq and Hold with Hope by rifting E of the IML (Larsen & Watt 1985). Outpouring of the voluminous tholeiitic plateau basalts in E Greenland was related to this rifting. Inland alkaline magmatism of group A Nephelinites and mela-nephelinites are suggested to be the parental liquids to the Gardiner complex, the alkaline lavas in Prinsen af Wales Bjerge and the nunataq zone at 74~ (group A).
506
T. F. D. Nielsen
TABLE2. Basic dyke-swarms of group B and tholeiite compositions 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
M a j o r elements (wt.~)
SiO2 AI203 Fe203 FeO MgO CaO Na20 K20 MnO TiO2 LOI
45.42 48.21 14.69 15.23 5.96 7.49] 7.12 3,69J 7.51 5.53 9.54 7.58 2.41 3.83 1.05 1.50 0.21 0.19 3.15 2.55 0.53 0.52 2.09 2.73
Total
99.68 99.05 99.84 97.71
P205
48.02 45.04 43.10 14.85 15.34 10.20 9.45 11,41 8.00 3.54 9.34 5.12 8.18 10.99 8.92 13.09 1.54 2.88 2.25 1,24 2.16 2.07 0.15 0.22 0.22 1.65 3,90 4.07 0.16 0.78 1.35 2.45 1.94 3.66
42.20 59.39 44.64 38.88 48.33 14.16 17.05 9.35 14.35 16.92 0.64 2,44 4.27 3,46 1.95 11,53 4.32 4.56 7.06 8.96 5.18 1,80 10.70 4.26 5.20 10.37 3.71 15.02 10.91 8.88 3.06 5.59 2.54 8.50 4.76 1.58 2.85 1,22 2.13 0.91 0.22 0.17 0,12 0.28 0.23 3.99 1.12 1.41 2.33 2.45 0.97 0.37 0,52 1.16 0.68 4.94 1.18 5,68 6.16 0.24
52.58 45.11 47.70 45.10 47.06 47.33 48.26 17.53 12.63 16.09 7.80 10.72 13.38 15.34 1.70 6.39 1.85 3,16 7,59 4,89 5.08 6.54 5.04 10.22 8.95 5.64 8.62 4.30 3.05 7,67 6,07 20.32 10.80 6.31 8,07 5.22 11.32 10.76 7.44 9.94 10.79 12,88 5.65 2.80 2.47 1.02 2.00 2.35 2.08 3.33 2.41 0.47 0.18 0.38 0.26 0.24 0.16 0.10 0.13 0.18 0.16 0.18 0.15 1.96 2.89 3.28 1.83 2.12 2.64 1.35 0.49 0.73 0.29 0.16 0.31 0.30 0.30 1.94 3.23 1.44 3.27 3.09 2.15 1.03
99.73 98.64 99.99 100.03 99.48 99.51 100.45 100.32 100.77 99.41 100.31
99.20 98.94
C IP W norms
Q or ab an lc ne ac ns di hy ol In mt il ap
. . . . . 5.12 6.21 8.85 7.51 13.30 12.23 9.34 17.02 20.39 32.41 13.39 22.06 6.59 15.00 47.90 26.17 19.94 30.69 23.48 11.62 20.24 13.16 .
.
--
.
--
.
--
.
.
1.81
.
.
.
6.74
.
.
5.90
.
.
.
.
.
2.00 4.84 1.20
.
1.85 3.21 0.40
.
2.27 7,72 1.92
.
.
2.06 7.73 3.13
--
11.77
2.51 7.68 --
51.87 . 10.54
.
. . . . . . . . 14.51 11.75 19.52 14.29 36.01 20.52 5.79 0.48 15.31 ---14.49 14.22 8.13 13.15 9.31 10.69 2.39 5.98 1.23
. 7.67 1.13 11.18
.
2.28 7.58 2.25
.
3.58 2.15 0.86
1.71 2.85 1.28
.
.
---10.60 36.05 4.18 1.39 22.36 . . 10.62 7.17 -4,75 2.89
. 5.88 30.47 22.12 . 5.13 . . 14.42 . 12.53 . 2.83 4.65 1.61
. 19.68 34.95 12.64 . . 6.97 . . . . 8.25 . 8.61 . . 2,46 3.72 1.16
. . . . 14.24 2.78 1.06 8.36 20.90 8.63 14.78 31.43 16.17 . . . 8.30 . . . . . . . . . 29.82 16.50 15.74 15.12 23.09 11.83 3.39 25.17 . . . 2.07 2.28 2.26 5.49 6.23 3.48 1,69 0.67 0.37
. 1.95 4.82 24.49 . . . . . . . 13.29 16.39 23.13 . . 2.47 4,14 0.37
1.60 1.46 20.48 18.05 25.97 32.62 .
22.48 25.87 19.46 10.60 !.25 6.67 2.57 5.17 0.72
1.74 2.63 0.38
1, Alkaline basalt, Alk-1 dyke, Kraemers O, Kangerdlugssuaq (Nielsen 1976, Table 10, column 3); 2, hawaiite, Alk-1 dyke, H~engefjeldet, Kangerdlugssuaq (Nielsen 1976, Table 4, column 6); 3, basalt, UPLS, Hold with Hope (Upton et al. 1984a, Table 8, column 2);
4, trachyandesite, UPLS, Hold with Hope (Upton et al. 1984a, Table 8, column 8);
5, trachybasalt, Alk-2 dyke,
Kangerdlugssuaq (Nielsen 1976, Table l l b , column 3); 6, trachyandesite, Alk-2 dyke, Kangerdlugssuaq (Nielsen 1976, Table l l a , column 8); 7, oversaturated trachyandesite, Alk-2 dyke, Kangerdlugssuaq (Nielsen 1976, Table l l b , column 14); 8, alkaline basalt ('lamprophyre'), Tugtilik (Rucklidge et al. 1980, Table 2, column 3); 9, nephelinite, Tugtilik (Rucklidge et al. 1980, Table 2, column 7); 10, hawaiite (?), Lilloise intrusion (Brown 1973, Table 1, column 7); 11, diorite (hybrid), Lilloise intrusion (Brown 1973, Table 1, column 4); 12, trachyandesite, Werner Bjerge complex (Bearth 1959, W W 159); 13, potassic tholeiite dyke, Werner Bjerge complex (Bearth 1959, W W 120); 14, tholeiitic picrite lava, Miki Fjord (Brooks & Nielsen 1982b, Table 4, column 1); 15, olivine tholeiite lava, Miki Fjord (Brooks & Nielsen 1982b, Table 4, column 4); 16, tholeiitic basalt, average Scoresby Sund area lavas (Brooks et al. 1976, Table 1, column 1); 17, MORB-like tholeiite, Thol-2 dyke, Kraemers O, Kangerdlugssuaq (Brooks & Nielsen 1978, Table la, column A). LOI, loss on ignition.
On the basis of data from the Gardiner complex and the nunataq zone at 74~ (Nielsen & Buchardt 1985) the less alkaline basanites and hawaiites with higher Sri (Fig. 13) are believed to have reacted with the Archaean basement during ascent. This is supported by Evans & Brown (1981), who found Nd and Sr isotope evidence for contamination with gneisses. Uncontaminated nephelinite liquids from both the nunataq zone and the Gardiner complex have Sri values of 0.7036-0.7038. Similar ratios are found in the lowest tholeiite lavas at the IML, and as suggested by Nielsen & Buchardt (1985) the LILE-rich tholeiitic picrites and the mela-nephelinites of the inland area may have the same mantle source.
The picrites could represent larger degrees of melting from the common source, or the melanephelinites could be high-pressure differentiates of the tholeiitic picrites (Fig. 14). The inland alkaline magmatism is accordingly regarded as a 'flank magmatism' to the tholeiitic activity along the IML. The alkaline dyke-swarms and lavas of group B The evolution from tholeiitic to alkaline compositions in the dyke-swarms S of Kangerdlugssuaq strongly suggests a genetic link between the tholeiitic and alkaline swarms. Clarke et al. (1983) described the origin of camptonitic and
Tertiary magmatism in E Greenland
507
FIG. I Y Summary of initial Sr isotope ratios Sri from various groups of alkaline and tholeiitic rocks from quoted references: GC, Gardiner complex; NZ, nunataq zone at 74~ US, undersaturated syenites; OS, oversaturated syenites. Unpublished data for lower lavas from P. M. Holm (personal communication), for plateau basalts from S of 71~ from Carter et al. (1979) and new analyses from the 'Astrophyllite Bay' complex. Grey field, Sri range of Icelandic tholeiites listed by Nielsen & Buchardt (1985). Pillow complexes: S, salic; M, mafic.
FIG. 14. Simplified petrogenetic scheme for the Tertiary magmatism in E Greenland. Only observed evolutionary trends are included. For explanation see the text. Roman numerals refer to Table 5 where occurrences of and references to each trend are given. Boxes indicate storage magma chambers and the possibility for interaction with host rocks (assimilation etc.). A box in a box indicates enriched-magma-chamber processes (O'Hara & Mathews 1981).
508
T. F . D . N i e l s e n
less alkaline dykes formed under very similar conditions on Ubekendt Ejland, W Greenland, during the opening of Baftin Bay. They suggested that the camptonites etc. formed over a period of about 25 Ma in the feeder system of the preceding picritic to basaltic tholeiite magmatism by replenishment of magma chambers as described by O'Hara & Mathews (1981). Their calculation suggested that even very alkaline liquids (monchiquites) could develop from the tholeiitic picrites. A similar process is suggested here for the formation of the dykes and lavas of group B. The genetic relationship was tentatively suggested by Vincent (1953). Upton et al. (1984a), however, suggested that the UPLS lavas (approximately Alk-1, potassic tholeiites) represent a re-initiation of melting from an enriched-mantle source along the flank of the tholeiitic activity at the developing spreading ridge to the E. This process could also lead to production of potassic tholeiites and alkaline lavas, but it would need increasing heat transfer to the upper part of the mantle over a period of 25 Ma while the area was removed further and further from the spreading axis. The process described by Clarke et al. (1983) has the advantage of not requiring a new energy source and demands only a slow cooling of the parental tholeiite liquids in the hot crust. The present author therefore considers the suggested process to agree better with the geodynamic situation in E Greenland during the continental break-up.
The intrusive complexes of group C (syenites) A spatial, chronological and possibly genetic relationship between the dyke-swarms of group B and the syenite complexes is apparent from the description in the previous sections. The exact nature of the relationship is not known, but new observations and interpretation from the 'Kangerdlugssuaq complex' may be applicable to other syenite complexes in E Greenland. As noted above, at least the eastern part of the Kangerdlugssuaq intrusion including 'Astrophyllite Bay' is underlain by a large kaersutite gabbro. The gabbros are described by Brooks & Platt (1975), who concluded that they crystallized from an undersaturated trachybasalt magma (group B magma) at a depth of 5-10 kin, which would have been at or just below the present level of exposure. The 'diorites' in 'Astrophyllite Bay' (Table 3) are accordingly believed to represent the top or marginal parts of that magma chamber. The anatectic melts (syenite, s.1.) at the margin of the diorites are more alkaline than expected (Fig. 5(a)), possibly owing to the introduction of alkalis by fluids from the diorite magma, but are
very similar to the quartz-nordmarkites of the syenite centres. Repeated assimilation of the often metasomatized basement by the magmas in the upper part of a trachybasalt magma chamber below the Kangerdlugssuaq intrusion produced saturated trachyte magma which crystallized as the nordmarkites of the intrusion. They crystallized downwards into the magma chamber, and further stoping and assimilation of increasingly alkaline roof rocks by the fractionating trachybasalt magma in deeper parts of the magma chamber resulted in the formation of pulaskites. The magma chamber progressively became shielded from contamination by the crustal gneisses. Similar processes were suggested by Thompson (1982) to have operated in the British Tertiary volcanic province. The low Sr~ (Pankhurst et al. 1976) and the high alkalinity of the foyaites in the centre of the complex suggest that they formed from uncontaminated undersaturated trachybasalt magma in the more central parts of the magma chamber. The model satisfies the gradual decrease in Sri towards the centre of the intrusion (Pankhurst et al. 1976), the increasing alkalinity and the lack of internal contacts but does not exclude intrusive relations in parts of the intrusion (see above). The model resembles that described by Brooks & Gill (1982). This model is also believed to apply to many of the syenite complexes in E Greenland owing to the spatial relationship between syenites and magmas of group B and because all the syenite complexes, in the areas that have been investigated, are associated with large positive magnetic anomalies which are believed to be caused by major marie intrusions at depth (H. C. Larsen, personal communication). Mixing of the salic magmas and the trachybasalt magmas during replenishment processes was often not complete and 'mechanical' mixtures intruded as the pillow dykes and complexes observed in most syenite complexes or as more 'mixed' hybrids, e.g. the hybrids of Kontaktbjerg. An independent group of high-silica rocks in the syenite centres includes granites, rhyolites and quartz porphyries (SiO2 ~ 72%-75~) (Table 4; Fig. 15). They form sub-volcanic complexes in the syenite centres, constitute the matrix in some of the breccias surrounding the syenite centres and occur as glassy alkali rhyolite dykes (Brooks & Rucklidge 1976). The large chemical differences suggest that the high-silica rocks and the syenites have different sources, although they are genetically and spatially related. The granites and rhyolites are compositionally close to anatectic melts (Fig. 15) and apparently
Tertiary magmatism in E Greenland
509
TABLE3. Syenites of group C 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
MgO CaO Na20 K20 MnO TiO 2 P2Os LOI
54.30 23.97 1.66 0.81 0.26 1.73 9.92 6.30 0.09 0.40 0.18 0.73
55.19 21.90 1.29 1.19 0.49 2.28 8.11 6.80 0.09 0.84 0.12 1.63
57.41 21.27 2.69 0.32 0.17 0.54 11.65 5.18 0.19 0.17 0.02 0.68
60.66 19.46 1.50 1.44 0.60 1.28 7.15 5.98 0.19 0.89 0.15 0.70
65.40 17.04 1.46 1.67 0.58 1.02 6.43 5.17 0.10 0.70 0.16 0.21
61.69 16.93 1.85 3.91 0.53 2.08 5.48 5.64 0.29 0.58 0.13 0.58
63.77 15.66 2.47 3.81 0.24 1.45 6.58 4.39 0.11 0.60 0.08 0.66
66.36 16.37 1.29 2.08 0.50 0.59 6.33 5.16 0.14 0.70 0.18 0.37
64.83 16.64 1.39 1.95 0.35 1.03 6.07 5.98 -0.54 0.09 0.47
60.78 65.14 17.61 17.09 1.41 2.58 3.74 0.81 0.30 0.23 1.68 0.83 7.84 6.02 4.86 5.89 0.18 0.06 0.44 0,38 0.09 0.05 0.73 0.16
58.52 16.28 2.43 4.89 2.31 3.94 5.09 2.77 0.14 1.74 0.64 0.93
60.26 15.88 2.39 4.31 1.56 3.23 5.63 3.36 0.21 1.34 0.43 0.56
52.91 72.41 1 5 . 1 1 14.51 4.85 0.64 5.92 0.53 3.37 0.35 6.48 0.91 4.50 4.66 1.79 4.63 0.16 0.04 2.62 0.09 0.94 0.03 0.83 0.26
Total
100.35
99.93
100.29
100.00
99.94
99.69
99.82
100.07
99.34
99.66
99.24
99.63
99.16
99.48
99.06
t.94 33.34 46.47 4.94 . . . 3.91 . 4.53 . . . 2.68 1.10 0.31
4.95 25.95 55.67 0.23 .
7.64 30.50 53.32 0.83 . 0.24 0.78
4.34 35.34 49.67 2.47
-28.72 48.50 -8.14 2.48 6.78 . -0.46
5.94 6.63 5.84 4.55 34.81 16.37 19.86 10.58 50.94 43.07 47.64 38.08 2.21 13.39 8.14 15.74 . . . . . . . . . . 1,23 1.62 4.21 8.23 . . . . -9.37 5.91 7.28 . . . . . 1.40 . . . . 1,70 3.52 3.47 7.03 0.72 3.25 2.54 4.98 0,12 1.48 1.00 2.18
25.03 27.36 39.43 4.32
M a j o r elements (wt.%)
SiOE AlzO3 Fe:O3
FeO
CIP W norms
Q or ab an
ne ac
di wo
. . . 37.26 40.19 18.50 21.23 2.28 3.27 35.44 25.68 --1.38 2.63 1.37 1.62
hy
.
ol hm mt
--
il ap
0.47 1.73 0.75 0,43
.
.
-0,12 1.69 1.60 0.28
. 30.61 28.19 -28.38 7.78 2.14 -.
0.01 --0.32 0,05
5.59 35.33 30.54 46.29 54.39 2.21 2.27 4.70 . . . . 2.33 1.23 -0.08 1.78 0.51 . 0.08 0.04 -2.06 1.70 1.33 0.29 0.39
5.62 .
. 1.83 .
.
. 3.58 1.14 0.19
. 2.97 . . 1.75 1.33 0.43
-2.17 . 1.22 . 2.02 1.18 0.28
0.80 0.84 0.2[
-1.24
0.93 0.17 0.07
1, Foyaite, Kangerdlugssuaq intrusion (Deer 1976, Table 16, column 5); 2, nepheline syenite, Werner Bjerge complex (Bearth 1959, Table 5); 3, nepheline syenite, Borgtinderne syenite (Brown et al. 1978, Table 4, column 1); 4, pulaskite, Kangerdlugssuaq intrusion (Deer 1976, Table 16, column 4); 5, nordmarkite, Kangerdlugssuaq intrusion (Deer 1976, Table 16, column 2); 6, nordmarkite, Kap Deichmann (Deer et al. 1984, Table 7, column 1); 7, nordmarkite, Kap Boswell (Deer et al. 1984, Table 2, column 3); 8, quartz-nordmarkite, Kangerdlugssuaq intrusion (Deer 1976, Table 16, column 1); 9, quartz-nordmarkite, Bagn~esset syenite complex (Larsen 1982, Table 6.1, column 15); 10, syenite, Lilloise intrusion (Brown 1973, Table 1, column 5); 11, rheomorphic basement, 'Astrophyllite Bay' complex, GM27110; 122 hybrid, Kialineq district (Brooks 1977, Table 1, column 2); 13, hybrid, 'Astrophyllite Bay' complex, GM27112; 14, hybrid, 'Astrophyllite Bay' complex, GM27108; 15, basement, granitic gneiss, 'Astrophyllite Bay' complex, GM27111.
LOI, loss on ignition.
have avoided the introduction of alkalis mentioned above. Variation in vapour pressure and composition could account for some of the compositional variation in Fig. 15. It thus appears that the geochemical range from undersaturated syenites through saturated syenites to granites includes magmas derived from the undersaturated group B magmas ('trachybasalts'), which may be contaminated with crustal material to variable extents, and anatectic melts that are 'contaminated' to variable extents by fluids from the undersaturated magmas as well as mixtures of these liquids. The magmatic environment of the syenite centres and the chemistry of the rhyolites and granites (Fig. 15) recalls those of the mineralized granites of porphyry Mo deposits in Colorado as described by White et al. (1981). The 'Flammefjeld' deposit in the Kangerdlugssuaq intrusion
and the Malmbjerg (or Erzberg) deposit in Werner Bjerge have previously been compared with the Mo deposits in Colorado (White et al. 1981; Geyti & Thomassen 1985). They lend support to the suggestions of Taylor (1979) regarding the occurrence of quartz-porphyry-type deposits both at convergent-plate margins and in continental-rift environments, which in the case of E Greenland developed into continental separation.
The petrogenetic models and final remarks The models for the evolutionary trends of the different groups of alkaline rocks in E Greenland and their relationship to one another are briefly described above. All the relationships are summarized in Fig. 14.
T . F . D . Nielsen
5 IO
TABLE 4. Rhyolites and granites of group C with comparisons 1
2
3
4
5
6
7
8
9
10
11
12
13
74.37 76.07 13.22 13.14 0.12 "~ 1.08 1.05 J 0.61 0.10 1.44 0.47 4.87 4.36 3.97 4.46 0.07 0.02 -0.06 -0.01 0.24 .
75.70 12.70 0.47 0.57 0.37 1.07 3.10 5.60 -0.56 -. .
74.08 13.20 0.87 0.34 0.19 0.61 4.00 4.80 -0.38 -.
75.20 12.20 0.53 0.50 0.10 0.75 2.32 5.80 -0.15 --
M a j o r e l e m e n t s (wt.%)
SiO 2 A120 3 Fe203 FeO MgO CaO Na20 K20 MnO TiO2 P2Os LOI
72.45 12.77 1.80 1.56 0.09 0.56 2.66 7.01 0.12 0.40 0.07 0.23
73.38 11.64 1.95 1.48 0.00 0.51 4.52 5.30 0.09 0.32 0.02 0.60
76.3 11.4 1.94 1.13 0.00 0.08 4.76 4.10 0.09 0.17 0.00 0.15
76.02 12.98 1.11 0.61 0.05 0.45 3.13 5.52 0.02 0.21 0.02 0.25
76.72 12.40 0.83 0.68 0.02 0.48 3.49 5.51 0.00 0.03 0.02 0.18
72.06 13.36 2.09 1.74 0.15 0.87 4.67 4.26 0.15 0.33 0.03 0.26
72.7 14.3 1.05 0.53 0.19 0.76 4.81 4.63 0.05 0.26 0.04 0.56
75.5 13.3 0.67 0.00 0.21 0.45 2.96 4.94 0.08 0.18 0.02 1.77
Total
99.72
99.86
100.00
100.37
100.36
99.91
99.88
100.08
99.96
99.77
100.14
98.47
97.55
an
28.61 41.43 22.51 2.20
27.38 31.32 30.62 --
33.2 24.3 35.7 --
35.70 32.63 26.48 2.10
34.17 32.57 29.53 1.90
26.35 25.18 39.51 2.91
25.3 27.4 40.7 3.5
38.2 29.2 25.1 2.1
26.96 23.46 41.21 2.49
31.87 26.42 39.39 2.28
33.63 33.10 26.23 4.20
30.63 28.37 33.84 3.03
37.38 34.28 19.63 3.72
ne
--
4.0
.
0.3 1.1
-0.12
0.31 0.40
1.03 0.93
-0.5
-0.5
3.89 1.54
-0.8 0.3 --.
0.13 1.42 0.40 0.05 1.09 . .
-1.20 0.06 --.
-3.03 0.63 0.07 --
-0.9 0.5 0.1 --
0.7 --0.1 2.3 0.2
. 0.17 ---.
CIP W norms
Q or ab
ac
ns di hy ol
hm
mt il ap
--0.10 1.11
--
5.64 0.28 2.14 1.22
- -
- -
--
--
2.61 0.76 0.16
-0.61 0.05
C
- -
- -
ru
- -
- -
.
.
.
.
.
.
.
.
. . 1.32 .
. 0.21 0.11 0.02 0.31 . .
. .
.
.
-. ----.
--
--
---0.32
---0.74
.
.
1, Acid matrix of breccia, K r a e m e r s O (Brooks & Nielsen 1982b, Table 6, column 6); 2, aphyric rhyolite dyke, K r a e m e r s O (Brooks & Rucklidge 1976, T a b l e 6, c o l u m n 2); 3, porphyritic rhyolite dyke, Uttendal Plateau (Brooks & Rucklidge 1976, T a b l e 6, column 8); 4 acid matrix of Sortskaer breccia, Kangerdlugssuaq (Deer e t al. 1984, Table 12, c o l u m n 4); 5, acid matrix of breccia, K o n t a k t b j e r g (Deer e t al. 1984, Table 14, c o l u m n 7); 6, microgranite, Barberkniven (Deer e t al. 1984, Table 6, column 9); 7, biotite granite, Kangerdlugssuaq (Brooks & Nielsen 1982b, T a b l e 8, column 6); 8, quartz porphyry, intrusive stock, ' F l a m m e f j e l d ' , Kangerdlugssuaq intrusion (Brooks & Nielsen 1982b, Table 6, column 9); 9, alkali granite, W e r n e r Bjerge complex (Bearth 1959, Table 3, WB6); 10, felsite, Kap Broer Ruys c o m p l e x ( U p t o n e t al. 1984b, Table V, column 1); 11, biotite-porphyry, C l i m a x stock, ore related (White e t al. 1981, Table 3, c o l u m n 3); 12, rhyolite porphyry dyke, C l i m a x (White e t al. 1981, Table 3, c o l u m n 4); 13, P r i m o s porphyry, Henderson deposit (White e t al. 1981, Table 3, column 8). LOI, loss on ignition.
As a generalization there are three source regions: (a) a deeper less depleted upper mantle, (b) a less deep but depleted upper mantle and (c) the continental crust. The magmatism was initiated in the deeper upper mantle where tholeiitic picrites and mela-nephelinites formed, the tholeiitic picrites below the future IML and the nephelinites on the flanks of the melting anomaly. The liquids could then (a) rise straight through the crust or (b) be subjected to fractionation, contamination etc. in magma chambers in the upper mantle and/or the crust. If the magma chambers were sufficiently large and the regional geothermal gradient sufficiently high, the crystallizing liquids in the magma chambers could
provoke melting of the country rocks (secondary magmas), e.g. the depleted tholeiites and granites indicated with the black arrows in Fig. 14. The liquids stored in the magma chambers in the upper mantle or crust could evolve, owing to processes in replenished magma chambers, to new magma types (e.g. alkaline magmas of group B). Coexisting magma types could also mix to form, for example, moderately Ti-enriched tholeiites and hybrids of syenite complexes. Nearly all magma types could be developed and many have been identified (Table 5; Fig. 14). On the basis of the magmatic chronology and the models described above, the magmatic evolution during continental break-up in E Green-
Tertiary magmatism in E Greenland
5II
TABLE 5. E v o l u t i o n a r y t r e n d s Trend
Rock types
Occurrence
Reference
I
Nephelinites to carbonatites
Gardiner complex
Frisch & Keusen 1977; Nielsen 1980, 1981 ; Nielsen & Buchardt 1985
II
Mela-nephelinites
Kangerdlugssuaq district Nunataq zone, 74~ Gardiner complex
Brooks & Rucklidge 1974 Brooks et al. 1979 Nielsen 1981 ; Nielsen & Buchardt 1985
III Nephelinites to nephelinehawaiites
Nunataq zone, 74~ Gardiner complex
Brooks et al. 1979 Frisch & Keusen 1977; Nielsen, unpublished data
IV Basanitesto hawaiites
Prinsen af Wales Bjerge Nunataq zone, 74~
Anwar 1955; Evans & Brown 1981 Fawcett et al. 1982; Brooks et al. 1979
V
Tholeiitic picrites
Lower lavas, Kangerdlugssuaq
Nielsen et al. 1981 ; Brooks & Nielsen 1982a
VI
Picrite-trend basalts and andesites
Lower lavas, Kangerdlugssuaq
Nielsen et al. 1981 ; Brooks & Nielsen 1982a
VII LILE-enriched tholeiites
Lower lavas, Kangerdlugssuaq
Brooks & Nielsen 1982a
VIII Fe-Ti and MORB tholeiites
Plateau basalts, E Greenland
Brooks et al. 1976; Brooks & Nielsen 1982a, b; Fawcett et al. 1982; Upton et al. 1984a; Larsen & Watt 1985
IX
Gabbro intrusions Contaminated lavas
Kangerdlugssuaq Scoresby Sund area
Brooks & Nielsen 1982b Larsen & Watt 1985
X
Re-mobilized basement
Miki Fjord macrodykes
Not described yet
XI
Trachybasalts to syenites and phonolites Kaersutite gabbros Syenites
Kangerdlugssuaq area intrusions Kangerdlugssuaq dyke-swarms Kangerdlugssuaq intrusions
Deer et al. 1984; present work Brooks & Platt 1975 Wager 1965; Kempe & Deer 1970; Pankhurst et al. 1976
XII Saturated trachyandesites
Kangerdlugssuaq dyke-swarms
Vincent 1953; Brooks & Platt 1975; Nielsen 1978
XIII Pillow complexes
Kialineq centre Kangerdlugssuaq area
Brooks 1977 Brooks & Nielsen 1982b
XIV Quartz porphyries Alkali granites Comendites
Flamme0eld Kangerdlugssuaq area, Werner Bjerge Kangerdlugssuaq area
Brooks & Nielsen 1982b; Geyti & Thomassen 1985 Bearth 1959; Deer et al. 1984 Brooks & Rucklidge 1976
land can be s u m m a r i z e d as follows" (a) upwelling of tholeiitic picrites and heating of the u p p e r mantle; (b) formation of plateau basalts, alkaline flank activity (group A) and heating of the crust; (c) continental break-up; (d) trapping of tholeiite l i q u i d s in the u p p e r m a n t l e a n d crust and the formation of alkaline m a g m a s (group B) and syenite complexes (group C). ACKNOWLEDGMENTS: This review would not have been possible without fruitful and inspiring cooperation and
discussion with a large number of colleagues over the last decade. Not all can be mentioned, but special thanks are due to C.K. Brooks, J. Gittins, J.C. Rucklidge, M. P. Gorton, P. M. Holm, L. C. Larsen, H. C. Larsen, H. C. Sch6nwandt, R. C. O. Gill, B. G. J. Upton and especially L. M. Larsen for much encouragement and critical reading of the manuscript, and to Nordisk Mineselskab A/S for the use of their internal reports. Tove Buus-Pedersen typed the manuscript, and the figures were prepared by Bente Thomas. The publication of this paper is authorized by the Director of Gronlands Geologiske Undersogelse.
T. F. D. Nielsen
512
QZ
i~: : .......
