The lowland American tropics have posed great challenges for archaeologists. Working in awkward terrain, in humid condit...
49 downloads
677 Views
13MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
The lowland American tropics have posed great challenges for archaeologists. Working in awkward terrain, in humid conditions where preservation is difficult, modern scholars pioneered new methods that increasingly influence archaeological practice internationally. The contributors to this volume all have substantial experience in the region. Their essays explore problems of site discovery, excavation, the preservation of artifacts and osteological and botanical remains, and methods of analysis. Specific technical innovations are discussed in relation to particular excavations. This book will be welcomed by all archaeologists, ecologists, and paleontologists working in the tropics.
Archaeology in the lowland American tropics
Frontispiece. Donald W. Lathrap (1927-1990).
Archaeology in the lowland American tropics Current analytical methods and applications
edited by
PETER W. STAHL Department of Anthropology, Binghamton University
CAMBRIDGE
UNIVERSITY PRESS
CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo Cambridge University Press The Edinburgh Building, Cambridge CB2 2RU, UK Published in the United States of America by Cambridge University Press, New York www. c ambridge. org Information on this title: www.cambridge.org/9780521444866 © Cambridge University Press 1995 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1995 This digitally printed first paperback version 2006 A catalogue record for this publication is available from the British Library Library of Congress Cataloguing in Publication data Archaeology in the lowland American tropics: current analytical methods and applications / edited by Peter W. Stahl; [contributors, James A. Zeidler . . . et al.]. p. cm. Includes bibliographical references. ISBN 0 521 44486 1 1. Indians of South America — Antiquities. 2. Tropics — Antiquities. 3. South America - Antiquities. 4. Archaeology Methodology. 5. Lathrap, Donald Ward, 1927-90. I. Stahl, Peter W. II. Zeidler, James A., 1949- . F2229.A73 1995 980'.01'028-dc20 93-51070 CIP ISBN-13 978-0-521-44486-6 hardback ISBN-10 0-521-44486-1 hardback ISBN-13 978-0-521-02738-0 paperback ISBN-10 0-521-02738-1 paperback
Dedicated to the vision of Donald W. Lathrap
Contents
List of figures List of tables List of contributors
page xi xv xvii
Introduction
i
PETER W. STAHL
1
Archaeological survey and site discovery in the forested neotropics
7
JAMES A. ZEIDLER
2
The archaeology of community organization in the tropical lowlands: a case study from Puerto Rico
42
PETER E. SIEGEL
3 Archaeological methods for the study of ancient landscapes of the Llanos de Mojos in the Bolivian Amazon
66
CLARK L. ERICKSON
4
Searching for environmental stress: climatic and anthropogenic influences on the landscape of Colombia
96
WARWICK BRAY
5
"Doing" paleoethnobotany in the tropical lowlands: adaptation and innovation in methodology
113
DEBORAH M. PEARSALL
6
Plant microfossils and their application in the New World tropics DOLORES R. PIPERNO
130
Contents 7
Differential preservation histories affecting the mammalian zooarchaeological record from the forested neotropical lowlands
154
PETER W. STAHL
8
Biological research with archaeologically recovered human remains from Ecuador: methodological issues
181
DOUGLAS H. UBELAKER
9
Interpreting dietary maize from bone stable isotopes in the American tropics: the state of the art
198
LYNETTE NORR
10
From potsherds to pots: a first step in constructing cultural context from tropical forest archaeology
224
J. SCOTT RAYMOND
11
Returning to Pueblo Viejo: history and archaeology of the Chachi (Ecuador)
243
WARREN R. DEBOER
References Index
263 305
Figures
Donald W. Lathrap (1927-1990) Frontispiece 1.1 Map of the Jama Valley showing three survey "strata" page 2.5 and various sub-strata 1.2 Histogram of the 130 sample quadrat returns from three survey 28 strata 1.3 Two STP sampling designs for a 1 ha quadrat: (a) five shovel32 probes per quadrat with staggered layout and specified spacing; and (b) eight shovel-probes per quadrat with hexagonal layout and optimal spacing 1.4 Site discovery probabilities in 1,000 simulation trials for two 38 STP sampling designs: (a) five probes per 1 ha quadrat; and (b) eight probes per 1 ha quadrat 2.1 Map of Puerto Rico showing the locations of the known early 44 Saladoid sites 2.2 Context of the Maisabel site 45 2.3 Distribution of the auger test pits 49 2.4 Isopleth map of prehistoric pottery, by weight, recovered from 49 the auger test pits 2.5 Topographic map of the Maisabel location 50 2.6 Three-dimensional plots of artifact densities at Maisabel 51 2.7 Isopleth map of prehistoric pottery, by weight, recovered from 52 the auger test pits 2.8 Distribution of macroblock features displayed by type 55 2.9 Portion of the ditch feature traversing one of the macroblock 56 excavation units 2.10 Density distribution of the macroblock artifacts as determined 58 by fe-means cluster analysis 2.11 Density distribution of macroblock artifacts, including the 59 contents of the ditch (six-cluster solution) XI
xii
List of figures
2.12
Ring model of the internal structure and organization of the Maisabel site 3.1 Location map of the Central Llanos de Mojos region with important sites and features mentioned in the text 3.2 Oblique low altitude aerial photograph showing prehispanic 3.3
3.4 3.5 3.6 3.7
3.8 3.9 3.10 3.11 3.12 3.13
3.14
4.1 4.2
4.3 4.4
raised fields at the El Villar site, 38 km east of San Borja Aerial photograph showing several complexes of prehispanic raised fields at the Santa Fe and La Envidia ranches, 19 km WSW of San Ignacio Digitally enhanced section of an aerial photograph of prehispanic raised fields near Santa Ana de Yacuma Aerial photograph of forested islas in the pampa west of San Ignacio Prehispanic raised fields at the La Envidia ranch Surface collections being made at La Asunta site, a large occupation site covering several hectares bisected by a logging road near San Ignacio Use of a coring device to find buried occupation sites and to determine the depth of agricultural features Clearing of vegetation from raised-field transects in preparation for topographic mapping at La Envidia A computer-generated topographic map of raised-field platforms and canals at the Beni Biological Station Excavation of Trench no. 1 at Santa Fe using natural stratigraphy Stratigraphic profile and soil descriptions of Trench no. 2/3/4 at the El Villar site Oblique aerial photograph of the experimental raised fields at the Beni Biological Station, Porvenir Ranch, 50 km east of San Borja Raised fields constructed for experiments by students of the Universidad Tecnica del Beni at the Beni Biological Station in 1992 Pollen diagram (abridged version) from the Hacienda El Dorado, Calima Region, Cordillera Occidental, Colombia Tentative temperature changes (F), humidity variations (G) and human influence on the natural vegetation (H), in the paramo zone of the Cordillera Oriental, Colombia (after Kuhry 1988) Changing extent of savannas (approximate) between the Gulf of Uraba and the Rio Magdalena (after Gordon 1957) Cultural development and climatic change in the lower San Jorge Basin (after Plazas et al. 1988)
63 69 76 76
JJ 77 81 82
83 85 86 88 89 93
94
98 100
103 105
List of figures 5.1 6.1 6.2 6.3 7.1 7.2 7.3
7.4
7.5 j.6 7.7
8.1 8.2 8.3 9.1
9.2 9.3
9.4
xiii
Diagnostic phytoliths from Pechichal, Feature 5 124 Location of the archaeological sites and lakes discussed in the 138 text (a)-(c) Phytolith and particulate charcoal records from Lake 143—4 La Yeguada sediments A summary pollen record from Lake La Yeguada sediments 145 showing absolute frequencies of major indicator taxa Map of the four major bioclimatic zones located within the 156 project area (after PRONAREG 1978) Jama Valley archaeofaunal totals 158 Relative weights and frequencies of Jama Valley mammalian 158 archaeofaunas arranged by arbitrary size classes (N = 7865, excluding human and recent remains) SEM imagery of selected rodent teeth identified in flotation 161 fraction. Top: Oryzomys spp., upper first molar (50X). Bottom: Proechimys, lower fourth molar (38X) Size classes of Jama Valley mammalian genera arranged 162 according to NISP (top) and MNI (bottom) Percentage skeletal representation of Jama Valley mammalian 163 size classes (a) Percentage of large mammalian Minimum Animal Units 164-5 compared to respective cervid bulk density measurements. (b) Observed NISP of White-tailed Deer Odocoileus virginianus compared to expected NISP survivorship based on relative density Geographical distribution of Ecuadorean sites yielding skeletal 186 samples Temporal change in adult stature 189 Temporal change in life expectancy at birth, age 5, and age 15 190 Isotopic composition of archaeological food resources in lower 201 Central America (after Norr 1990) (1.5% was added to the S13C value of modern foods to compensate for 12C enrichment of the atmosphere from the burning of fossil fuels, as per Tieszen 1991) Map of Panama with enlargement of study area in central 213 Pacific Panama Isotopic composition of human bone collagen from sites in 220 central Pacific Panama compared to those of human bone collagen from populations with isotopically distinct diets (data from Schoeninger et al. 1983) A 13Cca.co of human bone from sites in central Pacific Panama 220 compared to that of laboratory rats fed isotopically controlled diets (data from Ambrose and Norr 1993)
xiv
9.5
10.1 10.2 10.3
10.4 10.5 10.6 11.1
11.2 11.3 11.4
List of figures
Isotopic composition of prehistoric human diets in central Pacific Panama based on bone apatite carbonate S13C and bone collagen S15N compared to the isotopic composition of food resources The lower Apurimac and Upper Ene Valleys, showing locations of archaeological sites of the Sivia and Quimpiri complexes Diagrammatic illustration of rules which generate the most common vessel forms of the Sivia style Diagrammatic illustration of design rules which define the generation of incised and zone-painted design statements of the Sivia style (modified from Raymond et al. 1975: Figure 55(a), 55(b)) A sample of design statements of the Sivia style Diagrammatic illustration of rules which define the decoration of the most commonly decorated vessel forms of the Sivia style and of the generation of vessel categories Illustration of a Quimpiri style vessel with decoration (A) and two sherds from two pseudo-Sivia bowls which display Quimpiri style decoration (B and C) Upper left: late sixteenth-century map of northern Esmeraldas (based on Palop Martinez 1986: Figure 2). Lower left: location of sixteenth-century place names as plotted by Palop Martinez (1986: Figure 3) on a contemporary map of northwestern Ecuador. Right: portion of the "Mapa de la Zona Ecuatoriana de Integracion Fronteriza con Colombia" (SIFCE-DE 1973) showing major rivers and the location of Pueblo Viejo favored in the text Possible relationships among Chachi caciques (solid triangles) Chachi ceramics collected by Barrett in 1909 (left) and ceramics from Chachi archaeological sites (right) The Santiago and Cayapas Basins showing the distribution of Chachi sites (solid circles), Cantarana sites (open circles), and mixed Chachi-Cantarana sites (solid with open circles)
221
232 233 234
235 236 237 246
253 259 260
Tables
1.1
Sub-surface artifact densities from fifteen test pits at ten archaeological sites dating to the Ananatuba, Mangueiras, Formiga, and Marajoara Phases on Marajo Island, Brazil (compiled from Meggers and Evans 1957: 174-295) 1.2 Sub-surface artifact densities from seven test pits at the AGU 2 site, Aguatia River, eastern Peru (compiled from Myers 1972) 1.3 Effectiveness of quadrat sampling in the Jama Valley by survey strata 1.4 Mean densities, variances, and aggregation per 1 ha quadrat for upland zones of Jama Valley survey strata 1.5 Effectiveness of shovel test pit sub-sampling in the Jama Valley by survey strata 1.6 Sample output file (abridged) from Kintigh's (1988b) subsurface testing evaluation program (STP), showing simulation results from 1,000 trials using five-probes/ha sampling scheme 1.7 Sample output file (abridged) from Kintigh's (1988b) subsurface testing evaluation program (STP), showing simulation results from 1,000 trials using eight-probes/ha sampling scheme 2.1 Sampling strata areas and amounts excavated 5.1 Percentage occurrences of selected remains from the Pechichal site (M3B4-011, Feature 5) 6.1 Phytolith record from Panamanian archaeological sites 6.2Percentage of Heliconia phytoliths in modern and paleoecological assemblages, together with the proportions of these showing evidence of having been burned 6.3 Percentages of burnt Gramineae and arboreal phytoliths in modern forests 7.1 Jama Valley mammalian archaeofaunas arranged by relative size
14
15 29 29 30 34-5
35~6
54 124 140 147
148 159
xvi
7.2 8.1 8.2 9.1 9.2 9.3 11.1 11.2 11.3
List of tables
Natural histories of recent western Ecuadorean mammalian 178 genera identified in Feature 5 Samples of human remains from Ecuador, grouped into broad 182 cultural periods Temporal trends in biological attributes assessed from skeletal 188 samples Abbreviated chronology with central Pacific Panama sites 214 mentioned in the text Human bone collagen and apatite carbonate stable isotope 216—17 results from central Pacific Panama Summary statistics for central Pacific Panama sites 218 Comparison of six (I-VI) Chachi oral traditions pertaining to 248-50 their history Baptismal counts for the Chachi in 1597 (Monroy 1938: 252 334-348) A Chachi chronicle 256
Contributors
WARWICK BRAY,
Institute of Archaeology, University College London
Department of Anthropology, Queens College, City University of New York
W A R R E N DEBOER,
CLARK L. E R I C K S O N ,
Department of Anthropology, University of
Pennsylvania LYNETTE N O R R , Department of Anthropology, University of Illinois at Urbana-Champaign DEBORAH M. PEARSALL,
Department of Anthropology, University of
Missouri, Columbia DOLORES
R. PI PERN o, Smithsonian Tropical Research Institute, Balboa,
Panama j. SCOTT RAYMOND, Department of Archaeology, University of Calgary, Alberta, Canada P E T E R E. SIEGEL,
John Milner Associates, West Chester, Pennsylvania
w. STAHL, Department of Anthropology, Binghamton University, State University of New York PETER
H. UBELAKER, Department of Anthropology, Smithsonian Institution, Washington D.C.
DOUGLAS
A. Z E I D L E R , Department of Anthropology, University of Illinois at Urbana-Champaign
JAMES
Introduction PETER W. STAHL
The long neglected lowland American tropics have only recently received any significant amount of scientific attention. This unprecedented explosion of interest is the unfortunate product of a conspiracy of events whose global repercussions have forced us to confront the dramatic consequences of rapid ecosystemic degradation, declining biodiversity, and cultural extinction. As we race against time to learn as much as possible about these quickly transforming environments, we are consistently frustrated and humbled by how little we actually know. In our search for a clearer appreciation of the future implications that current policies and practices may hold, we are simultaneously compelled to reflect on the area's rich and complex archaeological past. The relative paucity of systematic archaeological investigation conducted throughout this vast region is generally attributed to some combination of logistical constraints, lack of ground visibility, meager preservation, and/or an historic deprecation of lowland environments. Much of our knowledge of lowland neotropical prehistory is built upon a scattered patchwork of museum pieces and fortuitous observations. Together with a healthy dose of speculation, these isolated bits of data are linked together via trait comparison, and correlated with reliable observations obtained from the few geographically disparate scientific excavations undertaken so far. Amongst the small group of intrepid archaeological pioneers whose efforts unearthed an early glimpse at the prehistoric world of the vast South American lowlands, Donald W. Lathrap occupies a prominent niche. Spanning four decades of research, Lathrap's career was marked by a bold rethinking of the role played by lowland areas in prehistoric New World developments. Frequently depicted as an area whose limited potential forfeited any claim to prehistoric achievement, lowland neotropical prehistory was subsequently marginalized as the degenerated consequence of externally derived influences. Lathrap's synthetic vision emphatically inverted this viewpoint. His research thrust the lowland neotropics into the forefront of consideration, and provided the intellectual stimulus for subsequent research into new and diverse geographical, temporal, and topical areas.
2
PETER W. STAHL
The impact of his legacy as a scholar and teacher has been addressed in a number of recent posthumous tributes, 1 and will be appreciated for many years to come. As a lasting recognition of Lathrap's vision, he motivated generations of researchers who continue to conduct archaeological investigations in the lowland neotropics. The often widely divergent but mutually complementary methodological perspectives taken by those he inspired, reflects yet another aspect of Lathrap's profound influence. Much of this analytical diversity can be attributed to the heightened appreciation which lowland-oriented archaeologists must bestow upon maximizing data recovery. This stems from a very practical concern for successfully extracting information in a burial environment repeatedly perceived as hostile toward the preservation and detection of all but the most durable remains. Preservation biases, whether perceived or real, were always at the forefront of Lathrap's attention. He chided those who cited the regular litany of problems associated with conducting archaeology in the neotropical lowlands as a justification for using simple or inadequate field techniques. He chose to view this substantial list of grievances not as an obstacle, but as a challenge. He dogmatically cajoled his colleagues and students to critically evaluate the data upon which they based their interpretations, and constantly coaxed them to maximize data recovery wherever, whenever, and however possible. Most of the contributors to this volume worked and studied closely with Lathrap; many are his former students. Each author has confronted the issue of maximizing data recovery and interpretation in neotropical lowland contexts. In addressing the crucial considerations of preservation and detection, each paper explicates methodologies designed to maximize, redefine, and/or perfect data recovery and interpretation. These concerns are illustrated from the complementary perspectives of different specializations, and through the inclusion of tangible applications derived from archaeological contexts within the lowland American tropics. The eleven contributions to this volume are sequentially arranged. They encompass: archaeological survey; site excavation; studies of regional landscapes and paleoenvironments; analysis of paleobotanical and osteological materials; approaches to ceramic analysis; and the critical application of ethnohistoric and contemporary oral records. The volume begins with two contributions which examine aspects of site survey, recovery, and definition in the lowland neotropics. Zeidler emphasizes the omnipresent problems of poor surface visibility and site accessibility, which exacerbate the systematic discovery of archaeological sites in a regional landscape. His contribution explores the nature of lowland neotropical archaeological sites, and the specific field methodologies used to achieve both broad and representative coverage. The effectiveness of probabilistic sampling, through the use of small quadrats and shovel test probes, is evaluated with data from the forested lowlands of western Ecuador. Siegel continues, by addressing the problems of visibility and deep site excavation. Adapting solutions developed by archaeologists facing similar problems in other areas, he demonstrates a multi-
Introduction
3
stage nested framework of site detection and sampling at the complex site of Maisabel, Puerto Rico. A circular pattern of community organization is convincingly demonstrated through a tiered recovery program of auger test-pitting and expanding horizontal excavations. Next, consideration is given to the study of regional archaeological landscapes in space and time. The difficulties of studying tropical lowland garden and field systems within a regional landscape are addressed by Erickson. Major limitations imposed by constantly reworked stratigraphy, poor preservation, and the inadequacy of direct historic analogs, are overcome through a combination of specific field methods and experimental archaeology. The massive precolumbian land modifications of the Bolivian Llanos de Mojos are effectively studied through integrating a mixed media approach to remote sensing, with aerial and pedestrian survey, and excavation. An ongoing program of agricultural experimentation not only provides direct clues for the interpretation of this prehispanic landscape, but also offers tangible recommendations for modern development and sustainable agriculture. Bray continues with a diachronic examination of the interplay between global, local, and human-induced environmental factors in the prehistoric landscape of Colombia. Stressing the need to interrelate events within lowland and adjacent highland areas, he demonstrates that subsistence and environmental data can be recovered for lowland settings. These can reveal massive regional landscape changes, including large-scale drained field constructions and irreversible degradation, brought about through the interaction between natural and anthropogenic environmental factors. Despite the repeated axiom that organic preservation is exceedingly poor to non-existent, significant information can be fruitfully recovered, analyzed, and integrated into archaeological research. Pearsall addresses the problems of botanical macroremain preservation, recovery, and identification in the diverse neotropics. She stresses the maximization of botanical data during all phases of project planning, excavation, and analysis. This can be achieved through the successful adaptation of recovery and sampling techniques, and appropriate identification in adequate comparative collections. Problems and prospects involved in optimizing pollen and phytolith microremain data for archaeology in the lowland neotropics are also discussed. The productive integration of macroand microremain data is advocated, and illustrated for northwestern South America. This theme is elaborated by Piperno, who provides a general review of pollen and phytolith analyses carried out under the limiting conditions of the American tropics. Significant improvements to paleobotanical technique are illustrated through the tandem application of microfossil analyses to the study of prehistoric subsistence and settlement in Panama. Pollen, phytolith, and charcoal records reveal the antiquity of slash and burn cultivation in the neotropical lowlands. She advocates the retrieval of data from culturally created, nonoccupational contexts, combined with the systematic construction of modern analog data for the comparison of natural and cultural effects on vegetation. The study of organic remains in lowland neotropical contexts is further
PETER W. STAHL
explored in three papers dealing with osseous residues. Stahl critically examines the basic axiom that poor bone preservation is principally the result of hostile burial conditions in lowland environments. The qualitative attributes of a large mammalian archaeofaunal collection from the western Ecuadorian lowlands, are used as an entry point to illustrate the complex set of variables that can affect the survivorship of animal bone assemblages. These variables, which are strongly affected by animal size, can combine to systematically distort preservation, recovery, and subsequent archaeological interpretation. The effects of intrinsic and systematic biases on inferences regarding ancient lowland environment and subsistence are considered. Ubelaker addresses the major problems associated with the analysis and interpretation of human skeletal remains in neotropical areas. His contribution focuses on the extent of representation in sampled skeletal populations, and offers recent analyses of human skeletal collections from adjacent lowland and highland areas in Ecuador. The temporal and spatial breadth of this uniquely large sample presents an unusual opportunity to examine ancient variation in biological information from these areas. Norr outlines current developments and methodology for dietary reconstructions, using the stable carbon and nitrogen isotope ratios in collagen and apatite fractions from archaeological human remains. She details the situations in which isotopic ratios are recommended for dietary reconstruction in the neotropics, the relationship between dietary isotopic composition and consumer tissue, and the selection of appropriate samples. Possibilities and limitations in the use of stable isotope ratios as one line of evidence for answering interrelated questions about early New World maize agriculture, are illustrated with data from Panama. Due to factors of preservation, analyses of durable ceramic fragments have overwhelmingly comprised the major source of interpretive data for archaeologists. Raymond explicates the methodology of modal analysis, which was earlier advocated by Donald Lathrap and fruitfully applied over the last three decades in the tropical lowlands of South America. Pointing out some of the limiting factors of typological approaches in these environments, he clearly explicates the underlying rationale and methodology of structural analysis, with illustrations from the Upper Peruvian Amazon. The meaningful integration of technological variables into this approach is further discussed with early examples from northwestern South America. Finally, despite many of the limiting conditions imposed by working in tropical lowland environments, DeBoer emphasizes the privileged position occupied by archaeologists working in the tropical forests. Here, thriving ceramic expertise and extant mythology and oral traditions are powerful tools in aid of archaeological interpretation. DeBoer critically synthesizes data from oral traditions, historic chronicles, and archaeology as they inform on the disputed origins of the Chachi who currently reside in the Santiago-Cayapas region of northern Esmeraldas Province, Ecuador. The independent role of archaeological data supports the Chachi version of their own origins.
Introduction
5
A recurrent theme emerges from these various attempts at maximizing and perfecting data recovery in neotropical lowland settings. This theme counters historically prevailing notions which have only served to increasingly isolate the area's prehistory, at least on methodological grounds. It is apparent that many of the difficulties associated with conducting lowland archaeology are basically similar to those faced by archaeologists working in other parts of the world. Although not distinctly unique to lowland environments, they may nevertheless be worse in matters of degree. Relatively inadequate infrastructure is a common feature encountered by archaeologists working in many areas, yet it is ubiquitous and generally more inadequate in the lowland tropics. Poor visibility caused by vegetation cover, or site destruction caused by looting or meandering rivers, are problems frequently confronted by archaeologists, yet are regular and often exaggerated facts of life in the tropical lowlands. Differential bias in material cultural preservation is a fundamental feature of all archaeological research, yet the lowland neotropical archaeologist must be particularly vigilant in controlling this pernicious variable in any attempt at interpretation. Potential solutions to all these conerns, and more, are found in the expanding repertoire of a robust archaeological discipline. They need only be judiciously adapted to local conditions, and their findings cautiously appraised. Donald Lathrap was perhaps best known for his hemispheric-wide perspectives which enabled a radical and occasionally controversial rethinking of lowland American tropical prehistory. The tropical lowlands were often geographically isolated through the use of heuristic dichotomies (for example, lowland/ highland, marginal/nuclear, "tribaP7"chiefdom") which frequently masked more than they facilitated, a point clearly brought out in a number of the contributions. It is not surprising that Lathrap, who trained and worked in various areas of North America, brought a hemispheric perspective to the practice of archaeology in the South American tropical lowlands. Perhaps less known or acknowledged, are the many significant methodological developments he either introduced or directly stimulated. These include undertaking tropical riverine surveys from a Mississippian perspective in Peru; introducing midwestern techniques of extensive aerial archaeological excavation in Ecuador; explicating ceramic analysis modeled after the methodology of descriptive linguistics; applying some of the earliest ethnoarchaeological studies to archaeological interpretations; and stimulating the use of flotation recovery and phytolith studies in the New World tropics. It is through his stimulus and encouragement to younger generations of archaeologists, that his legacy continues to discover and redefine prehistory in the lowland American tropics.
6
PETER W. STAHL
Notes The contributors to this volume kindly extend their thanks to Cambridge University Press and its staff for providing a venue in which to express their ideas, and to David Minor, of the Department of Anthropology, University of Illinois, who supplied the photograph, taken of Donald Lathrap in 1979.
1 Recent tributes include proceedings of a 1991 symposium at Cumana, Venezuala: Homenaje al Dr. Donald W. Lathrap, edited by Erika Wagner (1991), and published in Antropologica, 75-76; "Gifts to the Cayman: Essays in Honor of Donald W. Lathrap," edited by Evan C. Engwall, Margaret van de Guchte and Ari Zieghelboim (1992), and published in the Journal of the Steward Anthropological Society, 20; and the 1992 symposium "Model Building and Validation in New World Archaeology: Papers in Honor of Donald W. Lathrap," held at the 91st Annual Meeting of the American Anthropological Association in San Francisco. A thoughtful reflection on Donald Lathrap's scholarly accomplishments has been published by Jose Oliver (1992).
1
Archaeological survey and site discovery in the forested neotropics JAMES A. ZEIDLER
Until recent years, archaeologists seldom carried out large-scale surveys in regions having poor visibility and accessibility. Survey in such areas requires a variety of heroic and methodologically unlovely techniques ..., such as periodic shovel testing and use of local informants, to simply make site discovery possible (Schiffer 1987: 350).
While the preceding statement is true of numerous areas in New World archaeology, the dual problems of low surface visibility and limited accessibility are probably nowhere greater than in the tropical lowlands of South America, particularly Amazonia. These problems, combined with the sheer immensity of the area and small number of archaeologists working there, have led to widely conflicting interpretations of the archaeological record, not to mention opposing reconstructions of macro-regional prehistories (Gibbons 1990; Roosevelt 1991). In spite of these differences, all archaeologists working in this area have had to confront the severe logistical constraints and preservation biases imposed by the humid tropical environment, and in so doing, have been forced to employ "methodologically unlovely techniques" of one kind or another. My purpose in this chapter is not to enter into this protracted debate on Amazonian prehistory. Instead, I explore certain methodological themes relating to archaeological survey and site discovery in the forested tropics of lowland South America, using as my point of departure Donald Lathrap's seminal article entitled "Aboriginal Occupation and Changes in River Channel on the Central Ucayali, Peru" (1968a). This article is methodologically significant for two reasons. First, it outlined a specific field procedure for locating remnant archaeological sites in a broad meandering riverine environment where destructive fluvial forces commonly leave only a palimpsest of former prehistoric occupations. Using techniques common to contemporaneous archaeological investigations in the Mississippi floodplain, Lathrap employed aerial photographs to document the "horizontal" stratigraphy left by complex meander sequences. Remnant archaeological sites were found only in bluff areas of older alluvial terraces, but nonetheless revealed long stratigraphic sequences. Even though these sites have been spatially truncated by fluvial processes, "the ancient communities seem neither to be small nor particularly short-lived" (Lathrap 1968a: 75). He further notes that: Most of the sites which once existed within the Central Ucayali flood plain will never be seen by the archaeologist and, in most instances, have ceased to exist, for each meander loop completely destroys any previous sites within its limits.
8
JAMES A. ZEIDLER
Sites located on the bluff of old alluvium directly adjacent to the flood plain have a better chance of surviving and of being found by the archaeologist, but even here the odds are not good (Lathrap 1968a: 76). In some cases, cultural midden was buried by sterile alluvial deposition. The Cumancaya site (UCA-22), for example, "would be impossible to locate from surface indications were it not for the fact that wave action is continually cutting sherds out of the bank . . ." (Lathrap 1968a: 74). As an early documentation of natural formation processes in floodplain habitats of Amazonia and their effects on the archaeological record, Lathrap's study represents a methodological tour de force given the intellectual climate of the time. Second, and perhaps more importantly, in its critique of "traditional" archaeological survey methods, the article represents an early and valiant attempt at illustrating bias in the archaeological record of Amazonian vdrzea environments. It also demonstrated a cautious approach towards negative evidence, and warned against premature conclusions regarding site densities in areas where survey intensity was low. Two quotes adequately demonstrate these points: Any site more than 100 years old is bound to be in a different spatial relationship to the active river channel than when it was occupied ... One would predict that longitudinal site survey along the river, unless augmented by frequent trips back to the bluffs lining the floodplain, would be a most inefficient way to locate old archaeological sites. This has indeed proved to be the case (Lathrap 1968a: 75). . . . I would hold that the negative evidence from rapid site surveys, along the routes which are presently most accessible, is not reliable. The failure of such initial surveys to find remains of a particular culture is not good evidence that peoples of that culture did not migrate through or occupy the region in question. Furthermore, such surveys will certainly give a misleading picture of past population densities (Lathrap 1968a: 77; emphasis in original). By today's standards for conducting regional archaeological survey, Lathrap's discussion of site discovery may seem somewhat antiquated. Present-day archaeological field techniques and hi-tech wizardry for locating and examining sites far surpass those available to Lathrap and associates in the 1960s. Nevertheless, it would be a mistake to overlook his fundamental methodological insights on the nature of the archaeological record in neotropical floodplain environments, or his early recognition of biases inherent in contemporary archaeological surveys. Indeed, many investigators in Amazonia still conduct field surveys in essentially the same manner that Lathrap was criticizing, whether due to financial limitations, methodological ignorance, or both. In the remainder of this article, I pursue in greater detail two themes related to the foregoing: (1) the nature of archaeological "sites" in the forested neotropics (including various factors affecting their discovery probabilities); and (2) specific field methodologies designed for efficient, systematic, and representative discovery of neotropical sites in a regional landscape (as well as an explicit assessment of
Archaeological survey and site discovery bias in the evaluation of survey results). Finally, these themes are briefly illustrated through a case study from the Jama Valley in the western Ecuadorian lowlands. Emphasis is placed on the need to balance efficient and representative archaeological survey sampling with the logistical constraints of conducting fieldwork in neotropical environments. Here, surface visibility is generally low to nonexistent due to dense vegetation cover. Accessibility is severely constrained by characteristics of vegetational growth, geographical remoteness and lack of infrastructural support. Discovery probabilities and site constituents in the neotropics Earlier researchers in Amazonian prehistory have been quick to point out the strong preservation biases in neotropical archaeological sites which leave "only a few axe fragments and a vast mass of smashed pottery" (Lathrap 1970: 63; see also Meggers and Evans 1957; Hilbert 1968, for similar comments) to carry the interpretive burden in archaeological reconstruction. However, in spite of the general truth of such statements for certain sites, it is also true that feature contexts, charred organic remains and clear evidence of stratigraphic deposition have also been documented in lowland neotropical sites, such that earlier complaints of extreme preservation biases throughout the entire neotropics may be slightly exaggerated. As Roosevelt has recently argued, "it is a fact that archaeological sites of many periods abound in the tropical lowlands, and the great majority have abundant stratigraphic and structural patterning and numerous features of artifacts, carbonized macroscopic and microscopic plant remains, and faunal remains, as well" (1991: 118). Moreover, recent advances in archaeological data recovery techniques have helped mitigate, to a certain extent, some of these preservation biases. Perhaps the most important of these is flotation sieving of archaeological sediments for fine-fraction retrieval of macrobotanical, archaeofaunal, and artifactual materials (see Pearsall, Stahl, this volume). Likewise, recent advances in the study of soil micromorphology and sedimentation (Courty et al. 1989) would certainly aid in the interpretation of archaeological site formation processes and midden deposition rates, in instances where visible stratigraphy has been leached out of excavation sidewalls. Systematic refitting studies of conjoinable artifacts can also aid in the interpretation of midden deposition, depositional rate, and post-depositional formation processes (see Villa 1982; Villa and Courtin 1983; Schiffer 1987: 359-362). However, very little research of this nature has been carried out in the neotropics to date. Considerable headway has been made in overcoming at least some of the limitations imposed by severe preservation biases. Nevertheless, the lowland neotropical archaeologist is still confronted with a common set of factors or variables which determine discovery probabilities in the archaeological record, the effects of which have yet to be fully explored in this special environment. In a detailed treatment of site discovery procedures in areas of dense vegetation cover,
9
IO
JAMES A. ZEIDLER
McManamon (1984) describes a range of properties commonly found in archaeological sites, collectively referred to as site constituents. These primarily include artifacts, features, and anthropic soil horizons, although human-induced modifications of soil chemistry, magnetism, or other soil characteristics are also pertinent. Regarding all but the latter, McManamon (1984: 229) notes that data on their relative frequency of occurrence is often difficult to find in excavation reports. However, where they do exist, such data support a "general impression of many archaeologists about the relative intrasite abundance and spatial distribution of these three site constituents" (McManamon 1984: 232—233). Artifacts, defined as "the portable products and byproducts of human activities" (McManamon 1984: 228), are almost always the most widespread and abundant of site constituents. Cultural features, defined as "sharply delimited concentration^] of organic matter, structural remains, soil discoloration, or a mixture of these and artifacts" (McManamon 1984:229), generally fall far behind artifacts in abundance but are, nevertheless, detectable by subsurface testing procedures. In contrast, anthropic soil horizons are defined as "extensive deposits] that might be sharply or diffusely delimited . . . [which] result from deposition of large amounts of organic remains in a roughly delimited, relatively large (compared to features) area" (McManamon 1984: 229). These phenomena are less well reported in the literature. More often than not, their presence is simply noted and analytical attention is restricted to the artifacts or features contained within them. In any case, like features, they "do not commonly approach the extended spatial distribution of artifacts and in some cases might not even exist in a site area or large portions of it" (McManamon 1984: 233). In support of these relationships, McManamon cites three case studies from the eastern United States which employ different excavation techniques: (1) horizontal stripping of large areas to expose site structure (Illinois); (2) deep trenching in search of buried archaeological sites (Tennessee); and (3) sub-surface testing by shovel-probes and small test pits (Massachusetts). Given this set of site constituents, discovery probability can be "formally defined as the likelihood that cultural remains of interest will be detected within a sampling domain or sampling unit using a specified sampling procedure, given a certain level of sampling effort" (Nance 1983:292—293). As Nance and Ball (1986; see also Krakker et al. 1983) have pointed out, discovery probability is itself a product of two independent probabilities: intersection and productivity. The former is "the probability that a test pit intersects the site," while the latter is "the probability that a test pit yields artifacts, given that it has intersected a site surface" (Nance and Ball 1986: 459). Both of these are in turn influenced by sampling design (especially survey intensity) and the properties of the archaeological remains under study. For example, Schiffer et al. (1978) have enumerated at least three properties of the archaeological record which are important in this regard: abundance, clustering, and obtrusiveness. The effects of artifact abundance and relative clustering on site discovery probabilities have been well
Archaeological survey and site discovery studied for cases of surface inspection by systematic pedestrian survey (Nance 1983: 312-316). In general terms, assuming a constant level of survey intensity, as artifact abundance (average artifact density) decreases and/or artifact clustering (spatial aggregation) increases, discovery probability will decrease significantly. The obtrusiveness of the archaeological remains (in terms of size, shape, color, and so on) also contributes to this probability, in that site detection is greatly facilitated by easily recognizable cultural items (for example, large vessel fragments of highly decorated ceramics) or exposed features (such as, architectural remains, stone alignments, and so on). Thus, a short-term campsite made up of a series of light density clusters of lithic debitage scatter would have a considerably lower discovery probability. This stands in contrast to a site of comparable size that was composed of high density remains of pottery and large groundstone fragments dispersed evenly over the site surface in association with several cultural features. These effects become even more pronounced in cases where sub-surface sampling procedures, such as shovel test pits (STP) or augering are necessary due to low surface visibility or buried cultural deposits (see below; also Erickson, Siegel, this volume). Since the effective "inspection window" is drastically reduced by such procedures, the combined effects of low abundance and high clustering of cultural remains can result in large numbers of negative shovel probes, even though a site may have been intersected (see Wobst 1983: 68-71; McManamon 1994 for detailed discussion). Three other critical variables in evaluating discovery probabilities are: visibility; accessibility, and survey intensity, all of which are especially problematic in the neotropics. Visibility "refers to the extent to which a site has been buried or covered by soil aggradation and vegetation since its occupation (McManamon 1984: 224; see also Schiffer et al. 1978: 6-j). Low to nonexistent visibility due to dense vegetation cover is, of course, a common problem throughout the neotropics. This can be caused by old growth forest vegetation with dense understory, more recent secondary growth forest, or the dense groundcover of tall grasses used for pasturage. Often, even cultivated plots exhibit completely obscured ground surfaces due to thick leaf litter (for example, cacao and plantain), or lack of regular weeding. Deeply buried sites are common in aggradational floodplain settings, and the long-term effects of bioturbation and gravity in upland settings may also result in buried archaeological deposits (Michie 1990). In the neotropics, one source of bioturbation noted by Lathrap (1968a) in his Ucayali research, is the earth-moving ability of leaf-cutting ants. Their activity can deposit up to 15 cm of sterile sediment on top of an archaeological midden, even on a blufftop (see Evans and Meggers i960: 237 for a description of similar deposits overlying archaeological sites in Guyana). These land surface conditions have drastic consequences for discovery probability and logistic efficiency in regional archaeological survey. They require the use of laborintensive subsurface testing programs in lieu of, or in addition to, pedestrian surface inspection (see Siegel, this volume).
n
12
JAMES A. ZEIDLER
Accessibility is the ability to physically inspect a given area of terrain. Cases of low accessibility could be caused by difficult terrain or dense vegetational growth, which may impede or reduce mobility. Recent landscape modifications may completely destroy evidence of archaeological occupations or cover them entirely with large expanses of soil, water, or modern construction (see Erickson, this volume). Denial of access by landowners is another common occurrence, which can result in unsurveyed zones of a study area. Again, the effects of low accessibility on logistic efficiency and discovery probabilities can be considerable. Finally, survey intensity refers to the spacing between crew members during pedestrian survey, and the ensuing thoroughness with which the ground surface is visually inspected. In sub-surface surveys, it refers to the spacing and layout of test pits, shovel-probes or augers, and the thoroughness with which the test-pit fill is inspected (that is, whether trowel-sifted, screened, and so on) and the sidewalls examined for anthropic soil horizons. It is usually measured as the number of person-days expended per unit area surveyed (for example, person-days per square mile, Schiffer and Wells 1982). The degree of survey intensity employed in an archaeological survey is, of course, contingent upon the overall research objectives of the project and the resultant kinds of archaeological resources that are actively sought. For example, if the goals of the project do not include the discovery of preceramic sites, a program of systematic deep coring in search of such buried sites need not be implemented. However, if one is interested in locating archaeological remains representative of the entire archaeological sequence of a region, including all types and sizes of sites, then survey intensity must be commensurate with the complexity of the site discovery probabilities (that is, site discovery procedures must be fine-grained enough to intersect and detect the smallest of archaeological sites and lightest artifact density known for the region). If this cannot be achieved, then both the targeted minimal site size and artifact density should be explicitly stated. Additional factors to be considered include the aforementioned degree of artifact clustering and nature of artifact obtrusiveness. It is only through such bias assessment that the adequacy of survey coverage can be reliably evaluated. If the survey methodology is biased against the discovery of small and ephemeral sites, then the goal of representativeness is compromised. In the neotropics, few survey designs have had the recovery of a representative sample of sites in a region as their stated goal. In most cases, regional surveys have been conducted as a preliminary reconnaissance to locate any sites at all. The sites which are eventually discovered, and perhaps tested, then become "typical" of the region or even macro-region for purposes of constructing ceramic chronologies. Thus, it is difficult to evaluate neotropical surveys in terms of the validity and effectiveness of their field methodologies vis-a-vis the archaeological record. However, these pioneering efforts have generated a useful corpus of data pertaining to the nature and complexity of archaeological sites. Given the pessimism of Lathrap, Meggers, Hilbert and others regarding the preservation of
Archaeological survey and site discovery archaeological remains in lowland neotropical sites, it is worth reviewing briefly some of the corresponding site constituents as they have been recorded in the literature. In spite of the formidable logistical difficulties involved in locating these sites, it would appear that the constituent artifacts, features, and anthropic soil horizons of known archaeological sites provide some measure of optimism. Surface artifact densities and their spatial distribution are not often treated in the lowland neotropical literature. Here, sites can rarely be surface collected in a systematic controlled manner unless they coincide with modern settlements or large areas devoid of vegetation cover. In some cases, surface collections have been conducted, but have followed an uncontrolled "grab bag" approach for diagnostic artifact retrieval only, so that density measurements are impossible. Sub-surface artifact densities, on the other hand, can often be compiled from testpit data, thus providing valuable information on artifact abundance. However, little data is provided on artifact density-distribution, since extensive systematic test-pitting or shovel-probing over an entire site area has rarely been carried out (see Siegel, this volume). Myers (1973: 236) has lamented that, "a test pit at one site; two at another is characteristic of most of the work that has been done," although this situation has improved somewhat in the last twenty years. In spite of these limitations, the sparse evidence of sub-surface artifact densities that does exist is highly informative. Myers (1972: 543) has argued that "there are two basic kinds of sites in the Tropical Forest: multifunctional habitation sites that are cut out of the jungle, associated with agricultural fields and with good fishing or hunting resources; and unifunctional campsites, associated with the resources for a particular activity."1 If this dichotomous site typology is essentially correct, then the two kinds of sites should exhibit notable differences in such basic characteristics as size, depositional history, artifact assemblages, surface and sub-surface artifact densities, feature densities, and the development of anthropic soil horizons. Two such examples are briefly mentioned here as a means of illustrating relative artifact abundance and site variability in the neotropics. The first example is drawn from the pioneering research of Meggers and Evans (1957) on Marajo Island at the mouth of the Amazon, where numerous large habitation sites were tested with varying numbers of square test pits. These measured either i x i m , 1.5x1.5 m, o r 2 X 2 m i n size, with depths ranging from 0.30 to 2.25 m. Table 1.1 presents sub-surface artifact densities, for which the data were complete enough to compile density measurements in fifteen test pits from ten archaeological sites pertaining to the Ananatuba, Mangueiras, Formiga, and Marajoara Phases. For each site, the test-pit dimensions and volume are listed, together with total artifact count and artifact density per 0.25 m3 (following the convention of McManamon 1984: 232). As these density figures demonstrate, sub-surface artifact abundance in these sites is consistently high, with the lowest densities occurring in Marajoara Phase mound fill. Average sub-surface artifact
13
14
JAMES A. ZEIDLER
Table I . I . Sub-surface artifact densities from fifteen test pits at ten archaeological sites dating to the Ananatuba, Mangueiras, Formiga, and Marajoara Phases on Marajo Island, Brazil (compiled from Meggers and Evans 1957:174—295). Asterisks indicate artifact counts and densities for sherds only. Site
Test pit
Ananatuba Phase sites 1 J-7 (Sipo) 2 J-7 (Sipo) 1 J-9 (Ananatuba) J-10 (Sororoco) 1 Mangueiras Phase sites 1 J-5 (Croari) 1 J-13 (Bacuri) 1 C-3 (Porto Real) 2 C-3 (Porto Real) J-17 (Anajas) 1? Formiga Phase sites 1 J-6 (Formiga) J-6 (Formiga) 2 J-6 (Formiga) 3 1? J-18 (Coroca) Marajoara Phase sites J-15 (Os Camutins) 1 Mound 14 Mound 17 1
Total no. Density Dimensions (m) Volume (m3) artifacts (/0.25 m3) 2x2 2x2
2.40 3.60
3,026 6,833
474-5
2 X 2
2.40
4*724 2,660
492.0 281.7
0.90
1,470
0.75 0.40
3.O35 998
408.2 449.0 623.7
0.32
721
563-3
i-5
X
i-5
I X 1
i-5
X
i-5
I X I I X I
i-5
X
i-5
2.36
0.67
315.0
101.5
2x2
3.60
3.7O4
2 X 2
2.40
1,004
I X I
0.60
2,011
i-5 X i-5
1.69
644
i-5 X i-5 i-5 X i-5
5.06 2.70
1,2-75*
900 *
257.2 104.6 837.9 95-3
63.0 83.3
density for the first thirteen sites (the Marajoara figures are excluded because only sherds are tallied) is 384.9 artifacts per 0.25 m3 (SD = 222.3). Interestingly, none of the test pits executed at Marajo failed to produce cultural material, although the authors observed that in the case of the Formiga Phase sites, there is "an accumulation of sherds in several independent spots with sterile areas between them, rather than in one continuous area" (Meggers and Evans 1957: 249). At the remaining sites, refuse accumulation was apparently continuous over the site surfaces. The second example is the AGU 2 site on the Aguatia River, eastern Peru, tentatively interpreted by Myers (1972) as a dry season turtle-hunting and eggcollecting campsite. Here, a series of seven test pits of uniform size (5x5 ft) but varying depth (0.30 to 1.75 m), were placed in a purposive (non-systematic) manner along the right bank of the Aguatia River (Myers 1972: 542). Table 1.2
Archaeological survey and site discovery
15
Table 1.2. Sub-surface artifact densities from seven test pits at the AGU 2 site, Aguatia River, eastern Feru (compiled from Myers 1972). Site AGU 2 AGU 2 AGU 2 AGU 2 AGU 2 AGU 2 AGU 2
Total no. Density Test pit Dimensions (ft) Volume (ft3) artifacts Density (/ft3) (/0.25 m3) i 2
4 5 6 7 8
5x5 5x5 5x5 5x5 5x5 5x5 5x5
50.0 31.2
100
2.0
33
1.1
87.5
2-95 2-75
2.9 3.1
87.5
395
143.7 56.2
501 818
4-5 3-5 14-5
100.0
1.6 0.9 2.4 2.6
3-7 2.9 12.0
presents sub-surface artifact densities for these test pits, along with dimensions (in ft), volume (in ft3), and total number of artifacts. The density measurements are given both in ft3 (following the English system used by Myers), and in 0.25 m3 (to permit greater comparability with the Marajo data). Not surprisingly for this type of site, sub-surface artifact abundance is dramatically lower than that from the large habitation sites on Marajo. Average sub-surface artifact density at AGU 2 is only 3.7 artifacts per 0.25 m3 (SD = 3.8), over a hundred times lower than the average for the thirteen Marajo test pits! Although cultural material was encountered in all seven test pits, site discovery probabilities would be much lower in this case due to the small size of the site itself, and the generally low artifact densities. Assuming dense vegetation cover over the surface of such a small ephemeral site, the hypothetical placement of a large number of smaller test pits (for example, 40 x 40 cm shovel-probes) would probably have resulted in several negative probes due to the low overall artifact density. It is interesting to note that AGU 2 was not initially "discovered" through the test-pitting operation, but rather was located through observations of cut-bank erosion along the Aguatia River (Myers 1972: 543). Little systematic data is available on feature abundance and diversity in lowland neotropical sites. However, even a cursory examination of the literature demonstrates that features have been found at sites where investigations have moved beyond a testing phase and large area excavations been carried out. For example, urn burial features are common at sites in the Ucayali River area (Raymond et al. 1975; Raymond, this volume) and at Marajo Island (Meggers and Evans 1957; Roosevelt 1991). Roosevelt has found evidence of large multiple baked clay ovens or stoves, hearths, postholes, garbage-filled pits, and numerous prepared floors in her extensive excavations at Marajo (1991). Hearth features and posthole evidence were encountered in smaller excavations at the Taperinha site near Santarem in the lower Amazon (Roosevelt et al. 1991). From these few examples, we can conclude that features do indeed exist throughout the
l6
JAMES A. ZEIDLER
neotropics. In many cases, they are well preserved, but discovery procedures have generally not been intensive enough to locate them. Their presence seems to follow the general pattern noted by McManamon (1984: 233) for North American archaeological sites: that is, that they "do not commonly approach the extended spatial distribution of artifacts and in some cases might not even exist in a site area or large portions of it." In general, large habitation sites with long intensive occupational sequences should yield proportionately more evidence of cultural features, as well as a greater diversity of features, than small ephemeral sites having seasonal short-term occupations. Finally, anthropic soil horizons are site constituents to be considered (see Bray, Erickson, this volume). Roosevelt's (1991) intra-site geophysical prospecting on Marajo Island examined anthropic soils for magnetic and electrical anomalies, which were then selectively excavated as a form of "ground-truthing." Apart from locating the numerous feature contexts noted above in magnetically anomalous areas, secondary garbage and rubble middens were also located and subsequently excavated in areas of intermediate electrical conductivity. This provided complementary archaeological data to that recovered in the feature contexts. As she notes, "geophysical survey is as yet the only effective method for intensive and systematic subsurface investigation of very large archaeological sites . . . Only remote sensing can penetrate deposits and cover large areas rapidly enough to give high-resolution information for large archaeological sites" (Roosevelt 1991: 145). In spite of the relatively small amount of archaeological research carried out in the neotropical lowlands, compared to the temperate regions of the Americas, considerable attention has been given to one kind of anthropic soil horizon: the terra preta do Indio or "black Indian soils" found throughout the Amazon Basin (Falesi 1974; N. Smith 1980; Eden et al. 1984; Eidt 1984; Sombroek 1984; and more recently, Mora et al. 1991). In the most thorough treatment of these soils to date, N. Smith (1980) describes them as follows: Terra preta is an anthrosol, characterized by a distinctive anthropogenic epipedon. Intermixed potsherds and celts are a major distinguishing feature . . . [but] the dark color of terra preta is the most striking feature of the soil. . . The color ranges from jet black to dark gray-brown and is probably related to the time the site was occupied (N. Smith 1980: 553, 556). For present purposes, the important aspect of these anthropic soil horizons is that they are almost always isomorphic with large prehistoric settlements and/or their associated agricultural lands (Herrera et al. 1992; Mora et al. 1991). Moreover, they are found on a variety of geomorphological surfaces in both vdrzea and terra firme habitats, but are typically "either on the bank of a perennial water course, or within a few hundred meters of one" (Smith 1980: 562). N. Smith (1980) compiled a list of twenty-nine terra preta sites throughout the Amazon basin, ranging from 0.3 to 90.0 ha in areal extent and from 0.15 to 1.47 m in depth. He argues convincingly for a rate of deposition on the order 1 cm per 10 years, and
Archaeological survey and site discovery notes that many black earth sites represent multi-component occupations having considerable time depth. In contrast, then, to the North American data reviewed by McManamon (1984), at some neotropical sites the "black earth" middens may very well be as extensive as are the surface artifact assemblages which are traditionally used to define sites. In this case, the anthropic soils alone can serve as a useful indicator of archaeological sites, particularly when satellite imagery and remote sensing techniques are employed for regional analysis of archaeological distributions (see below, and Erickson, this volume). The distinctive properties of "black earth" sites should result in equally distinctive spectral signatures in the imagery when contrasted with surrounding terrain. We can see from this rapid survey of site discovery probability in the lowland neotropics and the nature of site constituents, that preservation biases, while present, may not be quite as limiting as previously thought. The fundamental constituents commonly recognized in archaeological sites are certainly present and often occur in admirable abundance. At issue are the appropriate field methods for locating archaeological sites in a regional lowland neotropical landscape. Survey methodologies and site discovery procedures for the neotropics Archaeological research in the neotropical lowlands involves many of the problems found in other areas of the world characterized by buried or otherwise obscured archaeological remains. What is different is the lack of methodological ingenuity for improving site discovery probabilities under these adverse conditions. In particular, Roosevelt (1987a, 1989, 1991) has repeatedly pointed out the methodological shortcomings of previous research at all levels of field investigation in Amazonia. Nowhere is this more apparent than at the regional level of survey and site discovery. Traditional survey of riverine sites in the tropical lowlands of South America has typically been carried out through "longitudinal" site surveys along a given watercourse (see, for example, Evans and Meggers 1968; Hilbert 1968; Lathrap 1968a; Meggers 1991,1992; Meggers et al. 1988). As Meggers (1991:199) observes: "most of the known sites are along the present courses of the major tributaries of the Amazon, including black, clear, and white-water rivers. They are thus 'riverine' in location, but the sustaining area is terra firme." Survey of inland localities in the terra firme zones has usually been accomplished through interviewingfieldguides (see, for example, Evans and Meggers i960),2 and, more recently, through right-of-way transects "where roads facilitated access inland" (Meggers 1992: 198). It is clear that little or no archaeological survey in the neotropical lowlands has been carried out from a probabilistic perspective where study areas are precisely defined, the area surveyed is a statistically defined sample of a larger sampling universe, survey intensity is explicitly defined, and rigid spatial controls are imposed over the surface and sub-surface inspection methods (see Weiland 1984
17
18
JAMES A. ZEIDLER
for an important exception). While the traditional survey methods were reasonable for a pioneering stage of research in the 1950s and 1960s, they are untenable by today's standards. If field sampling in neotropical ecology and biology is routinely conducted in a well-controlled probabilistic fashion, then there is no justification for archaeological survey and site discovery procedures to lag behind. As mentioned previously, perhaps the single most limiting logistical condition for carrying out archaeological survey in the neotropical lowlands is low accessibility. This takes two forms depending on the scale involved. One is due to geographical remoteness and lack of infrastructural resources which make transport and mobility difficult within a study area. It is not surprising that few terra firme zones have been intensively sampled, and that most surveys have been carried out in a purposive fashion along riverbanks accessible by watercraft. When sampling a regional study area, the severity of this problem depends on the number of sampling units and the representativeness of their spatial distribution over the landscape; the more units placed and the greater the spacing between them, the greater will be the travel time necessary to locate their geographical boundaries. The second form occurs at a smaller scale. It involves restricted accessibility to specific areas of interest within a given study area, such as sampling units (quadrats and so on) or landscape features selected for purposive sampling. This may be due to unusually difficult terrain, extremely dense multi-storied vegetation, or denial of access by landowners. In any case, logistical efficiency is greatly impaired while time and labor costs increase exponentially. For example, dense vegetation growth often requires labor-intensive machete clearing of "swathes" or transects within a sampling unit in order to properly locate subsurface shovel-probes. As we shall see below, the amount of time needed to complete this clearing is often equal to, or greater than the time required to execute the shovel-probes themselves. Directly related to the problem of low accessibility due to dense vegetational growth, is the issue of low surface visibility due to dense ground cover. In cases where surface visibility is less than about 25 percent, traditional pedestrian survey must be replaced by sub-surface testing such as shovel test pit (STP) sub-sampling or augering (see Erickson, Siegel this volume). Quite often a given sampling unit will exhibit variation in surface visibility, which permits pedestrian inspection in some areas while requiring shovel testing in another. Of course, visibility is also a problem in areas where archaeological deposits have been completely obscured by subsequent deposition from a variety of geomorphological processes. In these cases, deep coring or trenching may be necessary to determine the presence or absence of archaeological sites by locating buried anthropic soil horizons (see Erickson, Siegel, this volume). Given this host of logistical impediments to archaeological survey in the forested neotropical lowlands, how might an archaeologist proceed if his or her
Archaeological survey and site discovery
19
goal is to achieve broad and representative coverage of a given study area? Is regional archaeological sampling even an option in such an environment? If neotropical archaeology is to move beyond the stage of purposive reconnaissance and limited site testing as the sole basis of archaeological knowledge, then it is imperative that some form of regional landscape archaeology be implemented within well-defined study areas. Perhaps the most efficient way to grapple with the problem of defining a regional study area and planning an archaeological survey in the neotropics, is through analysis of multispectral digital satellite imagery by remote sensing techniques (Sever and Wiseman 1985; Limp 1989; Behrens and Sever 1991). Large areas can be rapidly assessed for their vegetational, geological, and ecological characteristics, without having to rely on ground-based infrastructure (Behrens 1991; see Erickson, this volume). If a digital elevation model (DEM) of the study area can be acquired, then the survey can be planned with full prior knowledge of local topography and hydrography, regardless of the quality of existing topographic or planimetric maps. Survey planning carried out with recent satellite imagery could also include detailed consideration of land cover, surface visibility, and accessibility, so that appropriate site discovery procedures and field logistics can be developed prior to initiation of fieldwork. Through use of previous information on human settlement derived from ethnographic and ethnohistoric sources, as well as prior archaeological reconnaissance, predictive models of prehistoric settlement behavior can be postulated and "ground-truthed" through systematic field survey. If archaeological sites have been previously documented in the study area, their particular spectral properties can be used to search for other probable site localities. Barring this, more traditional methods would be needed for study area definition and efficient survey planning, using high-quality aerial photographs as well as detailed topographic maps at scales of 1:50,000 or, preferably, 1:25,000. For many areas of the neotropics, such maps are simply not available. In some cases, aerial photographs can be scanned into digital format for analysis by remote sensing and GIS techniques (see Erickson, this volume). Also, many countries have produced geological, soil, vegetation, and modern land-use maps at reasonable scales which are based upon digital satellite imagery. However, coverage may be spotty, and the scales not always optimal (for example, 1:200,000). Thus, the distinct advantages of digital satellite imagery are its total coverage, commercial availability, and user-defined analytical potential. The imagery can be "ground-truthed" or geo-registered with Global Positioning System (GPS) technology, which has been shown to be reasonably effective even in tropical rain forest environments (Baksh 1991; Chagnon 1991; Wilkie 1989). Precise locational information on each archaeological site can be acquired as the field survey proceeds, and all relevant archaeological and ecological variables can be incorporated into a GIS format for statistical spatial analysis and long-term data management.