GRANITES, COLORADO (ORE R E L A T E D : I,A)
GRANITES LAND 'G
MINIMUM M E L T I N G AT WITH 0, 1 AND 2 * / , F (MANNING, 1981)
9
AVERAGE S Y E N I T E S LUGSSUAQ ( DEER 1 9 8 4 TABLE 17 )
O
SYENITE MARGINS, BAY C O M P L E X . ,~
50
50
'~ 1Kb
2
"
KANGERDET AL
ASTROPHYLLITE
GRANITES, WERNER ( ORE R E L A T E D )
BASEMENT, BAY.
ASTROPHYLLITE
AVERAGE SYENITES, KAP EDVARD H O L M C O M P L E X (, DEER ET A L 1 9 8 4 TABLE 16)
V ~2
V
AB
v NORMATIVE
ALKP, LI BJERGE
/~,/,~
3Kb :
GREEN-
PERALKALINE RHYOLITES EAST GREENLAND ( BROOKS AND RUCKLIDGE, 1974)
1KB
KAP BROER RUYS (GRANOPHYRES)
( S L ), EAST
50
v
v
OR
PROCENT
FIG. 15. Residual Qz-Ab-Or-H20 system (Tuttle & Bowen 1958; Carmichael et al. 1974) showing granites and rhyolites in the Tertiary of E Greenland in comparison with granitic and rhyolitic rocks related to quartz porphyry molybdenum deposits in Colorado (White et al. 1981).
References ABBOTT, D, & DEER, W. A. 1972. Geological investigations in East Greenland. Part X. The gabbro cumulates of the Kap Edvard Holm lower layered series. Meddr. Gronland, 190 (6). ANWAR, Y. M. 1955. Geological investigations in East Greenland. Part V. The petrography of the Prinsen af Wales Bjerge lavas. Meddr. Gronland, 135 (1). BEARTH, P. 1959. On the alkali massif of the Werner Bjerge in East Greenland. Meddr. Gronland, 153 (4). BECKINSALE, R. D., BROOKS, C. K. & REX, D. C. 1970. K-Ar ages for the Tertiary of East Greenland. Bull. geol Soc. Denmark, 20, 28-38. BENGAARD, H. J. & HENRIKSEN, N. 1984. Scoresby Sund, 1:500 000 Map, Sheet 12, Grenlands Geologiske Undersogelse, Copenhagen. BROOKS, C. K. 1973. Rifting and doming in southern East Greenland. Nature, Lond., Phys. Sci. 244, 235. 1977. Example of magma mixing from the Kialineq district of East Greenland. Bull. geol. Soc. Denmark, 26, 77-83.
--
& GILL, R. C. O. 1982. Compositional variation in the pyroxenes and amphiboles of the Kangerdlugssuaq intrusion, East Greenland: further evidence for the crustal contamination of syenite magma. Mineral. Mag. 45, 1-9. & NIELSEN, T. F. D. 1978. Early stages in the differentiation of the Skaergaard magma as revealed by a closely related suite of dike rocks. Lithos, 11, 1-14. -& -1982a. The E Greenland continental margin: a transition between oceanic and continental magmatism. J. geol. Soc. Lond. 139, 26575. -& -1982b. The Phanerozoic development of the Kangerdlugssuaq area, East Greenland. Meddr. Gronland Geosci. 9. --, NIELSEN, T. F. D. & PETERSEN, T. S. 1976. The Blosseville coast basalts of East Greenland: composition and temporal variation. Contrib. Mineral. Petrol. 58, 279-92. - - & PLATT,R. G. 1975. Kaersutite-bearing gabbroic -
-
Tertiary magmatism & E Greenland inclusions and the late dike swarm of Kangerdlugssuaq, East Greenland. Mineral. Mag. 40, 259-83. - - & RUCKLIDGE,J. C. 1973. A-Tertiary lamprophyre dike with high pressure xenoliths and megacrysts from Widemanns Fjord, East Greenland. Contrib. Mineral. Petrol. 42, 197-212. -& -1974. Strongly undersaturated Tertiary volcanic rocks from the Kangerdlugssuaq area, East Greenland. Lithos, 7, 239-48. & -1976. Tertiary peralkaline rhyolite dikes from the Skaergaard area, Kangerdlugssuaq, East Greenland. Meddr. Gronland, 197 (3). , PEDERSEN, A. K., LARSEN, L. M. & ENGELL, J. 1982. The mineralogy of the Werner Bjerge Complex, East Greenland. Meddr. Gronland Geosci. 7. & REX, D. C. 1979. The petrology and age of alkaline mafic lavas from the nunatak zone of central East Greenland. Bull. Gronlands geol. Unders. 133. BROWN, P. E. 1973. A layered plutonic complex of alkali basalt parentage: the Lilloise intrusion, East Greenland. J. geol. Soc. Lond. 129, 405-18. & FARMER, D. G. 1972. Size-graded layering in the Imilik gabbro, East Greenland. Geol. Mag. 108, 465-575. - - , VAN BREEMEN, O., NOBLE, R. H. & MACINTYRE, R. M. 1977. Mid-Tertiary igneous activity in East Greenland, the Kialineq complex. Contrib. Mineral. Petrol. 64, 109-22. - - , BROWN, R. D., CHAMBERS,A. D. & SOPER, N. J. 1978. Fractionation and assimilation in the Borgtinderne syenite, East Greenland. Contrib. Mineral. Petrol. 67, 25-34. BULLARD, E. C., EVERETT, J. E. & SMITH, A. G. 1965. The fit of the continents around the North Atlantic. Phil. Trans. R. Soc. Lond., Ser. A, 258, 41-51. BURKE, K. & DEWEY, J. F. 1973. Plume-generated triple junctions: key indicators in applying plate tectonics to old rocks. J. Geol. 81,406-33. CARMICHAEL,I. S. E., TURNER, F. J. & VERHOOGEN, J. 1974. Igneous Petrology, 739 pp. McGraw-Hill, New York. CARTER, S. R., EVENSEN, N. M., HAMILTON, P. J. & O'NIONS, R. K. 1979. Basalt magma sources during the opening of the North Atlantic. Nature, Lond. 281, 28-30. CLARKE, D. B., MUECKE, G. K. & PE-PIPER, G. 1983. The lamprophyres of Ubekendt Ejland, West Greenland: Products of renewed partial melting or extreme partial melting? Contrib. Mineral. Petrol. 83, 117-27. DEER, W. A. 1976. Tertiary igneous rocks between Scoresby Sund and Kap Gustav Holm, East Greenland. In: ESCHER, A. & WATT,W. S. (eds) Geology of Greenland, pp. 406-29. Gronlands Geologiske Undersogelse, Copenhagen. & KEMPE, D. R. G. 1976. Geological investigations in East Greenland. Part XI. The minor peripheral intrusions, Kangerdlugssuaq, East Greenland. Meddr. Gronland, 197 (4). - - , KEMPE, D. R. C. & JONES, G. C. 1984. Syenitic and associated intrusions of the Kap Edvard Holm -
-
-
-
-
-
513
region Kangerdlugssuaq, East Greenland. Meddr. Gronland Geosci. 12. ENGELL, J. 1975. The Kap Parry complex, central East Greenland. Rapp. Gronlands geol. Unders. 75, 1036. EVANS, I. B. & BROWN, P. E. 1981. Isotopic evidence from volcanic rocks for mantle heterogeneity beneath East Greenland in the lower Tertiary. Geol. Soc. Lond. Newsl. 10 (2), 39-40. FALLER, A. M. & SOPER, N. J. 1979. Palaeomagnetic evidence for the origin of the coastal flexure and dike swarm in central E. Greenland. J. geol. Soc. Lond. 136, 737-44. FAWCETT, J. J., GITTINS, J., BROOKS, C. K. & RUCKLIDGE, J. C. 1982. Petrology of Tertiary lavas from the western Kangerdlugssuaq area, East Greenland. Mineral. Mag. 45, 211-8. FRISCH, W. & KEUSEN, H.-R. 1977. Gardiner intrusion, an ultramafic complex at Kangerdlugssuaq, East Greenland. Bull. Gronlands. geol. Unders. 122. GEYTI, A. & THOMASSEN,B. (1985) Molybdenum and precious metal mineralisation at Flammefjeld, South-East Greenland. Econ. Geol. 79, 1921-29. GLEADOW, A. J. W. & BROOKS, C. K. 1979. Fission track dating, thermal histories and tectonics of igneous intrusions in East Greenland. Contrib. Mineral. Petrol. 71, 45-60. HALD, N. 1978. Tertiary igneous activity at Giesecke Bjerge, northern East Greenland. Bull. geol. Soc. Denmark, 27 (Spec. Iss), 107-15. HALLER, J. 1956. Geologie der Nunatakker Region von Zentral-OstgrSnland zwischen 72o30 ' und 74o10 ' n. Br. Meddr. Gronland, 154 (1). JOHNSON, G. L., MCMILLAN, N. J. & EGLOFF, J. 1975. East Greenland continental margin. In: YORATH, C. J , PARKER, E. R. & GLASS,O. J. (eds) Canada's continental margins and offshore petroleum exploration. Mem. Can. Soc. petrol. Geol. 4, 205-24. KAPP, H. 1960. Zur Petrologie der subvulkane zwischen Mesters Vig und Antartic Havn (Ost-GrSnland). Meddr. Gronland, 153 (2). KATZ, H. R. 1952. Ein Querschnitt durch die Nunatakzone ostgr6nlands (ca. 74 ~ N.B.). Meddr. Gronland, 144 (8). KEMPE, D. R. C. & DEER, W. A. 1970. Geological investigations in East Greenland. Part IX. The mineralogy of the Kangerdlugssuaq intrusion, East Greenland. Meddr. Gronland, 190 (3). & DEER, W. A. 1976. The petrogenesis of the Kangerdlugssuaq alkaline intrusion, East Greenland. Lithos, 9 (2), 111-23. - - , DEER, W. A. & WAGER, L. R. 1970. Geological investigations in East Greenland. Part VIII. The petrology of the Kangerdlugssuaq intrusion, East Greenland. Meddr. Gronland, 190 (2). KIRCHNER, G. 1964. Die molybdiinlagerstfitte des Erzberges bei Mesters Vig, Ostgr6nland. Berg. Hiittenmi~nn. Monatsh. 109, 162. LARSEN, H. C. 1984. Geology of the East Greenland shelf. In: SVENCER, A. M., HOME, P. C. & BERGLUND, L. T. (eds) Petroleum Geology of the North European Margin, pp. 229-39. Norwegian Petroleum Society, Graham & Trotman, London. LARSEN, L. C. 1982. Bagn~esset syenitkompleks, Kan-
-
514
T. F. D. Nielsen
gerdlugssuaq distriktet, Ost Gronland. Thesis, Institut for Petrologi, Kobenhavns Universitet (unpublished). LARSEN,L. M. & WATT, W. S. 1985. Episodic volcanism during break-up of the North Atlantic: evidence from the East Greenland Plateau basalts. Earth planet. Sci. Lett. 73, 105-16. LAYNE, G. D., RUCKLIDGE, J. C. & BROOKS, C. K. 1982. Astrophyllite from Kangerdlugssuaq, East Greenland. Mineral. Mag. 45, 149-56. MANNING, O. A. C. 1981. The effect of fluorine on liquidus phase relationships in the system Qz-AbOr with excess water at 1 kb. Contrib. Mineral. Petrol. 76 (2), 206-15. MATTHEWS, D. W. 1979. A buried Tertiary pluton in East Greenland? Bull. geol. Soc. Denmark, 28, 1720. MYERS, J. S. 1980. Structure of the coastal dyke swarm and associated plutonic intrusions of East Greenland. Earth planet. Sci. Lett. 46, 407-18. NIELSEN, T. F. D. 1976. Gangsva~rme i Kangerdlugssuaq omrgtdet, Ostgronland. Deres kronologi, petrografi og kemi saint deres relationer til ~bningen af det nordatlantiske bassin i terti~er. Thesis, Institut for Petrologi, Kobenhavns Universitet (unpublished). -1978. The Tertiary dike swarms of the Kangerdlugssuaq area, East Greenland. An example of magmatic development during continental breakup. Contrib. Mineral. Petrol. 67, 63-78. 1979. The occurrence and formation of Tiaegirines in peralkaline syenites. An example from the Tertiary ultramafic alkaline Gardiner complex, East Greenland. Contrib. Mineral. Petrol. 69, 23544. 1980. The petrology of a melilitolite, melteigite, carbonatite and syenite ring dike system in the Gardiner Complex, East Greenland. Lithos, 13, -
-
1 8 1 - 9 7 .
1981. The ultramafic cumulate series, Gardiner Complex, East Greenland. Cumulates in a shallow level magma chamber of a nephelinitic volcano. Contrib. Mineral. Petrol. 76, 60-72. -& BROOKS, C. K. 1981. The E. Greenland rifted continental margin: an examination of the coastal flexure. J. geol. Soe. Lond. 138, 559-68. & BUCHARDT, B. 1985. Oxygen, carbon and strontium isotope compositions, Gardiner Complex and tertiary of East Greenland. Chem. Geol. 53, 207-17. --, SOPER, N. J., BROOKS, C. K., FALLER, A. M., HIGGINS, A. C. & MATTHEWS, D. W. 1981. The pre-basaltic sediments and the Lower Basalts at Kangerdlugssuaq, East Greenland: their stratigraphy, lithology, palaeomagnetism and petrology. Meddr. Gronland Geosci. 6. NOE-NYGAARD, A. 1941. Syenitforekomsten ved Antarctic Havn (Ostgronland). Meddr. dansk Foren. 9 (for 1940), 550-56. 1974. Cenozoic to Recent volcanism in and around the North Atlantic basin. In: NAIRN, A. E. W. & STEHLI, F. G. (eds) The Ocean Basins and Margins, Vol. 2, pp. 391-443. Plenum, New York. -1976. Tertiary igneous rocks between Shannon
-
-
and Scoresby Sund, East Greenland. In: ESCHER, A. & WATT, W. S. (eds) Geology of Greenland, pp. 386-402. Gronlands Geologiske Undersogelse, Copenhagen. O'HARA, M. J. • MATHEWS, R. E. 1981. Geochemical evolution in an advancing, periodically replenished, periodically tapped, continuously fractionated magma chamber. J. geol. Soc. Lond. 138, 23777. PANKHURST,R. J., BECKINSALE,R. D. & BROOKS, C. K. 1976. Strontium and oxygen isotopic evidence relating to the petrogenesis of the Kangerdlugssuaq alkaline intrusion, East Greenland. Contrib. Mineral. Petrol. 54, 17-42. PARR, W. & PUNZENGRUBER, K. 1969. Montangeologischer bericht, Nordmine 1969. Der Forchhammer pluton/Traill-O, Ostgr6nland. Internal Rep., Nordisk Mineselskab A/S, Copenhagen. REX, D. C., GLEDHILL, A. R., BRIDGWATER, D. & MYERS, J. S. 1979a. A Rb-Sr whole rock age of 55 + 7 m.y. from the Nualik plutonic centre, East Greenland. Rapp. Gronlands geol. Unders. 95, 1025. - - - , GLEDHILL, A. R., BROOKS, C. K. & STEENFELT, A. 1979b. Radiometric ages of Tertiary salic intrusions near Kong Oscars Fjord, East Greenland. Rapp. Gronlands geol. Unders. 95, 106-9. RUCKLIDGE, J. C., BROOKS, C. K. & NIELSEN, T. F. D. 1980. Petrology of the coastal dikes at Tugtilik, southern East Greenland. Meddr. Gronland Geosci. 3. SCHAUB, H. S. 1938. Zur Vulkanotektonik der Inseln Traill und Geographical Society (Nordostgr6nland). Meddr. Gronland, 114 (1). 1942. Zur Geologie der Traill Insel (NordostGr6nland). Ecologae geol. Helv. 35 (with an appendix by REINHARD, M. Petrographishe Beschreibung der Gesteine der Traill Insel). SHEPPARD, S. M. F., BROWN, P. E. & CHAMBERS,A. D. 1977. The Lilloise intrusion, East Greenland: hydrogen isotope evidence for the efflux of magmatic water into the contact metamorphic aureole. Contrib. Mineral. Petrol. 63, 129-47. SOPER, N. J.., HIGGINS, A. C., DOWNIE, C., MATTHEWS, D. W. & BROWN, P. E. 1976. Late Cretaceousearly Tertiary stratigraphy of the Kangerdlugssuaq area, East Greenland, and the age of opening of the north-east Atlantic. J. geol. Soc. Lond. 132, 85104. TAYLOR, R. G. 1979. Geology of Tin Deposits. Development of Economic Geology, Vol. 11,543 pp. Elsevier, Amsterdam. THOMPSON, R. N. 1982. Magmatism of the British Tertiary volcanic province. Seott J. Geol. 18, 49107. TUTTLE, O. F. & BOWEN, N. L. 1958. Origin of granite in the light of experimental studies in the system NaAISi3Os-KA1Si3Os-SiOz-H20. Geol. Soc. Am. Mem. 74. TYRRELL, G. W. 1932. The petrography of some Kainozoic igneous rocks and the Cape Parry alkaline complex, East Greenland. Geol. Mag. 69, 520-7. UPTON, B. G. J. EMELEUS,C. H. & BECKINSALE,R. D.
Tertiary magmatism in E Greenland 1984a. Petrology of the northern East Greenland Tertiary flood basalts: evidence from Hold with Hope and Wollaston Forland. J. Petrol. 25, 15184. & MACINTYRE, R. M. 1984b. Myggbukta and Kap Broer Ruys: the most northerly of East Greenland Tertiary centres (?). Mineral. Mag. 48, 323-43. , & HALD, N. 1980. Tertiary volcanism in northern E. Greenland" Gauss Halvo and Hold with Hope. J. geol. Soc. Lond. 137, 491-508. VINCENT, E. m. 1953. Hornblende-lamprophyre dykes of basaltic parentage from the Skaergaard area, East Greenland. Q. J. geol. Soc. Lond. 109, 21-49. WAGER, L. R. 1934. Geological investigations in East Greenland. Part I. General geology from Angmagsalik to Kap Dalton. Meddr. Gronland, (2), 146. 1947. Geological investigations in East Greenland. Part IV. The stratigraphy and tectonics of Knud 105
-
-
515
Rasmussens Land and the Kangerdlugssuaq region. Meddr. Gronland, 134 (5). -1965. The form and internal structure of the alkaline Kangerdlugssuaq intrusion, East Greenland. Mineral. Mag. 34, 487-97. & DEER, W. A. 1938. A dyke swarm and coastal flexure in East Greenland. Geol. Mag. 75, 3946. WATT, W. S., LARSEN, M. L. & Watt, M. Volcanic history of the lower Tertiary plateau basalts in the Scoresby Sund region, East Greenland. Bull. Gronlands geol. Unders., in press. WHITE, W. H., BOOKSTROM, A. A., KAMILLI, R. J., GANSTER, M. W., SMITH, R. P., RANTA,D. E. & STEININGER, R. C. 1981. Character and origin of Climax-type molybdenum deposits. Econ. Geol. (75th Anniversary Vol.), 270-316. WILLIAMS,M. I. 1974. Geology report. In: Westminster East Greenland Expedition 1974. Unpublished report for private circulation. -
-
T. F. D. NIELSEN,Gronlands Geologiske Undersogelse, Oster Voldgade 10, 1350 Copenhagen K, Denmark.
Tertiary and Quaternary volcanism in the Massif Central, France Hilary Downes S U M M A R Y : Continental alkaline volcanics of the French Massif Central range in age from 65 Ma to 3450 years BP and can be divided by geography and age into 20 separate areas. There is no correlation between the age of an area and the geographic locality. Two magma series are present: a silica-undersaturated basanite-tephrite-phonolite series and an alkali basalt-trachyandesite-trachyte-rhyoliteseries, the evolved members of which are silica-saturated. Rare nephelinites are also found. Primitive magmas of both main series may be derived from 2~-20~ partial melting of a mantle enriched with light rare-earth elements with variable amounts of garnet in the residue; nephelinites were formed from a source which was isotopically and chemically distinct. Fractional crystallization of observed phases accounts for the diversity of intermediate and evolved products. Amphibole fractionation in basalts at depth causes the trend towards silica saturation, while alkali feldspar fractionation dominates the final stages of crystallization. Removal of small quantities of sphene and zircon explains trace-element depletions in trachytes, phonolites and rhyolites. Up to 30~ crustal contamination has occurred in the evolved magmas but contamination is generally absent from their basaltic parents. Sr, Nd and Pb isotope data indicate that the most probable contaminant is undepleted meta-sedimentary granulite-facies lower crust. Variations in Sr and Nd isotope ratios also constrain petrogenetic hypotheses such as amphibole fractionation and magma mixing. Mineralogical disequilibrium, compositional banding and emulsificationtextures indicate extensive magma mixing. Some flows are homogeneous hybrid lavas in which only the unusual composition and rare mineral disequilibria indicate the mode of formation.
Introduction The Tertiary and Quaternary alkaline volcanics of the French Massif Central have been discussed previously by Brousse (1971), Wimmenauer (1974) and Maury & Varet (1980). Recent advances in the acquisition and interpretation of petrological and geochemical data from the area are reviewed here in order to discuss four major processes which have produced the diversity of magmas: mantle partial melting, fractional crystallization, crustal contamination and magma mixing. The volcanics extend from Bourgogne-Charollais in the NE to Languedoc in the S (Fig. 1). The basement is composed predominantly of rifted and uplifted Hercynian metamorphic units and granites, underlain by granulite-facies lower crust (Dupuy et al. 1977; Leyreloup et al. 1977; Pin & Vielzeuf 1983; Downes& Leyreloup 1986). Fragments of the lithospheric mantle, i.e. spinel peridotites and pyroxenites, are commonly found in the basic volcanics (Hutchison et al. 1975; Brown et al. 1980). Geophysical evidence indicates no definable Moho beneath the volcanics but a region of anomalous mantle at a depth of 25-30 km (Perrier & Ruegg 1983) where both density and seismic velocities are low. The lithospheric mantle is thinned and an asthenospheric bulge is present (Lucazeau et al. 1984).
Seismic attenuation studies (Souriau et al. 1981) show low QB values indicating partial melting in the top 100 km of the mantle. Volcanic forms include Strombolian cinder cones aligned along fissures (Chaine des Puys), large basaltic plateaux (Aubrac, Deves) and two central volcanoes (Mont-Dore and Cantal). Maars and diatremes are common, indicating interaction of magma with surface water or concentration of volatiles at depth. Hydrothermal activity resulting from high heat flow, is sporadic in both the volcanic areas and the Hercynian terrains (Dupis & Gibert 1977; Lucazeau et al. 1984). Chemically the volcanics are silica saturated (trachyandesites, trachytes and rhyolites), silica undersaturated (alkali basalts, basanites, tephrites and phonolites) or extremely silica undersaturated (nephelinites sensu lato, which may contain leucite, hatiyne, melilite or analcime).
Geochronology Massif Central volcanism has been dated by K Ar and 14Cmethods (Fig. 2) with some additional fission-track and thermoluminescence data. There i s no correlation between geographic locality and age of volcanism in an area. Palaeocene and Eocene activity occurred in Bourgogne, Menat (Vincent et al. 1977) and
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 517-530.
517
518
H. Downes
Hercynian metamorphic complexes and granites
N
~-~
Post-Hercynian sediments
B
Tertiary - Quaternary volcanics
~ogne
l
Major faults
Charollais
§
9
. .
~ .
. ,
. ,
9
o n s
,
*0
:4
Coirons
Abbreviations:CP Chaine des Puys SH Sillon Houllier
Causses
j
Cez Cezallier
~
Scale:0 |
lOOkm I
e~t^
/
.
_.jr
Languedoc
Montpellier
v
,,
/.
MEDITERRANEAN
SEA
FIG. 1. Sketch map of the Massif Central, indicating the areas of volcanism under discussion.
Causses. Dispersed volcanism of lower Miocene age in Forez, Limagne and the Margerides was associated with the Oligocene graben system. Subsequently the central volcano Cantal (112.5 Ma) and the plateau-forming fissure volcanoes of Cbzallier, Velay, Coirons, Aubrac and Deves erupted in the Miocene. The Pliocene saw a continuation of this activity and renewal in the Causses, Chaine de la Sioule and Limagne, with new activity in Escandorgue and the eruption of the second central volcano Mont-Dore. The youngest activity (Recent) is in Vivarais (Berger et al. 1975) and the Chaine des Puys.
Recent geochronological studies have established a stratigraphy for Mont-Dore, including the recognition of two separate volcanoes MontDore sensu stricto and the younger Sancy centre (Besson et al. 1977; Santoire et al. 1977; Baubron & Cantagrel 1980; Mossand et al. 1982; Bourdier & Cantagrel 1983; Cantagrel & Baubron 1983). The ages of poorly known areas, e.g. Forez (Bellon & Hernandez 1979) and Margerides (Bellon & Gibert 1979), have also been determined. Dating of palaeomagnetic events and post-glacial activity has received much attention (Bonifay & Tiercelin 1977; Condomines 1978;
Massif Central, France Bourgogne
519
-
Charollais Menat Sioule
Ch. des Puys Forez Limagne Mont
Dore
Sillon
Houiller
Cezallier Cantal
Margerides Deves Velay Vivarais Coiron
Aubrac Causses Escandorgue Languedoc AGE (MA);
PAL EO OLIG 65 55 38 22
MIOCENE 5
PLIOCENE QUAT 1.8
FIG. 2. Age determinations for Massif Central volcanic areas. (Modified from Maury & Varet 1980.)
Gillot 1978; Hall & York 1978a, b; Deluzarch & Westphal 1981 ; Guerin etal. 1981).
Chemical and petrographic variation The nomenclature of the Massif Central volcanics is confused by local names and usages. Table 1 indicates the current consensus of nomenclature with a comparison of older terms together with phenocryst assemblages and groundmass mineralogies. An alkali-silica diagram (Fig. 3) demonstrates the relationship between nomenclature and composition. Data from Cantal (Downes 1983) are compared with trends from Velay (Berger et al. 1976). Chaine des Puys (Maury et al. 1980) and Mont-Dore (Brousse 1961). Several regions within the province erupted both the silica-saturated and silica-undersaturated magma series. In Mont-Dore and Cantal trachyandesite lavas and pyroclastics are abundant and unusually porphyritic; furthermore, rhyolites, in the form of small intrusions and widespread pumice flows, are confined to these two volcanoes. The differences between these volcanoes, with their central calderas and flank basalt flows, and the uncentralized activity of the other regions may be due to large crustal magma chambers in which
fractionation has occurred en route to the surface, producing crystal-rich flows and extreme differentiates. Magma mixing and crustal contamination are also well documented in the central volcanoes and may be related to the existence of crustal magma chambers. Rare occurrences of leucite basanites have been reported (Tournon & Velde 1971: Velde 1973; Magonthier & Velde 1976; Boudon & de Goer de Hevre 1978; Besson et al. 1979; Boudon & Cantagrel 1981). Other basanites contain ha/iyne (Churlet & Velde 1976) or analcime (Hernandez 1973). Nephelinites and melilitites are known from a few localities (Hernandez 1973; 1976; Boudon & de Goer de Hevre 1978). Systematic differences in the N a 2 0 / K 2 0 ratio are believed to occur between different areas (Maury & Varet 1980), but the data used were drawn from many sources and may therefore not be comparable. A study of representative samples from all the separate volcanic areas analysed under similar conditions is required. Massif Central basic lavas have N a 2 0 / K 2 0 ratios between 1.5 and 2.5. This ratio decreases in more evolved magmas owing to separation of amphibole and plagioclase. Nephelinites, however, have significantly lower N a 2 0 / K 2 0 ratios (around 0.7), indicating a different mantle source (Dowries 1983).
H. Downes
520
TABLE 1. Nomenclature and mineralogy o f M a s s i f Central volcanics Names*
Phenocrystst
Groundmasst
Saturated series
Alkali basalt (Ankaramite) Hawaiite Leucobasalt (Labradorite) (Trachybasalt) (Basalt demi-deuil) Mugearite Doreite (Mesocratic trachyandesite) Benmoreite Sancyite (Leucocratic trachyandesite) Latite Trachyte (Domite) Rhyolite
ol cpx mt (ilm)
ol cpx mt ilm plag ap foid
cpx Ca-plag ap amph mt ilm (ol)
cpx plag mt ilm ol ap alkfsp amph
cpx plag amph ap mt ilm (alkfsp) (bio) cpx plag mt ilm ap alkfsp (opx) plag alkfsp cpx amph bio ap mt ilm (opx)
cpx plag mt ilm ap alkfsp (crist) (trid)
alkfsp bio cpx plag amph ap mt ilm sphene zircon (opx) alkfsp cpx mt ilm plag (amph) (bio) (pyrr) (sphene)
alkfsp bio cpx plag mt ilm ap (trid) (crist) alkfsp plag cpx ap amph trid crist zircon
Basanite
ol cpx mt
ol cpx mt foid (amph) (bio) (alkfsp)
Tephrite Ordanchite Trachyphonolite Miaskitic phonolite (Subalkaline phonolite) Phonolite Agpaitic phonolite (Hyperalkaline phonolite)
cpx plag amph mt hatiyne ap (alkfsp)
cpx plag mt ap
alkfsp plag cpx amph haiiyne nosean sodalite zircon sphene mt
alkfsp foid ap cpx amph bio plag
Undersaturated series
alkfsp cpx sodalite analcime neph plag alkfsp ap cpx amph plag neph amph analcime mos lav
* Names in bold type are current usage (international); names in parentheses are no longer in use in the context of the Massif Central. t ol, olivine; cpx, clinopyroxene; mt, magnetite; ilm, ilmenite; amph, amphibole; bio, biotite; plag, plagioclase; alkfsp, alkali feldspar; opx, orthopyroxene; trid, tridimite; crist, cristobalite; ap, apatite; neph, nepheline; foid, feldspathoid; mos, mosandrite; lav, lavenite; pyrr, pyrrhotite.
Fractional crystallization Evidence for fractional crystallization in the evolution of Massif Central volcanics has been presented for the Chaine des Puys (Maury et al. 1980; Villemant et al. 1980a), Velay (Berger et al. 1976) and Cantal (Vatin-P6rignon 1968), and in many unpublished theses. Major-element variations for these regions are similar. TiO2, Fe203, CaO, MgO and P205 all decrease regularly with differentiation (Fig. 4). MgO shows an inflection point when olivine crystallization ceases at 50 wt. ~o SiO2. K 2 0 increases regularly from basic to evolved magmas, whereas N a 2 0 increases to 60-63 wt. ~o SiO2 whereupon a strong increase in the undersaturated series and a decrease in the saturated series are observed as alkali feldspar becomes a major fractionating phase. A1203 increases in the undersaturated series but has a maximum at 60 wt. ~ SiO2 in the
saturated series and then decreases in the rhyolites. Least-squares calculations have shown that basalt can fractionate to hawaiite by removal of a total of 25~ pyroxene and plagioclase in the Chaine des Puys (Maury et al. 1980). Traceelement calculations for the same stage (Villemant et al. 1980a) indicate that clinopyroxene is the dominant phase with subordinate olivine, FeTi oxides and plagioclase. In Cantal plagioclase does not appear in the basalts, and therefore olivine and clinopyroxene dominate the extract, possibly with the addition of small quantities of a granitic component (Downes 1983). The next stage is the evolution of trachyandesite which involved pyroxene, amphibole, plagioclase and Fe-Ti oxides (Maury et al. 1980) with the proportion of pyroxene decreasing and plagioclase increasing during fractionation. The total amount of liquid remaining from the original
Massif
Central, France
16 14 ._~ 12 <
10
Y
"5
I--
8
6;
_~ -
~ / ~ /;~d
.,,,,~./y
=
o~1 ~ I 40
45
50
.,7.
uu
~
Y \',
.s
I
(
.~o
~ 55 60 65 Wt % SiO 2
I 70
75
FIG. 3. Alkali - silica diagram showing data from Cantal volcano (Downes 1983): O, basanites; II, tephrites; I', phonolites; O, basalts; l--q, trachyandesites; A, trachytes; V, rhyolites; x, MontDore rhyolites and phonolites (Brousse 1961); Velay trend from Berger et al. (1976); Chaine des Puys trend from Maury et aL (1980); UC, Massif Central uppercrustal granites; LC, Massif Central lower-crustal granulites; (H. Downes and A. Leyreloup, unpublished data.) LC*, glass analyses in partially fused lower-crustal xenoliths from the Andes (R. S. Thorpe, personal communication, 1984). basalt is about 20% at the end of this stage (benmoreite). Villemant et al. (1980a) considered that the final stage of fractionation in the saturated series is dominated by amphibole, biotite, alkali feldspar and apatite. However, the decrease in P205 throughout the magmatic evolution shows that apatite must separate prior to this stage. Maury et aL (1980) suggested that plagioclase fractionation was responsible for the fractionation step from benmoreite to trachyte, but this alone cannot explain the trace-element patterns strongly depleted in middle rare-earth elements (REE) obtained for evolved magmas in both series. According to the extract calculations of Maury et al. (1980), over 88% fractional crystallization of the original basic magma is required to produce trachyte; these figures are in broad agreement with those calculated from trace-element considerations for the same samples (Villemant et al. 1980a, 1981) and for the Cantal saturated magma series (Downes 1983). Amphibole fractionation has been advanced as the mechanism whereby trachyandesites trend to silica saturation. This may occur at the level of mugearite (Maury et al. 1980) or basalt (Mervoyer et al. 1973). Amphibole from an amphiboleolivine cumulate in Cantal has an 87Sr/S6Sr ratio of 0.7033, which is similar to that of local uncontaminated basalts (0.7033-0.7038) and con-
52 !
siderably lower than that of Cantal hawaiites (0.7039-0.7043). Thus the amphibole must have separated either from a basalt (supported by the fact that it coexists with olivine which is not a prominent phenocryst phase of hawaiites) or from a hawaiite prior to crustal contamination. Amphibole megacrysts found within basic lavas in the Massif Central also have Sr isotopes ratios and REE patterns consistent with derivation from basaltic magmas (Liotard et al. 1983). Sparse phenocrysts of orthopyroxene and/or pigeonite have been observed in saturated-series lavas and are due to the increase in silica saturation accompanying amphibole fractionation (Maury & Brousse 1976) or crustal contamination (Berger et al. 1978). The origin of the rhyolites, especially the pumice flows which are found in Mont-Dore and Cantal (Brousse & Lefevre 1966), has been resolved by petrological (Menard et al. 1980) and geochemical (Villemant et al. 1980b) arguments. They are differentiation products of the saturated series and not crustal remelts as had been proposed (Letolle & Kulbicki 1969). The mineralogy of the pumice flows is similar to that of rhyolites whose origin by fractional crystallization has not been doubted (Menard et al. 1980). Villemant et al. (1980b) presented evidence from pumiceous glass which showed strongly middle REE-depleted patterns, similar to the highly evolved phonolites of Cantal (Downes 1984) (Fig. 5). This depletion is attributed to sphene fractionation, the amount of sphene removed being less than 0.5%. Estimates of distribution coefficients for REE between sphene and trachytic magmas are also given ir~ Fig. 5. The REE patterns of the Mont-Dore rhyolites indicate that they are derived from quartz trachytes, but in Cantal the trachytes are depleted in middle REE while the rhyolites are not, making a genetic link unlikely. Instead, the rhyolites and trachytes of Cantal may represent alternative terminations of the saturated series. Zircon is also an important phase in the crystallization of Cantal rhyolites since they have lower Zr contents than their parental trachyandesites. Further proof that the rhyolites are not crustal remelts is the Sr and Nd isotope ratios of unaltered rhyolites, which fall within the range of trachyandesites and not local granites (Downes 1984) (Fig. 6). Fe-Ti oxide geothermometry and oxygen barometry for the Grande Cascade Durbise nude ardente series of Mont-Dore (Gourgaud & Maury 1984) define a field of high oxygen fugacity and temperatures ranging from 1000 to 700~ (Fig. 7). Temperature estimates for the groundmass are generally lower than for phenocrysts but there is no regular decrease in temperature orfo2 with
522
H. Downes
o DE]
~'~
_~
8
~
o
aa~
O~ 0
i~
~o
0
o~
9
0
0~ ~D
D
~/)
o
.~o+
.~ttoo ,,.-
D
roll,"
o
o
,-
o ,. +, o
o
6
I>
o~
o
#
~.~
_
0
P-.-
z <~p~ o
._
d
~
<3
o
~ o
o
9 -~o ~
9
o
I~ ZL
aao~. o o ~ i~,-~_o
o m
~
0
In
,4"
0
.~
.m +. o
o
o~
~
P:
~>_o
I.J_
0o a,,
a~
o u P I''
+%'~ ,,. I.
1
+#+~ J,,"
|
~ 0
o
,..
+l~
o+ .--
I',,,.-
l--
ql~
1
,.-
~
-o
o CD
0
Y3
121[3
iT- I
I ~ i ~ .,,..-.
~
o ~
0
#'~"
,i,.-
0+ "~
0
s
I~
n~
0
,,,.-
~
~
g
~~
%0
I
I
~
;> ~
;
o
Massif Central, France
523
50O 200
200
100
100
9
50
50
20
20
._
c
+oJ"
9
kD
~176
x
O to x
10
c?
5
o
in 14;
llm ill I
9 I
I
I
La Ce
I
Nd
i
I
I
Srn Eu G d T b Dy
i
1
Er
Yb
/
o-o. / /
9"
16-
/
o ~ 9
/
/
/
/
/
/
~-O23
Io,
~
//
9
] 0 Nephelinites 9 Basalts / Basanites 0 Xenocryst-rich
basalts
[]I n t e r m e d i a t e
9 tl
-,5
i
.~
D.I. 50
FIG. 5. REE patterns for two phonolites from Cantal
(i, 0 ) and pumiceous glass from Mont-Dore rhyolite (11). The distribution coefficientsfor REE between sphene and trachytes are also shown (line A from Villement et al. (1980b) and line B from Downes 0983)).
7~
,,oo~-
800
70
90 IOOO
TEMPERATURE ~
FIG. 7. Temperature and oxygen fugacity determinations for mixed magmas from Mont-Dore (Gourgaud and Maury 1984): O, acid host groundmass; II, basic inclusion groundmass; O, acid host phenocrysts; O, basic inclusion phenocrysts. The inset shows temperature versus differentiation index for Chaine des Puys lavas (Maury et al. 1980).
Crustal contamination
CNd
,z. I
I
-20
-10
O CONTINENTAL CRUST /
0
10
I
I
I
[
20
30
40
50
s
FIG. 6. Sr-Nd isotopic variation in Massif Central volcanics (Alibert et al. 1983 Chauvel & Jahn 1984; Downes 1984). composition owing, perhaps, to the mixing processes which operated in the evolution of these flows. Using the igneous plagioclase geothermometer and assuming that Pn2o = 1 kb, Maury et al. (1980) found a regular decrease in temperature of eruption from 1175~ in hawaiite to 900~ in trachyte (Fig. 7), with an anomalous basaltic sample. From the disagreement between the temperature obtained by this method and the values obtained using the Fe-Ti oxide geothermometer on the same samples, Maury et al. (1980) inferred a high PH,O (3--5 kb).
The extent of crustal contamination in the Massif Central volcanics has long been debated. Early workers called upon it to explain the existence of rhyolites or the abundance of silica-saturated compositions (Glangeaud & Letolle 1962; VatinP6rignon 1968; Brousse 1971). Rhyolites are now considered to be differentiates of trachyandesites or trachytes, and Mervoyer et al. (1973) and Maury et al. (1980) have demonstrated that the abundance of trachyandesites could be due to amphibole fractionation in crustal magma chambers under high PH2O conditions. Studies of the interaction of crustal xenoliths and basic magmas (Maury & Bizouard 1976; Maury et al. 1978) have shown that diffusion, mixing and solution of a crustal component can occur on a hand-specimen scale. Isotopic systems have been employed in order to determine the presence and extent of crustal contamination. Sr, Nd and Th initial isotope ratios and 6180 values all indicate assimilation of a crustal component (Javoy 1970; Morand et al. 1978; Stettler & All6gre 1979; Condomines et al. 1982; Downes 1984). It has been argued
H. Downes
524
(Maury & Varet 1980) that the high 87Sr/S6Sr values may be caused by hydrothermal alteration. This is indeed the case for some rhyolites, but combined Sr and Nd data show trends to high 87Sr/S6Sr and low 143Nd/144Nd with increasing differentiation (Downes 1984); Nd isotopes are unaffected by hydrothermal activity. Mantle heterogeneity has also been suggested as an explanation of the isotopic variation (Chauvel & Jahn 1984). The variation in eSr and e n d for basalts and basanites from the region is shown in Fig. 6 (Alibert et al. 1983; Chauvel & Jahn 1984; Downes 1984) and compared with a field of basalts which show petrographic evidence of digestion of crustal rocks (quartz and feldspar xenocrysts). Intermediate lavas and those showing evidence of crustal interaction overlap with the basic lavas but also extend to higher eSr and lower eNd. If mantle heterogeneity were the sole cause of isotopic variation, this difference should not occur. Maury & Varet (1980) have claimed that the high STSr/S6Sr initial ratios are similar to those obtained for mantle xenoliths from the Massif Central by Leggo & Hutchison (1968). However, re-analysis of Sr isotope ratios both in these bulk nodules and in separated clinopyroxenes from Massif Central spinel lherzolites indicate a highest 87Sr/86Sr ratio of only 0.70452, which is considerably lower than typical trachyandesite values.
Assimilation of up to 30~ crustal material has been suggested for the most contaminated lavas of Cantal (Downes 1984). The most probable contaminants are meta-sedimentary granulites from within the lower crust, but evidence from all isotopic systems is equivocal; partial melts of upper-crustal granites and gneisses may also be involved. Pb isotope data for Cantal volcanics and possible contaminants are reported in Table 2 and Fig. 8. The field of contaminated evolved rocks is offset from that of uncontaminated basic volcanics towards crustal compositions. The nephelinite sample has a very low 2~176 ratio compared with those of both the contaminants and the other volcanics, and thus the nephelinite cannot be contaminated. Its source must have undergone an early U - P b fractionation event and therefore may have a lithospheric mantle component. Figure 8 demonstrates that undepleted meta-sedimentary granulites such as RP6 and RP41 may be involved in contamination, confirming the conclusions drawn from Sr and Nd data (Downes 1984). 2~176 is correlated with both eSr and eNd. Contaminated magmas do not generally lie on the bulk mixing hyperbolae between basic magmas and possible contaminants, but are offset towards a straight line joining the two endmembers (indicating equal Sr/Pb or N d / P b ratios
TABLE 2. Pb isotope data for Cantal volcanics, local upper crust and four granulite xenoliths from Roche
Pointue Sample 42442 42451 42510 42450 42447 42511 43149 43510 42476 42466 GLC1 BG1 RP6 RP41 RP50 RP12
Rock-type Basanite Basanite Basalt Basalt Trachyandesite Trachyandesite Trachyandesite Rhyolite Phonolite Nephelinite Granite Gneiss Metasedimentary granulite Metasedimentary granulite Metasedimentary granulite Meta-igneous granulite
2~176 19.470 + 6 19.720+__7 19.178___6 19.428-t- 12 19.046 _ 4 19.062___4 19.192__+ 10 19.428+__2 19.475 + 7 18.637__+7 18.558+6 18.789+25 18.560 __+5 18.261 ___11 18.244 ___5 18.011___7
2~176 15.600 + 5 15.632+5 15.605___4 15.624__+8 15.637 + 4 15.678-t-2 15.630__+6 15.673___3 15.643 ___3 15.523+6 15.632+6 15.633+28 15.680 __+4 15.620___ 9 15.657 -I-4 15.623+10
2~176 39.183 ___9 39.424+ 12 39.049___21 39.199+22 39.137 ___9 39.222___10 39.021 __+38 39.327___7 39.435 ___5 38.813.__+14 38.428+19 38.488_+50 39.436__+ 11 38.618 ___26 38.372 + 11 38.067+30
Pb (ppm) 4.79 nd 3.05 nd 9.68 nd 17.59 nd nd nd nd 24.31 11.00 nd 5.63 2.76
Pb separated on 0.5 ml and 0.05 ml anion-exchange columns using HBr eluent. Chemistry blanks averaged 2 ng. Pb loaded onto Re filaments with silica gel and H3PO 4. Analyses were performed at S.U.R.R.C., East Kilbride, using a V.G. Isomass 54E automated mass spectrometer and were corrected for mass fractionation by the addition of 0.1 ~ per a.m.u. ; nd, not determined.
Massif Central, France 15.7
.=
r 15.6
-tr
]i~ RP6
[2
~.
~
-kBG1
~7
9
9
RP12
9
15.5 I 18.0
I 18.2
t 18.4
I 18.6
181.8
~ 19.0
l 19.2
I 19.4
I 1 19.6 19.8
2 0 6 Pb / 2 0 4 Pb
,~ RP6 [] []
39.2
%
39.0 38.8 ~'
38.6
~"
9 Nephelinit e
3~"
38.4
~1~
3~" BG1
9 Basalts/Basanites ~2 I n t e r m e d i a t e
38.2
lavas
~7 R h y o l i t e 9 Phonolite RP12
1 8 . 0 18.2
"I~ G r a n u f i t e s
.oo 18.4
18.6
18.8 19.0
19.2
......
19.4
t
19.6 19.8
2 0 6 Pb / 2 0 4 Pb
FIG. 8. Pb isotope variations in Cantal volcanics and possible upper- and lower-crustal contaminants (Table 2). in both). Thus an assimilation-fractional-crystallization model is required to explain the data. In this the Sr/Pb or Nd/Pb ratio of the magma will change with fractionation; application of this model is hampered by a lack of distribution coefficients for Pb.
Magma mixing Field and petrographic evidence for magma mixing and hybridization processes is abundant in Mont-Dore (Gourgaud & Vincent 1980; Gourgaud et al. 1980, 1981), Chaine des Puys (Gourgaud & Camus 1984) and Cantal (Milesi 1977; Fontaine-Vive 1981). Heterogeneous trachyandesitic magmas occur in Mont-Dore (Cantagrel et al. 1984); felsic porphyritic trachyandesites generally enclose decimetric inclusions of vesiculated basic magma with chilled and crenulated margins. A mechanism of mechanical mixing, chilling and vesiculation which triggers massive eruptions is envisaged (Gourgaud & Taupinard 1983). Three types of mixed magmas have been described in the Grande Cascade area
52 5
of Mont-Dore (Gourgaud & Maury 1984), including the common basic inclusions, 'emulsified lavas' containing small (1 cm) basic globules disseminated in the acid host and banded lavas with alternating zones of acid and basic compositions. Chilled margins, quench textures, banding and emulsification, the cauliflower-like appearance of the basic inclusions and the transfer of phenocrysts testify to the liquid-liquid nature of the interaction. The final stage of the process would be a completely homogenized magma with no textural evidence of mixing, although mineral disequilibria and compositional peculiarities would persist. This type of magma should be termed 'hybrid'; those showing prominent mixing features are termed 'mixed'. The mineralogy of Mont-Dore mixed magmas is complex; large numbers of phases are present and many show disequilibrium textures, e.g. olivine rimmed by orthopyroxene, indicating transfer of phenocrysts. Microprobe analyses of the groundmass of mixed magmas form linear trends, although compositional gaps may occur where mixing is incomplete. Although the best-known examples of magma mixing are between basalt and trachyandesite, and between trachyandesite and rhyolite, basanite-phonolite mixing also occurs. Unlike hybridization theories in which the acid member was thought to be a crustal melt (Glangeaud & Letolle 1962), magma mixing is now considered to occur between end-members which are related by fractional crystallization. Magma mixing contributes to the diversity of volcanic products because mixing trends between end-members are always straight lines on inter-element diagrams, while fractional crystallization trends often contain inflections. Thus a mixing trend which is not coincident with a fractionation trend may produce hybrid magmas which differ chemically from the normal products of fractional crystallization. Figure 9 shows trace-element concentrations of two mixed magmas from Mont-Dore which form straight-line trends compared with the curved fractional-crystallization trends of the Chaine des Puys (Maury et al. 1980) and Cantal. Magma mixing is a major volcanological process which implies the existence of large stratified magma chambers or smaller homogeneous shallow magma chambers periodically refilled with basic magma. Mixing between evolved magmas takes place within the crust after high STSr/S6Sr ratios have been acquired, since there is a linear relationship between SiO2 and 87Sr/86Sr for mixed magmas (Cantagrel, personal communication, 1983). However, the phonolite-basanite mixing of Puy Griou, Cantal, is unusual in that the Sr and Nd isotopic ratios of
526
H. Downes
,oo[F
]
...]
~
300 k
~
[450~/"I ~
02
FRACTIONATION
..... ..
,0o)
"
0
20
40
/, 0 ""~ ) )
"L
60
80
100
120
i 140
160
1 0 0 0 ~-
9
~
/ OGRANOE
4)~) r ,~ eoo?
"% ~-~-~-~
300
'
i__
CASCADE-"
/e DURBISE NUEE
~ o
o t-
Puys ~",~ "--q I ~ 4
60
,~o
I
i"
~
,C.]
,
=
Chame des N~,(
L
_~:
~
20/8 ~ o
'
""~ 9
~
..,..,Z
T
,o
20/~/~/
ppm Rb
~ I~
-I /
i
I
1
I
I
80
100
120
140
160
/
ppm Rb
FIG. 9. Magma mixing in Mont-Dore volcanics (after Gourgaud & Maury 1984). Cr-Rb and Sr-Rb variation diagrams showing straight-line mixing trends; curved fractionation trends from Cantal (Downes 1983) and Chaine des Puys (Maury et al. 1980) shown for comparison. the phonolite component are identical with those of typical Cantal basanites, indicating that the magma mixing occurred before the magmas entered the crust (Downes 1984).
i 2
Yb
5
10
20
N
FIG. 10. (La/Yb)N against (Yb)N (chondrite normalized) showing REE data for Massif Central basic volcanics (O, basalts; O, basanites; 0, nephelinite) from Chauvel & Jahn (1984) and Downes (1983): field X, calculated source of volcanics (Chauvel & Jahn 1984); field Y, calculated source (this work); field Z, basanite source (Frey et al. 1978); curves A, B and C, trends of liquid composition in equilibrium with garnet residues (details in text) (percentage partial melting indicated); point K, typical REE content of Kiama amphibole-apatite pegmatite (Wass & Rogers 1980); curves D and E, mixing lines between depleted mantle and K (percentage of K indicated); shaded area, depleted mantle (analysed spinel peridotites from Dreiser Weiher (Stosch & Seck 1980)).
Petrogenesis of basic magmas Chauvel & Jahn (1984) used REE data to model the genesis of Massif Central alkali basalts and showed that the most likely process was 10%20% partial melting of a metasomatized mantle. Figure 10 shows (La/Yb)N against (Yb)N for uncontaminated primitive magmas from Chauvel & Jahn (1984) and Downes (1983), plotted in logarithmic form, with the trends of liquids derived by partial melting using various garnetbearing residual assemblages. The advantage of the logarithmic form of this diagram is the ease with which traces of the liquid trends can be moved around the diagram, facilitating backcalculations of the mantle source. The residual assemblages used were as follows: 0.65ol, 0.20 opx, 0.14 cpx and 0.01 gt (trend A); 0.65 ol, 0.20 opx, 0.12 cpx and 0.03 gt (trend B). These two trends encompass the entire range of (La/ Yb)N-(Yb)N for Massif Central basalts and
basanites, and indicate derivation from a mantle source represented by field Y. This source is not identical with that proposed by Chauvel & Jahn (1984) (field X) but the differences are due to the distribution coefficients used in the two calculations. The residual asemblage of trend C is 0.63 ol, 0.225 opx, 0.095 cpx and 0.05 gt (Frey et al. 1978); this melting trend, projected back from the nephelinite (Fig. 10), also passes through fields X and Y, indicating that, despite the chemical and isotopic differences between the nephelinite and all other Massif Central basic magmas, the REE of the source could have been similar although the residual mineralogy would have been quite different. The model indicates that the alkali basalts could have been derived from a mantle strongly enriched in light REE ((La/Vb)N=7; (Yb)y=2) by 8%-20% partial melting, while the basanites require a lower
Massif
Central, France
degree of melting (2%-8%) of the same source. About 3% melting of a similar source but with a greater gt:cpx ratio could give rise to the nephelinite. Although there is no unique solution, Fig. 10 shows that sources such as that proposed by Frey et al. (1978) would be inappropriate for the Massif Central volcanics. The model proposed by Chauvel & Jahn (1984) invokes the action of a metasomatizing fluid strongly enriched in light REE but with the same isotopic characteristics as the depleted mantle which it invades. It is suggested here that such a mantle source could be produced by mixing amphibole-bearing mantle pegmatite material of a type similar to that described in xenoliths from Kiama, SE Australia (Wass & Rogers 1980; Menzies & Wass 1983) with depleted mantle represented by Dreiser Weiher spinel peridotites (Stosch & Seck 1980). In the calculations mantle pegmatites (point K) with (La/Yb)N=20 and (Yb)N = 5 are mixed with depleted mantle ((La/ Yb)N= 1 and (Yb)N=I-2) (lines D and E). Melting would take place in the garnet lherzolite field, so the spinel lherzolite xenoliths found in the alkali basalts of the Massif Central would not be the source of the volcanism. However, evidence for light REE enrichment in the mantle beneath the region comes from the presence of amphibole in the spinel lherzolites (Hutchison et al. 1975; Brown et al. 1980) and the low Sm/Nd of clinopyroxenes within certain mantle nodules (Downes & Dupuy, in press). This model explains the coherence of the alkali basalts and basanites as different degrees of partial melt of the same source with similar residua, but does not explain the unusual chemistry and isotopic compositions of the nephelinites. Isotopic variability in the other basic magmas (Fig. 6) can be explained as a combination of small differences in isotopic composition of the source, incipient contamination or slight alteration.
527
Conclusions The Massif Central volcanic province is a widespread area of Tertiary to Recent continental alkaline volcanism comprising alkali basalts, basanites and their derivatives. Four major processes have operated to produce the diversity of compositions in saturated and undersaturated magmas. 1 Variable degrees of partial melting of a garnet lherzolite source which had undergone recent light REE enrichment gave rise to the spectrum of primitive magma types. 2 Amphibole fractionation at depth in the basalts or hawaiites causes the trend to silica saturation which appears in the mugearites. Fractionation of clinopyroxene, olivine, plagioclase and Fe-Ti oxides in shallower-level magma chambers causes differentiation to either tephrites or trachyandesites. Olivine ceases to be a major phase at 50 wt. % SiO2, but removal of clinopyroxene, plagioclase, amphibole and opaques drives the liquid composition to phonolire or trachyte. Alkali feldspar crystallization, with accessory sphene and zircon, dominates the final stage of fractionation causing concavity of the REE patterns. Pumice-flow rhyolites are part of this differentiation sequence, and not crustal remelts. 3 Crustal contamination has occurred in the differentiated magmas of both series, as witnessed by isotopic variations. The total amount of contaminant is less than 30% and the effects are not discernible in the bulk chemistry. The probable contaminant is undepleted lower crust. 4 Magma mixing betweenend-members related by fractional crystallization is common and gives rise to mixed and hybrid magmas which are chemically distinct from the products of fractionation.
References ALIBERT, A., MICHARD-VITRAC,A. & ALBAREDE,F. 1983. The transition from alkali basalts to kimberlites: isotope and trace element evidence from melilitites. Contrib. Mineral. Petrol. 32, 176-86. BAUBRON, J. C. & CANTAGREL,J. M. 1980. Les deux volcans des Monts Dore (Massif Central fran~ais): arguments chronologiques. C.R. Acad. Sci. Paris, 290, 1409-12. BELLON, H. & GIBERT,J. P. 1979. Ap~rcu g~ochronologique (K-Ar) du volcanisme nord-Margeride, Massif Central, France. Bull. Soc, gkol. Ft. 21, 1619.
--
& HERNANDEZ,J. 1979. Chronologie du volcanisme dans le Forez (Massif Central ffan~ais); place dans l'activit6 magmatique tertaire de France. C.R. Somm. Sdances Soc. gkol. Ft. 4, 175-9. BERGER, E., BROUSSE,R. & MAURY, R. C. 1975. Age r6cent (11770+ 270 ans B.P.) des derni6res bruptions du Vivarais. C.R. Acad. Sci. Paris, S&. D, 280, 419-22. & -1978. Pr6sence de ferropigeonite dans le trachyte quartzif6re des Dents du M6zenc (Velay). , OLIVES,J. & RIVOIRARD,J. 1976. Etude des laves
H. Downes
528
alcalins du secteur des Bouti&es, Velay oriental. Travaux de l'Option Sciences de la Terre, 100 pp. Ecole Nationale Sup6rieure des Mines, Paris. BESSON, J. C., BOUDON, G. & BOURDIER, J. L. 1979. Fr6quence des laves/t leucite sur le versant oriental du Mont-Dore (Massif Central franqais). C.R. Acad. Sci. Paris, S&. D, 288, 1639-42. , LY, M. H., CANTAGREL, J. U., DE GOER DE HEVRE, A. & VINCENT,P. M. 1977. Une coul6e de ponces post-villafranchienne sur le versant oriental du Mont-Dore (Massif Central franvais). C.R. Acad. Sci. Paris, S&. D, 284, 1875-8. BONIFAY, E. & TIERCELIN, J. J. 1977. Existence d'une activit6 volcanique et tectonique au d6but du Pleistoc+ne moyen dans le bassin du Puy (Haute Loire). C.R. Acad. Sci. Paris, S&. D, 284, 2455-7. BOUDON, G. & CANTAGREL, J. M. 1981. Donn6es g6ochronologiques sur quelques laves ~ leucite des 'plan6zes' du Cantal (Massif Central fran~:ais). C.R. Somm. Skances Soc. gbol. Ft. (5-6), 171-4. & DE GOER DE HEVRE, A. 1978. Leuciton6ph61inites et basanites ~. leucite dans le Nord et l'Est du Cantal (Massif Central franqais). C.R. Acad. Sci. Paris, S&. D, 286, 1645-8. BOURDIER, J. L. & CANTAGREL, J. M. 1983. Le volcanisme aux confins des Monts-Dore et du C6zallier (Massif Central franqais): nouvelles donn6es chronologiques. Bull. Soc. gbol. Fr. 25(7), 277-80. BROUSSE, R. 1961. Analyses chimiques des roches volcaniques tertaires et quaternaires de la France. Carte Gdol. Fr. 58, 160 pp. 1971. Magmatologie du volcanisme n6og+ne et quaternaire du Massif Central. Syrup. J. Jung, pp. 377-478. Plein Air Service, Clermont-Ferrand. & LEFEVRE,C. 1966. Nappes de ponces du Cantal et du Mont-Dore. Leurs aspects volcanologique, p&rographique et min6ralogique. Bull. Soc. gkol. Fr. 8(7), 223-45. BROWS, G. M., PISSEST, R. H. & CoIsv, P. 1980. The petrology of spinel-peridotite xenoliths from the Massif Central, France. Am. J. Sci. 280A, 471-98. CANTAGREL,J. M. & BAUBRON,J. C. 1983. Chronologie des 6ruptions dans le massif volcanique des MontsDore (m6thode potassium-argon). Implications volcanologiques. Gbol. Fr. (2) 1, (1-2), 123-42. , DIDIER, J. & GOURGAUD, A. 1984. Magma mixing, origin of intermediate rocks and 'enclaves', from volcanism to plutonism. Phys. Earth planet. Inter. 35, 65-76. CHAUVEL, C. & JAHN, B. M. 1984. Nd-Sr isotope and REE geochemistry of alkali basalts from the Massif Central, France. Geochim. cosmochim. Acta, -
-
-
-
-
-
48,
9 3 - 1 1 0 .