2O
JAMES A. Z E I D L E R
Once a study area has been defined, using either satellite imagery or traditional cartographic analysis, a specific regional sampling design can be developed in accordance with the overall goals of the investigation. The design should ensure discovery of a representative sample of archaeological sites, or certain kinds of sites, over a regional landscape. This requires that a series of procedural decisions be made in accordance with the local landscape conditions and the nature and complexity of the archaeological record (Plog et al. 1978; Schiffer et al. 1978). As Read (1975: 60) has noted, "there is no single best sampling procedure for regional surveys. The sampling procedure must take into account at least these important parameters: the information desired, the distribution of that information in space, cost of obtaining samples, and degree of precision needed, etc." In spite of the recognized difficulties of logistical mobility and accessibility in the lowland neotropical environment, where total survey coverage is an obvious impossibility, some form of probabilistic sampling is still feasible and is certainly preferable to purposive reconnaissance along major waterways and roadcuts. The central issue is how to maximize survey efficiency in light of these logistical impediments. The question of survey efficiency has two complementary aspects: "first, how one samples in a manner that satisfies the demands of statistical inference and, second, how one samples in a manner that satisfies the archaeological requirements" (Read 1986: 480) of a given project. Nance (1983) has referred to these, respectively, as statistical precision sampling and discovery model sampling. The latter is concerned with "the effective discovery of cultural remains," while the former is concerned with the "efficient estimation of the quantitative properties of the remains discovered" (Nance 1983: 291; emphasis in original). Thus, the two perspectives have different goals and different measures of efficiency or success. In discovery model sampling, success is measured as a ratio of discovery probability to cost (for example, number of sites discovered per person-day). In statistical precision sampling, efficiency is measured in terms of the ratio of precision (expected sampling error) to cost, with a low ratio representing greater efficiency (Nance 1983). The best way to ensure that the requirements of statistical inference are met, is through probabilistic sampling techniques such as simple random sampling in which each sampling unit is selected independently of all other units in the sample. Quadrats or transects of varying size and shape have traditionally been employed as a form of cluster sampling (Scheaffer et al. 1986; Read 1975), since the sites themselves are not being sampled directly. Rather, the landscape is being sampled by small spatial units, within which sites may or may not be discovered. The probability of discovery depends on the intensity of the inspection methods within the sampling units, and on the size, density, and spatial distribution of the sites relative to sample unit size and number (Read 1986). Where the items of interest are distributed in an aggregated or clustered manner, such as archaeological sites, then small sampling units are desirable in order to maximize statistical
Archaeological survey and site discovery
21
precision (Green 1979; Scheaffer et al. 1986). For regional archaeological sampling, "one can approximate a direct random sample of the population of sites by using quadrats sufficiently small in area so that the quadrat contains only one site or zero sites" (Read 1986: 481). Read goes on to note, however, that "the drawback is pragmatic: the much greater time cost for locating, surveying, and traveling among many small quadrats versus a smaller number of larger quadrats. The differences in time and cost affect the site sample size that can be obtained for a fixed time budget" (Read 1986: 481). For reasons outlined above, this pragmatic drawback can be quite severe in the neotropics, thus necessitating the use of long transects as the basic sampling unit (for example, Weiland 1984). Transects are much easier to lay out as well as traverse when penetrating dense vegetation. If transect size is relatively large and fewer transects are needed, logistical efficiency increases correspondingly; however, some loss of statistical precision will occur. In the end, "a balance between size and number of plots (sampling units) must be achieved . . . [but] there are no good rules that always hold for making this decision" (Scheaffer et al. 1986: 200). In areas where considerable ecological heterogeneity could affect the relative distribution of archaeological sites, careful attention should be given to regional stratification of the study area into subareas or survey "strata." Independent samples are then drawn from each stratum, thus maximizing within-stratum homogeneity and reducing the error in estimations based on a single heterogeneous study area. In forested neotropical lowlands such as Amazonia, past researchers have often assumed that little ecosystemic heterogeneity exists. This is typified by the common use of the basic dichotomy between varzea and terra firme environments. However, Moran (1990) has shown that when one is working at finer local scales, considerable heterogeneity does exist, and must be taken into account. Reliance, until recently, on the simple dichotomy between thefloodplainand the uplands, comprising 2% and 98% of the area respectively, is at a level of generality not likely to generate systematic scaling of the Amazonian regional system and implies that each of the two areas are more homogeneous than is the case . . . When one changes level from the Amazon as a whole to specific subregions, the homogeneity suggested at the regional level rapidly yields to extreme local variability (Moran 1990: 287-288). Any microenvironmental differences that may be significantly correlated with the distribution of archaeological sites, such as ecological zones, soil zones, certain drainage patterns, elevations, and so on, should be isolated into survey strata for independent sampling. A final consideration in developing sampling designs for lowland neotropical environments, has to do with the problem of reduced or nonexistent surface visibility. As mentioned previously, this problem is not unique to the neotropics. Where archaeologists have not chosen to ignore it completely, some form of systematic sub-surface testing has been employed, such as test pits, post-holing,
22
JAMES A. ZEIDLER
or augering (see Erickson, Siegel, this volume). Post-holing and augering provide extremely small sub-surface tests, usually resulting in a bulk sediment sample but little or no cultural material (McManamon 1984). However, augering can be especially effective and logistically efficient for the discovery of deeply buried sites in alluvial settings (for example, Muto and Gunn 1981). Test pits provide a much larger "inspection window" and are thus preferable for sub-surface artifact recovery. They often measure as much as 1 x 1 m, but smaller "shovel probes" measuring from 0.25 x 0.25 m up to 0.50 x 0.50 m are more common. They can be executed in fairly rapid succession by small crews. When taken to a uniform depth, shovel-probes provide standardized data on probe volume, and hence, sub-surface artifact density where cultural materials are found. Sidewalls can often reveal valuable data on anthropogenic soil horizons. Most often, probe fill is carefully screened, but in some cases trowel-sorting is employed. Several studies exist which evaluate the overall effectiveness, reliability, and validity of STP survey (for example, Hasenstab 1986; Kintigh 1988a; Krakker et al. 1983; Lightfoot 1986; McManamon 1984, 1992; Nance 1983, 1990; Nance and Ball 1986; Shott 1985; Wobst 1983). However, as "they involve a very critical examination of the landscape" (Lightfoot 1986: 500), perhaps the most salient feature of intensive sub-surface testing programs is their labor-intensiveness. Shovel-probes provide a relatively small "inspection window" on the land surface when compared to pedestrian survey in areas of high visibility. Since practically all site discovery depends on the shovel-probe results, a large number of probes must be executed for effective site discovery, whereby confidence estimates can be placed on the minimal site size expected to be found with a given probe size and probe spacing. Moreover, when an isolated find is recovered from a single probe, additional probes must be executed in its vicinity either to verify the presence of a "site" through the discovery of additional artifacts, or to provide sufficient negative evidence to conclude that a site is not present. In comparing his intensive STP survey on Long Island to pedestrian surveys in the Southwest United States, Lightfoot (1986) has observed that the pedestrian field crews cover between sixteen and thirty-two times more space per unit of time (for example, acres/person-day) than field crews in STP surveys. What is noteworthy for present purposes is that the Long Island environment is relatively benign in terms of vegetation cover, accessibility, and logistical infrastructure. In the neotropical lowlands, STP costs rise dramatically where dense vegetation requires that machete crews precede the shovelprobe crews in all of their movements across the landscape. Topographers in these environments have labored under such conditions for years, however, so that the costs in time and manpower are well understood (A. Iturralde, personal communication 1991). In spite of the generally high labor costs of STP sub-sampling, it provides the only systematic solution to site discovery in the forested neotropics, enabling the
Archaeological survey and site discovery
23
archaeologist to move beyond purposive reconnaissance and informant-driven prospecting as the sole sources of archaeological knowledge. Some form of rapid geophysical prospecting along transects would be a tempting substitute for the excavation of multiple shovel-probes. However, these techniques have traditionally, and most effectively, been employed for intra-site prospecting once a site has already been defined (for example, Roosevelt 1991). Their use in an extensive landscape survey may be prohibitively expensive, especially since large areas devoid of sites would have to be surveyed as well. Also, any anomalies would have to be "ground-checked" through test pits anyway. However, as a site discovery tool, certain techniques as electromagnetic (EM) conductivity could productively be used in tandem with STP techniques, in order to locate the extent of midden areas in the vicinity of positive shovel-probes only. Conductivity also has the advantage of high portability and rapid execution time. Magnetometry techniques, while fairly portable and rapid, would be less effective for site discovery purposes. They are more suited to detection of fixed cultural features, which are much rarer occurrences in a regional landscape than dense concentrations of artifacts (McManamon 1984). It is clear from the foregoing discussion that regional archaeological survey in the neotropical lowlands poses unique problems in terms of logistical efficiency and appropriate site discovery procedures. However, neither the importance of probabilistic sampling techniques, nor the necessity of STP sub-sampling, can be overemphasized in areas of low visibility. In spite of their difficulty, it is only through such methods that a truly representative sample of archaeological sites can be recovered across a regional landscape. Perhaps more importantly, by controlling the spatial coverage of the study area, they provide the statistical means by which objective bias assessment can be carried out. In this way, the survey methods employed can be rigorously assessed in terms of the kinds of archaeological resources that they are not likely to find. Regional survey and site discovery in the Jama Valley, coastal Ecuador
As a brief illustration of probabilistic sampling techniques and shovel-probe testing in a lowland neotropical setting, we can turn to a case study from the Jama River Valley in northern Manabi Province, coastal Ecuador (Zeidler and Pearsall 1990; 1994). The life zone ecology of this area ranges from dry tropical forest along the coastal strip, to humid pre-montane forest at higher elevations farther inland (Canadas 1983). Three summer field seasons (1989-91) were dedicated to regional survey and selective site-testing within a 785 km2 study area, covering the central axis of the Jama drainage (Figure 1.1). The study area cross-cuts three distinct physiographic zones, including "semi-arid coastal plain," "sub-humid coastal cordillera," and "humid upland valley," ranging in elevation from sea level to c. 600 masl (Zeidler and Kennedy 1994).
24
JAMES A. ZEIDLER
Our overall goal was to study the interrelationship between prehistoric settlement dynamics, social complexity, and agricultural production of the poorly studied Jama-Coaque chiefdoms and their Formative Period precursors. Since little prior research had been carried out in this region, another goal of the project was to establish a radiocarbon chronology for the valley, and to develop a reliable ceramic sequence for these successive occupations (Zeidler and Pearsall 1994). The cultural sequence encompasses over 3,000 years of coastal Ecuadorian prehistory, ranging from the Terminal Valdivia Phase of the Early Formative Period (c. 1650 BC) up to a final Integration Period Jama—Coaque II occupation, which was truncated by the Spanish Conquest in AD 1531. Archaeological sites are thus highly variable in terms of size, function, and density across the valley landscape. Our survey design employed a probabilistic sampling strategy along the lines suggested by Read (1986) for maximizing statistical representativeness and survey effectiveness. This involves: (1) "fine-grained stratification of a region that includes relative density and spatial clustering of sites as part of the criteria for defining strata" (Read 1986: 491); and (2) the differential coverage of high- and low-density areas in order to maximize site recovery. This strategy recognizes that "differential density and clustering of sites represents a minimal feature of spatial distributions that can be related to the structure of settlement systems" (Read 1986: 491). Important in this regard are the magnet sites (Altschul and Nagle 1988), defined as sites whose location affects the location of other sites within a regional settlement system (that is, large civic-ceremonial centers with monumental architecture). In order to retrieve information on the spatial patterning of site clusters and magnet sites, greater survey effort is given to these areas (up to 100 percent coverage). The more extensive, lower density areas are sampled by randomly placed quadrats at whatever sampling fraction is deemed necessary for a representative sample of sites (Zeidler 1991). Accordingly, the stratification scheme of the Jama survey employed the three physiographic zones mentioned above, which respectively comprise Strata I, II, and III in our probabilistic sampling design (Figure 1.1). Strata I and III were further subdivided into a series of sub-strata of known high site density and a greater proportion of large sites and principal magnet sites. As Read (1986: 491) notes, "fine-grained stratification of a region may require strata consisting of noncontiguous segments of space." In Stratum I, these include three large pockets of alluvial bottomland (I/A/1-3) along the main channel of the Jama River, and two shoreline areas on either side of the river mouth (I/B/1-2). In Stratum III, a total of fourteen pockets of alluvial bottomland were isolated as sub-strata, along the main channel of the Jama and major tributary streams. All nineteen of these sub-strata were surveyed with 100 percent coverage for maximum retrieval of information on the spatial distribution of principal magnet sites in the settlement system. The upland or non-alluvial areas in Strata I and III were assumed to have lower
Archaeological survey and site discovery '^90
z5
'©10
00
Jama Valley Study Area
9980
onvento 99 7 o
'60
alluvial deposition (silty clays)/ (fluventic hapludoll and/or tropofluvent) alluvial deposition (silty clays)/ (vertic ustropept and vertic ustifluvent) recent alluvium (silty sands)/ (typic ustifluvent)
10
=1=
20km
=1
Figure 1.1. Map of the Jama Valley showing three survey "strata" and various sub-strata.
site densities, less aggregation, and generally smaller sites than the alluvial areas. The non-alluvial upland zone, which comprises the entirety of Stratum II> was assumed to have even lower site densities and smaller site sizes due to a more inhospitable landscape, and total lack of floodplain development. Accordingly, these upland areas in all three strata were randomly sampled by numerous small
26
JAMES A. ZEIDLER
quadrats measuring i ha each. Some 40 quadrats were placed in Stratum I, 30 in Stratum II, and 60 in Stratum III, giving a total of 130 quadrats (or 130 ha) of randomly sampled area. Sampling fraction was roughly equivalent in Strata I and II (approximately 0.21 percent), but somewhat lower (0.14 percent) in Stratum III. Since some degree of site clustering was expected, even in these upland zones, small quadrats were employed rather then large quadrats or long transects, so as to increase statistical precision (Nance 1983; Plog 1976; Read 1986).3 This decision was made at some cost in logistical efficiency, however, since numerous quadrats fell in relatively remote areas, in areas of rough terrain, or both. The resulting constraints on travel time and survey accessibility often hindered efficient and timely execution of quadrat inspection. In spite of these constraints, the small quadrats were reasonably effective at discovering archaeological sites in proportion to their probable densities across the three survey strata (see below). Space does not permit a thorough description of the sampling design and field procedures carried out in each stratum, nor the results obtained. Only two aspects of this research will be explored, so as to illustrate both the special methodological constraints on regional sampling in the lowland neotropics, and the general effectiveness of quadrat sampling and STP sub-sampling methods in spite of these constraints. The effectiveness of quadrat sampling in upland landscapes In evaluating the effectiveness of quadrat sampling in the Jama study area, it is important to point out that uniform surface inspection procedures could not be carried out in the 130 upland quadrats due to variable surface visibility conditions. Five quadrats were visited but not surface inspected because they fell over extremely precipitous terrain where sustained human occupation or even occasional use is precluded. Thus, it was assumed that no archaeological sites would be discovered in these cases. Of the 125 remaining quadrats falling on terrain hospitable for human use, some 34 quadrats presented visibility conditions sufficiently high (that is, greater than 80 percent) to permit inspection of the ground surface solely by traditional pedestrian techniques. This involved two crew members carefully walking the terrain in parallel swathes across the quadrat, while maintaining an optimal spacing of 20 m between swathes. Assuming that sites manifested themselves on the surface with moderate artifact density and visual obtrusiveness, this spacing interval ensured that extremely small sites would be intersected and detected. These quadrats almost always occurred in the immediate vicinity of modern settlements, but sometimes fell in large tracts of recently cleared or cultivated land. In the remaining 91 quadrats, however, dense surface vegetation resulted in very low to non-existent visibility in the entirety of some quadrats (17) and only partial visibility in others (74). In a few cases, this was due to primary stands of
Archaeological survey and site discovery
27
dry tropical forest (Stratum I) and humid pre-montane forest (Strata II and III). In most cases, however, it was due to dense secondary forest growth in fallow fields, or to dense cover of an artificially introduced pasture grass, Panicum maxima (Zeidler and Kennedy 1994). In these cases, the quadrat was examined by means of STP sub-sampling with shovel-probes uniformly measuring 40 cm x 40 cm x 40 cm. In the 17 zero-visibility quadrats, five such shovel-probes were employed in an X-shaped pattern. In the partial-visibility quadrats, varying numbers of shovel-probes (ranging from one to four probes) were employed, depending on the degree of vegetation cover, but generally in the same locations provided by the "X" pattern. The remaining areas of the quadrat where visibility was relatively high were inspected using pedestrian techniques described above. In these cases, then, site discovery procedures involved a combination of pedestrian surface inspection and sub-surface shovel testing, and in some quadrats a given site was discovered concurrently by both techniques. A uniform inspection method for all sampling units is a desirable goal of any sampling design. Several studies have shown that "the ability to perform an adequate search must be consistent from quadrat to quadrat" (Nance 1983: 312). While such uniformity is often achievable in areas of high surface visibility and little physiographic variability, combinations of the above described surface and sub-surface discovery procedures must be employed in other areas not characterized by these conditions. For example, Spurling (1980), productively employed both surface and sub-surface inspection techniques for substantially larger quadrats (500 m x 500 m) in a zone of mixed vegetation cover in the temperate forests of western Canada. In such cases, greater uniformity could be achieved by simply executing all shovel-probe sub-samples for a given level of survey intensity, regardless of whether high visibility in certain areas would permit surface inspection by pedestrian methods. Visual inspection of the surface would then provide a supplementary search procedure in open areas of some quadrats, while uniformity in sub-surface inspection would be maintained for all quadrats. While this is a laudable goal, in the interest of minimizing person-days expended per quadrat so as to maximize the total number of quadrats covered, it was not carried out in the Jama Valley, nor in Spurling's (1980) study area in western Canada. The overall effectiveness of the shovel-probe sub-sampling will be dealt with in the next section. Here, quadrat effectiveness is briefly examined in terms of overall discovery probabilities between the three survey strata, regardless of the particular inspection technique(s) used. The effectiveness of random quadrat sampling in the upland zones of the study area can be examined both from a site discovery and a statistical perspective (Nance 1983). With regard to site discovery, the relatively small quadrats far exceeded our expectations in locating archaeological sites. Given the binomial nature of the outcomes, where a quadrat finds either zero sites (negative return) or one site (positive return), it was expected that far fewer than 25 percent of the quadrats in each stratum would result in the discovery of a site. It was also
28
JAMES A. ZEIDLER
Stratum I (n=40)
Stratum (n=30)
Stratum I (n=60)
40-i
40 n
40-,
30-
30
30-
20-
20-
10
10-
53 i_
•D
(0
o
- 20H o .Q
E 13
10-
1
0
1
0
1
Sites per Quadrat Figure 1.2. Histogram of the 130 sample quadrat returns from three survey strata.
expected that if the results differed in the three strata, the results obtained in Stratum II would be considerably lower than those from Strata I and III, where site densities were perceived to be higher. Figure 1.2 graphically illustrates the binomial quadrat returns from each of the three survey strata. In all three cases, negative quadrat returns outnumber positive returns, but the margins are much smaller than expected on the basis of our previous knowledge of the study area. Table 1.3 gives the relative discovery probabilities of quadrat sampling for the
Archaeological survey and site discovery
29
Table 1.3. Effectiveness of quadrat sampling in the Jama Valley by survey strata (p = probability of encountering a 'positive3 quadrat [that is, one site]). Stratum no. I II III Total
Total no. quadrats 40
Positive quadrats
P
Approx. 95% interval
17
0.4250
0.2718—0.5782
8
0.2667
0.1085—0.4249
0.4000
0.2760-0.5240
0.3769
0.2936-0.4602
30 60 130
49
Table 1.4. Mean densities, variances, and aggregation per 1 ha quadrat for upland zones of Jama Valley survey strata. The variable vim is the variance! mean ratio and the variable k is equal to the negative binomial index of aggregation. Stratum no.
n
X
s2
v/m
k
I II III
40
0.425
3.294
0.270
1.179
60
0.400
3-515
7-751 4-355 8.787
0.063
30
0.081 0.051
three strata and for the entire study area, along with the 95 percent confidence interval for each probability. As these figures indicate, the quadrats in Strata I and III were much more successful at locating sites (42 percent and 40 percent, respectively) than their Stratum II counterparts (27 percent), largely due to different site densities in these areas. For the Jama Valley study area as a whole, roughly 38 percent of the 1 ha quadrats located an archaeological site; this implies both a relatively high density of archaeological sites across the upland landscape and a reasonably effective site discovery procedure. From a statistical perspective, the quadrat returns are of interest for the differences they exhibit between the three survey strata. Table 1.4 shows the mean site density, variances, and measures of aggregation for the 1 ha quadrats in the upland zones of the three strata. Strata I and III show close similarities for these statistics, while Stratum II is clearly differentiated. For Stratum II, site density is lower and both the variance/mean ratio and the k parameter indicate slightly less clustering in site distribution than is the case with Strata I and III. This pattern is interesting in view of the fact that Strata I and III contain the alluvial floodplain soils where large magnet sites are concentrated. Thus, their adjacent upland landscapes would be expected to contain relatively higher densities of sites than the dissected upland landscapes in Stratum II where no floodplain soils are located and no large magnet sites have been identified. In this sense, Stratum II may have served as a less densely populated buffer zone between the middle and lower reaches of the Jama Valley.
3O
JAMES A. Z E I D L E R
Table 1.5. Effectiveness of shovel test pit sub-sampling in the jama Valley by survey strata (p = probability of encountering a 'positive' shovel probe). Stratum no. I II III Total
Quadrats w/STPs
Total no. STPs 81
Positive STPs
P
Approx. 95% interval 0.0730-0.3468 0.0657-0.4477 0.1691-0.4695 0.1955-0.3485
20
39
17 10
37
119
38
0.2099 0.2564 0.3193
239
65
0.2720
34
The effectiveness of shovel test pit (STP) sub-sampling Like the small quadrats discussed above, shovel-probe effectiveness can also be examined as a binomial experiment where a given probe has only one of two possible outcomes: (1) finding no artifacts (negative return); or (2) finding one or more artifacts (positive return). In sampling terms, "the practice of examining test units within primary survey quadrats is an example of subsampling (i.e., twostage cluster sampling)," and the number of clusters (shovel-probes) per quadrat can be variable or constant (Nance 1983: 320). Since the Jama Valley quadrats varied in terms of the specific search procedures employed in their inspection (visual walk-over versus shovel-probe sub-sampling, or a combination of both), the total number of shovel-probes executed does not correlate with number of quadrats and thus cluster sizes are variable. A detailed statistical treatment of STP returns for each stratum, and the corresponding parameter estimates of stratumwide archaeological remains, lies beyond the scope of the present discussion. Here, shovel-probe results will be treated in cursory fashion as aggregate probabilities by survey stratum only. Table 1.5 shows the number of quadrats having one or more shovel-probes (column 2), as well as the total number of STPs executed by stratum (column 3) and the total number of positive shovel-probes (column 4) by stratum. For the entire study area, some ninety-one quadrats were at least partially sub-sampled by shovel-probes. A total of 239 shovel-probes was executed, 65 of which yielded cultural material. The probability (p) of encountering a positive shovel-probe is given for each stratum and the total study area in column 5, along with their corresponding 95 percent confidence intervals. Here, artifact discovery probabilities gradually increase as one moves inland from Stratum I to Stratum III, and range from 21 percent to 32 percent. For the study area as a whole, then, roughly 27 percent of the shovel-probes executed were successful at detecting archaeological remains. This is a fairly high return for the labor invested in locating a quadrat in areas of dense ground cover, clearing the vegetation necessary for locating the shovel-probes, excavating the probes, and recording the results. The relatively high p values obtained in the Jama Valley shovel-probe results are no
Archaeological survey and site discovery doubt partially due to the phenomenon of spatial autocorrelation, where a positive probe is likely to occur near other positive probes, simply because they have detected the same archaeological site (Nance 1983). Still the results demonstrate that shovel-probe sub-sampling can be an effective means of initially locating archaeological sites in areas where the ground surface is obscured by dense vegetation. What becomes difficult at this stage of investigation, depending on visibility and accessibility conditions, is site assessment (that is, the determination of the areal extent and intra-site artifact densities of a site). Once a site is located in the field by sub-surface testing procedures, assessing the nature and significance of the site by further vegetation clearing and sub-surface testing can be extremely costly in time and labor. Ideally, shovel-probing should extend out from initial positive probes so as to track sub-surface artifact densities across a site surface and eventually determine site boundaries (for example, Lightfoot et al. 1987). At large sites, this may be impractical, and in these cases some type of geophysical prospecting (such as electromagnetic conductivity) in long transects may be of value if vegetation can be cleared and accessibility permits equipment transport. A related aspect of the effectiveness of sub-surface testing has to do with the question of bias assessment (that is, what kinds of archaeological resources are likely to have been missed by implementation of a given sampling design and survey intensity). What is the minimal site size that a sub-surface testing scheme is likely to detect with an acceptable degree of confidence? In the case of pedestrian surface inspections, this is a relatively straightforward question of the spacing between crew members as they cover the sampling unit in systematic swathes. It also involves consideration of the thoroughness with which the surface is inspected. In sub-surface testing, however, bias assessment is more complex. Here discovery probability is the product of two independent but related probabilities (Krakker et al. 1983; Nance and Ball 1986): (1) the probability of intersecting a site with one or more shovel-probes (site intersection); and (2) the probability of encountering cultural material in the shovel-probes (test-pit productivity). These in turn are affected by properties of the sampling design (test-pit geometry and inspection method), and by properties of the archaeological record (site configuration, artifact density, and density-distribution) (Krakker et al. 1983; Nance and Ball 1986). Of these properties, the most crucial for maintaining spatial control over bias assessment is test-pit geometry (that is, testpit interval and pattern). For example, Krakker et al. (1983) have explored the way in which test-pit size, spacing, and layout affect discovery probabilities for sites of a given size and artifact density. Obviously, test-pit spacing is the most fundamental variable affecting the minimal site size that can be detected. The smaller the test-pit interval, the more likely it is that sites larger than that diameter will be detected by sub-surface testing, with smaller sites missed. They also demonstrate that substantial gains in effectiveness can be made through the use of optimal or staggered spacing of test-pits in a quadrat, rather than even spacing in
31
32
JAMES A. ZEIDLER
a. 5 Probes/1 ha. Quadrat
b. 8 Probes/1 ha. Quadrat
Figure 1.3. Two STP sampling designs for a 1 ha quadrat: (a) five shovel-probes per quadrat with staggered layout and specified spacing; and (b) eight shovel-probes per quadrat with hexagonal layout and optimal spacing. Probe size not depicted to scale.
Archaeological survey and site discovery a square-grid pattern. This can be illustrated by data from the Jama Valley survey. It was mentioned previously that when dense vegetation completely covered a 1 ha quadrat, a maximum of five shovel-probes was executed in an "X" pattern (one probe in each of the four corners and one in the center of the quadrat; Figure 1.3(a)). In this case, the layout is staggered and the spacing is specified as 99.0 m along a side and c. 71.0 m along the diagonal axes of the quadrat. Thus, the maximum diameter of an untested site in this scheme would be 99.0 m and the site discovery procedure is considered biased against the detection of small sites. By increasing the number of shovel-probes to eight, and arranging them in an hexagonal layout with optimal spacing (Figure i.3(b)), significant gains are made with a modest amount of extra field effort. Spacing between probes is reduced to 43.30 m and the maximum diameter of an untested site is reduced to 50.0 m; almost half that of the five-probe scheme. The eight-probe scheme was not employed in the Jama Valley study due to logistical difficulties and constraints on time and labor, but the resulting bias against the detection of sites smaller than 100 m are at least specifiable. Apart from the site size variable, the other critical archaeological properties for assessing the effectiveness of sub-surface testing are artifact density and densitydistribution. As discussed previously, sites having high densities of surface artifacts are much more likely to be detected than those having low densities. The spatial dispersion of artifacts can also affect discovery probabilities (Kintigh 1988a; Nance and Ball 1986). In order to assess the variable effects of these archaeological properties on discovery probabilities for different sampling designs, Kintigh (1988a, 1988b) has recommended a Monte Carlo simulation approach. This approach examines the interactions of the multiple probabilistic factors "by operationalizing the random processes with a computer simulation. The average interaction of the random processes can be 'observed' in repeated runs of the simulation" (Kintigh 1988a: 689). Using Kintigh's (1988b) STP Program and the two examples of test-pit geometry discussed above, we can simulate hypothetical parameters of the archaeological record, and make useful comparisons regarding discovery probabilities under the two sampling schemes. These parameters include: (1) site size; (2) artifact density (number of artifacts per m2); and (3) the function shape of the artifact density (that is, uniform, hemispherical, conical, sinusoidal, and negative binomial with different degrees of clustering). The number of repeated simulation trials can also be specified, and depends on the level of accuracy and reliability desired in the simulation results. As an example, Tables 1.6 and 1.7 illustrate sample output files from the STP Program pertaining respectively to the fiveprobe and eight-probe sampling schemes discussed above for a 1 ha quadrat. Several archaeological pararheters are included in this exercise, and in each case 1,000 simulation trials were run in order to calculate the intersection and detection probabilities (Tables 1.6 and 1.7, column 6). Five site sizes (measured in
33
34
JAMES A. ZEIDLER
Table 1.6. Sample output file (abridged) from Kintigh's (1988b) sub-surface testing evaluation program (STP), showing simulation results from 1,000 trials using five-probes/ha sampling scheme. Each row represents a Monte Carlo evaluation for a specified combination of site size, artifact density, and artifact density-distribution. Column headingsa are as follows: 1. File number; 2. Site diameter; 3. Artifact density-function; 4. Artifact density mean; 5. Artifact density k value; 6. Number of sites; 7. Number of sites intersected; 8. Percentage of sites intersected; 9. Intersected site hits; 10. Number of sites detected; 11. Percentage of detected sites; 12. Number of detected site hits. I
2
I
IO
I
IO
I
IO
I
IO
I
10
I
IO
I
IO
I
IO
I
IO
I
IO
I
3O
I I
3O 3O
I
3O
I
3O
I I
3O 3O
I
3O
I
3O 3O
I I I
50 5O
I
5O
I
5O
I
5O 5O 5O 5O
I I I I I I
5O 50
I
7O 7O
I
7O
I
7O
3
S S
s s s
N N N N N S
s s s s
N N N N N S
s s s s
N N N N N S
s s s
4
5
0.1
0
0.5
0
1.0
0
5.0
0
10.0
0
O.I
I
0.5
I
1.0
I
5.0
I
10.0
I
O.I 0.5
0 0
1.0
0
5.0
0
10.0
0
O.I 0.5
I I I I I
1.0
5.0
10.0 O.I 0.5
0 0
1.0
0
5.0
0
10.0 O.I 0.5
0 I I
1.0
I
5.0
I
10.0 O.I 0.5
0
1.0
0
5.0
0
I 0
6
7
8
1,000 1,000
14
!-4
J
12
1.2
1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
17 17
i-7 i-7
17 17
5 4
20
2.0
20
22
2.2
22
1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
9
10
11
12
4
0
0.0
0
12
1
O.I 0.5 0.4
1
5 4
10
1.0
10
0
0.0
0
O.I
1 2
19 19
i-9 i-9
19
1
19
2
0.2
12
1.2
12
4
O.4
4
16
1.6 14.7
16
10
1.0
10
147 154
3 9
0.3 0.9
3 9
151
22
2.2
22
147 137 135 144 163
63
6.3
^3
82
8.2
82
2
0.2
2
7
0.7 2.3 6.2
7
*3
9-3
93
0.9 2.6
26
J
47
154 151
147 137
135 144 163
15.4 15.1 14.7 13-7 13-5 14.4 16.3 16.1 16.9 41.0
161
2-3 62
169 410
93 9 26
43.0
397 416 409 430
412
41.2
412
398 370 418
39.8
398
161 169 410
397 416 409 430
387
39-7 41.6 40.9
238
5 28
37.0
370
50
41.8
418
38.7
387 807 786
184 248
806
80.6
784 794
78.4
796
56 *75
79-4 79.6
794 796
5.6 *7-5 23.8 0.5 2.8 5.0 18.4 24.8
62
9 56 175 238
5 28 50
184 248
11
1.1
11
55
5-5
55
121
12.1
121
335
33-5
335
Archaeological survey and site discovery
35
Table 1.6. (cont.) I
2
I
70 70 70
I I I I
70 70 70
I
100
I
100
I
100
I
100
I
100
I
100
I
100
I
100
I
100
I
100
I
3
4
S N N N N N S S
10.0
0
O.I
I
0.5
I
s s s
N N N N N
5
1.0
I
5.0 10.0
I
O.I
0
0.5
0
1.0
0
5.0 10.0
0
O.I
I
0.5
I
1.0
I
5.0 10.0
I
I
0
I
6 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
7
8
814
81.4 78.4
784 795
79-5
9
10
11
12
815
448
44.8
448
784 796
13 61
i-3 6.1
13 61 in
810
81.0
811
in
11.1
798
79.8
353
763
76.3
798 764
450
1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
100
1,621
100
1,587 i,599 i,575 1,617 1,580
29 92
35-3 45.0 2.9 9.2
216
21.6
216
662
66.2
695
820
82.0
905
17
i-7 11.7 19.0 56.8 76.1
100 100 100 100 100 100
1,615 !,593 1,621
100
1,589
100
117 190
568 761
353
450 29 92
17 119 201
696
978
Note: a
See text for explanations
Table 1.7. Sample output fde (abridged) from Kintigh's (1988b) sub-surface testing evaluation program (STP), showing simulation results from 1,000 trials using eight-probes/ha sampling scheme. Each row represents a Monte Carlo evaluation for a specified combination of site size, artifact density, and artifact density-distribution. Column headingsa are as follows: 1. File number; 2. Site diameter; 3. Artifact density-function; 4. Artifact density mean; 5. Artifact density k value; 6. Number of sites; 7. Number of sites intersected; 8. Percentage of sites intersected; 9. Intersected site hits; 10. Number of sites detected; 11. Percentage of detected sites; 12. Number of detected site hits. I
2
3
4
I
IO IO
0.5
I
IO
I
IO
I
IO
I
IO
I
IO
I
IO
I
IO
I
IO
I
30
S S S S S N N N N N S
O.I
I
6
7
8
9
0
1,000
62
6.2
0
1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
55
5-5
67 62
5
1.0
0
5.0 10.0
0
O.I
1
0.5
1
1.0
1
5.0 10.0
1
O.I
0
0
1
1,000 1,000
12
10
11
62
1
O.I
1
55
6.7
67
5 4
0.5 0.4
5 4
6.2
62
2-5
6.9
69
43
^•5 4-3
15
69
74
7-4
74
2
0.2
2
71 81
7-i 8.1
71 81
6 6
0.6 0.6
6 6
74 63 479
7-4 6.3
74 63 479
41
31
3-i 4.1 0.6
4i
47-9
6
43
3i
6
JAMES A. ZEIDLER
Table 1.7. (cont.) I
2
3
4
5
6
I
30
0
30
1.0
0
I
30 30 30
S S S S N N N N N S S
0.5
I
5.0 10.0 O.I 0.5
0
1.0
I
5.0
I
1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
I I I I I I I I I I I I I I I I I
30 30 30 30 50 50 50 50 50 50 50 50 50 50
I
70 70 70 70 70
I
70
I
70 70
I I I
I I I
70 70
I
100
I
100
I
100
I
100
I
100
I
100
I
100
I
100
I
100
I
100
s s s
N N N N N S
s s s s
N N N N N S
s s s s
N N N N N
0 I I
10.0
I
O.I 0.5
0 0
1.0
0
5.0 10.0 O.I 0.5
0
1.0
5.0 10.0 O.I 0.5
0
I I I I I 0 0
1.0
0
5.0 10.0
0
O.I 0.5
I
0 I
1.0
I
5.0 10.0
I
O.I 0.5
0
I 0
1.0
0
5.0 10.0
0
O.I
I
0.5
I I I I
1.0
5.0 10.0
0
Note: a See text for explanations.
7
8
9
10
11
12
504
50.4
504
496
49-6
496
34 73
3-4 7-3
34 73
505 519
50.5
505 5i9
229 320
22.9
5i-9
32.0
229 320
497
49-7
497
8 *9
0.8 2.9
29
470
47.0
470
52.1
52.1
5*i
480
48.0
480
496 976 911
49-6 97.6
1,129
91-1
982
98.2
969
9^-9
982
985 983 975 919 984
98.2
98.5 98.3 97-5 91-9 98.4
496 !.i43 1,138 1,132 1,142 1,121 1,148 1,116 1,136 1,172
71 183 303 18
95
188
8
7-i
7i 183 303 18
18.3 30.3
1.8
95
9-5
188
18.8
538
53.8
538
706
70.6
707
13
16
13
i-3
7.6
19
134
13.4
136
454
45-4 62.2
486 683
622
1,985 1,969
161
2,014 1,981
296 808
29.6 80.8
163 306 923
100
1.955
96.7
1.2-54
1,000 1,000 1,000 1,000 1,000
100
1,920
967 36
3.6
36
100
138
13.8
147
100
1,976 1,966
252
25.2
2,005
649
2-77
100 100
2,001
802
64.9 80.2
1,200
1,000 1,000 1,000
100
3.555
100
3.531
5i 271
100
1,000 1,000 1,000 1,000 1,000 1,000 1,000
100
3.577 3.544 3.576 3.583 3.589
1,000 1,000 1,000 1,000 1,000
100 100 100 100
100 100 100
*7
458 95i
996 5O 228
100
3.58i
100
3,620
403 861
100
3.565
95*
2-7 16.1
2.7
877
5-i 27.1 45.8 95-* 99.6 5.0
5i
285 532I
.7 I 5
2,183
51
22.8
M5
40.3 86.1 95.2
1.585
505 2,169
Archaeological survey and site discovery
37
diameter) were explored: 10 m, 30 m, 50 m, 70 m, and 100 m (column 2). For each of these site sizes, ten simulations were run for each combination of artifact density and density-function. Thus for each site size, two density-functions were tested: sinusoidal and negative binomial with clustering parameter k = 1.0 (column 3). Then, for each of these density-functions, five different artifact densities were tested: 0.1, 0.5, 1.0, 5.0, and 10.0 artifacts per m2 (column 4). The resulting output files produced 50 different Monte Carlo simulations, each consisting of 1,000 trials. The results for the intersection probabilities are given in columns 7, 8, and 9, while those for the detection probabilities are given in columns 10, 11, and 12. The two output files show marked differences in discovery probabilities as a result of the three-probe difference between the two sampling schemes. Within each output file, however, general results are comparable. Within each site size category (column 2), substantial gains are made in intersection and detection probabilities as artifact density increases (column 4). In all cases, sinusoidal density-functions resulted in slightly elevated probabilities, when compared to those of the negative binomial function. This is due to the fact that sinusoidal distributions are considerably more clustered around a centroid, while negative binomial distributions are likely to have more open space between groups of smaller clusters (see Kintigh 1988a: 693), and thus would escape detection in more cases. As is intuitively obvious, the larger the site size, the greater the probability that it will be detected. Note, however, that in the eight-probe scheme (Table 1.7), even a site measuring 100 m in diameter has only a 5 percent chance of being detected if the associated artifact density is 0.1 artifacts per m2 (Table 1.7, column 11), regardless of the density-function. Thus, even for relatively large site sizes, light artifact scatters do not stand a very good chance of detection in the eightprobe sampling scheme, and are even worse (3 percent) for the five-probe scheme employed in the upland sub-strata of the Jama Valley (Table 1.6, column 11). These variable relationships can be better comprehended graphically in Figure 1.4, using data extracted from the two simulation output files. The two bar graphs show changes in site detection probabilities (percentages) as site size increases and artifact density changes from 1.0 artifact/m 2 to 10.0 artifacts/m2. In both cases, the negative binomial density function was used with k = 1.0. In the five-probe scheme (Figure i.4(a)), discovery probability is relatively high (76 percent) only for the largest site size (100 m) and the highest artifact density (10.0 artifacts/m2). At smaller site densities, detection probabilities drop sharply even for high artifact densities. At low artifact densities, the five-probe scheme is not very successful (less than 20 percent) at detecting sites as large as 100 m, and drops steadily for smaller site sizes. As expected, the eight-probe scheme produces better results, especially for sites with high artifact densities. Low density sites do not show much gain over the five-probe scheme, however. The implications of these simulation results are inescapable. If the goal of a given survey is to find evidence for small sites and/or sites with low artifact
a) 100
0
b) 100
5 probes/1 ha. quadrat (1000 trials) • density=1.0 artifacts/sq.m. 01 density=10.0 artifacts/sq.m. negative binomial dist. (k=1.0)
30 50 70 Site S i z e (diameter)
100m
8 probes/1 ha. quadrat (1000 trials) • density=1.0 artifacts/sq.m. EE density=10.0 artifacts/sq.m. negative binomial dist. (k=1.0)
30 50 70 Site Size (diameter)
100m
Figure 1.4. Site discovery probabilities in 1,000 simulation trials for two STP sampling designs: (a) five probes per 1 ha quadrat; and (b) eight probes per 1 ha quadrat. Site diameter varies from 10 to 100 m. Artifact density varies from 1.0 to 10.0 artifacts/m2 for a negative binominal density function (k= 1.0).
Archaeological survey and site discovery
39
densities regardless of their size, then shovel-probe spacing and layout are critical variables of sampling design that must be addressed in a statistically informed manner. Survey intensity must be adjusted accordingly to avoid sampling bias against the discovery of small ephemeral sites. In the forested neotropical lowlands, where dense vegetation and low surface visibility is the norm, this problem is especially acute. It is important, then, that the neotropical archaeologist carefully weigh temporal and financial constraints against the need to avoid bias in archaeological sampling designs. Where biases against the discovery of certain kinds of archaeological remains cannot be avoided, they should be made explicit so that survey results can be objectively evaluated. Monte Carlo simulation provides a useful method for experimenting with hypothetical outcomes of sampling design geometry and various properties of the archaeological record. Various options exist for the application of such exercises. As Kintigh notes, "(t)he method can either be used ex post facto to evaluate the results that are achieved by a testing program or, perhaps even more usefully, it can be used to examine 'what if scenarios in order to better plan a testing program" (Kintigh 1988a: 689). In either case, the method permits objective assessment of survey effectiveness (that is, discovery probability) as well as sampling bias. As such, it should become an integral part of the neotropical archaeologist's methodological repertoire in the planning and execution of regional archaeological surveys and intra-site testing programs. Conclusion Archaeological survey and site discovery in neotropical lowland landscapes pose special logistical problems for those interested in recovering representative samples of sites and material remains from broad regional areas. Once the fundamental decision has been made to conduct a regional archaeological survey under these environmental conditions, a host of "methodologically unlovely techniques" (Schiffer 1987: 350) present themselves for careful scrutiny. In the end, significant compromises must often be made between desired survey intensity and the limited resources available to carry out the research. Meggers and Evans (1957: 6) commented on this same point some years ago in the following manner: Archeology in the tropical forest of South America presents, in addition to the usual problems, many difficulties that are not encountered in the more arid or more accessible parts of the New World. Manuals of field procedure and precision methods of excavation technique frequently cannot be followed, and thefieldsituation must be met with an understanding of what is pertinent and what is unprofitable in order to gain the maximum of information in the shortest possible time. This statement is as true today as it was over thirty-five years ago. Today, however, collective wisdom tells us that effective and reliable sampling methods
40
JAMES A. ZEIDLER
exist for conducting regional surveys in the forested neotropical lowlands, and that specific procedures exist for objectively evaluating sampling bias in these situations. It is only through greater attention to methodological rigor that significant advances will be made in lowland South American archaeology. As regional sampling designs and systematic sub-surface testing programs become the norm, a clearer picture of regional archaeological distributions will certainly emerge and new interpretations of those distributions will be possible. In the words of one neotropical specialist, "more cultural surprises await beneath the forest mask" (Smith 1980: 566). Notes The ideas expressed in the foregoing article have had a long period of gestation during which numerous individuals have contributed, often unwittingly, to their present form. First, I must express a long-standing debt to the late Donald Lathrap whose writings and lectures on the prehistory of the neotropical lowlands provided much of the impetus for the present discussion. Although he would very likely take exception to aspects of the approach advocated here for regional survey and site discovery, he was resolute in his belief that past archaeological practices for site survey in the neotropical lowlands are woefully inadequate for the level of knowledge we seek. Secondly, the Achuar peoples of the Makuma and Huasaga drainages of Amazonian Ecuador provided me with first-hand knowledge of site formation processes in a tropical lowland environment, and forced me to ponder the problem of effective site discovery in a forested terra firme zone. Jorge Yambik was especially helpful in pointing out the nuances of abandoned Achuar settlements, as well as archaeological site locations and surface manifestations in areas of dense tropical vegetation cover. Archaeological fieldwork in the Jama Valley of western lowland Ecuador was generously funded by grants from the National Science Foundation (BNS-8709649, BNS-8908703 and BNS-9108548) awarded jointly to the author and to Deborah Pearsall. This work was undertaken under the kind auspices of the Instituto Nacional del Patrimonio Cultural (Guayaquil Office). The continued support of these institutions is gratefully acknowledged. The success of that fieldwork between 1988 and 1991 was largely due to the diligent efforts of project field personnel from United States and Ecuadorian institutions, and from the parroquias of San Isidro and Jama, Manabi Province, Ecuador. Finally, I am grateful to John Isaacson for productive discussions on survey intensity, variable landscape conditions, and appropriate site discovery procedures, and to Michael J. Shott and Peter Stahl for their critical editorial reading of the text. Marie Zeidler kindly prepared the figures. I alone remain responsible for any shortcomings. 1 This bi-partite site typology is undoubtedly an oversimplification. Further subdivision of these two categories into several discrete variants is probably warranted for many areas, especially where complex chiefdoms have been identified (for example, Marajo Island in Brazil, the Sangay site in the Upano Valley of eastern Ecuador, and the Llanos de Mojos in eastern Bolivia [see Erickson, this volume]). Indeed, in a subsequent article, Myers (1973) effectively demonstrates the wide variability which exists for size and complexity of habitation sites throughout the Amazon Basin. A third general category can be added to Myers' list of basic site types: cemeteries. Very often these are special purpose burial mounds or isolated interments of urn burials not located in close proximity to habitation sites (Meggers 1991). 2 The following passage is a particularly telling account of informant-based site discovery techniques in the neotropical lowlands of British Guiana:
Archaeological survey and site discovery
41
Locating sites in this area would have been difficult and slow without a guide. The river banks are densely forested, and vegetation conceals the ground so effectively that hills are often not visible from the river. An interview shortly after our arrival with the Wai Wai chief, the oldest man in the tribe, produced a long list of places that he said were "old villages". On questioning, he was firm in his identification, although he had never seen potsherds at any of them. We were consequently somewhat dubious as we proceeded to the first such spot. Testing revealed sherds, however, and we found this to be true of all the places listed by the chief with rare exceptions . . . Tests on high spots above the flood level not mentioned by the chief always proved sterile (Evans and Meggers i960: 6). 3 To quote Read (1986: 488) on this important point: . . . the relative efficiency of small quadrats in comparison to large quadrats (holding other factors constant) varies inversely with the spatial distribution of sites as that distribution varies from uniform to clustered. In other words, small quadrats (defined as quadrats for which there will be either one or zero sites per quadrat) are generally more efficient in the sense of yielding more precise estimates of the number of sites in the region than are large quadrats (defined as quadrats for which there will either be zero or several sites) when sites are spatially clustered.
The archaeology of community organization in the tropical lowlands: a case study from Puerto Rico PETER E. SIEGEL
The systematic analysis of prehistoric community organization has been slow to develop in the lowlands of South America and the West Indies. This is a product of two factors: (i) a predominant concern among archaeologists with the origins, spread, and complexity of the populations that inhabited the region; and (2) the logistical constraints on doing archaeology in the tropics. In recent years, archaeologists working in the lowlands have expanded their range of investigations to include ancient communities as legitimate foci of concern. In doing so, serious methodological challenges must be confronted in delineating site sizes, settlement layouts, and occupational histories. It is not acceptable to blindly import methods or techniques developed for other world-areas; we must experiment with and design approaches that are uniquely appropriate for getting at community organization in the tropical lowlands. In this paper, I review the problems, constraints, and challenges of studying community organization in the neotropics. A case study is offered using the Maisabel site, a large Saladoid/ Ostionoid village on the north coast of Puerto Rico. Problems and constraints
The major factor hindering community-oriented archaeology in the tropics can probably be narrowed down to one word: visibility. Tree growth and ground cover vegetation mitigate the effectiveness of surface surveys in the tropics (see Zeidler, this volume). In this regard, archaeologists working in the tropics share some of the problems and challenges faced by their colleagues working in temperate-forest settings (for example, Lovis 1976; Wobst 1983). Site discovery and boundary definition techniques are major methodological issues that must be confronted in systematic fashion if we are to have confidence in our discussions of community organization. In the following case study I present a strategy for delineating site size and internal spatial structure in a tropical setting. This example is meant to serve as a point of departure for other similar studies in the tropics. 42
The archaeology of community organization
43
The Maisabel site, Puerto Rico Horticulturally adapted groups of Indians entered the West Indies from northeastern South America roughly 2,500 years ago (Rouse 1989,1992; Siegel 1991b). These groups had a complex ceramic tradition, which is remarkably consistent from Venezuela to the eastern edge of the Dominican Republic (Rouse 1982; Rouse and Allaire 1978; Rouse and Cruxent 1963). Within this tradition, or more correctly, series, local variations provide the basis for defining ceramic complexes or styles (Rouse 1982). The tradition as a whole is termed the Saladoid series of cultures, named after Saladero, the site where the complex was initially identified (lower Orinoco River, Venezuela) (Rouse and Cruxent 1963: 112-125). The motivating circumstances for expansion into the West Indies by Saladoid groups is not well understood, but may be a product of interethnic feuding and competition for prime alluvial land in the lowlands of South America (Carneiro 1961; Lathrap 1970; Siegel 1991a). Socially and politically, it has been assumed that these groups were tribally based with trade and social networks extending from Puerto Rico to at least Venezuela and perhaps into Brazil (Boomert 1987; Cody 1993; Vescelius and Robinson 1979). Other than detailed descriptions of ceramic styles, however, discussions of Saladoid social systems have been based on little to no physical evidence. Loose analogies to extant Amerindian groups in Amazonia are frequently made. At this stage in West Indian archaeology it is crucial to provide linkages between our ideas of Saladoid social and community organization and a solid empirical base. As such, many of our assumptions should be treated as testable hypotheses. The overall framework for the excavation of the Maisabel site was designed to address explicitly the nature of Saladoid community organization (Siegel 1988). Maisabel is located on the north-central coast of Puerto Rico, 30 km from San Juan (Figure 2.1). This portion of the coast is characterized by a low alluvial plain dotted with lagoons and swamps (Kaye 1959:53-54) • The coastal plain is roughly 3 to 5 km wide, from the shore to the foothills of the Cordillera Central (Beinroth 1969). The shoreline alternates between sandy beaches, cemented dunes, and Pleistocene reef rock (Kaye 1959: 54). The portion of the shore line directly fronting the Maisabel site consists of Pleistocene reef rock. The climate of the north coast is tropical marine. Temperatures range from 23 to 27°C with 60 to 70 inches (1,524-1,778 mm) of annual precipitation (Fassig 1909, 1911; TorresGonzalez and Diaz 1984: 12, Figure 2.4-1). Maisabel was discovered in the mid-1970s by pothunters, referred to locally as saqueadores (see also Ubelaker, this volume). As a result of their activities it became apparent that a Saladoid site with several "middens" (depositos) existed in this location. In years subsequent to its discovery, Maisabel was sporadically visited by pothunters and both amateur and professional archaeologists. All of these individuals limited their work to the thickest midden deposits, which form distinct topographic eminences.
66°45'
0
C
E
A
N
A
0
T
L
A
N
T
I
C
O
KILOMETERS
M
A
R 66°45'
C
A
R
I
B E
66°I5'
Figure 2.1. Map of Puerto Rico showing the locations of the known early Saladoid sites. The perimeter of the island is defined by a flat coastal plain and the interior is mountainous.
The archaeology of community organization
45
Atlantic Ocean
i
^ _ > _ v P t a . Puerto Nuevo
>
— -
/
S
—~^-^y
"a
Maisabel ^Site
Pta.
/
N\
»
GarzaH
Boca
•
—
^7—••;
^V_del Cibuco/^; 'SLaguna de
(.
NU#vo
^;: :r"° v
<'••••'$
•'•'••
f~
,
'
~
,-'-'
/ \
'^'''^,-15->,"";
/'~X'O
N
,'
'
/
i* " f
* *
Jk,
ilL
•Ml
L
-Jk
*/
/
' - - ^
M Jit
(
^
'
/
0 **
*
A
*
0
1 E
*
I Kilometres
Figure 2.2. Context of the Maisabel site. In 1985, the Centro de Investigaciones Indigenas de Puerto Rico initiated a major excavation of Maisabel under my direction. My research interests on the site centered around internal spatial structure and community organization. Prior to the Maisabel project, no early ceramic period site in the West Indies had been systematically tested in the horizontal dimension. For chronology-building studies, investigators generally searched for the thickest midden deposits in which to place trenches. Given this lack of background information, it was difficult to establish a set of expectations for the structure and organization of a Saladoid community. In order to obtain suitable information for addressing site structure and community organization, a three-staged approach was followed in the excavation. Stage one consisted of a survey, in which the areal extent of the site and a coarse-grained view of its internal structure was obtained (Siegel 1986, 1989). In deciding upon a sampling method a number of considerations were evaluated. First, the relatively flat to undulating terrain in the area does not represent a natural obstacle in constraining the size of a prehistoric village. The only natural boundaries for settlement expansion were the Atlantic Ocean to the north, and the Cibuco River roughly 1 km to the southeast (Figure 2.2). The foothills of the Cordillera Central are 4 km south of the site. Therefore, in the horizontal dimension I needed to consider a very large area for testing.
46
PETER E. SIEGEL
Second, for the vertical dimension more background information was available. As mentioned above, most studies of Saladoid sites have systematically investigated the thickest midden deposits. Therefore, these studies were used as a guide to determine the approximate vertical extent which I needed to consider. Hacienda Grande, located on the northeast coast of Puerto Rico, is one of the most intensively sampled Saladoid sites (Rouse and Alegria 1990). Based upon numerous excavations, the culturally sterile soil is reached at an average depth of n o to 150 cm below ground surface (Alegria 1965; Bullen and Bullen 1974; Rouse and Alegria 1990: 29-33). Chanlatte Baik (1979), excavating at the large Saladoid site of Sorce, on the island of Vieques, encountered some midden deposits in excess of 3 m in depth (Chanlatte Baik, personal communication 1990). Our requirements therefore were to test a large areal extent to a depth of roughly 2 to 3 m.1 Vegetational coverage mitigated the effectiveness of surface survey. Surficial examination in a few cow paths which traverse much of the site, yielded small pieces of pottery and shell. Further, by walking along the beach, potsherds were seen in the roots of palm trees and eroding out of the sand embankment. This information yielded only a cursory view of the site size. Two alternative sampling methods were considered in order to define the site size and to yield information on internal structure. Most of the literature dealing with intrasite sampling problems have dealt with shallow or surface sites, where it is relatively easy to conduct controlled surface collections, or to remove entire sections of overburden revealing underlying features (Binford 1964; Binford et al. 1970; Reid et al. 1975; Shott 1987). In these situations archaeologists have often employed probabilistic sampling designs. Before presenting the sampling design, it is useful to review the project goals and what we want to learn about the site. Settlement organization and community patterning are the primary interests. Specifically, how did the prehistoric occupants of Maisabel structure their community? Given the multi-occupational aspects of the site, I am interested in ascertaining whether or not there is continuity in the structure of the ancient settlement through time. It is important to define how settlement space is partitioned, the nature of the activities in the various partitions, and how these activities relate to the organization of the community. Given these questions and what little background information is available for ceramic-age settlement organization in the Caribbean, it was deemed most productive to structure the sampling within a multi-stage or nested framework (Binford 1964; Redman 1973, 1975: 148-152; Redman and Watson 1970). A logistical complication, affecting the manner in which the site could be sampled, is the depth of the midden. As discussed earlier, Saladoid middens generally may be up to 2 m thick. Although this is not as deep as sites like Koster, it does represent a problem in deep-site sampling: "Unlike the surface or shallow site that can be sampled using a repertoire of sampling strategies, the deeply buried occupation imposes extremely difficult problems in the effective use of sampling controls" (Brown 1975: 160).