CHURLET, M. A. & VELDE, D. 1976. Description d'une roche vitreuse (fi composition chimique de basanite) b, olivine, hauyne, clinopyrox6ne et analcime, mais sans plagioclase, affleurant au sud d'Apchon (Cantal). C.R. Acad. Sci. Paris, S&. D, 282, 21314. CONDOMINES, M. 1978. Age of the Olby-Laschamp geomagnetic polarity event. Nature Lond. 276, 257. , MORAND, R., CAMUS,G. & DUTHOU, J. L. 1982. Chronological and geochemical study of lavas
from the Chaine des Puys, Massif Central, France; Evidence for crustal contamination. Contrib. Mineral. Petrol. 81,296-303. DELUZARCHE,M. & WESTPHAL,M. 1981. Pal~omagn&isme de quelques syst6mes +ruptifs r6cents du Massif du Mont-Dore. Bull. Soc. g~ol. Ft. 23, 1115. DOWSES, n. 1983. The petrogenesis of coexisting saturated and undersaturated continental alkaline magma series. PhD Thesis, University of Leeds (unpublished). 1984. Sr and Nd isotope geochemistry of coexisting alkaline magma series, Cantal, Massif Central, France. Earth planet. Sci. Lett. 69, 321-34. & DuPuY, C. Textural, isotopic and REE variations in spinel peridotite xenoliths, Massif Central, France. Earth planet. Sci. Lett. in press. & LEYRELOUP,A. 1986. Granulitic xenoliths from the French Massif Central--Petrology, Sr and Nd isotope systematics and model age estimates. In: DAWSON,J. B. (ed.) Nature of the Lower Continental Crust. Geol. Soc. Spec. Publ. 24, 319-30. DuPts, A. & GIBERT, J. P. 1977. Apport de la magn&otellurique de faille profondeur aux ~tudes hydrog6ologiques: l'exemple de l'aire d'6mergence d'eaux hydrothermales de Chaudes Aigues (Cantal, France). C.R. Acad. Sci. Paris, S&. D, 285, 139-42. DUPUY, C., LEYRELOUP,A. & VERNIERS, J. 1977. The lower continental crust of the Massif Central (Bournac, France) with special reference to REE, U and Th composition, evolution, heat-flow production. Phys. Chem. Earth, 11,401-16. FONTAINE-VIVE,M. 1981. Volcanologie et p&rologie de la haute vall6e du Brezons et des plan6zes adjacents (Massif du Cantal). Thkse 3eme Cycle, Universit~ de Grenoble I (unpublished). FREY, F. A., GREEN, D. H. & ROY, S. D. 1978. Integrated models of basalt petrogenesis" a study of quartz tholeiites to olivine melilitites from south eastern Australia, utilizing geochemical and experimental petrological data. J. Petrol. 19, 463-513. GILLOT, P. Y. 1978. K-Ar dating of the Laschamp magnetic excursion; a method for dating young volcanic rocks. In: R. E. ZARTMAN(ed.) U.S. Geol. Surv. Open File Rep. 78-701, pp. 140-2. U.S. Geological Survey, Washington D.C. GLANGEAUD, L. & LETOLLE, R. 1962. Evolution gbochimique et g6odynamique des magmas du massif volcanique du Mont-Dore et de l'Auvergne. Methodes et consbquences g6n6rales. Bull. Soc. ft. Mineral. Cristallogr. 85, 296-308. GOURGAUD, A. & CAMUS, G. 1984. Magma mixing at La Nug6re volcano (Chaine des Puys, Massif Central, France). Role in trachyandesite genesis. Bull. Volcanol. 47, 781-805. & MAURY, R. C. 1984. Magma-mixing in alkaline series: an example from Sancy volcano (MontDore, Massif Central, France). Bull. Volcanol. 47, 827-47. -& VINCENT, P. M. 1980. Coexistence de magmas rhyolitique et trachy-andbsitique au d6me de la Gacherie (Monts-Dore, Massif Central fran~ais). C.R. Acad. Sci. Paris, S&. D, 290, 81-4. -
-
-
-
Massif Central, France --,
BOURDIER,J. L. & VINCENT,P. M. 1980. M6lange de magmas sur-satur& et sous-satur& dans le volcan du Sancy (Monts-Dore, Massif Central fran~ais). C.R. Acad. Sci. Paris, Sbr. D, 291, 1758. -& CANTAGREL, J. M. & VINCENT, P. M. 1981. M61ange de magmas et p6trog6n6se des trachyand6sites du Mont-Dore (Massif Central franqais). C.R. Acad. Sci. Paris, S~r D, 293, 711-6. GUERIN, G., GILLOT, P. Y., LE GARREC, M. J. & BROUSSE, R. 1981. Age subactuel des derni6res manifestations 6ruptives du Mont-Dore et du C6zallier. C.R. Acad. Sci. Paris, S~r. D, 292, 8557. HALL, C. M. & YORK, D. 1978a. K - A r and 4~ studies of the Laschamp geomagnetic reversal. Eos, 59, 1038. -& 1978b. 4~ dating of the Laschamp event and associated volcanism in the Chaine des Puys. Eos, 60, 244. HERNANDEZ, J. 1973. Le volcanisme tertaire des monts du Forez (Massif Central fran~ais): basanites /~ analcime, ~ leucite et n6ph61inites fi m61ilite. Bull. Soc. fr. Minbral. Cristallogr. 96, 303-12. 1976. Donn~es nouvelles sur la composition min6ralogique de la n6phblinite de Marcoux (Forez). Bull. Soc. Jr. MinOral. Cristallogr. 99, 616. HUTCHISON, R., CHAMBERS, A. L. PAUL, D. K. & HARRIS, P. G. 1975. Chemical variations among French ultramafic xenoliths; evidence for a heterogeneous upper mantle. Mineral. Mag. 40, 153-70. JAVOY, i . 1970. Utilisation des isotopes de l'oxyg6ne en magmatologie. Th~se d'Etat, Paris (unpublished). LEGGO, P. j. & HUTCHISON, R. 1968. A Rb-Sr isotope study of ultrabasic xenoliths and their basaltic host rocks from t h e Massif Central, France. Earth planet. Sci. Lett. 5, 71-5. LETOLLE, R. & KULmCKI, G. 1969. G6ochimie des laves du massif volcanique plio-quaternaire du Mont-Dore. Bull. Centre Rech. Pau SNPA, 3, 4017; 4, 191-233. LEYRELOUP, A., DUPUY, C. & ANDRIAMBOLOLONA,R. 1977. Catazonal xenoliths in French Neogene volcanic rocks; constitution of the lower crust; 2, Chemical composition and consequences of the evolution of the French Massif Central Precambrian crust. Contrib. Mineral. Petrol. 62, 283-300. LIOTARD, J. M., BOIVIN, P., CANTAGREL, J.-M. & DUPUY, C. 1983. M6gacristaux d'amphibole et basaltes alcaline associ&. Probl6mes de leurs relations p6trog6n6tiques et g6ochimiques. Bull. Mineral. 106, 451-64. LUCAZEAU,F., VASSEUR,G. & BAYER, 1984. Interpretation of heat flow data in the French Massif Central. Tectonophysics, 103, 99-119. MAGONTHIER, M. C. & VELDE, D. 1976. Mineralogy and petrology of some Tertiary leucite-rhonite basanites from central France. Mineral. Mag. 40, 817-26. MAURY, R. C. & Bizouard, H. 1976. Melting of acid xenoliths into a basanite: an approach to the
529
possible mechanisms of crustal contamination. Contrib. Mineral. Petrol. 48, 275-86. -& BROUSSE, R. 1976. Pr6sence de pigeonite et d'orthopyrox6ne dans certains laves du Massif Central fran~ais: leur repartition et leur origine. Bull. Mineral. 101, 10-21. & VARET, J. 1980. Le volcanisme tertaire et quaternaire. In: Colloque C7, G~ologie de la France, Evolutions gbologiques de la France, AUTRAN, A. & DERCOURT, J. coord., Mem. B.R.G.M. 107, 13859. --, BROUSSE, R., VILLEMANT, B., SORON, J. L., JAFFREZIC, H. & TREUIL, M. 1980. Cristallisation fraction6e d'un magma basaltique alcalin" la s6rie de la Cha]ne des Puys (Massif Central, France). I. Petrologie. Bull. Minbral. 103, 250-66. --, DIDIER, J. & LAMEYRE, J. 1978. Comparative magma--xenolith relationships in some volcanic and plutonic rocks from the French Massif Central. Contrib. Mineral. Petrol. 66, 401-8. MENARD, J. J., CLOCCHIATTI, R., MAURY, R. C. BROUSSE, R. 1980. Origine des ponces rhyolitiques du Mont-Dore (Massif Central, France): Arguments p&rologiques. C.R. Acad. Sci. Paris, Skr. D, 290, 559-62. MENZIES, M. A. & WASS, S. Y. 1983. CO2- and light REE-rich mantle below eastern Australia: a REE and isotopic study of alkaline magmas and apatiterich mantle xenoliths from the Southern Highlands Province, Australia. Earth planet. Sci. Lett. 65, 287-302. MERVOYER, B., MAURY, R. C. & VARET, J. 1973. Un m&anisme possible d'6volution des trachyand6sites du Massif Central par cristallisation fractionnb sous pression d'eau. C.R. Acad. Sci. Paris, S~r. D, 277, 9-12. MILESI, J.-P. 1977. Dynamique des 6coulements pyroclastiques du Cantal oriental sur l'example de ceux de la vall6e de l'Alagnon, Massif Central fran~ais. Th~se 3brae Cycle, Grenoble (unpublished). MORAND, R., CONDOMINES,M. t~ ALLI~GRE,C. J. 1978. D6s6quilibres 23~ dans quelques laves de la Chaine des Puys. C.R. Acad. Sci. Paris, SOt. D, 286, 1845-9. MOSSAND, P., CANTAGREL, J.-M. & VINCENT, P. M. 1982. Le Cald6ra de Haute-Dordorgne: ~ge et limites (Massif des Monts-Dore, France). Bull. Soc. gOol. Fr. 24(7), 727-38. PERRIER, G. & RUEGG, J. C. 1973. Structure profonde du Massif Central fran6ais. Annls Geophys. 29(4), 435-502. PIN, C. & VIELZEUF, O. 1983. Granulites and related rocks in Variscan median Europe: a dualistic interpretation. Tectonophysics, 93, 47-74. SANTOIRE, J. P., BROUSSE, R. & BELLON, H. 1977. Le Puy de Bessoles, enregisteur des 6pisodes volcaniques du Massif du Mont-Dore. C.R. Acad. Sci. Paris, S~r. D, 285, 19-22. SOURIAU, A., CORREIG, A. M. t~ SOURIAU, M. 1981. Attenuation of Rayleigh waves across the volcanic area of the Massif Central, France. Phys. Earth. planet. Inter 23, 62-71. STETTLER, A. & ALLI~GRE, C. J. 1979. STRb/S7Sr constraints on the genesis and evolution of the
53o
H. Do wnes
continental volcanic system Cantal (France). Earth planet. Sci. Lett. 44, 269-78. STOSCH, H.-G. & SECK, H. A. 1980. Geochemistry and mineralogy of two spinel peridotite suites from Dreiser Weiher, West Germany. Geochim. cosmochim. Acta, 44, 457-70. TOURNON, J. & VELDE, D. 1971. On the presence of leucite in basaltic rocks from Central France. Contrib. Mineral. Petrol. 30, 291-5. VATIN-PI~RIGNON,N. 1968. Les formations 6ruptives et la structure de l'6difice volcanique au centre du Cantal (Massif Central franqais). Bull. Volcanol. 32, 207-51. VELDE, D. 1973. A propos des basanites du Massif Central fran~ais: la pr6sence de feldspathoide potassique (leucite) ou sodique (nepheline ou analcime) correspond 5. une diff6rence de composition chimique. C.R. Acad. Sci. Paris, S~r D, 276, 3257-60. VILLEMANT,B., JAFFREZIC,H., JORON, J.-L. & TREUIL, M. 1981. Distribution coefficients of major and trace elements: fractional crystallisation in the alkali basalt series of Chaine des Puys (Massif
Central, France). Geochim. cosmochim. Acta, 45, 1997-2016. , JORON, J.-L., JAFFREZIC,H., TREUIL, M. MAURY, R. C. & BROUSSE, R. 1980a. Cristallisation fractionnbe d'un magma basaltique alcalin: la s6rie de la Chaine des Puys (Massif Central, France). II. G6ochimie. Bull. Mineral. 103, 267-86. , & TREUIL, M. 1980b. Origine des quartz des nappes de ponces du Mont-Dore (Massif Central frangais): Arguments g~ochimiques. C.R. Acad. Sci. Paris, Sbr. D, 290, 687-90. VINCENT, P., AUBERT, M., BOIVIN, P., CANTAGREL,J.M. & LENAT,J. F. 1977. Dbcouverte d'un volcanism Pal~ocene en Auvergne: les maars de Menat et leurs annexes. Bull. Soc. g~ol. Ft. 19(7), I057-70. WIMMENAUER, W. 1974. The alkaline provinces of Central Europe and France. In: H. SORENSEN(ed.) The Alkaline Rocks, pp. 238-70. John Wiley & Sons, New York. WASS, S. Y. & ROGERS, N. W. 1980. Mantle metasomatism--precursor to continental alkali volcanism. Geochim. cosmochim. Acta, 44, 1811-23.
HILARY DOWNES, Grant Institute of Geology, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, U.K. Present address: Department of Geology, Birkbeck College, 7-15 Gresse Street, London W1P 1PA, U.K.
Alkaline rocks of the eastern part of the Baltic Shield (Kola Peninsula) L. N. Kogarko S U M M A R Y : The Kola Peninsula contains a remarkable assemblage of alkaline rocks. These can be sub-divided into (a) those of Proterozoic age (1900-1600 Ma), including the Yelet'ozero and Gremiakha-Vyrmes complexes and (b) those of Palaeozoic age (about 365 Ma). The latter include numerous alkaline ultramafic intrusions and the larger agpaitic complexes at Khibina and Lovozero. The Proterozoic activity involves syenites and nepheline syenites in an alkaline gabbroic association. In both the Yelet'ozero and the GremiakhaVyrmes complexes the magmas become more salic with time. In the Palaeozoic associations alkaline gabbroic rocks are absent and the assemblages range from ultramafic (olivinites, pyroxenites, melilitites etc.) to highly differentiated agpaitic rocks and also to carbonatites. Low initial STSr/86Srratios (0.7035-0.7039) suggest that the magmas were co-genetic and of mantle origin. Experimental studies coupled with investigation of micro-inclusions suggest that the agpaitic magmas crystallized over a wide temperature interval (more than 300~ under highly reducing conditions. The associated gas phase had a lowjH2obut relatively high CH~, H2 and HzS fugacities. The apatite deposits in the Khibina complex are uniquely large and are associated with ijolite, melteigite, urtite, juvite and malignite. The phosphate deposits are interpreted as apatite cumulates produced by suspension of apatite from the host magma after precipitation of clinopyroxene and nepheline.
Proterozoic alkaline complexes The Kola Peninsula in the eastern part of the Baltic Shield is a classic region for the study of alkaline magmatic rocks of widely differing ages and compositions (Fig. 1). The first alkalic complexes were emplaced during the early consolidation of the Baltic Shield, and include the Yelet'ozero and Gremiakha-Vyrmes complexes in the Soviet part of the Baltic Shield and the Almunge and Norra-Karr complexes in Sweden. The age of these intrusions is 1900-1600 Ma (Kukharenko 1971). Comprehensive analysis of the geological-tectonic location of the Yelet'ozero and Gremiakha-Vyrmes complexes has shown that they were formed contemporaneously with the late orogenic depressions of Karelian age, bounded by horsts.
3 A layered complex of peridotite, pyroxenite, anorthosite and gabbro. 4 Essexites. 5 Dykes of porphyrite, diabase and spessartite. Stage 2
6 Nepheline syenites and miaskitic alkali syenites. 7 Dykes ofbostonite, albitite, and metasomatic and post-magmatic rocks. The average composition of the Yelet'ozero rocks is close to that of olivine basalt (Table 1). A detailed description of the magmatic evolution at Yelet'ozero has been presented by Kukharenko et al. (1965). During the differentiation of the alkaline magma the content of SiO2, A1203, alkalis and trace elements increased, with a concomitant decrease in mafic components.
The Yelet'ozero complex The Yelet'ozero complex (50 km z) was formed in two stages and comprises seven phases (Kukharenko et al. 1965 (Fig. 2) during which the following rocks were emplaced. Stage 1
1 Various types of pyroxenites. 2 Olivine gabbro, gabbro-diorite and gabbronorite.
The Gremiakha-Vyrmes complex The Gremiakha-Vyrmes complex (130 km 2) was formed in two phases (Kukharenko et al. 1965) (Fig. 3). Of these, the first intrusive phase involved the emplacement of peridotite, pyroxenite, gabbro, syenite and pulaskite. The most abundant rocks, however, are gabbroic. The second intrusive phase affected the central part of the complex and gave rise mainly to biotite-
From: FITTON,J. G. & UPTON, B. G. J. (eds), 1987, Alkaline Igneous Rocks,
Geological Society Special Publication No. 30, pp. 531-544.
531
532
L. N. Kogarko
\ :' :'.'i'i"-'..':
\ \ 9.,:?.,..
",~
1''~1
::.~
"
_
_
~
,
~-~
~
- ..~ .--~.~
~,'/11 I' I~ \'g ~ , ' - x
FIG. 1. Sketch map showing the distribution of alkaline massifs on the Kola Peninsula and N Karelia: I, Lower Archaean complex of gneiss, granite-gneiss and migmatite; II, undivided complex of Archaean gneiss, granitegneiss and granite; III, Upper Archaean Belomorsk gneisses, granite-gneiss, granulite and metabasites; IV, Lower Proterozoic supracrustal and volcanogenic series; V, gabbro and ultrabasites; VI, granitoids; VII, massifs of the alkaline gabbroid association rocks (1, Yelet'ozero; 2, Gremiaha-Vyrmes); VIII, massifs of the alkaline granitoid association (3, massifs of the Ponoi river; 4, Western Keivy; 5, white tundras); IX, porphyritic granites and rapakivi granites; X, Upper Proterozoic sedimentary metamorphic rocks; XI, Caledonian ultramafic alkaline massifs (6, Turii mys; 7, Kovdozero; 8, Vuoriyarvi; 9, Sallanlatvi; 10, Pseotchniy; 11, Ingozero; 12, Salmagorskiy; 13, Lesnaya Varaka; 14, Ozernaya Varaka; 15, Afrikanda; 16, Mavrgubinskiy; 17, Kovdor; 18, Sebl'yavr; 19, Kurga); XII, Palaeozoic sedimentary volcanic rocks (alkaline basaltoids) (20, Lovozero formation; 21, Kontozero); XIII, Palaeozoic nepheline syenites (22, Khibina; 23, Lovozero; 24, Soustova massif). aegirine nepheline syenites. Subordinate rarer components include ijolite-melteigite, urtite, juvite, malignite and alkali syenite. The relationships of the alkaline rocks indicate their
syngenetic character. Some of the alkaline rocks display primary layering. A number of workers (e.g. Polkanov & Gerling 1960) have considered that the Gremiakha-
533
Alkaline rocks of the Kola Peninsula TABLE 1. Average chemical composition of the
Proterozoic alkaline rocks (Kola Peninsula) Alkaline gabbroid association Yelet'ozero wt. %
1
2
SiO2 TiO2 A1203 FeO MnO MgO CaO Na20 KzO P205
43.19 2.58 17.61 4.57 0.17 6.78 12.98 1.63 0.63 0.67 0.20 0.20 0.77
55.45 0.37 21.32 1.54 0.06 0.18 0.95 6.99 8.39 -0.15 0.15 0.06
H2O+ H20LOI No. of analyses
11
1
Gremiakha-Vyrmes 3
4
5
47.06 44.01 50.15 2.25 1 . 7 6 0.84 14.30 14.99 21.93 4.48 6.15 2.20 0.14 0.22 0.10 6.96 4 . 2 1 1.23 9.32 9.15 2.72 3.54 6.26 9.76 0.59 2.92 4.98 0.60 0.35 0.04 0.70 1 . 0 7 0.98 0.28 0.16 0.17 ---10
4
From Kukharenko 1971. 1, Gabbro; 2, nepheline syenite; 3, gabbro; ijolite-melteigite; 5, nepheline syenite. LOI, loss on ignition.
6 4,
The compositional evolution of the Grem i a k h a - V y r m e s alkaline m a g m a s was similar to that of the Yelet'ozero complex (Table 1). During the differentiation of the alkaline basalt m a g m a , the content of SiO 2, AI20 3, alkalis and incompatible trace elements increased while the mafic c o m p o n e n t s decreased.
FiG. 2. Geological sketch map of the Yelet'ozero massif: 1(a), bostonites; 1(b), alkaline pegmatites; 2, nepheline syenites (~Pt); 3, porphyritic dolerite, spessartite and vogesite dykes; 4(a), leucocratic essexites; 4(b), leucocratic gabbro and orthoclase plagioclasites (V4Pt); 5, olivine gabbro, ore (V3Pt); 6(a), coarse-grained and pegmatitic gabbro (massive and streaky); 6(b), fine-grained gabbro (gabbro, olivine gabbro, gabbro-norite, gabbro-diorite) (VzPt); 7, pyroxenites and peridotites (V1Pt); 8, plagioclasemicrocline granites and granodiorites (A1); 9, igneous lamination and primary banding; 10, crystalline schistosity and banding; 11, zones of schistosity and mylonitization; 12, tectonic dislocations; 13, inferred geological boundaries. Vyrmes complex also includes alkaline granites and syenites. However, new data ( K u k h a r e n k o 1971) imply that the granites belong to a separate intrusive complex w h i c h is only spatially related to the G r e m i a k h a - V y r m e s alkaline rocks.
Palaeozoic alkaline complexes After the final stabilization of the Baltic Shield about 1.2• 103 Ma ago alkaline complexes of Palaeozoic age were intruded. These include some 23 ultramafic alkaline intrusions in addition to the large K h i b i n a and Lovozero agpaitic complexes. Recent isotopic investigation has led to a revision of the ages of the K h i b i n a and Lovozero massifs and they are now k n o w n to have been contemporaneously intruded at 365 Ma, i.e. Mid to U p p e r D e v o n i a n (Kogarko et al. 1981, 1983). The ultrabasic alkaline intrusions of the Kola Peninsula are the same age. Isochrons from rocks and minerals (Fig. 4) showed that all the rocks are co-magmatic with initial 87Sr/86Sr ratios ranging from 0.7035 to 0.7039; the lower limit is characteristic for the K h i b i n a and some of the ultramafic alkaline intrusions and the higher values are characteristic
534
L. N. Kogarko 0.718 ) .7 14 0.714 ./86Sr ~ LOVL%lv0996 87Sr/S6Sr
.... LOV24o~
KH9436
1%
2
9445
0.710 HIIfLOV31 )-3~'K H9439 0.706 ) .706 ill%-------o'S" ~\ \ J/'KH94O2 ~,;~;A.# o. 0.702 ,.7O2,o
o[,
,'
i
212
87Rb/86Sr
FIG. 4. Whole-rock isochrons for agpaitic rocks: LOV, Lovozero (age, 362 + 17 Ma; initial isotope ratio, 0.70381 • KH, Khibina (age, 364___15 Ma; initial isotope ratio, 0.70348 ___0.00015). ment rather than three as previously supposed. There was thus a Proterozoic stage with formation of an alkaline gabbroic association and a Palaeozoic stage in which the ultramafic alkaline and agpaitic complexes were emplaced. The similarity of the initial Sr isotopic ratios in the agpaitic and ultramafic alkaline rocks suggests that they were genetically related. These complexes (0.835 km 2) form polyphase central-type intrusions whose contacts with country rocks are essentially vertical (Kukharenko 1958); Kukharenko et al. 1965) (Fig. 5). The alkaline ultramafic intrusions
FIG. 3. Geological sketch map of the GremiakhaVyrmes massif: 1, Quaternary deposits; intrusive rocks (2, alkaline granite (~Pt); 3, ijolite and nepheline syenite (V4Pt); 4, gabbro, pulaskite and syenite (V4Pt); 5, leucocratic gabbro (V3Pt); 6, ore-bearing peridotite (V2Pt); 7, gabbro and gabbro-norite (V1Pt)); basement rocks of the Kola suite (8, pyroxene diorite and gneissose diorite (A 0; 9, garnet andalusite, garnet-biotite and other gneiss (K2A1); 10, microcline-biotite gneiss (K 1A1); structural elements (11, orientation of crystalline schistosity and banding; 12, orientation of igneous lamination and primary banding; 13, contacts (a) observed and (b) inferred; 14, tectonic dislocations). for the Lovozero complex. These data imply that crustal sources did not contribute to the generation of these alkaline rocks nor to the world's largest apatite deposits which are contained in them. Comparison of the age and initial 87Sr/ 86Sr ratios of the alkaline complexes at the eastern part of the Baltic Shield demonstrates that this alkaline magmatism had two stages of develop-
The ultrabasic alkaline intrusions are very similar with respect to their geological structures. The form of the geological map for these intrusions is apparently controlled mainly by the depth of erosion. The ultrabasic rocks in these complexes are represented by olivinites, sometimes by dunites and rarely by peridotites. However, the presence of pyroxenites and melilite-bearing rocks is very characteristic. The more alkaline rocks are dominantly melteigites and ijolites with relatively rare urtites. Malignites and alkaline syenites (with nepheline and cancrinite) are commonly present as dykes and veins. Carbonatites are also present, usually as stocks. Fenites in the surrounding gneisses are fairly common with the areas affected usually being proportional to the size of the intrusion. The structure of these ultramafic alkaline intrusions has been studied by many workers and the chemical compositions of some of the rocks are presented in Table 2. The main petrochemical peculiarity of the assemblage is the successive formation of (i) undersaturated ultramafic rocks,
A l k a l i n e r o c k s o f the K o l a P e n i n s u l a
535
FIG. 5. Block diagram illustrating the geology of the Kovdor massif (after Kukharenko et al. 1965) : 1, carbonatite; 2(a), Ti-magnetite ore; 2(b), apatite-forsterite rocks; 3, hydrothermal veins (tremolite and tremolite--clinohumite rocks, phlogopite and vermiculite rocks); 4, vein rocks (ijolite and pegmatite); 5, melilitebearing rocks (turjaite and melilitite); 6, melteigite and ijolite; 7, pyroxenite; 8, olivine-pyroxene rocks; 9, olivinite; 10, country rocks ((a) fenite; (b) granite and gneiss). (ii) a nepheline-pyroxene alkaline rock series and (iii) nepheline syenites and carbonatites. All the Kola ultramafic alkaline rocks are characterized by high contents of CaO, TiO,, Fe203, Nb, Zr and other incompatible trace elements, and low MgO/(FeO + Fe203) ratios. The average composition of the assemblage is close to that of Daly's melilite basalt (Kukharenko et al. 1965).
The Khibina complex The Khibina complex (1350 km 2) is a centraltype polyphase intrusion controlled by the tectonic contact between metamorphic rocks of Archaean and Proterozoic age (Fig. 6). It was formed in eight intrusive phases (Eliseev et al. 1939; Galakhov 1975). The oldest rocks tend to occur towards the margins with successively younger intrusions encountered towards the centre. From the oldest to youngest, the components are as follows. 1 Nepheline and alkali syenites; nepheline syenite-porphyries. 2 Massive khibinites(coarse-grainednepheline syenites). 3 Trachytoid khibinites.
4 Rischorrites (K-richnepheline syenites). 5 Ijolites, melteigites and urtites (this intrusion composes the well-known stratified complex of rocks which contains the world's largest apatite deposits.) 6 Medium-grained nepheline syenites. 7 Foyaites. 8 Carbonatites. A stock-like lode of carbonatite and carbonatized silicate rocks was recently discovered near the eastern boundary of the Khibina massif (Dudkin et al. 1984). The diameter of the carbonatite stock is approximately 800 m. The carbonatities are closely associated with lamprophyre, tinguaite and trachyte porphyry dykes. Carbonatization of the foyaites and the presence of late carbonatite veins have been observed. Isotopic studies carried out in cooperation with Dr. U. K r a m m have shown the age of the carbonatites to be 363 Ma. They are somewhat enriched in radiogenic Sr relative to the rocks of the Khibina massif ( S 7 S r / 8 6 S r i = 0.70410). The average composition of the Khibina complex nepheline syenites is very close to Daly's average nepheline syenite. There were no important compositional fluctuations during the formation of the massif (Table 2) with the exception of the rischorrites, ijolites-melteigites-urtites and carbonatites (Gerasimovsky et al. 1974).
536
L. N. Kogarko
.-~ m ~
~
~
~
,
~
~
2 ~ d d d 2 d 2 M d ~ d
~>~ ~ 2 ~ 2 M 2 d d 2 ~ d
~ R ~ - ~ ~ ~ o ~m
~~
~
_
~
-
<~ ~ .
.
.
.
.
.
.
.
.
.
.
b,i
r~ 9
~
~
~
~
~~
r-i <
c~c~o~'~176 ~ 9 #oo o~
0
Alkaline rocks of the Kola Peninsula
+
. ~ = ~ M t .
"S
Valepa~
+
i
"~YI At ~
9
Mt. Lyavochorr /
/
/
/
~~t
+"r + +§ + ~.J
537
Mr.
I Ii1;
+ + I I -4- "f + Mtl Yumechorr + t ~\ ~
+ t
Jr +
+
+
+-++ \+ +
+
-+
9
+ "t-
+
~.
+ +
~'~ ,,,. ,
~ s
~ //
~ - - ~ /
/ ~/
~ /f
~
" ~ rA=
=.~ '
Mr. Yukspor
+ \
Mt. Rasvumchorr
ffTq14 r lS
,4.,..
~1
FI~. 6. Geological sketch map of the Khibina massif: 1, Quaternary rocks; 2, fine-grained mica-aegirineamphibole syenites; 4, trachytoid foyaite; 5, massive foyaite; 6, ijolite, urtite and malignite; 7, apatite-nepheline rocks; 8, rischorrite; 9, alkaline syenite porphyry; 10, trachytoid khibinite; 11, massive khibinite; 12, alkaline and nepheline syenites; 13, plagioclase-pyroxene hornfels; 14, quartz gabbro-diabase; 15, pillow lavas, greenschist, tuff and sedimentary rocks; 16, gneiss.
The Lovozero complex The Lovozero complex (650 km 2) was formed during four intrusive phases (Gerasimovsky et al. 1966) (fig. 7). The overall shape of the complex is lopolithic. The first phase is represented by miaskitic nepheline syenites and poikilitic nepheline and nosean syenites. The second phase produced a sill-like stratified intrusion with a thickness exceeding 1500 m. This rhythmically layered body is built up from regular layers of nepheline syenite with varying composition (urtites, foyaites and lujavrites). Three zones can be distinguished within this intrusion : (i) an upper zone 300 m thick predominantly represented by three units (urtite, foyaite
and lujavrite); (ii) a middle unlayered zone 650 m thick of predominantly lujavritic composition; (iii) a lower zone over 500 m thick which consists of units including two (foyaite-lujavrite) or three (urtite-foyaite-lujavrite) members. The third intrusive phase gave rise to a complex of eudialyte lujavrites which formed a sill-like intrusion cutting and overlying the stratified rocks of the second phase (Bussen & Sakharov 1972). This intrusion is also layered and consists mainly of mesocratic, melanocratic and leucocratic eudialyte lujavrites. The rhythmic repetition of units and their persistence in space, however, is not as regular as that in the differentiated second-phase intrusion. There are nest-like and lenticular bodies of
538
L. N. Kogarko
L
/ L/./ /
t..
,
v..
~'.
x.
x
~
O~
x
L
I.
Iluayv~ /
L
'~
L L
L
L L
L
x
x
x
x
-
~
Y~
t
I, N inchurt
- ,Ji
M,
---4
k-"
,
v
L
~
L
t_ L
\ \
r-qT U ls IS-Tl,
':q12
FIG. 7. Geological map of the Lovozero massif: 1, Lower Archaean granite gneisses; 2, Proterozoic sillimaniteandaiusite schist; 3, Proterozoic ultrabasic rocks; 4, Palaeozoic magmatic complex, volcanic and sedimentary rocks of the Lovozero formation. Intrusive rocks of the Lovozero Massif: first phase (5, metamorphosed nepheline syenite; 6, poikilitic nosean syenite, nepheline-nosean syenite and even-grained nepheline syenite); second phase (7, urtite, foyaite and lujavrite; 8, poikilitic sodalite syenite); third phase (9, eudialyte lujavrite; 10, porphyritic lujavrite; 11, porphyritic lovozerite-murmanite lujavrite; 12, poikilitic sodalite syenite and tawite).
eudialyte-rich facies with thicknesses ranging from several to tens of metres among the eudialyte-bearing lujavrites. Euhedral eudialyte crystals compose 70%-80% (modally) of these rocks. Poikilitic sodalite syenites and tawites occur as lensoid bodies within the rocks of the second and third intrusive phases (Gerasimovsky et al. 1966). The latest components of the Lovozero complex are numerous veins and dykes of porphyritic
murmanite-bearing lujavrite (Bussen & Sakharov 1972). The average composition of the Lovozero complex is close to that of an agpaitic phonolite but is characterized by (i) an extreme excess of alkalis with respect to alumina ( ( N a 2 0 + K20)/ A1203 = 1.40), (ii) the predominance of Na over K, (iii) a high Fe203/FeO ratio and (iv) a high concentration of rare elements and volatile components. The chemical evolution of the Lovozero complex was unusual in that the
Alkaline rocks o f the Kola Peninsula melanocratic (Fe-rich) components increase during differentiation. The residual products are agpaitic and have relatively high contents of silica, alkalis and rare elements (Gerasimovsky et al. 1966).
Crystallization of the agpaitic magmas
Attempts have been made to determine the physical and chemical conditions under which the agpaitic alkaline magmas crystallized, i.e. to ascertain the temperature and pressure of the volatile components, the redox conditions and the genesis of the associated mineral deposits. Two independent methods were used to determine the crystallization temperatures: (1) investigation of phase equilibria from melting experiments; (2) thermometric investigation of primary inclusions in the rock-forming minerals. The experiments were carried out on a variety of compositions from the Lovozero and Khibina complexes at high temperature and pressure under controlled oxygen and water-vapour conditions (Kogarko 1977) (Fig. 8). These experiments indicated the following. 1 The agpaitic magmas, in contrast with other rock types, crystallized over very wide temperature intervals (more than 300~ which facilitated extreme differentiation processes. This conclusion is in good agreement with that deriving from experimental work by Piotrovsky & Edgar (1970).
bar lOOO
539
2 The agpaitic magmas proved to be surprisingly dry (Fig. 8). It is inferred from the P-T melting curve of the average composition of the agpaitic nepheline syenites that, in the initial stages of crystallization, the water-vapour pressure was fairly low (100-300 bars) resulting in the earlier separation of the leucocratic minerals relative to the melanocratic minerals. At higher water-vapour pressures the miaskitic order of crystallization is observed, i.e. aegirine becomes the liquidus phase and the nepheline field is significantly reduced. 3 Agpaitic magmas were formed under very reducing conditions (approximating the QFM buffer), which is characteristic of mantle-derived basalts, ultramafic rocks and kimberlites. The paragenetic association of agpaitic rocks (i.e. aegirine, aenigmatite and arfvedsonite) is stable only under the conditions of the QFM buffer. At higher oxygen activity aenigmatite becomes unstable, while at lowerJO2 aegirine decomposes and diopside crystallizes. The temperatures established for the agpaites from the melting experiments were confirmed by investigation of micro-inclusions in the minerals (Kogarko 1977). The primary inclusions were shown to be the result of crystallization of trapped melt and to rehomogenize at temperatures in the range 1010-730~ High temperatures of inclusion homogenization in the nepheline of the agpaites also confirm the anhydrous character of the initial magma. The homogenization temperatures of inclusions in the nephelines are commonly in the range 980-1000~ corresponding to water-vapour pressures of 100-200 bars.
u~
Volatile components in the alkaline magmas
500 PH20
'
',..,,
100 600
Q-1
700
800
900
1000
~
T ~-5
A-2
O-6
Q-3
$-7
FIG. 8. Melting relations of a typical agpaitic rock. 1, liquidus; 2, solidus; 3, nepheline; 4, pyroxene; 5, Kfeldspar; 6, aenigmatite; 7, average homogenization temperature of micro-inclusions.
The experimental investigations of the agpaitic nepheline syenites revealed, as indicated above, that their formation took place at low watervapour pressures. Experimental investigations of the solubility of water at different values of T and Prho in a foyaite corresponding to the average composition of agpaitic nepheline syenites and in an agpaitic pegmatite showed that the water content of the hyperalkaline melts approaches 1.5 wt. % (Kogarko et al. 1977a). The activities of acid volatile components such as HF during crystallization of the agpaitic nepheline syenites can also be evaluated (Fig. 9). Fugacities for these components have been calculated using data for the natural paragenesis of agpaitic nepheline syenites and for activities obtained both experimentally and theoretically for Na2 O, SiO 2 and A1203. The HF and HC1
540
L. N. Kogarko cH4 H2S H2 ,
~
500
T,~
-10
log -20-
SO 3
-30
02
FIG. 9. Relationship between the fugacity of the volatile component and the temperature of the magmatic gas phase. fugacities were calculated according to the following reactions: N a F + 89 log(filE)T=
--* 89
AGT(1) 2.302RT
NaC1 + 89 H20 ~ log (fHC1)7
89
HF
(1)
+ ! logfH2o
2
89N a 2 0 + HC1
(2)
AGr(2) 2.302 RT - k log aNa2o+ log aNacl + 1Ogfn20
Water activity has been calculated from the constant of the monovariant equilibrium 89
+ 89
+ H20 --~ NaA1SiEO6.H20
since the paragenesis of nepheline-analcimealbite is very common in agpaitic syenites. The NaF activity was assumed to be unity since villiaumite is present in the agpaites. The NaC1 fugacity was calculated from the reaction 3NaA1SiO4 + NaC1 ~ 3NaA1SiO4. 89
These data suggest that, despite the significant enrichment of the alkaline rocks in F and C1, the activities of HF and HC1 in the equilibrium gas phases of these magmas were extremely low (Fig. 9). This may explain (i) the absence of greisenization and (ii) the lack of pneumatolytic transfer of ore and volatile components in alkaline rocks. Such processes are more generally confined to more acid rocks such as granites. Loss of ore and volatile components across contacts (both internal and external) is untypical, despite their high concentrations, in such agpaitic complexes as Lovozero, Khibina and Ilimaussaq. Minerals containing volatile components (e.g. villiaumite, sodalite and alkali amphibole) are present as accessory or rock-forming minerals. The accumulation of volatile components in the melts (but not in separate vapour phases) leads to a parallelism in the behaviour of volatile components (e.g. F and Cl) and non-volatile lithophile elements (e.g. Na, A1 and Si) which is reflected by the close correlations of these groups of elements in alkaline rocks (Kogarko 1977). Thus a substantial reduction in the activity of the volatile components with the increase in alkalinity in the silicate liquids gives rise to the coupled migration of alkalis with volatiles in such processes as thermodiffusion and diffusion in gravitation fields and under pressure gradients. It is known that alkaline rocks are associated with hydrocarbons (Petersilie 1964). The high content of hydrocarbon-rich gases has been known to cause explosions in mines within alkaline intrusions. The main components of these gases are CH4, heavy hydrocarbons and H E. On the basis of data yielded by experiments on the fugacities of O2 and H20, the fugacities of CH4, H:, CO, CO2, H2S, SO2 and SO3 in the magmatic gas phase of alkaline rocks have been quantitatively estimated from the following reactions: H24- 8902 ~ H 2 0
(3)
log (fH~)r=2A.3Go;(RT+log fH2o- 89 (fo2)r C + 2H 2 ~ CH 4
(4)
log (fcH4)r= -2A.3G2(RT+21og fH2 C + 02 ~ CO 2
log (fco2)7
(5)
Alkaline rocks of the Kola Peninsula 2CO + O2 ---* 2CO 2
(a)
1:10000
AGT(6) 2_302 x 2 R T + log fc02 -- 89log f o 2
log (fco)r
89 + H 2 0 ---* H2S [- 89
-
89l o g
Or ~2 ES[]J
(7)
r~6
AGT(7) 2.302 R T
log (fH2s)7 -
f% + 89 log fs2 + log fR2o
82 Jr- 202 = 2802
log (fs%)r = - o z.Juz x z 1 ( 1 -
~
)
~
1..25ooo L'~,
(8)
89log fs2 + lo8/o2
S2 "~ 302 ---* 2SO3
log (fso3)r =
(6)
541
~C)
(9)
1:5000
, ]"~'-"]'~
Ir~,,
AGT(9) Jr • log fs2 + 3 log fo2 2.302x2RT 2
The data show that the gas phase associated with agpaitic rocks was characterized by very high CH4, H 2 and HzS (Fig. 9); CH 4 fugacity shows a rapid increase with drop in temperature, while the fugacities of CO2 and SO2 are substantially reduced. Comparison of these data with those reported by Petersilie (1964) shows that the gas phase is a product of agpaitic nepheline syenite differentiation and is formed at relatively low temperature. With increase in temperature (other conditions being equal), the contents of CO, CO2, SO2 and SO3 in the gas phase rise. The juvenile nature of H2S in the gas phase of agpaitic alkaline rocks has been confirmed by isotope studies which also show a relatively high content of the heavier carbon isotopes in Lovozero, Khibina and Ilimaussaq (Galimov & Petersilie 1967). The investigations have also solved the interesting problem of the coexistence of highly-alkaline mineral associations (aegirine etc.) in agpaitic nepheline syenites in which Fe 3 + dominates over Fe z + with a distinctly-reducing gas phase whose reducing character increases as the temperature falls.
The Khibina apatite deposits The problem of the genesis of the Khibina apatite deposits is of great interest. With their enormous resources of P, A1, Sr and rare-earth elements these deposits are unique and are of great industrial importance. The apatite-rich rocks have been classified into three sub-phases (Zak et al. 1972): I pre-ore, II ore and III post-ore. The rocks of sub-phase I consist of an ijolite series interlayered with subordinate melteigite, urtite, juvite and malignite, the whole having a total thickness of less than 700 m. Sub-phase II consists of massive feldspathic urtite, ijolite-urtite and
F]o. 10. Geological sections of the Khibina apatite deposits (1, medium-grained nepheline syenite; 2, lujavrite; 3, malignite; 4, melteigite; 5, ijolite; 6, sphenite; 7, patchy apatite ore; 8, breccia; 9, banded ore; 10, lenticular-banded ore; 11, block ore; 12, netlike ore; 13, ijolite lenses; 14, massive urtite; 15, trachitoid urtite; 16, rischorrite; 17, khibinite): (a) Kukisvumchorr; (b) Uksporr; (c) Rasvumchorr. (After Zak et al. 1972.) apatite ore with a total thickness of 200-700 m. The units of sub-phase III are from 10 to 1400 m thick and include urtite, ijolite, melteigite, juvite, malignite and lujavrite. The principal phosphate-ore deposits are found in sub-phase II where the apatite-rich rocks occur in the hanging wall of an ijolite-urtite intrusion (Fig. 10). The deposit is both petrographically and geochemically zoned. The upper zone (apatite-rich ore) has been called patchy and patchybanded ore (Zak et al. 1972) and consists of 60%90% euhedral apatite crystals ranging up to several tenths of a miUimetre across. Clinopyroxene, sphene, feldspars, titanomagnetite and nepheline occur as intergranular minerals. Monomineralic nepheline layers sometimes alternate with monomineralic apatite layers. The lower zone (apatite-poor ore) is composed of lenticularbanded, net-like and block ore. The lenticularbanded ore consists of fine-grained ijolite separated by layers of apatite and fine-grained urtite, whereas the net-like ore is texturally and structurally similar to lenticular-banded ore and differs from it only in the scarcity of urtite and apatite bands. The block ore appears to be pegmatitic. Occasional large crystals of nepheline (up to 15 cm across) occur in the nepheline-apatite rock and in the apatite segregations. The lower zone grades down into massive urtite (Fig. 10) which
L. N. K o g a r k o
542
consists of 75%-90% large euhedral nepheline crystals with intergranular acmitic clinopyroxene, feldspar, titanomagnetite and aenigmatite. Small grains of euhedral apatite are also found in the mesostasis. An apparent late-stage eruptive breccia, containing xenoliths from the lower zone of the apatite ore in an ijolite-urtite matrix, also occurs in the ore zone. The micro-inclusions in apatites and nephelines of the basic apatite deposits have been comprehensively studied. Thermometric studies of the primary inclusions in the rock-forming minerals of the apatite-nepheline rocks indicate that the Khibina apatite ores were formed during the magmatic stages of crystallization. Typical homogenization temperatures for the microinclusions in the apatite and nepheline of the apatite ores and underlying urtites are of the order of 1000-700~ The inclusions homogenize to form alumino-silicate melt (Kogarko 1977). The morphological peculiarities, phase compositions and homogenization temperatures for the inclusions in the rock-forming minerals from the apatite ores and the underlying urtites show similarities. This suggests that the minerals of the apatite ores and their underlying massive urtites crystallized from an alumino-silicate melt under similar physico-chemical conditions and are syngenetic. The phase equilibria of the apatite-bearing ijolite-urtite rocks of Khibina can be approximated in the system NaA1SiO4-CaMgSizO6Cas(PO4)3F (Kogarko et al. 1977b). Composition A in Fig. 11 represents a weighted average of the N a A I S i O4
,//~ 526~
bulk composition of the apatite-bearing complex under consideration (Table 2). Nepheline would crystallize first from a melt of such a composition. Apatite and nepheline would be the next to crystallize as the temperature falls. This crystallization sequence is in agreement with petrographic observations on the rocks. It was shown that the sequence of crystallization of nepheline and apatite is unaltered by applying water pressure to the system (Kogarko & Lebedev 1968). Thus, from the viewpoint of experimental phase equilibria, an alumino-silicate melt with about 2 wt. % P205 could have been the parental magma of the apatite-bearing urtite-ijolite complex. Such a magma would crystallize 10%-15% nepheline before reaching the nepheline-apatitepyroxene cotectic. One outstanding problem presented by the apatite ore-bodies concerns the manner in which segregation of nepheline, apatite and pyroxene occurs. There is a surprising uniformity of mineral grain-sizes in the various zones of the apatite deposits (Fig. 12). Distinct sorting of the minerals in vertical sections is observed, with larger nepheline crystals forming the lower part of the intrusion and smaller crystals of pyroxene and apatite composing the upper part of the deposit. Such sorting suggests the participation of convective flow during formation of the apatitenepheline rocks. This is confirmed by clearly defined primary-flow structures in apatite ores. The distribution of minerals in the magma chamber is the product of crystal settling and convective flow. The model proposed by Bartlett (1969) can be used to determine the distribution of crystals in a magma chamber. Bartlett's equations can be applied to the results of careful grain-size analysis of the minerals in the Khibina
400l ~300 8~I
~ 2~176 I 100
a
d
E=4o[
o 20~i / ) //':; J Y/~/: ~ 1648 ~
1391"
Ca5(POa)3F
CaMgSi206 wt.
%
FIG. 11. Liquidus phase boundaries in the nephelinefluorapatite-diopside system at 1 atm pressure: Cg, carnegieite; Ne, nepheline; O1, olivine; Me, melilite; Di, diopside; Sph, silicophosphate; Ap, apatite; L1 § L2, immiscibility field. Point A represents a weighted average of the bulk composition of the apatite-bearing rocks of Khibina.
-2
-2
log x
-1
0 log x
FIG. 12. Granulometric analysis of apatite deposits (x is the diameter of the mineral in mm) : curve a, apatite of rich ores (Kukisvumchorr, Uksporr and Rasvumchorr); curve b, pyroxene of poor ores (Kukisvumchorr and Uksporr); curve c, nepheline of poor ores (Kukisvumchorr and Uksporr); curve d, nepheline of underlying urtites (Kukisvumchorr, Uksporr and Rasvumchorr).
1
Alkaline rocks of the Kola Peninsula apatite deposits to estimate the distribution of crystals in the magma chamber as a function of melt viscosity (Fig. 13). Unless the melt viscosity approximates to 103 poise, all nepheline crystals of a size corresponding to those in the massive urtite will settle, whereas minerals of grain-sizes corresponding to the other rock units will remain in suspension and move up with the rising liquid. The viscosity calculated for the apatite-ore magma according to the method of Bottinga & Weill (1972) is (2-3) • 103 poise. Convection decreased during cooling, resulting first in the deposition of pyroxene and nepheline of the apatite-poor ores and finally in the deposition of apatite. The coincidence of the size distributions of apatite, nepheline and pyroxene in vertical sections of the intrusion with the position of the hydraulic curves (Fig. 12) suggests that crystal settling in conditions of convective overturn was the main process governing the distribution of minerals during the differentiation of the apatite-bearing ijolite-urtite magma. Many peculiarities of the micro- and macro-structures in the patchy and patchy-banded apatite-rich ores and in the lenticular-banded net-like apatite-
543
poor ores, as well as in the massive urtites, are explained well by this model. The geochemical distribution of the elements in the apatite deposits also accords with the hypothesis of gravitational crystallization differentiation. N2
f
N] 0.8
0.6
0.4
0.2-
,
,
,
,
,
,
1
2
3
4
5
6
Ig~
FIG. 13. Distribution of crystals along a vertical section of the magma chamber: curve a, apatite of rich ores; curve b, pyroxene of poor ores; curve c, nepheline of poor ores; curve d, nepheline of underlying urtites. N2/N 1 is the ratio of the abundance of crystals in the upper part of the magma chamber to the abundance in the lower part; rl is the viscosity.
References BARTLETT, R. W. 1969. Magma convection, tempera-
ture distribution and differentiation. Am. J. Sci. 267, 1067-82. BOTTINGA, H. & WEILL, D. 1972. The viscosity of magmatic silicate liquids: a model for calculation. Am. J. Sci. 272, 438-75. BUSSEN,I. V. & SAKHAROV,A. S. 1972. The Petrology of the Lovozero Alkaline Massif. Nauka, Leningrad (in Russian). DUDKIN, O. B., MINAKOV, F. V., KRAVCHENKO,M. P., KRAVCHENKO, E. V., KYLAKOV, A. N., POLEGOEVA,L. P., PRtPACHKIN,V. A., PUSHKAREV,U. D. & PUNGENEN, G. I. 1984. Carbonatites of Khibina, Apatiti. Nauka, Moscow (in Russian). ELISEEV, N. A., OJINSKY, J. S. & VOLODIN, E. N. 1939. Geological map of the Khibina tundra. Trudy Leningr. Geol. Upr. 19, 1-38. GALAKHOV,A. V. 1975. Petrology of the Khibina alkaline massif Nauka, Leningrad (in Russian). GALIMOV, A. M. & PETERSILIE, I. A. 1967. Isotopic composition of carbon in hydrocarbon gases and CO2 in alkaline rocks of the Khibina, Lovozero and Ilimaussaq massifs. Dokl. Akad. Nauk S.S.S.R. 176(4), pp. 914-17 (in Russian). GERASIMOVSKY,V. I., KOGARKO, L. N. & POLYAKOV, A. I. 1974. Kola Peninsula. In: SORENSEN, H. (ed.) The Alkaline Rocks, pp. 206-21. J. Wiley, New York. --, VOLKOV, V. P. & KOGARKO, L. N. 1966. The Geochemistry of the Lovozero Alkaline Massif.
Nauka, Moscow (in Russian). (English translation by D. A. Brown, Australian National University Press, Canberra, 1968.) KOGARKO, L. N. 1977. Problems of Genesis of Agpaite Magmas, 295 pp. Nauka, Moscow (in Russian). & LEBEDEV, E. B. 1968. Equilibrium in the nepheline-apatite-water system. Geochem. Int. 5, 341-3. - - , BURNHAM,C. & SHETTLE,O. 1977a. Water regime in alkalic magmas. Geochem. Int. 14(3), 1-8. --, KRAMM, U., BLAXLAND, A., GRAUERT, B. & PETROVA, E. N. 1981. Age and origin of alkaline rocks of the Khibina massif (isotopes of rubidium and strontium) Dokl. Akad. Nauk S.S.S.R. 260(4), 1001-4 (in Russian). --& GRAUERT, B. 1983. New data on age and genesis of alkaline rocks of Lovozero Massif; Rb and Sr isotopes. Dokl. Akad. Nauk S.S.S.R. 268(4), 970-2 (in Russian). --, KRIGMAN,L. D., PETROVA, E. N. & SOLOVOVA,I. P. 1977b. Phase equilibria in the fluorapatitenepheline-diopside system and the origin of the Khibiny apatite deposits. Geochem. Int. 14(1), 2738. KUKHARENKO, YU. A. 1958. A Palaeozoic complex of ultrabasic and alkaline rocks in the Kola Peninsula and associated metal deposits. Proc. VMO, Zap. vses. miner. Obshch. 87(3), 304-15 (in Russian). 1971. Metallogenetic features of alkaline formations of Eastern part of Baltic Shield. Trudy Leningr. Obshch. Estest. 72(2), 1-267 (in Russian). -
-
544 --,
L. N. Kogarko
ORLOVA, M. P., BULAH, A. G., BAGDASAROV,E. A., RIMSKAYA-KORSAKOVA,O. A., NEFEDOV,E. I., ILINSKEY, G. A., SERGEEV, A. S. & ABACUMOVA, N. B. 1965. The Caledonian Complex of UItrabasicalkaline Rocks and Carbonatites of the Kola Peninsula and Northern Karelia, 772 pp. Nedra, Moscow (in Russian). PETERSILIE, I. A. 1964. Geology and Geochemistry of Natural Gases and Dispersed Bitumens in Certain Geological Formations of the Kola Peninsula, N au ka, Moscow (in Russian). PIOTROVSKY, J. M. & EDGAR, A. D. 1970. Melting
relations of undersaturated alkaline rocks from South Greenland compared to those of Africa and Canada. Meddr. Gronland, 181, 9, 1-62. POLKANOV, A. A. & GERLING, E. K. 1960. The preCambrian geochronology of the Baltic Shield. 21st Int. Geol. Congr. Report Part 9, 183-91. Copenhagen. ZAK, S. L., KAMENEV,E. A., MINAKOV,P. V., ARMAND, A. L., MmHEICHEV,A. S. & PETERSlLIE,J. A. 1972. The Khibina Alkaline Massif, 175 pp. Nedra, Leningrad (in Russian).
L. N. KOGARKO,Vernadsky Institute of Geochemistry and Analytical Chemistry, U.S.S.R. Academy of Sciences, Kosygin Street 19, Moscow 117975, U.S.S.R.