The archaeology of community organization
47
The two major logistical problems in sampling the Maisabel deposits were the heavy ground cover vegetation and midden depth. To rectify the visibility problem I considered plowing selected portions of the site and conducting controlled surface pick-ups in these areas. There are two problems with this strategy. First, the limits of the site were unknown, thus it would be difficult to decide where or where not to plow. Second, since the depth of the midden could be relatively great, plowing would only enable me to sample the upper portions of the midden. In recent years, archaeologists have been experimenting with remote sensing techniques for site discovery and testing (Carr 1982; Roosevelt 1991). Roosevelt (1991: 186-229) has had some success in identifying buried archaeological features using geophysical prospecting methods at the Teso do Bichos site in Brazil. Geophysical testing, however, is not sensitive to variable densities of artifacts. Assemblage composition is a major class of information that must be discussed when evaluating prehistoric community organization. An alternative testing technique is to place a number of sub-surface soundings across the area of concern, monitoring the presence/absence of artifacts. In this way site boundaries, resolvable to the interval used between soundings, may be established. Furthermore, the frequency of artifacts recovered in each pit may provide a coarse-grained view of the internal structure of the site. Finally, these test pits sample the vertical, as well as the horizontal dimensions of the midden. For these reasons, I decided on a sub-surface testing program to delimit the boundaries of the site and to obtain a view of its internal structure. A Pitman Polecat auger mounted on the back of an ex-firetruck was used to make the soundings. Each auger hole is 55 cm in diameter and uniformly dug to 2 m. The major disadvantage of using this system is an unfortunate loss of vertical control, as stratified deposits are collapsed to produce plowzone quality data. The one advantage to excavating these test pits by machine instead of by hand is that in several sections of the site the deposits were quite thick. In these areas shovel test pits would have been somewhat more difficult to excavate than was the case with the auger. Using the information obtained in the stage one survey, I stratified the site into three zones for further intensive excavation. A set of 2 x 2 m excavation units was placed systematically across each stratum. It was decided not to distribute these units randomly, because the primary interest was to obtain an idea of the kinds of artifacts and features located spatially across the strata. The rationale used in the judgmental placing of the excavation units varies with the stratum. Finally, the third stage of excavation consisted of a follow-up on the information obtained during Stage 2. In certain excavation units particularly useful data were retrieved for the stated project goals. In these instances, additional squares were excavated to follow up on the units. My approach to the Maisabel sampling problem is nonprobabilistic for many of the reasons described in excavating deep sites (Brown 1975). Different occupations in a multicomponent site are likely to have different sampling
48
PETER E. SIEGEL
parameters. In a general sense the strategy outlined by Brown was followed in the Maisabel excavation: (1) the collecting of information relevant to the number, depth and extent of the subsurface archaeological zones within the site limits; (2) the creation of a firstorder sample stratification of the site sample space; (3) the excavation of the set of sample excavation units; (4) the classification of sample units for each layer to recover activity categories held in common among the set of layers; and (5) the expansion of excavation as a result of creating a second-order sample to improve on the representation of activity types in each layer (Brown 1975: 169). I will now present the results of the stage one survey, discussing the site size and general site structure.
Stage 1 survey Stage 1 survey involved mechanical excavation of a series of auger test pits (ATP). Given the potentially enormous size of the site, a relatively coarse-grained approach was selected for the stage one survey. A 30 m interval between ATPs was deployed (Figure 2.3). This sampling interval provides a rough approximation of the site boundaries as well as a general idea of the artifact distributions within the site. A total of 256 points was tested in this way (except for a handful that were excavated manually). A considerable amount of material was recovered in this testing. Counts and weights were recorded for ten artifact categories: ceramic; unmodified shell; ground stone; chipped stone; unmodified crab parts; unmodified bone; beads; unmodified coral; unmodified rocks; and miscellaneous items. Of the ten categories, ceramic artifacts are the most abundant and continuously distributed class of materials. For these reasons, ceramic distribution is used in this discussion of initial test results. Using the ceramic weight distribution, an isopleth map based upon 500 g contour intervals was produced (Figure 2.4). Due to communition by the auger bit, ceramic weight rather than frequency reflects a more accurate distribution of the pottery. For instance, a group of 100 potsherds weighing 100 g is very different from a group of 5 potsherds weighing 100 g. The map displays several clear patterns. First, the site is roughly oblong in shape and measures approximately 500 x 400 m, with the longest dimension parallel to the coast line. Two mounded midden areas easily discernible in the terrain (Figure 2.5), show up clearly on the ceramic distribution map (Figure 2.4). The north mounded midden is round with a rough diameter of 70 m. The south mounded midden is oblong and approximately 100 m long by 30 m wide. Second, two large areas of low ceramic weight (less than 500 g per auger test pit) are located within the central portion of the site. Area 1 (east area) is roughly
The archaeology of community organization
49
Distribution of Auger Test Pits
»- Augtr ttst pits without artifacts.
Legal property limit. 4444^444 4 4
4 4
4 4 4 4
4 4 1 1 4 4 4 1 |4 4444--^-$-4. 4 4 4 4 4
0
50
100
meters
Figure 2.3. Distribution of the auger test pits.
site boundary
v /
POND
J ^ \
• ^//ffi)\
/ / /
/
\ / / site
boundary
^
v
.
/
° ^
^ >
V
Ceramic
Map for Weights
(grams)
H SOO- IOOO g
J
'
Isopleth
n<5°°fl
/ \
°
°
LEGEND
/
\POND J
o
^ IOOO-1500 g |
1500-2000 g
H 2000- 2500 g £3>25OOg 0
50 meters
Figure 2.4. Isopleth map of prehistoric pottery, by weight, recovered from the auger test pits.
100
PETER E. SIEGEL
Topographic contour lines at I.Omts interval. Topographic contour lints at 0.25mt« interval. Legal property limit. lote : All elevations are referred to mean sea level.
Figure 2.5. Topographic map of the Maisabel location. rectangular in shape and is about 90 x 70 m. Area 2 (west area) is amoebic to oval in form, and is approximately 110x90 m (Figure 2.4). Furthermore, the low ceramic weight area completely encircles the site. Finally, separating the mounds from the areas of low ceramic weight, is a 500 to 1,000 g contour interval. This zone roughly encircles the central portion of the site, in addition to crisscrossing it in narrow (c. 15 m wide) bands. The other artifact categories, for which contour plots were made, also conform to the overall circular pattern (Figure 2.6). Within this pattern, however, particular concentrations do not always correspond to the concentrations displayed in the ceramics map. Stratification scheme
Using the ceramic distribution map, the site was stratified into three zones for intensive excavation (Figure 2.7). Stratum 1 is represented by the low ceramic weight area (less than 500 g), and is 13.393 n a i n area. If Maisabel structurally corresponds in any way to most relatively sedentary, Amerindian villages in lowland South America, then this portion of the site should yield architectural and residential remains. The South Amerindian context is used as a rough ethnographic analog because of the conjectured similarity in lifeways and activity
The archaeology of community organization
House Area
ALL ARTIFACTS RECOVERED FROM AUGER TEST PITS
Mounded Midden 1
Midden 2
CERAMICS RECOVERED FROM AUGER TEST PITS Mounded Midden 1 Midden 4
House Area Cemetery Midden 5
Mounded Midden 2
Figure 2.6. Three-dimensional plots of artifact densities at Maisabel. The top map shows all artifact classes combined and the bottom map shows prehistoric pottery. The z axes represent weights in grams.
51
52-
PETER E. SIEGEL
Isopleth Map for Ceramic Weights (grams) LEGEND
•
< SOOg 500 g - 1500 g
ffl
X X
X X
Macr Hou ' . ^ / - i t c h Oth Ceme tery
X X
X X
Midd en 3 Midd en 4 Midd en 5
X X X
X X X
Mounded Midden 1
X X X X ?
X
? ?
Figure 2.7. Isopleth map of prehistoric pottery, by weight, recovered from the auger test pits. The major site sections discussed in the text are labelled. The insert chart gives the chronological affiliations of the site sections. patterns of these groups with the prehistoric tribal societies of the West Indies and lowland South America. Further, ethnographic descriptions and previous archaeological excavations among forest societies suggest that villages of sedentary groups are organized around and within a matrix of cleared plaza space (Crocker 1985: 31-33; DeBoer and Lathrap 1979; Gregor 1977: 48-60; Lathrap et al. 1977; Marcos 1978: 2; Myers 1973; Nimuendajii 1946: 37-38; Siegel and Roe 1986; Stahl 1984: 153-167; Zeidler 1984: 8-9, 54, 59-60). Stratum 2 is composed of the 500 to 1,500 g contour intervals. This is the zone separating the low ceramic weight areas (Stratum 1) from the mounded midden areas discussed earlier, and is 3.202 ha in area. The third stratum is defined by the ceramic weight contours greater than 1,500 g. These locations are represented essentially by mounded midden areas 1 and 2, yet there are small pockets of this stratum located in three other portions of the site (east center, north northeast, and north northwest). This is the smallest of the three sampling zones, at 0.633 n a - If w e examine the ceramic "hotspots" in total there is roughly a circular to horse-shoe pattern to their distribution. This is reminiscent of the Monserrate site pattern reported many years ago by Rainey (1940: 77). The topographic characteristics of Mounded Middens 1 and 2 do not conform exactly to the spatial patterns produced by plotting the weights of ceramic
The archaeology of community organization
53
artifacts recovered in the auger pits (Figure 2.5). This is partly a factor of the 30 m auger interval employed. If a smaller sampling interval had been used in the augering program, then the ceramic contour map would probably conform more closely with the topographic map. The three remaining ceramic "hotspots" do not show up as topographic eminences. I believe this is due primarily to the smaller volumes of each of these artifact deposits compared to Mounded Middens 1 and 2. Thus, there is a minimum amount of material that needs to be deposited in a circumscribed area before it has an effect on the actual topography of the location. If the amount of material deposited is not great enough to produce a visible "mound," it may still show up in an artifact distribution map, and have behavioral significance. Stages 2 and 3 results: intensive excavations The three sampling zones vary considerably in area (Table 2.1). The fraction of each stratum excavated is quite small, with the greatest proportional attention being paid to Stratum 3. Of the total 280 m2 excavated at Maisabel, 190 m2 are in Stratum 1, 40 m2 are in Stratum 2, and 50 m2 are in Stratum 3. Stratum 3 was sampled by placing units in the two largest mounded middens. An L-shaped trench, consisting of twelve 2 x 2 m squares, was excavated on the large round north mound. A small 2 x 4 m (two 2 m squares) trench was placed on the oblong-shaped south mound. These mounded midden units produced the highest density of elaborate artifacts found at the site. Further, the sizes of the ceramic sherds generally are much larger than sherds found off the mounds (except for ceramic vessels associated with burials). No structural features or burials were recovered, except for two isolated crania from Mounded Midden 1. Stratum 2 was sampled with a trench near Mound 2. The artifacts differ from mound fill in that the density of elaborate material is lower and the sherd sizes are smaller in Stratum 2. Excavation units measuring 2 x 2 m were placed systematically across Stratum 1. This zone yielded considerable information regarding site structure and settlement organization. In every 2 x 2 m square that was placed in the large eastern low ceramic weight area of Stratum 1, at least one human burial was encountered (Figure 2.4: Area 1). In many cases a burial extended into a wall of a unit, thus necessitating an extension, or the excavation of an adjacent unit. In those cases, additional burials were usually encountered. In total, 23 burials were recovered from this part of the site. Early during this stage of field work, a postmold was found in a 2 x 2 m square placed between Mounded Middens 1 and 2. In order to search for structural remains and related features, considerable attention was therefore paid to this portion of the site. Excavation in this area eventually resulted in thirty-two 2 x 2 m squares, producing what I have termed the "macroblock."
Table 2.1. Sampling strata areas and amounts excavated. Stratum excavated
Area (m2)
Area excavated (m2)
% Stratum area excavated
% Total excavated area
Volume excavated (m3)
% Total excavated volume
1 2
190 40 50
0.14 0.12 0.78
67.85 14.28 17.85
144.155
3
133,930 32,020 6,330
53-337 70.617
53-76 19.89 26.33
Total
172,280
280
0.16
99.98
268.109
99.98
The archaeology of community organization
55
Figure 2.8. Distribution of macroblock features displayed by type. A considerable number of features were found in the macroblock, consisting of postmolds, hearths, pits, trash areas, and burials (Figure 2.8). The most intriguing feature (F 101) located in the macroblock is a long (roughly 22 m), narrow (80 cm wide), curving deposit of debris containing all of the artifact classes found at Maisabel, with the exception of crab shell.2 Given that most of the postmolds, hearths, pits, and burials in this portion of the site are located within the arc of this feature, it appears that we have the remains of a large structure surrounded by a ring of refuse. In contrast to the mounds mentioned previously, this refuse clearly is deposited within a ditch (Figure 2.9). Again, drawing on analogies with lowland South Amerindian groups (Roe and Siegel 1982: Figs. 8 and 29; Siegel 1976, 1990: 405, Fig. 23), it appears that this feature is a drainage ditch surrounding a residential structure. In the absence of one of these ditches, high tropical precipitation will rapidly flood a house. The north coast of Puerto Rico is part of the humid tropics and the Amerindians here would have had to deal with the same problems of water control as their South American counterparts. Summary The most efficient procedure for sampling the Maisabel site, given the stated goals, is to employ a multi-stage nested testing program. The results of each stage feed directly into the implementation of the next stage. Stage 1, the survey, established the boundaries of the site and provided a
PETER E. SIEGEL
Figure 2.9. Portion of the ditch feature traversing one of the macroblock excavation units. The feature fill is recognizable by the dark colored soil and the high density of artifacts. coarse-grained view of its internal structure. This was accomplished by deploying a set of sub-surface borings every 30 m. Using the resulting artifact distributions the site was stratified for further intensive excavation. Three zones were identified based upon artifact densities: high, medium, and low. Based upon analogy to South Amerindian lowland village organization, these zones were isolated as the sampling strata. Results of the intensive excavations in Stages 2 and 3, indicate that the sampling zones represent distinct spheres of activity. Stratum 1, the low artifact density zone, produced a large number of structural features and burials. The evidence at this point reveals that Stratum 1 was partitioned into at least two distinct regions: a cemetery and a house area. Given the enormity of this stratum, it is likely that other portions also contain distinct areas. Stratum 3 is the high artifact density zone and corresponds to five discrete locations in the center of the site. Two of these locations are large enough to produce topographic eminences, referred to here as mounded middens. The greatest quantity of elaborate artifacts recovered at the site comes from units excavated in Stratum 3. Stratum 2 is an intermediate zone between Strata 1 and 3. The nature of the artifacts recovered from this zone suggests that it is closely related to the activities of Stratum 1, although additional excavation may be required to substantiate this idea. Exploring the social significance of the spatial patterns is beyond the scope of
The archaeology of community organization
57
this paper. I will discuss here the behavioral organization of the site, focusing on formation processes that resulted in the archaeological record and how the different site sections were used. Behavioral organization The Maisabel occupants partitioned the settlement into a number of discrete functional areas. In my field work I have defined several of these zones or areas. Given the large size of the site and the amount of excavation conducted, it is likely that additional zones of behavioral importance are present. The archaeological record is a product of numerous processes, many of which are unrelated to each other. Schiffer's (1972, 1976, 1987) distinction between cultural and natural formation processes in the archaeological record is now well known, and provides a useful framework for evaluating prehistoric deposits. Much of the discussion concerning assemblage composition is subsumed within the conceptual framework of cultural formation processes. I will now examine the defined site sections in terms of disposal processes and refuse types. Macroblock Features, artifacts, and burials dating to all of the major occupations of the site are present in the macroblock, resulting in a thin palimpsest assemblage. It appears, however, that the (or some of the) final occupants (Santa Elena people) inhabited a large longhouse and were primarily responsible for the disposition of the archaeological record as it appears today. Regular cleaning activities within the structure resulted in high densities of all artifact classes accumulating along the perimeter of the house. Upon abandonment, the building was used as a refuse zone. The conjectured entryway for the structure yielded the highest densities of artifacts in the macroblock, except for the ditch feature. This finding accords well with ethnographic observations of refuse disposal in abandoned structures. Figures 2.10 and 2.11 display artifact densities within the macroblock. The artifact-bearing deposit in the macroblock is 40 to 50 cm thick. The upper 20 cm is represented by the plowzone, leaving 20 to 30 cm of relatively undisturbed midden. There are no clear stratigraphic distinctions, culturally or naturally, within the macroblock midden. Plowzone material was deleted from the analysis of horizontal distributions of artifacts. To include it with the sub-plowzone artifacts would potentially blur patterns or create spurious ones due to the effects of plow-drag, especially at the scale of the macroblock units (Odell and Cowan 1987: 477-479; Roper 1976). No excavation unit was larger than 4 m2. Artifacts were collected by grid units within the macroblock. The grid unit data were converted into densities in order to facilitate spatial comparisons of artifact distributions. Examining raw frequencies of artifacts by excavation unit would
PETER E. SIEGEL
W24
W22
W20
W18
W16
W14
W12
W10
W8
W6
W4
W2
EO
E2
E4
Legend (g/m 3 ) Cluster 1.x = 5,178 Ditch Feature
^ |
Cluster 2, x = 66,492
^ ^
Cluster 3, x = 20,675
H H
Cluster 4, x-= 28,266
tX\N
Cluster 5, x= 11,695
U.71
Cluster6, x = 34,078
Figure 2.10. Density distribution of the macroblock artifacts as determined by kmeans cluster analysis. The six-cluster solution is displayed, not including the ditch feature artifacts. be misleading, since not all units were of uniform size. Furthermore, depths of units varied, depending on such circumstances as the presence of features or variable thickness in the midden across the macroblock. Grid unit densities were used as input data for fe-means cluster analysis (Kintigh and Ammerman 1982). This method of analyzing the spatial distribution of artifacts is a form of unconstrained clustering (Whallon 1984). Clusters are defined as sets of excavation units with similar densities of artifacts. The analysis may be performed either separately for each artifact class, or with all artifact classes combined, or both. Clusters are not necessarily comprised of spatially contiguous excavation units. This analysis, therefore, classifies units of space rather than artifacts (Whallon 1984: 248). Units of space (in my case excavation units) are characterized by such attributes as overall artifact density; individual artifact class densities; and cluster size, shape, and distribution. In this paper, I present the densities of all artifact classes combined (Figures 2.10 and 2.11). In a separate study I also present density distributions by artifact class (Siegel 1992: 327—361). Figure 2.10 displays the six-cluster solution of total artifact densities across the macroblock. Although the ditch feature is plotted on this map, the clusters are derived from artifacts recovered only from the midden and not the feature.
The archaeology of community organization
59
Legend (g/m 3 ) Cluster 1,x = 6,872 B f |
Cluster 2, x = 357,734
| ^ |
Cluster 3, x = 72,779
[7T1
Cluster 4, x = 181,850
H H
Cluster 5, x = 106,238
IVsXl
Cluster 6, x = 25,314
Figure 2.11. Density distribution of macroblock artifacts, including the contents of the ditch (six-cluster solution). Figure 2.11 shows the six-cluster map including the contents of the ditch feature. We see the contents of the ditch and the deposits around it exerting the greatest effect on these two distribution maps. The densest deposits are associated with the ditch. Elsewhere I suggested that much of this debris resulted from regular cleaning activities within the house (Siegel 1992). It is likely, however, that while the house was occupied, the ditch would periodically be cleared of garbage, otherwise its drainage function would be hindered. In planned gradual abandonments of places (houses, villages, regions, and so on), refuse tends "to accumulate in areas usually kept free of such debris as the need to redeposit it elsewhere [is reduced] . . . within enclosed living areas, refuse on sites which have been abandoned gradually, as opposed to rapidly, will be more abundant and distributed in a more clustered or orderly manner" (Stevenson 1982: 246-248). Furthermore, upon abandonment, the house is likely to become a refuse zone. An open ditch becomes a convenient place to deposit trash. Not all sections of the ditch were filled with refuse, however. In the southwest corner of the macroblock the ditch feature contained very little material. This area coincides with the hypothesized back of the house, which faces away from the main portion of the village. It is conceivable, therefore, that only those sections of the ditch facing the village center were filled in with refuse, with the more remote ditch sections left open and allowed to fill in naturally.
60
PETER E. SIEGEL
Mounded middens Five dense concentrations of artifacts were defined during the stage one survey (Figure 2.7). Two of these form distinct topographic features in the landscape (Figure 2.5), referred to as mounded middens. These mounded middens are complex features that defy simple interpretations. They represent dense accumulations of the full range of artifact classes found at Maisabel. When examining the mounded middens in terms of disposal processes and activity organization, a different set of constraints enters the discussion, compared to the macroblock. In the house area, the primary concern was to remove unwanted debris from the central portions of the house to the periphery. The limited excavation evidence from the areas immediately outside of the structure also suggest that refuse removal from these areas was a major concern. The mounded middens are surface deposits of refuse, which include inorganic and organic remains. The problem with interpreting these features as simply large refuse heaps is that they are also filled with the most elaborate and symbol-laden items fabricated during the Puerto Rican early ceramic-age. In these mounded middens we are observing the products of multiple disposal processes. To a degree, the three-part disposal model developed by Hayden and Cannon (1983) is intersected by the Maisabel mounds. That is, they reflect ease of disposal considerations, value of material items (provisional disposal or storage), and hindrance factors. Ease of disposal considerations are reflected by the great quantities of faunal refuse and such utilitarian items as cooking vessels, firecracked rock, unmodified local stone, and fragments of broken stone and shell tools. In sedentary communities, this kind of mundane debris is periodically collected from various dispersed locations within a village, and deposited in such discrete areas as pits, ravines, lakes, and heaps (Hayden and Cannon 1983; Schiffer 1987). In addition to everyday debris, a considerable amount of "expensive," high energy-investment items is found within the mounds. These include such objects as: finely crafted ground and polished stone celts and adzes that have no macroscopic evidence of use; delicately carved shell and bone ornaments; a wide range of lapidary items; and elaborately decorated ceramic vessels, many of which are adorned with effigies. The great quantities of these sumptuary items in the mounds indicate that they are not present as a result of accidental loss and disposal along with everyday debris. Further, a number of the elaborate ceramic vessels are nearly complete suggesting that they were carefully deposited in the mound. In terms of the Hayden and Cannon (1983) model for refuse disposal, one might interpret the presence of the sumptuary remains as high value items in provisional storage to be used at a later date. Alternatively, this class of material may represent what Schiffer calls "offertory caches" (Schiffer 1987: 80). That is, the items are placed, purposely, as offerings in a ritual context, similar in status to things that are interred with burials.
The archaeology of community organization
61
Additional evidence that the mounds consist of more than simple heaps of refuse, is the presence of two unusual features in Mounded Midden 1. Deep in one of the units (N106W10) were found two isolated crania. One was an elderly male (Feature 1; 55 years or older; 70 to 80 cm below ground surface, orbits directed north) and the other was a dog (Feature 2; n o to 120 cm below ground surface, orbits directed south). Both crania were deposited with the dorsal surfaces up. They do not appear to have been casually discarded in a refuse pile. To the contrary, they were carefully positioned, probably in a ceremonial context. In this unit, too, a small fragment of guanin (gold, copper, silver alloy) was recovered (Siegel and Severin 1993). The Spanish observed the Taino Indians treating this material with great reverence (Vega 1979). It is likely that the material was highly valued during Saladoid times as well, especially given the labor invested in producing the alloy. The mounds, like the macroblock, were used throughout the occupations of the site. However, instead of all occupations represented in each mound, there appears to have been a pattern of shifting use (or disuse) of these features through time. Mounded Midden 1 is filled predominantly with Hacienda Grandecomplex artifacts. The Cuevas period is lightly represented in some of the excavation units. Mounded Midden 2, located about 80 m south of Mounded Midden 1, contained Hacienda Grande, Cuevas, and Monserrate period artifacts. The mounded middens are not haphazardly located, but are placed in a circular configuration around the center of the site. This suggests that the placement of these features was part of the planning process in the layout of the settlement by the earliest occupants, and was followed by all succeeding inhabitants. Evidence for this process is most clear in Mounded Middens 1 and 2, where Hacienda Grande period artifacts are present in the lowest levels of all excavation units. An excavation unit (N112W88) placed near the easternmost midden (Figure 2.7: Midden 3) yielded quite a bit of Hacienda Grande-style material, suggesting that this midden, too, was begun in the earliest occupation. Finally, artifacts retrieved from auger test pits associated with the remaining two middens (4 and 5) also included Hacienda Grande-style ceramics. Most of the auger test pits deployed at the site did not contain Hacienda Grande artifacts. The presence of this complex in all of the spatially discrete middens, therefore, indicates that these features were selected for during the earliest time period. Following Hacienda Grande, the use of the middens varied, depending on the time period. Cemetery The section of the site termed the cemetery, was used as a burial zone throughout the prehistoric occupations of Maisabel. Except for a few cases, the burials are generally devoid of obvious offerings. In a few instances graves contained such
6z
PETER E. SIEGEL
special offerings as complete ceramic pots, nicely polished stone tools, or carved shell items. Like the macroblock, the cemetery constitutes a low artifact density section of the site. Unlike the macroblock, the cemetery yielded very few features, other than burials. Depth of the midden is relatively shallow, being roughly 40 to 50 cm including the plowzone. In general, the cemetery and macroblock middens are very similar, except for the excavation units associated with the ditch feature surrounding the longhouse. Thus, in the cemetery there appears to have been an effort to keep the area clear of large accumulations of debris. Much of the Stratum 2 midden surrounding the cemetery may be refuse that was centrifugally cleared from the cemetery. Life, death, and disposal: a ring model of site formation The preceding discussion of the various settlement areas, disposal processes, and archaeological residues may now be examined, as a whole, to characterize the behavioral organization of Maisabel. Behaviorally, the configuration of the site, and its internal organization, may be characterized by a ring model (Figure 2.12). At the center of the ring, space is partitioned into residential and burial zones. These areas are kept relatively clear of refuse, with debris swept towards the village periphery. DeBoer and Lathrap (1979: 128) observed that this pattern of village maintenance resulted in a doughnut-shaped midden accumulating at the settlement edge. In my ring model, DeBoer and Lathrap's doughnut midden conforms to sampling Stratum 2 at Maisabel. Beyond this zone of refuse, however, an additional ring to the settlement is present. Like the cemetery and house areas, this zone is low in artifact density. Smole (1976: 68), in describing Yanoama settlements, indicates that "trash heaps" accumulate outside of palisaded villages. Roosevelt (1980: 218), in studying prehistoric sites in the middle Orinoco, refers to a "band of refuse that might have extended beyond the clearing." In my ethnographic observations of refuse disposal in Shefariymo (southern Guyana), I noted that the Waiwai and Wapisiana had distinct areas for depositing considerable amounts of trash. In addition, however, careful examination of the ground surface beyond the edge of the village as defined by the tree line, revealed a relatively continuous, but light density, layer of debris. These observations were not made systematically in order to distinguish categories of refuse in different contexts. Stahl (1984:157-158) noted a "ring-like deposition of living deposits" surrounding a cleared central area at the Early Valdivia site of Loma Alta. Sites excavated in the West Indies revealing oval or circular zones of refuse include Indian Creek, Antigua (Rouse 1974), Monserrate, Puerto Rico (Rainey 1940), and Carrier, Haiti (Rainey 1941). For Indian Creek, it was observed that: it has the appearance of an oval ring, colored grey against a dark background. On the ground, it can be seen that the grey color is due to a concentration of shell
The archaeology of community organization
63
Cleared central portion of village. Dense midden accumulation surrounding central cleared area.
Mounded middens.
Low dense midden accumulation defining site periphery.
Cemetery.
Ill
H0USe>
Figure 2.12. Ring model of the internal structure and organization of the Maisabel site.
64
PETER E. SIEGEL
refuse in a series of middens extending around the periphery of an oval area measuring approximately 283 by 165 m. The refuse in the remainder of the oval area appears to be sparse and shallow . . . (Rouse 1974: 167). Based on the observations made by DeBoer and Lathrap (1979), Smole (1976), Roosevelt (1980), myself (Siegel 1985), Stahl (1984), Rouse (1974), and many others (Crocker 1985: 32-33; Gregor 1977: 50-56, ^y; Levi-Strauss 1948: 301; Maybury-Lewis 1985: xi; Myers 1973: 243-245; Nimuendajii 1946: 37; Riviere 1969: 38; Wilbert 1972: 32), a concentric ring model may be applied also to the Maisabel disposal patterns. The outer "band of refuse" (Roosevelt 1980: 218) defines the true boundary of the ancient settlement. Conclusions Some years ago, Thomas Myers (1973) offered an insightful review of Amazonian community patterns based upon archaeological and ethnographic data. Marshaling data from a variety of sources, Myers postulated demographic shifts in relation to the evolution of sociopolitical complexity. The major flaw of this study is that Myers was rather accepting of site sizes that were determined unsystematically, and there is very little discussion of the internal spatial structure of the communities. Nevertheless, the paper by Myers is important, especially for the era when it was published, because it directly addressed issues of community organization. Today's challenge is to evaluate prehistoric community organization in the tropics based on assemblage composition, feature distributions, and site structure. Judicious use of remote-sensing techniques combined with sub-surface testing is proving to be a productive way to proceed in tropical-forest settings (Roosevelt 1991). In the Maisabel cemetery I retrieved twenty-three burials. An additional ten burials were recovered by a local amateur archaeologist. If we consider this recovery rate as representative of the overall burial density in this section of the site, then there could be as many as 2,500 burials present. Demography, diet, and social status are domains of information that can be used to address directly issues of community organization. In order to retrieve an adequate mortuary sample, it may be productive to investigate the Maisabel cemetery (c. 6,300 m2) using remote-sensing methods. Burials, as sub-surface anomalies, potentially may be discovered prior to digging, thus precluding unnecessary excavation and site destruction. Continued advances in community-oriented archaeology in the neotropics hinge on careful considerations of sampling, site formation processes, ethnographic analogy, and critical review of ethnohistoric documents. Sound middlerange linkages between the archaeological record and our interpretations of it will result in more believable, and less speculative, reconstructions of lowland culture history and process.
The archaeology of community organization
65
Notes This research was supported by the Centro de Investigaciones Indigenas de Puerto Rico (San Juan, Puerto Rico). The CIIPR, founded and directed by Gaspar Roca, was an anthropological research center devoted to the study of Amerindian lifeways in South America and the West Indies. Comments on earlier drafts of this paper were provided by Albert Dekin, Randall McGuire, Anna Roosevelt, Irving Rouse, and Peter Stahl. I am entirely responsible for the analysis, interpretations, and conclusions presented here. 1 The deep 3 m deposits at Sorce were encountered only on the topographic eminences referred to here as mounded middens. Areas off these mounded deposits were considerably shallower. 2 Crab shell (Cardisoma guanhumi and Gecarcinus lateralts) is ubiquitous in Saladoid deposits (Rainey 1940; Rouse 1952a, 1952b). Its absence suggests that the ditch was dug during the Ostionoid period. This suggestion is supported by a C 14 date and artifact styles.
Archaeological methods for the study of ancient landscapes of the Llanos de Mojos in the Bolivian Amazon CLARK L. ERICKSON
The archaeological study of agricultural systems
Archaeological gardens and field systems are notoriously difficult to study. They tend to be "artifact poor" contexts, and thus, extremely difficult to date with accuracy. Stratigraphy tends to be heavily reworked and eroded, the result of continual cultivation and mixing of soil structure by humans and nature, both during the time of use, and after abandonment. Due to the poor preservation of botanical remains, there is usually no direct evidence for the crops which were cultivated. Technological information on cultivation practices and tools is limited, and rarely is there direct evidence for labor and social organization, land tenure, and efficiency of the system. Ethnographic analogy can be useful in many contexts, but it is usually difficult to determine direct historical ties between contemporary farmers and their previous counterparts. In many situations, ancient field and garden systems have been completely abandoned, breaking any continuity between past and present. Even in cases where ties can be demonstrated, the social, political, economic, and environmental situation has changed so much that the usefulness of direct analogy is limited. Historical records can sometimes be extrapolated back into the past, but agricultural practices are not often discussed in sufficient detail. Despite these limitations to research, certain archaeological field methods, combined with experimental archaeology, can provide the detailed information lacking in cases where historical and ethnographic analogy is inadequate and preservation is poor. In this chapter, I discuss research techniques which have been useful in our study of raised field agriculture in the Bolivian Amazon. Archaeologists and geographers have dispelled the "pristine myth" of the natural environment of the Americas before the arrival of Europeans (Denevan 1992). Ancient landscapes of the Americas show evidence of massive landscape modification projects undertaken by prehispanic populations. In raised field construction alone, there may have been 173,000 ha of fields and canals throughout the wetlands of the Americas (estimate based on Denevan 1982). 66
The study of ancient landscapes
67
When this figure is combined with estimates on terrace construction, artificial drainage networks, causeways and roads, and forest management, the effect on the environment is awe-inspiring. The effect of such massive landscape modification on the local ecology is unknown, but probably great. These are truly anthropogenic landscapes. Landscape archaeology has the potential to provide alternatives for sustainable agriculture for small farmers. Archaeology can provide details on the long-term history of local and regional landuse by native tropical peoples over thousands of years. Today, many of these areas such as the Llanos de Mojos of Bolivia are abandoned and depopulated. In other areas, large expanses of land have been appropriated for use by ranchers; thereby, suffering severe degradation through over exploitation by non-appropriate technologies. Some of this knowledge of past landuse technologies could be put into use by development agencies for ecologically sound agricultural production. It can not only aid native peoples, but can also relieve pressure on fragile landscapes like the tropical forests, by providing sustainable and appropriate alternatives to western models of development. Much of the available historical data regarding the neotropics has an extremely shallow time depth, extending back only to the arrival of the Spanish and Portuguese. The archaeological record for the Amazon extends back at least 12,000 years and much of this includes extensive landscape modification (Denevan 1992; Piperno and Pearsall 1990; Roosevelt 1989, 1991; citations in Denevan 1992). Archaeological methods are probably the best means by which we can address many of these issues regarding ancient landuse and humanenvironmental interaction. Limitations and potential of tropical lowland archaeology The most commonly cited limitations to traditional archaeology in the tropical lowlands, are preservation factors and problems of field logistics (for example, Meggers and Evans 1957, 1983; Roosevelt 1989, 1991: 100—155; and see papers, this volume). The majority of the material culture inventory used by native peoples of the tropics are organic and are not commonly preserved in hot humid contexts. Sites are often deeply buried below more recent sediments, covered with dense vegetation, or redeposited by erosion in downstream locations (see Siegel, Zeidler, this volume). Easily preserved stone artifacts are rarely found in riverine and wet savanna locations in Amazonia. The heavy leaching of soils makes delineation of stratigraphy and households and settlement pattern difficult. The consensus among many archaeologists is that little can be done beyond studying the more durable pottery, and developing chronological sequences through the analysis of ceramic style. Theoretical issues tend to be traditionally grounded in environmental and ecological paradigms such as carrying capacity, agricultural potential, and the appropriateness of projecting contemporary ethnographic
68
CLARK L. ERICKSON
models back into the past (Meggers 1971; Meggers and Evans 1983; Roosevelt 1991; Steward and Faron 1959). An alternative to this ecological-chronological focus is the work of Donald Lathrap and students. Lathrap focused on "big models" of Amazonian prehistory based on archaeological excavations in the Ucayali River Basin of the Upper Amazon (Brochado and Lathrap 1982; Lathrap 1962, 1970, 1977; Lathrap et al. 1985). Some of his best work focused on agricultural origins of tropical cultigens, complex modeling of population dynamics and migrations, ecological dynamics, and prehispanic landuse, which successfully integrated archaeological, genetic, agronomic, ecological, ethnohistorical, ethnographic, and linguistic databases (Oliver 1992). Due to advances in archaeological methods and research design, the traditional limitations cited for the study of tropical prehistory never bothered Lathrap. Excavation at the Real Alto site changed the face of lowland studies in South America and introduced major breakthroughs in methods and analytical techniques (Damp 1984; Marcos 1978; Lathrap et al. 1977; Pearsall 1979; Zeidler 1984). Using large-scale horizontal excavations developed in the midwestern United States, sampling strategies for survey, and botanical recovery techniques (pollen, opal phytoliths, and flotation), Lathrap and colleagues were able to address issues of households, community plan and structure, ceremonial architecture, and ritual activities in a prehispanic tropical context (Damp 1984; Lathrap et al. 1977; Marcos 1978; Stahl 1986; Zeidler 1984). Recent excavations at Marajo Island by Roosevelt (1989, 1991) have introduced a number of productive techniques for sub-surface detection of tropical settlements, dietary analysis, and mapping. It is now more difficult to use the traditional "lack of preservation argument" with the current theoretical perspectives and recovery techniques available to archaeologists working in the neotropics (see various papers, this volume). Prehistory of the Llanos de Mojos
The Llanos de Mojos The Llanos de Mojos of the Bolivian Amazon is one of the largest areas of seasonally inundated grassland savanna in the world, covering some 145,000 km 2 (Figure 3.1). An additional 55,000 sq km2 of the zone are composed of patches of dense tropical forest, meandering rivers and oxbow lakes, river levees, large shallow lakes, and permanent swamps (Denevan 1966a). During the four to six month wet season, a large proportion of the low-lying terrain of the Llanos de Mojos is covered by a sheet of surface water ranging from only a few centimeters to several meters in depth. The savannas gradually drain during the dry season when surface water becomes scarce in many areas. The alternation of marked wet and dry seasons, and the waterlogged and heavy soils, have a significant impact on the present landuse patterns of the region. The indigenous populations
The study of ancient landscapes
1
66°
f
• Rio Yata Area
r\
69
1
65°
•
San Joaquin
Trinidad
r •
15 0
Central Llanos de Mojos Region population centers major raised fields
0 kilometers
Figure 3.1. Location map of the Central Llanos de Mojos region with important sites and features mentioned in the text.
70
CLARK L. ERICKSON
developed a complex system of massive earthworks to solve the problems of water control and agriculture (raised fields, canals, dikes, reservoir impoundments, drainage systems), transportation and communication (large raised causeways and canals), and shortage of good locations for domestic occupation (raised platform mounds). The Amazonian basin includes several very diverse environments. Although primarily seasonally flooded savanna and wetlands, the Llanos de Mojos bears some resemblance to the riverine systems of the central Amazon. The major "white water" rivers (Mamore, Beni, Maniqui, Guapore) crossing the savannas provide a microcosm of floodplain environments common to the larger Amazonian rivers (point bars, levee formations, oxbow lakes, gallery forest, and backwater wetlands) (Denevan 1966a). This is where the largest and apparently longest-occupied sites are located (Dougherty and Calandra 1981,1981-2,1984). The savannas of Mojos are distinct from the drier savannas common in the interfluvial zones of the Brazilian Amazon due to the many months of seasonal inundation and vast areas of permanent wetlands and shallow lakes. Within the savannas, the gallery forests and wetlands provide a rich bounty of wild resources for exploitation (fishing, hunting, gathering) not commonly present in interfluvial zones of Amazonia. Under intensive cultivation, these zones may have sustained higher population densities than even the better-documented riverine zones of the central and lower Amazon. Raised fields
Raised fields are "any prepared land involving the transfer and elevation of soil above the natural surface of the earth in order to improve growing conditions" (Denevan and Turner 1974). Data obtained from experimental archaeology (Erickson 1985, 1986a, 1986b, 1988a, 1988b; Garaycochea 1986a, 1986b, 1987; Gomez-Pompa et al. 1982; Kolata 1991; Muse and Quintero 1987; Puleston 1977; Ramos 1986, 1990), ethnographic analogy (Denevan and Turner 1974; Jimenez and Gomez-Pompa 1987), agro-climatological modeling (Grace 1983; Knapp 1991; Kolata and Ortloff 1989), remote sensing (Adams et al. 1981; Lennon 1982, 1983), and archaeological survey and excavation (Bray et al. 1987; Culbert et al. 1991; Eidt 1984; Erickson 1987,1988a; Graffam 1989,1990; Hammond et al. 1987; Knapp and Ryder 1983; Kolata 1986, 1991; Kolata and Graffam 1989; Mathewson 1987; Parsons et al. 1985; Pohl 1989; Siemans 1989; Smith et al. 1968; Stemper 1987; Turner and Harrison 1983; Zucchi and Denevan 1979; and others) have provided insights into some of the functions of raised-field technology in the Americas.1 The benefits of raised cultivation platforms include: (1) drainage of excess water; (2) improvement of soil and cultivation conditions through aeration, mixing, and doubling of the depth of topsoil; and, (3) improved local microclimatic conditions. Canals and ditches between fields: (1) conserve mois-
The study of ancient landscapes
71
ture to extend the growing season and counteract drought conditions; (2) act as heat sinks to minimize frost damage; and, (3) provide a medium for aquaculture and for the capture, production, and recycling of organic sediments and nutrients. Raised fields in the Llanos de Mojos Passing references to evidence for the vast and sophisticated prehispanic agricultural complexes in the Llanos de Mojos were made in publications by Nordenskiold (1910) and Metraux (1942). The importance of these prehispanic earthworks was demonstrated independently by Plafker (1963) and Denevan (1963). Denevan continued this research for his dissertation and published an excellent summary monograph in 1966. Through aerial and ground survey and the interpretation of aerial photography, Denevan (1963, 1966a) and Plafker (1963) located and described massive complexes of abandoned earthworks including raised fields, canals, causeways, reservoirs, oriented lakes, surface alignments features, and mounded occupation sites within the open pampa grasslands. Other field systems reported since then indicate that the prehispanic earthmoving efforts may have been even more substantial (Bustos 1976a, 1976b, 1976c, 1978a, 1978b, 1978c, i978d; Dougherty and Calandra 1984; Erickson 1980; Erickson et al. 1991; Lee 1979; Pinto Parada 1987). Recent research has demonstrated that the densely forested zones within the Llanos de Mojos also show evidence of intensive prehispanic occupations and earthworks (Arnold and Prettol 1989; Erickson 1980; Erickson and Faldin 1979). Denevan (1966a, 1980,1982) has conservatively estimated the existence of 100,000 raised field platforms within an area of 72,000 km2, and approximately 503 linear kms of raised causeways in an area of 3,900 km2. The number, size, and density of mound (lomas) and old levee occupation sites (islas) apparently associated with raised-field farming in the savanna zones and in the river gallery forests, is astonishing (Bustos 19760^ 1978a, 1978c, i978d; Erickson 1980; Erickson et al. 1991). The origins of these massive landscape modifications are unclear, owing to a lack of archaeological research. Our preliminary dating (five radiocarbon dates) from fieldwork in 1990, indicates that some raised fields date to between 800 and 2070 years BP (Erickson et al. 1991). The vast majority of the earthworks have been abandoned at least since the arrival of the Spanish, although there is some evidence that limited construction and use of causeways and canals extended into the historic period (Denevan 1966a; Pinto Parada 1987). Population within Amazonia declined rapidly after contact with the Europeans (Denevan 1966b, 1970b, 1976). Warfare, disease, exploitation for labor, population reorganization, and ethnocide swiftly took their toll on the native population of the Llanos de Mojos. This post-conquest population collapse is believed to have been responsible for the abandonment of the raised-field systems (Denevan 1966a).
72-
CLARK L. E R I C K S O N
Regional cultural development and raised-field agriculture The lack of general archaeological interest in Amazonia has included the Llanos de Mojos. Despite evidence of the area's importance in prehistory (that is, vast remains of raised-field and causeway-canal systems), relatively few investigations have been carried out in this zone. Most research has been limited to investigations of the large mound occupation sites (lomas) located along the course of the Rio Mamore near Trinidad (Bustos 1976a, 1976b, 1978a, 1978b, 1978c, i978d; Dougherty and Calandra 1981, 1981-82, 1984; Hanke 1957; Kuljis et al. 1977; Nordenskiold 1910, 1913, 1930; Ryden 1941). Other investigations have concentrated on the frontiers of the Llanos de Mojos (Arnold and Prettol 1989; BeckerDonner 1956; Bustos 1978b; Cordero 1984; Dougherty 1985; Kelm 1953; Nordenskiold 1924b; Reister 1981). Archaeological site surveys have been limited to zones along the recently constructed road between San Borja and Trinidad where numerous large precolumbian settlements and agricultural earthworks have been recently discovered (Bernardo Dougherty: personal communication; Bustos and Faldin 1978; Erickson 1980; Erickson and Faldin 1979; Erickson et al. 1991; Faldin 1984). A coherent and adequate synthesis of the prehistory of the zone has yet to be presented, although several attempts have been made to place the limited data available within the broader context of Amazonian prehistory (Bennett 1936; Brochado 1984; Brochado and Lathrap 1982; Denevan 1966a; Howard 1947; Lathrap 1970; Lathrap et al. 1985; Meggers and Evans 1983; Nordenskiold 1910, 1930; Portugal Ortiz 1978; and others). Much of the previous work in the Llanos de Mojos focused on developing ceramic chronologies and excavation of the largest mound sites (Bustos 1976a, 1978b; Dougherty and Calandra 1981-82, 1984). The Smithsonian Institution excavations focused on the excavation of the largest mounds along the Rio Mamore near Trinidad. Although survey work involving surface collections and small-scale excavations was carried out on a variety of sites, this was undertaken in the context of developing and refining the ceramic chronology for the region. Small excavation units of up to 15 m depth were dug in artificial levels with concern for applying quantitative seriation (Dougherty and Calandra 1981-2, 1984). The cultural development and social complexity of the Llanos was systematically downplayed in the reports, and the importance of the prehispanic raised fields and other evidence of agricultural engineering was often dismissed (Dougherty and Calandra 1981, 1984). Traditionally, the raised fields of the Llanos de Mojos have been associated with the prehispanic occupation sites found throughout the region. In particular, these included the "mound cultures" along the Rio Mamore, and the ethnohistoric and ethnographic populations recorded for the area (Bustos 1976a, 1976b; Denevan 1966a; Lathrap 1970; Metraux 1942; Nordenskiold 1924a; Pinto Parada 1987; and others). Few of these sites are near known raised-field blocks. It is highly likely, however, that the scores of archaeological sites recorded during our
The study of ancient landscapes
73
1979 survey in the savanna and forest between Trinidad and San Borja where major zones of raised fields are reported, were the occupation sites of raised field farmers (Erickson 1980; Erickson and Faldin 1979; Faldin 1984). This remains to be demonstrated through direct archaeological association with the agricultural earthworks. Of all the archaeological research conducted in the zone, only three projects have focused on the impressive agricultural remains and causeway earthworks. The results of the excavations of a moat-embankment structure in the far north of the Llanos de Mojos were presented recently (Arnold and Prettol 1989), and a large agricultural complex of raised fields, canals, and causeways near San Borja has been briefly studied and described (Erickson 1980). The joint University of Pennsylvania/Instituto Nacional de Arqueologia exploratory project conducted in 1990 at the El Villar site on the Rio Matos provided the first excavations, detailed topographic mapping, ethnobotanical and soil samples, stratigraphic profiles, and radiocarbon dates for raised fields and causeways in the central Llanos de Mojos (Erickson et al. 1991; Jacob 1991a; Jones 1991b). Any understanding of settlement systems, socio-political organization, or regional development must take into account the advantages and limitations of the agricultural infrastructure which supported the population in this region. The overall project goal of our archaeological investigation of the raised fields of the Llanos de Mojos is to provide important information relating to: (1) the regional cultural development; (2) the role of wetland agriculture in developing precolumbian tropical lowland societies; and (3) the relationship between intensive agricultural and socio-political organization. In many areas, causeways and associated canals commonly co-occur with raised fields (Denevan 1963, 1966a, 1991; Erickson 1980, Erickson and Faldin 1979; Pinto Parada 1987; Plafker 1963). Several hypotheses regarding this association are being investigated. Causeways and canals are believed to have been used for transportation and communication between residential sites and between sites and fields (Denevan 1966a; Pinto Parada 1987; see also Garson 1980). I have proposed that the massive causeway and canal networks are integral to the proper functioning of raised fields (Erickson 1980; also see Bustos 1978a; Lee 1979; Pinto Parada 1987). In addition to serving for transportation, the raised causeways may have also functioned as flood-control dikes, artificial levees, and reservoirs to control water at optimal levels during the wet season and to conserve moisture for dry season cultivation (potentially allowing double or triple cropping). The associated canals would also have provided a means of diverting water to where it was needed. Both causeways and canals may have been important in aquaculture as well, in particular the raising of fish in these artificial water bodies. Our study focuses on a much different scale of investigation than the past work in the Bolivian Amazon. Our investigation is regional in scope, following the pioneering work in cultural geography of William Denevan (1963, 1966a). The
74
CLARK L. ERICKSON
multidisciplinary, landscape archaeology approach used here is very different from the previous strategies used by the projects mentioned above. Methodologies for a landscape archaeology In order to address these broader issues, the Agro-Archaeological Project of the Beni investigated a number of raised-field complexes using a variety of archaeological techniques. These included: interpretation of remote sensing imagery (LANDSAT digital imagery and associated photographic products and standard stereo pairs of aerial photographs); reconnaissance from small aircraft; ground survey; topographic mapping with laser theodolite; archaeological excavation; and agricultural experimentation. Most of these techniques for application in the Llanos de Mojos were developed, tested, and shown to be successful in my recent research on precolumbian raised-field agriculture in the Lake Titicaca Basin (Erickson 1985, 1986a, 1986b, 1987, 1988a, 1988b, 1991, 1992a, 1992b, 1993; Erickson and Brinkmeier 1991; Erickson and Candler 1989; Garaycochea 1986a, 1986b, 1987) and during exploratory research conducted in the Llanos de Mojos in 1990 (Erickson et al. 1991). Although the use of experimental methods in the study of raised-field agriculture is a relatively new approach (Erickson 1985,1988a; Gomez-Pompa et al. 1982; Kolata 1991; Muse and Quintero 1987; Puleston 1977; Riley and Freimuth 1979), the methods used in this study are not new, and have been used in numerous projects in South America and the Mesoamerican tropics. Elements of these techniques have been used successfully by many archaeologists and geographers in the investigations of prehispanic raised fields and agricultural landscapes in the lowland American tropics (Bray et al. 1987; Darch 1983; Denevan 1966a; Mathewson 1987; Parsons et al. 1985; Pohl 1989; Puleston 1977; Siemans 1989; Stemper 1987; Turner and Harrison 1982; Zucchi and Denevan 1979; and others; general sources of information on methodologies appropriate for a landscape archaeology can be found in Darch 1983; Denevan et al. 1987; Farrington 1985; Gleason and Miller 1993; Harrison and Turner 1978; Killion 1992).
Aerial photographic interpretation A major limitation to the study of raised fields in the Llanos de Mojos, and other zones of the humid tropics, is the lack of accurate maps for guiding survey and for locating sites. Because the area is considered to be of low priority by the government, maps are either not available for large areas, or they are inaccurate and of poor scale. The use of remote sensing, in particular aerial photographs and satellite imagery, has been very important in overcoming this limitation (see Zeidler, this volume). Interpretation of aerial photographs and remote sensing has had a long and
The study of ancient landscapes
75
important history in the study of raised-field archaeology in the Llanos de Mojos and elsewhere (Adams et al. 1981; Dahlin and Pope 1989; Denevan 1963, 1966a; Harrison and Turner 1978; Plafker 1963; Siemans 1989; Siemans and Puleston 1978; Turner and Harrison 1982; and others). These resources become especially important both for locating potential raised-field locations and for use as base maps. The raised fields of the Llanos de Mojos were discovered independently by two individuals using aerial photographs (Denevan 1963; Plafker 1963). On the basis of previous documentation, my study of aerial photographs in the collection in the archives of the Instituto Geografico Militar and the Bolivian Air Force's Oficina de Aereofotogramia, and on-ground survey within the Llanos de Mojos, the most extensive and best preserved raised-field complexes and canalcauseway networks are located: (1) along the Rio Apere in central Mojos southwest of San Ignacio, and near the ranches of El Peru and La Esperanza (Denevan 1963, 1966a; Plafker 1963); (2) south of Laguna Rogoaguado in northern Mojos (Denevan 1966a); and (3) in the savannas between San Ignacio and San Borja along the San Borja—Trinidad Highway (Denevan 1966a; Dougherty and Calandra 1984; Erickson 1980; Erickson et al. 1991; Metraux 1942; see Figure 3.1). This highway provides relatively easy access to large blocks of earthworks and can be used as a survey transect across the central Llanos de Mojos. Our procedure is to locate prehispanic earthworks on the photographs through stereoscopic analysis and high magnification. Cultural features are photographically or xerographically enlarged, with features traced onto base maps for closer investigation in the field (Figures 3.2-3.5). Whenever possible, series of contact prints (9 in x 9 in) that can be viewed as stereopairs under low powered stereoscopes are purchased. Selected photographs with dense remains of raised fields are often enlarged to 1 m x 1 m. Although expensive, these enlargements bring out additional detail of earthworks, and are better for tracing earthwork patterns and for guiding fieldwork. Photographs are examined with additional low power magnification. Despite the low topographic relief (rarely more than 2 m), raised fields and causeways can easily be detected. High contrast black and white enlargements of original photograph sections are often made from negatives taken with a camera macro lens (Figure 3.3). We have found that the xerographic copy machine is an excellent tool for producing high contrast enlargements of aerial photographs to distinguish fields and earthworks. These are also useful for making inexpensive field copies, and for creating photo mosaics without having to damage the original stereopairs. Visibility can be a problem at times. Raised fields and other earthworks have been abandoned for at least 500 years, and have undergone considerable erosion. Cattle grazing, construction of roads and drains, plowing by tractors, and natural factors such as the annual floods and sediment buildup, have greatly reduced the earthworks and filled canals with sediments. Despite this destruction, raised fields, settlements, and other earthworks can be located on aerial photographs of
76
CLARK L. ERICKSON
Figure 3.2. Oblique low altitude aerial photograph showing prehispanic raised fields at the Arizona ranch, 60 km east of San Borja.
Figure 3.3. Aerial photograph showing several complexes of prehispanic raised fields at the Santa Fe and La Envidia ranches, 19 km WSW of San Ignacio. The dark linear and curvilinear features are large causeways.
The study of ancient landscapes
77
Figure 3.4. Digitally enhanced section of an aerial photograph of prehispanic raised fields near Santa Ana de Yacuma.