Index Page numbers in italics refer to figures, those in bold to tables Abontorok ring complex 365-6 absarokite 104 acid porphyrites 196 acmite 63, 149, 341,405,417 actinolite 387 adamellite 449 Admapluton 388, 389 Adrar Bous ring-complex 366 aegirine 71, 75, 77, 156, 338,386, 387,405,409, 411, 452, 479,484, 539, 541 Zr-rich 461 aegirine-augite 71, 75, 337, 338, 339, 405,409, 411 aenigmatite 387, 417,452, 461,479, 539, 541 African Plate, movement of 282-3 agglomerate 136, 403,404, 502 agglutinate tuff see piperno agpaites 473,476, 539 agpaitic rocks (Ilimaussaq intrusion) bottom rocks 481-3 mineralogy and layering 481-2 theories of formation of kakortokite layering 482-3 roof rocks 478-80 layering 479-80 mineralogy 478-9 AillikBay, Labrador 128, 163 aillikite 128, 198,465 akerite 437 gtkermanite 31,32 albanites 107 albite 48, 65, 72, 77, 109, 137, 149, 164, 198, 370, 373, 387, 484 destabilization of 374 secondary formation 465-6 albite-aegirine schists 77 albitites 77 albitization 78,370, 374, 387 alkali enrichment 92, 93 alkali gain 317 alkali loss 72, 76, 78, 86, 87, 88, 92, 282, 317, 324, 420, 465 Zomba-Malosa complex 341 alkali-depletion, volume requirements of 89 alkaline lamprophyre branch 196, 198, 200, 203,204, 205,212, 214, 215,218,219 allanite 305, 387, 388 zoned 387 Aln6 complex (Sweden) 53, 66 aln6ite 68, 128, 136, 138,219, 337, 339, 349, 442, 463, 464-5 CO2-rich 444 Ile Bizard 442 incompatible element concentrations 443 alteration 255,257, 337 hydrothermal 135, 271,370, 407, 465,503, 524 K-silicate 374 metasomatic 170, 351 Na-silicate 370
post-magmatic 138 by seawater 394 alumina saturation 370 alvikite 53, 72, 73, 75, 77, 78 AmbaDongar, carbonatite 71 amphibole 2, 3, 7, 8, 9, 34, 45, 72, 77, 90, 92, 108, 137, 138, 160, 161, 170, 193, 198,202, 243,258, 271, 337, 338, 341,364, 386, 387, 388, 405,409-11, 420, 437,439, 451,466, 520, 521,527 alkali 133, 135, 156, 160, 479, 540 Ba-K-amphibole 198 Ca-amphibole 387 crystallization of 387 pargasitic 63 secondary formation 465-6 sodi-calcic 370 stability of in basaltic magma 228 Ti-amphibole 198 amphibole fractionation 523 amphibolite 360 amygdales 213 analcime 108, 120, 123, 132, 133, 138, 160, 163,202, 217, 338, 405,484, 517,519 origin of 48 primary 29 analcite basalt 140 anatase 131,133, 135, 138, 163 anatectic melts 508, 509 anatexis 48,468 crustal 468 andesine 388 andesite generation, models for 394 andesite 68, 104, 128, 381,504 basaltic 233 continental margin 261 potassic 140 subduction-derived 132 ankaramite 234, 495 ankaratrite 295, 301,303 annite 371,373 anorogenic complexes, within-plate 395 anorogenic magmatism effect of late-stage fluids on 370-6 acid metasomatism 374-6 biotite granites and their mineralization 370 potassic metasomatism 373-4 sodic metasomatism 370, 373 petrogenesis 366-70 Phanerozoic anorthosite association, Air, Niger 363-6 tectonic and structural controls 358-70 anorogenicmagmaticevolution 361-3 influence of Pan-African orogeny 358, 360-1 tectonic constraints on alkaline magmatism 361 anorthite 149 anorthoclase 420 anorthoclase trachyte 234 anorthosite 361,363, 364, 450, 466, 531
545
546
Index
anorthosite (cont.) Air 363,364 labradorite-andesine 461 massif-type 449 anorthositic complexes, evolution of 366 anorthositic enclaves 368 apatite 57, 66, 67, 68, 71, 72, 73, 77, 90, 91, 97, 108, 109, 114, 122, 129, 13l, 132, 133, 138, 139, 164, 166, 168, 234, 243,305, 337, 338,339, 349, 387, 388,405,409, 451,462, 465, 468,475,479,492, 521,541-3 early-crystallizing 466 Na-REE-rich 462 apatite deposits 534, 541-3 apatite glimmerite 139, 163 aplite 405 aplogranite 407, 409 albite 373 aqueous fluid phase, separation of 92, 93 aqueous fluids 23 and alkali loss 86, 88 development from magma 88 arfvedsonite 71, 72, 77, 337, 386, 387, 411,417,479, 481,482, 483,484, 539 K-Ti-Mg-arfvedsonite 156 Mg-arfvedsonite 77 arfvedsonite-riebeckite 452 Argorintrusion (Canada) 90 Argyle, E Kimberleys, W Australia 131,142, 144, 148, 149, 151,163 diamondiferous tuff 104 armalcolite 108, 123, 163, 165 Ascension Island 253-61,269-71 Ascutney Mountain complex 441 ash-flows 421,423 comendite 303 assimilation 313,425,452, 508,524 asthenosphere 4, 21,284 generation of parental magmas 357 partial melting of 64, 65, 283-4 as primitive mantle source 398 asthenospheric bulge 517 asthenospheric plumes 24 astrophyllite 387,405,452 augite 61, 71,107, 128, 135, 137, 234, 282, 339, 386, 409 Fe-augite 387 sodian 410 T-Al-augite 62 augite syenite 451,452, 474, 499 Australia, generalized map 129 autometasomatism 370 Azores 10-11 Azzaba, Algeria 136-7 Ba mineralization 57 Ba-K-amphibole 198 Ba-K-feldspars 198 Ba-K-phlogopite 198 Ba-Ti-phlogopite 198 back-arc spreading 427 baddeleyite 71,405,424 Bagnold effect 306, 369
Baikal-Aldan Belt, USSR 140, 143 Baltic Shield (Kola Peninsula), alkaline rocks of 53143 Palaeozoic alkaline complexes 533-43 Proterozoic alkaline complexes 531-3 Bambouto, Cameroon 278 barite 71, 74, 78, 109, 133, 138,462 Sr-barite 120, 122, 164 barroisite 387 bartonite 200 baryte see barite basalt 5, 8, 35, 68, 104, 107, 139, 154, 255,259, 271, 273,278, 279,280, 281,282, 284, 309, 314-15, 317, 318, 321,323,423, 442, 452, 454, 459,465, 504, 520, 521,524 alkali 19, 30, 45, 103-4, 148, 151,167, 179, 196, 232, 258, 262, 293,294, 296, 297,298, 300, 301, 302, 307, 417,444, 493, 517 evolved 37 fissure-fed 303 genesis of, Massif Central 526-7 alkali olivine 15, 32, 33, 34-6, 298,442, 444 alkaline, transitional 296, 298-9, 303, 304, 307, 309 analcite 140 continental margin 261 ferrobasalt 296, 298,303 Gardar 456, 460 genesis of 30 K-rich 128 leucite 140 oceanic 381 olivine 294, 298 olivine-augite 303 primary 443 shoshonitic 135 subduction-derived 132 tholeiitic 230, 232, 241,435,444, 495 trachybasalt 315 Walvis Ridge 263,264 basalt-trachyte relationships 320-3 fractional crystallization of basaltic magma 320-2 partial melting of mafic deep continental crust 323 significance of composition gaps 322-3 basanite 2, 3, 4, 10, 15, 32, 36-7, 196, 232, 277, 278, 280, 282, 294, 296, 297, 301,317, 442, 504, 506, 517, 519, 524, 526 leucite 135 nodules in 6 olivine 39 bastnaesite 71, 73-4 batholiths 381 calc-alkaline, Iforas 398 Sierra Nevada 127 unroofing of 394-5, 398 W Iforas (composite calc-alkaline) 382, 388,398 White Mountain 433-4, 435 beforsite 53,465 Belknap Mountain complex, geology of 434-5 Benfontein (S Africa) 65 Benin-Nigeria shield 360 Benioffzones 144, 381 fossil 143, 168 benmoreite 37, 234, 296, 298,303, 304, 321,322, 323, 417, 521
Index Benue trough 362 related to Cameroon line 283 bimodal suites 278 bimodality 65, 67, 304, 307, 308, 309, 341 basalt-trachyte volcanism (Gregory rift) 299 magmatic 5-6 transitionally alkaline suites (Kenya) 293, 303, 304 Bioko (Fernando P6o) 280 biotite 37, 71, 72-3;73, 77, 90, 104, 137, 200, 214, 320, 337, 338,339, 349, 383, 387, 405,420, 451, 465, 521 annite-rich 462 castellated 193 F-bearing 466 green 388 Mg-biotite 387 biotite granite 436, 452 mineralization of 370, 371,372 biotite mafurite 39, 42, 165 biotite peridotite 464 biotite pyroxenite 464 biotite ugandite 165 biotite-annite 387 biotite-phlogopite 193,202 Biu Plateau, Nigeria 277-8 blairmorite 335 block faulting 144 Blosseville Kyst shelf 489-90 Bobi-Sequela, Ivory Coast 136, 144, 149, 163 Bohemian Massif, Czechoslovakia 135 Borrolan syenite (NW Scotland) 75 bottom cumulates 479 boundary layer 308 boundary-layer liquids 306 boundary-layer migration 331 Bouvet Island 253-61 Brazil 140 Brazilian Precambrianshield 403 breccia 122, 123, 136, 139, 144, 212, 405,437, 498, 499, 502, 503, 508,542 autolithic 142 contact intrusive 124 hyaloclastite 280, 282 igneous 437 intrusive 436 lamprophyric 8 pyroclastic 142 quartz porphry 499 rheomorphic 441 syenite 365, 501 volcanic 503 breccia pipes 436 breccia plugs 503 brecciated greisens 375 brecciation 67, 79 of roof 484 roof cumulates 474 breunnerite 203 britholite 405,409 BultfonteinMine 100, 170 Bultfontein pipes, S Africa, inclusions from 24 buoyant ascent 306, 308 buoyant transport, of magma 233
547
Ca-amphibole 403 calc-alkaline lamprophyre branch 196, 198, 200, 202, 203, 204, 205,212, 214, 215, 217, 218, 219 calcite 54, 66, 67, 69-70, 71, 73, 89, 90, 91,133, 135, 138,203,462, 467 in lamproite 97 calcite carbonatite 53, 72, 73 caldera collapse 228, 234, 278, 305,308, 309, 315, 320, 322, 324, 325,330, 331 caldera complexes 433 caldera formation 314 Krakatoa-type 315, 331 caldera volcanoes 313-15 calderas 271,278,299,421,422, 423,425,428 trachytic 315 Cameroonline, W Africa 273-89 age 273-4, 277 geology and petrology 277-82 continental sector 277-80 oceanic sector 280-2 mantle sources and generation of the basic magmas 283-6 origin of 282-3 salic rocks, origin of 286-9 Cameroon, Mount 279 camptonite 212, 219, 339, 349, 352, 442, 493, 508 Canary Islands 54, 63 Canaveseline 133 cancarixite 132 cancrinite 65, 66, 338,405,467, 534 Canning Basin, W Australia 129, 130 Cape Verde Islands 54, 64, 67, 69, 143 carbonate 3, 7, 8, 9, 132, 193, 203, 338,349, 468 primary magmatic 467 secondary formation 465-6 stability of 166 carbonate-saturated system ((Si + A1) - (Na + K) - Ca) 68 carbonatite olivines 90 carbonatites 29, 38, 44, 68, 69-74, 86, 109, 128, 196, 215,296, 297, 306, 336, 337, 339, 354, 376, 383, 442, 444, 462-5,492, 534, 535 alkalic, development of 92 calcitic-dolomitic 89 calcite 53, 72, 73 calcitic toankeritic 86 defined 53 dolomite 53, 72, 73, 75 intrusive 71-3 liquid immiscibility and genesis of 45, 47 mineralization 73-4 natrocarbonatite magma 69-71 and nephelinites 53-79 normal, development of 91-2 carbonatization 535 Cargillcomplex (N Ontario) 65 cassiterite 277, 375, 376, 405 catapleiite 417,424 cauldron subsidence 383,452, 499, 503 cedricite 108 centralcomplexes, Gardar province 450-2 Cerro Manomo, carbonatite complex, silicified 403 CH~ fugacity 541 chalcopyrite 376
548
Index
(~hannel-Island lamprophyres 135 Chaone complex 336, 337 metavolcanic rocks 339 nepheline syenite and syenite 338, 341-4, 345,347, 351 charnockitic rocks 360 Chelima, India 139, 141, 142, 144, 149, 151,163 Chikalacomplex 336, 337, 339 nepheline syenite and syenite 338,341-4, 345,347, 351 quartz syenites 352 Chilwa alkaline province (N part) petrochemistry of 335-54 petrography 337-9 Chilwa Island, carbonatite complex 337, 339, 351 dykes 339 nepheline syenite and syenite 338,345-6, 348, 349, 352 Chinduzi complex 336 nepheline syenite and syenite 338, 341-4, 345, 347, 351 Chino Valley, Arizona 128 chkalovite 484 chlorite 108, 132, 164, 369, 374, 387, 388 chromite 120, 123, 133, 135, 136-7, 136, 137, 138, 163 Ti-Mg chromite 131 Zn-Mg-chromite 204 cinder cones 120, 142, 143, 234, 277,278,279, 280, 282, 517 monogenetic 298 clinohumite 77 clinopyroxene 3, 7, 8, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 44, 45, 56, 108, 120, 133, 149, 156, 164, 165, 167, 200, 202, 218, 258, 324, 326, 364, 365,368, 387, 388,420, 426, 458,465,467, 520, 524, 541 acmitic 424, 541 alkali 499 aluminous 279 analyses of 62 clinopyroxenite 4 C1, role of 466-7 CO2 61, 64, 142, 349 and magma genesis 36-44 CO2 fugacity 86, 541 COdH20 ratio 239 Coc Pia and Sin-Cao, Upper Tonkin, N Vietnam 138-9 cocite 107, 138 Colima Graben, Mexico 128, 146 collision 360, 395 and calc-alkaline magmatism 398 Tuareg shield and W African craton 394 columbite 370, 373, 376 comendite 6, 269, 299, 331,341,452, 462 composition gaps 5-6,10 mafic and felsic Si-undersaturated rocks 417-18 significance of 322-3 compositional gradients 478 concentration gradients 478 cone sheets, alvikite 73 cones (volcanic), overlapping 491 conglomerates 280 conjugate immiscible liquids 72 contact metamorphism 218
contamination 437 crustal see crustal contamination of mantle 23 of OIB source region 265 radiogenic Sr from groundwater 425 sea-water 271 sedimentary 264 selective 412 continental margins 106 continental regions, isotopic data 19-20 convecting upper mantle, as source of MORB 285 convection 450, 482 double-diffusive 4 7 8 , convective flow 542 convective mixing 324-5 convective overturn 325,326, 482, 543 Cr-diopside 61, 62, 122 Cr-pyrope 122 Cr-spinel 114, 123, 132 Crater of Diamonds State Park, Prairie Creek, Arkansas 97, 104, 107, 122-3, 125, 142, 144, 146, 148, 156, 160, 163, 168 crater lakes 278 cratons 21, 22 crust-mantle system, models for chemical evolution of 261-2 crustalassimilation 166, 452 crustal attenuation 453 crustal contamination 24, 63, 154, 155, 165, 168, 253,273, 286, 287, 293, 323,332, 394, 407, 408, 412, 435,437, 439, 441,468,475, 509, 519 agpaitic magma 475 fractionating magmas 287 mantle-derived melts 441 Massif Central magmas 523-5 progressive 407, 408 and silica-undersaturation 440 crustal doming 361 crustal extension 215, 330 crustal intrusive processes 63-4 crustal melting 321,435 crustal shear zone (Ngaound6r6 fault) 283 crustal shortening 427 crustal thinning 427 crustal weakness, pre-existing zones of, controlling magmatism 444-5 crust, involved in magmatic processes 441 cryolite 374, 466, 467 cryophyllite 370 crystal accumulation 277 crystal boundaries, migration along 73 crystal fractionation 63, 92, 168, 170, 234, 255, 269, 279, 281,285, 331,332, 351,352, 435,452, 461,467,468,485,499 advanced 459-60 closed-system 450 low-pressure 287 mechanisms 323-4 crystallayering, sub-vertical 411,412 crystal settling 306, 313,482, 542, 543 differential 478 crystal-liquid fractionation 304, 305, 306, 309, 351, 428 crystal-liquid processes, low-pressure 255
Index crystallization 308, 329 of amphiboles 387 of anhydrous minerals 91 downward 480, 508 intermittent 482 low-pressure 32 shallow 318 Ti-magnetite 461 crystallization processes 323 crystallization rate, increase in 485 crystals flow sorting of 366 gravitational separation of 482 cumulate blocks, hornblende-rich 288-9 cumulate layering 466 cumulates 62, 72, 77, 219, 236, 269, 387, 437 anorthosite 368,461,468 bottom 474 clinopyroxene 460, 492 dunite 492 flotation 461,479 gabbroic 331,451 layered 451 leuconorite 368 olivine 237,249, 492 plagioclase-rich 320-1 pyroxene 237 pyroxenite 64, 492 roof 474 sodalite (naujaites) 467 syenogabbroic 451 troctolitic 461 wehrlite 492 cumulus augite 421 cumulusnepheline 421,428 cumulus olivine 436-7 cumulus processes and carbonatite rock types 92 cumulus sequences 442 cumulus tabular perthite 405 cumulus textures 411,438,439 current bedding 482 Cuttingsville complex, geology of 436, 440 dacite 104 Daly gaps 296, 299, 301,303, 305 dalyite 135, 164, 269 davanite 123 Deep Springs Valley, California 127, 155-6 degassingofH20 271 dehydration 262, 264 down-going oceanic lithosphere 394 density control 461 density hump 307, 308 ferrobasalt 308 density inversion 39 density sorting 482 depletion events 20,168 desilication 405,406 deuterium, depletion in 271 diamond sources, primary 198 diamonds 24, 95, 97, 104, 114, 120, 122, 131,136, 139, 164, 167, 168 diapiric segregations 418 diapiric uprising 39
549
diapirs 47 asthenospheric 427 diatom ooze 264 diatremes 114, 129, 142, 144, 193,436, 517 inIndia 139 kimberlite 54, 95, 142, 143 lamproite 122, 142, 143 diatremic eruption 8 diatresis liquid 9 mafic melt 8-10 differentiates, high-pressure 506 differentiation 5, 39, 86, 219, 233,249, 341,408,437, 462, 541 carbonatite magmas (computer modelling) 90 continuous 6-7, 11 crystal-liquid 407 due to crystal fractionation 235 extreme 539 geological evidence for depths of 318-20 insitu 451 influence of magma chamber processes 306 liquid-state 303,450 low-pressure 293 shallow 320 differentiation mechanisms 331 differentiation processes, complex 168 diffusion 481,483 diopside 21, 30, 32, 49, 96, 97, 108, 109, 114, 123, 129, 131,132, 133, 135, 136, 138, 139, 148, 156, 165, 166, 168, 170, 217 Cr-diopside 61, 62, 122 ferri-diopside 200 Ti-diopside 114 diorite 360, 366, 433, 434, 436, 437, 439, 441 porphyritic 436 disaggregation 321 disequilibrium melting 217 disequilibrium textures 525 distension 361,395 djerfisherite 200 dolomite 39, 71, 73, 89, 90, 91 ferroan 139,203 dolomite carbonatites 53, 72, 73, 75, 78 dolomitization 203 domal uplift 361 domes, diorite 499 doming 353-4 resurgent 422 dunite 61,167, 460, 534 recrystallized 235 dyke-swarms 335, 381-2, 382, 383, 388-9, 395, 398, 426, 451,456, 476, 491,495,498 acid 398 basic 454 Fox Bay 455 Gardar, distribution and chronology 453-7 lamprophyre 215 radial 422 tholeiitic and alkaline, genetic link 508 dykes 129, 138, 142, 236, 280, 365,403,421,433, 503 agpaitic 461 albitite 531 alkali olivine basalt 438
550 dykes (cont.) alkaline syenite 534 aln6ite 339, 438 aln6ite-monchiquite 454 alvikite 73 basaltic 282 basanite 282, 438 bostonite 531 camptonite 339, 355,438,493, 508 carbonatite 47,467 ChannelIslands 135 Chilwa Island 337 concentric 339 diabase 531 doleritic 454-5 felsic 438,440 felsite 383 Fortification dyke, Colorado 127 giant, Gardarprovince 450-2 hawaiite 426, 451 Holsteinsborg 116 ijolite 339 kersantite 454 kimberlite 114 lamproite 114, 123-4, 127, 142 lamprophyre 195,219, 495,499, 502, 535 lamprophyre dyke, Helam mine 96, 97 late-Gardar 456 lujavrite 478, 538 malignite 534 melasyenite 138 mica peridotite 120, 131 microfoyaite 338 microgranite 337 microsyenite 337 minette 128 monchiquite 438,442 nephelinite 339 Nunarssuit-Isortoq 460 okaite 436 olivine basalt 451 olivinelamproite 131 orientation of, Trans-Pecos 426 oversaturated 341 peralkaline rhyolite 501 phlogopite-olivine lamproite 120 phonolite 337, 338, 339, 349, 350, 354 pillow 499, 504, 508 porphyrite 531 porphyry 383 potassic 493 prowersite 128 quartz microsyenite 323 rare-earth melasyenite 408-9 rhyolite 303,383, 508 ring dykes 337, 422, 434, 451,492, 503 monzonite-syenite 436 nordmarkite 437 porphyritic quartz syenite 436 syenite 495 syenite-quartz syenite 433 saturated and oversaturated 337-8 sheared, meta-kimberlites 136 s61vsbergite 337, 341
Index spessartite 454, 531 syenite 120, 135 tholeiitic 493 tinguaite 535 trachyte porphyry 535 ultramafic lamprophyre 463,464, 465 ultrapotassic 133,463,464, 465 Velasco province 405 E African rift system 6, 136, 293, 335 tectonic evolution of 293-4 eclogite 217 edenite 387 element mobility 257 emeleusite 462 Emuruangogolak 305, 306, 314, 315, 316, 317, 320-1, 323, 324 trachytes 321-2 enclaves, anorthositic 368 Enoree Vermiculite district, S Carolina 109 enriched zone 8 enrichment events 100, 168,237 double-enrichment event 265 enrichment processes 4 enstatite 32, 40, 41, 44, 122, 165, 166 epididymite 462 epidote 369, 387, 388 epithermal neutron activation analysis (NAA), trace elements, OIBs 254-5 equilibrium melting, MORB source 286 erdite 200 erosion 278,418 erosion, selective 142 eruption sequence, Hawaiian 227, 230 post-caldera alkalic stage 228, 230, 234-7 post-erosionalalkalic stage 230, 237-8 pre-shield alkalic stage 227, 230, 232-3 tholeiitic shield-building stage 227-8,233-4 eruptions basalt 308 diatremic 8 explosive 234, 271,307-8,309, 331 fissure, basaltic 320 high-velocity 4, 8, 9, 10 OldoinyoLengai 86, 87, 88 preferential, of salic magmas 308 pyroclastic 309 shield-building 309 small-volume 232, 237 basalt 233 trachytic 309 tufts, air-fall and ash-flow 315 essexite 437, 531 Etinde, Cameroon 279-80 eudialyte 269, 338,405,417, 424, 452, 461,479, 481, 482, 483,484, 538 Exetervolcanics 135 explosion craters 278,279 explosion craters, recent 297 extension 22, 426, 427 extensional tectonics 320 Farallon Plate 427 fault lines 144
Index fault reactivation 357, 361 fault swarms, Gregory rift 319 faulting, Basin and Range 426 faults 335 normal 319 recent, distribution of (Kenya) 319-20 tensional 361 transcurrent 357, 361,362, 376, 383,389, 395 wrench 398 fayalite 45, 47, 324, 337, 370, 386, 387 Fe enrichment 307 Fe-augite 387 Fe-dolomite 139, 203 Fe-Ti tholeiites 504-5 feldspar 7, 29, 49, 73, 75, 138-9, 144, 146, 202, 279, 387, 475,479, 481,482, 541 alkali 77, 337, 338, 349, 373,420, 451,462, 479, 521 Ba-K-feldspars 198 K-feldspar 71,160, 193, 198,217, 339, 387 feldspar (alkali) fractionation 304 feldspathoids 7, 29, 144, 146, 149, 160, 198,202-3, 338,438 felsic alkaline rocks 45-49,417,419-20, 433,435, 438 felsite 503 Fen (S Norway) 66, 74 fenites 66, 72, 76, 86, 342, 534 aegirine-albite 77 feldspathic 77 K-rich 77, 78, 339 mobilization of 351 syenitic 77, 78 fenitization 2, 53, 69, 72, 74-7, 78, 92, 341,351,412, 436 carbonatite 77 earlyGardarlavas 465 N Chilwaprovince 336 potassic and sodic 86 silicate 77 fenitizing fluids 65-6, 93 ferri-diopside 200 ferrisyenite 368 ferro-pseudobrookite 163 ferroactinolite 373 ferrobasalt 296, 298, 303 ferrocarbonatite 53, 71, 72, 73, 75, 78 ferrodiorite 449,450 ferrosalite 475 ferrosyenite 450 ferruginous clay 264 filter-pressing 417-18 fitzroyite 97, 107, 108, 136 flood basalt 449,454 continental 459 Tertiary, Greenland 458 flood lavas 423 flotation cumulates 461,479 flow differentiation 365 flow sorting, crystals 366 fluid evolution, stable isotope constraints 271 fluid immiscibility 270 fluid inclusions 66, 67, 68, 72, 376, 467, 481 Ascension I., granite xenoliths 270
55I
fluid phase, boiling of 375 fluid phases 477 fluid reactions, in biotite granites 370 fluids, late stage 74 fluoramphiboles 77 fluorapatite 71,77 fluorides 466 fluorine 386, 467 fluorine fugacity 77 fluorite 71, 78, 92, 198, 338,374, 383, 387,420, 462, 466, 467 precipitation of 74 secondary formation 465-6 flux melting 7 flux migration 7, 8 Fort Portal(SW Uganda) 72-3 Fortification dyke, Colorado 127 forsterite 44,97, 149, 217 fortunite 103, 132 foyaite 338, 342, 343, 349, 353,405,407,408,410, 411,412, 436, 437, 451,462, 466, 467, 474, 475, 478,499, 508, 535, 537 sodalite 478,479 fractional crystallization 5, 34, 68, 86, 87, 93, 104, 255, 303, 313, 331,423, 525 of basaltic magma 320-2 closed-system 318 hornblende-bearing assemblages 288-9 low-pressure 365 Massif Central magmas 520-3 nephelinitic and ijolitic magmas 65 olivine s6vite magmas 90, 92 problems 326 fractional melting 6, 167,423 fractional resorption 170 fractionation 56-7, 60, 62, 63, 64, 68, 72, 77, 104, 105, 148, 166, 167, 217, 301,305,351,354, 428, 434, 510 alkali feldspar 304, 341,351,408,412 amphibole 34, 351,521,523 carbonatite 75, 76 centripetal 365 clinopyroxene, extended 460 closed-system 321,492 of comendite magma 270 complex 33 dry, basaltic magmas 460 dynamic 61 early mafic minerals 387 eclogite 39 extreme 409 fayalite 351 feldspar 326 high-level 343, 352 in situ 466 intense, of K-feldspar 387 K/Na 44 low-pressure, benmoreite residues 461 and magma chamber layering 368 magnesian magmas 281-2 nephelinite, olivine-poor 55 olivine 36, 37, 38, 39 extended 460 phlogopite 149
552
Index
fractionation (cont.) plagioclase 260, 301,304, 307, 412, 521 pyroxene 61-3,351 sphene 521 titanomagnetite 307 U-Pb 524 see also crystal fractionation; liquid fractionation; crystal-liquid fractionation fractionation trends 32 fracturing 361 gabbro 71,196, 273, 361,363,433,436, 437,439, 466, 493,498, 502, 531 kaersutite 499, 508 leucogabbro 361,363, 366, 369,437 Lilloise 501-2 miarolitic 318 olivine 531 tholeiitic 495, 497-8 gabbro-diorite 503, 531 gabbro-norite 531 galena 376 galena deposits, Pb isotope compositions 504 Galore Creek, British Columbia 128 Gardar Province, S Greenland 449-68 alkali metasomatism in vicinity of Gardar intrusions 465-6 anorthosite cumulates and benmoreite residue 461 carbonatites and ultramafic lamprophyres 462-5 central complexes and giant dykes 450-2 compositions of Gardar basic rocks 457-60 cycles of Gardar magmatism 453 distribution and chronology of Gardar lavas and dyke-swarms 453-7 highly differentiated salic magmas 461-2 petrogenic summaryand conclusions 467-8 stratified magma chambers 452 volatile components of Gardar magmas 466-7 garnet 33, 34, 36, 38, 41, 45, 97, 123, 164, 165, 218, 258 melanitic 198 residual 442-4 Ti-Zr-garnet 200 garnet lherzolite 40 garnet peridotite 2, 16-17, 21, 25 gascontent oflavas 233 gas coring 320 gasexsolution 9, 10 gas phase 541 gaseous transport 166 gases hydrocarbon-rich 540-1 importance of 4, 9 Gaussberg, Antarctica 106, 137-8, 141,142, 149, 154, 156, 160, 163 gaussbergite 107 geochemistry, lithospheric control on 468 geothermal gradients 4, 7, 8, 10 geothermometry in carbonatites 89 Fe-Ti oxide, Mont-Dore, France 521,523 geotherms, upward migration of 354
gieseckite 405 glimmerite 16 apatite 139, 163 globular structures felsic 193 lamprophyres 213-15 gneiss 77, 128,278,337, 360, 506, 534 Archaean 498,499 crustal 508 Gondwanacoalfields, India 139, 142, 151 Gondwanaland 137, 357, 360 fragmentation of 361 Gough Island 253-61 grabens 294 Graciosa (Azores) 6, 7, 11 grain boundaries 2, 8 granite 104, 105, 273,278, 336, 337, 339, 341,346, 351,361,363,383,386, 405,408,421-2, 433, 435,441,462, 495,498, 499, 502, 503, 508, 509 A-type 398,462 albite 370, 373 alkali 384-5,433,451,533 biotite 361,434, 436, 437,452 mineralization of 370, 371,372 of crustal origin 440 crystallization of 270, 271 foliated 360 green, Ilimaussaq 462 hastingsite biotite 434 hypersolvus 383, 385 metaluminous 384 peralkaline 386 metaluminous 376, 387, 389 biotite 433-4 monzogranites 361,388,389 peralkaline 269, 341,361,376, 384, 387, 389, 499 porphyritic 360 riebeckite 412 S-type, absence of 394, 398 sub-solvus metaluminous 391 two-mica 360 granite liquid permeation 370 granite porphyries 383, 385, 387,434 granitic minerals, instability of 374-6 granitic plutonism 218 granitoids 360, 381,388,398, 399 Iforas 394 granodiorite 360, 388,436 granophyres 399, 503 granulites 524 charnockitic 336 Iforas 393 lower crustal 467 graphite 467 graphitization, lamproite diamonds 168 gravitational crystallization differentiation (hypothesis) 543 gravitational separation, of crystals 482 gravity anomalies, positive 360, 381,474 gravity separation 91 Greenland, East: Tertiary alkaline magmatism 489512 alkaline dyke-swarms and associated lavas 493-5, 504, 506, 508