Figure 3.5. Aerial photograph of forested islas in the pampa west of San Ignacio. Evidence indicates that most of these features are small prehispanic village mound sites where raised-field farmers lived, or artificial agricultural earthworks. scales lower than 1:40,000. Even larger scales of high resolution aerial photographs can be useful with photographic enlargement or optical magnification. The rectilinear features of the artificial earthworks contrast sharply with more random forms of natural features on the landscape. The artificial topographic relief created by the construction of causeways, canals, and raised fields is also easily distinguished on stereopairs of aerial photographs. Vegetation, moisture, and soil differences produce sharp contrasts on the aerial photographs. The field platforms, causeway surfaces, and mounds tend to support vegetation (trees and shrubs) adapted to drier environments. Large termite mounds are also common
78
CLARK L. ERICKSON
on earthworks. The canals between fields, or alongside causeways, are colonized by aquatic plants which appear as darker areas in the photographs. Moisture differences, sometimes in the form of standing water, are also good indicators of canals and ditches. Other earthworks can be seen in disturbed areas of contemporary settlements (for example, the annually burned pampa, and slash and burn fields within the forest). Because of the low topographic relief and annual flooding, the most common locations to find occupation settlements and prehispanic human agricultural activity are in areas of naturally higher ground. These areas are generally active or abandoned river levees, or artificial accumulations of fill and midden from many generations of inhabitants. In many cases, barrow pits and circular canals ring these settlement sites. Aerial photographs are also important in monitoring the erosion and destruction of raised fields and associated features. The earthworks have suffered from extensive destruction recently. Photographs taken between 1959 and the present, show profound changes in the region. Fields, causeways, and canals have been destroyed through road construction and associated causeways and drainage features. Cattle grazing in the pampas has inflicted massive damage through the leveling of topography by cattle hooves. Some mechanized farming has also begun to level large areas of pampa. Logging activities and associated infrastructure (roads, bridges, camps) have caused considerable damage to archaeological features. Urban development has also taken its toll. For example, the town and airport of Santa Ana de Yacuma was constructed on top of a large raised-field block. The structure and morphology of raised fields vary widely throughout the Llanos de Mojos (Figures 3.2, 3.3, and 3.4). The nature of this variation is an important focus of the project. The internal differentiation within field blocks may be related to land tenure and social groupings responsible for the construction and maintenance of the field blocks. The structural differences between discrete field blocks may be related to ethnic or larger scale social groupings, or to chronological differences. Preliminary analysis of aerial photographs in the central Llanos de Mojos has shown that, between major field blocks, there tends to be several kilometers of areas without fields, possibly representing frontiers or boundaries between social groups. Aerial survey As noted above, coverage is limited for key zones, and eroded raised-field remains are not always clearly visible on some large-scale black and white aerial photographs. A small aircraft was rented for aerial survey in order to cover as much of the central Llanos de Mojos as possible. Field systems and causewaycanal earthworks were photographed at various scales in order to produce field maps, guide the ground reconnaissance, document changes in landuse and vegetation cover, and aid in interpretation of the archive aerial photographs
The study of ancient landscapes
79
(Figures 3.2 and 3.5). The visibility of earthworks changes seasonally throughout the year, due to rainfall and moisture conditions. We found that the best conditions were after heavy rains when canals offieldsand causeways held water, thereby highlighting patterns. We were also told that several weeks after the burning of the pampa, when new grasses are beginning to return, is a good time for delineating canals between raised fields, as their retained moisture permits more rapid growth of new grasses. Each step of the flight had to be carefully planned, with duties divided among the team members. The flight plan was discussed with the pilot and signals were developed to communicate during the noisy flight. We used four cameras loaded with different types of film, plus a hand-held video camcorder. A tape recorder, and the sound on the video, were used to keep a continuous record of the flight and photographed locations. A Geographic Positioning System (GPS) was used to record locations when the flight was beyond the area covered by aerial photographs. We found that low flights were good for areas where we had photographic coverage to guide us; higher flights were more useful for new areas. A series of tight curves were made above interesting sites to obtain close, near vertical photographic coverage. Luckily the pilots of the Beni know the region well and can be trained to identify archaeological features on their own. One pilot later found and reported two small blocks of fields on his regular routes after flying with us on a survey flight. Digital remote sensing Remote sensing has become an important tool for archaeologists studying ancient raised-field agriculture (Adams et al. 1981; Dahlin and Pope 1989; Siemans 1989). We are in the process of preparing computer-generated classifications of key zones from digital LANDSAT imagery using the IDRISI software (a Geographic Information System and remote sensing software package developed by the Department of Geography, Clark University). These will be used to produce a series of large-scale environmental base maps for guiding ground reconnaissance, and for locating earthwork features found in the photographs and field. The computer can efficiently produce rough maps of vegetation communities, landuse, forest—pampa boundaries, river courses, and roads over large areas. The computer-generated vegetation and landuse classifications will be checked and refined through incorporation of ground data collected during the field project. UTM and longitude/latitude coordinates can be projected onto these computer-generated maps. Another technique used by this project is the computer scanning and digitization of aerial photographs using the IDRISI. The scanning of black and white aerial photographs permits detailed manipulation of the scanned image to increase contrast, emphasize certain features of the landscape, and delineate earthworks. The system allows minute features to be enlarged for easier study
80
CLARK L. ERICKSON
and fieldmap production. Gray scales of the black and white images can be controlled to define cultural and natural features (Figure 3.4). With ground truth, UTM or longitude/latitude coordinates can be projected on these images. The images can also be used as illustrations with the addition of text, legends, and classifications of landuse features. A preliminary study has indicated that a computer can recognize signatures of raised fields (Erickson 1981). Even on the coarse resolution of LANDSAT imagery, the distinct patterning of reflected light created by the undulating raisedfield platforms and canals and the alternating wet—dry land surfaces, produces distinct patterns that can be distinguished from non-raised-field areas. Whether these preliminary data can be extrapolated over the entire Llanos de Mojos is not yet clear. A mixed media approach (black and white photographs, color photographs, digitized images, satellite digital imagery from LANDSAT and SPOT, and topographic survey on the ground) appears to be the best means of studying these earthworks. The integration of multiple techniques for finding and mapping raised fields and other earthworks provides flexibility and efficiency. The digital nature of the topographic and excavation mapping by EDM, GPS point locating, computer-enhanced aerial photographs, and LANDSAT imagery can be integrated for analysis through a Geographical Information System or other data management. Ground survey of agricultural earthworks and settlements Ground survey in the Llanos de Mojos is logistically difficult and expensive. Throughout most of the area, visibility is limited by dense vegetation cover in the forests and high grasses in the pampas. Active geomorphological processes have buried many cultural remains. Despite these limitations, ground survey conducted on foot, horseback, oxcart, canoe, and four-wheel drive vehicles can locate many sites. The recording of sites through survey (Figure 3.6) is aided by natural and modern anthropogenic disturbances which provide "windows" in the vegetation cover and sediment overburden. Road cuts are the most useful, and provide long survey transects across the pampa and forest. During survey on foot and by canoe, we have utilized river cuts to find sites. Sherds can be found in situ, buried in these banks (Cordero 1984; Michel and Lemuz 1992). We have found that with an understanding of the local geomorphology, site locations can be predicted more efficiently and accurately than by time consuming 100 percent survey (a difficult endeavor in the humid tropical lowlands, see Zeidler, this volume). Occupation site locations are similar to those traditionally reported in the archaeological literature for riverine Amazonia - large mounds on active or abandoned river levees, or on adjacent uplands (Brochado 1984; Lathrap 1970, 1968b; Lathrap et al. 1985; Meggers and Evans 1983; Meyers 1992; Roosevelt 1991). Because of the prehispanic importance of the pampas for
The study of ancient landscapes
81
Figure 3.6. Prehispanic raisedfieldsat the La Envidia ranch. Recent burning of grasses by ranchers has exposed the raised-field platforms and canals. agricultural production, many large and small sites are located some distance from rivers. These sites are commonly islas, or low circular or oval mounds (of up to several hectares and one to two meters tall), distinguished by tree and shrub cover (Figure 3.5). The sites within the pampa are commonly established on low river levee formations of abandoned channels or edges of permanent water bodies such as swamps, marshes, and lakes. Raised fields are found on the edges of permanent water (swamps, lakes, marshes), or on the back slopes of abandoned levees within the waste pampa. In many riverine areas, the geomorphologically active river floodplains have destroyed or deeply buried all prehispanic sites, making total coverage survey useless. Occupation sites are likely to be covered with dense vegetation. Without human disturbances, it is often impossible to make adequate surface collections. Most modern settlements and fields in the pampa are located on areas of slightly higher topographic relief {islas) to prevent flooding. Many are old prehispanic settlement mounds. These open and cleared areas can be easily surveyed, and the disturbances caused by the construction of post-structure ranch and farm buildings, pits, fences, cattle grazing, burning of pasture, and swidden cultivation, make adequate surface collections possible. Tree falls are also potential areas for surface collections, as the uprooting of large shallow root systems of buttressed tropical forest trees can open up many square meters for collections.
82
CLARK L. ERICKSON
W^HS
Figure 3.7. Surface collections being made at La Asunta site, a large occupation site covering several hectares bisected by a logging road near San Ignacio. The site is eroding out of the bank and roadbed recently cut by heavy equipment. One of the most efficient means of systematically locating a large number of sites, is to use the various new roads as survey transects. These tend to be deep disturbances, as soil is excavated by bulldozer from ditches to construct the modern raised causeway roadbed. The clearing of vegetation and the excavation of ditches have exposed many archaeological sites and earthwork features. Surface collections are relatively easy on these sites because of the initial construction disturbance, and periodic re-excavation during maintenance of the roads (Figure 3.7). The seasonal rains and erosion expose sherds in the road cuts. Roads also provide stratigraphic profiles of occupation mounds, causeways, and raised fields, some running the length of the site. In addition, many of the roads have been well surveyed and accurately mapped for construction and improvement. The large multicomponent mound settlements along the Rio Mamore near Trinidad, cover several hectares with artificial fill to a depth of 15 m, representing long cultural occupations (Dougherty and Calandra 1981-82). None of these large sites have been excavated with horizontal techniques. In addition to their large size, many sites are deeply buried. The Llanos de Mojos is an active geomorphological landscape involving annual flooding of the pampas, deposition of riverine sediment loads, and the periodic re-working of riverine flood-
The study of ancient landscapes
Figure 3.8. Use of a coring device to find buried occupation sites and to determine the depth of agricultural features. plains through river channel changes. In excavations at the site of El Villar, evidence of domestic occupation dating to 800-120 BC, was found at 0.75 m to 1 m below the present-day pampa surface, and below later raised fields construction. A manual soil probe is used to locate buried sites, test the depths of midden deposits, verify artificial nature of mounds, and recover artifacts from vegetation covered sites (Figure 3.8; see also Siegel, Zeidler, this volume). A 9 cm diameter core is sufficient for rapidly testing sites up to 3 m in depth. Soil from these cores is screened and/or troweled for evidence of occupation. Soil color and texture changes are also mapped to document sub-surface stratigraphy. Our tentative data indicates that the coring program works as an efficient survey method because of the high density of sherds in most occupation sites in Mojos. At the Santa Fe/Tacuaral complex of raisedfieldsand causeways, a 2.5 m deep core from
84
CLARK L. ERICKSON
a causeway produced sherds associated with a possible living floor. The coring device can also be used to rapidly collect large samples of organic soils from the bases of canals. These are currently being processed to date the organic content of the samples using radiocarbon analysis. In areas lacking maps or aerial photographic coverage, sites recorded during survey can be very difficult to geographically locate. In addition, the relatively low, featureless relief topography, combined with frequently changing forest/ savanna boundaries, can make it difficult to locate sites, even with the aid of photographs. Our project has recently begun to use a global positioning system (GPS) (Magellan NAV 5000) which tracks up to 5 satellites at a time, providing site locations with an accuracy of approximately 15 m. The system can also be used to guide ground survey crews to specific sites that were detected during interpretation of aerial photographs. The limitations of reliance on aerial photographic interpretation were made clear during the ground survey. At the Santa Fe and La Envidia ranches near San Ignacio, large blocks of well preserved raised fields which do not appear on the aerial photographs (Figure 3.9), were found within forested zones. The slightly higher ground provided by the linear raised-field surfaces provided appropriate drained soils for the establishment of large forests. Under the continuous tree canopy, trees grew in rows, thus mirroring the structure of the fields. It is apparent that over time, the forest has gradually advanced over the open grassland. Much of this may be due to grazing cattle which eat grasses but leave the inedible woody plants that gradually replace the grass cover. Topographic mapping Accurate topographic mapping is necessary for assessing the degree of integration between raised fields, canals, and causeways. Of particular importance are the mapping of: (1) discrete hydrologic units of raised fields and canals; and (2) causeways and associated canals. The generally low topographic relief and extremely low gradient of the savanna and streams necessitate mapping of the relative elevations of the aggraded river channels, levees, and the earthworks with very precise instruments. For this part of the project, an EDM Topcon Total Station laser theodolite, cabled to a hand-held computer, is used to accurately record long topographic transects and thousands of survey points for each hydraulic unit. Computer-generated topographic maps are produced daily from these points (Figure 3.10). Ground features found in survey (vegetation boundaries, fields, modern landuse, and settlements) are precisely located on maps and aerial photographs using the portable Global Positioning System (GPS). The use of the Total Station has many advantages. The theodolite readings are extremely precise over long distances (up to 2 km). This precision is necessary to address the hydraulic functions of raised-field systems. Automatic computer
The study of ancient landscapes
Figure 3.9. Clearing of vegetation from raised-field transects in preparation for topographic mapping at La Envidia. The eroded canals hold water from rainfall.
85
CLARK L. ERICKSON
0.00
-10.00
-20.00
-30.00
-40.00
-50.00 0.00 10.00 20.00 30.00 Figure 3.10. A computer-generated topographic map of raised-field platforms and canals at the Beni Biological Station.
The study of ancient landscapes
87
calculations and recording allow hundreds of survey points to be taken in a few hours, with less chance of human error (Weiss and Traxler 1991). This enables us to map large areas of earthworks. The setup can also be used for purposes of general piece plotting during excavation and basic artifact data management (McPherron and Dibble 1989). Combining digital aerial and LANDSAT imagery with computer-assisted topographic surveying on site, has a great potential for analytic and presentation purposes. The resolution of earthworks for study can be greatly improved by using graphic presentation such as CADD and other 3-D imaging systems. Our recent work at the Santa Fe Ranch has documented the systemic integration of causeways, canals, and raised fields. Field blocks are bounded by encircling causeways, forming low dikes. Causeways ringing a large permanent swamp, may have been used for flood control. During periods of rains in 1992, we observed causeways channeling runoff and excess water. Other causeways blocked the flow of water across the flat plains, forming large shallow reservoirs. Our maps of the microtopography of the fields have begun to document additional complexities of the ancient hydraulic engineering (Figure 3.10). Excavation of trenches A main focus of the fieldwork is to excavate key agricultural structures within the mapped zones. Stratigraphic trenches are excavated across raised fields, causeways, and canals in order to determine their original soil stratigraphy or earthwork morphology, individual building stages and constructional sequence, post-depositional erosional history, chronology of use and abandonment, and to collect samples for dating and paleoethnobotanical study (Figures 3.11-3.12). Excavating long units perpendicular to the orientation of raised fields and canals is the most effective technique. Trenches are excavated from the center of one raised-field platform to the center of the adjacent field platform. The intervening canal is excavated to sterile soil below the base of the deepest part of the canal, usually no more than 1 m. Causeway/canal excavations were similar. We find that trenches 1 m in width are large enough for us to draw stratigraphic profiles and sample the earthworks. Trenches have to be bailed or pumped periodically, due to flooding caused by the high water table. Stratigraphy within trenches is remarkably well preserved. The organic matter of the sediment-filled canals stands out in sharp contrast to the lighter soils of the raised fields and subsoil. In some cases, different stages of field construction could be noted within fieldplatform fill. Soil from the cultural strata within these excavations is carefully screened with \ in mesh, for the recovery of artifacts and macro-floral and faunal remains. In the future, several horizontal excavations will be made within raisedfield blocks for better definition of field and canal form. These trenches also provide models for reconstructing experimental fields (discussed below). These excavations are also important for the recovery of artifacts, organic remains, and soil samples. Within fields excavated at the El Villar site, we
CLARK L. ERICKSON
Figure 3.11. Excavation of Trench no. 1 at Santa Fe using natural stratigraphy. The excavation unit includes half of two adjacent raised-field platforms and the intervening canal. recovered potsherds, bone, and charcoal in situ. Samples for ethnobotanical, soil, and dating analyses are extracted either during excavation, or directly from the stratigraphic profile after section mapping. Samples are collected for the recovery of pollen (Pearsall 1989, and this volume), opal phytoliths (Piperno 1988, and this volume), other ethnobotanical and small faunal remains through flotation (Pearsall 1989; see also Stahl, this volume), and for analysis of soil composition and fertility. Samples are collected from all natural and cultural strata within every stratigraphic trench. For comparative purposes, additional samples are taken from areas where fields are absent. During the 1990 preliminary project at El Villar, excavations conducted in fields, canals, and causeways demonstrated the excellent preservation of original soil stratigraphy within the agricultural earthworks, despite humid tropical conditions (Erickson et al. 1991). Artifacts, especially ceramics, and carbonized botanical remains, were abundant within the earthen features, making stylistic dating, radiocarbon dating, and flotation recovery possible. Permanently waterlogged locations deep within raised-field and canal profiles may also provide
El Villar Site Raised Fields and Canals Southwest of Isla El Villar Trench 2/3/4 East Profile
S8
I
I
I
I
I
I
I
S22
I
raised field
I raised field
-
181.00
— 180.00
-
197.00
Strata Descriptions
dark grayish brown (10YR3/2) very fine sandy loam, humus, A Horizon, grass roots, reddish mottling, orange linear vertical stains, lots of organic matter, few fragments of ceramics and charcoal.
F
brown (10YR5/3) loam, large orange/brown mottling, very humid, B Horizon, clear boundary.
brown/dark brown (10YR4/3) loam, lot of rootlets, continuation of A Horizon, red/orange oxidized rootlet mottling, gradual boundary, upper fill of raised field, some fragments of ceramics and charcoal.
G
dark gray (10YR4/1) clay loam, humus, high percentage of organic material, lots of grass roots and decomposed plants, very humid.
H
dark gray (10YR4/1) clay loam, few rootlets, very homogeneous, B Horizon, some fragments of ceramics and charcoal, thin vertical cracks (0.5 - 3.0 cm wide) filled with dark gray (10YR4/1) silty clay, gradual boundary, lower fill of raised field.
gray (10YR5/2) clay loam, orange/brown mottling, rootlets and oxidized soil, thin vertical cracks filled with grayish brown (10YR5/2) clay, upper canal sediment fill, gradual boundary.
I
very dark grayish brown (10YR3/2) clay loam, dark yellowish brown (10YR5/8) hard concretions, produced by pennenantly humid and inundated conditions, thin vertical cracks filled with grayish brown (10YR5/2) clay, base of canal sediment fill, clear boundary.
J
brown (10YR5/3) clay loam, lots of yellowish brown (10YR5/8) mottling and concretions, thin vertical cracks filled with grayish brown (10YR5/2) clay, B Horizon, clear boundary.
dark gray (10YR4/1) loam, lots of oxidized brown/orange mottling giving the soil a yellowish color, thin vertical cracks filled with dark gray (10YR4/1) silty clay, some fragments of ceramics and charcoal, gradual boundary. very dark gray (10YR3/1) loam, cultural stratum, occupation floor, lots of fragments of ceramics, charcoal, ash, and burned clay, probably a buried A Horizon (paleosol), thin vertical cracks filled with dark gray (10YR4/1) silty clay, clear boundary.
Figure 3.12. Stratigraphic profile and soil descriptions of Trench no. 2/3/4 a t Villar site.
tne
90
CLARK L. ERIKSON
opportunities for the recovery of highly perishable cultural materials such as wood and bone. Dating of earthworks Results from our exploratory investigations indicate that raised-field agriculture has evolved over a considerable period of time (at least 1,000 years) (Erickson et al. 1991). The field systems appear to be the result of an accumulation of landscape capital by Mojos farmers, an incremental process of accretionary growth over time through the piecemeal construction of fields as needed, or changes made during the routine use and maintenance of the system (Doolittle 1984). Raised fields are dated using a variety of techniques. Relative dating of construction phases, use periods, and abandonment, is based on stratigraphic analysis from the excavation trenches. Absolute radiocarbon dates are obtained from in situ charcoal in field fill, and/or the rich organic sediments of canal fillin excavation trenches. It may also be possible to cross date ceramics recovered from field and canal contexts with the ceramic chronologies already established for excavated mounds along the Rio Mamore (Dougherty and Calandra 1981, 1981-82). It is often difficult to directly date raised fields. As they are used over considerable periods of time, agricultural fields and earthworks represent continually reworked soils. Thefillincorporated into platforms often comes from various locations. Canals and platforms are subject to periodic re-excavation and continual erosion, both through use and natural factors. Vertisol formation and leaching of soil colors, animal burrowing, and termite mound construction often erase stratigraphic boundaries. Diagnostic artifacts recovered in situ in filland canals can potentially be used for comparative dating, but chronological sequences have only begun to be established and unfortunately cannot be extrapolated over the whole of the Mojos region. Direct thermoluminescence dating of ceramics recovered from trenches has not been attempted yet, although it was a successful means of dating fields in the highlands of Peru (Erickson 1987). Radiocarbon methods have been very useful in dating field systems. Adequate samples of charcoal were recovered from raised-field fill and occupation levels below raised fields at the El Villar site. Bulk soil dates run on the organic content of sediments from canal bases between raisedfields,provide information on canal use or abandonment. AMS dating may make it possible to work with even very small organic samples. We have found that a variety of dating techniques must be used in raised-field research. Sites along the Rio Mamore span nearly 3,000 years of continuous or intermittent occupation; therefore, indirect dating of raised fields through use of associated sites is probably unreliable, due to the multicomponent nature of occupations. Certain features of design and proximity between causeways/canals
The study of ancient landscapes
91
and field blocks indicate contemporaneity as they were probably constructed as a single unit or over a relatively short period of time. Stratigraphic relationships can also give clues to relative chronology, where younger fields have been constructed on top of older fields, or where fields have been gradually improved through the addition of fill over time. In addition to vertical stratigraphic analysis, horizontal stratigraphy may also be useful for the relative dating of raised fields, as field blocks are gradually expanded in area over time. The connection of discrete sites by large causeways and canals, indicates communication and transportation between presumably contemporaneous settlements. Raised fields are often crossed by causeways and canals which connect them with sites on river levees or occupation mounds in the pampa. In addition to hydraulic functions, these causeways and canals provided easy access to fields for farmers, and a means for transportation of agricultural produce from raised fields to settlements. It should be possible to date fields through the establishment of a seriation of field and earthwork structure and morphology. The stylistic differences between field blocks in morphology, scale, and formal structural pattern (wavelengths, orientations, internal patterning, engineering techniques, size of field platforms, canals, and blocks) may provide an efficient means of dating the construction, either intra- or inter-regionally. Such a seriation would prove immensely useful for addressing current questions of field development and culture history in the area, and for guiding future excavations. Alternatively, the major differences among raised-field blocks in the Llanos de Mojos may represent environmental, functional, or ethnic differences in field construction by contemporary farmers. Regional ethnic groupings, internal group social organization, and land tenure may be mapped into formal spatial structure of raised-field blocks. At both El Villar and Santa Fe, spatially discrete field blocks (with different wavelengths of canals and platforms) are bounded by large causeways and canals. Fields are not continuous over the landscape in Mojos, but rather form independent units or regions which suggest social organizational units. Aerial photographic interpretation and low altitude flights over raised fields in Central Mojos also suggest that areas of unoccupied and unutilized land exist between field blocks. Ethnobotany The poor preservation of botanical/organic remains was commonly lamented in the traditional perspectives on the limitations of tropical archaeology (see Pearsall, Piperno, Stahl, this volume). In addition to samples from soil probes in field canals and occupation sites at the sites of El Villar and Santa Fe, samples for soil, pollen, opal phytolith, and flotation analyses were collected during excavations of trenches in raised fields, canals, and causeways. Surprisingly large amounts of macroremains of charcoal (primarily wood) were found within
92-
CLARK L. E R I C K S O N
raised-field fill, canals, and within occupation sites at El Villar. This indicates that flotation could be successfully used in this tropical context (see Pearsall, this volume). Tentative results from analysis of archaeological pollen recovered from raisedfield contexts have identified Xanthosoma (possibly gualusa or papa china cultivated for its taro-like corm), Bixa (possibly urucu or achiote, used as body paint and food coloring), and Ilex (possibly guallusa or mate, a strong, caffeinerich, ceremonial drink) in addition to a wide range of grasses, trees, and aquatic plants (Jones 1991b; Erickson et al. 1991). The opal phytolith analysis is incomplete, but preliminary results indicate the presence of similar plants (Piperno: personal communication). Experimental archaeology Experimental construction as a method of raised-field investigation, provides important insights into how the system functioned, the kinds of crops grown, labor input in construction and maintenance, nutrient production and cycling, dynamics of field hydrology, crop productivity, potential carrying capacity, sustainability of the system over time, and other important issues (Erickson 1985, 1988a, 1988b; Puleston 1977; Muse and Quintero 1987). Because the raised fields of the Llanos de Mojos have been completely abandoned (possibly for over 500 years), and there is no reference to this technology in the written ethnohistorical or ethnographic record, archaeological methods are the only means available for investigating raised-field agriculture. Experimentation or an "applied archaeology," based on what the archaeological research tells us about the technology, can also be a useful adjunct approach to understanding raised fields. Based on information collected through mapping and excavation of trenches in fields at the El Villar site (Figure 3.12), models for three different forms of raised fields were developed for construction in an experimental plot. One half of a hectare of raised fields was constructed during 1990—91 at the Biological Station of the Beni in central Mojos (Figures 3.13—3.14). In 1992, agronomy students of the Universidad Tecnica del Beni expanded these fields to include nearly another 0.5 ha of fields. The experimental fields are planted in native crops that were recorded as being important during the Colonial period. Fields are constructed and maintained by a group of farmers from a local community. Plans are to expand and continue these experiments in the contexts of native farming communities, in collaboration with the Bolivian National Academy of Sciences, Interamerican Foundation, and the Universidad Tecnica del Beni. Preliminary results indicate that raised-field farming can be labor efficient, very productive, and potentially sustainable. Despite heavy rains and massive flooding during 1991-92, the raised-field platforms remained dry. Water for irrigation was maintained in the deep canals during much of the dry season. Several older farmers remarked that it was the first time in their lives that they had seen the
The study of ancient landscapes
93
Figure 3.13. Oblique aerial photograph of the experimental raised fields at the Beni Biological Station, Porvenir Ranch, 50 km east of San Borja. The fields are being prepared for the 1992-93 growing season. pampas produce agricultural crops. Previous attempts at non-raised-field agriculture failed because of inundation by flooding and a high water table. Manioc and maize did exceptionally well the first year on the experimental raised fields. A dry season crop is being attempted to gauge the feasibility of year-round multicropping. Manual labor requirements for field construction are considerable (nearly 800 person-days per ha), but the fields can be farmed continuously for many years with little labor input for maintenance. When considered over the long run, this technology appears to be very efficient and productive. Integration of data from field and laboratory Many of the traditionally cited limitations to archaeological research in the humid tropics have been overcome by new technological breakthroughs for the recovery of data. The costs of labor-intensive archaeological survey and excavation of raised fields is offset by the massive amounts of topographic data collected quickly and efficiently in digital format. These digital data can easily be integrated with the impressive amounts of information on remote sensing using LANDSAT and SPOT, digitized aerial photographs, and piece plotting of artifacts.
94
CLARK L. ERICKSON
Figure 3.14. Raised fields constructed for experiments by students of the Universidad Tecnica del Beni at the Beni Biological Station in 1992. Field morphology is based on information obtained by archaeological excavations of prehispanic raised fields at the nearby El Villar site. These fields will be planted with a variety of crops native to the zone.
Information from remote sensing and aerial photographic analyses, correlated with ground survey and excavations, will provide the basis for the establishment of a regional, spatially organized Geographical Information System (GIS) data base containing details on natural, prehispanic, and contemporary landuse, and the environment. This data base will also incorporate recent soils and vegetation studies undertaken by various institutions in the Department of the Beni in association with the Beni Biosphere Reserve. GIS data base and digital imagery will be used to create a large-scale map and inventory of natural and cultural features of the Llanos de Mojos. The Department of Landscape Architecture at the University of Pennsylvania is in the process of setting up a Laboratory for Remote Sensing and Geographic Information Systems, and the new Computer Resource Center of the University Museum will also have this capability.
The study of ancient landscapes
95
Ancient landscapes and an applied archaeology A landscape archaeology, such as that described here, can provide a long-term perspective on intensive uses of local environments. It is possible to begin to investigate and monitor: (1) potential prehispanic population dynamics; (2) the productivity and sustainability of indigenous technologies; and (3) cultural landscape stability, change, and degradation over thousands of years. Given the current environmental devastation caused by contemporary agricultural practices in Amazonia, a study of the prehispanic agricultural technologies capable of supporting dense populations is of critical importance. These systems may provide alternative and less destructive strategies for sustainable development in the wetlands of Amazonia and other endangered tropical environments. The Llanos de Mojos also provides a striking example of differences between prehistoric and post-conquest period landuse. The density and size of occupation mounds, and the extent of large-scale engineering and intensive agricultural production stand in sharp contrast to the contemporary pattern of dispersed populations, low agricultural and cattle-raising yields, and abandonment of the engineering infrastructure. There is considerable potential for the re-introduction of time-tested technologies developed by farmers of the Americas. Notes I would like to acknowledge the Instituto Nacional de Arqueologia (INAR) and the Instituto Boliviano de Cultura (IBC) for granting the necessary permits and for collaboration in all aspects of the project. The ex-director, Juan Albarracin, and the current director, Oswaldo Rivera, of INAR, were instrumental in the planning and success of the project. Juan Peredo of the Bolivian Consulate in Washington D.C. helped with project visas. The archaeological team consisted of my coinvestigators Jose Esteves (1990) and Wilma Winkler (1992-94), Marcos Michel, Kay Candler, John Walker, and Andy Pelcin. The landowners of El Villar, Miguel Rea, and Santa Fe, Roberto Parada, are to be thanked for their hospitality and collaboration at thefieldsites.Funding for various phases of the project have come from research grants awarded by the Interamerican Foundation (1992), National Science Foundation (1992-93), H. John Heinz III Charitable Trust of Pittsburgh (1992), and the Research Foundation of the University of Pennsylvania (1990). CORDEBENI provided technical assistance and the use of a vehicle in the 1990 field season. The Beni Biological Station provided land for the experimental raised-field plots. Computer data base and EDM/total station surveying assistance in Philadelphia was provided by Andy Weiss (MASCA), Shannon McPherron (Anthropology), and Harold Dibble (Anthropology). I want to especially thank Kay Candler, John Walker, and Peter Stahl for editorial help in preparing this chapter. 1 Raised fields have also been found in other parts of the world. Numerous recent studies document the forms, distribution, and chronology of prehispanic raised-field agriculture in many locations in the Americas (for example, Armillas 1971; Bray, this volume; Bray et al. 1987; Darch 1983; Denevan 1970,1980, 1982; Denevan et al. 1987; Erickson 1987, 1988a; Farrington 1985; GomezPompa et al. 1982; Graffam 1989, 1990; Harrison and Turner 1978; Kolata 1989; Parsons and Denevan 1967; Pohl 1989; Turner and Harrison 1983; and others) and archaeological and ethnographic cases in the Old World (Denevan and Turner 1974; Farrington 1985; and others). The long-term perspective provided by archaeology is an excellent means to study this neglected technology.
Searching for environmental stress: climatic and anthropogenic influences on the landscape of Colombia WARWICK BRAY In memory of Don Lathrap, who was never afraid to paint with a broad brush.
This chapter explores two interfaces: one between human culture and the natural environment; the other between highland and lowland regions of Colombia. The starting point is a series of archaeological-environmental studies in the Caribbean lowlands and in certain highland areas of the country. These investigations are chosen because they draw attention to a number of methodological issues, the first of which is that (contra much romantic mythology) there is no essential difference between tropical lowland archaeology and archaeology as practiced elsewhere. The work of Roosevelt (1991) on Marajo Island, of Mora et al. (1991) in Colombian Amazonia, and of several contributors to the present volume, demonstrates that most techniques of field investigation are practicable in the American tropics, and that information about subsistence patterns and environmental conditions can be recovered from rainforest excavations. Conceptually, too, there is little that is unique to lowland archaeology. The comprehension gap that exists between "lowland" and "highland" specialists seems (to an outsider, at least) to have more to do with personal psychologies that with philosophical incompatibility. If these assertions are true, it follows that most of the central questions of archaeology are potentially testable against lowland data. One of these classic themes is the relationship between human culture and the global climate, in particular the notion that culture adapts to environment, and that climatic change is a prime mover that may cause the rise or collapse of both lifestyles and polities. These arguments have been much in fashion among Peruvianists (for example, Paulsen 1976, 1981; Conrad 1981; Isbell 1978; Shimada et al. 1991). To ecologists, much of the archaeological literature seems simplistic. The impetus for my essay was a request from a skeptical colleague to: (1) summarize the evidence for environmental change in lowland Colombia from the end of the Pleistocene to the Spanish Conquest; (2) examine the ways in which the ancient landscape was modified by human activity; and (3) identify, from the archaeological evidence, any episodes of environmental deterioration or ecological mismanagement severe enough to cause cultural stress and political collapse. For several reasons, some of which I return to in the concluding section, the attempt 96
Searching for environmental stress
97
to answer this challenge raises more questions than answers. In the process, it draws attention to some major flaws in archaeological reasoning, and emphasizes that lowland archaeologists cannot ignore events in the adjacent highlands. To establish an environmental baseline we must learn to distinguish between three different kinds of phenomenon: (1) (2) (3)
Changes in the global climate, that is, worldwide oscillations in temperature and humidity. Purely local events, that is, volcanic eruptions, changes in river courses, tectonic or eustatic changes, and so on, which affect only limited areas. Changes resulting from human activities. At the top of this list are the interconnected phenomena of agricultural intensification and forest clearance, or in other words, the conversion of the natural landscape into an artificial manmade one.
These three factors are difficult to separate in the paleo-record. They may all pull in the same direction (for better or worse), or they may work in opposition. We should also remember that a single phenomenon may have widely different consequences (for example, an increase in rainfall may benefit a marginal desert farmer while causing floods elsewhere). Nor are climatic and human factors independent of each other. The tendency among archaeologists has been to assume that climate (that is, the environment), is the prime mover. However, with the growth of world population, and its increasing capacity to modify the landscape, humankind is in a position to influence regional (and perhaps global) climates (see the simulation by Lean and Warrilow 1989, for the effects of Amazonian deforestation). If today's destruction of forests can influence local rainfall patterns by altering evapo-transpiration, runoff, and albedos, then any archaeological reconstruction must consider the effects of massive forest clearances for agriculture in prehispanic times. Before making broad generalizations, we first need good regional studies which look at the interplay of all factors, and which do not assume that the global climate is the only driving force. Another source of difficulty is the problem of identifying cultural or economic stress from the archaeological evidence. I cannot think of any region of lowland Colombia for which we have adequate data on settlement patterns, population size, prehistoric crop regimes, agricultural productivity, carrying capacity, and all the other statistics required for numerical models. We cannot therefore calculate whether any particular society was coming under stress or not; all we can do is make impressionistic observations. This is not to admit utter defeat. As a rule of thumb in this chapter, I am assuming that populations have a general tendency to grow larger. If this trajectory is interrupted, if growth stops or numbers fall, then some kind of stress (not necessarily environmental) may be the explanation. This approach, based on relative rather than on absolute numbers, and on impressions rather than real figures, is the best we can do in many neotropical regions. Nevertheless, in archaeology as in child medicine, any 'failure to thrive' may indicate a potential crisis.
98
WARWICK BRAY
AGE B.P
720*90 1870*140
6680*230
9590*150
H11111II Quercus
|
H^H
Ililllll Gramineae
Andean/Subandean forest
| Subandean forest
Figure 4.1. Pollen diagram (abridged version) from the Hacienda El Dorado, Calima Region, Cordillera Occidental, Colombia. Analysis by Jose G. Monsalve (after Bray et al. 1987).
Landscape and man in the Cordilleras (7500 BC to the present)
Middle altitudes: the Calima region As a first example, to illustrate the ways in which natural and anthropogenic factors interact to bring about landscape change, I offer a simplified pollen diagram (Figure 4.1) from the Calima region in the western Cordillera of
Searching for environmental stress
99
Colombia at an altitude of just over 1,200 masl. The precise locality is the Valley of El Dorado which, at certain times in its history, was a closed basin containing a lake (Bray et al. 1987). The processes governing the ecology of lake-basin drainage systems are outlined in Binford et al. (1987). The diagram shows the influence of the global climate; for example, in the upward displacement of vegetation belts (marked by the retreat of oak) at the end of the last glaciation, and in a temporary return of cooler and drier conditions (Zone 5B) about 700 years ago. But the pollen history also shows the effects of anthropogenic influence on the landscape. The first maize pollen appears soon after 5000 BC, during the preceramic period (see Pearsall, Piperno, this volume); however, for the next five millennia, the vegetation continues to be dominated by forest. Then, fairly suddenly, and long after the introduction of pottery a little before 700 BC (uncalibrated), we see a major ecological change. At about the time of Christ (the start of Pollen Zone 5 A3) the Yotoco Period begins, with a massive episode of forest clearance. Grasses become prominent; there is a sudden jump in the percentage of maize pollen; and soon afterwards some localized erosion and alluviation can be recognized both in soil profiles on the slopes and in valleybottom sediments (Bray et al. 1988: 24-34). At much the same time a minor "local event" occurred, and also contributed to the restructuring of the landscape. The outlet of the Valley of El Dorado was unblocked, and a former lake drained away through the newly opened outlet. On the now marshy bottomland, the Yotoco people constructed a system of drained fields (see Erickson, this volume). In the new, more open, landscape, the numerous house sites, cemeteries, and road system suggest a fairly large population. The discovery of a few very rich tombs indicates social ranking and by most criteria, the Yotoco people had attained a chiefdom level of organization. The trends established in the early centuries AD were maintained into the Sonso Period (c. AD 1200 until the Conquest), when the population seems to have been at its maximum and most of the land was in use for settlement or cultivation. Finally, with the population loss under Colonial rule, much of the landscape was reclaimed by forest, and remained in that state until the trees were once again cleared and burned in the present century. In summary, we can identify two phases of agricultural activity: (1) an earlier phase (when population levels were low, or farming poorly developed) which minimally affected the natural landscape; and (2) a later one which fundamentally changed it. There were also cultural changes, including the possible replacement of one ethnic group by another at the break between the Yotoco and Sonso Periods, but without any signs of collapse. Population levels (a simple biological measure of efficiency) maintain themselves, and there is no evidence of traumatic stress until the Conquest. Finally, the speedy regeneration of the forest shows that Indian land management did not cause permanent damage to the cordillera soils. I will now argue that many elements of this pattern are also recognizable in the high Andes.
WARWICK BRAY
IOO
Age in yr BP
Chronostratigraphic units of the Colombian Cordillera
Relative temperature fluctuations
Humidity variations
Human influence in the study area
z
5,000-
Y
Holocene period'
X
10,000Late' Late Glacial EarlyLate Glacial
El Abra sladial
\
stadial La CiegaNstadial Susaca stadiol
r
15,000Late
Fuquene Stadial Middle
20,000Early
Figure 4.2. Tentative temperature changes (F), humidity variations (G) and human influence on the natural vegetation (H), in the paramo zone of the Cordillera Oriental, Colombia (after Kuhry 1988).
The high Cordilleras Our most complete evidence comes from a series of pollen studies in the eastern Cordillera, mainly from peat bogs and lakes in the paramo zone above 2,500 m. (Figure 4.2; Hooghiemstra 1984; Kuhry 1988; Schreve-Brinkman 1978; van Geel and van der Hammen 1973). The sequence reaches back into the Pleistocene, but I will pick up the story in the early Holocene, around 7500 BC.
Searching for environmental stress
101
7500-5500/5000 BC. (Andean Pollen Zones IV and V, Kuhry's Holocene X Interval). The climate is somewhat warmer than today. The landscape shows minimal human influence, though there is a single record of maize pollen at Paramo de Pena Negra I, c. 6200 BC. This consists of one maize pollen grain, presumably carried by wind from lower altitudes. 5000-1000 BC. (Andean Pollen Zones VI and VII, part of Kuhry's Y Interval). The climate is at its warmest, and the forest line at its highest. In the paramo zone there are fairly abundant records of maize pollen from about 4650 BC onwards, though Kuhry (1988: 127) believes this pollen to be derived from agricultural activities at lower altitudes on the eastern slope of the Magdalena valley. 1000 BC—AD 150. The date of 1000 BC (3000 BP) marks the start of Andean Pollen Zone VII, a climatic boundary that has been recognized also in the central Cordillera (Salomons 1986) and the Sierra Nevada de Santa Marta (van der Hammen 1979). Temperatures everywhere fall to their present levels, and human influence on the landscape becomes more marked. In the lowlands, near the Laguna de Agua Sucia in the Llanos Orientales, the creation of open savannas took place at about 1000 BC and seems at least partly caused by forest felling and burning (Wijmstra and van der Hammen 1966: 82). In the central and eastern Cordilleras, human intervention is recognizable from the start of Pollen Zone VIII, intensifying during the final centuries before Christ, when agriculture had expanded from the lower slopes to the high intermontane basins. These events are marked at high altitudes by a decline in forest elements, an increase in grassland, presence of maize pollen, and the appearance of the first pottery in the archaeological record. At the Zipacon I rock shelter at 2,500 masl, pottery, cultivated maize, sweet potatoes, and avocados occur in a stratum whose base is dated to 13201b 320 BC (GrN-11, 125; Correal and Pinto 1983). Pottery typology indicates that this deposit continues through most of the first millennium BC. The pollen data show that the environment was still largely forested. The excavators suggest that the cultigens (and one of the pottery types) derive from the Magdalena Valley, the major communication route to the Caribbean. By the mid-first millennium BC, sites with pottery and/or evidence for agriculture are too numerous to list individually (see Bray 1984: 318). AD 150 to the present. With the establishment of villages of pottery-using farmers in the high plains, the predominant influence on the vegetation becomes man, rather than climate. Around the time of Christ, the montane forest all but disappears and is replaced by grassland and Myrica shrub. ChenopodiacaeaAmaranthaceae increase, and there is a rise in the curve of Dodonaea (a pioneer species, and indicator of soil erosion) in many of the pollen diagrams from the eastern Cordillera. This rise in Dodonaea seems an excellent marker for intensive Indian agriculture in these areas (van Geel and van der Hammen 1973: 88).
IO2
WARWICK BRAY
In summary, in Andean Colombia there is evidence of human presence from Paleoindian times onward. Signs of maize cultivation appear in pollen diagrams from the fifth millennium (see Pearsall, Piperno, this volume), but somewhere around the time of Christ (or a few centuries before) there is massive and fairly rapid forest clearance. I suggest that in the Andes, this marks the emergence of a man-made agricultural landscape. As a footnote, we should add that, in all areas for which we have data, there seems to have been no major interruption in population growth. Large populations and intensive farming may have caused some local damage, but not universal or permanent degradation of the environment. In those areas where the Indian population all but disappeared after the Conquest, the soils were in good enough shape to allow the forest to regenerate. This regeneration can be seen in pollen diagrams from Calima, the central Cordillera (Salomons 1986: 152-156) and the Sierra Nevada de Santa Marta (Herrera de Turbay 1985). I make no apology for spending so much time on the prehistory of the highland zone. In the first place, it provides some of the best information that we have. Secondly, ecological events in the Andes have direct repercussions on lowland environments. For the middle Magdalena Valley, Jungerius (1976: 123) argues that widespread Holocene erosion can be attributed to continuing stream incision and human removal of stabilizing vegetation. It hardly needs emphasizing that the great lowland rivers have their headwaters in the Andes, and that water flow and sediment load (two of the critical factors for agriculture in lowland environments) are directly affected by deforestation and soil erosion at higher altitudes. With this preamble, we can now turn our attention to the Caribbean lowlands. Savanna and floodplain adaptations: the Sinu and the Mompos Depression The prevailing easterly airflow along the Caribbean coast of South America produces atmospheric divergence and subsidence that contribute to the neardesert conditions of much of the Colombian and Venezuelan coasts. A short distance inland, lies a broad belt of lowland, with alternating wet and dry seasons (Parsons 1980). Today, these Caribbean lowlands are open savanna, with extensive tracts of river floodplain; however, this has not always been the case. Gordon (1957) has brought together ecological and historical data for the region between the Gulf of Uraba and the Rio Magdalena. His map (Figure 4.3) summarizes the changing distribution of grassland and forest from prehispanic times to the present. Gordon's basic argument is that, irrespective of minor climatic fluctuations, the natural vegetation of this area is broadleaf forest. The savannas, he maintains, are anthropogenic. These resulted from prehistoric agricultural clearance, which caused the formation of iron concretions and subsoil hardpans, which led to the replacement of original tree cover by bunchgrasses and fire-resistant species such as palms (see Piperno, this volume).
Searching for environmental stress
103
100 Km
Figure 4.3. Changing extent of savannas (approximate) between the Gulf of Uraba and the Rio Magdalena (after Gordon 1957). The three different patterns shown in Figure 4.3 thus summarize the effects of changing human populations and modes of landuse during precolumbian, Colonial and recent times. With this, the theme of deforestation enters our discussion once again, and with it the question of how Indian communities of this area adapted to a semiaquatic environment, much of which was flooded for several months of the year. The story involves the construction and abandonment (well before the Spanish Conquest) of the most impressive and largest hydraulic systems of native America. Ancient drained fields (see Erickson, this volume) have been noted in the floodplain of the Rio Sinii (Plazas and Falchetti 1986), but have as yet not been tied into the archaeological sequence for that area (Reichel-Dolmatoff and Reichel-Dolmatoff 1957). Our most complete evidence comes some 90 km further east, from the San Jorge Basin in the Mompos Depression, where a multidisciplinary team is studying long-term cultural development in the context of an unstable and changing environment. The Mompos Depression is a vast interior delta, 6000 km2 in extent, and on average, only 20 masl. In this depression, the waters of the Cauca, Magdalena and San Jorge Rivers converge, overflowing or breaking through their banks and flooding much of the area between the months of April and November. For eight months of the year this is an aquatic landscape, with a constantly changing
104
WARWICK BRAY
complex of shallow flood-lagoons (cienagas), swamps, flood channels (canos), and river courses of various ages. The tectonic depression is a huge sediment trap, receiving the run-off of almost 25 percent of the country. Deposition has averaged 3 to 4 mm of sediment per annum over the last 7,500 years. In the center of the basin, c. 30 m of sediment have accumulated during the last seven millennia, though the figure is somewhat lower at the margins of the depression. In compensation, the weight of sediment has caused subsidence to the extent of 1.8 to 5 m since the time of Christ (Eidt 1984: 73-86; Plazas et al. 1988, 1992). The rate of sedimentation has not been uniform; its variability is linked with the climatic oscillations, changes in sea level, alteration in the course of the Rio Magdalena, and also with the quantity of material carried in suspension by the rivers flowing from the cordilleras. It is this factor that links events in the Andes with the history of the floodplain. High rainfall in the mountains causes increased sediment deposition in the Mompos Depression, and this trend is accentuated by deforestation and mining. In recent memory, forest clearance at the headwaters of the Rio San Jorge has led to accelerated erosion and the silting up of river beds in the lower San Jorge drainage, causing widespread and prolonged flooding (Parsons and Bowen 1966). These events are a pointer to what may have happened during the prehispanic deforestation discussed in the previous section. However, man is not the only influence on this landscape. The climatic changes in the area during the last 3,000 years are shown in Figure 4.4. All over the basin, are bands of peat formed during times of low rainfall in the Cordilleras and of increased dryness in the lowlands. During dry epochs, lake and river levels fall, sediment transport is reduced, flood areas shrink and peat deposition becomes more extensive, backswamps increase in extent, savanna and savanna woodland invade former open marshes, and dry organic soils derived from swamp vegetation form in the man-made canals (Plazas et al. 1988,1992; Wijmstra 1976; van der Hammen 1986a). The presence of three or four well-developed soils separated by thick layers of clay, indicates periods of relative stability (drier episodes) separated by times of rapid sedimentation. Superimposed on these major climatic oscillations, is a weaker cycle with a periodicity of about 120 years, sufficiently marked to produce dark soils in the canal system, but not severe enough to cause widespread peat deposition (Plazas etal. 1988: 64). The major dry periods are due to fluctuations in the global climate. Corresponding episodes have been recognized in the high Cordilleras (van der Hammen 1986a), the eastern llanos (Livingstone and van der Hammen 1978; Wijmstra and van der Hammen 1966), Brazilian Amazonas (Absy 1979), and in the Quelccaya ice cores of Peru (Thompson et al. 1985). The sequence of wetter and drier periods presented in Figure 4.4, may therefore be valid for the whole of northern South America. The problems of coping with this fluctuating environment over the past 3,000 years have been studied in detail for the San Jorge Basin (Plazas and Falchetti
Searching for environmental stress
105
Approximate curve of relative changes in river level AD 0
•ligh
•
1900 -
1 I 1 /
100 1800 200 • 1700 • 300 •
/
1600 400 • 1500 -
Wet period
500 • (^
1400 • 600 -
Malibu occupation No hydralic works
1300 •
Malibu pottery widespread
Dry period
1200 • 800
Disappearance of Zcnu population
1100 • 1000 • 000 • 900 •
Wet period
1100 •
C
800 • 200 •
<^
700 •
300 •
600 -
400
500 •
1500 •
^-. .
t
\ Canals. /
"
Dry period s—--^
400 •
600 •
Great population density. Continuous linear settlement along rivers Massive drainage schemes.
Col 512 (Carate 25)
+
+ Col 515 (Caralc 26)
Wet period
200
1800 -
I
100 0
•
100 •
Population growth
•—
Ancestral Zenu pottery still in use Increase of Zcnu pottery
Nucleated settlements Canals'
y
2000
Disappearance of pottery of ancestors of Zcnfics Appearance and diversification \> of Zenu ceramics /
Zenu expansion. Burial Tumuli. Goldwork. Extensive canal systems.
300 •
1700
Zenu pottery exclusively
2100
Predominance of ancestral Zenu pottery
Dry period
Start of Zenu ceramic tradition ) )
300 2300
+
+
Col 596 (Carate 19)
Relatively wet period
400 • 2400
500
2500
600
2600
^
^
Extremely dry period
700
2700
800
2800
Construction of hydraulic systems
900
2900
1000
3000
1100
3100
1
1200
3200
1300
3300 3400
Dispersed settlement Progressive occupation of seasonally-flooded depression
1400 BC
Tentative curve of flood levels for the lower San Jorge (Cauca/Magdalena) showing relatively humid and dry epoch •+- C"> dates associated with canal u
Figure 4.4. Cultural development and climatic change in the lower San Jorge Basin (after Plazas et al. 1988). 1981,1986; Plazas et al. 1988,1992). The San Jorge drainage lies near the western border of the Mompos Depression, overlooked by rolling savanna uplands some 30 to 100 masl, stretching westwards as far as the Sinii. These uplands form part of the anthropogenic savanna belt, and were virtually treeless at the time of the first Spanish entradas (Gordon 1957). On the floodplain of the San Jorge and its tributaries, Plazas and Falchetti (1987) have mapped approximately 500,000 ha of ancient fields and canals. The creation of this vast hydraulic system is not only for flood control, but also improves the nutrient content, drainage, and aeration of the mounded 'ridged fields' where crops were planted (Eidt 1984: 80-84; see
IO6
WARWICK BRAY
Erickson, this volume). Maize pollen was obtained at one locality along the Cano Carate (Parson 1977); however, phosphate fractionation data from ancient soils suggest mixed cropping of maize, sweet manioc, and other tubers (Eidt 1984). This is also the practice today. The first occupation of the San Jorge floodland has been dated to c. 800 BC, which coincides with the onset of a dry period that lasted for more than 400 years. These immigrants to the San Jorge may have their origins further west, for their simple Granulosa—Incisa pottery is related to that of Momil and Cienaga de Oro, in the Sinii region. Soon after their arrival in the San Jorge Basin, these people began to construct canals, and one of these at Cano Pimienta has a radiocarbon date of 8 i o ± i 2 o BC (GrN-14472). Another canal, associated with an extinct course of the Cano Carate, is dated 330± 80 BC (GrN-14475). A century or two later, there is a date from the habitation platform at Cogollo. It is associated with an early canal system, part of which is stratified underneath later canals, and part now covered by the waters of the Cienaga de los Patos. Therefore, by the time of Christ, hydraulic systems were well established, and population was starting to build up on the floodplains. From the first century AD, there is evidence for nucleated villages of about 600 inhabitants, and a new style of pottery (Tradicion Modelada—Incisa) made its appearance alongside the earlier wares. This marks the beginning of a cultural tradition associated with the protohistoric Zeniies, which was still flourishingin some areas at the time of European contact. In the Bajo San Jorge, the period from the third to the ninth centuries was a time of cultural climax. Settlement was almost continuous along the major waterways, rich burials with elaborate goldwork were deposited in tumuli, and the hydraulic systems were in full operation. Plazas and Falchetti (1987: 498) have counted more than 400 habitation platforms in a sector of 1,400 ha along Cafios Marusa and Barrancuda. They estimate a population density of about 160 people per km2 in this zone. Long canals allowed water to flow through the system more quickly, thereby reducing sedimentation in the canal zone, and increasing deposition in the lower sectors of the basin. In those areas which flooded, zones as large as 1,500 to 2,000 ha were converted into farmland by the construction of short, closely-spaced camellones (ridged planting surfaces). This was not a static system. Rivers changed their channels, and old canals were buried under thick layers of flood sediment, with new construction based on different patterns. Abandoned field and canal zones became settlement areas with house platforms, and remodeling was continuous. For centuries, the social and cultural organization of the Zeniies were technologically adapted to accommodate climatic fluctuations and changing water regimes; however, from the tenth century onwards, archaeology shows a gradual abandonment of the flood zone. This abandonment does not mark the end of the Zeniies as a cultural and ethnic tradition. Remnants of the Zemi population survived and maintained many of their old customs in higher, non-
Searching for environmental stress
107
flooding localities such as Ayapel and Montelibano in the San Jorge Basin, and in the savanna lands of the Sinii drainage. Spanish cronistas describe these late Zenu towns with their temples and funerary mounds, but Colonial sources make no mention of drainage works or hydraulic systems. Once abandoned, the old floodlands remained empty until about AD 1300, when they were reoccupied by new groups related to the protohistoric Malibues of the Magdalena Valley. These people brought with them a new form of pottery (belonging to the Tradicion Incisa-Alisada), and built their settlements on the available high ground, while cultivating the surrounding area but not using canals and camellones. When the Spanish expeditions first saw the area, the floodland itself was almost without inhabitants. The reasons why the Zenu peoples abandoned the San Jorge floodplain are not at all clear. Conquest and replacement can be ruled out, as archaeology shows a progressive and voluntary withdrawal. This abandonment coincides with the onset of the thirteenth-century dry period; however, the system had survived an even more severe one around AD 600. Sudden military or ecological trauma does not seem to fit the case. We may have to think of a more insidious kind of stress: the rising investment costs of trying to feed an ever-growing population in an unstable environment. Perhaps the continuous maintenance and reconstruction of the hydraulic system was simply more than society could bear. Be that as it may, what eventually came to replace the great hydraulic schemes of the Zenues was the low cost, lower investment farming of the Malibii communities. In the San Jorge, what came under stress was a man-made ecological system, not a natural one. The Conquest brought about the usual fall in population, and settlement relocation. Forest returned to the floodplain (as it did to the Cordilleras), and the Mompos Depression was not opened up again for settlement until the nineteenth century, when the modern transhumant cattle economy was developed (Parsons 1980). If current experiments with refurbishing the old field systems are successful (Plazas and Falchetti 1986, 1992: 184; see also Erickson, this volume), we may once more see active camellones in the San Jorge Basin. The Rancheria Valley and the Guajira: a problem of desertification
The strongest case for irreversible environmental degradation does not come from the major floodplain and savanna zones; rather, from a peripheral region, the Rancheria Valley, close to the Venezuelan border. Today, the middle and lower Rancheria zone is an impoverished area, basically an extension of the Guajira Desert, with high temperatures, poor sandy soil, xerophytic scrub vegetation, and no large game animals. Rainfall is no more than 500 to 800 mm per year, most of which falls in the wet season with heavy showers that accentuate erosion. The area is sparsely inhabited; before the opening up of the coal deposits at El Cerrejon, goat-raising and smuggling provided the basic subsistence.
IO8
WARWICK BRAY
Conditions were very different in the past. Archaeological surveys by Gerardo and Alicia Reichel-Dolmatoff in the 1940s (1951), and more recently by Gerardo Ardila (1983, 1984, 1986, 1990), have revealed a long sequence going back to the fifth century BC (Loma Period), and reaching a sort of cultural climax during the Horno Period of the first millennium AD. The Loma and Horno styles together constitute the 'First Painted Horizon' of the Reichel-Dolmatoffs. Related styles are found right across the Caribbean lowlands (Bray 1984) and into Venezuela, where they are included within the 'Tradicion Hornoide' (Ardila 1986, 1990; Tartusi et al. 1984). During the Horno Period, sites in the Rancheria Valley were at their most numerous. Some of these sites were large and permanent villages, with up to 2 m of archaeological deposit. Occupation along the river bank was virtually continuous for several kilometers, but the settlements were also found well away from the river. The presence of milling stones in these villages suggests maize cultivation (impossible under present conditions), and this in turn implies an improved water regime. Dry depressions, which mark former lagoons, indicate that the ancient alluvial terraces and fans were formed by a river much bigger than today's. The deepest archaeological deposits lie directly on top of a humus layer (not the sterile sandy soil of today), and some of the Horno sites are dark terra preta soils, rich in organic matter (see Zeidler, this volume). Pollen analysis (EPAM 1982) indicates a climate more humid than at present, and old land surfaces of Horno age exposed in barranca cuts contain shells of tree-dwelling snails. In the Venezuelan Guajira, too, this period represents a climatic optimum. At the shell mound of La Pitia, the Hokomo occupation (with pottery closely related to Loma-Horno) seems to represent a fairly large and permanent settlement close to a large river, which is now extinct (Gallagher 1976). The archaeological deposit is dark with organic matter. Manos and metates appear for the first time, and may indicate maize cultivation. The inhabitants collected Pomacea and Marisa snails, which are characteristic of slow-moving rivers. In the Guajira and the Rancheria Valley, then, the early centuries of the Christian era were a time of relative prosperity, with better climate, available water, and more fertile soils. In the Rancheria Valley, the Horno Period ended in about the eighth century AD. Horno pottery was replaced by that of the Portacelli Style (belonging to Reichel-Dolmatoff's 'Second Painted Horizon,' and with close relatives in the Venezuelan Ranchoid Styles). Large villages disappeared from the archaeological record. Portacelli sites were fewer, smaller, and more dispersed than their Horno counterparts, and were associated with sandy rather than loam soils. Perhaps environmental deterioration in the Rancheria Valley had already begun. There is support for this view from the later strata of Gallagher's (1976) site at La Pitia. Here, the Siruma Phase (derived from Hokomo) was distinctly impoverished. The Siruma strata are no longer dark and organic. Manos and woodworking axes
Searching for environmental stress
109
are absent. The former river seems to have become a brackish swamp. River snails are replaced by the arid-adapted terrestrial variegated tree snail, and land tortoises become significant food items for the first time. Gallagher (1976) suggests that the transition from Hokomo to Siruma marks the onset of the bleak and inhospitable conditions characteristic of the Guajira today. In the Rancheria, too, conditions were becoming insupportable. There are no Contact-Period sites, and the valley seems to have been all but depopulated by the time of the Conquest, though a few scattered descendants of the Portacelli people held out on the flanks of the Sierra (Reichel-Dolmatoff 1965:121). The exact time of this disintegration is unknown, but the final C14 dates for Portacelli in the Rancheria (from Paredon I; Ardila 1986: 67) fall between AD 1250 and AD 1300. At one standard deviation, the figure could be as early as 1200 or as late as 1350. It may be no more than a coincidence, but the terminal Portacelli dates equate with the dry period which also coincides with the abandonment of the hydraulic systems in the San Jorge Basin (Figure 4.4). I doubt whether simple climatic determinism is an adequate explanation for these events, and we should give serious consideration to the idea put forward by the Reichel-Dolmatoffs in their original report of 1951. They note that imported Tairona sherds occur in late Portacelli sites from the tenth century onwards, and argue that the degradation and desiccation of the Rancheria environment is directly linked with the rise of the Tairona chiefdoms in the adjacent Sierra Nevada de Santa Marta (see also the next section). The Rancheria River has its origin in the Sierra Nevada. Along its upper reaches, above 1,500 m, are several modern Kogi villages, as well as ancient Tairona sites dating from the last five centuries before the Conquest. The southeast slope of the Sierra Nevada is the driest and least steep of the massif. Today, much of it has no forest. Instead, dry savanna (which the ReichelDolmatoffs believe to be anthropogenic) extends up to 2,000 m. All the main valleys of the sierra have Tairona agricultural terraces. The authors suggest that the irreversible damage to the Rancheria environment began with forest clearance in the Sierra Nevada, by a growing population of Tairona colonos. This eventually led to savanna formation, soil erosion, and the drying up of rivers, and terminated with the nibbling of remnant vegetation in the Rancheria by rabbits and, more recently, by goats. In contrast to most other areas, where landscape recovered once human pressure was removed, the Rancheria Valley and the Guajira have suffered permanent damage. Reichel-Dolmatoff's theory has its attractions. It relies on the interplay of human and climatic factors, and fits with my contention that the lowlands and the sierras are part of a single macro-system. It is also testable, and there is a need for a multidisciplinary program in the southeastern Sierra Nevada. Unfortunately for the present argument, research in the Sierra has concentrated on the northern and western sectors, which have a slightly different history.
HO
WARWICK BRAY
The Sierra Nevada de Santa Marta: the rise of the Tairona chiefdoms
The northern slope of the Sierra Nevada is ecologically different from the southeastern side. It is steeper, more humid, dissected by river valleys with patches of fertile land, and is forested almost down to sea level. There is no broad coastal plain and, because of the juxtaposition of mountains and coast, these two areas share in a common cultural tradition. The drier episodes around AD 450— 600 and AD 1200—1250 (Figure 4.4) have also been recognized in the sierra, as has the Little Ice Age of AD 1600-1850 (Oyuela 1987b; van der Hammen 1986b). The early stages of the story belong to the coast, with the establishment of communities using pottery in the Malamboid Tradition (Langebaek 1987a). At these localities, as in other Malamboid sites from Venezuela to the Magdalena, budares are present, and grinding stones absent. This is usually taken as evidence for bitter manioc, rather than maize, as the staple crop. The Malamboid occupation is followed, on the coasts north and west of the Sierra Nevada, by the Nahuange Phase, which takes up most of the first millennium AD (Langebaek 1987a, 1987b; Oyuela 1986, 1987a, 1987b). Maize now seems to replace bitter manioc as the main storable crop, though sweet manioc and other tubers were still present in the Caribbean lowland repertoire at the time of European contact (Bray 1984). Chronologically, the Nahuange Phase is equivalent to the Horno Period on the inland side of the Sierra. Both these ceramic styles belong with the 'First Painted Horizon,' and Horno sherds have been found on Nahuange sites. In other respects, however, the Nahuange Phase is clearly proto-Tairona in its level of development, and range of artifacts. This shows most clearly at the type site, Nahuange I, excavated by Alden Mason (1931-39). The site is a burial tumulus surrounded by a stone kerb and containing a stone-lined grave. Other burials were deposited in the body of the mound. Contents of these burials included pottery, stone figurines, approximately 8,000 beads, and 30 winged pendants (many of nephrite), as well as gold items. A radiocarbon date of AD 310 ±70 (OxA-1577) was obtained from the casting core of one tumbagua figure. This item is transitional between the International Style (Cooke and Bray 1985) and the full Tairona Style, and several of the other metal objects also show Tairona traits. During the sixth to tenth centuries, other Tairona characteristics appeared at coastal Nahuange sites. Population appears to increase and there are hints of a two-level hierarchy. Some localities had megalithic structures, including roads, canals, and stairways. There was a progressive development of ritual and ceremonial paraphernalia, and pottery became more standardized and began to assimilate the elements that constitute the mature Tairona Style. At this point, the Sierra Nevada was colonized, apparently for the first time, beginning with the lower slopes and spreading to progressively higher altitudes. This pattern is reflected by the radiocarbon chronology. Between 360 and 500 masl, there are dates of AD 580 ± 120 (Beta-3563) from Las Animas, and 660 ± 90
Searching for environmental stress
111
(Beta-13,948) from Frontera. The oldest high sierra dates are AD 970 ± 260 (GrN11,887) from La Estrella at 670 to 950 masl, and AD 1000 ±70 (Beta-12,994) from Buritaca 200 (La Ciudad Perdida) at 950 to 1,300 masl. From AD 1000 onwards, the Sierra Nevada rapidly filled up with settlements and remained densely populated until a century or so after the Spanish Conquest (Cadavid and Herrera de Turbay 1985). These archaeological events are recorded in a pollen study carried out by Herrera de Turbay (1985) at La Estrella and Buritaca 200. The early stages of her diagrams show natural forest, with now human intervention. Then, in levels corresponding with the Tairona occupation, Gramineae and Compositae increase (though tree pollen does not disappear entirely), and cultivated maize, avocado, and perhaps yucca, make their appearance. Finally, with the depopulation of the sierra that followed the suppression of the last Tairona rebellions in about AD 1600, the forest regenerated, though its components were not quite the same as before. From this evidence, Herrera argues that six centuries of Tairona farming caused no permanent damage to the soils of the north slope of the Sierra Nevada (no matter what may have happened in the more vulnerable southeast sector). This contrasts with the activities of recent, non-Indian colonists, whose indiscriminate cutting and burning is already causing erosion and environmental degradation. Concluding remarks This collection of case studies will disappoint anyone who hoped for a general model applicable to the whole of Caribbean Colombia. Nor will it give much comfort to those sentimentalists who believe that "primitive peoples" never make ecological mistakes (Ellen 1986). The interplay between all sorts of environmental factors (global, local, and human) is complex and variable. The attempt to find a single all-embracing explanation for change may be as futile as the quest for the philosopher's stone. In reaction — perhaps overreaction — against deterministic models which generally emphasize the role of the global climate, I have given more attention to changes brought about by man, in particular by deforestation and its secondary consequences. In doing so, I have tried to show that neither lowland nor highland prehistory should be studied in isolation. In spite of all the regional variability, it seems possible to identify certain widespread trends. After millennia during which man had relatively little effect on the landscape, there is a major change somewhere around the time of Christ, though it is not synchronous everywhere. At this critical time, a number of things happened more or less simultaneously. With population growth, agriculture became more intensive (and maize may largely have replaced bitter manioc as the staple Caribbean crop), large-scale forest clearance was initiated in many parts of Colombia, and a "managed" agricultural landscape replaced the natural one. At about the same time, ranked societies and chiefdoms emerged, and some of these
112
WARWICK BRAY
societies invested a good deal of labor in agricultural works. All these phenomena are interconnected, and none of them can be singled out as the prime mover. From this time onwards, man is a major ecological factor in his own environment, and the means exist for leaders to take political decisions which may or may not be influenced by climatic change. To model this complexity, some kind of systems approach is required; prime mover explanations are simply inadequate. From the archaeological data it is clear that, in some places and at certain times, the process of population growth and socio-cultural evolution was interrupted, but I am not convinced that we can distinguish between environmental stress and other kinds of pressure. The traditional approach to the question has been to line up all the column-sequences side by side and to see whether a reorientation in the cultural record coincides with a wetter or drier episode in the global climate. This produces a correlation, not an explanation. The timing may be nothing more than a statistical coincidence. There are, after all, cultural breaks that did not coincide with any obvious climatic change, and there are environmental fluctuations that produced no recognizable effects in the archaeological record. A contributory difficulty is poor chronological control, in particular for severe climatic events that are of short duration. In Colombia (where we have neither ice core information nor tree ring dates), existing radiocarbon chronologies are coarse-grained and imprecise. Lining up the columns is not, however, the main problem. Until archaeologists develop some kind of bridging theory to show how we can identify climatic or environmental stress in the archaeological record, and until we can specify the mechanisms by which environmental change brings about cultural change, or vice versa, we cannot even begin to tackle questions of causality. In the end, what this chapter emphasizes is the lacunae in our knowledge and the weakness of our explanatory models. Note An earlier version of this chapter was delivered at the symposium 'Non-imperial polities in the lands visited by Christopher Columbus' (Smithsonian Tropical Research Institute, Panama). For information and helpful advice, I am indebted to Gerardo Ardila, Mark Brenner, Luisa Fernanda Herrera, Jose Oliver, and Peter Stahl.
5
"Doing55 paleoethnobotany in the tropical lowlands: adaptation and innovation in methodology DEBORAH M. PEARSALL
Archaeological research in the New World tropics has many different foci. Investigating the nature of subsistence practices in this diverse region, especially the transition to agriculture,figuresprominently among these. However, recovering direct evidence of subsistence (for example, the remains of plants used by people), is a formidable challenge facing archaeologists who work in the neotropics. Preservation of macroremains (seeds, tubers, wood, corn cob fragments and the like) is limited to charred materials in all but the most arid settings. Even when charring occurs, macroremains may be highly fragmented due to expansion and contraction of soils, making their detection and recovery difficult during excavation. The problems of identifying such fragmented materials are rendered more complex by the high species diversity of the tropical flora. This necessitates a large botanical comparative collection and the occasional application of specialized identification techniques, such as scanning electron microscopy (Pearsall 1989). Problems of preservation, recovery, and identification of botanical remains are not unique to paleoethnobotanists working in the neotropics, but the environments of the moist lowlands seem to "conspire" to create the worse possible conditions for recovering subsistence data. As Mangelsdorf remarked in his review of Agricultural Origins and Dispersals, in which Carl Sauer (1952) proposed the riverine zone of the moist tropical forest as a likely hearth of agricultural origins: His two principal hearths occur in regions where few archaeological remains have so far been found and where the climate almost precludes the long-time preservation of herbaceous cultigens. . . Indeed if one sought, as an exercise in imagination, to design a completely untestable theory of agricultural origins and dispersals, it would be difficult to improve upon this one (Mangelsdorf 1953). Donald Lathrap (1973a) objected to this characterization of the potential for studying agricultural origins in the lowland tropics. It is largely as a result of his influence that I have focused on the problems of recovering botanical data needed
114
DEBORAH M. PEARSALL
to address this issue and other questions of plant-people interactions in the neotropics. Is it possible to recover botanical data from the moist lowlands? Can wellidentified remains be recovered in quantities sufficient to assess their contribution to diet? In this paper I argue that the answers to these questions are "y es ?" but only if steps are taken during project planning, excavation, and analysis to maximize the potential of the botanical data. The key issue is to utilize multiple archaeological indicators of plant use. The necessary correlates are to build an adequate comparative collection, and design a sampling strategy in consultation with a paleoethnobotanist to maximize recovery of plant remains. I focus here on utilizing multiple indicators, namely, macroremains, pollen, and phytoliths, and adapting those techniques to the lowlands. Issues of sampling and comparative collections will be discussed throughout. To illustrate how our understanding of past human—plant interrelationships in the neotropics has been enhanced by application of good macroremain recovery techniques and the analysis of pollen and phytoliths, I present a case study — the antiquity and importance of maize in northern South America. Utilizing what macroremains there are: recovery, sampling, and identification Paleoethnobotany is the study of the interrelationships between humans and the plant world through the archaeological record. For many, this means the study of the dried, charred, or waterlogged pieces of plants discarded by their users and recovered during archaeological excavation. In order to distinguish them from microremains or microfossils (for example, pollen and phytoliths; refer to Piperno, this volume), these materials may be referred to as macroremains (Pearsall 1989). In my view all these are archeobotanical remains; the data of paleoethnobotany. However, it is useful to consider the larger materials separately, since their recovery is usually in the hands of the field archaeologist, and they often hold the greatest potential for genus or species level identification and quantitative analysis. Recovering macroremains Paleoethnobotany began in the nineteenth century with the study of wellpreserved, conspicuous botanical materials from Egyptian tombs (Kunth 1826), lakeside Swiss villages (Heer 1866,1878), and Peruvian mummy bundles (Saffray 1876) (refer to Pearsall 1989; Renfrew 1973; Towle 1961 for further discussion). For nearly a hundred years, research remained focused on conspicuous finds, including desiccated remains from the Mediterranean region, the Near East, the American southwest, dry caves or rockshelters from a number of regions, and on waterlogged materials from Europe. Publication of Excavations at Star Carr by British archaeologist J. G. Clark (1954) convinced many of the importance of
"Doing" paleoethnobotany in the tropical lowlands
115
biological remains for archaeological interpretation. The late 1950s and the 1960s saw increased emphasis by both American and European archaeologists on reconstructing subsistence and environment, and consequently on recovering and analyzing macroremains. An outgrowth of this interest in biological data recovery was the "flotation revolution." Elsewhere (Pearsall 1989) I have reviewed the history of flotation, and described methods currently in use (manual systems, such as IDOT; machine-assisted systems, such as SMAP; and froth flotation systems; refer to Wagner 1988 and Watson 1976 for other perspectives on flotation). Water flotation utilizes the differences in specific gravity between botanical materials and soil matrix to separate and concentrate botanical remains from soil. All sizes of materials, not just conspicuous larger finds, are recovered. Wetting desiccated plant remains can lead to their deterioration; therefore, fine sieving is substituted for flotation in arid settings. Fine sieving is also useful for concentrating waterlogged botanical materials which are too heavy to float. The impact of recovering all sizes of plant remains on our understanding of plant-people interrelationships in the New World tropics cannot be overstated. Many examples could be cited; the following is an interesting case from coastal Peru. Here the problem for understanding the role of various plants foods in diet is, ironically, an abundance of conspicuous plant material preserved in the arid coastal desert. This abundance makes fine sieving seem unnecessary, yet without recovering material of all sizes, the record is biased towards larger remains. Therefore, the relative importance of local crops, wild resources, and maize cannot be assessed with confidence. In a joint analysis of plant remains from El Paraiso, Ojeda and I generated almost mutually exclusive plant lists from remains recovered in situ or by screening and by fine sieving (Pearsall and Ojeda 1988; Quilter et al. 1991). Fine sieving adds an entirely new suite of plants to the record (small-seeded annuals such as grasses, chenopod, amaranth, and members of the Solanaceae, Portulacaceae, and Aizoaceae, among others; Pearsall and Ojeda 1988; Quilter et al. 1991; Umlauf 1988). Similar taxa are recovered when coprolite samples are analyzed (Jones 1988). Sedge has also been reported in more recent analyses (Bird et al. 1985; Pearsall and Ojeda 1988; Quilter et al. 1991; Umlauf 1988). Is this a small seed and sedge horticultural complex, to which introduced maize and tubers (achira, sweet potato, jicama, manioc) were later added? As more researchers employ techniques to minimize size bias in botanical data, we will be able to evaluate this, and other hypotheses. Advances in understanding neotropical plant-people interrelationships brought about through flotation and fine sieving, are too numerous to summarize here. Success stories include coastal Ecuador (Pearsall 1988; 1992), Parmana in Venezuela (Roosevelt 1980), Cuello and Pulltrouser Swamp in Belize (Miksicek 1983; Miksicek et al. 1981), Copan in Honduras (Lentz 1991), and various sites in Panama (C. Smith 1980; see also Piperno, this volume). The quality of botanical recovery from the lowlands is often disappointing, however. Flotation was
Il6
DEBORAH M. PEARSALL
developed in the United States and England, and all commonly used systems evolved during archaeological research in the American midwest and southwest, the British Isles, and the Near East (Pearsall 1989). Flotation systems must be adapted to conditions in the tropical lowlands, not merely transported there from the temperate zone (see Siegel, this volume, for similar points regarding site excavation strategies). The key for adapting flotation successfully to the tropical lowlands is to evaluate the soil matrix and alter recovery technique accordingly. Loamy or sandy soils are easiest to float. The higher the clay content, the more difficult it is to separate botanical materials from soil, since dry, clay-rich soils will often not dissolve (disperse) during the 15-20 minutes of standard flotation. There is no monopoly on clay soils in the tropics, but riverine settings are often of special interest there. Alluvial soils may be very high in clays and metallic oxides, which bond to clays, further impeding dispersion. In addition, clay-rich soils expand and contract dramatically upon wetting and drying, which breaks up charcoal. The problem is therefore two-fold. Clay soils increase fragmentation and destruction of remains, reducing quantity per soil volume, and what remains can be difficult to extract from the soil matrix. The solution is to increase the volume of processed soil (to increase overall recovery), and to disperse clays (to release remains from matrix). To evaluate what recovery technique should be used, test samples should be floated. If clay peds do not break up during flotation, then samples should be deflocculated (soaked in water with a dispersal agent such as sodium hexametaphosphate, sodium bicarbonate, or hydrogen peroxide), and processed by fine sieving or flotation with an IDOT system followed by chemical flotation (refer to Pearsall 1989 for sample procedures). Soil volume should be increased until approximately twenty pieces of wood charcoal are present in each test sample. In our research in the Jama River valley of Ecuador, we increased soil volume from 10 1 to 40 1 per sample (nearly the entire contents of a test pit level) to get adequate samples (Zeidler and Pearsall 1994). Soaking samples to disperse clays waterlogs charcoal, thereby impeding flotation. Dense plant remains, such as palm fruits and certain types of wood, also resist recovery by flotation. Any materials which sink during flotation and are smaller than the mesh used in the bottom of the flotation apparatus can be lost. This loss can be reduced by using a fine mesh (such as 0.5 mm) in the flotation barrel insert. Non-buoyant remains can then be siphoned off the insert screen during flotation using the Gumerman and Umemoto (1987) procedure, handsorted from heavy fractions, or recovered by chemical flotation. The three types of flotation systems (manual, SMAP-type, and froth) have all been successfully adapted to the moist tropical lowlands. The original froth system (Crawford 1983; Jarman et al. 1972) utilizes bubbling air to float materials. It does not perform well in clay soils, and should be modified to incorporate water flow (see Pearsall 1989: Figure 2.2 for such a machine in use at
"Doing" paleoethnobotany in the tropical lowlands
117
the Salango site in Ecuador). SMAP machines, and the similar water-sieve system used in Britian, have been widely used in the tropics. Designed to handle clayey soils (Watson 1976), these machines rely on the pressure of water bubbling from beneath the sample to separate and raise botanical materials, and work well in all but the most clay-rich soils. If flotation tests reveal that clay peds are not breaking up, the it in barrel insert screen (SMAP) or 1.0 mm flexible mesh (water-sieve) can be replaced with finer screen. This will catch small remains which become waterlogged during extended flotation, or allow the system to be used for finesieving treated samples. An alternative to replacing the insert screen for processing wet samples is to enclose the soil sample in a mesh bag, which is suspended in the flotation tank and "washed" until soil is removed. The residue is then chemically floated (Neal Lopinot, personal communication 1992). If cost or transportation difficulties mitigate against machine-assisted flotation, a manual system of the IDOT-type (Wagner 1976, 1977; see also description in Pearsall 1989) works well, either for standard flotation, or as afinesieve to process treated samples. The disadvantage of a manual system is that large samples, such as the 40 1 samples floated in the Jama project, can require an hour each to process. Operator fatigue leads to somewhat lower recovery rates. Poppy seed recovery in Jama samples is nearer 80 percent than the usual 90-100 percent reported for IDOT systems (Wagner 1982). As a final note on enhancing recovery of macroremains in the tropics, it is important to know when not to use water flotation or water sieving. A test of the effect of floating desiccated (uncharred) remains showed that some materials were damaged. Maize cupules became soft and weak, beans became discolored and began to shed seed coats and mold growth began (Pearsall 1989: 79-81; see also Andrews 1990: 22, for effects on micromammalian dental remains). When charred remains are very dry, as in samples taken from rarely wetted soils, they may disintegrate upon contact with water (Jonathan Damp, personal communication, 1985, concerning flotation of soil from deep strata at the Real Alto site, Ecuador). Under these conditions, fine sieving dry matrix, using a 0.5 mm screen to catch small seeds and fragments, may be preferable to flotation. The key to good recovery is to process test samples by flotation early in a project, evaluate recovery, and modify the procedure if needed. Sampling The title of this section, utilizing what macroremains there are, reflects the fact that remains of seeds, fruits, and tubers may be scarce in neotropical sites. As discussed above, this problem can be remedied somewhat by increasing sample sizes and using efficient recovery systems, but sometimes there are simply few remains present. Yearly cycles of wetting and drying of soils can lead to near destruction of charred botanical remains; even ceramics may become soft and weathered (see also Stahl, this volume for effects on bone remains).
Il8
DEBORAH M. PEARSALL
Table 5.1. Percentage occurrences of selected remains from the Pechichal site (M3B4-011, Feature 5). Elem. C Elem. F Elem. 1 Elem. M. Elem. R Totals Fl. 89-23 Fl. 89-17 Fl. 89-15 Fl. 89-20 Fl. 89-34/35 Liters floated Wood, Ct. Wood, Wt. Corn, Ct. Corn, Wt. Bean, Ct. Bean, Wt. Cotton, Ct. Cotton, Wt. Palm, Ct. Palm, Wt. Dense, Ct. Dense, Wt. Porous, Ct. Porous, Wt. TOTAL Ct. TOTAL Wt. Ct./liter Wt./liter
14
66.9 89.2 25.2 6.7
61.2
6 6.7
51
4.2
21
9-2. 7
80.7 79.2
3-i
4-5
2.2
8.1 2
2.2
64.5
107
21.6 22.3
29.2
316
43.1
46.2
32-5 13-7
15-5 2-7
29.2 7.8 9.8
3
i-5
5-9
4-9
1.4
7-9
0.6
1.8
2.1
7.8 4.4
4-3
3-i
0.5
3i
127
4.3686 9.07 0.31
98
0.9837 4.67 0.05
8 7.8 779
9.7804
129.83 1.63
i-3
11.9
7-9 5i
0.3969
34 0.26
5^ 1.1887
2-5
O.I
24.5
9.187
9
1
1
863
0.8 0.3
1.6
2
9.163
69.8
370
6.1033
0.081 0.33 22
0.1061 134
I-577I 1425 21.6329
5-74 0.09
Two approaches to sampling can improve macroremain recovery: blanket sampling and opportunistic sampling. The commonly recommended practice of blanket sampling, in which flotation samples are taken from every level in every excavation unit, and from all features, should be followed rigorously in the lowland tropics. This will insure that "good" contexts for botanical samples, often missed during excavation, will be sampled. When contexts are encountered that do contain obvious plant remains (charcoal concentrations, hearth areas), extra soil should be taken for flotation. For example, if one encounters larger botanical materials during excavation, taking a flotation sample from that location, in addition to the "general sample" for that level, will probably yield other, less visible remains. During survey in the Jama River valley, Zeidler and crew discovered a large, bell-shaped pit eroding out of the stream side near Pechichal (Zeidler and Pearsall 1994). Charcoal was obvious in the soil. The pit was excavated and large quantities of soil were floated, yielding a wealth of botanical data (and faunal remains, refer to Stahl, this volume). As Table 5.1 illustrates, this pit yielded high counts/weights of material per liter of soil and a diverse assemblage of materials.
"Doing" paleoethnobotany in the tropical lowlands
119
Opportunistic sampling of this feature provided a glimpse of early Jama-Coaque subsistence unparalleled in our sample, and a valuable guide for interpreting phytolith assemblages (see below). Role of the comparative collection Identifying fragmentary remains of seeds, fruits, wood, tubers, and other macroremains requires direct comparisons of unknown material to known comparative specimens. While a necessary component of all paleoethnobotanical research, a good comparative collection is especially important in the tropics. Species diversity is high. There are more taxa to compare to unknown remains than is the case when working in temperate zones. For example, the North American wood collection in the University of Missouri Paleoethnobotany Laboratory (for use in the midwest) contains some sixty-five specimens. In comparison, the collection for west Ecuador, a much smaller geographic area, contains 150 taxa, and there are many unknown remains in archaeological samples from this region. High diversity also makes it difficult for an anthropologically trained paleoethnobotanist to build a good comparative collection. Numerous plant collecting trips may yield only a fraction of the specimens needed; some regions lack complete floras, and few illustrated keys exist. There are several strategies which enhance the quality of comparative collections, and thus final results. The ideal situation is to work with a professional botanist in order to build a regional comparative collection and document present-day vegetation. If this is impossible, collecting with a local person known for their folk knowledge of plants can result in good coverage of useful plants, and specimens identified by common names. If care is taken to collect flowering and fruiting specimens (refer to Pearsall 1989 for collecting and pressing procedures), one set of voucher specimens can be donated to an herbarium in exchange for specimen identification. Herbarium collections can be a valuable source of comparative materials, especially seeds, fruits, leaf material (for phytolith analysis), and flowers (for pollen). Material can be sketched or photographed, and one may also request permission to sample extra material curated with the voucher. Specialists in particular plant groups, for example wood, grasses, or cultivated taxa, are usually happy to assist with identifying specimens at the species or varietal levels. A good comparative collection greatly facilitates identification of fragmentary remains. It is important to bear in mind, however, that material may not be identifiable beyond a very general level (for example, to the botanical family) due to the high redundancy of form within the group, or because of diagnostic characters missing from fragmentary specimens. Precision of identification should always be made clear: is a seed a cheno-am {Chenopodium or Amaranthus), Chenopodium sp., or the cultivated Chenopodium quinoa? The high
I2O
DEBORAH M. PEARSALL
species diversity of the lowland tropics necessitates a conservative approach to identification. Finding pollen in the tropical lowlands Palynology is a widely applied technique for reconstructing past vegetation and human subsistence patterns. From late nineteenth- and early twentieth-century applications for reconstructing Quaternary period vegetation and climate in northern Europe, palynology has spread to all regions of the world (refer to Birks and Gordon 1985; Bryant and Holloway 1983; Dimbleby 1985; Faegri and Iversen 1975; Holloway and Bryant 1986 for overviews of the technique). Archaeologists were quick to take interest in pollen analysis; by the 1960s analysis of soils from habitation sites was widespread. In the New World, archaeological applications were first focused on the southwestern United States, and soon spread to other regions in North America, and the neotropics. Palynological research in Latin America was first focused on vegetational reconstruction. The work of van der Hammen and associates in northern South America (van der Hammen 1963, 1966, 1981; van der Hammen and Gonzalez i960, 1964, 1965; van der Hammen et al. 1973; van Geel and van der Hammen 1973; Wijmstra 1967; Wijmstra and van der Hammen 1966), Wright and colleagues in Peru (Hansen et al. 1984; Wright 1983), and Bartlett and Barghoorn in Panama (Bartlett and Barghoorn 1973), illustrated that pollen was well preserved in lake and bog sediments in the tropics, and could be used to investigate the response of tropical vegetation to the global climatic changes of the Pleistocene. As Piperno's overview of recent research (this volume) demonstrates, sedimentary pollen analysis continues to contribute to our understanding of past vegetation, and of human impact upon that vegetation. Applications of pollen analysis at archaeological sites in the lowland tropics have proven less successful than lake studies, due in large part to preservation problems. I will focus the remainder of my remarks on the problems of archaeological palynology in the lowlands, and how archaeologists can maximize the potential of these data. Pollen preservation Pollen grains are produced in the anther and consist of three concentric layers. It is the outermost layer, the exine, which often preserves because it is composed of sporopollenin, one of the most resistant natural organic substances known. Along with shape and size of the grain and number and shape of apertures, the surface features of the exine are used to identify pollen (Faegri and Iversen 1975). A number of factors determine how long pollen grains are likely to survive once they are released from the flower and deposited in soil or in lake sediments (Bryant and Holloway 1983). Mechanical degradation takes a high toll on pollen
"Doing" paleoethnobotany in the tropical lowlands
121
grains, especially those which fall on land surfaces. The surface of the exine is abraded by soil particles, and once the exine is breached, fungi and bacteria more easily destroy the grain. The exine is also weakened by alternating periods of wetting and drying of soil. Chemical destruction of grains occurs in either soil or sedimentary environments. Preservation is enhanced in acid settings (pH less than 7) and in reducing environments (low oxidation potential, Eh). The more alkaline the soil, the more likely is pollen destruction (contrasted with preservation of osseous remains; see Stahl, this volume). Not all types of pollen are equally susceptible to chemical destruction, however. Even in base-rich soils some resistant types survive. A degraded pollen assemblage is characterized by dominance of resistant types and low counts (Dimbleby 1985). Biological agents (fungi, bacteria, earthworms) also destroy pollen. Fungi and bacteria cause extensive destruction of pollen unless their activity is checked (for example, in waterlogged sediments or in very dry sediments). If samples taken from waterlogged sediments are allowed to dry, or dry sediments wetted, bacterial and fungal activity resumes and pollen may be destroyed. Waterlogged sediments should be stored at 5°C, whereas moist soil samples can either be refrigerated or dried (oven drying at 90—ioo°C, or sun-dried) (Dimbleby 1985). Sampling to maximize recovery As should be clear from the above discussion, habitation sites in the lowland tropics may often occur in settings poor for pollen preservation. For example, sites in flood-plains where alternate wetting and drying of soils routinely occur, are likely to produce only degraded assemblages. This was Zimmerman's experience in analyzing samples from sites tested in 1989 in the Jama River valley (Zimmerman 1994). All fifty-four samples analyzed from test excavations in four alluvial sites yielded degraded assemblages. These were characterized by: significant numbers of fungal spores; many indeterminate (damaged) grains; durable pollen types; and low concentration values (pollen abundance per unit of sediment). As a result of this experience, we sought better pollen preservation environments in the valley. Knowing that waterlogged sediments often produce good samples, Zimmerman extracted cores from a swamp and a small pond. No large lakes are present in the study area. Unfortunately, this strategy proved unproductive as both the swamp and the small pond dried up with enough regularity that pollen was destroyed by fungal and bacterial action (Zimmerman and Bryant 1991). Since both areas were inundated when sampling was undertaken, there was no way to predict this outcome. As demonstrated by the work of Delcourt and colleagues (Delcourt et al. 1986), ponds located in alluvial settings can produce good samples, and should be considered prime sampling locations. As Piperno (this volume) stresses, off-site sampling locations can provide data which
122
DEBORAH M. PEARSALL
complement those recovered from cultural contexts, and may often be the best data available. Wiseman's pollen analysis at Pulltrouser Swamp in Belize (Wiseman 1983) illustrates another potentially productive environment for pollen sampling: raised fields (agricultural fields constructed in swampy terrain by mounding up earth in long rows or in platforms above the level of water inflow, see Bray, Erickson, this volume). In this study, four microenvironments within the field system, each with limitations and advantages for pollen deposition and preservation, were recognized. These included: canal bottoms (the post-agricultural sequence); canal sides (a poor environment, subject to oxidation); raised-field structure (matrix; reflecting plants grown on the fields in the past); and raisedfield surface (modern vegetation). It was hoped that the clay and sascab matrix of the fields would hold water by capillary and ionic means, thus preventing alternating wetting and drying. This was the case in some, but not all strata in the fields. Six of thirteen samples taken from a field profile yielded enough pollen to count. Sufficient data were recovered, however, to reconstruct vegetation at the time of field construction and to identify weeds and crops grown on the field, including maize. In general, if waterlogged sediments exist on a site (strata below the current water table, inundated areas, wet agricultural fields, canals or ditches), these may be good environments for pollen preservation, and should be sampled. Preservation is also enhanced if pollen is protected from abrasion and wetting and drying (for example, by being enclosed in a pottery vessel or protected under a rock slab). In addition, Dimbleby (1985) observed that pollen from samples taken under earthworks was often well preserved. Any rapidly buried surface may produce better pollen samples. Even if samples from most contexts have poor pollen preservation, opportunistic sampling of protected environments and localized wet areas on sites may eventually yield results. Since it is nearly impossible to predict in thefieldif a given context will produce good pollen preservation, the best overall strategy is to submit samples for testing early in a project. Then, if preservation is poor, sampling should focus on contexts with the highest chance of success, and on testing new sites. Off-site sampling in lakes, ponds, and swamps should be a high priority (see Piperno, this volume). Finally, as we learned during the Jama project, even if ancient pollen proves elusive, worthwhile contributions can still be made by studying modern surface samples to develop vegetation analogs, and by collecting comparative samples to build a pollen key. The role of phytolith analysis in tropical paleoethnobotany Applications of phytolith analysis, the study of plant opal silica bodies, have grown dramatically in archaeology since the early 1970s, when the modern research period began (for historical overviews refer to Bryant 1993; Pearsall
"Doing" paleoethnobotany in the tropical lowlands
123
1989; Piperno 1988a, 1991). This has been especially true in the moist tropics, where before the advent of phytolith analysis, knowledge of subsistence was often limited by poor preservation of organic remains of plants. Unlike macroremains or pollen, phytoliths are well suited to the environments of the tropical lowlands. One could argue that archaeological phytolith analysis has matured as a discipline, "come into its own," in the lowland tropics of the New World. The nature and production of phytoliths are discussed in some detail by Piperno in this volume. A number of other recent reviews of phytolith production patterns and research approaches also exist (that is, Pearsall 1989,1993; Piperno 1988a, 1991; see also papers in Pearsall and Piperno 1993; Rapp and Mulholland 1992). I will focus on how phytolith analysis may be integrated into an overall paleoethnobotanical research design. This topic is usefully approached from two perspectives: (1) the role phytoliths play in subsistence reconstruction in combination with other botanical data; and (2) the strengths and limitations of phytoliths when they are the sole source of data in contexts where organic preservation is lacking. Integrating phytolith data An effective approach for maximizing the potential of phytolith data is to sample the same archaeological contexts for phytoliths, flotation, and pollen. Soil phytolith samples are small and easily transported and stored for later selection of proveniences for analysis. Phytoliths are not harmed by refrigeration or low temperature drying; samples taken for pollen analysis can be divided, with half reserved for phytolith processing (refer to Pearsall 1989 for field sampling procedures). If pollen preservation is good, than the same proveniences should be selected for analysis of phytoliths and pollen. Typically many more flotation samples than microremain samples are studied, but macroremain data should be examined for all contexts selected for microremain analysis. This approach to field sampling and selection of phytolith samples for analysis maximizes the potential of phytolith data. Gaps in the record of plant utilization, resulting from a lack of charring or post-depositional destruction of remains, can be filled using phytoliths. In turn, problems of redundancy and over- and underrepresentation of phytolith types can be addressed by comparison to the other data. These points may be illustrated by examining the Pechichal pit discussed above. A comparison of Table 5.1 (macroremains), and Figure 5.1 (diagnostic phytoliths), from the same strata in this pit, illustrates clearly what phytolith data bring to interpretation of subsistence. First, phytoliths document edible plants which left no clearly identifiable macroremains (for example, arrowroot, Marantaceae and achira, Canna). "Porous" material, probably tuber remains, was found, but precise identification is difficult. In addition, the phytolith record illustrates that debris from various tropical trees (such as Cordia, Chrysobalanaceae, Croton)
H
CTQ
ft tr en £-
3 «^ O rt
5 3 ao
« b
C
n
-•
Canna
Aracaceae, conical Marantaceae
Aracaceae, sphere
Gramineae
Total (counts)
• "
inn
Spongespicules
Diatom
Cystoliths
Euphorbiaceae Tapura Cleidon castanneifolium
Croton fraseri
Epidermal appendages
Irregular faceted
Folded spheres
Rugulose spheres
Trichomanes
Cordia lutea
*" Cordia
Chrysobalanaceae
~~ Heliconia " " ~
m
" • i
—
TIT]
CD CD CD CD CD CD
E3Q3EDQ3 03 ED
"Doing" paleoethnobotany in the tropical lowlands
125
decayed in the pit, or were deposited there during flooding (see below). Some of these taxa produce useful fruits; phytoliths thus provide "clues" for identifying unknown charred wood and fruit fragments (dense). Macroremain data, which represent charred debris thrown into the pit, in turn provide insights for interpreting phytolith abundances. Maize kernel fragments and cupules occur commonly in the macroremains (9.2-80.7 percent by count, 60.6 percent overall; 6.7-79.Z percent by weight, 42.5 percent overall), at higher percentages than do cross-shaped phytoliths produced by maize leaves (0-60 percent by level; 19 percent overall of all crosses are maize). Maize crosses make up a very small percentage of all phytoliths (less than 1 percent overall, N = 23), yet clearly maize was a common component of garbage disposed of in the pit. These data indicate that maize leaves were less frequently deposited in the pit than charred remains of kernels and cobs; a finding which suggests that frequency of maize leaf phytoliths underestimates the abundance of the crop in such contexts. By contrast, in an agricultural field setting, one would expect maize leaves to be common, and maize cross frequency to be a better indicator of maize abundance. In the case of palm (Aracaceae), phytoliths occur in somewhat higher frequencies than do charred remains (phytoliths: o—10 percent; charred: 0.1—5.9 percent by count, 0.6—7.9 percent by weight). Since palm phytoliths are produced by all parts of the plant (leaves, stem, and fruits), but macroremains consist only of palm nut fragments, over-representation of phytoliths would be expected. The data thus appear to be in agreement that palm was a minor element in the pit. Examples of how phytolith and pollen data complement each other are presented in this volume by Piperno. The differing sources of botanical remains in archaeological contexts also have an impact on interpretation. In the case of the Pechichal pit, three processes resulted in deposition of botanical materials: (1) trash disposal (deposition of burned plant tissues, leaf debris, flowers, wood, and so on, followed by decay and release of phytoliths); (2) airborne "rain" (pollen, wind-blown dust carrying phytoliths, volcanic ash); and (3) flooding (phytoliths and pollen redeposited by soil movement). Although pollen and phytoliths may both be deposited by wind and water, the source area for pollen is generally larger, since many pollen types are designed to be transported by wind (that is, wind-pollinated taxa). This is not the case for phytoliths, which move with the matrix in which they are deposited. Phytoliths deposited by flooding would thus come from the watershed of the river, while pollen (if preserved) would represent regional and local vegetation. Charred macroremains likely represent in situ deposition of trash, and many phytoliths and some pollen grains also share this source. This complex depositional situation must be taken into account when data are compared and interpreted. One way to accomplish this is to compare assemblages derived from cultural contexts to those from natural deposits. Continuing with the Jama example,
126
DEBORAH M. PEARSALL
Veintimilla and Pearsall's (1992) analysis of phytoliths from a soil profile in the river channel, revealed which phytolith taxa were derived from "background" vegetation, and what abundances should be expected. This allowed us tofilterout "background" phytoliths in the cultural deposits. Phytolith assemblages derived from comparative surface soil samples are also useful in this regard. Phytoliths as sole data source The most impressive contributions of phytolith analysis to paleoethnobotany in the lowland tropics have resulted from those situations where phytoliths are the sole source of data on subsistence or vegetation. Pollen preservation may be poor, flotation recovery inadequate, post-depositional destruction high, or the context such that no other remains were deposited. Lack of burning activity, and thus lack of charred remains, is often encountered in man-made landscape features such as raised fields and agricultural terraces. Phytolith analysis may reveal not only what crops were grown on fields, but document the history of their construction, use, and abandonment. Phytolith analyses of samples from two raised field complexes in the Guayas Basin, the Penon del Rio complex, and the Yumes complex, for example, provided a direct opportunity to study late prehistoric crop production (Pearsall 1987a, 1987b). Results indicated that temporal and regional diversity in cropping practices existed. Maize was grown on thefieldtested in the Penon complex, while analysis of samples from two Yumes fields revealed a focus on mixed cropping of perennials and long-season annual crops, and secondarily short-season annuals such as squash. On-site contexts may also lack charred macroremains, either because no burning activity occurred (for example, in burial features) or because remains were destroyed after deposition. The latter is the case for many contexts at the San Isidro site in the Jama River valley (Pearsall 1991). Phytolith analysis suggests that the first inhabitants of the site, peoples producing terminal Valdivia pottery (1600 BC), brought with them a subsistence system which included maize, squash or gourd, and two root crops, arrowroot and achira. Other utilized plants included palms, sedges, and bamboos. The list of cultivated and utilized plants present during the subsequent Chorrera (1000—500 BC) and Jama—Coaque (500 BC^AD 1500) Periods remained much the same. Finally, while lakes, ponds, or swamps are excellent contexts for sampling phytoliths for vegetation reconstruction (see Piperno, this volume), waterlogged conditions are not necessary for phytolith preservation. This opens up many other natural deposits for sampling. For example, while contexts for good pollen preservation are rare in the Jama valley, phytoliths are abundant, in both on- and off-site contexts, including the fluvial profile studied by Veintimilla and Pearsall (1992). Representing buildup of alluvium from before human occupation to recent times, this profile, exposed during river downcutting, reveals a pattern of vegetation change which can be linked to modification of the valley environment
"Doing" paleoethnobotany in the tropical lowlands
127
by its human occupants. Veintimilla found, for example, that clearing of the forest began when people first occupied the valley, and that this process accelerated rapidly through time. The earliest indication of forest clearing may even predate terminal Valdivia by a few hundred years. Because the base of the profile is dated only by extrapolation from dated tephra deposits, solid evidence for earlier Valdivia occupation awaits discovery of sites. No sites dating before terminal Valdivia have yet been discovered (Zeidler 1991, and this volume). When geological sequences can be precisely dated, phytolith analysis provides us with a way to "see" people on the landscape before occupation sites are discovered. In environments where ancient sites may be deeply buried, this is an important contribution of phytolith analysis to archaeology. "Doing" paleoethnobotany in the tropical lowlands: assessment and future directions To illustrate how our understanding of past human-plant interrelationships in the New World tropics has been enhanced by application of good macroremain recovery techniques and the analysis of pollen and phytoliths, I will first present a case study - the antiquity and importance of maize in northern South America then discuss problems and future directions in tropical paleoethnobotany. The question of maize Because botanical preservation is far from ideal in northern South America (Colombia, Ecuador, and Venezuela), few conspicuous botanical remains were recovered from early excavations. For example, prior to the introduction of flotation in the mid-1970s, there were few direct data on subsistence for any prehistoric time period in Ecuador (Pearsall 1979). It was assumed that maize cultivation was the mainstay of subsistence late in prehistory (Meggers 1966), but even this was speculation. When maize came to assume this role could only be inferred by comparison to better known regions, such as the central Andes. Subsequently, flotation and phytolith analyses have been carried out at sites in both the Ecuadorian sierra and the coast, with foci on the late preceramic Vegas tradition (8000-4400 BC) and the Formative, or early ceramic, period (3300-500 BC, coast; 1500-500 BC, sierra). The earliest evidence for maize comes from late Las Vegas (6000-4400 BC) strata at site OGSE 80 on the Santa Elena Peninsula (Piperno 1988b; Stothert 1985,1988), where maize phytoliths occur in association with silica bodies produced by squash. Whether the squash was domesticated or wild is not known. No evidence of maize occurred in early Las Vegas strata. A preceramic date for the introduction of maize into northern South America is also supported by data from Colombia and the Ecuadorian Amazon. Maize appears in a long pollen core from Hacienda El Dorado in the Calima region of Colombia at 4600 BC. This initial appearance is followed by evidence for deforestation and a sharp rise in Gramineae pollen and in maize (Bray et al. 1987;
128
DEBORAH M. PEARSALL
Monsalve 1985; Bray, this volume). Maize pollen and phytoliths occur in a core from Lake Ayauch1 in Amazonian Ecuador in strata dated to 3300 BC (Bush et al. 1989; Piperno 1990). Maize occurs in association with disturbance indicators and abundant carbon particles, indicating the beginning of shifting cultivation. As the result of macroremain and phytolith analyses from a variety of Formative period sites in the Ecuadorian coast and sierra, we now know that maize, jack bean {Canavalia plagiosperma), common bean (Phaseolus vulgaris), the root crop achira {Canna edulis), cotton {Gossypium), and gourd (Lagenaria siceraria) were cultivated by the end of the Formative (500 BC) (Pearsall 1988, 1992). Maize is present from the beginning of the Valdivia Period (3300 BC). A host of other plants also occur in Formative Period sites, including fruits like palm, soursop, and cactus on the coast, lupine in the sierra, small-seeded annual weeds (chenopod, amaranth, and similar plants in the Aizoaceae and Portulacaceae), and sedge tubers (Cyperus or Scirpus). Maize thus occurs in prehistoric Ecuador in association with a multitude of other plants, both wild and cultivated. Given the evidence that maize was introduced into northern South America before 5000 BC, nearly contemporary with its appearance in western Panama (Piperno 1988a; Piperno and Husam-Clary 1984; Piperno et al. 1985; see Bray, Piperno, this volume), what role did the crop play in diet? As a "foreign" plant, introduced into existing foraging/horticultural subsistence systems, maize likely underwent a long period of low utilization as a vegetable or curiosity. While the type of maize introduced cannot be reconstructed directly from the available phytolith and pollen data, it was probably similar to the early maize of central Mexico; small, fragile cobs with few rows of small kernels. Low productivity makes it unlikely that maize would immediately supplant other, more productive resources. Testing this scenario is a challenge, however. It is difficult to estimate relative abundance of plants using phytolith data, and macroremains are relatively scarce. There is, however, some evidence from coastal Ecuador for increasing variety, or richness, in plant taxa from early preceramic through late Valdivia times (that is, 8000-1500 BC). This is partially an artifact of preservation or recovery, since no macroremains were identified from the Vegas site. Looking at just the Valdivia sequence, and beginning with a cultivated plant assemblage which included cotton, maize, and Canavalia beans, Canna was added in Valdivia III times, and Maranta at the end of Valdivia, suggesting increasing variety. Maize becomes more common over the sequence. It occurs in many loci at the Real Alto site at the same time Canna appears. These trends suggest that while maize became more widespread, subsistence was still broad-based, with tuber resources, tree fruits, wild small-seeded annuals, and cultivated annuals, along with fish, shellfish, and terrestrial animals, all part of subsistence. There is no paleoethnobotanical evidence for intensive use of maize in northern South America until much later in prehistory (for example, in coastal Ecuador, early in Jama-Coaque).
"Doing" paleoethnobotany in the tropical lowlands
129
This tentative conclusion is supported by the apparent late dietary importance of maize as measured by carbon isotope studies (see Norr, this volume) in areas as diverse as the Orinoco Basin, where it is important only at AD 400 (van der Merwe et al. 1981), and highland Peru, where its importance post-dates Chavin (that is, after 250 BC) (Burger and van der Merwe 1990). These findings suggest that indigenous subsistence systems based on cultivation of tubers and use of wild plant and animal resources sustained cultural development in South America for many millennia. The status of paleoethnobotany in the lowland tropics As the example of maize illustrates, it is possible to recover botanical remains from the lowlands, and to apply those data to questions of importance in archaeology. The key is to design and execute a research strategy focused on recovering macroremains, pollen, and phytoliths from both archaeological site contexts, and from off-site sampling locations where direct evidence of plant cultivation and human impact on the environment can be obtained. Reconstruction of past plant-human interrelationships depends on precise identification of remains. Comparative collections are essential; identification techniques must take into account the high species diversity of the tropics, and the high redundancy of form within some plant groups. A conservative approach to identification is crucial. The issue of quantitative analysis of botanical materials remains a difficult one. Too few macroremains may be recovered to calculate ratios or diversity measures to compare occurrence of taxa among contexts or sites (but see Lentz 1991). Presence/absence data (ubiquity) may be the only alternative, and these data are not free of sample size constraints (Kadane 1988). Low sample counts will continue to constrain the testing of models for subsistence change. Returning to the Jama example, we are applying Rindos's (1984) model of agricultural evolution to the Jama case, but there may be insufficient data to calculate diversity (the measure which best fits changes predicted in the model) for some time periods, forcing us to rely on ubiquity measures to characterize subsistence change. What strengthens the Jama results are the phytolith data; phytolith analysis is the most important development in paleoethnobotany since the flotation revolution. Most phytolith studies to date, however, have used these data only to argue for the presence of certain crop plants or to propose changes in vegetation based on changes in frequencies of indicator species. Multivariate statistical analysis of phytolith assemblages holds considerable potential for applying these data more rigorously to questions of subsistence and environmental change. While it is still difficult to recover direct evidence of plants used by people in the lowland tropics, and while not all data recovered may be of sufficient strength to test every question we have, we have come a long way.
6
Plant microfossils and their application in the New World tropics DOLORES R. PIPERNO
The Neotropics are home to over 50,000 species of plants, which have existed together in their modern geographic assortments since shortly after the end of the Pleistocene, 11,000 years ago. Various lines of evidence, including modern plant distributions, accounts of the first Europeans to see tropical America, and the archaeological record from deserts occupying tropical latitudes, tell us that many species were domesticated or manipulated extensively in prehistory (for example, Oviedo 1959; Plowman 1984; Ugent et al. 1986). In addition, ethnobotanical studies provide a living testament to the impressive knowledge of plants held by native tropical people (for example, Duke 1968; Posey and Bailee 1989; Prance and Kallunki 1984). However, these many affiliated lines of evidence stand in contrast to the limited data on prehistoric plant exploitation accumulated from archaeological sites. The lowland humid tropics, by definition, encompass a set of environments that are inimical to the preservation of organic material. Year-around high temperature and humidity, or fluctuating rainy and dry periods in regions with marked seasonality, ensure that the recovery of plant remains from archaeological sites presents formidable challenges to prehistoric researchers. Donald Lathrap's brilliant insights into native tropical society included prominent attention to the ways in which culture and the plant world articulate (Lathrap 1970, 1973). He stressed that botanical study is crucial in tropical archaeology, not only for explicating the developmental history of plant domesticates and subsistence practices, but also for understanding patterns of cultural interaction and change. Lathrap advocated an archaeological research design that stressed maximum recovery of plant materials, using both established and new techniques. Recent programs in Central America, western Ecuador, and the Amazon Basin, directed toward recovery of macrobotanical remains using improved flotation procedures, have been very successful in documenting some aspects of plant subsistence (for example, Miksicek et al. 1981; Pearsall 1989, this volume; Roosevelt 1980; Turner and Miksicek 1984). Yet, the list of utilized/domesticated 130
Plant microfossils
131
species can be nowhere near complete, and preserved plant remains do not appear to be of very great antiquity. At many other sites, particularly those bearing cultural deposits of early and mid-Holocene age, poor preservation of fruits, seeds, and other types of macrobotanical material is the rule (see Mora et al. 1991; Pearsall 1985; Piperno and Clary 1984). Fortunately for archaeology, plants commonly fossilize more than one of their constituent body structures or tissue types. This leads to the creation of both macro- and microfossil remains and, thereby, increased potential for recovery of significant botanical data. This paper reviews the contributions of microfossil plant remains to major questions in the prehistory of the lowland humid tropics. Two classes of plant microfossils, pollen and phytoliths, are useful in archaeological botany. Pollen grains are the male element in flowering plants, produced in large number in the stamen of the flower. Phytoliths, literally meaning "plant stones," are secretions formed of either opaline silica or calcium that usually develop in living plant cells of non-flowering organs, and that subsequently are liberated into the environment when the plant dies and decays. In this paper, only those phytoliths composed of opaline silica are considered, since they are taxonomically more diverse than the calcium products, occur in more plant taxa, and are found well preserved in many tropical sites. Research on the different classes of plant remains preserved in archaeological sites has developed very unevenly. For example, serious attention to recovery and identification of pollen has a longer history in tropical archaeology than does phytolith analysis. However, over the last decade the latter methodology has made rapid strides as an accepted tool of paleoethnobotany, and the two techniques are increasingly being carried out in tandem. This paper starts with a general review of pollen and phytolith analysis in the American tropics, emphasizing recent developments and addressing some questions that have arisen relating to microfossil plant analysis in archaeology. The following section discusses how application of phytolith and pollen analysis has contributed to study of prehistoric subsistence and settlement in lowland Panama. Finally, an argument is presented for developing a new approach in archaeological praxis, which emphasizes the retrieval of data from non-occupation site contexts relating to the environmental correlates of prehistoric behavior. It is believed that this approach can become an important component of tropical archaeology.
General aspects of pollen and phytolith analysis in tropical archaeology Pollen analysis Pollen analysis has long been an accepted technique in archaeology and paleoecology (for example, Bryant and Holloway 1983; Dimbleby 1985; Faegri and Iversen 1975). Compared with analyses in temperate regions of the world, the application
132
DOLORES R. PIPERNO
of palynological techniques in lowland tropical regions is a younger and taxonomically rudimentary science. This situation exists because the tropical flora present greater complexities for recovery, quantification, and interpretation of fossil pollen. For example, there are large numbers of silent taxa, those present in the flora but unrecorded in pollen diagrams because their pollination mechanisms (usually by insect or bird) result in low pollen production. In addition, the great diversity of tropical species requires that extensive reference collections of modern plants be made by palynologists before identification of the fossil pollen record may proceed. Consequently, pollen diagrams have often contained high proportions of unknown (and often highly distinctive) taxa. Another limiting factor in tropical palynology is that pollen grains are often prone to destruction in depositional environments that are not protected or waterlogged and, hence, may not survive in the highly weathered soils characteristic of many archaeological sites. While these problems should not be underestimated, recent studies of archaeological sites and associated lakes and swamps have proven that the palynology of the lowland tropics is a robust area of study. Pollen grains from many important economic taxa, including maize {Zea mays L.), manioc {Manihot esculenta Crantz), squash {Cucurbita spp.), and chile pepper {Capsicum spp.) have been found well preserved in open-air sites (for example, Crane 1986; Mora et al. 1991; Piperno and Husum-Clary 1984), as well as in deep levels of stratified rock shelters (Piperno et al. 1985). Working on terra preta sites in the Colombian Amazon, Mora et al. (1991) have constructed a 5,000-year history of tropical forest adaptation, showing the sequential introduction of maize and manioc along with trends in the exploitation and modification of the surrounding forest. Studies such as these emphasize that the palynology of lowland tropical archaeology can be a productive endeavor. Other important findings made recently in tropical palynology relate to the records of lakes and swamps. Despite the large numbers of insect- and birdpollinated plants in the lowland tropical forest, pollen influx into lakes from these biomes has been found to be comparable with that of temperate regions (Bush and Colinvaux 1988; Bush et al. 1992; Jones, personal communication, 1992). Preservation is excellent, and concerted efforts to build reference collections and study the pollen rain of extant forests have resulted in the identification of a large number of taxa, with concurrent reduction in the proportion of unknown pollen types (Bush and Colinvaux 1988; Bush et al. 1992; Jones 1991a; Liu and Colinvaux 1988; Rue 1987, 1988). Thus, fears that the lowland tropical forest would leave little in the way of a useful pollen record because of poor production, preservation, and taxonomic specificity (for example, Faegri 1966) are unfounded. The analysis of lakes and swamps located in archaeological study regions has recently become a focus of palynological attention. Traditionally, such studies, carried out by biologists whose primary concern was with climatic and vegetational history, were not considered directly relevant to the goals of archaeology.
Plant micro fossils
133
However, pollen analysis of geological sites, undertaken as part of the archaeological research design and carried out by investigators trained in both anthropology and botany', is providing crucial information on the prehistory of several regions of Central and South America. It seems pertinent here to provide a brief review of the major findings of these studies, and also to address some criticisms that have been directed at them recently. The criticisms, which will be noted first, seem to derive from unfamiliarity with paleoecological techniques, leading to overstatement of the problems and lack of appreciation of results by some archaeologists. For example, in a discussion of potential avenues of data accumulation for Amazonian prehistory, Roosevelt (1989) comments that lake sediment cores offer poor conditions for stratigraphic interpretation, even poorer chronological control than normally achievable with datable remains in archaeological middens, and inconclusive botanical results. These claims seem to be based on her uncritical reading of pollen records from vdrzea lakes in the Amazon (Absy 1982), and they are particularly spurious when applied to permanent lakes. Permanent bodies of water have been found to be much more common in the lowland tropics than previously thought. A basic feature that has long made them the subject of paleoecological enquiry elsewhere is that they accumulate sediment and associated microfossils with each passing year. Bioturbation is inherently unlikely due to the rarity of organisms that promote sediment mixing in anoxic conditions. The retrieval of undisturbed cores from lakes with a Livingstone sampling device is a tried and true technique (for example, Colinvaux et al. 1985). Subsequent X-ray and visual analyses of core sections are capable of revealing precise details of sedimentation, such as fine, undistorted laminae. When present, such details strongly suggest continuous and undisturbed sedimentation. Finally, chronologies derived from lake sediments are arguably of finer quality than those from archaeological middens, which are subject to inaccuracies caused by the numerous confounding effects of past cultural activities (see Siegel, this volume). The majority of tropical lakes are not subject to the hard water error (Deevey et al. 1954); therefore, it is often possible to directly date small sections of core sediment. In these circumstances the generation of long and internally consistent C14 records has been achieved (e.g., Bush et al. 1989; Piperno et al. 1991b). Lakes and swamps will also frequently have sedimentary zones consisting of peat or partially decomposed wood, which provide reliable dates even when the hard water effect is in play (Jacob 1991b). In sum, geological sites are capable of providing high-quality botanical information of considerable relevance to the goals of tropical archaeology. Some of the recent studies exemplifying this progress will now be discussed. Rue (1987, 1988) examined the pollen records from a lake and swamp located near Copan, a Mayan center in western Honduras. He recovered earlier evidence for prehistoric occupation of the region than had been revealed with archaeological survey and excavation. The presence of maize pollen along with other indicators of slash and burn agriculture, which were evidenced from the very
134
DOLORES R. PIPERNO
beginning of the core sequence and dated to c. 4500 BP, revealed the considerable antiquity of seed cropping and associated forest modification in the Mayan area. These results highlight the important questions of Mayan origins (extrusive or in situ development) and the antiquity of human settlement and agriculture in the lowland tropical forest. Rue's studies also shed light on the cultural and demographic factors associated with the Classic Mayan collapse. Most interestingly, the vegetation records provided no evidence for forest regeneration that would be expected had the region been abandoned during the ninth century AD. It seemed that the rural lands outside of Copan continued to be occupied after the collapse, afindingin accordance with the most recent archaeological evidence. Jones' (1991a; personal communication, 1992) recent pollen studies of swamps located in Belize have greatly amplified knowledge of prehistoric settlement and subsistence in that area. At Cobweb Swamp, adjacent to the Mayan center Colha, a 10,000 year old sequence revealed that maize and squash were grown, and forests cleared, starting at c. 4700 BP. The long record of another site, Cob-III, reveals an even earlier presence of maize, which occurs well before the first signs of forest clearing, dated again to c. 4600 BP. In South America, Monsalve's (1985) two long pollen sequences from swamps located 3 km apart in the Calima Valley, Colombia record the continuous presence of maize between 6680 BP and 720 BP. While located at an elevation of 1,400 m and thus not embedded firmly in the lowlands, these sequences corroborate the others in indicating early maize dispersals out of Mexico. Finally, Bush and colleagues' (1989) pollen results from a lake in Amazonian Ecuador evidenced maize pollen and associated forest modification in a section of the core 10 cm below a 2 cm section dated to 4510 BP, and 29 cm above a 2 cm section dated to 7010 BP, yielding an interpolated date of 5300 BP. This finding constituted both the earliest evidence for maize in the Amazon Basin and the first record of prehistoric occupation in the forests of this region. Colinvaux and Bush's palynological efforts in Panama, which are part of a multidisciplinary study of a tropical watershed, will be explained in detail in the following section of this paper. There are two unifying themes emerging from these records. The first is that paleoecological evidence for human occupation of the regions in question is predating available archaeological evidence. Apparently, the environmental correlates of early foraging and farming in tropical forest are more visible than the stones and other implements used by people in their daily lives. Of course, many of the regions have not seen systematic archaeological survey, so the antiquity of settlement is typically based on the excavation of one or two large sites which reached their zenith later in time. The other major finding of these records is that all are evidencing an early pattern of seed cultivation associated with forest clearing. In the Colombian, Ecuadorian, and Belize records, uncorrected dates relating to the earliest maize (6680 BP, 5300 BP, and 4700 BP, respectively) are earlier than or coeval with those
Plant microfossils
135
of the earliest finds of maize in Mexico (Long et al. 1989). The skeptic would argue that a direct date on an archaeological cob is a more reliable determination than one on geological sediment. However, the paleoecological age determinations were carried out, respectively, on polliniferous sandy peats, gyttyas, and peats. All of these types of sediment are reliable indicators of age, and it is unlikely that the sediments are much older than the pollen grains they contain. The most parsimonious explanation of these findings is that the earliest Mexican maize remains do not represent the earliest Mexican maize. All lines of botanical evidence are indicating that the Tehuacan Valley was the recipient of a maize that was domesticated elsewhere (for example, Doebley 1990; Benz and Iltis 1990). Many of the discrepancies between the archaeological and geological data sets will undoubtedly be erased as more information is compiled on early tropical forest settlement and subsistence. In the meantime, serious attention to retrieval of "paleoecological" data should continue to provide information otherwise difficult to obtain from archaeological contexts. Phytolith analysis The emergence of phytolith analysis in tropical archaeology owes a direct debt to the foresight of Donald Lathrap. He encouraged one of his graduate students, Deborah Pearsall, to pursue the technique after reading a little-known manuscript (Matsutani 1972) about a phytolith study carried out on the Kotosh site, Peru (see Pearsall, this volume). Lathrap realized that phytolith analysis was an unexplored discipline with considerable potential. Pearsall (1978) then carried out one of the first modern applications of archaeological phytolith research, by developing a technique to identify maize phytoliths in Early Formative contexts in southwest Ecuador. Phytolith analysis, though still classifiable as a "new" technique if its development is compared to that of palynology, has matured considerably over the last ten years. Much of the basic research that has proven its reliability in paleobotany has been carried out in the American tropics (for example, Piperno 1988; Pearsall 1989a, this volume). Phytoliths are probably the most durable kind of plant fossil known to science. They are found faithfully preserved in a wide range of depositional environments, including highly leached and weathered soils characteristic of archaeological sites, as well as the anoxic conditions of lakes. Thus, they appear to outrival pollen grains and charred macroremains in suvivability. It seems that only a very high soil pH (measuring nine and above), a circumstance characteristic of archaeological shell middens, may cause the dissolution of phytoliths in tropical sites (Piperno 1988a). But, they sometimes have been found in high number even in these kinds of contexts from the temperate zone. Phytoliths are likely to be added to ceramics as another category of artifactual remain consistently preserved in lowland tropical archaeological sites.
136
DOLORES R. PIPERNO
Establishment of the validity of phytolith analysis in tropical paleobotany rested on the same tests that earlier had been applied to pollen analysis. Namely, phytoliths were shown to: (1) be formed in a wide range of plant taxa; (2) be produced repeatedly in morphologically consistent shapes by different local populations of individual taxa; and (3) exhibit shapes that identify plants at precise (family, genus, species) taxonomic levels (Piperno 1985, 1988a, 1989, 1991; Pearsall 1989). Additionally, and very importantly, a strong correspondence was found among the shape of a phytolith, the species producing that shape, and the evolutionary relationship of that taxon to other species (see Piperno 1991 for a full discussion of this issue). Such patterns pointed to a strong genetic component influencing phytolith morphology, indicating that phytolith formation and morphology were not likely to be subject to the whims of local environmental variability and change over time. Phytolith formation does not occur in many plants, and all plants in which phytoliths are produced do not contribute taxonomically useful shapes. Also, amongst phytolith producing taxa the silicification of several different organs of the plant and of vegetative and reproductive structures may occur. This factor makes the compilation of a reference collection a somewhat more arduous task than that required for palynology. However, there is a long list of plants important in tropical subsistence and environmental reconstruction, including such major plant domesticates as maize {Zea mays L.) and squash {Cucurbita spp.), that produce discrete, identifiable phytoliths in large number. Phytoliths are routinely found in many kinds of depositional environments. They have already been studied in prehistoric habitation sites, swamps, and lakes. The geological phytolith record is proving to be highly informative (for example, Piperno et al. 1991a, 1991b). Indeed, in the Panamanian and Ecuadorian lake sites discussed above, where phytolith and pollen analysis were carried out in tandem, the inferences drawn on the basis of the pollen data were independently derived from the phytolith records. They will be reviewed in detail in the following section. Phytoliths may also be preserved in such artifactual remains as pottery, stone tools, and teeth (Piperno 1988a, 1991), creating the potential for a large data base with relevance to technology and stone tool function, as well as dietary and ecological reconstruction. There are still some doubts in the archaeological community about how useful phytolith analysis will prove to be in tropical paleobotany. Some prehistorians note that phytolith analysis has not yet benefitted from the years of taxonomic research needed to standardize identification keys, and owing to the impressive richness of tropical flora, the possibility exists for misidentification of phytoliths in sediments. In answering these questions it should be emphasized that extensive (thousands of species) modern reference collections, leading to a very good understanding of phytolith morphology, have been constructed for regions of the American tropics
Plant microfossils
X
37
where phytolith research has been carried out (that is, Piperno 1988a, 1991; Pearsall 1989). As described above, predictable patterns of phytolith production and morphology that parallel the taxonomic affinities of plants have been demonstrated, and these patterns also bear close relationships to those observed in related plants from different regions of the world (for example, Tomlinson 1961, 1969; Bozarth 1992; Kealhofer and Piperno n.d.; Ollendorf 1992). There is obviously a fundamental rhyme and reason to phytolith morphology, a fact that makes it even more unlikely that phytoliths characteristic of, say, squash rinds or bamboo leaves will be found in some rare, unanalyzed tropical tree. It should also be noted that phytolith analysts have very good information on production patterns in modern plants. This makes it instantaneously possible to selectively research those families of plants that will produce a phytolith record, while de-emphasizing the many taxa that for reasons currently unknown simply do not silicify their organs. Lack of phytolith production by tropical taxa leads to the absence of important plants in the fossil record, but it also reduces the number of species to be examined by many orders of magnitude. Finally, the phytolith record of Panama has been examined in conjunction with detailed pollen records from the same contexts (see below). Comparison of the two data sets shows a very close correlation of results, a situation hardly possible if serious errors were biasing a phytolith record. I now turn to the application of phytolith and pollen analysis in this important low-lying region of Central America. Phytolith and pollen studies in Panama: the evolution of tropical subsistence and settlement during the past 11,000 years Over the last ten years the drainage of the Rio Santa Maria in the central Pacific watershed of Panama has been a focus of archaeological and paleobotanical study (Figure 6.1). The 3,500 km2 area stretches from the Continental Divide, at times reaching an altitude of only 1,000 m, to the zone of mudflats and mangrove bordering the Pacific Ocean. Beginning shortly after the end of the last glacial epoch 11,000 years ago to before the onset of intensive human disturbance, lowland tropical forest adapted to varying levels of annual precipitation, and therefore, forests exhibiting different degrees of species diversity and canopy closure (evergreen, semi-evergreen, and deciduous types) graced virtually the entire area. The "Proyecto Santa Maria" (PSM) was a multidisciplinary project initiated in 1982 by Anthony Ranere and Richard Cooke, which studied the evolution of settlement and subsistence in this region. Project goals were accomplished by means of systematic settlement survey, excavation of selected sites, and analysis of surface remains and excavated materials. The closely allied botanical efforts, directed by the author, were initially formulated to provide tangible evidence on subsistence trends over time, particularly the transition to agriculture. These
DOLORES R. PIPERNO
i38
\ COSTA RICA
CARIBBEAN SEA
c'i
•^.CONTINENTAL DIVIDE
)
°
«
^ ^ ^ ^ ^ A i L VALLE ,^^^^-LAYEGUAOA R0CK SHELTERS
^Qsz^^S^ s
f V T ^ ^ ^ ^ ^
^
^ >^
SITIOSIERRA
"
CUEVA DE LOS VAMPIROS " LAMULA
3
PACIFIC OCEAN
• Lake and Archaeological Site Locations ty$$i.
Q \ \ Q, i ^
Proyecto Santa Maria Study Area
0 km
\
Figure 6.1. Location of the archaeological sites and lakes discussed in the text. efforts were eventually broadened to address early patterns of foraging and the environmental contexts of human occupation. Summaries of the archaeological sequence may be found in Cooke (1992), Cooke and Ranere (1984, 1992b), Ranere (1989), Ranere and Cooke (1991), and Weiland (1984). In brief, a continuous 11,000 year long record of human occupation has been uncovered. Surface finds of Paleo-Indian artifacts occur at La Mula, a site that now lies near the coast. A date of 10,440 BP on carbon associated with lithic implements, including bifacial thinning debris, has been recovered from deep in the deposits at the Corono rock shelter, which lies further inland at about 220 m (Figure 6.1). Beginning around 10,000 BP, rock shelters located at elevations between 200 m and 900 m were more regularly occupied by people who continued to work stone bifacially. An edge-ground cobble/boulder milling stone complex was introduced by 8040 BP. Seven thousand BP seems to mark an important shift in regional human adaptation. Settlement number and size appear to increase substantially. Bifacial thinning is no longer used as a core reduction strategy, and the edge-ground cobble/boulder milling stone complex becomes common. The manufacture of ceramics in central Panama was underway by 5000 BP, and shortly thereafter the well-known Monagrillo pottery is found in all parts of the region. At 2400 BP the first signs of sedentary agricultural villages are evident. These were positioned to exploit the fertile alluvial zones of the coastal plain and were numerous by a few centuries after the time of Christ. Phytoliths and pollen in the archaeological sites Phytoliths were examined in archaeological soils from eight rock shelters (occupations C14 dated between 10,400 BP and 3000 BP), two shell middens
Plant microfossils
139
(including the Monagrillo type-site), and two open-air agricultural villages where C14 dates place their occupations between c. 3000 BP and AD 500. The pollen analysis, which is still underway, has been completed at two of the rock shelters, the Monagrillo site, and one of the agricultural villages. Pollen preservation was generally very uneven, manifesting what may be described as a typical "lowland tropical pattern." Many levels from sites contained very little pollen; however, unpredictability was the rule as older levels of preceramic units sometimes contained identifiable grains in high amounts. Maize and other types of economic pollen were identified in preceramic levels from the Cueva de los Ladrones, C14 dated with associated carbon to 7000 BP. Also present along with preceramic maize were grains from economically useful trees (for example, Byrsonima and Hymenaea). Another site which contained abundant identifiable pollen was the agricultural village Sitio Sierra, whose deposits are C14 dated between 3000 BP and AD 500. Here Cucurbita and maize pollen were common. The pollen spectra were dominated by taxa from invasive, weedy growth typical of anthropogenic savanna. The phytolith record from the archaeological sites offered more information than did pollen, in that preservation was nearly uniformly excellent, and a greater number of taxa were evidenced. The only three sites examined that did not yield identifiable phytoliths in large number were the Monagrillo and Cerro Mangote shell middens, and La Mula, where large amounts of shell were also present in ceramic-phase deposits. With little doubt, tropical shell middens seem to be inimical environments for phytolith preservation because contexts into which phytoliths are deposited are continually bathed by warm water of high pH (8.5 and above), causing dissolution of silica. The degree of preservation exhibited by the phytolith record enables comparisons of within-site and between-site assemblages that are seldom possible with other botanical data. This is not to say that differential preservation of phytoliths does not influence the final composition of assemblages. There are many classes of phytoliths that apparently do not survive in tropical soils (often those in which silicification was confined to the wall of the living, parent cell). Nevertheless, we have considerable understanding of the types of phytoliths that are resistant to dissolution, and it is these types that form the basis of the following interpretation of the data. Table 6.1 presents a summary of the phytolith record from the archaeological sites. Phytoliths document the presence of two plant domesticates (maize and squash), and point to the early Holocene use and possible cultivation of a third plant, arrowroot, which is present at several sites. Arrowroot (Maranta arundinacea), which possesses a smallish tuber and is casually grown by campesinos today, is native to Panama and all of the Central American region. Arrowroot phytoliths are most common in preceramic-phase horizons and then decline in abundance until they are barely visible in deposits associated with sedentary village occupations. The earliest record of arrowroot is from levels dated to 8600 BP at the Cueva de las Vampiros, a site where the phytolith record
Table 6.1. Phytolith record from Panamanian archaeological sites. Period
Subsistence
Environment
Paleo-Indian ? 11,400-9500 BP
A single, reliable record from the rock shelter Corona indicates the exploitation of some tropical forest plants Utilization of tropical forest plants (Palmae, Chrysobalancaceae, bamboos) at rock shelters located at elevations between 600 m and 900 m. At Cueva de los Vampiros, located a few kilometers from the coast, arrowroot is present in deposits dated to 8,600 BP Maize phytoliths are present at two rock shelters. One of them also contains arrowroot. At the Aguadulce shelter, located on the coastal plain, maize is absent in preceramic units and arrowroot is abundant. A large number of phytoliths characteristic of oil palm (Elaeis) here (carbonized Elaeis fragments were also present in number) suggests heavy exploitation of local seasonal swamps. At other rock shelters utilization of a variety of tropical forest taxa continues Maize enters the record of the Aguadulce shelter. Arrowroot nearly disappears here and also appears to become less important at the Cueva de los Ladrones. Other shelters show little evidence for change in subsistence practices. Unfortunately the Monagrillo shell midden did not contain a phytolith record Maize phytoliths are present for the first time in the uppermost levels of two rock shelters, which probably date to this period. Other sites show an increase in the number of maize phytoliths. The first archaeological record of Cucurbita occurs, at Sitio Sierra
Too few data to draw inferences concerning the environment
Early Preceramic (Archaic) 9500-7000 BP
Early Preceramic (Archaic) 7000-5000 BP
Early Ceramic 5000-3000 BP
3000 BP-AD 400
The presence of moist tropical forest near the interior rock shelters is indicated. At La Mula, abundant grass and fragmented sponge spicules plus presence of marine diatoms indicates a landscape like the modern one, where arrowroot does not grow naturally Vegetation near some shelters, including one with maize, now contains substantial amounts of disturbance taxa. Records of others show little sign of human interference. Phytolith assemblages point to variability of subsistence orientations at sites, located in different types of forest and microhabitats
Continued signs of habitat interference near some sites. At others, trends indicating little environmental modification continue
Signals of serious deforestation and habitat modification are now present. Markers of anthropogenic savanna appear in sites occupying the coastal plain
Plant microfossils
141
indicates an environment unlike habitats presently favored by this plant. Arrowroot thus might have been taken outside of its natural range and planted there. Arrowroot's place as a significant carbohydrate supplier (Piperno 1992) probably fell victim to the availability and increasing importance of maize (and most likely manioc, though it cannot be directly documented). Phytolith and pollen data are concordant in registering preceramic maize at the Cueva de las Ladrones by c. 7000 BP. Maize phytoliths were also recovered from the bottom of the preceramic deposits at another rock shelter, SE-189, whose deposits are also dated to early in the seventh millennium BP. Squash phytoliths are not recorded archaeologically until c. 2000 BP, when they appear in deposits from the agricultural village site of Sitio Sierra. The preeminence of maize in subsistence systems between 3000 BP and AD 500 is amply evidenced in the phytolith record. An interesting feature of the phytolith record is that maize and arrowroot are not uniformly present in the preceramic-phase deposits from all sites investigated. In fact, four rock shelters showed no evidence of either preceramic or ceramicphase maize. These differences probably have much to do with the seasonality of occupation, functional variability, and ecological setting of the sites. They also tell us that it might be a mistake to assume that all residential groups in Panama between 7000 BP and 5000 BP were cultivating crops, the same crops, or the same crop mixtures. Other plants recorded in the phytolith record speak of the exploitation of tropical forest species. Palms, members of the Chrysobalanaceae, Heliconia, Chusquea, and Trichomanes, are among such taxa. These plants are present in number in the earliest cultural strata of the rock shelters, indicating an ancient pattern of tropical forest occupation and resource use. Significant differences in the site-specific distribution of major ecological indicator species suggested differences in their environmental settings (evergreen forest vs. deciduous forest; coastal vs. foothill zone). Over time, taxa indicating human disruption of the surrounding forest increased at several sites, particularly those located within the deciduous forest zone of the coastal plain where the dry season is long and marked. Pollen and phytolith records from lake cores The archaeological botanical program was successful in documenting some important trends in plant subsistence, and strongly hinted at what appeared to be the significant modification of forest near some habitation sites. However, spotty preservation of pollen combined with the characteristics of the archaeological strata themselves, which spoke of local deposition of plants by seasonal and otherwise discontinuous occupations through time, prevented us from really studying the human/forest association. In 1988 the author and Paul Colinvaux started a program of lake coring in
142
DOLORES R. PIPERNO
central Panama. It was designed to elucidate long-term trends in man—land relationships on the Isthmus, especially the late Pleistocene environmental parameters associated with the initial colonization of the region and the subsequent transition from foraging to food production. We had the good fortune to find a large lake situated in the archaeological study region that held a continuous sequence of deposits dating from 14,000 years ago to the present era. The pollen and phytolith records from this lake, La Yeguada, have been presented and discussed in detail elsewhere (Bush et al. 1992; Piperno et al. 1990, 1991a, 1991b). Here, I will summarize the major findings of these records as they pertain to the archaeology of the region. I will also discuss what may be referred to as the methodology of our interpretation, or how we constructed and analyzed various data sets to arrive at our conclusions. This information is often not included in journal publications due to space limitations. In stratigraphic order from bottom to top, the two long cores of 17.5 m and 13 m raised from La Yeguada consisted of laminated clays, silty-clays, and gyttyas. The presence of: (1) fine undistorted laminae through time; (2) the repeated gross stratigraphy in parallel cores; and (3) a twin series of internally consistent C14 dates achieved directly on lake muds (see Bush et al. 1992; Piperno et al. 1991a, 1991b), all indicate that sedimentation was continuous and undisturbed within the lake, and that undisturbed cores were retrieved successfully. Both pollen and phytoliths were well preserved in the La Yeguada sequence. Figures 6.2 and 6.3 present the phytolith and pollen profiles. In addition to the siliceous remains of plants, a number of other particles of plant origin are routinely isolated in phytolith preparations. One class, particulate carbon, can provide useful insights into the fire history of watersheds. Results of this analysis are also presented in Figure 6.2. Pollen and phytolith identification were achieved by comparison to extensive modern reference collections (3,000 species plus for pollen, 1,500 for phytoliths), and published pollen descriptions (Roubik and Moreno 1991). In addition, Bush and Piperno are carrying out studies of modern pollen, phytolith, and charcoal "rain" in forests of Panama, Ecuador, and Brazil where botanists have identified, mapped, and censused the vegetation (Bush, personal communication, 1991; Piperno 1988a, 1993a, 1993b). This work has resulted in the generation of distinctive signatures for different vegetational types, a reduction in the proportion of unknown pollen and phytoliths, and vital information about vegetational patterns under natural and human disturbance. It can be seen that a large number of taxa are recorded in the Lake La Yeguada phytolith and pollen profiles. Taxa occurring in the phytolith record are quite often not present in the pollen record. Consequently the two lines of evidence complement each other, helping to lend representation to otherwise "silent" species. The increase in species visibility made possible by carrying out the two microfossil analyses in tandem translates into a significant improvement in the
Plant microfossils
lM I 10
I 30
I 10
I 30
50
70
I 10
' 30
bH
hhl
50 I I I 10
30
50
70
I I I 10
30
50^ ^ 70^ ^ I I I 10
.......1... 30
50
70
I 10
I 30
50
143
,„,... I 10
30
50
70
100
150
500 1000
3000 3500 5000 6000 -
9000 10000 11000 12000 13000 1601
14000
l-LU L I |1f.
I, I i1.0, . |3.°i i . . [??
3
°X
7
°
I
I | 10[
30
50
70
I 10
30
50
I 10
50
100
150
1600
14000 I 10 30 i [ I 11 t H I
50 I 10 30 50 70 90 i i i i l i i i i i i i i n i ill m i n i
I I 10 30 L Li i i i i M M
50 n
70 ii i n
90 I 10 i ii i t i n
30 50 I 10 30 ii ii n ii i h i m n
Figure 6.2(a)-(b). Phytolith and paniculate charcoal records from Lake La Yeguada sediments.
50 i n
DOLORES R. PIPERNO
144
1600-
14000
Figure 6.z(c).
overall capability of paleobotanical technique, as well as in the precision of reconstruction. For example, the phytolith record appears to be more sensitive to the presence of, and perturbations occurring in, the primary evergreen and semievergreen forest, because some herbs (for example, Chusquea, Heliconia), ferns (Trichomanes) and trees (for example, Magnolia, the Chrysobalanaceae) characteristic of these formations contribute diagnostic phytoliths in large number. The pollen record of primary forest is typically an assortment of various taxa represented by a few grains of each, hence, their changes in fossil spectra are difficult to track and quantify. On the other hand, arboreal taxa of successional growth (for example, Cecropia, Pilea, Trema) appear to be much more visible in pollen profiles than in their phytolith counterparts. The La Yeguada records show that the late Pleistocene climate (14,000 BP— 11,000 BP) was substantially cooler and drier than present. Magnolia and Quercus (oaks) were important constituents of the montane forest that moved about 800 m downward and replaced the lowland forest during glacial periods (Figures 6.2a, 6.3). Adiabatic lapse rates were used to estimate a 4.50 to 6° temperature depression during the late-Glacial period. The types and distribution of clays and other minerals in the lake sediments indicated that this very substantial cooling went hand-in-hand with a reduction in precipitation, although it is difficult to estimate by how much rainfall was lowered (Bush et al. 1992; Piperno et al. 1991a). A Pleistocene/early Holocene sequence is also available from the El Valle lake,
Plant microfossils
145
Local biozones
YEG4
YEG3
10,000YEG2 12,000YEG1 14,00010 -1
2-1
Pollen influx grains cm + = <100
yr
3
x10
Figure 6.3. A summary pollen record from Lake La Yeguada sediments showing absolute frequencies of major indicator taxa. which while not in the PSM study region, is located only 50 miles to the east. A cool and dry late-Glacial period with montane forest replacement of lowland forest is again evident in the pollen and phytolith records here (Piperno et al. 1991a, 1991b).
146
DOLORES R. PIPERNO
At c. 11,000 BP, a rapid climatic warming and increased precipitation was initiated, as evidenced by declining frequencies of montane taxa, appearance and/or increase of representatives of the lowland forest, and shifts in the lake sediment cation chemistry. The end of rapid vegetational succession and resortment following northern hemisphere deglaciation appears to have occurred by 8600 BP, when montane elements are no longer recorded and lowland forest appears to be fully established. Long considered to have been inured to dramatic environmental perturbations, the humid lowland tropics together with its human inhabitants, as this evidence clearly demonstrates, experienced environmental oscillations at the close of the Pleistocene no less profound than those which occurred elsewhere in the world, and which have long been associated with major economic shifts (for example, Henry 1989; Moore and Hillman 1992). The first signal of a human factor at La Yeguada is coincident with the Pleistocene/Holocene perturbation of climate. It is manifested at c. 11,050 BP by a sudden and large rise of particulate carbon along with plants typical of forest gaps, like Heliconia and grasses (Figures 6.2(b), 6.2(c)). Over 75 percent of the Heliconia phytoliths and a substantial proportion of the grass phytoliths exhibited direct evidence of burning. Now, the interpretation of charcoal in lake sediments is a tricky undertaking. Natural fires, though they are infrequent and of minimal intensity, do occur in tropical evergreen forest (Uhl et al. 1988). They could well contribute something in the way of measurable charcoal to a paleoecological profile. However, we feel that the likelihood of the La Yeguada disturbance patterns being attributable to natural fires is negligible for several reasons. After being virtually absent during the initial three thousand years of the lake's history, particulate carbon suddenly increased by more than several orders of magnitude at 11,000 BP, with high levels sustained across the Pleistocene/Holocene boundary of increasing precipitation. Carbon influx at 11,000 BP is similar to later levels when human influence is certain. Particulate carbon influx at 11,000 BP can also be compared to its counterpart at the surface of La Yeguada (Figure 6.2), and in another Panamanian lake, where human burning does not presently occur in the watershed forest (Piperno n.d.). The amount of carbon annually injected into these sediments by natural fire is currently far lower than that recorded at 11,000 BP at La Yeguada. Other factors weighed heavily in our interpretation. The frequencies of the other disturbance indicators at 11,000 BP plus the proportions of these same phytoliths that are burnt, are much higher than in our samples taken from modern forests whose sole causes of alteration are from natural processes (Tables 6.2 and 6.3; Piperno 1988a, 1993a, 1993b). The evidence from the important indicator taxon Heliconia shows that this plant invades the edges of human clearings much more frequently than the limited open area typically created by
Plant microfossils
147
Table 6.2. Percentage o^ Heliconia phytoliths in modern and paleoecological assemblages, together with the proportions of these showing evidence of having been burned. Modern assemblages from Panama Old evergreen forest from El Cope, Panama. Four samples taken in an area c. 5 km2. Semi-evergreen forest, c. 500 and 200 years old, from Barro Colorado Island, Panama. Twenty-five samples taken every 20 m along 500 linear m of two transects through the forest.
Phytoliths abundant in soils. Over 50,000 phytoliths were scanned and no Heliconia was observed. Phytoliths abundant in soils. Isolated, unburnt Heliconia phytoliths were observed in only a few samples, where extended counting indicated they occur with a frequency of one in every 9,800 to 11,000 phytoliths (0.01%). Their overall frequency in the forest is less than 0.0002%.
Old and successional deciduous forest in Santa Rosa National Park, Guanacaste, Costa Rica. Seven samples taken every 40 m along a linear transect through the forest. Lake assemblages from Panama Lake La Yeguada, Panama. A sediment level dated toe. 11, 000 BP Lake La Yeguada, Panama. A sediment level dated to c. 8600 BP
Phytoliths abundant in soils. A single, unburnt Heliconia phytolith was observed in one sample. Overall frequency of Heliconia is less than 0.0002%.
Heliconia frequency is 2.1%. Over 75% of these are burnt. Heliconia frequency is 7.0%. Over 75% of these are burnt.
Modern assemblages from the Amazon Basin, Brazil Mature evergreen forest on terra firme Phytoliths abundant in soils. Single, soils about 64 km north of Manaus. unburnt Heliconia phytoliths were Fifteen samples taken every 400-800 m observed in two samples, where extended counting indicated a frequency of 0.05%. along several transects through an area c. 10 km2. Overall Heliconia frequency is 0.004% •
natural tree-fall gaps, and is quickly replaced by woody successional taxa in one to three years after the onset of regeneration. In order to maintain a large proportion of these members, which are also burned, it seems that forest openings of a substantial size (when compared to treefall gaps) must be frequently made and then also subjected to further disturbance, including fire, a relatively short time later. It can be seen in Table 6.2 that Heliconia is nearly invisible in modern phytolith assemblages from various kinds of mature tropical forest in Panama and Brazil. (It should be mentioned that Heliconia is not a lake-edge plant in our Panamanian or Brazilian basins, and its
148
DOLORES R. PIPERNO
Table 6.3. Percentages of burnt Gramineae and arboreal phytoliths in modern forests. El Cope, Panama
Barro Colorado Island, Panama Guanacaste, Costa Rica. The only forest where human-ignited fires are believed to have affected the sampling areas over the last hundred years. Fires set in nearby pasture resulted in variable degrees of penetration into the forest at ground levelsTrie Central Amazon, Brazil
No burnt arboreal phytoliths were observed out of c. 2,000 scanned for this trait. Grass phytoliths are very rare. None observed were burnt. Frequency of burnt arboreal phytoliths is 0.2%. Grass phytoliths are rare. None observed were burnt. Frequency of burnt arboreal phytoliths in the samples range from 0% to 4%. Grass phytoliths are uncommon. The frequency of burnt grass phytoliths ranges from 6% to 15%. Frequency of burnt arboreal phytoliths is less than 0.02%. Grass phytoliths are very rare. None observed were burnt.
Note: * Thefirehistory was provided by Daniel Janzen presence in cores almost certainly derives from the watershed forest). It seems that beginning 11,000 years ago the forest was repeatedly subjected to burning and to the formation of small clearings, whose microfossil manifestations are much more consistent with anthropogenic than with natural perturbation. This disturbance pattern continued and intensified into the early Holocene, with particulate carbon and Heliconia reaching peak frequencies shortly after 8600 BP. Remember that the first signs of archaeological arrowroot occur in the region at this time, opening the possibility that such modification was associated with small-scale (perhaps garden) horticulture. Signs of human occupation in the watershed during this time are unmistakable. A bifacial point was found on the La Yeguada lake shore itself, and bifacially worked stone tools and scraper types are present in quantity near the lake. Associated paleoecological and archaeological data clearly point to an established, early Holocene tropical forest adaptation. The continuity and similarity of the disturbance manifestation over the Pleistocene/Holocene boundary and then into the Holocene itself, suggests that it is the result of a unitary cause - culture carried out by the same cultural tradition. If accepted as anthropogenic, the earliest La Yeguada disturbance horizon indicates that occupation and modification of New World tropical forest are ancient patterns, dating to Paleo-Indian times. It indicates that Paleo-Indians
Plant microfossils
149
were not limited to chasing horses and mastodonts in the more open environments of the coastal plain to the south of La Yeguada (Piperno et al. 1991a), but were also exploiting the plants and smaller faunas of dense, evergreen forest at least part of the year. The major perturbations in the record occurring between 8600 BP and the modern era pertain to changes in cultural use of the landscape as subsistence practices intensified. The first of these perturbations occurs at c. 7000 BP when there is a sharp decline of arboreal phytoliths (Figure 6.2(a)). Concurrently, there is a rise in pollen of secondary forest taxa, with levels of particulate carbon remaining high (Figures 6.3, 6.2 (c)). These patterns continue over the next three thousand years. It is important to point out that many of the arboreal phytoliths emanated from the canopy and understory of primary forest, not from taxa typically associated with successional forest after gap creation or other disturbances. Therefore, it appears that the period between 7000 BP and 4000 BP witnessed a reduction in the number of plants from the mature forest. Pollen definitely attributable to Zea mays occurs in the sediments by 5700 BP (remember that maize is also first recorded in the regional archaeological record at c. 7000 BP). The horizon of declining tree taxa may then be interpreted as representing the first removal of trees by slash and burn cultivators. However, the chronology of the earliest signals of slash and burn activity needs to be interpreted with caution, as the sedimentation rate in the cores slows between 7000 BP and 5000 BP, quite possibly as a result of reduced annual precipitation (Bush et al. 1992). Hence, reduction in the number of arboreal phytoliths could be an artificial consequence of fewer phytoliths having arrived at the lake in stream-flow. A consideration of the pollen and charcoal evidence helps to clarify the picture. The pollen record, which is less susceptible to fluctuations brought about by stream-flow changes, shows a large increase of woody secondary growth taxa like Trema, Pilea and Cecropia between 7000 BP and 4000 BP, also suggesting the removal of older forest growth (Figure 6.3). One would not expect these taxa to remain consistently abundant over a three thousand year period as a result of forest change brought about by a drier climate. They are not typical components of primary semi-evergreen or deciduous forests, the vegetation that would be expected to replace the present, potential evergreen forest under a reduced annual precipitation. Also, particulate carbon, which apparently is arriving at the lake both from aerial transport and stream-flow, fluctuates somewhat in abundance between 7000 BP and 5000 BP but its levels generally remain high. And by 5000 BP when phytolith influx has rebounded, levels of arboreal taxa are still very low, a result most probably not influenced by changes in sediment input. Another major boundary in the record occurs at c. 4000 BP. At this time both pollen of woody taxa and particulate carbon influx decline sharply, while levels of arboreal phytoliths remain very low. Pollen and phytoliths from grasses
150
DOLORES R. PIPERNO
increase. Saddle-shaped phytoliths from Chloridoid grasses, which are conspicuous components of present-day agricultural fields in Panama (Piperno 1988a), attain their highest frequencies of the entire sequence. These changes probably indicate that agriculture was intensified to the point where even second growth woody taxa became sparse in a severely deforested landscape. Such an intensification would also account for the loss of particulate carbon. Taken together, the phytolith, pollen, and charcoal records indicate that slash and burn cultivation was practiced in the Lake La Yeguada watershed from the seventh millennium BP onward. By c. 4000 BP the landscape appears to have been severely deforested. Paleoecological data are consistent with the archaeological settlement data. The PSM survey, which was carried out just south of the La Yeguada watershed (Figure 6.1), indicated that there was a fifteen-fold increase in the number of sites occupied between 7000 BP and 3000 BP (Cooke and Ranere 1992b; Weiland 1984). These settlements were small, generally less than 1 ha in size, and were situated on promontories overlooking streams or on interfluvial spurs. They suggest dispersed hamlet clusters of several families, analogous with the organization of modern shifting cultivators. These various lines of data leave little doubt that major changes in the composition of plant taxa around La Yeguada are primarily a result of growing numbers of people engaged in slash and burn agriculture. Human pressure on the landscape apparently continued throughout the prehistoric period, as few trees existed in the La Yeguada watershed until c. 350 BP. At this time a strong forest resurgence is registered in the phytolith record (Figure 6.2). It is undoubtedly associated with European incursion and the rapid demographic decline of indigenous populations. Summary The existence of long-term, detailed paleoecological and archaeological data is rare in tropical studies. Both data offer different kinds of information. Archaeological botanical remains, together with their associated inventory of material culture, are more likely to tell us about the diversity of species manipulated, how they were manipulated and changed morphologically, and how they were processed into food. The strength of paleoecological reconstruction lies in its relevance to the historical landscape. It determines the context of human activity, and provisions information on relationships between the environment and the evolution of subsistence strategies, organization of labor, and demographic trends. None of the data logically aligned with each discipline are easily extractable from the other. The major conclusions drawn from each data set described in this paper are remarkably concordant in indicating the deep antiquity of foraging and farming in the lowland Panamanian forest. If our interpretations concerning the continuity and intensification of land usage are correct, they would indicate in situ
Plant microfossils
151
developments of small-scale (garden) horticulture and slash and burn agriculture, and the presence of societies already organized as shifting cultivators during the seventh millennium BP. These results should hopefully stimulate multidisciplinary studies of a similar nature elsewhere in the Neotropics. Toward seeking the environmental correlates of human activity in the lowland tropics It is apparent that cultural activity in the lowland tropics leaves visible traces on the environment by altering the types and quantities of plant species present in any area. In addition to the obvious effects brought on by slash and burn cultivation, shifts in plant species composition also frequently result from such other activities as the preparation of a living site, transplantation and protection of useful species, and burning of vegetation to increase the density of important game species (e.g., Posey and Balee 1989). In attempting to describe the myriad types of vegetation associated with human activity, as well as driving home the point that, yes, humans living at tropical latitudes have exerted and continue to exert major changes on the landscape, we can lose sight of the fact that all of these major and minor vegetational shifts are a manifestation of different types of cultural behavior. However, if we exploit this point, we can embark on a potentially valuable methodological exercise. As the field archaeologist may locate and assign social or functional attributes to the ceramic and stone implements left by prehistoric tropical dwellers, the paleobotanist with access to improving and diversifying techniques may recover a variety of data that contain those marks of human behavior especially driven by settlement and economic concerns. It is my intent in this section to suggest an approach to tropical archaeology that may provide evidence for cultural activity apart from the availability of any archaeological record. The approach posits that certain past behaviors left identifiable marks on the environment that are present in a microfossil (pollen, phytolith, particulate carbon) record, and seeks to retrieve and decipher these environmental correlates of culture. The evidence should be obtained from lakes, swamps, and other depositional environments where a culturally-created habitat, and not the culture itself, has been the main determinant behind the assemblage of plants that have come to rest in the sediments. Thus, archaeological sites themselves are generally not suitable. However, this is of little importance as one of the objectives is to provide a history of tropical culture independent of an archaeological record. The few records obtained thus far from Panama suggest that the Isthmian region is amenable to this kind of analysis. To widely carry out the approach will require an ability to consistently differentiate cultural interference from natural perturbations. It is hypothesized that on many occasions human interference with the forest is both a qualitatively and quantitatively different phenomenon than natural mechanisms of alteration. For example, logic suggests that there is
152
DOLORES R. PIPERNO
probably a great deal of difference between: (i) the amount of carbon liberated from an annual series of burns associated with game drives or the creation of agricultural plots; and (2) the very long fire return intervals (one in every several hundred to thousand years) associated with natural processes seemingly characteristic of forests studied in the American tropics (Sanford et al. 1985; Saldarriaga and West 1986; Kauffman et al. 1988). The fire return interval is not the only factor that could cause substantial differences in charcoal records. Cultural, especially agricultural, fires can be expected to both penetrate and burn through more of the litter, and include substantially more of the vegetative and woody debris made available from the slash, than would natural fires (Kaufmann et al. 1988; Uhl et al. 1988). This situation would possibly be reflected in the number of charred tree leaf phytoliths present in soil horizons, as well as an increase in the total amount of charcoal present. Modern phytolith records compiled thus far seem to support these premises (Tables 6.2 and 6.3). None of the forests studied are currently under pressure from human activity. Levels of burnt arboreal phytoliths are very low in all of them. The only forest exhibiting any substantial amount of burnt grass phytoliths is one in which human fires started in adjacent pasture are known to have periodically penetrated into the ground level (and/or caused transport of burnt grasses into the sampling areas). Modern carbon production cannot be calculated from the data at hand, but amounts of particulate carbon seem to be very low in all of the forests. Significantly different patterns of plant succession are also to be expected after natural and human impact on vegetation. Many resultant species associations have already been described (for example, Uhl et al. 1982; Uhl, Murphy et al. 1982; Foster 1982). It is worth repeating that a "human" look to plant succession is by no means limited to the initiation and abandonment of agricultural activity. It occurs in contexts associated with persistent human interference of many kinds (for example, Foster 1982; Posey and Balee 1989). In this analysis, the translation from the hypothetical to the real will be accomplished by the systematic construction of modern analogs, which will provide the means for comparing and measuring natural and cultural effects on vegetation. The "modern analog" method has long been used in palynology to compare assemblages deriving from known, natural vegetational formations with fossil pollen assemblages, and to identify the latter on the basis of overall similarity to the former (Wright 1967). The "modern cultural analog" method will attempt to isolate and define phytolith, pollen, and charcoal spectra that correspond with known, contemporary patterns of human-induced vegetation. If close relationships exist between extant cultural vegetation and their resultant microfossil assemblages, then human activity can be identified in fossil records. Extensive records from modern, undisturbed vegetation will be also be necessary to determine that consistent and measurable differences exist today between the
Plant microfossils
153
natural and cultural landscape under any given type of pressure. The significance of parallels and deviations evident in the records will be amenable to verification by formal, numerical analysis (for example, Birks and Gordon 1985). The significant deviations from patterns expected under natural disturbance exhibited by the particulate carbon and phytolith records from Panamanian sites have already been mentioned. To these we can add a set of modern and fossil records from the central Amazon Basin that have recently been compiled (Piperno 1993b). The ability to identify a cultural component to forest disturbance in this and other regions where maize and squash may never have been important would be most useful as the indigenous cultivars are for the most part not detectable in botanical analysis. In a series of modern soil profiles taken from underneath mature, terra firme forest 40 miles north of Manaus, Heliconia and grass phytoliths were nearly invisible and unburnt, as in the Panamanian sequences, and charred tree phytoliths occurred with a frequency of one in every 5,000 to 7,000 siliceous bodies (Tables 6.2 and 6.3). These records can be compared to a 7,300 year old core sequence taken from Lake Geral, which is located a few miles north of the Amazon River near Prainha, Brazil (Piperno, unpublished data). In levels dated to between 7300 BP and 5800 BP, phytoliths from trees dominate, particulate carbon and grass phytoliths are rare, and quantities of burnt arboreal phytoliths are very low (range: 1-5 percent). Heliconia phytoliths are not present. Then, in a level dating to 5800 BP, the amount of particulate carbon suddenly increases by ten-fold. At the same time, phytoliths from grasses and Heliconia increase markedly. Some 19 percent of the grass and 75 percent of the Heliconia phytoliths are charred. These patterns continue over the next two thousand years until c. 4000 BP, when Heliconia phytoliths and particulate carbon increase substantially again and a far higher proportion (40 percent) of the grass phytoliths are burnt. High proportions (20 percent) of burnt arboreal phytoliths are also first recorded at this time. These patterns continue over the next several thousand years. The disturbance manifestations in these initial records from the Brazilian Amazon are very similar to those in the Panamanian Lake La Yeguada. One can posit that they represent an initial human impact on vegetation at 5800 BP (either the earliest occupation in the lake's watershed, earliest modification with fire, or modification and fires persistent enough to register a strong signal), followed by some sort of intensification of landscape use at 4000 BP. Increase of burnt grass and burnt tree phytoliths at this time suggest slash and burn activity, although other interpretations are certainly plausible at present. The missing link in this analysis is obviously analog data from a series of cultivated plots, fallow fields, and other environments that are subject to perennial cultural manipulation. When the necessary analogs have been constructed, they may become the future basis for the generation of a series of testable hypotheses concerning the history of tropical forest occupation.
7
Differential preservation histories affecting the mammalian zooarchaeological record from the forested neotropical lowlands PETER W. STAHL Of all the complex and specialized material culture which the Tropical Forest peoples had developed to cope with their difficult environment only a few axe fragments and a vast mass of smashed pottery remain (Lathrap 1970: 63)
As one of the foremost practitioners of archaeological research in the forested neotropical lowlands, Donald Lathrap was preoccupied with preservation biases in the buried record. He inspired us to judge the effects of differential preservation on the archaeological record, view negative evidence with healthy skepticism (for example, Lathrap 1968a: 77), and avoid conferring undue emphasis on enhanced preservation in arid contexts (for example, Lathrap 1973a: 92; 1974: 115). Lathrap's intense interest in the origin and dispersal of early agricultural systems led him to champion the use of indirect evidence (for example, Lathrap 1970: 48; 1973a: 91; 1973b: 174-176; 1974: 115, 130; 1977: 740), and encourage us to make the most of rare lowland contexts in which preservation was favorable (for example, Lathrap 1973a: 91). A basic axiom in tropical archaeology is that organic residues perish quickly in lowland environments. Archaeologists traditionally dismissed the possibility that organic remains might be recovered in significant quantities. Even where preservation is excellent and recovery intensive, archaeofaunal samples are still often small. The few surviving bones are substantially fragmented, thereby compromising identification of larger animals. It is often only possible to identify these remains to broad zoological categories. These factors severely hamper the interpretive potential of archaeofaunal data in lowland environments, and are almost universally attributed to preservation bias brought about by some combination of excess humidity, temperature, leaching and/or soil acidity (that is, principally due to aspects of the diagenetic environment). Are lowland archaeofaunal collections adequate samples upon which to base inferences about prehistoric environments and faunal exploitation, or are they characterized by intrinsic and systematic biases? The answer has obvious implications for our assessments of lowland neotropical prehistory. For example, zooarchaeological data are directly relevant to recent debates on the alleged deficiency of protein resources in prehistoric Amazonia. It is argued that human population growth, sedentism, and the development of cultural complexity were all impeded by a limited availability of protein from larger mammals (see the
The mammalian zoo archaeological record
155
various arguments in Beckerman 1979; Chagnon and Hames 1979; Gross 1975; Johnson 1982; Meggers 1971; Ross 1978; Sponsel 1986). Before we can begin to evaluate the importance of these issues in prehistoric contexts, we must assess potential biases within the archaeological record. Zooarchaeological samples also contribute to our understanding of prehistoric biodiversity. These data assume crucial importance as we increasingly fathom the extraordinary degree to which lowland environments have been degraded by recent human activity (for example, Coniff 1991; Southgate and Whitaker 1992; Zeidler and Kennedy 1994). This underscores the importance of critically examining the evidence we use to reconstruct prehistoric biodiversity, as the archaeologists' trowel is unfortunately one of the few tools left for assessing a native landscape so dramatically altered since the arrival of Europeans. In this chapter, I explore important variables which can distort the preservation, recovery, and subsequent interpretation of prehistoric mammalian faunal remains in forested lowland neotropical contexts. I use a recently recovered archaeofaunal data set from the western Ecuadorian lowlands to examine the assumption that poor preservation results from hostile conditions of the burial environment. This large sample is highly fragmented, taxonomically rich, but unevenly dominated by small faunas. I suggest that the specific quantitative and qualitative attributes of lowland archaeofaunal samples are not the sole result of diagenetic alteration. Rather, they are more likely the product of complex, interrelated variables. Tropical forest archaeofaunal samples are strongly biased toward a natural abundance and richness of small-bodied faunas. Once transformed into a death assemblage, preservation of the mammalian component in the zooarchaeological record is differentially affected by animal size. These biases can strongly affect the disposition of a bone assemblage both before and after it enters a hostile diagenetic environment. They include the various cultural and natural factors shaping the archaeofaunal record from its initial accumulation to eventual burial, and strongly condition its ultimate reaction to an inhospitable burial regime. Potential variables involve: indigenous cultural strategies of food acquisition, transport, distribution, and consumption; how disposed bone is subsequently altered; and how microvertebrate remains accumulate in buried contexts. Generally overlooked in the literature on lowland archaeology, these factors can systematically control the ultimate deposition of mammalian faunal assemblages. They may alter the proportion or completely mask the presence of larger mammals in buried contexts, thereby biasing our assessments of prehistoric cultural and biological diversity in the forested neotropical lowlands. Faunal assemblages from the Jama Valley, Manabi, Ecuador
The goal of the Jama River Ethnobotanical/Archaeological Project is to investigate the evolution of agricultural systems and sociopolitical organization in one
156
PETER W. STAHL
VERY DRY TROPICAL 3
DRY SUB-TROPICAL DRY TROPICAL SUB-HUMID TROPICAL CONTOUR INTERVAL ANNUAL PERCIPITATION
Figure 7.1. Map of the four major bioclimatic zones located within the project area (after PRONAREG 1978).
tropical lowland area of the northern Andes (Zeidler and Pearsall 1994; also see Pearsall, Zeidler, this volume). The study area is focused on the Jama River drainage of northern Manabi Province in western lowland Ecuador (Figure 7.1). Centered around the modern town of San Isidro, it encompasses at least four distinct life zones, each conditioned by factors of elevation and humidity (PRONAREG 1978; Zeidler and Kennedy 1994; s e e Zeidler, this volume).
The mammalian zoo archaeological record
157
Despite the natural biotic diversity encountered throughout the middle and upper reaches of the Jama Valley, recent human activity has drastically altered the landscape. Selective logging, cattle grazing, and cultivation continue to remove both forest canopy and understory. Today, patches of less diverse secondary forest remnants are confined to isolated ridgetops and slopes, whereas riverine and groundwater forests are all but obliterated by cultivated land (Zeidler and Kennedy 1994). Since 1987, project personnel have systematically collected data on over 150 archaeological sites encompassing a comprehensive 3,300 year sequence of prehispanic habitation. One of the largest and most diverse collections of archaeofaunas yet recorded for the lowland neotropics has been amassed. By the beginning of the 1991 field season, over 24,161 elements representing at least 52 separate taxa, had been identified from twenty-four sites. In particular, the assemblages include a diverse list of mammalian taxa, consisting of at least nine orders, twenty-two families, and twenty-eight genera. Nevertheless, the entire sample weighs slightly over 3 kg, and can fit inside a small suitcase. The assemblage is fairly typical of archaeofaunal data sets from humid lowland settings, in that it is dominated by tiny fragments with relatively poor diagnostic acuity. In the following sections, I discuss characteristics of the entire Jama Valley archaeofaunal assemblage from a variety of chronological and depositional contexts across twenty-four sites. I examine abundance, diversity, size, and fragmentation within the Jama sample as a means to explore the myriad variables which might affect the disposition of archaeofaunal assemblages in the humid lowland neotropics. Numerical abundance In terms of sheer abundance, the Jama Valley sample is relatively unique amongst lowland archaeofaunal collections. The systematic use of intensive water flotation recovery strongly affects the size and composition of the Jama Valley collection (Figure 7.2). Its character is entirely consistent with results obtained by others who have examined the effects of meticulous sampling on differential recovery (for example, Barker 1975; Casteel 1972,1976; Chace 1969; Clason and Prummel 1977; Cooke 1992: 40; DeMarcay and Steele 1986; Falk and Semken 1990; Gillespie 1989; Grayson 1984: 168; Kobori 1979; McKenna 1962; Payne 1972, 1975; Sanchez and Shaffer 1991; Semken and Falk 1991; Shaffer 1992a; Stewart 1991; Streuver 1968; Szuter 1989; Thomas 1969; Watson 1972). A detailed evaluation of how recovery has affected mammalian remains is published elsewhere (Stahl 1992b) and summarized here. The Jama sample is dominated by a category of fragments so small as to preclude their identification to the level of zoological Class. The remaining set of identifiable fragments is comprised principally of tiny skeletal and dental elements from small-bodied taxa. Mammalian specimens (N = 7,943) constitute
iS8
PETER W. STAHL
Indeterminate 10926
Amphibia 387 Mollusca 175 Reptilia 449 Chondrichthyes 3 Crustacea 681
Osteichthyes 3351
Aves 246 Mammalia 7943
Figure 7.2. Jama Valley archaeofaunal totals.
Large
Medium
Small
•n i
Indeterminate 0
500 1000 1500 2000 2500 3000 3500 4000 4500
Frequency(n)
Weight (gm)
Figure 7.3. Relative weights and frequencies of Jama Valley mammalian archaeofaunas arranged by arbitrary size classes (N = 7865, excluding human and recent remains). the majority of identified (to Class) remains. Figure 7.3 illustrates the distribution of taxon size categories within the mammalian subsample. The collection is comprised primarily of small fragments which could not be reliably assigned to any size group. A clear inverse relationship between quantity and total weight exists for those fragments that could be assigned to a size class. In other words, large mammal remains are few in number yet higher in weight. Numerically dominant small mammal remains are relatively light. This is partly explained by the uneven distribution of taxa within the collection. Diversity of mammalian taxa The Jama Valley sample is taxonomically rich, and thus fairly typical of neotropical lowland faunal assemblages. An inventory of the identified taxa is
The mammalian zoo archaeological record
159
Table 7.1. Jama Valley mammalian archaeofaunas arranged by relative size. Large Primates Carnivora Perissodactyla Artiodactyla
Hominidae Felidae Tapiridae Suidae Tayassuidae Cervidae Bovidae
Homo Felis concolor Panthera onca Tapirus
Human Puma Jaguar Tapir
Sus
Pig
Tayassu Mazama Odocoileus Bos
Capra Medium Marsupialia Xenarthra Primates Carnivora
Rodentia
Small Marsupialia Chiroptera Rodentia
Lagomorpha
Didelphidae Bradypodidae Dasypodidae Myrmecophagidae Cebidae Procyonidae Mustelidae Felidae Erethizontidae Agoutidae Dasyproctidae Didelphidae Phyllostomidae Sciuridae Muridae
Echimyidae Leporidae
Didelphis Dasypus Tamandua Cebus Alouatta Felis pardalis Felis yagouaroundi Coendou Agouti Dasyprocta Marmosa Artibeus Oryzomys Rhipidomys Akodon Zygodontomys Sigmodon Rattus Proechimys Sylvilagus
Peccary Brocket deer White-tailed deer Cattle Goat Opossum Sloths Armadillo Anteater Capuchin Howler Raccoons Weasels Ocelot Jaguarundi Porcupine Paca Agouti Mouse opossum Fruit-eating bat Squirrels Rice rat Climbing rat Grass mouse Cane mouse Cotton rat Rat
Spiny rat Rabbit
arranged according to arbitrary size categories (Table 7.1). Excavation strategy clearly influenced the degree of taxonomic richness (see Pearsall, this volume, for effects of recovery strategy on botanical remains). During initial phases of analysis, the majority of identified taxa included smaller vertebrates. The overwhelming majority of diagnostic elements was retrieved from flotation
160
PETER W. STAHL
fraction, including abundant isolated and highly diagnostic dental remains, particularly of rodents (Figure 7.4; Stahl 1992b, 1994). However, regardless of retrieval technique employed, virtually the entire range of faunal taxa was eventually recovered through increased sample size (see Wolff 1975: 198—199; Stahl 1992a, 1992b). The degree of taxonomic richness expressed in the sample is also a factor of many interrelated pre-excavation variables, the effects of which are examined in detail below. The mammalian sample is unevenly distributed across taxonomic categories. Small-bodied faunas numerically dominate the sample at each increasingly exclusive level of identification. Figure 7.5 illustrates: (1) total numbers of identified specimens (NISP); and (2) minimum numbers of individuals (MNI), discriminated at the generic level by separate excavation proveniences (excavation unit and level). Despite minor variations between the different counting techniques, the sample is dominated by small- to medium-sized taxa, particularly rodents. Larger mammals (especially deer and peccary) are relatively well represented, yet not in proportion to their expected dietary contribution. Below, I explore the potential hazards of basing conclusions on this differential representation of mammalian taxa in lowland archaeofaunal collections. Dimensional size and fragmentation The Jama sample is comprised of small and highly fragmented bone, and is thus typical of lowland neotropical archaeofaunal collections. Almost one half (45 percent) of the total sample could not be microscopically identified to the level of zoological Class (Figure 7.2). Mammal bones comprise over 78 percent of the faunal sample by weight, yet total a mere 2,395.07 g. Figure 7.3 illustrates that over one half (52 percent) of the sample identified as mammalian, could not be assigned to an arbitrary size category. Where size categories were reliably determined, bone weight is inversely correlated with frequency across size class. Bone remains grouped by arbitrary size category show a distinct pattern of survivorship. Smaller mammalian taxa are represented by proportionately more complete skeletons (Figure 7.6). The histograms represent percentages; therefore, certain obvious discrepancies are the products of small sample sizes (for example, clavicle, N = 1, patella, N = 4). An indeterminate (skeletal part) category was not established for small taxa. These fragments were so small that they could not be microscopically identified beyond the level of zoological Class. Although it is reasonable to suggest that part of the unidentified (to Class) sample contains very small fragments of larger mammals, the potential for fragments to be assigned to arbitrary size classes obviously increases with skeleton size (Watson 1972; Lyman and O'Brien 1987). An examination of the two major dietary contributors in the sample demonstrates how large mammalian remains are preserved. As previously illustrated, thirty-eight White-tailed Deer (Odocoileus sp.) and twenty-three Peccary
The mammalian zooarchaeological record
Figure 7.4. SEM imagery of selected rodent teeth identified in flotation fraction. Top: Oryzomys spp., upper first molar (50X). Bottom: Proechimys, lower fourth molar (38X).
161
162
PETER W. STAHL
(M) Tamandua (M) Coendou (M) Cebus (L) Fells (S) Zygodontomys (M) Artlbeus (L)Taplnis (M) Agouti (L) Mazama (S) RhlpMomys (M) Fells (L) Panthers (S) Marmosa (M) AJouatta (M) Dasyprocta (S) Akodon (S) Sylvllagus (L) Tayassu (S) Slgmodon (S) Proechlmys (M) Didelphls (L) Odocolleus (M) Dasypus (S) Oryzomys
* 4
*
2 2 2 3 5 g 5 6 g 10 12 13 15
$54 54 54
156,^
244
NISP
(M) Tamandua (M) Coendou (M) Cebus (L) Fells (S) Zygodontomys (M) Artibeus (L)Taplrus (M) Agouti (S) RhlpkJomys (M) Fells (M) Didelphis (M) AJouatta (L) Mazama (L) Panthera (S) Marmosa (M) Dasyprocta (S) Akodon (S)Sylvilagus (L) Tayassu (S) Slgmodon (S) Proechlmys (M) Dasypus (L) Odocolleus (S) Oryzomys
MNI Figure 7.5. Size classes of Jama Valley mammalian genera arranged according to NISP (top) and MNI (bottom).
The mammalian zoo archaeological record
0
10
20
Large
30
40
50
60
Medium
70
80
163
90 100
Small
Figure 7.6. Percentage skeletal representation of Jama Valley mammalian size classes. (Tayassu spp.) were computed where MNI counts were discriminated by separate excavation proveniences (Figure 7.5, bottom). However, these MNI counts decrease dramatically (two each) if the entire sample is considered as one provenience. This suggests that large taxa are dispersed across archaeological proveniences as isolated finds. Figure 7.7(0) suggests that the rare survivorship of large mammal remains is linked to the relative structural density of skeletal portions.1 This is estimated for: (1) the entire large mammal sample by plotting proportionate element survivorship against a reliable set of density measures for a large artiodactyl (Figure 7.7(3)); and (2) the sample of White-tailed Deer via comparison with a hypothetical population of remains predicted to survive on the basis of density (Figure 7. Factors affecting bone preservation in lowland neotropical forests What factors contribute to the differential survivorship of a few, dense and large mammal fragments, and abundant remains of relatively complete small mammals? In the following sections of the paper, I examine some significant variables implicated in this pattern of preservation. I begin by discussing extrinsic aspects
164
PETER W. STAHL
100
A
MAN
:
80
A AST
60-
M A U
A"TPHA'L""
DMTC DM.T T ^
40
DHUMA
A 5
;•
AAP
WTT
:
RAD DTIB
PHL PFEM
20-
:
AIL
i
P
"4
A AHIP
2PHAJ,. .., BCP'
:
:
QAL
A
:
RIB
J 3PHAL A A LUMV
^D
:
^THRV
I
02.
0.4 VD
I
I
I
0.8
Figure 7.7(3). Percentage of large mammalian Minimum Animal Units compared to respective cervid bulk density measurements. of the tropical burial environment. These conditions are generally inimical to the preservation of organic materials, particularly as they interact with intrinsic features of bone material (for example, high fragmentation and small particle size; see also Pearsall and Piperno, this volume). Next, I examine important preburial features in the neotropics which can regulate the entry of abundant and potentially complete remains of small taxa, and increase the overall fragmentation and potential absence of progressively large taxa in the archaeological record. They include a diverse mammalian fauna predominated by small taxa, and indigenous cultural strategies of food acquisition, transport, distribution,
The mammalian zoo archaeological record I
165
1 1 1 1 I 1 • 1 • 1 1 1 1 1 1 . 1 1 1 I : : : • ;
15 ..;
1PHAL :
12
9-
-
N I S P
2PHAL
6-
3-
: :
; SAC
DFEIt
4
3PHAL
k
DMCP
. .m fWITT PTIB 'L DMTT J ! DTIB AkAAST PAT
CALp
:PFEM
S(^ c
HIP
b, r
D
iPMTr1
l i i i i l i i i i i i i i i i i i i i
0
1
2 3 EXPECTED NISP
4
Figure j.y(b). Observed NISP of White-tailed Deer Odocoileus virginianus compared to expected NISP survivorship based on relative density. consumption, and deposition. I then examine buried accumulations of microvertebrate remains, which may be the end products of different variables, and whose lasting signature may be one of potentially greater skeletal completeness. Bone diagenesis in humid forested environments After burial, bone material adjusts to the prevailing conditions of its immediate environment, when new products are derived from reciprocal interactions between buried matter and its encasing matrix. The process and rate of diagenesis
l66
PETER W. STAHL
involve a complex interaction between physical, chemical, and biological variables, both extrinsic to the matrix environment and intrinsic to the buried object (see Buikstra et al. 1989: 172; Henderson 1987: 44; Von Endt and Ortner 1984: 248). Extrinsic burial conditions in humid forested settings can be relentless in their destruction of organic remains. High humidity and temperature combine with acid soil conditions and significant biological activity to create a relatively hostile preservation environment. Both the process and rate of bone decay is further intensified by diurnal and seasonal fluctuations. Water is an essential ingredient in the diagenetic destruction of bone materials. Protein degradation begins as hydrolyzed amino acid bonds are broken and polypeptide chains begin to uncoil (Von Endt 1979: 90). This is further exacerbated by the introduction of waterborne minerals which can weaken the protein mineral bond through substitution in the crystal lattice (Von Endt and Ortner 1984: 248). The amount of water needed to affect protein degradation may be entirely intrinsic to bone material itself (Hare 1980: 212; Von Endt and Ortner 1984: 248). However, increasing the input of extrinsic water exposes the soluble materials of denatured collagen to leaching, thereby altering both the rate and extent of diagenesis (Garlick 1970: 504; Hare 1980: 213; Ortner et al. 1972: 514). High mean daily temperatures accelerate the rate of diagenesis. With environmental acidity and water kept constant, an approximate 70 increase in mean annual temperature affects a quadrupled rate of nitrogen loss, an obvious indication of protein degradation (Ortner et al. 1972: 517). Increased temperature also promotes the rate of racemization, in which amino acid molecules structurally change with diagenetic detachment and reattachment of hydrogen atoms (Von Endt 1979: 78). It has been suggested as a "rule of thumb," that reaction rates are doubled for each io° raise in temperature (Von Endt and Ortner 1984: 249)Humidity and heat combine with external soil acidity and significant biological activity to further enhance the rate of bone diagenesis. High vegetational turnover in forested environments contributes to low soil pH, and changes in soil chemistry can enter the hydrolytic reaction, thereby accelerating bone diagenesis (Von Endt and Ortner 1984: 249). Hydroxyapatite crystals become more soluble with increased acidity, exposing calcium and phosphates to ground water leaching (White and Hannus 1983: 316, 319, 322). The exchange of elements between the inorganic phase of bone material and soil is related to environmental acidity, as trace elements become more mobile (Buikstra et al. 1989:159; Lambert et al. 1985b: 91). Water, temperature, and pH also govern the growth and activity of Clostridium bacteria, responsible for the production of collagenase enzymes which degrade the organic collagen of buried bone (Garlick 1970: 504; Rottlander 1976: 86). Microorganisms in moist soils infiltrate buried bone and excrete organic acids which leads to further destruction and dissolution of apatite (Sillen 1989: 220; Garland 1987: 122). Macroscopically, plant roots can seek out buried
The mammalian zoo archaeological record
i6y
bone materials as nutrient sources (Morlan 1980: 36), and are notorious for acidic surface destruction (Schiffer 1987: 187). Two intrinsic characteristics of buried mammalian bone assemblages further exacerbate the extent and rate of bone degradation in the humid neotropical lowlands: (1) a potentially high proportion of small animal remains; combined with, (2) intense fragmentation. Fragment size is crucial to bone diagenesis, as hydrolytic destruction, loss of protein, and elemental exchange proceed at a rate which is inversely proportional to bone size (Hare 1980: 214; Von Endt 1979: 96; Von Endt and Ortner 1984: 252). Extensive fragmentation affects the degree and rate of diagenetic reactions, due to the relative ease with which extrinsic environmental factors can enter internal bone structure (Von Endt and Ortner 1984: 252). Element exchange and leaching proceed at different rates for different materials. Permeable bone, especially highly porous cancellous material exposed through fragmentation, is significantly degraded through diagenesis. Less permeable dental enamel, relatively dense cortical structure, and intact bones, all weather diagenesis much better (see Buikstra et al. 1989: 174; Klein 1989: 374— 375; Lambert et al. 1982: 291; 1985a: 478; 1985b: 87; Marshall and Pilgram 1991: 160; Parker and Toots 1980: 199; Sillen 1989: 212; Von Endt 1979: ^j\ White and Hannus 1983: 318). The mammalian community of the forested neotropical lowlands Prior to entering the burial environment, the mammalian zooarchaeological record is affected by unique factors of the neotropical forest community. Large mammals are relatively small, rare, unpredictable, and elusive. Small mammals numerically dominate nonvolant arboreal and terrestrial niches. The sheer abundance and richness of small-bodied mammals has important implications for archaeological recovery and estimates of prehistoric diversity (Stahl 1992a, 1992b). The majority of potentially diagnostic bone fragments originate from taxa with average Head-Body dimensions below 200 mm. Small mammals in the forested neotropics. Tropical forest habitats are taxonomically rich. Fleming (1973: 55^~557; see also Hershkovitz 1972: 332) estimates that the neotropical mammalian forest community can accommodate twice as many species as temperate communities. The majority of neotropical species are small. For example, over y<) percent of mammalian forest species in Central and South America are under 500 g in total live weight, with threequarters of these animals well under 100 g (Stahl 1992b). Numerous factors are implicated in the relationship between richness and small body size, and the physical constraints imposed by a mechanical adaptation for life in arboreal habitats are particularly crucial. Within tracts of mature tropical forest, the bulk of energetic activity is locked into tree crowns. In the mammalian community, terrestrial activity decreases as
l6S
PETER W. STAHL
the extent of canopy exploitation increases (Eisenberg 1980: 41; 1989: 432; 1990: 361; Eisenberg and Thorington 1973). Lower weight and efficiency of movement are critical for effective exploitation of arboreal habitats. These factors limit optimum body size (Eisenberg 1978: 138). Not surprisingly, the diverse arboreal component of the neotropical forests is dominated by bats (Fleming 1973: 558). In terms of abundance and biomass, rodents and to a lesser extent marsupials, dominate the nonvolant component of small mammals (Fleming 1975: 271-273, 275). Large mammals in the forested neotropics. The few mammals that achieve any appreciable biomass are characterized by low density, solitary or cryptic behavior, and a natural susceptibility to rapid depletion through overpredation. Overall size diminution is a characteristic shared by all neotropical mammals, and is particularly notable for large taxa (Hershkovitz 1972: 372). Even the larger canopy dwellers, though capable of contributing significantly to overall standing biomass, are only intermediate in size. Mammals that incorporate a large amount of leafy matter in their diets (for example, Howler Monkeys, or obligate folivores like Sloths) must maintain a minimal body size in order to digest large amounts of an energetically inefficient and potentially toxic diet. Nevertheless, arboreal habitats impose biomechanical constraints on the upper limit of body size (Eisenberg 1978: 138,146; 1983: 268; Kay and Hylander 1978: 175; McNab 1978: 158; Parra 1978: 209). Large mammals are rare, elusive, and susceptible to overpredation. They are few in number and widely dispersed. Increased body mass tends to be inversely correlated with numerical density (Eisenberg 1980: 40; 1990: 360), and positively related to increased home range size (Eisenberg 1989: 432). Scarcity of large terrestrial mammals is augmented within mature tropical forests where high plant diversity is combined with decreased vegetational activity beneath the canopy. Terrestrial habitats are therefore less capable of sustaining significant numbers of large mammals, and with few exceptions, large herbivore density decreases with the need to increase foraging territory. Neotropical forest mammals are difficult to capture because of their solitary and often cryptic nature. Preferred human quarry such as the terrestrial tapir, deer, capybara, paca, and agouti, are inclined toward solitary behavior or minimal aggregation. They are also cryptic, shy, easily alarmed, and crepuscular or nocturnal, especially when pursued by humans (Eisenberg 1989; Emmons and Feer 1990). Exceptions include the peccaries, notably the White-lipped variety, which can accumulate in herds of up to 300 individuals (Emmons and Feer 1990: i59;Sowles 1984:153). Optimal herds are prized sources of terrestrial protein, yet appear infrequently as the amount of food needed to maintain herd size demands a wide foraging space (Emmons and Feer 1990: 159-160; Kiltie 1980: 542; Sowles 1984: 152). Large mammalian canopy dwellers like sloths and monkeys are also highly cryptic, particularly when viewed from the ground. As many of these
The mammalian zoo archaeological record
169
species tend to utilize the uppermost canopy and/or remain generally inactive, they become notably inconspicuous (Emmons and Feer 1990: 35, 125, 130). Large terrestrial forest mammals are markedly susceptible to rapid depletion via predation. In particular, the large herbivores tend toward longer life spans and repetitive reproductive periods with comparatively small litter sizes and extended intervals between births (Eisenberg 1980: 40; 1989: 432). These factors, in combination with low abundance and density, increase the danger of rapid overpredation. In contrast, populations of small mammals are relatively rich, abundant, dense, and subject to rapid biological turnover. The discussion to this point has focused on mammalian communities within stands of mature, interfluvial forests. Mammalian diversity may differ widely as a result of numerous local variables (Emmons 1984: 215—216; Gentry and Emmons 1987:226); however, species richness is strongly related to the increased structural diversity of mature forest (for example, MacArthur 1972: 177; August 1983; Eisenberg 1990: 359). Overall species richness is decreased via removal of the canopy, reduction in habitat complexity, and elimination of arboreal niches. Llanos habitats, where arboreal mammalian biomass reaches its lowest point, are dominated by terrestrial herbivores (Eisenberg et al. 1979: 197). Second growth forests can support unique species and community structures whose mammalian biomass is dominated by terrestrial herbivores (Eisenberg et al. 1979). Potential maximum diversity can thus be achieved in areas where second growth and mature forest intermesh (Eisenberg 1989: 434). Significantly, such areas often include abandoned gardens and residential clearings. Indigenous strategies of mammalian resource exploitation in northwestern South America1 The cultural selection of game items from a menu of available faunas, strongly affects how the mammalian zooarchaeological record is eventually comprised. Small mammals often constitute a significant dietary contribution. In particular, these include various forms of smaller arboreal and terrestrial rodents, as well as lagomorphs, and the smaller callitrichid and cebid monkeys. Both in number and weight, medium to large mammals comprise the most frequently killed class of animal food in the lowland neotropical forests (Vickers 1984: 370). Notable inclusions are the White-lipped and Collared peccaries, tapirs, pacas, agoutis, deer, larger cebid monkeys, capybaras, sloths, anteaters, and armadillos. I discuss below the numerous factors which influence decisions governing prey selection. Indigenous exploitation and consumption practices can lead to consistent biases in the patterning of mammal remains. Small game has a greater chance of being transported and processed either whole, or with relatively greater skeletal integrity. Large game is subjected to intensive dismemberment and fragmentation. Consequently, larger prey items have a greater chance of exhibiting diminished skeletal integrity. Variables governing the extent of dismemberment
170
PETER W. STAHL
for purposes of transport, processing, and eventual consumption, vary consistently with animal size. The hunt. An interrelated set of variables influences the decisions involved in acquiring animal protein. Where and when is game pursued, and over what length of time? What method is used? How many hunters are involved? The aggregate answer is an entry point which affects subsequent trajectories of mammal exploitation. Game animals may be pursued at varying distances from the principal residence. Young specimens or mature wild game may be intentionally maintained in residential context, often in anticipation of eventual consumption. Functioning and abandoned gardens are prime areas for both the specific and opportunistic pursuit of game, as certain mammalian species are attracted by agricultural produce and successional conditions. Decisions concerning the pursuit of game in these and other areas may be further affected by the need to perform additional tasks like land clearance and garden maintenance. Also, seasonal variations in rainfall patterns can hinder hunting success through increasing the cryptic qualities of game, or facilitating capture through enhanced visibility of tracks. Under certain conditions, animal populations may concentrate on high ground above seasonally inundated areas, where concentrations of mammalian game are hunted closer to the residence. Studies conducted in interfluvial forests suggest that settlement age and differential resource depletion are major factors affecting the numbers, kinds, and locations of animal game (for example, Hames 1979, 1980; Hames and Vickers 1982; Vickers 1980). Local game resources, particularly large mammals, are relatively abundant during the initial years of residence. However, they rapidly decline due to over-exploitation. Hunters are required to travel greater distances in pursuit of large mammals while simultaneously intensifying the acquisition of smaller-sized game in the intervening hunting grounds. The net result is that with increased settlement age, large game is procured at increasingly greater distances where biomass concentrations are higher. Small game is increasingly derived from localized areas. The location in which a hunter chooses to pursue a particular game item interacts with other variables such as the method of capture, the number of participating hunters, and the temporal length of the hunt. With some overlap, the traditional hunting arsenal of the northwestern lowland areas tends to be species specific. The kind of game a hunter sets out to pursue generally governs the choice of weapon he takes on the hunt. The chosen weapon guides the hunter's choice of game. The species specificity of portable hunting equipment is loosely organized around prey size. The largest terrestrial mammals, now hunted with shotguns, were once pursued with specialized, and occasionally poisoned, lances and spears. As this arsenal demands strong thrusts at close range, dogs were often used to flush out and drive prey toward the hunter. Large game is also
The mammalian zoo archaeological record
171
opportunistically captured with blowgun and poisoned darts, or bows and arrows with poisoned or specialized tips. Large terrestrial species like tapirs, whose movements can be predicted through the location of habitual runs, paths, or salt licks, are also procured with pit and blade, trip, and deadfall traps. Medium-sized arboreal mammals are the predilected quarry of blowguns and specialized bows and arrows. Blowgun hunting is a solitary pursuit, carried out in mobile hunts, or from blinds. Throughout the region, muscle-relaxing dart and arrow poisons like curare are used to overcome the death grip of arboreal primates. Shafts are also notched to avoid the dart being pulled out. The silence of this technology not only permits multiple shots at solitary prey, but also does not scare away other members of gregarious taxa. The terrestrial component of medium-sized mammals is more flexibly pursued with lances, clubs, and an assortment of slip, tunnel, fence, deadfall, pit, guillotine, trigger, and net traps. Nightly torch-lit pursuit of nocturnal terrestrial and aquatic game is also described. Small arboreal game can be pursued by solitary blowgun hunters, with or without the use of muscle-relaxing poisons. Small terrestrial faunas can be incidentally captured by hand, with clubs, or purposefully pursued with a host of ingenious pit, slip, snare, tension and spring, deadfall, fence and noose, tunnel, trigger, and net traps. The choice of game to be pursued, and the selection of an appropriate weapon, are crucial determinants in hunting group size. Blowgun hunting of smaller taxa is best undertaken as a solitary pursuit, or in nominal groups. Large game, like tapir, peccary, deer, and anteater, is most effectively captured through cooperative group effort. Certain behavioral characteristics of a particular game item can be crucial determinants in hunting party size. The chance location of gregarious peccary herds is an immediate excuse for the organization of communal hunting parties. Festive occasions which demand the procurement of greater amounts of meat, also require organized trips by larger aggregates of hunters. Any combination of these decisions can affect the duration of the hunt. For example, an organized communal peccary hunt at a distant hunting ground, or the festive procurement of large amounts of meat, both demand protracted periods of hunting effort. In contrast, localized procurement of game animals is less demanding of large blocks of time. Very often, hunters set out in search of large game, with expectations diminishing as the tropical day progresses. Differential transport. Decisions involving the location of the successful hunt, the choice of prey, the quantity taken, the duration of hunting effort, and the size of the hunting party, affect the degree of game dismemberment. Important factors include simple transport logistics, the application of preservation technologies, and the even distribution of prey shares amongst all participants. Locally acquired animals present fewer logistical concerns for transport,
172-
PETER W. STAHL
especially as this implicates smaller taxa. Small animals are relatively easier to return whole, or with increased skeletal integrity. Both the increased size of game, and the greater distance travelled by the hunter to procure it, are crucial factors determining the degree to which the animal is disarticulated. These factors contribute to the greater differential processing of larger animals. Organized hunts generally take place at increased distance, and focus on large game hunted over longer periods of time by a greater number of hunters. Large game is most easily transported in smaller, often selected pieces. Both solitary and group hunters who bring down large and/or gregarious animals far from the residence (especially late in the day), gut their quarry and carry back what they can by themselves. The remainder is often concealed until the hunter(s) return to complete butchering and transport. On lengthy trips to distant hunting grounds, captured game must be preserved in the hunting camp prior to returning home. This is particularly crucial in the humid tropical lowlands, where the onset of decomposition and spoilage is rapid. As discussed below, preservation technologies always involve the intensive processing of large animals. The larger the animal, the more comprehensively it is cut into pieces and smoked, resulting in the discard of a substantial portion of osseous and non-edible tissue. In contrast, small animals need not be processed as intensively prior to smoking and drying. The increased number of hunting participants can also necessitate the degree to which larger prey is dismembered. This is often more a factor of game distribution than of simple transport logistics or processing technology. Each participant of a group hunt, which by definition is usually in the pursuit of larger quarry at greater distances, is entitled to a share of the game whether he took part in the kill or not. Issues of ownership must also be considered, as when a game item may stray into the hunting territory of others. This could necessitate the further subdivision of larger prey through obliged sharing with its owners. Distribution. The supply of desired meat protein is highly punctuated, thereby contributing to its status as the shared food par excellence. Large, visible game items are an immediate cause for excitement when they appear in residential context. The communal sharing of large game amongst kin and village members alike confers prestige, and minimizes accusations of niggardliness. This food is invariably partitioned and dispersed throughout residential contexts. The communal consumption of abundant meat sources is certainly important in an environment where rapid consumption can offset immediate spoilage. In contrast, whole specimens of small game can enter residential context less visibly. Though generally shared, they may be surreptitiously consumed in localized nuclear familial settings. Consumption. The differential processing of large and small animal food packages becomes pronounced during consumption. The smaller the animal, the
The mammalian zoo archaeological record
173
higher the chance that it will survive with greater skeletal integrity. The larger the animal, the greater the degree to which it is subjected to intense processing and decreased skeletal integrity (for example, Beckerman 1980: 94, Chart 8). The relative abundance of meat generally governs the kind of processing technology employed. When abundant, either at hunting sites or in residential contexts, meat must be preserved or it will quickly spoil in the tropical heat and humidity. Preserved meat, which is commonly stored under roof rafters or suspended receptacles, may last for some months under optimal conditions. Smoking and drying are the most common preservation technologies employed. Prior to drying, large game is dismembered and/or filleted, and individual pieces are smoked on a rack suspended over the fire. During consumption, preserved items are further processed into smaller edible portions. Small game is often preserved whole, or with far greater skeletal integrity. Roasting of meat for immediate consumption also contributes to a differential pattern of size-related processing. The entire gutted or ungutted carcass of smaller game is simply impaled on a stick or wrapped in a leaf and roasted. Large game must be more thoroughly processed. Prior to butchery, large animals are commonly singed, washed, and gutted. Certain parts of the gut content may be roasted in leaves and eaten. Invariably, all large meat items must be reduced to appropriate serving portions. Throughout the area, meat is commonly consumed in the ubiquitous soup, stew, or gruel. In these cases, it is important to reduce the size of the portion to fit the appropriate pot, thereby favoring the increased integrity of smaller items. Whereas small meat packages may enter whole, large items are invariably segmented, tossed in as pieces, flaked, or pulverized. Further fragmentation of bone for removal of marrow and osseous nutrients is variable. The Ache of eastern Paraguay smash marrow-bearing bones of large animals at mid-diaphysis or near their articular ends. Shaft portions of small animals are opened either via mid-shaft breakage, or through removal of articular ends by biting or with a knife blade (Jones 1984: 103). The Shuar of eastern Ecuador and Peru never eat or chew bones, nor do they suck marrow. Osseous remains of meals are carefully collected by the women, and disposed of in the river (Bianchi 1988: 20, 21, 49, 145, 150).
Further trajectories of bone material in residential contexts. After consumption, bone material may be discarded or retained for ornamentation, magical aids, medicine, and tool or weapon manufacture. Within the Jama sample, dental and associated cranial elements from variously-sized jungle cats were specifically curated. These include the rostrum of Felis pardalis in a burial association, as well as basally drilled feline canines. Discarded bone remains are affected by a range of variables, particularly regular and intense episodes of trampling and sweeping (for example, DeBoer and Lathrap 1979; Stahl and Zeidler 1990). Remnants discarded in and around areas of consumption are exposed to intense scavenging. Lowland forested
174
PETER W. STAHL
contexts are characterized by efficient nutrient cycling, thereby accentuating the degree and speed at which released nutrients are reincorporated into plant biomass. Competition over exposed organic detritus is severe. Ubiquitous canine scavengers may play a decisive role in the spatial movement and differential reduction of bone remains (e.g., Lyon 1970; Smole 1976: 179). They are routinely implicated in the density-dependent destruction of larger mammalian bone items (for example Binford 1981: 60; Brain 1981: 19; Lyman 1984: 279; Marean and Spencer 1991; Marean et al. 1992: 111, 117). Conversely, certain lowland groups never allow their hunting dogs to eat meat or bones, further restricting pigs and chickens from accessing discard areas (for example, Bianchi 1988: 21,49; Oberem 1971: 163). Within the entire Jama sample, only ten fragments of bone exhibit some sort of carnivore-related damage. It may be significant that one half of these marks were located on fragments minimally large enough to be identified as large mammal remains. Although predators need not leave visible marks on bone remains, this might suggest that the rarity of such evidence may be related to extreme fragmentation and small particle size. Exposed bone remains may be further subjected to patterned destruction and dispersal through exposure to the intense sub-aerial weathering of tropical climates. The nature and extent of destruction is potentially controlled by a number of intra-skeletal and inter-taxon variables. These include both relative density and the ratio of surface area to volume and body size (for example, Behrensmeyer 1978: 152—153; Gifford 1981: 417; Lyman and Fox 1989: 297, 301; Marean 1991: 685). The extent of weathering destruction predictably intensifies along a gradient of dense bones with low surface area from larger mammals, to light bones with high surface area from smaller mammals. Visible weathering of bone within the Jama sample is restricted to seven proveniences, at least one of which is recent. The density-mediating nature of weathering may be reflected in the predominance of cortical, diaphyseal, and relatively durable long bone segments, all from large mammals. Augmented destruction via weathering may also contribute to an increased capacity for dispersal by hydrodynamic sorting (Lyman and Fox 1989: 303). Susceptibility to sorting is further correlated with variation in shape, size, and relative density (for example, Behrensmeyer 1975: 499; Boaz and Behrensmeyer 1976: 59; Voorhies 1969: 69). Microvertebrate accumulations in buried contexts Preserved remains of small mammals are not necessarily indicative of human exploitation. These remains can enter burial contexts in a variety of ways, thereby requiring zooarchaeologists to establish distinct criteria for identifying different agents of accumulation. Many non-cultural mechanisms may be responsible for autochthonous and/or allochthonous accumulations of micromammalian remains in buried assemblages. These mechanisms also can contribute to the increased skeletal integrity of small taxa. Microvertebrates can enter
The mammalian zoo archaeological record
175
buried contexts with only minor skeletal modification via accidental death (Andrews 1990: 2—4), such as entrapment in pitfalls (Whyte 1988: 79, 86; 1991). Increased skeletal integrity of microvertebrates may be produced through allochthonous transport under slight hydrodynamic conditions. This is usually associated with patterned sorting and bone breakage, depending upon water velocity, shape, size, and osteologic construction (Andrews 1990: 18-19; Dodson 1973:17; Korth 1979: 249,253,262; Pratt 1989:130; Wolff 1973). Microvertebrate remains may accumulate through non-human predation. Typical patterns of breakage, corrosion, and elemental loss are governed by interactions between predator and prey. These can involve size and age related phenomena, prey availability and capture, method of consumption, differential bone densities, and the area and manner of deposition (for example, Andrews 1990: Chs. 2-3; Andrews and Evans 1983; Brain 1981; Dodson and Wexler 1979; Emslie 1988; Fernandez-Jalvo and Andrews 1992; Fisher 1981; Hockett 1991; Hoffman 1988; Korth 1979: 240; Kusmer 1990; Levinson 1982: 115-117; Maas 1985: 125; Mellett 1974; Pratt 1989). Unfortunately, most criteria for identifying mechanisms of accumulation and modification in microvertebrate assemblages are equivocal. The use of any single criterion for establishing human exploitation (for example, Vigne and MarinvalVigne 1983), is often confounded by cultural and non-cultural factors alike (for example, Grayson 1988: 26; Hockett 1989: 41-42; Payne 1983: 151; Szuter 1984: 156; 1989: 222, 312). Multiple lines of evidence provide the most reliable insight into microvertebrate exploitation by humans. These include patterned element representation and damage, comparison with non-human accumulated assemblages, associations with butchery scars, charring, and artifact manufacture, patterns of age and size, natural history and ecology, and contextual associations (for example, Dansie 1984; Grayson 1991:108; Hesse 1985: 43; 1986: 79; Hockett 1991: 674; Jones 1983: 182-184; 1984: 103, 107; Mengoni Gofialons 1983: 325; Payne 1983; Semken and Falk 1991; Shaffer 1992b; Simonetti and Cornejo 1991: 94; Szuter 1984: 154-156; 1989: 217-222; Thomas 1971: 368-369). The majority of small mammal remains in the Jama sample were carefully excavated from one archaeological context (see also Pearsall, this volume). Feature 5 (Pechichal, M3B4—on) is a 1.5 m deep bell-shaped storage pit with steeply outsloping sides and a restricted (70 to 80 cm) orifice. The pit appears to have been dug into a layer of volcanic tephra and originally used for storage. Eventually, it was rapidly infilled with as many as nineteen separate refuse lenses, clearly associated with a single first millennium AD occupation. Comprehensive flotation recovery yielded a rich and abundant faunal component including crustaceans, amphibians, reptiles, birds, and mammals from at least ten Orders, eleven Families and fifteen Genera. The small rodent faunas are characterized by greater skeletal completeness, whereas the medium and large mammals survive as isolated elements distributed throughout the pit fill. The small rodent remains are unevenly distributed throughout the vertical lenses,
176
PETER W. STAHL
with no apparent concentrations either at the top or the bottom of the feature. Numerous remains of small terrestrial rodents like Oryzomys spp., are located virtually throughout every level. This might suggest their accidental blundering into the pit (for example, Whyte 1988,1991); however, remains of bats, birds, and obvious food animals (that is, Alouatta, Cebus, Tayassu, Odocoileus, and Mazama) which are less susceptible to pitfall entrapment, are also vertically associated throughout the numerous lenses. Burned Dasypus cranial and body armor are concentrated toward the bottom of the pit, which may suggest infilling of consumed refuse. However, the precise mechanisms of microfaunal accumulation, whether through accidental entrapment, secondary removal, or multiple agency, still remains equivocal. This emphasizes the need for caution in interpreting patterned survivorship of mammal remains. Summary and discussion
Intensive flotation recovery from archaeological contexts throughout the Jama Valley is singularly responsible for a large archaeofaunal assemblage dominated by small mammal remains. Otherwise, the Jama River sample is similar to other taxonomically rich, numerically modest, dimensionally small, and highly fragmented neotropical archaeofaunal collections. Archaeologists customarily attribute these characteristics to extrinsic aspects of the burial environment (notably soil acidity, intense humidity, and high temperature). I suggest that blanket statements, which attribute poor preservation exclusively to diagenetic phenomena, can conceal what is in all likelihood a very complex issue. If we rigidly adhere to this basic axiom of neotropical archaeology, we effectively suppress further avenues for exploring potentially informative portions of the buried record. The Jama River archaeofaunas are a case in point. Had intensive collection techniques not been systematically applied throughout the excavated proveniences, the axiom would have remained a selffulfilled prophecy. Carefully analyzed flotation fraction yields a substantial faunal component which contrasts with a small population of remains recovered through dry-screening. This begs a central question: Why has the Jama River archaeofaunal sample survived in this way? Abundant small mammal remains survive with greater skeletal representation. The relatively few large mammal remains survive in fragmented, densitydependent fashion. A complex set of pre-burial variables can precondition size and density-related survivorship. The community of forest mammals includes a rich and abundant small mammal component. Large mammals are relatively small, rare, elusive, and unpredictable. After the animal is selectively removed from the living assemblage, certain factors can further systematically bias the nature of survivorship up to the point of burial. These factors may regulate the entry of bones into buried contexts by: (1) accumulating abundant and potentially complete remains of small taxa; and (2) increasing the overall fragmentation, density-mediated preservation, or total absence of large taxa. The positive
The mammalian zoo archaeological record
177
identification of implicated mechanisms often remains equivocal; however, a number of interrelated variables may be suggested. They include indigenous human game acquisition, carcass transport, preparation, distribution and consumption, bone curation, sweeping and trampling, intense scavenging of exposed remains, subaerial weathering, hydrodynamic sorting, accidental death and entrapment, natural transport, and non-human predation. The substantive issues raised in this paper are not new. Their methodological orientation can be traced to an earlier series of seminal articles written by a vertebrate paleontologist who sought to recognize and interpret patterns in differential animal bone preservation at archaeological sites. Theodore E. White (1952,1953a, 1953b, 1954,1955,1956) related patterns observed in the archaeological record to anatomical differences both within skeletons, and between varying sizes of animal carcasses. He considered patterned discrepancies as the potential products of differential transport, carcass disarticulation, processing and consumption (particularly of lighter meat and grease-bearing elements), carnivore attrition, tool manufacture, sampling, and preservation. Throughout his publications, White consistently stressed the potential role that "accidents of preservation" might play in shaping the nature of zooarchaeological samples. These "accidents" are of fundamental importance to archaeological interpretation, and are currently a major focus of research; however, they have as yet not been seriously considered in the virgin study of lowland neotropical zooarchaeology. Can we reliably infer the richness and/or relative abundance of prehistoric animal taxa from lowland archaeofaunal collections? Both measures of biodiversity are confounded by variability in sample size. This is particularly relevant for our meager neotropical archaeofaunal samples, as small samples are generally skewed toward the most abundant class of items (Grayson 1981a: 82; 1984: 116). Usually, this consists of unidentifiable fragments, with the remaining fraction dominated by a restricted number of taxa (Stahl 1992a). Augmenting sample size increases the likelihood of adding rarer taxa to the assemblage, for scarcer categories are only gradually recovered with increased sampling (for example, Grayson 1984: 132; Leonard and Jones 1989; Wolff 1975). This can be achieved through expanding excavation, or intensifying recovery techniques. However, in both cases, if we fail to consider biases which affect the different components of the archaeological assemblage, we might be led to erroneous conclusions. The often equivocal identification of mechanisms involved in assemblage accumulation and modification seriously hampers our ability to make reliable interassemblage comparisons (Grayson 1981b: 33). This is certainly magnified by the likelihood of a systematic pattern of differential survivorship based upon overall body size. Inferring that large mammals contributed little to prehistoric subsistence could further support a basis for potentially groundless claims about environmental limitations. These inferences could then be spuriously connected to unreliable assertions about prehistoric population density, sedentism, and cultural complexity. We can avoid some of the problems caused by small, variable, and biased
i78
PETER W. STAHL
Table 7.2. Natural histories of recent western Ecuadorean mammalian genera identified in Feature 5. Taxon
Typical natural history
Didelphis
Variable, including multistratal forest, arboreal, scansorial, terrestrial, solitary, nocturnal, omnivore, nests in tree cavity or burrow Secondary, gallery, to multistratal forests, arboreal and terrestrial, solitary, diurnal and nocturnal, insectivore, nests in hollow trees Open areas to forest, terrestrial and fossorial, small groups, diurnal and nocturnal, omnivore Dry deciduous to multistratal evergreen forest, arboreal, groups, diurnal, frugivore and folivore Mature and disturbed forest, arboreal, groups, diurnal, omnivore Semi-arid brush to forest, croplands, terrestrial, solitary, nocturnal, browser Deciduous to evergreen forests, commensal, semi-arboreal, scansorial and terrestrial, solitary, nocturnal, omnivore, nests arboreally and in roofs Dry deciduous to multistratal forest, commensal, arboreal and scansorial, solitary, nocturnal, omnivore, nests in trees Open grass and scrubland, croplands, terrestrial to shallow subterranean, nocturnal and diurnal, omnivore Open grass and scrubland, cropland, commensal, terrestrial, diurnal and nocturnal, omnivore Forest clearings, croplands, terrestrial, nocturnal, omnivore, nests in trees or burrows Humid forest, croplands, terrestrial, solitary and paired, diurnal and crepuscular, herbivore and frugivore Forest and riverine areas, terrestrial, solitary, nocturnal, omnivore, food caches in burrows Variable, including savanna, forest, and cropland, terrestrial, gregarious, nocturnal, diurnal in rainforest, omnivore Mixed cover, cropland, terrestrial, solitary and small groups, diurnal, nocturnal, crepuscular, grazer and browser Shrub savanna to multistratal forest, terrestrial, solitary or pairs, diurnal and nocturnal, browser and frugivore
Tamandua Dasypus Alouatta Cebus Sylvilagus Oryzomys Rhipidomys Akodon Sigmodon Zygodontomys Dasyprocta Proechimys Tayassu Odocoileus Mazama
The mammalian zoo archaeological record
179
sample sizes. If we consider the chance preservation of large mammal remains in the buried record as part of an underlying pattern concealed by a complex set of preservation pathways, and consider the mammalian archaeofaunas as attributes, we can focus on spatial and/or temporal variation amongst "suites" of taxa (for example, Grayson 1981b: 34; Lyman 1986; Lyman and Livingston 1983: 362; Stahl 1991). If we further bear in mind some requisite cautions (Grayson 1981b: 35-36), particularly the non-randomizing effects of various cultural and natural modes of assemblage accumulation, we can capture a broad and significant glimpse of prehistoric biodiversity. The Feature 5 mammalian assemblage is an excellent case in point. Table 7.2 provides natural histories for some analogous genera currently inhabiting the western lowlands of Ecuador (compiled from Eisenberg 1989; Emmons and Feer 1991; Janson and Emmons 1990: 317-321). Considered together, this suite of identified genera provides a comprehensive (albeit biased) impression of varied habitats potentially occupied by roughly coterminous mammalian taxa. The thoroughness of this list is further increased with the addition of associated freshwater crabs, frogs or toads, turtles, snakes, tinamous, herons or bitterns, hawks, rails, pigeons, cuckoos, bats, and carnivores. The habitats represented include an impressive assortment of potential niches arrayed along spatial, vertical, behavioral, and trophic dimensions. Environmental inferences based upon archaeofaunal data are further open to future corroboration through comparison with associated botanical materials (Lyman and Livingston 1983: 364; see Pearsall, this volume). It is in this sense that we might borrow from Donald Lathrap's attitude toward all archaeological finds which managed to survive the severe preservation biases of lowland tropical environments. He wisely chose to emphasize "rare and unique items, a procedure not in line with the current emphasis on quantification and statistical significance" (Lathrap 1973b: 176). Acknowledging an apparent uneasiness amongst archaeologists "if they cannot handle actual things and sort them into neat piles to be counted" (Lathrap 1974: 130), he was nevertheless guided by a practical rule of thumb whereby, "one swallow doth a summer make. It is far more probable that a unique item in an archaeological sample represents an established . . . pattern than it records a unique and idiosyncratic event" (Lathrap 1973b: 176). Notes Funding for this research was made possible through grants from the National Science Foundation (BNS-8709649, BNS-8908703, BNS-9108548) awarded to James A. Zeidler and Deborah M. Pearsall. I thank Jim and Debby for their continuing support and encouragement. Identification of the materials was partially undertaken at the American Museum of Natural History, and Museum of Vertebrate Zoology, University of California, Berkeley, and I thank the generosity of respective museum staffs, particularly Dr. Jim Patton for his kindness. Finally, I thank Ann B. Stahl for her comments, criticisms, and editorial prowess, and acknowledge an abiding debt to the stimulating writings of Donald Grayson and R. Lee Lyman. I alone remain responsible for the contents of this paper.
180
PETER W. STAHL
In Figure 7-7(a) Minimum Animal Units (MAU) are computed and scaled (%MAU) according to standardized procedures (see Grayson 1988: 64-65; 1989: 644; Lyman 1991: 128-129; 1992: 9; Lyman et al. 1992: 558) and compared to bulk density (g/cc.) estimates established by Lyman (1984; 1985: 227). The percentage of surviving large mammalian elements (%MAU) and their respective density estimates are significantly correlated (Spearman's Rank Correlation: 0.4069, Two-tailed P value = 0.0419). Figure 7-7(b) compares the observed NISP of White-tailed Deer in the sample, with a hypothetical population of remains (Expected NISP) surviving completely on the basis of density ([MNI {2}] x [{g/cc} for each skeletal portion] x [occurrence of each portion in the skeleton]). The observed and expected populations are significantly correlated (Spearman's Rank Correlation: 0.4481, Two-tailed P value = 0.04). Data for this section are derived from a variety of ethnographic studies situated in lowland areas of northwestern South America and contiguous areas. They include: Alvard and Kaplan (1991); Barrett (1925); Beckerman (1980); Beckerman and Sussenbach (1983); Bennett (1962); Bergman (1980); Bianchi (1988); Carneiro (1970); Ceron (1986); Costales (1956); Dumont (1976); Fejos (1943); Goldman (1963); Gordon (1957, 1982); Hames (1979, 1980); Hames and Vickers (1982); Henley (1982); Jones (1983, 1984); Kaplan (1975); Karsten (1935); Linares (1976); Nietschman (1972, 1973); Numuendaju (1952); Nordenskiold (1938); Oberem (1971); Ross (1978); Ruddle (1974); Setz (1991); Silva (1962); Siskind (1973); Smole (1976); Vickers (1980, 1984, 1989); Von Hagen (1939); Wafer (1934); West (1957); Whitten (1976); and Yost and Kelley (1983).
8
Biological research with archaeologically recovered human remains from Ecuador: methodological issues DOUGLAS H. UBELAKER
Although publications with biological information gleaned from archaeologically recovered human remains in Ecuador appeared sporadically during the early twentieth century (for example, Duckworth 1951; Saville 1913), intensive research on these materials did not begin until the early 1970s. Following Munizaga's important work on human remains from sites of Valdivia and Machalilla phases (Munizaga 1965), I initiated excavation and analysis of a variety of well-dated human skeletal remains, initially from precontact sites mostly on the coast, but later including historic sites in the highlands. This work formally began in 1973 with the excavation and analysis of the Late Integration cemetery site at Ayalan (Ubelaker 1981a) in coastal Ecuador and continued in 1974 at the nearby site of San Lorenzo del Mate (Ubelaker 1983a). Since 1974, excavations by many archaeologists have produced important well-dated samples of human remains that offer an unusual opportunity to examine temporal trends and geographic variation in biological information (Ubelaker 1988a). The extent of the research outlined above is nearly unique in Latin America. Partly to blame for the paucity of such work in other areas is the historical shortage of well-trained physical anthropologists. Although significant samples of human remains have been discovered and excavated over the years, the samples have not been saved or otherwise have lost their provenience and thus much of their scientific value, owing either to a lack of awareness regarding their importance, or a lack of professional interest by physical anthropologists. The other factor in the dearth of such studies in Latin America is preservation. In many regions of the neotropics, postmortem taphonomic factors are such that either the bones do not survive or they survive in poor condition (see Stahl, this volume), prompting decisions against curation. Using our new techniques and methodologies, much can be learned even from such fragmentary material, yet bone preservation continues to be a major limiting factor of such research in the neotropics. In spite of our technological advances in research methodology, sampling remains the major limiting factor in neotropical research in human skeletal 181
Table 8.1. Samples of human remains from Ecuador, grouped into broad cultural periods. Number in Sample
Period
Sample
Date
Location (Province)
Culture
Early Precontact Intermediate Precontact
Santa Elena Cotocollao La Libertad (OGSE-46) Cumbaya OGSE-MA-172 La Tolita Ayalan Non-urn La Florida Agua Blanca Ayalan Urn San Francisco, Zaguan Santo Domingo San Francisco, strata cut, upper level San Francisco, strata cut, lower level San Francisco, Atrium San Francisco, Church San Francisco, superficial collection, lower level San Francisco, Main Cloister San Francisco, superficial collection, upper level San Francisco, Boxes
8250 BP-6600 BP 1000 BC-500 BC 900 BC-200 BC 400 BC-AD 100 100 BC AD 90-AD 190 500 BC-AD 1155 AD 800-AD 1500 AD 730-AD 1730 AD 1500-AD 1570 AD 1500-AD 1650 AD 1540-AD 1650 AD 1580-AD 1700 AD 1600-AD 1725 AD 1535-AD 1858 AD 1670-AD 1709
Guayas Pichincha Guayas Pichincha Guayas Esmeraldas Guayas Pichincha Manabi Guayas Pichincha Pichincha Pichincha Pichincha Pichincha Pichincha Pichincha
Vegas Complex 192 199 Cotocollao Engory 24 20 Cumbaya Guangala 30 18 Tolita Tardio Milagro 51 Chaupicruz 7^ Manteno 7 384 Milagro Historic 3O Historic 46 Historic 74 Historic 46 Historic 19 119 Historic Historic
AD 1730-AD 1858 AD 1770-AD 1890
Pichincha Pichincha
Historic Historic
21
AD 1850-AD 1940
Pichincha
Historic
33
Late Precontact
Early Historic
Late Historic
AD 340
33
Biological research with human remains
183
biology. Limitations of sampling take the form not only of poor preservation, but also the extent of representation of even the well-preserved bones that are recovered. To properly interpret the skeletal samples, one must know: (1) the extent to which an individual is skeletally represented; (2) the extent to which the individuals present represent the once living population that contributed to the cemetery; and (3) the extent to which that population is representative of those from broader geographic areas or time periods (Ubelaker 1974, 1981b). This essay explores these issues in reference to research conducted on such samples in Ecuador since 1973. The data base To date, biological information has been published or formally reported for 1,474 human skeletons (Table 8.1) from twenty different samples ranging in age from 8250 BP to AD 1940. This list does not include the sample from San Lorenzo del Mate (Ubelaker 1983a) or others (for example, Real Alto, Salango) for which biological analysis has not been reported in full. In order to examine long-term trends with meaningfully large sample sizes, the samples in Table 8.1 have been grouped artificially into five broad temporal categories: Early Precontact, Intermediate Precontact, Late Precontact, Early Historic, and Late Historic. Although these categories generally follow a temporal sequence, they are overlapping and do not correspond directly with well-defined cultural changes. The Early Precontact Period extends from 8250 BP to about 2750 BC, and is represented by only one sample, the 192 skeletons from the Vegas Complex, Santa Elena Peninsula in Guayas Province. This site was excavated by Karen Stothert with support from the Banco Central, Guayaquil and represents a largely preceramic, pre-agricultural population which was probably involved in some horticultural activity (Stothert 1985, 1988; Ubelaker 1980a, 1988b). The Intermediate Precontact Period ranges from about 2750 BC to 50 AD and represents a period with shifts toward an agriculturally-based subsistence and increasing village permanence with augmented population density. Samples from five different sites are included in this period. The Cotocollao sample of 199 individuals from Pichincha Province in suburban Quito dates from about 1000 BC to 500 BC and was excavated and published by the Banco Central Quito (Villalba 1988). The La Libertad sample from coastal Ecuador was excavated by Karen Stothert with support from the Banco Central, Guayaquil and includes the remains of twenty-four individuals dating from 900 BC to about 200 BC (Ubelaker 1988c). The Cumbaya sample of twenty individuals was excavated by J. Buys and V. Dominguez from a highland site near Quito and dates from about 400 BC to about 100 AD (Buys and Dominguez 1988a, 1988b, 1990; Ubelaker 1990a). The La Tolita sample (eighteen individuals) was excavated by a Banco Central Quito team headed by F. Valdez. The La Tolita sample includes both primary and
184
DOUGLAS H. UBELAKER
secondary remains from a site that functioned as a ceremonial center and has been heavily looted (Ubelaker i988d; Valdez 1987). The La Tolita site symbolizes an additional important limiting factor with archaeology and human skeletal biology in the New World neotropics: "huaquerismo." Although the problem of unauthorized digging in archaeological sites for marketable artifacts is a worldwide problem, it is especially acute in Latin America where public interest in artifacts is high, and economic problems are especially acute. Valdez (1987) documents generations of families who have, for economic reasons, dug for artifacts at La Tolita. Reacting to destruction at the coastal Ecuadorean site of Salango, Norton (1987: 205-206) notes: The most damaging of man's activities to the archeological heritage of Salango (and indeed of Latin America) is undoubtedly the willful and illegal destruction of the patrimony by treasure hunters or huaqueros. These people are highly skilled in the location of prehistoric sites from which they remove objects of maximum contemporary market value. The desecrated sites, usually tombs, are left as gaping holes, with the surrounding stratigraphy disturbed and relics considered to be of no intrinsic value tossed aside or broken. The huaquero is more often than not only the first in a long line of mercenaries who trade in artifacts solely for financial gain. The objects are sold on the black market to tourists or to connoisseurs, collectors, and even museums. Beginning with the pillaging by the Spanish in the sixteenth century, huaquerismo has resulted in the loss of unaccountably large amounts of priceless information regarding the preColumbian societies of Latin America. In the area of Ecuador covered by the Salango Research Centre, huaquerismo has long been a result of ignorance and poverty. The attraction of large sums of easy money to a people whose material standards of living are low, coupled with a lack of education regarding their national heritage, has resulted in a certain laissez faire attitude toward the numerous sites on their land. Human remains are rarely sold, but the bones are usually discarded or otherwise destroyed in the unlawful scramble for artifacts. The loss of information to science is staggering. The Late Precontact Period ranges from AD 50 to AD 1615 and includes 518 skeletons from four samples. Two of the samples in this period originate from the Ayalan site in Guayas Province, coastal Ecuador (Ubelaker 1981a). The "nonurn" component consists of primary and secondary burials thought to date from about 500 BC to AD 1155. The "urn" component includes 384 skeletons all found within large ceramic urns dating between AD 730 and AD 1730. The seventy-six skeletons from La Florida were excavated by L. Doyon with support from the Banco Central Quito. They date from about AD 340 and originate in deep, highstatus shaft tombs in a site located within suburban Quito (Doyon 1988). The Historic Period samples are all drawn from excavations at two Historic churches in Quito, San Francisco and Santo Domingo. The San Francisco excavation was sponsored by the Instituto Nacional de Patrimonio Cultural del Ecuador (INPC) an4 the Agencia Espanola de Cooperacion Iberoamericana
Biological research with human remains
185
(AICI) (Teran de Rodriguez 1988). The Santo Domingo sample originates from an excavation in the church sponsored by the government of Belgium and the Instituto Nacional de Patrimonio Cultural of Ecuador. The samples from these two churches have been differentiated according to radiocarbon dates and those estimated from architectural features and artifact analysis. For details of these samples see Teran de Rodriguez (1988) and Ubelaker (1993a, 1993b) and Ubelaker and Ripley (1993). The "Early" Historic sample dates from AD 1615 to about AD 1675 a n d includes 2.15 skeletons. Four of these are from various components of San Francisco church and one is the Santo Domingo sample. The "Late" Historic sample dates from AD 1675 t o about AD 1940. All 258 skeletons are from five components of San Francisco church. Materials in this group likely temporally overlap considerably with the "Early" sample. Geographic representation As Figure 8.1 shows, most of the Precontact samples discussed above originate from coastal sites, especially the southern coast. The others are from the highlands, in the area of Quito. Two factors have largely influenced this distribution: ground cover and bone preservation. Although a small country, Ecuador contains remarkable geographic diversity. Bisected by the north—south running Andes mountain chain, Ecuador presents three major environmental zones: a mountainous high altitude zone, a dry southern coastal area, and tropical forest areas on the interior and northern coast. The collision of two ocean currents (the Nino and Humboldt) near the midpoint of Ecuador's coastline creates a sharp division in rainfall. High rainfall and tropical forest environment dominate north of the line, while low rainfall and arid environment predominate in the south. While this pattern of rainfall and vegetation fluctuates somewhat (Norton 1987), it could be argued that we know more about the southern coast simply because it is easier to work there. The present sparse ground cover facilitates site location and excavation. For these reasons, much of the archaeological work in Ecuador has focused on the dry coastal areas as well as the highlands. The tropical forested areas probably also contain rich resources, but have received minimal archaeological attention (see Pearsall, Stahl, Zeidler, this volume). These factors present sampling problems for archaeological interpretation. Our inferences about past biological patterns are biased in favor of those from the coast and limited areas of the highlands simply because we only have skeletal samples from those areas. These limitations must be kept in mind in making broad inferences about Ecuador's past. Preservation affects the sample in much the same way. Archaeological investigations in the wet tropical forest areas are not likely to produce skeletal samples for analysis (see Stahl, this volume). The meager work that has been done so far indicates that human bone preservation is very poor. Again, the few bones that survive offer a very limited view of the people they represent.
186
DOUGLAS H. UBELAKER
ECUADOR Cape San Francisco
Bahia de Caraquez r&loneR-
AYALA Gulf of Guayaquil
|Vmpo-6ooo
J
200 k.
Figure 8.i. Geographical distribution of Ecuadorean sites yielding skeletal samples.
Biological research with human remains
Individual
187
representation
As shown in Table 8.1, the published samples utilized in this report represent at least 1,474 individuals. This number reflects the minimum number of individuals reported for each of the samples included and it, too, is affected by problems of sampling. Inventory of individual primary skeletons is relatively straightforward. However, assessment of the number of individuals present in secondary deposits is more complicated and can affect demographic calculations. The Ayalan analysis (Ubelaker 1981a) offers an example of the problems encountered in assessing individual representation within secondary deposits. This site included primary skeletons as well as secondary deposits found within large ceramic vessels. To determine the minimum number of individuals within each urn, a detailed inventory of all bones within each urn was made. For example, the inventory of adult (mature) remains from Feature 7 reveals eleven individuals represented by the right humerus and left femur, only six by the left humerus, three by the right femur, and only one by many of the small bones of the hand. This suggests that at least eleven individuals are present, but raises questions regarding what happened to the other bones. Similarly, immature bones from Feature 7 showfiveindividuals represented by the right femur, four by the left femur, and only one by the right calcaneus. Clearly, immature remains also show variable representation, reflecting the complex multistage mortuary procedures employed. To compute the total number of individuals present within the entire urn burial sample, the analysis assumed that each urn burial was a distinct feature and that bones of different individuals did not overlap among different urns. With this logic, the maximum number of individuals within each of the urns may be summed to produce the total number of individuals in the entire sample. This procedure indicates at least 194 adults (over the age of twenty) and 190 subadults (under age twenty), for a total of 384. Fewer individuals are suggested from the totals of the inventories. In contrast to the above, if we assumed that individuals could be deposited in more than one urn, then the total sample size would be less. The total inventory of all adult bones from all features shows that the most commonly present bone is the right humerus (147) while only 19 right lesser multangulars are present. For immature individuals, 140 are represented by right femora and only 7 by the right talus. Thus, the assumption that the bones of each individual could be deposited in different urns produces a total sample (287) that is 97 individuals (25 percent) smaller than the figure assuming that the contents of each urn are distinct. Although the total size of the sample varies with assumptions made about the mortuary customs, demographic reconstructions are only minimally affected. For example, the ratio of adults to subadults in the larger sample is 1.02. The same ratio in the smaller sample, assuming that individuals are not confined to one urn, is a very close 1.05.
Table 8.2. Temporal trends in biological attributes assessed from skeletal samples. Biological Variable Female stature (cm) Male stature (cm) Life expectancy at birth Life expectancy at age 5 Life expectancy at age 15 Ratio of immature individuals to adults Maximum longevity Ratio of bones with periosteal lesions to individuals Ratio of bones with periosteal lesions to adults Ratio of bones with trauma to adults Percentage of permanent teeth with carious lesions Percentage of permanent teeth with alveolar abscesses Percentage of permanent teeth lost antemortem Percentage of permanent teeth with hypoplasia
Early Precontact
Intermediate Precontact
Late Precontact
Early Historic
Late Historic
149
151 161
I
151 155
161
161 15 28 22
O.57 6O
26
2-7 2-3 0.32
54
5I
161 2-3 3i 2-5 0.78
68
154 20
15 2-3 28 1.29
2-3 23
55
61
O.O6
O.II
0.06
O.O9
0.14
0.12
0.13 0.28
O.O9
0.06
0.12
0.12
3.O
1-7 i-9
8.7 3-1
5.6
0.66 0.22 0.37 0.29 16.6
4.6
9.8
i-4
6.O
11.8
20.0
I.O
2.0
4.4
2-5-4
3.0
0.5
I.O
Biological research with human remains
189
175-r 150125C
100-
D)
Female Male Figure 8.2. Temporal change in adult stature.
Patterns of change
Grouping the skeletal data into thefivevery general temporal categories shown in Table 8.1 offers an opportunity to assess long-term trends in health and disease experience. Recognizing the limitations discussed earlier, the samples provide data not only regarding a few of the specific diseases that were present but, more importantly, the frequency of these diseases. If correctly assessed, such patterning of biological attributes offers evidence of how the broad cultural changes in the Ecuadorean past have affected the biology of the people themselves. Table 8.2 lists many of the biological variables and how frequencies changed through time. All of these variables provide useful information about morbidity and mortality, but each has its own limitations and perspective. Stature represents estimates of living height and must be calculated from long bone lengths (Ubelaker 1989a). The data in Table 8.2 show comparatively little change through time. In fact, the values reported are similar to those documented for contemporary Indian populations in Ecuador (Gillin 1941). This trend is seen graphically in Figure 8.2 for temporal increase in stature. Values show a slight decrease for males during the Early Historic period, probably reflecting elevated disturbances in growth due to poor nutrition and disease during this time. The values for both males and females (Table 8.2) increase slightly during the Late Historic. This increase probably reflects relief from the growth problems of previous periods. Sampling could also be a major factor here since the Late Historic samples contain mostly European remains while all of the Precontact and some of the Early Contact remains are of non-European Indian remains. Additional methodological factors affect interpretation of these data. Bone preservation again represents a limiting factor in stature reconstruction. In the Ayalan sample of 194 adults, preservation factors limited stature estimates to
190
DOUGLAS H. UBELAKER
At Birth
At Age 5
t
1 At Age 15
35 -f
E. Precontact
Late Precontact
Early Historic
Late Historic
Figure 8.3. Temporal change in life expectancy at birth, age 5, and age 15. only 72 adults (38 females and 34 males), or only 37 percent of the total sample. If fragmentation and these limiting preservation factors operated at random then the stature values are adequate and representative. However, if mortuary procedures or taphonomic factors selected for either tall or short individuals then the estimates would be misleading. The type of bone preserved can also affect conclusions about stature. Probably the most accurate formulae for Ecuadorean Indian populations are those of Genoves (1967: 76) developed from studies of Mexican cadavers classified as "indigenous." Genoves published formulae for only the femur and tibia for both males and females. In the studies cited above, I relied upon the formulae of Genoves whenever femora and tibiae were available, but utilized the formulae of Trotter (1970) when only the fibula or other bones were present. The Trotter formulae are available for the humerus, radius, ulna, femur, tibia, andfibulaand were developed from United States samples. Formulae are provided for Mongoloid and Mexican males. For female remains, the Trotter equations for White females must be used if the bones necessary for the Genoves equations are not present. Obviously, using formulae developed from United States samples are probably less accurate for Ecuadorean Indians than those developed from Mexican cadavers. In such situations, it appeared preferable to augment the sample using formulae with more potential error than to rely upon the comparatively smaller sample of individuals represented by intact femora and tibiae. This problem is compounded in samples such as Cotocollao and OGSE-80, Vegas Complex with extensive fragmentation. It is possible that such factors have obscured some past variability in Ecuadorean stature.
Biological research with human remains
Demographic
191
change
The demographic temporal trends suggested by Table 8.2 generally reveal a pattern of increasing mortality through time. Life expectancy at birth (Figure 8.3) was highest during the Early and Intermediate Precontact Periods and subsequently declined to a low during the early historic, followed by a moderate recovery during the Late Historic Periods. These values are calculated from the life table and are heavily influenced by the infant mortality rate. The increasing mortality through time during the Precontact Periods probably represents increasing morbidity due to increased population density and the accompanying increase in infectious disease. Changes from the varied diet of Las Vegas times to the more restricted agricultural period diets may also be factors. The low value in the Early Historic Period probably reflects the mortality inflicted by the new diseases introduced from Europe and the nutritional problems associated with urban life during that period. Life expectancy at age five actually increased during Late Precontact, only to fall dramatically during the Early Historic, and remain low during the Late Historic Periods. This statistic excludes infant mortality, but includes child mortality. Comparing this statistic with the life expectancy at birth values suggests that the decrease in life expectancy at birth figures are largely due to an increase in infant mortality. Additional interpretation is possible from examining values for life expectancy at age fifteen. This value gradually increases from Early Precontact through Early Historic, and then decreases during the Late Historic Periods. The data suggest that much of the elevated mortality during the Late Precontact Period was aimed at the individuals less than fifteen years of age. This interpretation is reinforced with data presented on the ratio of immature individuals to adults in the samples examined. This value is highest during the Early Historic, followed by the Late Precontact Period. The Late Historic value is still higher than the values for the Early and Intermediate Precontact Period. Maximum longevity records the age of the longest living individual in the sample. For Table 8.2, this value was computed by recording the age of the oldest individual for each of the individual samples listed in Table 8.1. The mean of these values was then computed for all of the samples within each of the broader groups utilized in the analysis. This procedure suggests that longevity was greatest during the Late Precontact and least during Intermediate Precontact and Early Historic Periods. Of all of the variables in Table 8.2, the demographic factors are most susceptible to sampling problems. To begin with, the data are affected by changes in growth and fertility in the populations represented (Horowitz et al. 1988). We can reasonably suggest that positive population growth occurred during the Early Precontact and Intermediate Periods. Native population size declined during the
192-
DOUGLAS H. UBELAKER
Early Historic, and probably also in the Late Precontact Periods. This would suggest that life expectancy values would be underestimated for the Early and Intermediate Precontact, and overestimated for the Late Precontact and Early Historic Periods. Such adjustment would mostly augment the differences discussed above. Bone preservation and representation can also be limiting factors in demographic analysis. All of the sampling problems discussed earlier can dramatically affect demographic reconstruction if not properly accounted for. Mortuary procedures or taphonomic factors that result in any non-random reduction of the skeletal sample (see Stahl, this volume) will introduce error into the analysis. Loss of infants is an especially troublesome problem since they are at times targeted for special separate mortuary procedures and are more subject to postmortem destruction than are the more sturdy bones of older individuals. Population representation is more of a problem with the historic samples. All of these originate from historic churches in Quito. Only the Zaguan sample from San Francisco church is known to be entirely of Indian origin. The other historic samples probably represent mostly Ecuadoreans of European descent, but some Indians and even some individuals of African descent are also present. The church samples do not represent all Ecuadoreans or even all of those living in Quito. Rather, they represent those individuals who for whatever cultural reasons were selected for burial in the church. Physical characteristics of all of the samples except the superficial collection from the lower chamber of San Francisco church suggest they represent the general population affiliated with the church (all ages, both sexes). The exception (superficial collection, lower chamber) consists mostly of adult males of European origin that presumably represent officials of the church whose remains were deposited in the chamber after about AD 1770 (Ubelaker and Ripley 1993). Demographic reconstruction also depends upon accurate estimation of age at death which also can be negatively impacted by taphonomic factors. The most accurate methods of estimating age at death rely upon the formation of teeth in immature individuals (Ubelaker 1987a, 1989a, 1989b) and on microscopic age changes in the cortical long bones of adults (Ubelaker 1978; Kerley and Ubelaker 1978). In practice, the most reliable age estimates are those generated from all age indicators on the skeleton. Obviously, reduction in the preservation and completeness of the skeleton will reduce the accuracy of the age estimates and subsequently the demographic reconstruction. In large secondary deposits, like the historic church ossuaries and the Ayalan urns, individual bones are present in varying frequencies and in varying degrees of preservation. Ages may have to be estimated for the individual bones and then compared within the total sample to produce a list of expected ages. The accuracy of this procedure can be greatly reduced if key morphological markers are not present or are unclear due to postmortem change.
Biological research with human remains
193
Evidence of infection
Although the mortality figures discussed above may largely represent the impact of infectious disease, some relevant observations can be made from the bones themselves. In particular, periosteal lesions probably provide such direct evidence of infectious disease. These lesions represent the result of abnormal stimulation of the periosteum. In the usual scenario, infection will stimulate the periosteum to produce new bone on the outside surface of the normal bone cortex. If the individual survives, this new bone formation will be remodeled and will gradually change its appearance. Eventually, this remodeled bone may be difficult to distinguish from the normal bone surface. Although usually produced by infection, these lesions may also be caused by other factors, such as trauma, that can also stimulate the periosteum (Ortner and Putschar 1981). In the Ecuadorean data summarized in Table 8.2, periosteal lesions are expressed as the number of bones with lesions to both individuals and to adults in the sample. Both of these values show increases from Early Precontact to Intermediate Precontact, reduction in Late Precontact, and then dramatic increases during the Historic Periods. Presumably, these temporal trends represent the general increase in infectious disease that in Precontact times is associated with increased population size and density, and in Historic times with the introduction of infectious disease and the urban population density of Quito (Alchon 1991). Although a temporal increase in morbidity best explains the data, sampling may also be a factor. The lesions are more likely to survive to be noticed in the relatively well preserved skeletons of Ayalan then in the fragmented samples of OGSE-80 or Cotocollao. It is also important to remember that such lesions require time to form. In that sense, they probably record exposure to infectious disease rather than mortality. Clearly, an adult skeleton showing multiple evidence of bony response to infectious disease was better off than a child who did not survive the exposure to infectious disease. If the child died immediately from the disease, he or she would show no lesions on the skeleton. One could even argue that an adult showing extensive evidence of well-remodeled periosteal lesions is healthier than an adult counterpart who lacks such evidence. Presumably, surviving the disease experience would offer the adult with the lesions some immunological resistance to additional exposure. Trauma Evidence for trauma in Ecuadorean past populations is mostly confined to adults. Frequencies are relatively low during Early and Intermediate Precontact, increase to higher levels during the Late Precontact and Early Historic, and then dramatically increase during the Late Historic Periods. Presumably, the elevated
194
DOUGLAS H. UBELAKER
levels of trauma during the Historic Periods reflect the violence and accidents of urban life. Once again, the frequencies of trauma reported are affected by problems of sampling and preservations. In Precontact times, it seems reasonable that some severely traumatized individuals might be excluded from the sample either through exclusion in mortuary ritual or because their bodies were never recovered. Poor preservation and fragmentation could also affect interpretation. It is difficult to detect perimortem fractures in extremely fragmented remains. Antemortem fractures with callus formation and considerable remodeling may go undetected if the remains are extremely fragmented. Dental disease As shown in Table 8.2, most measures of dental disease in the populations examined show temporal increases. The dental problems discussed here include carious lesions, alveolar abscesses, dental hypoplasia, and antemortem dental loss. Carious lesions represent the collapse of dental tissue following necrosis and decay. Although some deciduous teeth do show carious lesions, especially during the Historic Periods, most are found in the permanent dentition. Carious lesions are produced by a process that begins with plaque formation resulting from the buildup of food residue. Sticky, starchy foods are excellent producers of plaque. Colonization of bacteria within the plaque contributes to the release of chemicals that break down the surface of the tooth. Although the frequency of dental caries can be related to the structure of the teeth and to habits of dental hygiene, in past populations it primarily reflects dietary factors (Larsen et al. 1991). In particular, the development of agriculture with its intensive use of starchy foods has been linked to increased caries frequency in many parts of the world (Cohen and Armelagos 1984; Turner 1978). In the Ecuadorean data, frequencies of carious lesions remain relatively low until Late Precontact, and then increase substantially again during the Late Historic Periods. The jump during the Late Precontact appears to represent the increasing reliance upon agricultural products during that time. The Late Historic increase probably corresponds with the availability of refined sugar. Sugar was also present in Quito during the Early Historic Period, but was not economically available to the general population. Since teeth are very durable and are among the skeletal elements most likely to survive, taphonomic factors affecting the data are minimal. The general sampling problems discussed in reference to other variables would apply, however. It is interesting that the configuration of the temporal groupings of the samples also seems to affect the pattern. An earlier study (Ubelaker 1984) that temporally arranged the data from five individual sites that had been reported on at that time, suggested a very direct temporal increase in carious lesions through time. With
Biological research with human remains
195
the inclusion of more data from additional sites reported here, the pattern is still one of temporal increase, but not as direct as indicated previously. The correlation with agriculture is also not as direct as previously believed. The highland site of Cotocollao (Ubelaker 1980b, 1988c) which dates well into the agricultural period shows a caries frequency similar to that of the pre-agricultural site of OGSE-80. Alveolar abscesses represent lesions in the bony alveolus near the roots of adjacent teeth. Such lesions are produced when the pulp cavity of a tooth becomes infected. This results when an abnormal opening is formed in the surface of the tooth, allowing the bacteria to penetrate. Such an opening can be produced by carious lesions, trauma, or by excessive dental attrition due to mastication. In the Ecuadorean data, the temporal trend in the frequency of abscess formation is very similar to that reported for dental caries, probably reflecting the strong role of dental caries in abscess formation. The data suggest that Early Precontact mouths had relatively low levels of alveolar abscesses. The frequency of abscesses increased to the Late Precontact, decreased slightly during the Early Historic, and then increased again to an all time high during the Late Historic Periods. The frequency of permanent teeth lost antemortem displays the clearest temporal trend of any of the reported variables. The values increase steadily from 6 percent in Early Precontact to 25.4 percent in Late Historic. This variable reflects the increases in dental caries and alveolar abscesses as well as the increasing adult life expectancy in the Late Precontact and Historic Periods. As with other variables, data on the loss of teeth will be affected by abnormal sampling of the population. Since the loss of teeth represents the cumulative effects of dental disease and other factors, age is an important factor. Any reduction in older adults within the samples will negatively affect the frequencies of tooth loss. Dental hypoplasia represents an area of faulty formation of the tooth enamel (Kreshover i960). Many workers believe that hypoplasia results from disturbance in the growth process during dental formation. Such disturbances frequently may be traced to nutritional problems, illness or other morbid conditions. Frequencies of hypoplasia have been used extensively as stress indicators (Goodman et al. 1984,1988). The locations of the defects on the tooth crown also provide information on the age of the individual when the defect formed, and thus when the individual was stressed. Temporal trends of enamel hypoplasia within the Ecuadorean data present an interesting pattern. The lowest frequencies clearly are from the Early Precontact Period. The value increases steadily until Late Precontact, and then drops dramatically during the Early Historic and Late Historic Periods. Since hypoplasia is widely regarded as a sensitive indicator of stress, the decrease during the Historic Periods is problematic since other variables clearly show that the Historic Periods were times of high morbidity, high mortality, and, presumably, high stress. One likely explanation is that the enamel defects that characterize
1$6
DOUGLAS H. UBELAKER
hypoplasia also serve as ideal locations for plaque formation and thus also for dental caries (Bibby 1943). The frequency of hypoplasia may be reduced because the defects are destroyed by dental caries which greatly increases in frequency during the Historic Period. Porotic hyperostosis Porotic hyperostosis represents a condition usually observed on the cranial vault and on the superior interior surfaces of the orbits. The condition involves an expansion of the spongy diploe with increased surface porosity and thinning of the outer cortex (Angel 1966). In extreme newly forming cases, the condition shows marked abnormal porosity in a pattern resembling a "honeycomb" (Stuart-Macadam 1985,1987a, 1987b). With time and remodeling, the porosity is less apparent but the bone appears abnormally thickened. In the cranial vault, a cross-section of such lesions frequently reveals a hair-on-end effect, that clinically is known to describe adult thalassemia and other forms of extreme anemia. This condition has been observed throughout the Americas. Since the hereditary anemias are confined to the Old World, the occurrence of porotic hyperostosis has been linked to acquired anemia. Many workers have suggested that New World porotic hyperostosis may be caused by iron deficiency anemia, triggered by maize consumption (El-Najjar et al. 1976; Lallo et al. 1977). Maize is low in iron and also contains the enzymes phytates that can inhibit the absorption of iron. Others have noted that although diet may be a key factor in the formation of porotic hyperostosis, parasitism, other diseases, and even methods of food preparation may be important as well (Walker 1985, 1986). In the Ecuadorean samples discussed above, porotic hyperostosis is confined to several individuals from the Guangala site of OGSE-MA-172 (Ubelaker 1983b), both components of the Ayalan site, La Tolita, and an isolated Jama-Coaque cranium from Manabi (Ubelaker 1987b). The distribution of the sites with porotic hyperostosis does not match well with the pattern expected if diet is the principal causal factor. Notably, no lesions were found among the skeletons from the highland site of Cotocollao, a population which surely had a high dietary input of maize. In contrast, the lesions were found in high frequency among the Guangala sample of OGSE-MA-172, representing a population which utilized agriculture but which also relied heavily upon fish. The site is located within sight of the Pacific Ocean and abundant evidence of fish remains were recovered during the excavation. This pattern suggests that although dietary factors may have been present, a key factor in the formation of porotic hyperostosis was probably parasitism, especially hookworm. The current geographical distribution of hookworm (Necator americanus and Ancylostoma duodenale) matches the geographical distribution of porotic hyperostosis in the Ecuadorean samples. Namely, these hookworms are health problems on the Ecuadorean coast but not
Biological research with human remains
197
in the highlands because they require a warm climate and moisture to complete their life cycle (Ubelaker i988f, 1990b, 1991, 1992). Again, the limitations of sampling preclude a complete study of the distribution of this phenomenon. It would be interesting to learn if the pattern of porotic hyperostosis is also present in the inland tropical areas as predicted from this model. Unfortunately, the samples are not available. Summary In spite of the environmental limitations of the neotropics, considerable research on human skeletal biology from archaeologically recovered samples has been advanced. In general, a broad range of different indicators suggests a pattern of increasing morbidity through time. Collectively, the data follow the pattern suggested by Kent (1986). Throughout the Precontact Period, disease conditions were stimulated by increasing sedentism and population density. Poor sanitation led to the increase in infectious disease and general morbidity in the population. During the Historic Period, this trend continued and was augmented by the new diseases introduced from Europe and the general cultural disruption that followed European contact. This situation probably resulted in a negative impact for both European and Indian communities. Although the patterns of change described above are firmly rooted in the biological evidence available at this time, interpretation remains tentative because of the many sampling factors discussed throughout this essay. The general pattern revealed by skeletal analysis has changed with the availability of new samples and additional research. The pattern apparent in 1984 (Ubelaker 1984) based on the analysis of five published samples, is substantially different from the more comprehensive analysis reported here. Our understanding will continue to evolve with additional research and the availability of new samples. The neotropical environment of Ecuador imposes significant limitations on research in skeletal biology. Data are available for only certain areas that are conducive to bone preservation and archaeological excavation. These data must be interpreted cautiously because of the many possible errors introduced by problems of sampling, complexity of mortuary procedures, and age determination. These limitations do not preclude valuable inference about the nature of human adaptation in this varied neotropical area, an adaptation that led to increasing population size and density but with the accoutrements of increasing morbidity. Only additional research on more skeletal samples from additional areas and time periods will allow the complexity of this adaptation to be more fully known.
9
Interpreting dietary maize from bone stable isotopes in the American tropics: the state of the art LYNETTE NORR
Archaeologists approach the reconstruction of prehistoric subsistence and ecology using macrobotanical, palynological, phytolith, and faunal evidence, along with functional studies of utilitarian tools and iconographic analyses of ceremonial artifacts. When a dependable subsistence staple such as storable maize or processed manioc flour is identified, this has further implications for population dynamics and social complexity. Inquiries about prehistoric settlement and subsistence patterns in the tropics often focus on the inclusion of maize as a dietary component, and when, if ever, it became a dietary staple (that is, a sustaining or principal food source). A series of interrelated questions about prehistoric maize are pertinent: When and where were its origins? Was it only a minor dietary component, or was it relied upon as a subsistence staple? Is its consumption linked to patterns of poor health or disease? Was it differentially consumed by certain social classes or age/sex categories? Are particular patterns of settlement associated with agricultural subsistence? In complex environments like the American tropics, where species diversity is high and subsistence alternatives can be many, multiple lines of evidence are necessary to answer questions of subsistence, settlement, and agricultural origins. The stable isotopes of carbon and nitrogen in archaeological human remains can be one line of evidence used to provide additional information about these topics. Natural variations in the stable isotope ratios of certain categories of food resources allow for several different inquiries into paleodiet, one of which is the importance of dietary maize. To date, bone collagen has been the most commonly analyzed tissue fraction in applications of stable isotope analysis to questions of paleodiet. Where human diets are more diverse, however, interpretations of subsistence data become more challenging. In such cases, additional analyses may be needed to provide insight into patterns of subsistence. The carbon isotope ratios of carbonate in bone and tooth enamel apatite will also provide dietary information that complements data from bone collagen. Stable isotope analysis of human bone must be carried out carefully and dietary interpretations must be
198
Interpreting dietary maize from bone stable isotopes
199
well founded. To accomplish this, several important aspects of the distributions of isotopes in nature, sample preservation, analytical procedures, and interpretations need to be addressed. When using the stable isotopes of human skeletal remains from tropical sites for dietary reconstruction, both the isotopic results and the dietary interpretation can be dramatically affected by the research design and the methodology used. One must carefully choose the dietary questions to be investigated, based on knowledge of the isotopic composition of the food resources in the area of study. In many tropical environments, poor preservation of archaeological bone is a serious problem (see Stahl, Ubelaker, this volume). Often there is no bone preserved, and when bone is found the collagen may not be preserved or the carbonate may be contaminated. The appropriate tissues and their fractions, such as bone collagen, bone or tooth apatite carbonate, or hair, must be selected and prepared in such a manner that the fractions analyzed represent the isotopic composition of the foods consumed rather than that of the burial matrix. Interpretation of the results depends on our understanding of how the isotopic composition of the diet is recorded in human bone and tooth tissues. These aspects of dietary reconstruction with stable isotopes are addressed. Following this discussion, a case study using the stable isotopes of carbon and nitrogen in human bone collagen and apatite from prehistoric Panama is presented to illustrate some of the possibilities and limitations of this analytical approach to dietary reconstruction in the American tropics. Stable isotopes and paleodiet The analysis of archaeological bone for ratios of the stable isotopes of carbon and nitrogen is now a well-established technique for reconstructing paleodiet and deciphering prehistoric subsistence patterns. This is possible because specific food resources have distinct ratios of the stable isotopes of carbon (13C/12C) and nitrogen (15N/14N). Dietary isotopic composition is incorporated into body tissue fractions such as bone collagen and bone and tooth enamel apatite, and may be preserved for thousands of years. Thus, the stable isotope ratios of carbon and nitrogen in body tissues can be useful in the reconstruction of prehistoric diet (DeNiro 1987; van der Merwe 1982). The stable isotope ratios of carbon or nitrogen are expressed using the delta (S) notation as parts per thousand (per mil, %o) difference from a standard, a method that will give more accurate results than will absolute determinations. For carbon, the standard is the Pee Dee Belemnite (PDB) limestone fossil (Craig 1957), and S13C values are usually negative because the PDB limestone has more 13C relative to 12C than most other substances. For nitrogen, the standard is atmospheric nitrogen, AIR (Mariotti 1983), and S15N values are usually positive. The delta values of a sample are calculated relative to the appropriate standard, using the following equations:
2OO
LYNETTE N O R R
r( 13 c/"C) sample -("C/"C) PDB -|
S C(PDB) = ^
(13C/12QpDB
j x 1000/00
and r(15N/"N)sampic-("N/^N)AIRi 3 *N(AIR) = ^ j x 1000/00 rN/1
Interpreting dietary maize from bone stable isotopes
201
Marine and Estuary Fish
+15
+10 Sea ! Turtle Water Fowl
C3
To
+5
Terrestrial Vertebrates
c4 Maize
0 -30
-20
5
13
c%<
-15
-10
Figure 9.1. Isotopic composition of archaeological food resources in lower Central America (after Norr 1990) (1.5% was added to the 613C value of modern foods to compensate for 12C enrichment of the atmosphere from the burning of fossil fuels, as per Tieszen 1991). Tieszen (1991: Table I) and Ambrose (1993: Table IV). In hot dry regions, such as some parts of East and South Africa, drought tolerant herbivores will conserve water by recycling their urea. Excreted urea is depleted in 15N, thereby significantly concentrating 15N in body tissues (Ambrose and DeNiro 1986a). This physiological factor in nitrogen metabolism has not been demonstrated for temperate or neotropical herbivores, as those thus far analyzed have S15N values typical of terrestrial fauna (Norr 1990). In closed-canopy forests, such as the Amazon rainforest (van der Merwe 1989; van der Merwe and Medina 1989,1991) and the Costa Rican cloud forest (Sternberg et al. 1989), high rates of C3 plant decomposition, the recycling of 13C-depleted CO 2, and high CO2 concentrations can significantly lower S13C values in the understory plants and the animals feeding on them (also, Ambrose and DeNiro 1986a). These responses of plants
2O2
LYNETTE NORR
and animals to environmental conditions result in regional variations in the isotopic compositions of food resources. The specific variability in the isotopic composition of the local food resources ultimately affects the utility of stable isotope analysis for dietary reconstruction in a particular part of the world. Situations in which stable carbon and nitrogen isotope ratios in human remains are recommended for paleodietary reconstructions in the New World tropics are outlined below. Isotopic variation in food resources Maize versus nonmaize terrestrial diets. Where all local resources are C 3, as in the temperate woodlands of eastern North America, the consumption of maize, a tropical C4 plant, can be identified by high stable carbon isotope ratios in human bone. The wild terrestrial faunas of eastern North America are in the C3 food chain, and therefore do not contribute high 13C/12C ratios to prehistoric human diets (Katzenberg 1988, 1989; Land et al. 1980). The first dietary study using stable carbon isotopes in human bone collagen compared amounts of maize in prehistoric diets from Woodland and Mississippian sites in eastern North America (van der Merwe and Vogel 1978; Vogel and van der Merwe 1977). The identification of a C4 plant with a high 13C/12C ratio in a predominantly C3 diet, is a fairly uncomplicated application of stable isotope analysis of archaeological human bone to studies of prehistoric diet. Several similar studies followed in North America (for example, Broida 1984; Buikstra et al. 1988; Buikstra and Milner 1991; Lynott et al. 1986; Matson and Chisholm 1991; Schwarcz et al. 1985) and Amazonia (van der Merwe et al. 1981). High altitude regions of the Neotropics provide a similar environment for paleodietary reconstruction based on stable carbon isotope ratios in human skeletal tissues. Grasses that grow at cooler and moister high altitudes are C3, just as in temperate zones (DeNiro and Hastorf 1985; Tieszen et al. 1979). Therefore, maize can be identified as a dietary component from its stable carbon isotopic composition because it would again be the single predominant C4 plant in an otherwise C3 diet. Burger and van der Merwe (1990) analyzed human bone collagen for stable carbon isotopes from the Andean site of Chavin de Huantar and concluded that maize was not the important dietary staple previously thought (Lanning 1964). The consumption of maize at Chavin de Huantar was probably of greater significance as a ritual food than as a dietary staple (see Lathrap 1973a). Marine versus terrestrial nonmaize and maize diets. Stable carbon isotope ratios of human bone collagen can distinguish a marine diet from a terrestrial diet in temperate and arctic regions of the world (Chisholm et al. 1982; Schoeninger et al. 1983; Tauber 1981), as well as in temperate and tropical regions where maize
Interpreting dietary maize from bone stable isotopes
203
or other C4 plants (or animals that grazed on C4 grasses) were not eaten (Sealy 1986; Walker and DeNiro 1986; Yesner 1988). Marine vertebrates generally have S13C values in the range of —19 to — 9%o (Schoeninger and DeNiro 1984), compared to the average of — z6%o for C3 plants and — 2i%o for terrestrial fauna in a C3 environment. These differences in stable carbon isotope ratios are sufficient to allow prehistoric marine or nonmaize terrestrial diets to be identified. Because marine organisms generally fall into a S13C range that is intermediate to C3 and C4 plants, some maize and marine diets will be similar in their carbon isotope ratios. In a situation where maize and marine organisms are both potential food resources, another line of evidence must be used to determine whether an increased 13C/12C ratio in human bone collagen is a result of a marine diet, a maize diet, or a mixture of the two. Nitrogen isotope ratios in human bone collagen also will distinguish between most marine and terrestrial diets (Schoeninger et al. 1983; Walker and DeNiro 1986). S15N values of flesh from terrestrial herbivores in the New World typically average + 4%o and carnivores average + 6%o, while marine fish average + i4%o and marine mammals + i8%o. Therefore, nitrogen isotope ratios used in conjunction with stable carbon isotope ratios of bone collagen will distinguish between marine and terrestrial nonmaize and maize diets (Schoeninger et al. 1983) and possible mixtures of those foods (Larsen et al. 1992; Norr 1990). When the 813C values of humans are within the range of a marine diet but the S15N values indicate that the diet was a terrestrial one, then the observed S13C values most likely represent a diet that included a C4 plant, such as maize. In contrast to Old World tropical environments, where the consumption of meat from certain terrestrial herbivores will elevate both S13C and S15N values (Ambrose and DeNiro 1986a, 1986b; Sealy et al. 1987), neotropical diets that can produce elevated S13C and S15N values include marine vertebrates in conjunction with maize (Norr 1990, 1991). Neotropical reef fish, however, may have low §15N values (DeNiro and Epstein 1981; Keegan and DeNiro 1988; Schoeninger and DeNiro 1984). The marine organisms that live in a coral reef environment will have lowered S15N values due to nitrogen fixation in the reef environment (Capone and Carpenter 1982). Reef fish may have S15N values similar to terrestrial animals, and thus terrestrial and marine diets would be indistinguishable using nitrogen isotope ratios. But, if the terrestrial diet is only C3, then only carbon isotope ratios are necessary to distinguish between terrestrial and marine diets in reef environments. Legumes in a terrestrial diet. Legumes have S15N values near o%o, similar to that of air from which symbiotic bacteria fix N 2 into nitrate compounds that are usable by the plant (Shearer and Kohl 1986). Other terrestrial plants have S15N values that average + 4%o because their source of nitrate compounds derives from the nitrification process by bacteria in the soil. While it may be difficult to identify legumes as a dietary component from the S15N value of bone collagen, the
ZO4
LYNETTE N O R R
presence of a large quantity of legumes in the diet would lower the average dietary S15N, thereby leading to the underestimation of foods higher in 15N, like marine vertebrates or meat from carnivores. Because of differences in the amount of protein in plants and animals, the effect of dietary legumes on the S15N of consumer tissues is likely to be small when the diet includes animal protein. Forest canopy effect on carbon isotopes. Plants and animals living in the understory of the canopy and on the floor of a closed canopy forest will be depleted in 13C; that is, their §13C values will be more negative than would otherwise be expected of organisms in a C3 environment. This is because CO 2 formed by the decomposition of organic matter on the forest floor cannot readily escape into the atmosphere. As the CO 2 is recycled by forest plants, animals, and bacteria, the S13C value of the CO 2 becomes more negative, resulting in more negative than average values for the entire forest floor biomass. This is known as the "canopy effect." Forest animals in Africa (Ambrose and DeNiro 1986a; van der Merwe et al. 1988; Vogel 1978b) and in Amazonia (van der Merwe and Medina 1989,1991) have S13C values more negative than animals living in nearby savanna habitats. As with the case of identifying dietary legumes by carbon isotope ratios, the small difference between forest and open habitat animals makes it difficult to distinguish between only those two food sources in highly varied human diets. Nevertheless, the canopy effect on S13C values is important to take into consideration with forest dwelling populations, as the assumption of a global average S13C value of — z6.$%o representing a C3 plant biomass (Vogel et al. 1975) would be inappropriate for a forest habitat with S13C values between — 30 and — 37%o (Medina and Minchin 1980; Medina et al. 1986; Sternberg et al. 1989; van der Merwe and Medina 1989, 1991). Freshwater and estuary habitats. Inland freshwater lacustrine and riverine environments usually have a C3 foodweb, while estuaries may have C3 and C4 foodwebs. Various species of sea grasses have photosynthetic pathways resulting in carbon isotope compositions similar to C4 plants, and are largely responsible for the C4-like character of some nearshore and estuary habitats. A large number of animal species (such as turtles, shore birds, manatee) feeding in estuarine and coastal environments will have S13C values intermediate to C3 and C4 food chains (Norr 1990; Schoeninger and DeNiro 1984). Therefore, interpretations of the S13C values of archaeological bone in freshwater and estuary habitats are highly dependent upon knowledge of the isotopic compositions of the local food resources. Nitrogen isotope composition also may vary in these environments. Schwarcz et al. (1985; Katzenberg 1989) traced unusually high S15N values in prehistoric human bone collagen from southern Ontario to a diet of Great Lakes fish with high S15N values. A similar phenomenon may be the reason for consistently elevated S15N values at the Contact period central Gulf site of Tatham Mound in
Interpreting dietary maize from bone stable isotopes
205
Florida (Hutchinson and Norr 1993). Estuarine animals, including shore birds, invertebrates, and nearshore-feeding fish will often have 615N values that fall intermediate to terrestrial and marine animals (Norr 1990; Schoeninger and DeNiro 1984). Dietary carbon isotopes in consumer tissue fractions The incorporation of natural isotopic variation of producers into consumer body tissues is well demonstrated (DeNiro and Epstein 1978, 1981; Schoeninger et al. 1983). Exactly how this is accomplished is less clear. Accurate dietary interpretation of the isotopic composition of body tissues is dependent upon a clear understanding of how different types of foods (for example, meat versus grains) and their biochemical fractions are incorporated into those body tissues. The current debate. How are the stable isotopes of carbon in different nutritional components of the diet incorporated into the different body tissues and tissue fractions of the consumer? While many researchers acknowledge that this is an important issue (for example, Chisholm et al. 1982; Klepinger and Mintel 1986; Krueger and Sullivan 1984; Lee-Thorp et al. 1989; Parkington 1991; Schoeninger 1989; Schwarcz 1991; Sillen et al. 1989), few studies have been undertaken to answer the question. One theoretical model assumes that the carbon atoms of collagen have an equal chance of coming from any digestible portion of the diet, including proteins, carbohydrates, and lipids (fats). Those who use a linear mixing model to attempt calculations of "percent dietary C 4 " or "percent dietary marine" from the S13C value of bone collagen are making this assumption (for example, Schoeninger 1989; Schwarcz et al. 1985; Spielmann et al. 1990; van der Merwe 1982; White and Schwarcz 1989). This is an appropriate exercise for the dietary reconstruction of herbivores consuming some proportion of C3 versus C4 grasses (Tieszen 1991; Vogel 1978a), or for environmental reconstructions of C3 woodland versus C4 grassland environments from soil organic matter or soil carbonates (Ambrose and Sikes 1991; Ceding et al. 1991), because the diets or plants being compared are similar in the quality of their protein. This model for dietary reconstruction may be seriously inappropriate for omnivores consuming protein, carbohydrates, and fats/lipids that differ in their carbon isotopic compositions (Klepinger and Mintel 1986). A different theoretical model assumes that the carbon atoms of collagen come mainly from dietary protein (Chisholm et al. 1982; Chisholm 1989; Krueger and Sullivan 1984). At the very least, the essential amino acids come from dietary proteins since they cannot be synthesized by the body. It is not known which dietary fractions (for example, proteins, lipids, carbohydrates) are used to synthesize non-essential amino acids. Essential amino acids comprise only 12% of collagen, but contain 18% of its carbon atoms (Ambrose 1993; Sillen et al. 1989). Along with their model for routing dietary protein carbon atoms to collagen,
2O6
LYNETTE NORR
Krueger and Sullivan (1984) proposed a model to explain the observed differences between S13C values of bone apatite carbonate and bone collagen (A13Cca.co) among animals at different trophic levels. Bone carbonate is derived from blood bicarbonate, which is generated by the cellular metabolism of energy substrates (Poyart et al. 1975). Therefore, Krueger and Sullivan (1984) proposed that the S13C value of bone apatite carbonate is a function of the S13C value of the energy substrate, usually carbohydrates and fats. The use of bone apatite carbonate for dietary reconstruction was proposed by Sullivan and Krueger (1981, 1983), but discouraged by Schoeninger and DeNiro (1982, 1983) because of problems with bone diagenesis. Subsequent research (Krueger 1991; Lee-Thorp 1986,1989; LeeThorp and van der Merwe 1987,1991; Lee-Thorp et al. 1989) demonstrated that the carbon isotope ratios of carbonate in archaeological bone apatite can be used for reconstructing paleodiet. Experimental results. The relationship between dietary carbon atoms and those of consumer body tissues is under investigation by Ambrose and Norr (1993). Following the model suggested by Krueger and Sullivan (1984), a series of controlled diet experiments was designed to determine the degree to which dietary protein may be routed to bone collagen and dietary energy to bone apatite carbonate. Laboratory rats were raised on purified, pelletized diets for which the isotopic composition of proteins, lipids, and carbohydrates were known and their proportions controlled. Casein, a milk protein from cows fed either C3 or C4 diets, was used as the protein source. C3 energy was provided by beet sugar, rice starch, and cottonseed oil, and C3 fiber from wood cellulose. C4 energy was provided by cane sugar, corn starch, and corn oil, and C4 fiber from corn bran. The diets were formulated to produce combinations of pure C3 protein and energy, pure C4 protein and energy, C3 protein with C4 energy, and C4 protein with C3 energy, with very low (5%), normal (20%), and very high (70%) levels of protein. Spermpositive albino mother rats were started on each diet one day after conception, and four to six of their offspring were raised beyond sexual maturity on the same diet. The bone collagen, bone apatite, muscle, and hair from the experimental animals were analyzed for their stable carbon and nitrogen isotopic compositions. Two experiments are still in progress: (1) a diet with 20% C4 protein and C4 energy; and (2) a diet with 20% C4 protein plus C3 energy. The initial results of the laboratory experiments by Ambrose and Norr (1993) demonstrate that the S13C value of rat bone collagen largely reflects that of dietary protein rather than that of the whole diet. For example, rats on diets with only 5 % protein had 42% to 51% of the carbon in their bone collagen derived from the carbon in their dietary protein. Therefore, it was concluded that the carbon isotope composition of bone collagen disproportionately reflects that of the dietary proteins rather than being "scrambled" with those from carbohydrates and lipids. Because of "routing" of dietary protein to bone collagen, the relationship between the S13C of the whole diet and that of the collagen varied
Interpreting dietary maize from bone stable isotopes
207
considerably (— 2.2%o to + 9.6%o, a range of n.8%0) as the isotopic composition of the protein differed from that of the whole diet. The monoisotopic C3 diet had an intermediate A13Cdiet-coiiagen of 3.8%o, and it is expected that the monoisotopic C4 diet will have a similar value (Ambrose and Norr 1993: Table 4). This difference between the S13C of the whole diet and that of the collagen is often referred to as the diet-to-collagen "fractionation factor" (A13Caiet-coiiagen)- The A13Cd_co is considered to be 4.5 to 5.o%o for humans and large herbivores (van der Merwe 1982; Vogel 1978a), but ranged from 3.5 to 4.4%o for mice fed different diets when lipids were removed from bone prior to the analysis of the collagen (DeNiro and Epstein 1981). Now it has been demonstrated that this is not a species difference in A13Cd-CO9 but a variable that is dependent on the isotopic composition of the dietary protein relative to that of the whole diet (Ambrose and Norr 1993). In contrast to collagen, bone apatite carbonate S13C values of the rats raised on experimental diets closely followed the carbon isotope values of their whole diet. The mean difference between the S13C values of the diets and the carbonate (A13Cdiet-carbonate) averaged +9.5±o.6%o. This small variation in the A13Cd-ca suggests that bone apatite carbonate is a better isotopic measure of whole diet than is bone collagen (Ambrose and Norr 1993: Table 4). Because there are general similarities in the digestive physiology of rats and humans, the relationships between the isotopic composition of the diet and bone collagen and carbonate in humans are likely to be similar (Karasov and Diamond 1988). If this is the case, bone apatite carbonate from humans will more accurately reflect the isotopic composition of the whole diet than will bone collagen. Preliminary examination of isotopic results from archaeological human bone (below) suggests that humans follow the same patterns predicted by the controlled diet study with laboratory rats. The use of S13C values from both bone collagen and bone apatite carbonate will give a more accurate indication of diet than will collagen alone. The relationship between S13C values of carbonate and collagen (A13Ccarbonate-coiiagen) varies predictably with the isotopic composition of the protein and energy sources because bone collagen disproportionately represents dietary protein, and bone carbonate accurately represents the whole diet. When the dietary protein and energy sources have similar isotopic values, then the differences between the bone carbonate and collagen S13C values have an intermediate spacing (experimentally, a monoisotopic C3 diet gave a mean A13Cca.co value of 5.7%o±o.4%o). When the dietary energy is more negative than that of dietary protein, for example, a C 3 plant and a marine fish diet, the difference between the S13C values of the bone carbonate and collagen is small (experimentally, a 5% C4 protein with C3 energy diet gave a mean A13Cca.co= i.2%o± o.i%o; and a 70% C4 protein with C3 energy diet gave a mean A13Cca_co = 2.i%o±o.2%o). When the dietary protein is more negative than that of the dietary energy, for example, a diet of terrestrial C 3 protein, such as deer, along with maize, the difference between
2O8
LYNETTE NORR
the S13C values of the bone carbonate and collagen is large (experimentally, a 5% C3 protein with C4 energy diet gave a mean A13Cca.co = io.8%o ± o.4%o; a 20% C3 protein with C4 energy diet gave a mean A13Cca.co = n.3%0 ± o.4%o, and a 70% C3 protein with C4 energy gave a mean A13Cca.co = j.z%o ± o.3%o; Ambrose and Norr 1993: Table 4). While the carbonate phase of bone or of tooth enamel provides an accurate record of the carbon isotope composition of the diet, the issue of diagenetic effects in buried bone, such as isotopic exchange with groundwater carbonates (Schoeninger and DeNiro 1982, 1983; Sillen 1989), must always be addressed. Of the two, tooth enamel apatite carbonate is the least susceptible to diagenetic alteration.
Methodological considerations Choosing an appropriate sample for analysis The research question. Before a group of human bone samples is used for stable isotope analysis, it should be determined whether the results will provide meaningful dietary information. It is advisable, if not necessary, to know the actual or theoretical isotopic composition of the local or regional food web so that the human bone isotope results can be interpreted more effectively. Food isotope ratios can vary in such a manner that an accurate interpretation for one region may not be accurate for another. For example, a S15N value of + i5%o and a S13C value of — 8%o in human bone collagen from a neotropical context would suggest a coastal agricultural diet of marine fish and maize (Norr 1990). The same human bone collagen isotopic results from East Africa could be the result of a diet of meat from large herbivores that grazed on C4 grasses (Ambrose and DeNiro 1986a, 1986b). Another interpretive problem arises when attempting to identify a C4 plant, such as maize, where CAM plants were eaten. CAM plants usually have S13C values intermediate to C3 and C4 plants (Bender 1968; Szarek and Ting 1977), and include cacti (for example, nopal, prickly pear), agaves (such as maguey), epiphytes (for example, vanilla), and bromeliads (such as pineapple). In hot, arid regions, CAM plants have S13C values like those of C4 plants. If the research question is to determine maize consumption in a hot and dry region where CAM plants may have been eaten (for example, parts of the southwestern United States and in northern Mexico), stable carbon and nitrogen isotope analysis of bone collagen and carbonate may fail to be informative. Unlike the large difference in S15N between a marine diet and a terrestrial diet, which can distinguish between 13 C-rich marine fauna and C4 plants, CAM and C4 plants do not differ in their 15N abundance. Thus, there is currently no demonstrated way to isotopically identify maize in a diet that included significant amounts of CAM plants. Using such a population for stable carbon and nitrogen isotope analysis could prove to be a waste of time and money until appropriate methods are developed.
Interpreting dietary maize from bone stable isotopes
209
When it can be demonstrated that the local food resources differ in isotopic composition (for example, terrestrial C3 versus terrestrial C4 plants, or terrestrial versus marine protein), the stable isotopes of carbon and nitrogen can be used to reconstruct paleodiet. Potential research questions are many and varied. They may include, among others: (1) the transition from horticulture to maize agriculture (Norr 1990; van der Merwe and Vogel 1978); (2) the importance of coastal resources in the diet (for example, Chisholm et al. 1982; Johansen et al. 1986; Norr 1990; Sealy 1986,1989; Sealy and van der Merwe 1985; Tauber 1981; Walker and DeNiro 1986); (3) changes in subsistence patterns over time (Hutchinson and Norr 1994; Larsen et al. 1992; Norr 1990,1991; Schwarcz et al. 1985); (4) within-group dietary differences, for example, between males and females (Lovell et al. 1986; Norr 1990: 270-271), individuals of different status (Murray and Schoeninger 1988; van der Merwe, personal communication, 1992), different age groups (Lovell et al. 1986; Fogel et al. 1989), the local population and "foreigners" (for example, Andean highland versus coastal; Verano and DeNiro 1993); and (5) the relationship between diet and health (Norr 1984,1990,1992). In complex ecosystems, such as those found in the American tropics, it is advisable to limit interpretations to relative differences in diet rather than to attempt calculations of percent maize or marine foods in the diet. As we improve our understanding of how different kinds of foods are incorporated into body tissues, and as our knowledge of the variation of the isotopic composition of tropical foodwebs increases, more specific and quantitative questions can be addressed. The individuals. Human skeletal remains selected for stable isotope analysis should be from secure archaeological contexts. Dietary information from stable isotopes will be of little use to the archaeologist if the age and provenience of the sample are not known. The accurate temporal placement of human burials is particularly important when approaching the period of European contact and expansion in the Americas. Sugarcane, a C4 plant, was introduced into the Americas by Europeans during the early 1600s and was widely cultivated in the tropical regions. The isotopic composition of cane sugar and molasses is very similar to that of maize. Moreover, since diet and nutrition among individuals of a population may vary due to gender, age, status, health, or origin, it is important to know the biological age and sex of the individuals, their health status, their cultural affiliation, and their economic status (see Ubelaker, this volume). Analyses of isolated fragments of bone from test trenches and small excavation units can be difficult to interpret if considerable within-population variation and questionable cultural context exists. Skeletal elements. For many years, rib portions were preferentially selected from human skeletal remains for chemical analysis (for example, Katzenberg 1984; Lambert et al. 1979,1982). If skeletal remains were coated with a chemical preservative during removal from the ground, a few ribs were left untreated for
2IO
LYNETTE NORR
subsequent analysis. A few missing ribs from an otherwise complete and intact human skeleton seldom interfered with other osteological and paleopathological investigations. The rib, however, is probably one of the poorest choices for chemical analysis, especially from burials in tropical climates. While there is little variation in the carbon and nitrogen isotopic composition of modern collagen extracted from different skeletal elements from the same individual (DeNiro and Schoeninger 1983), thin and fragile bones like ribs are highly susceptible to increased rates of decomposition and soil organic contamination (humic acids) during burial. A better choice is the midshaft of a long bone, such as the femur, where cortical bone is thick, making both the mineral and organic phases of the bone less susceptible to degradation and contamination (see Stahl, Ubelaker, this volume). Evaluating sample quality The method of sample preparation and its resulting purity can affect the isotopic results and, ultimately, the dietary interpretation. If the sample or fraction analyzed (for example, collagen, apatite carbonate, hair) is not pure or clean, and if the contaminant has an isotopic composition that is different from the tissue fraction being analyzed, then the delta value of the sample will represent in part the contaminant rather than the tissue fraction and the diet of the individual. Contaminants of collagen may be other biochemical fractions found naturally in bone (such as lipids), or they may be external contaminants (for example, rootlets or soil organics). Bone carbonate contaminants can come from isotopic exchange from ground water carbonates or organic carbon. Hair is least likely to be contaminated, but it is seldom recovered in archaeological contexts. Hair can be found with mummified remains in arid environments, where there is little vegetational biomass for soil organic contamination. For many years bone collagen has been the preferred tissue fraction for dietary reconstruction with stable isotopes. Collagen is a complex protein molecule with a slow turnover rate (Tieszen et al. 1983) that can maintain its isotopic integrity over many hundreds or thousands of years. Collagen, however, is susceptible to contamination from soil organic carbon, which must be removed prior to analysis (Masters 1987; Nelson et al. 1986; Schoeninger et al. 1989; Stafford et al. 1988; Tuross et al. 1988). The simplest way to remove soil humic acid from collagen is with a method proposed by DeNiro and Espstein (1981), where bone demineralized with 0.2 to 1.0 molar HCl is then treated with 0.125 molar sodium hydroxide (NaOH) prior to solubilization in 10~3 molar HCl at 95°C (Ambrose 1990; Norr 1990). This method has been criticized by Chisholm (1989), because the NaOH may degrade the collagen while removing the soil organic contaminants. Kennedy (1988), however, found that the benefits of using NaOH far outweighed the problem of small amounts of sample loss. Chisholm (personal communication, 1992) is testing an alternative centrifuge filtration method of
Interpreting dietary maize from bone stable isotopes
2.11
collagen separation and purification, but the effectiveness of this method remains to be demonstrated. Collagen contamination and degradation can be identified for most samples by using simple determinations: the C:N ratio and the percent-by-weight of carbon (%C/wt) and nitrogen (%N/wt) in the sample. The C:N ratio of the extracted residue ("collagen") can be determined from manometric measurements of CO 2 and N 2 . The C:N ratio of the sample should be similar to that of modern collagen, which is theoretically 3.21 (Ambrose 1990, 1993; Kennedy 1988). DeNiro (1985) has shown that collagen extracted from ancient animal bones with C:N ratios between 2.9 and 3.5 most often had S13C and S15N values reflecting the diets of the animals. When the sample C:N ratio is highly divergent from that of modern collagen, there is a greater chance that the delta value of the sample will be different from that of pure collagen. This is particularly true when the isotopic composition of the contaminants is different from that of the bone collagen. Bone samples from tropical climates are particularly susceptible to collagen degradation due to high temperatures and contamination from soil organics (see Stahl, this volume). If such samples are poorly purified, the result can be a contaminated "collagen" residue that can yield highly unreliable results (Ambrose 1990; Ambrose and Norr 1992; Kennedy 1988; Norr 1990). Bone lipids are also considered a contaminant because they have S13C values 5~io%o more negative than bone collagen, but only small amounts are preserved in bone buried in the tropics. The NaOH used to remove humic acids will remove residual lipids from archaeological bone. Polyvinyl acetate (PVA) and other preservative compounds applied to bone complicate isotopic analyses. Preservatives often can be removed through solubilization or scraping prior to the preparation of collagen, with subsequent isotopic analysis yielding acceptable results (Moore et al. 1989; Norr 1990).
The mineral phase of bone is more susceptible to isotopic exchange of apatite carbonate when most or all of the collagen has been lost after burial for long periods of time (Koch et al. 1990; Lee-Thorp and van der Merwe 1987, 1991; Sillen 1989). Apatite carbonate from tooth enamel is a preferable alternative to that of poorly preserved bone, because it is less susceptible to diagenetic alteration (Koch et al. 1990; Lee-Thorp and van der Merwe 1987,1991; Sillen and LeGross 1991; see Stahl, Ubelaker, this volume). Postmortem carbonate contamination and soluble biological carbonates must be removed from the apatite sample prior to isotopic analysis of the carbonate. This can be accomplished through a series of pretreatments of either bone or tooth enamel powder with organic solvents such as ether, or chloroform and methanol (Bligh and Dyer 1959; Williams 1992) to remove lipids (if bone is fresh or recently buried), 50% Clorox bleach (sodium hypochlorite) to remove organics, and dilute (1 molar) acetic acid to remove soluble carbonates (Krueger 1991; Lee-Thorp et al. 1989; Nelson and Featherstone 1982). The selection of either tooth enamel, or bone with most of the original collagen, is recommended for dietary reconstruction with apatite
212
LYNETTE NORR
carbonate. When bone is used as the source of apatite carbonate, animal bones from the same archaeological contexts can serve as controls because herbivores and carnivores on natural diets will have a limited range of variation in their A13Cca.co values. In samples where collagen is not preserved, it is more difficult to assess the validity of bone carbonate carbon isotope ratios (Ambrose 1993). Analytical standards for interlaboratory comparisons The stable isotope data obtained from archaeological samples analyzed by different laboratories should be comparable. Not only are the archaeologists and the geochemists (sometimes one and the same) responsible for presenting the methods employed in the preparations and the analysis of the samples, they are also responsible for reporting or making available to other laboratories the results of isotopic analyses performed on universally available standard materials. Laboratory precision and accuracy are determined through the periodic isotopic analysis of a variety of laboratory and international standards. Through the comparison of results obtained on the same substances in different laboratories, the interlaboratory comparability of archaeological results can be determined. The International Atomic Energy Agency (IAEA) and the National Bureau of Standards (NBS, now NIST, National Institute of Standards and Technology) will supply laboratories with small amounts of standard reference materials whose isotopic compositions have been determined in several different laboratories. Some of the available carbon standards include carbonates NBS 18, 19, and 20, graphite NBS 21, and hydrocarbon oil NBS 22, and nitrogen standards include ammonium sulfate IAEA Ni and N2. Since these standards are available only in small amounts, "working standards" can be run more frequently. Those used in the stable isotope lab in the Department of Anthropology at the University of Illinois are thiourea (Sigma Chemical Co.), for the production of CO 2 and N 2 from combustion (C:N = 0.5), and Corydon calcite, supplied by the biogeochemical laboratory at Indiana University and used as a carbonate standard for reactions. The laboratory's working standards are prepared with each batch of ten to fifteen archaeological samples, and the working standards are checked against the IAEA and NBS standards with a similar frequency. A case study from Panama
The small region of central Pacific Panama is geographically diverse, with mangrove-lined shores, tidal mudflats, and estuaries along the coast, and navigable rivers, fertile floodplains, rockshelters, and caves between the coast and the foothills of the continental divide (Figure 9.2). Evidence for human occupation of this region dates back at least 10,000 years to the Paleo-Indian Period (Ranere and Cooke 1991). At some point in prehistory, a sedentary lifestyle and agricultural subsistence pattern were adopted, leading to the complex system of
Enlarged Area
Caribbean Sea
Panama
ta
d!/
Sitio Sierra f?arita Bay teZZzT^- _ A ir; e r r o Mangote La Mula onagrillo
c
Azuero Peninsula
Pacific Ocean Figure 9.2. Map of Panama with enlargement of study area in central Pacific Panama.
214
LYNETTE NORR
Table 9.1. Abbreviated chronology with central Pacific Panama sites mentioned in the text Period
Description
Period dates
Sites
IV-VIII IIIB IIIA
Village chiefdoms Large Formative village Early Ceramic Preceramic Paleo-Indian; Archaic
300 BC — Conquest
Sitio Sierra La Mula Monagrillo Cerro Mangote quarry, shelter sites
IIB
I, IIA
1200-300 BC 2500-1200 BC 5000-2500 BC 9000-5000 BC
chiefdoms encountered by the Spanish in the sixteenth century (Cooke and Ranere 1992a). But, when in prehistory did maize agriculture first become a part of the subsistence pattern in central Pacific Panama? Stable carbon and nitrogen isotopic data from human bone recovered at the Preceramic coastal site of Cerro Mangote suggest the date is an early one, between 5000 and 3000 BC (see also Piperno, this volume). The cultural sequence for central Pacific Panama begins with Paleo-Indian and Archaic at the bottom of the chronological chart, from about 9000-5000 BC (Table 9.1). At the very top of the sequence are village chiefdoms, which are identified in the archaeological record by 300 BC and continue to the time of conquest by the Spanish (Cooke 1984). The middle of the sequence is marked by cultural advances in subsistence, technology, and settlement. During the Preceramic Period (5000—2500 BC), a number of lines of evidence suggest a horticultural and early agricultural adaptation (Cooke and Ranere 1992a; Norr 1990; Piperno et al. 1985; Piperno, this volume). This is the period when the Preceramic coastal site of Cerro Mangote was occupied (McGimsey 1956, 1957; Ranere 1993b; Ranere et al. 1980). The location and midden contents of Cerro Mangote suggest that coastal and estuarine resources were important dietary items. The subsequent Early Ceramic Period (2500-1200 BC) was similar in midden composition to the Preceramic, but with the addition of Monagrillo Style ceramics to the artifact assemblage (Bird and Cooke 1978; Cooke and Ranere 1984; Ranere 1993b; Willey and McGimsey 1954). Unfortunately, no intact human burials were excavated from Monagrillo (Willey and McGimsey 1954). During the period 1200-300 BC the first large village, La Mula-Sarigua, was settled (Hansell 1988). The stable carbon and nitrogen isotope ratios of human bone collagen and apatite carbonate from the Preceramic coastal site of Cerro Mangote, the coastal Formative village of La Mula, and the upper-estuary village site Sitio Sierra are used to reconstruct patterns of subsistence in central Pacific Panama prehistory. Stable isotope analysis for paleodiet A sample of individuals buried at Cerro Mangote, La Mula, and two components at Sitio Sierra have been analyzed for their stable carbon and nitrogen isotope
Interpreting dietary maize from bone stable isotopes
215
composition of collagen, and the stable carbon isotope composition of bone apatite carbonate (Tables 9.2 and 9.3). The analysis of tooth enamel apatite from a small number of individuals is currently in progress. Bone preservation was fairly good at most of the sites. La Mula was an exception to this because wind erosion had compressed the site deposits and exposed some human skeletal remains on and near the ground surface (Hansell 1988). Remains from Cerro Mangote included both intact primary and secondary burials, and scattered bones from earlier disrupted burials (McGimsey et al. 1966; Ranere 1993b). Most of the Period IV Sitio Sierra remains were recovered from a small cemetery and included intact primary flexed burials. Some of the Period VI Sitio Sierra burials were intact as primary burials, others were bundles or disturbed (Cooke 1984). Collagen preservation at Cerro