Index Blosseville Kyst 501-2 discussion and petrological models 504-12 evolutionary trends 510-11 Gardiner complex 492-3, 504, 506 geochemistry 503-4 Hold with Hope 503 Kangerdlugssuaqdistrict 498-501,504, 508 Kap Gustav Holm complex 495 Kialineq centre 495, 497-8 Mestersvig area 502-3 Nualik centre 498 Prinsen af Wales Bjerge 491,504, 506 tectonic framework 489-91 Traill 503 Greenland, South: Gardar Province 449-68 Gregory rift 297, 301,302, 303, 318 gregoryite 53-4, 69 greisen 373 brecciated 375 greisenization 374-5 absence 540 groundwater 74 contamination by 425 Guinea, Gulf of 280-2 Haleakala, lavas 242, 243 half-graben 294 halides 193 halite 467 halogen loss 341 halogen-bearing systems 47-8 halogens 30, 77, 321,328,467 harzburgite 40 hastingsite 387, 451 hastingsite biotite granite 434 hastingsite quartz syenite 434 hafiyne 109,279, 517, 519 Hawaiian alkaline volcanism 227-50 eruption sequence 227-38 experimental constraints--depth of origin 238-9 geological setting 227 isotopic constraints on mantle sources 239-41 rare-gas constraints on mantle sources 241-2 trace-element constraints on magma sources 242-6 volume relations--magma storage at depth-xenolith assemblages 246-9 Hawaiian eruptive products 229 Hawaiian Islands 227 Hawaiian volcano (hypothetical), evolutionary stages 248 Hawaiian-Emperor volcanic chain 228, 229, 231, 241 hawaiite 5, 15, 230, 234, 235,243, 255,259, 278, 279, 280, 323,417, 426,454, 459,465,493, 495,506, 520 nepheline 37 3He/4He ratios 241-2 hedenbergite 370, 386, 387, 479 Helam Mine, Transvaal 96, 97 hematite 74, 133, 373 heterogeneities, caused by mantle metasomatism 44 heteromorphism 215 heulandite 164 HF and HC1 fugacities 539-40
553
HFS (high field strength) elements 425 HighwoodMountains, Montana 127, 163 Hills Pond, Kansas 114, 121, 142, 144, 149, 156, 160, 163 H20 in Ascension I. granite xenoliths and magmas 270-1 H20 and magma genesis 36-44 H:O fugacity 86, 540 H:O/COz ratios, low 165 Holmeade Farm, UK 133 Holsteinsborg, W Greenland 114, 116, 120, 141,142, 148, 149, 156, 160, 163 Homa Mountain (W Kenya) 57, 69 hornblende 288, 369, 387, 388,405 paragasitic 410 hornblende monzonite 495 hot spots 253,262, 263, 264, 361 transient 169 HREE 149 hydration 257 hydrocarbons 540-1 hydrothermal activity 517 hydrothermal alteration 135,271,370, 407, 465,503, 524 hydrothermal discharge 230 hydrothermal fluids 74 F-bearing 387 hydrothermal overprinting 370 hydrothermal veins 426 hydrous fluids (LIL- and LREE-enriched), crystallization of 16 hypersthene 132, 419 icelandite 233 Ichoualenring structure 383 Iforas (Adrar des Iforas), alkaline magmatism in 381-99 ages, isotopic and main geochemical characteristics of the calc-alkaline and alkaline transition 389-92 general geology 381-2 Pan-African orogenic context 388-9 Iforas-Ahnet area, structural map 382 igneous lamination 321 igneous systems, with K, Na and A1 104-6 ignimbrite 278, 363, 382, 383, 389, 423,426 comenditic 361 rhyolitic 274, 278 Iivaara (Finland), ijolite intrusion 67 ijolite 54, 60, 63, 66, 68, 77, 88, 296, 306, 339, 349, 436, 534, 535, 541 ijolite-melteigite 532 ijolite-urtite 542 Ilimaussaq intrusion 473-85 the agpaitic magma 475-8 crystallization of 478 development of 475-6 properties of 476-8 bottom rocks 481-3 final stages: the lujavrites 484-5 geological setting 473-4 roof rocks 478-80 structure of the intrusion 474-5
554
lndex
ilmenite 73,90, 120, 132, 133, 135, 138, 139, 163, 164, 170, 198, 203,204, 234, 364, 408 magnesian 97 Mn-ilmenite 131 Imilik gabbro 497-8 immiscible liquids 440 immiscible separation 72, 86, 88, 90, 93 inclusion-bearing alkaline rocks 15-25 in continental rift valleys 18-19 in ocean basins 16-18 in other continental regions 19-20 inclusions apatite-amphibole pyroxenite 16 basalt 501 garnet peridotite 16 glimmerite 16 high-density 24 kaersutite gabbro 499 mafic and ultramafic 18, 20, 21 mantle 16 MARID (mica-rutile-ilmenite-diopside)rocks 16 mica clinopyroxenite 16 naujaite 484 picroilmenite 123 saline 270 spinel lherzolite 16 spinel peridotite 16 vapour-rich 270 incompatible elements 459 abundances 284-5,443 depletion in 167 enrichment 237,439, 440 India, generalized geology 140 Indonesian arc volcanics 140 intra-plateactivity, Niger-Nigeria 358 intra-plate volcanism 284 intrusions 68,219 agpaitic 48 alkali gabbro 45 alvikite 69 augite syenite 461 bimodality of in E Africa 67 breccia 499 carbonatite 65, 67, 72 ferrocarbonatite 69 gabbro 450, 497 hypabyssal 464 ijolite 65, 67, 75, 76 Kangerdlugssuaq 499,501 leucite lamproite 144 phonolite 72 polyphase 535 Prairie Creek 122 quartz syenite 426 rhyolite 426 salic 450, 489 peralkaline 452 s6vite 69 syenite 497,499 Trans-Pecos 422-3 intrusive complexes and Cameroon line 274, 277 carbonatite 301 ijolite 301
nepheline syenite 301 Iskou complex, anorthositic enclaves 368 island arcs 106, 360 isotope studies, co-magmatism of Kenyan alkaline series 305-6 isotopic data, Loihi lavas 232 isotopic systems, to determine crustal contamination 523-5 Italy, NW 133, 142, 148, 149, 151,156, 160 Roman Province 135, 155, 163, 165 Tuscany 135, 155 Jacupiranga, Brazil 73 jacupirangite 73,436 jeppeite 108, 129, 163,200 jumillite 103, 107, 108, 132, 133, 138 Jungunicomplex 337,352 nepheline syenite and syenite 338, 348 petrochemistry 344-5 juvite 532, 541 K enrichment 40, 41,42 K-arfvedsonite 108, 138 K-feldspar 71,109, 160, 193, 198, 217, 339, 387 K-for-Na exchange 374 K-rich rocks, experimental studies 164-6 K-rich volcanism 144 K-richterite 114, 120, 123, 129, 132, 133, 138, 160, 170 K-riebeckite 108, 123, 156 K-Ti-Mg-arfvedsonite 156 K-Ti-richterite 108, 156, 200 K-trachyandesite 136 Kaangankunde (Malawi),carbonatite 74 kaersutite 36-7, 37, 116, 169,202, 437 kaersutitegabbro 499, 508 Kaiserstuhl(S Germany) 54, 70 Kajan River, Kalimantan (KAJ) 139 kajanite 139, 140 kakortokite 474, 481-3 kalsilite 32, 36, 44, 104, 109, 127, 135, 136, 149, 166 kamafugite 166 kamafugitic rocks 106, 140, 148 Kamas and Moon Canyon, Utah 120, 142, 144, 148, 151,160, 163, 164 kaolinite 374 Karoo dolerites 352 Karoo volcanic cycle 335 kataphorite 338,411 katungite 39, 41,45, 136, 165 Kenya rift 293-309 associations of alkaline igneous rocks 294-6 caldera volcanoes 313,314 distributions and volcanological characteristics of the suites 296-300 general characteristics of the province 293-4 influences of magma chamber processes on differentiation 306-9 isotope ratios 305-6 petrology and geochemistry 301-4 basanitic and alkali basaltic suite 301-2 mixed volcanic associations 303-4
lndex nephelinite-carbonatite suite 301 transitionally-alkaline series 303 silicic rocks and caidera volcanoes 313-29 trace element characterization of volcanic suites and series 304-5 Kerguelen Islands 143 Kerguelen-Gaussbergaseismic ridge 137 Kerimasi (N Tanzania), carbonatite 69, 70, 71 kersantite 204 khibinite 535 Kidal massif, ring-complex 383-5,387, 389 Kilauea 230, 241 Kilimanjaro 298,299, 303 Kilombe, syenite nodules 324 Kimberley Basin, W Australia 129 kimberlite clan 95-100 kimberlite pipes 104 kimberlite-lamproiterelationships 169-70 kimberlite 2, 3, 4, 15, 29, 32, 38-9, 39, 44, 45, 47, 54, 65, 66, 68, 86, 103-4, 104, 107,109, 120, 124, 136, 139, 140, 144, 146, 148, 151,156, 166-7, 191,193, 198,215,284, 539 definition 95 distinguished fromlamproite 97 Group I (non-micaceous) 19--20, 25, 96, 99, 170 enrichment event 100 Group II (micaceous) 20, 23, 24, 25, 95-6, 97, 99, 105, 136, 168, 170 chemistry 98 source region 100 relationship with lamproite 97-100 relevance to upper-mantle metasomatism 99-100 trace elements 98 varieties of 95-7 King Leopold mobile zone 129, 130 Kishalduga (Kenya rift valley) 58 Kisingiri volcano (W Kenya) 56, 57 Knox County, Maine 128 Kola Peninsula, USSR 72, 163, 531-43 alkaline ultramafic intrusions 534-5 crystallization of the agpaitic magmas 539 Gremiakha-Vyrmes 531-3 Khibina apatite deposits 534, 541-3 Khibinacomplex 532, 535 Lovozerocomplex 532-3,537-9 volatile components in the alkaline magmas 5394l Yelet'ozerocomplex 531 Kruidfontein(S Africa) 69 labuntsovite 135 laccoliths 127, 128,421,422 Lake CargeIIigo Area, N S Wales 131-2, 163 lamproite pipes 167 lamproite terminology, historicalsummary 107 lamproite 5, 8, 15, 19, 20, 23, 24, 25, 29, 32, 39, 191, 195, 196, 198, 200, 202, 203,204, 205, 213, 215, 217,218 age of emplacement 141-2 controlled by intra-plate tectonic processes 144 defined 108-9 diamondiferous 104, 142---4, 151-2, 166 W Australia 97
555
geochemistry 146-56 major-elementchemistry 146-9 trace-element chemistry 149--52 intrusive-extrusiveforms 142 isotopic data 212 kimberlite-lamproite relationship 169-70 leucite 97,98,129 localities and provinces 110-11 locations in N America 117 mineralchemistry 156-64 nomenclature and classification 106-9 occurrences of lamproites and other potassic to ultrapotassic rocks 109-46 olivine 97, 98, 99, 107, 114, 136, 144, 166-7, 198, 166, 167 diamondiferous 129 petrogenesis 166-9 relationships between lamproites, MARID and other mica-rich xenoliths 170 Sr and Nd isotopic compositions 154 tectonic-geological environment 142-4 trace elements 98, 99 lamprophyre associations, summary 195 lamprophyre branches, major petrological contrasts 194 lamprophyre clan, hierarchical structure 192 lamprophyreversuslamproite 144-5 lamprophyre-kimberlite relations 198 lamprophyres 5, 19, 31, 49, 103-4, 106, 109, 124, 139, 148, 151,156, 160, 166, 167 differentiation in 219 feldspathic 465 leucite 136 mafic and felsic enriched relatives 196 nature and origin of 191-219 comparative whole rock chemistry 204-12 criteria for identification of as a clan 193-6 discrimination between the four branches 196-8 global distribution 215 nomenclature 191,193 petrological characteristics (miscellaneous), whole clan 212-15 whole clan mineralogical diagnostic features 198-204 as parental magmas? 219 petrogenetic significance 215-19 plutonic and volcanic relatives 195-6, 196 as primary-mantle melts 215 ultrapotassic 133 volatile-poor relatives 195-6 lamprophyres, ultramafic 462-5 lapillae sprays 9 lapilli, autholithic lamproite 124 larnite 465 latite 104, 128, 135 laurdalite 437 lava plateaux 395 Cameroon and N Nigeria 273 lavas 64, 120, 136, 142, 298,421 alkali 227,230, 232, 239, 242,243,246, 493, 508 alkali basalt 227, 243,491 alkalic carbonatite 87, 88 basalt 260, 269, 277, 278 basanite 127,227,237,280, 491
556
Index
lavas (cont.) basic 281,282 benmoreite 299, 322 carbonatite 71, 72-3 composition and evolution of (Ascension Island) 269 differing sources 242-3 emulsified 525 fissure-fed 302 Gardar, distribution and chronology 453-7 Haiwaiian, trace-elementdata 242-6 hawaiite 234, 260,491 kamafugitic 135 lamprophyre 193 leucitelamprophyre 137 leucite-bearing 131 Loihi Seamount 232-3 major-element analyses 232 mafic 301 mafic and ultramafic 31 mugearite 234 natrocarbonatite 297 nepheline melilitite 237, 243 nephelinite 72, 237, 280, 504 ocean islands, S Atlantic, analyses 256 Oldoinyo Lengai 45, 47, 87, 88, 89, 90, 91, 92, 93 olivine-phyric 491 phonolite 280, 301-2, 303,465 P-phonolite 303 post-erosional 237, 240, 243,246 primitive 237 rhomb porphyry 303 rhyolite (peralkaline) 269 salic 298, 303,306 tholeiite 227,230, 243, 506 Hawaiian 240 tholeiitic basalt 233 tholeiitic picrite 233 trachyandesite 519 trachyte 274, 299, 303,309, 465 ultramafic 464 layering 475 agpaitic rocks 479-80 Ilimaussaq rocks 473 inverted 482 kakortokite, theories of formation 482-3 of magma 476 primary 532 rhythmic 537 layers, double-diffusive 482 Leeward Islands 234 lepidolite 370, 374 lepidomelane 387 leucite 31, 32, 36, 44, 45, 49, 97, 104, 106, 108, 131, 136, 137, 138, 139, 146, 149, 160, 164, 166, 167, 198,202, 213, 345, 351,465, 517 leucite basalt 107, 140 leucite basanite 104, 135, 519 Leucite Hills, Wyoming 39, 44, 103, 106, 107, 120-2, 124, 141,142, 144, 146, 148, 149, 154, 155-6, 160, 163, 164, 165, 167 leucite lamproite 97 leucite lamprophyre 136 leucite phonolite 104
leucitetephrite 104 leucitite 39-44, 107, 128, 138 olivine 136, 165 leucogabbro 361,363, 366, 369,437 leuconorite 368 leucorhyolite 436 leucotroctolite 368 lherzolites, mantle-type 61 LIL elements 36, 57, 133, 260, 262, 284, 285,425, 504 depletion in 434 enrichment in 39, 166, 167-8, 283,449 limburgite 135, 140, 296 liquid diatresis 9 liquid fractionation 307, 308, 309, 46l, 468 liquid immiscibility 30, 45, 47, 48, 60, 64, 65-8, 6971, 77, 78,214, 301,370 liquid-fluid distribution processes 485 liquids, unmixing of 64 lithophile elements 2, 7-8 lithosphere 21 continental 22-3 lithosphere focussing 353 Loihi Seamount 227, 230, 232, 248 isotopic studiesoflavas 241,242 pre-shield alkalic lavas 243 Longonot volcano 315, 322-3 syn-caldera ash flow 326 lopoliths 127 LREE 149 decline in 301 depletion in 17, 18, 19, 21 enrichment in 21, 59, 243,259, 262, 444, 459, 5267 Luangwa Graben, Zambia 136 lujavrite 48,474, 484-5,537, 541 agpaitic 466 eudialyte 537 maars 517 macrocrysts 156, 203, 213 in kimberlite 95 olivine 109, 114 phlogopite 114 picroilmenite 114 madupite 39, 108, 122, 164 mafic melt diatresis 8-10 mafic rocks 317 mafurite 31,32 magma bodies, compositionally stratified 478 magma chambers 63, 64, 271,329, 330, 478,484, 508, 511 agpaitic 480 crustal 63,368, 369, 519, 523 deep-seated 476 and density hump 308 distribution of minerals 542 Ilmaussaq intrusion 273-4 open 492 pre-caldera 320 replenishment of 508 rupture of 309 shallow 233,234 smaller homogeneous 525
Index Skaergaard 478 stratified 303,452 large 525 sub-crustal 368 sub-volcanic 264, 306 trachyte 322 zoning within 313 magma densities 477 magma evolution, Menengai volcano 324-5 magma and fluid evolution (Ascension Island) 26971 magma fractures 23 magma mixing 63,305,313,315, 329, 331,332, 407, 423,427, 508, 511,519 back-mixing 309 Massif Central volcanism 525-6 salic magmas 324 magma plumbing 59-64 crustal intrusive processes 63-4 pyroxene fractionation 61-3 magma reservoirs 422-3,479 convecting and slowly cooling 480 oceanic 288 shallow 248-9 magma rise, rate of 93 magma sources 250 E Greenland alkaline rocks 507, 510 harzburgite 467 lherzolite 41 lower crustal granulites 467 spinel lherzolite 467 trace-element constraints 242-6 magma storage 235 magma storage reservoirs 249 magma storage sites 250 magma storage system, evolution of 246, 248-9 magma storage zones 237, 249 magmas 277,298,443 acid 362, 363 agpaitic 461,467,485 crystallization of, Kola Peninsula 539 progressive crystallization and formation of layering in 473-85 alkali basalt 30, 435,442 parentalto Gardar rocks 475-6 alkali carbonatite 53 alkali mafic and ultramafic--background studies 30-4 rock systems--the pyrolite model 32-4 synthetic systems at low-pressure 30-2 alkali mafic and ultramafic--recent experimental studies 34-44 alkali olivine basalt 34, 351,444 alkali picrite 442, 443,444 alkalic carbonatite: parental or derivative 85-93 alternative petrogenetic scheme 91-2 critique of parental-status scheme 87-9 origin of 89-91 alkalic/alkaline 21, 22, 273,425,444, 450 genesis of 29-49 interaction with NaCI brine 440 Oldoinyo Lengai type 86 volatile components in 539-41 alkaline basaltic 143
557 aln6itic 443 aluminous (troctolitic) 467 anchieutectic (multiply saturated) 477 andesitic 34 asthenospheric 358, 361 augite syenite 476 basaltic 34, 315, 322, 330, 362, 363,450, 451,457-8 anhydrous 466 mantle sources 246 basanite 444 basic 452, 459, 523,525 Gardar 467 generation of 283-6 petrogenesis of 526-7 vesiculated 525 benmoreite 322, 461,468 residual 450 calcitic-dolomitic 92 camptonitic 443 carbonated nephelinite 86 carbonatite 72, 86, 89,450 possible stylesoforigin 65 CO2-rich 467 comendite 270 contaminated 524 diverse, assimilation of 39 Fe-free K-rich alkaline 165 felsic 8, 10, 11,418 alkaline 5 fractionated 308 Gardar, volatile component of 466-7 granite 467 contamination by sea water 271 halogen-rich 77 hawaiite 450, 451,453,457,461,467,468 high alkalinity 36 high-alumina 450 troctolitic 461 highly-mobile 468 ijolite-urtite 543 intratelluric 217 K-rich 45 genesis of 44 kakortokite 481,482 kimberlite 21, 22, 32, 167, 168, 170, 351 differentiation of 167 lamproite 97, 149, 167, 168, 169 3-stage evolution model 170 lamproite-kamafugitic 167 lamprophyre 217-18,450 layered 482, 483 leucitelamproite 167 lujavrite 484 mafic 10, 300, 308,426, 437,444 mafurite 32, 40 magnesian 282 mantle-derived 30, 441 contaminated 412 interaction with crust 434, 435 mica-peridotite primary 351 monchiquitic 443 mugearite 461 Na-rich 44 natrocarbonatite 69-71, 77, 78
558
Index
magmas (cont.) nephelinite 55-9, 66, 67, 70, 74, 77, 88, 92 carbonated 64 olivine basalt 453,457 genesis model 467 olivinelamproite 167 olivine melilitite 38 olivine nephelinite 90,93 olivine s6vite 90, 92, 93 olivine ugandite 40 parental 86, 306, 352, 357,449, 451 anorthosite complexes 366 carbonated melanephelinite 301 olivine s6vite 91, 93 pulaskite 407 peralkaline granite 475 peralkaline silicic 328 peralkaline trachyte 328 phlogopite-lamproite 167 phonolite 66, 70 picritic 37, 217 potassic, genesis of, modelled 41 primary 217-8, 354, 468 primary syenite 341 quartz syenite 428 residual 476, 480 troctolitic 460 volatile-rich 483 salic 306, 307,451,452, 508 alkaline 309 highly differentiated 461-2 secondary 510 shoshonite 106 silica-undersaturated 440 silicate 68, 91 sodic carbonate 86 s6vite 75-6 sub-silicic 450 subduction-zonerelated 261,262 syenitic 351 tholeiite 505 trachybasalt 508 trachyte 324, 331,412, 508 Trans-Pecos, HFS/LILratios 427 troctolitic 467 ultramafic alkaline 3 ultrapotassic 164 stable and radiogenic isotope geochemistry 1526 undersaturated 407 volatile-charged, ascent of 320 volatile-rich, Ilmaussaq 478 volatile-saturated 481 water-saturated 271 wolgidite 165 magmatic activity, within plate 361 magmatic alteration 105-6 magmatic bimodality 5-6 magmatic centres, migratory 357, 358 magmatic differentiation 148 magmatic evolution anorogenic 361-3 stages of, Niger anorthositic centres 368-9 trend, MonteregianHills 439
White Mountainplutons 442 magmatic foci, migration of, Gardar province 455 magmatic layering 368 magmatic reservoir systems 423 magmatic systems, time-space evolution of 318 magmatism 4 alkaline, Baltic Shield 534 alkaline mafic 5 alkaline, tectonic constraints on 361 anorogenic 366 effect of late-stage fluids on 370-6 and megashears 361 tectonic and structural controls of 358-70 within-plate 398 basanitic 4 calc-alkaline 398,427 collision-related 395 Cameroon line 282 carbonatitic, evolution of 65 cratonic 5 felsic 427 Hawaiian 237 intra-plate 215, 376, 449 lamprophyric, distribution through geological time 216 Mesozoic, N America 444-5 Phanerozoic alkaline 358 plate-margin 215,218 subduction-related 381 syenitic 493 tholeiitic 491,504-5 tholeiitic basalt 444 tholeiitic rift 489 magnesite 39,44, 166 magnetite 71, 72, 73, 90, 123, 133, 138, 163,405,462, 492 Ti-magnetite 451 magnetite deposits 74 Malawi, carbonatitecentres 335 malchites (semi-lamprophyres) 196 Mali (Adrar des Iforas) 381-99 malignite 532, 541 MandaraMountains, Cameroon 278 Manengouba, Cameroon 278-9 mantle asthenospheric 398-9, 398 carbonated 443 depleted 441,443, 527 composition of Iforas 393-4 enriched 284 low-density 353 metasomatized 443 sub-continental, geochemical evolution of 23 mantle composition, primordial 262, 263 mantle contamination 23 mantlediapirism 445,456 mantle enrichment 2 mantle evolution 169 mantle heterogeneity 460, 524 mantle metasomatism 29, 44-5, 284, 353, 467 mantle mobilization 394 mantle pegmatites 170 mantle plume 241 mantle processes, importance of 32
Index mantle sources 263, 506 alkaline magmas 273 basaltic magma characteristics 246 convecting upper mantle 285 depleted lithospheric 398 enriched 508 Fe-rich 460 and generation of basic magmas 283-6 isotopic constraints on 239-41 lherzolite 29 for OIBs 255 partial melting of 246 rare-gas constraints 241-2 similar for oceanic and continental basalts 285 undegassed 232-3 mantle wedge 354 margins, active and passive 370 MARID rocks 16 Marquesas Islands 143 Massif Central (France) 58, 517-27 chemical and petrographic variation 519 crustal contamination 523-5 fractional crystallization 520-3 geochronology 517-19 magma mixing 525-6 petrogenesis of basic magmas 526-7 Mauna Loa 230 mechanical segregation 369 megacrysts 10, 15, 62-3,154,425-6 alkali feldspar 8 amphibole 8, 278, 521 anorthoclase 277,278,426 apatite 426 biotite 426 clinopyroxene 61,277, 278, 282 Cr-diopside 218 feldspathic 461 garnet 277-8 ilmenite 278 kaersutite 37,426 olivine 129,218 orthopyroxene, Al-rich 426 plagioclase 426, 461,466 soda-rich 8 spinel 442 Ti-phlogopite 213 titanomagnetite 426 zircon 278 megashear zones 361,381,388 lithospheric 398 reactivation of 395 melanephelinites 55, 57, 60, 61, 62, 63, 68, 90, 279, 295,296, 297, 301,307, 309, 506, 510 melanite 109, 305, 338,339, 349,405,465 melaspessartite 196 melasyenite 405,407, 408, 412 melilite 29, 31, 72, 109, 136, 193, 198, 200, 213,464, 465,492, 517 strontian 279 melilite nephelinite 68 melilitite 3, 5, 8, 29, 30, 32, 37-8, 68,295, 296, 301, 307, 309, 463,519 olivine 39 melilitolite 492
559
melt channels 7-8 melt diapirism 10 melt enrichment 10 melt mixing 236 melt percolation 7-8, 9, 10 melteigite 120, 534, 535, 541 melting models 442-4 melting, large scale 468 melts alkaline 29 alkaline ultramafic 7 alumino-silicate 542 anatectic 508,509 asthenospheric 21 basaltic 260 basanitic 9,16 carbonate 65, 66 diamondiferous, primary mantle-derived 154 felsic 7, 8 kimberlite 170 lamproite 168, 170 mantle-derived 167 large-percentage 444 mafic, Navajo-Hopi Buttes 127 mantle 435 migration of 7 mixing with mantle 262 nephelinitic 16, 38 phlogopite peridotite 166 primary-mantle 218 silica-undersaturated 239 silicate 23, 66, 78 infiltration of 21 sub-alkaline 324 Menengai 315-16, 319, 324-6, 327 mercury deposits 426 Merrymeeting Lake, geology of 436, 441 mesolite 441 metacarbonatites (silicocarbonatites) 71 metamorphic aureoles 436 metamorphism 109 metasediments 128 metasomatic alteration 351 metasomatic aureole 370 metasomatic enrichment 306, 443 metasomatism 6, 10, 72, 74, 78, 85, 165, 167, 168, 337, 349, 354, 445,485,499 acid 370, 374-6 conditions of 2-4 definition 1-2 development of separate felsic sources 7-10 Fe-Ti 16 gas-phase 217 geological factors 4-5 K-rich 44 lamproite-source 168 large-scale 4 Na-rich 44 occurrence and distribution 2 potash 375 potassic 370, 373-4 sodic 370, 373, 375 ofsub-Kenyan mantle 317-18 timing 2
560
Index
metasomatism (cont.) in the vicinity of Gardar intrusions 465-6 metasomatism-modellingexperiments 44-5 metavolcanic rocks 351 metavolcanics 128 Meugueur-Meugueur ring-dyke 368 Mg-arfvedsonite 77 Mg-biotite 387 Mg-Mn-spinel 200 miaroliticcavities 383,418 miarolitic texture 422 miaskitic sequence 387 mica 34, 71, 91, 93, 170, 338, 341,370 green 374 mica peridotites 139 micaschists, eclogitic 381 micro-inclusions, apatites and nephelines 542 microcline 72, 76, 135, 137, 138, 373, 388,484 destruction of 374 microclinization 374 microdiamonds 139 microfoyaites 338, 345,405 microgabbro 363 picritic 368 microphenocrysts 61,131 biotite 302 ferroaugite 302 magnetite 73 titanomagnetite 303 micropulaskite 437 microsyenite 363 mid-Atlantic Ridge 269 mid-ocean ridge basalts (MORBs) 4, 15, 21-2, 169, 212 mid-ocean ridge basalt (MORB) source, depleted 246 migmatites 77, 370 mineral crystallization, sequence of 34 mineral layering 366 mineral sorting, vertical 542 mineral stabilities 3, 4, 5, 7 felsic 10 mineralization 73-4, 450, 465 mineralizing fluids 78 minette 49, 104, 105, 124, 127, 128,132, 133, 135, 139, 140, 144, 146, 148, 149, 163, 165, 166, 193, 195, 196, 198,212 missourite 107, 127 mixing processes 263-4 mixing, magmas and crustal material 441 Mn-ilmenite 131 mobile belts, Pan African 9360, 361 molybdenite 405 molybdenum deposits 426, 502 monazite 71,109, 168 Th-rich 373 monchiquite 128, 198,212, 218, 219, 442, 463,464, 495 picritic 196 Mongolowe complex 336 nepheline syenite and syenite 338, 341-4, 345,347, 351 nepheline syenites 341 Monteregian Hills and White Mountains Older White Mountains 433-5
tectonic setting 444-5 Younger White Mountains and Monteregian Hills 435-44 chemistry and isotope geology 438-41 dykes 438 geochronology 438 Monteregian Hills 6, 436-8 petrogenesis 441-4 Younger White Mountains 435-6 monticellite 72, 95, 108, 198, 200, 465 montmorillonite 374 monzoanorthosite 363, 369-70 monzodiorite 434-5,436 monzogabbro 361,366 monzogranite, porphyritic 388,389 monzonite 269,433,436, 437,440, 441,503 hornblende 495 layered 495 quartz 361 MORB 264 Mulanje complex 352-3 peripheral fenitization 341 Murcia-Almeria province, SE Spain 124, 132-3, 142, 146, 148, 149, 154, 156, 160, 163, 164, 167 muscovite 387 N American plate 420 and alkaline magmatism 445 Na-for-K exchange 370 NaCI fugacity 540 nahcolite 467 Namibia 54 nannofossil ooze 264 narsarsukite 462 natrocarbonatite 53-4, 67, 69 modes 70 Oldoinyo Lengai 47 parental 69,71-2 natrocarbonatite magmas 69-71 natrolite 338,484 naujaite 467,474, 477,478,479, 480, 484 naujakasite 484 Navajo-HopiButtes, Arizona 127, 146 Nb depletion 132 143Nd/14*Ndratios 16-25 nepheline 30, 34, 36, 37, 48, 49, 109, 136, 149, 202, 279, 301,338, 339,405,408,416, 440, 479, 481, 482, 484, 534, 535, 541,542 nepheline hawaiite 504 nepheline syenite-porphyries 535 nepheline syenite 66, 68, 107, 123, 138,215, 296, 306, 336, 337, 341-4, 417-19, 425,437, 440, 466-7, 502, 531,535 miaskitic 537 peraluminouspotassic 127 nephelinite magma 55-9 nephelinite-carbonatitecomplexes 54-79 nephelinite 3, 5, 6, 8, 15, 29, 30, 32, 36-7, 68, 90, 196, 273,294, 295,296, 300, 303, 304, 307, 336, 337, 339, 349, 351,352, 354, 504, 506, 517, 519 analyses 58--9 and carbonatite 53-79 immiscible with alkalic carbonatite lava 87
Index melanephelinite 349 nosean leucite 279 olivine 39,297, 349 olivine-poor 55-8, 59-60, 61, 62 olivine-rich 37-8, 58, 59, 62 world average 58 net-veining 418 New England Seamounts chain 444 Ngaound+r6 Plateau, Cameroon 278 Niger-Nigerian alkaline ring complexes 357-77 effect of late-stage fluids on anorogenic magmatism 370-6 tectonic and structural controls of anorogenic magmatism 358-70 Niggli values 106 Nkalonjecarbonatitecentre 353 nodules 5, 8, 45 amphibole 10 diamond 218 gabbro 6, 320-1 harzburgite 493 kaersutite eclogite 36-7 kimberlite 4 kimberlite peridotite 4-5 lherzolite 37, 45, 59 mantle 2, 64 mantle peridotite 218 peralkaline granite 6 peridotite 6, 7 phlogopite-clinopyroxene 45 plutonic 313, 318, 320, 323 pyroxenite 45 sanidine-coesite eclogite 7 spinellherzolite 9, 35 spinelperidotite 10 syenite 6, 320, 324 transformed 4 ultramafic 2, 3, 4, 5, 9 nontronite 109, 164 noonkanbahite see shcherbakovite nordmarkite 361,405,437, 439,499, 502, 503 nosean 109 nyerereite 53-4, 69, 70, 464 Nyos, Lake, Cameroon 278 obduction 360 Obontorok ring-complex 367 obsidian 303, 320, 328 comenditic 269-70 rhyolitic 316 ocean basins, isotopic data 16-18 ocean crust, subduction of 262, 264 ocean island basalt (OIB) 15, 21-2 alteration 255,257 chemically anomalous, hypotheses for origin of 253 geochemistry 255-61 major elements 255 source characteristics 258-61 trace elements 255 ocelli 202-3, 213,214, 219 ocellitextures, syenites 45 odinite 196 Ofoud ring complex 363-4, 366
56I
Ojin Seamount 230 Oka carbonatite complex 436, 440, 441 Oku, Mt., Cameroon 278 Oldoinyo Lengai 45, 47, 66, 67, 90, 92, 93, 296 eruptions of 86, 87, 297 natrocarbonatites 69, 70 oligoclase 133,404 olivine 30-1, 32, 34, 35, 36, 38, 40, 44, 45, 61, 71, 72, 95, 107, 114, 123, 129, 131,133, 136, 137, 139, 148, 149, 156, 164, 165, 167, 196, 200, 234, 258, 282, 301,349, 364, 365, 368, 369, 370, 425,437, 465,520 carbonatite 90 cumulus 436-7 fayalitic 420, 451,475 stabilized by H20 37 olivine basalts 294, 298 olivine gabbro 531 olivine lamproite 99, 107, 114, 136, 144, 166, 167, 198 olivine leucitite 136, 165 olivine melilitite 39 olivine nephelinite 39, 68, 349 olivine tholeiite 280 olivine ugandite 39 olivine-glimmerite 107 olivinites 67, 534 Olorgesailie 300 eruptive sequence 303-4 Oman Mountains, Arabia 106 ophiolites 360, 381 Orciatico, Pisa, Italy 133 ore deposits, Nigeria 376 ore metals, development of important economic sources 375 orendite 39, 40, 41,103, 106, 107, 108, 120, 121-2, 164 orogenesis, collisional 143 orogeny 360 orthoclase 72,76,149 orthoclase effect 387 orthopyroxene 32, 34, 35, 36, 41, 44, 45, 149, 156, 164, 165, 167, 202, 258, 388,425-6 stabilized by CO2 37 Ossipee ring complex 435-6, 442 oxidation 484 oxides Fe-Ti 131,133, 138, 139, 324, 361,370, 383, 387, 388,465, 520 crystallization of 323 Fe-Ti-Cr-Mg-A1 163 iron 73, 77 K-Fe-Ti-Ba 163 precipitation of 56-7 oxygen fugacity 68, 73, 97, 233,477 oxygen isotopic studies 154-5,271 Pagalu (Annobon) 282 pahoehoe surfaces 454 Paka, Kenya 314, 315, 318,321-2 palagonite 142, 232 Pamir, Mongolia, USSR 140 Pan-African magmatic evolution, proposed model, the Iforas 3 9 6 - 7 Pan-African orogeny 357, 358, 360-1,388-9
562 pantellerite 6, 326, 331,423 phase equilibrium studies, central Kenya 320 paragneisses 326 pargasite 169, 412 ferro-pargasite 451 ferroan 410 titanian 202 parisite 71 partial melting 6, 20, 30, 33-4, 39, 44, 45, 104, 164, 167, 169, 170, 242, 255,257-8,259, 261,306, 313, 354, 361,425,443 alkalineultrabasic silicate 65 amphibole peridotite 38 of asthenosphere 64 clinopyroxenite 165 garnet lherzolite source (Honolulu Group) 243 granites and gneisses, upper-crustal 524 granulitic material 441 harzburgite 170 lherzolite 36, 39, 167 Fe-rich 34 Mg-rich 34 LILE-rich streaks 285 ofmafic lower continental crust 323,331 mantle 165, 166, 246, 354, 460, 517 metasomatized 323,526 metasomatized spinellherzolite mantle 444 N-MORB source mantle 263 olivine pyroxenite 39 peridotitemantle 39, 65, 165, 167 primitive mantle 263 pyrolite 38 refractory mantle 167 small percentage 241,243, 284, 468 MORB source 286 spinel lherzolite 444 sub-continentallithosphere 170 subductedocean crust 262 upper mantle 394 of wall-rocks 243 partial melts, mixing of 243 partially molten zones 319-20 PAWMS (Pacific authigenic weighted mean sediment) 263-4 Pb isotopic data, Hawaiian volcanism 241 Pbisotopic studies 155-6 pegmatite 450, 501,527 Ivigtut 466 layers 479-80 marginal 475 minerals 467 veins 483 pegmatitic-hydrothermal trend 389 pelagic sediments 263,264, 265 Pendennis, Cornwall, UK 132, 133, 149, 160 peralkaline rocks 104, 149 peralkaline volcanic rocks, Menengai 328-9 peralkalinity 316, 317, 324, 326, 339, 343, 345,352, 417,421 percolation, melt 7-8 periclase 71 peridotite 7, 15, 108,368, 531,534 biotite 464 garnet-bearing 237
Index isotopic variability 16, 18 mica 122, 139,463,464 micaceous, diamond-bearing 104 phlogopite 166 spinel 517 univariant melting 44 peridotite massifs 2 perknites 108 perovskite 61, 71, 97, 108, 114, 116, 122, 123, 129, 131,133, 139, 164, 198, 301,305,409, 465 perpotassicrocks 105 perthite 337, 383, 386, 387 perthosites 338 oversaturated 342 phenocrysts 61,108, 144, 146, 160, 269, 338,405,410 aenigmatite 417 alkali feldspar 302, 326 amphibole 193, 339, 417 kaersutitic 417 analcime 48 ankerite 139 anorthoclase 322, 417, 421 apatite 72, 139, 417 augite 138, 301,339, 404, 421 biotite 193 calcite 69-70 chrome-spinel 139 clinopyroxene 55,417 diopside 114, 131,138 fayalite 421 manganoan 417 feldspar 298,324 fractionation of 304 gregoryite 70 hedenbergite 326 hornblende 288 K-richterite 114 leucite 129, 138 magnetite 72-3 microcline 388 nepheline 301,302, 324, 417 nyerereite 70 olivine 114, 120, 123, 131,138,301,339,417,459 orthopyroxene 422, 521 perthite 404 perthitic feldspar 128, 337 phlogopite 120, 131,135, 138, 139, 165 pigeonite 521 plagioclase 196, 303, 404, 417,460 pseudoleucite 114 pyroxene 62, 63,301,339 quartz 320, 421 resorbed 139, 217 sodalite 417 Ti-phlogopite 114 titanaugite 417 titanomagnetite 417 zoned 313 phlogopite 2, 3, 5, 7, 9, 31, 39, 40, 41, 44, 45, 49, 72, 76, 77, 97, 108, 109, 123, 127, 129, 131,132, 133, 135, 136, 138, 139, 148, 160, 162, 165, 167, 218, 243, 258,468 Ba-Ti-phlogopite 198 biotite-phlogopite 193
Index primary 32 tetraferriphlogopite 139, 160, 202 Ti-phlogopite 114, 123, 131,138, 160, 198 phlogopite peridotite 166 phlogopite-biotite 91 phonolite 5, 6, 15, 37, 55-6, 60, 61, 63, 66, 77, 107, 128, 234, 273,278,281,286, 293,294, 301,302, 305, 306, 335,351,417-19, 425,452, 454, 517, 521 agpaitic 468, 538 G-phonolites 295, 301,304, 307, 309 LREE depleted 305 immiscible with alkalic carbonatite lava 87 P-phonolites 296, 297-8, 300, 301,309 LREE enriched 305 peralkaline 349,424 phosphate ore deposits 541 picrite 495,504, 506 alkali 442 tholeiitic 33,510 picroilmenite 96 pigeonite 34, 35, 422 pillow dykes 499, 504, 508 pillow lavas 232 pillows, diorite 499 piperno 124, 142 plagioclase 7, 30, 31, 66, 107, 109, 135, 136, 202, 234, 317, 363, 365, 368,416, 437,460, 491,520 calcic 420 sodic 339, 386, 426, 458 zoned 369 plagioclase effect 324, 329, 387 plagioclase fractionation 260, 301,304, 307, 412, 521 plate motion 55 plate movements, directional changes 283, 361 plateaubasalts 277, 504 tholeiitic 495, 504 plateau lavas 383 Pliny Range, geology of 434 plugs 129, 132, 193,337 biotite granite 499 diorite 498 diorite-gabbro 436 granite 436, 495 lamproite 124 leucite lamproite 142 monzonite 436 trachyte 282 tristanite 282 plume ascending 168 deep mantle 282, 285 source, degassed 241 plutonic complexes, sub-volcanic 64 plutonism basic to salic 491 granitic 218 Kangerdlugssuaq district 498-501 plutons 64, 127, 129,219, 382, 405 Adma 388 carbonatite 78, 163 durbachitic 135 granite 422 ijolite 74-5
563
lamproite 114, 142 mafictoultramafic 439 peralkalinegranite 335,433 post-tectonic 398 pyroxenite 433 quartz syenite 422 ring 403 syenite 6, 325 syenogranite 388-9 ultramafic 106 Velasco 403 Younger White Mountains and Monteregian Hills 435-44 zoned 6 pneumatolytic-hydrothermal processes 374 Pniel, Postmasburg, Swartruggens etc., S Africa 136 Pollen (N Norway) 71 polymerization 66 polyvalent cations 37 porphyries 196, 389 hydrothermal alteration 370 quartz 495 porphyrites 196 potassic-ultrapotassic rock suites compositionally related to lamproites 112-13 potassic-ultrapotassicrocks 127, 140 potassium 30 Prairie Creek see Crater of Diamonds precipitation of augite 61 of olivine 61, 64, 72 primary magmatic 74 of apatite 72 pressure release 361 pressure-release melting 444 priderite 108, 114, 122, 123, 128, 129, 135, 163, 165, 200, 203 Priestly Peak, Antarctica 138, 142, 156, 160 Principe, Gulf of Guinea 280 protolithionite 374 provinces (alkaline) Angola 54 Appalachian 337-8 carbonatite 198 E Africa 54, 55, 59 E Siberian 54 Gardar, S Greenland 449-68 Iforas 381-99 Karoo 57-8 Kenya 293-309 MonteregianHills 219,433-45 N Chilwa, petrochemistry 335-54 Niger-Nigeria 357-8 Roman 107, 108, 135, 155, 163, 165 Tadhak, Permian 383, 398 Velasco, petrology, chemistry and crystallization history 403-12 White Mountains 433-45 provinces (ultra-alkaline), Kola 67 provinces (ultrapotassic), Toro-Ankole 135, 136 prowersite 103 pseudobrookite 120, 122, 123 ferro-pseudobrookite 163 pseudoleucite 29, 48, 120, 138,203
564
Index
pseudomorphs 69-70, 138, 139, 337 carbonate 464 pulaskite 338, 342, 343,405,407,408,409,410, 437, 451,475,478,499, 502, 508,531 differentiation of 412 pyrite 114, 337, 338,462 pyrochlore 71, 72, 73, 337, 370, 376 pyroclastic units, zonation in 325-6 pyroclastics 193 trachyandesite 519 pyrolite 34 pyrope 166 Cr-pyrope 122, 164 pyroxene 55, 57, 61, 72, 74-5, 137, 193, 301,338, 339, 349, 386, 405,409-11,412, 520, 542, 543 alkali 409 hedenbergitic 337 omphacitic 44 reverse zoning 63, 64 secondary formation 465-6 Ti-rich 198 pyroxene cores (green) 63, 64 pyroxene fractionation 61-3,351 pyroxene zoning 410, 411 pyroxenite 16-17, 25, 77, 86, 436, 437, 517, 531,534 alkali 61,502 biotite 137,464 cumulative 63 mica 463,464 quartz 49, 71,108, 123, 132, 133, 137, 138, 149, 163, 164, 166, 217, 337, 339, 342, 374, 375, 383,388, 416, 437 quartz diorite 360 quartz mangerite 449 quartzmonzonite 361,434 quartz porphyry 370, 495,508 quartz syenite 269, 336, 337, 339, 341,361,383, 386, 405,407, 419-21,422, 426, 433,451,495, 503 quartz trachyte 278, 279, 282, 286, 419-21,423,439, 425 quartzite 360 fenitized 77 quench crystals 138 quench textures 525 radiogenic heat 169 rare gases, isotopic analyses of 241-2 rasvumite 200 Rb depletion 16, 17, 18, 19 Rb enrichment 17-18 reaction rims 55, 61 recrystallization partial 388 secondary 338, 387 Red Hill complex 435 reducing conditions 539 REE depletion 160, 521 enrichment 53, 78, 217, 316, 391-2, 405, 439, 440, 484 inlamproites 149
in lamprophyres 212 REE data, Trans-Pecos 425 REE distribution, Oldoinyo Lengai lavas 87, 88 REE fractionation 258 REE minerals 69, 73-4 REE partitioning 71 REE patterns, Hawaiian alkalic lavas 242, 244-5 regional uplift 282 replenishment processes 508 residua differentiated 331 silica-oversaturated 439 silica-undersaturated 440, 408 upward migration 476 residuasystem 286-7,407 residual fluids 370 boiling of 373 resorption 61, 62, 72, 108 fractional 166 rheomorphism 351 rhomb porphyry 302 rhyodacite 233 rhyolite 6, 68, 104, 105, 274, 278, 279, 286, 287, 299, 303, 307, 315, 323,382, 383, 389, 398, 421-2, 426, 508, 509, 517, 519, 523, 524 alkali 273,294, 296, 298, 303,309,425 ignimbritic 362 leucorhyolite 436 metaluminous 422,424, 425,428 origin of Massif Central 521 peralkaline 303, 313,423,424, 425,428,434, 468 rhythmic layering 364 richterite 97, 135, 383, 387 K-richterite 114, 120, 123, 129, 132, 133, 138, 170 zoned grains 160 K-Ti-richterite 108, 156, 200, 202 Ti-richterite 131 richterite-arfvedsonite 135 riebeckite 337, 338, 387,405 riebeckite-arfvedsonite 133 rift valleys, continental, isotopic data 18-19 rifting 505 E Africa 6, 136, 293-4, 335 intra-continental 449 SAmerica-Africaplate 403 within plates 361 ring complexes 382 9Iforas, oversaturated alkaline 381-99 Niger-Nigeria 357-77 rischorrites 535 rock-fluid interactions 370, 375-6 Rodberg iron deposit (Fen, S Norway) 74 Roman co-magmatic province 107, 108, 135, 155, 163, 165 roof collapse 484-5 rosenbuschite 424 Ruri Hills (W Kenya) 69 rutile 97, 132, 133, 135, 138, 139, 165, 168, 170 St Helena 253-61 salic rocks 303 fractional crystallization origin 323 incompatible trace-element enriched 304
Index origin of 286-9 under- and over saturation 287 salite 156 sandstone 280 sanidine 41, 44, 49, 104, 108, 114, 120, 123, 131,132, 133, 135, 137, 138, 160, 163, 165-6, 234 sanidine trachyte 234 Sgo Tom6, Gulf of Guinea 280-2 sapphire 124 scapolite 198 schists 360 schlieren 74, 219 disrupted 72 scoria cones 131 sea-floor spreading 381,489-90, 505 seawater 271 sediments pelagic 263,264, 265 terrigenous 263 selagite 107, 133 sequential tapping 452 sericite 339, 374 serpentine 114, 123, 132, 164 serpentinization 167 Seward Peninsula, Alaska 127 Shaki and Okeho, SW Nigeria 137 shales 280 shcherbakovite 108, 129, 164, 200 shear forces 306 shear zones 360, 361,382 lithospheric, reactivation of 361 Pan-African 357, 361,376 shonkinite 104, 107, 124, 127, 138, 140, 144 shoshonite 106, 142, 144, 166 siderite 462, 467 siderophyllite 370, 373, 374 sidewall crystallization 313, 451,476 Sierra Nevada, California 127, 155-6 Silali 314, 315-6, 318 faults and fissures 320 nodules 320, 321,323, 330-1 silica 45 silica enrichment 34, 44, 132 silica saturation 39, 40, 41, 42, 122, 167, 303, 345, 349, 416, 420, 427, 437 silicate melts, infiltration of 20 silicic rocks of Kenya 313-32 basalt-trachyte relationships 320-3 the caldera volcanoes 314-15 generalized models 329-32 geological evidence for depths of magmatic differentiation 318-20 rock types 315-18 trachyte-trachyte and trachyte-rhyolite relationships 323-9 silicification 74 sills 129, 132, 193, 236, 503 lamproite 114, 142 mafic 422, 426 stacked 422 Sisco, Corsica, France 133, 134, 135, 142, 149, 163 slab, down-going 169 Smoky Butte, Montana 123-4, 127, 142, 144, 146, 149, 154, 155-6, 160, 163
565
SO2 fugacity 541 sodalite 48, 109, 202, 301,338, 343, 344, 345,437, 440, 466-7,477, 479, 480, 484, 540 upward-migrating crystals 479 sodalite foyaite 478,479 sodalite syenite 538 sodic peralkaline trend, Iforas rocks 387 Sokli carbonatite (N Finland) 75 sorting, by convection currents 478 source characteristics, ocean island, S Atlantic 25861 source reservoirs, potential for OIBs, constraints on 261 s6vite 53, 71, 72, 73, 75, 76, 77, 78, 90, 339, 436, 463, 465 s6vsbergite 337 spessartite 204, 212 sphalerite 375, 376, 462 sphene 61, 71,108, 109, 131,133, 135, 137, 138, 164, 166, 168, 301,305,338, 369, 383,387, 388,405, 408,498, 541 sphene fractionation 521 spinel 3, 41, 45, 97, 108, 116, 122, 129, 133, 136, 163, 203-4, 205, 425 Cr-spinel 114, 123, 132 Mg-Mn-spinel 200 Zn-spinel 198 spinel lherzolite 524 spinelperidotite 2, 16-17, 517 Spotted Fawn Creek, Yukon 128 Sr-barite 122 Srednogorie 135 87Sr/S6Sr isotopic ratios 16-25, 63 mantle and crustal types 15 steenstrupine 484 stilpnomelane 387 stocks 193,382 diorite 434-5 syenite 434-5 stoping 451,452, 461,508 and emplacement of Ilimaussaq intrusion 475 strato-volcanoes 60 stress fields, within-plate 361 subduction 15, 144, 253, 264, 360, 381,394, 395,427 andcalc-alkalinemagmatism 398 of oceanic crust, chemical modification during 262 subduction zones 360 subsidence 384, 385 Suiko Seamount 230 sulphates 193 sulphur fugacity 477 Sunnfjord, W Norway 135 Suswa 313, 315-16, 319 syenite plutons 6 syenite porphyries 128 syenite 68, 104, 105,273, 324, 337, 339, 341-4, 346, 351,352, 363,387, 434, 435-437,451,466, 498, 502, 503, 531,537 agpaitic 452, 492 agpaitic nepheline 474 alkali 108,433,532, 535 augite 451,452, 474, 499 biotite-aegirinenepheline 531-2 of crustal origin 440
566
Index
syenite (cont.) differentiation of 341 Emuruangogolak 318 ferrosyenite 368 hastingsite quartz 434 Klokken, layeringin 483 laminated 451 nepheline 66, 68, 107, 123, 138, 215,296, 306, 336, 337, 341-4, 417-19,425, 437,440, 466-7, 502, 531,535 miaskitic 537 peraluminous potassic 127 nepheline-sodalite 435 quartz 269, 336, 337, 351,361,383,386, 405, 407, 419-21,422, 425,433,451,495, 503 rare-earthmelasyenites 412 sodalite 538 ultrapotassic 137 syenitic fenites 78 generated by carbonatite 75-7 generated by ijolite 74-5 syenogabbros 502 sylvine 467 symplektite, quartz feldspar 388
Taguei anorthositic complex 369-70 Takellout ring complex 389 talc 136 Tanzania, isotopic data 19 tchevkinite 387 tephra 132, 137 Laacher See, zoned 305 tephrites 140,294, 296, 301,302, 307,517 Terceira rift 11 terrigenous sediments 263 tetraferriphlogopite 139, 160, 202 Texas, Trans-Pecos Province 415-28 early work 415-16 field relations 422-3 igneous rock types 416-22 isotope geochemistry 425 location and extent 425 megacrysts and mantle-derivedxenoliths 425-6 mineralization 426 relations between tectonic setting and magmatism 426-7 structure 426 trace element geochemistry 423-5 texturelampropyres 193 Th, as general index offractionation 258 thermal anomaly 7, 361 thermal aureole 72 thermal contraction 308 thermal divide 412 low-pressure 287-8 thermodiffusion 316, 461,540 tholeiite 232, 243,493,504 Fe-Ti 504-5 K-rich 495 olivine 280 potassic 493, 508 tholeiitic shield-building 227-8,233-4
thorium 376 Thornton-Tuttle differentiation index 416, 422 Ti-Al-augite 61 Ti-amphibole 198 Ti-diopside 114 Ti-K-richterite 202 Ti-magnetite 451 Ti-Mg-chromite 131 Ti-phlogopite 123, 131,138, 198 Ti-pyroxenes 198 Ti-richterite 131 Ti-Zr-garnet 200 Timedjelalen ring complex 385-6, 387, 389 tin mineralization 398-9 Tinderet (Kenya) 70, 297 tinguaites 128,437,499 TiO2 depletion 132 TiO2 enrichment and depletion 149 titanaugite 336-7 titanomagnetite 57, 116, 138, 234, 364, 365,405,425, 475,479, 541 tonalite 388 topaz 370, 374 Toro-Ankole and Birunga volcanic field, Uganda 136, 143, 155-6, 163, 165, 168 trace elements 90 lamprophyres 205, 207-12 OIBs 254-5,257-61 trace elements, incompatible 258, 261,263, 321 abundances 238, 305,315 correlation with peralkalinity 317-18 enrichment in 236, 253, 299 trace-element abundances 151 trace-element chemistry, lamproites 149-52 trace-element data, Hawaiianlavas 242-6 trace-element enrichment 19, 90-1,167 trace-element geochemistry, Trans-Pecos province 439-41 trace-element partitioning 30 trace-element signatures 58-9, 60, 370, 374 trachyandesite 419,454, 504, 517, 523, 525 evolution of 520-1 trachybasalt 315,419,454 trachyphonolite 63,278,294, 296, 300, 303, 304, 309 LREE enriched 305 trachyte 5, 6, 7, 11,104, 105, 108, 135, 234, 243,255, 273, 278, 279, 280, 281,287, 294, 296, 298, 303, 304, 305,307, 309, 315,318,320, 323, 326, 331, 351,408, 417-19, 426, 434, 444, 452, 454, 495, 517, 521,523 alkali 107 anorthoclase 234 peralkaline 278,299, 313, 322 phase equilibrium studies, central Kenya 320 porphyritic 403,404-5 quartz 278, 279,282, 286, 419-21,423,425 sanidine 234 trachyte-trachyte mixing 324 'trap door intrusions' 422 tremolite 109 Tristanda Cunha 143, 253-61 tristanite 280 trough banding 482 Tuareg shield 360, 362, 369, 381,388,398
Index tufts 123, 129, 136, 144 agglutinate 142 air-fall 315, 325 ash-flow 299, 303,307, 308, 309, 315,320, 426 brecciated and welded 403 carbonate 297 hydrothermally altered 404 nephelinite, ScoresbySund 493 phonolitic 305 post-caldera 331 pre-caldera, Menengai 324 sandy 131 welded agglutinate 124 Tundulu complex 353 Turkanadepression 297, 301 Two Buttes, Colorado 128 ugandite 31, 32, 140 biotite 165 olivine 39 ultramafic lamprophyre branch 196, 198,200, 202, 203,204, 205, 212, 213,214, 217, 218 ultrapotassic and perpotassic rocks 129, 148 ultrapotassic and potassic rocks, Sr and Nd isotopic compositions 155 ultrapotassic rocks 103-4, 106, 107-8, 154, 155 univariant melting 44 uplift 144, 361,382, 418 epeirogenic 54-5 upwelling 22, 25 uranium 363, 376, 426 uranium deposits 485 urtite 66, 67, 68, 532, 534, 535,537, 541,543 Usaki complex (E Africa) 67 ussingite 484 vapour phase 352 buffering by 317 halogen-bearing 321 halogen-rich 328 vapour pressure, variation in 509 vapour-phase transfer 167 vapour-phase transport 313, 329, 331-2, 345, 347-8 varioles 213 Velasco alkaline province (E Bolivia), petrology of 403-12 petrogenesis 411-12 pyroxene and amphibole relations in the undersaturatedrocks 409-11 rock chemistry 405-9 verite 103, 108, 132 vesicular texture 422 vesiculation 233,525 villiaumite 466,484, 540 viscosity decrease in 307 high 322 increase in 9 lowering of 450 unusually low, related to F content 466 vlasovite 269 volatile accumulation 484
567
volatile complexing 313,326, 328, 329, 330, 331-2 volatile components accumulation of 479, 480, 540 acid 539-40 alkaline magmas 539-41 C-rich 467 dissolved 476-7 volatile content, and nephelinite magmas 60-1 volatile degassing 305, 317 volatile elements 232 volatile enrichment 218,481 volatile flux 7, 9, 10 volatile fugacities, lamprophyre melts 198 volatile loss 32, 308,483,484 volatile phase 323 volatile pressure release 482 volatile release 354 volatile saturation 308 volatile-phase transfer 167, 170, 321,461 volatiles 4, 7, 34, 37, 40-1, 72, 73,325 channelling of 361 and genesis and evolution of magmas 29 influx of 444 residual 8 upward concentration of 478 volcanic migration 237 volcanic rocks basic (Cameroon line), analyses of 275 intermediate and evolved (Cameroon line), analyses of 276 volcanic rocks, Trans-Pecos 423 volcanics mafic 8 potassic 135-6 shoshonitic 140 volcanism alkaline, Cameroon line 273 alternatingmafic-felsic 6, 10-11 anorogenic, within-plate 358 bimodal, basalt-trachyte 297 Biu Plateau 277 central, Kavirondorift 294 dispersed, Massif Central 518 fissural 388 intra-plate 273, 284 K-rich 144 nephelinitic 61 oceanic 273 Pan-African domains 361 parasitic 303 volcanoes bimodal 294 caldera 313, 314-15 basalt-trachyte 294 bimodal, Emuruangogolak 305,306 Quaternary 299, 303, 305 trachytic 303 central 518 nephelinite-carbonatite association (Kenya) 296-7 central, mixed 295,296, 298,299-300 Kilimanjaro 299, 303 Mount Kenya 299, 303 Olorgesailie 300, 303-4
568
Index
volcanoes (cont.) composite, Mount Cameroon 279 felsic 11 fissure 518 peralkaline 320 shield 423,495 basalt-trachyte 294, 309 basaltic 298 bimodal 303 limburgite and nephelinite 302 W Africa, geological map 274 W African craton 360, 381,388, 398 W Kimberleys, W Australia 39, 124, 129-30, 142, 144, 146, 148, 149, 154, 156, 160, 163, 164, 165, 167, 168 wadeite 108, 122, 123, 129, 163-4, 198, 200 wall-rock metasomatism 450 wall-rock reaction 89, 92, 284, 460 Walvis Ridge 263 Wasaki complex (E Africa) 67 water saturation 92, 93 water vapour, partial pressure of 288 wehrlite 40, 460 Wilson cycle 55 Pan-African 397 Pan-African Trans-Saharanorogenic belt 398 Winnett alnoite sill 127 wolframite 376 wolgidite 39, 40, 41,108, 164, 165 wollastonite 339, 349,465 wyomingite 103, 107, 108, 120, 121,133, 136, 164, 165 xenocrysts 95, 109, 164, 195 Cr-diopside 61 K-feldspar 388 kimberlite 167 olivine 233 plagioclase 368 xenoliths 3, 71, 95, 109, 154, 195,218,337,403,465, 495, 542 anorthositic 461,466 Ascension Island 269-71 basalt 303 basic 383 biotitite 109 biotite-clinopyroxenite 109 COz-rich inclusions 233 cognate 156, 282 cognate and accidental 138
crustal 212, 218,250, 523 cumulate 237 gabbroic 235 dunite 130, 167,236,237,242, 249 mineral chemistry of 233 feldspathic 461 gabbro 249 garnetlherzolite 109 glimmerite 170, 465 granite 269 granulite 116, 287 harzburgite 116, 130, 167, 169 Hawaiian, distribution of 249 kimberlite 168, 169 lamproite 123 Leucite Hills 122 lherzolite 116, 169, 233, 249-50 mantle-derived 213, 237,425-6, 524 mantle peridotite 202 MARID 169, 170 melasyenite 135 metasomatized 100, 170, 284 metasomatism of 169 olivine cumulates 233 olivine-rich 130 peridotite 160, 167, 170, 277, 278, 279, 282 phlogopite-pyroxenite 163 plutonic, composition and evolution of (Ascension Island) 269 Precambrian rocks 425 pyroxenite 109 spinellherzolite 109, 237, 249, 527 ultramafic 23,233,425,465 wehrlite 235, 236, 249 yamaskite 437 zeolites 109, 132, 163, 164, 193, 324, 338, 339, 349 zinnwaldite 370, 373,374 zircon 71,109, 133, 138, 164, 166, 168, 337, 338, 369, 376, 383, 387, 388,405,420, 424, 521 Zn-Mg-chromite 204 Zn-nordite 462 Zomba-Malosa complex 336, 339, 341,343, 345, 351 granite, quartz syenite and syenite 337, 352 zone of melting, ascent of 309 zone refining 166, 284, 306 zoning 160 of magmas 476 Zr/Ti, as index of fractionation 57 Zr, as index of fractionation 288
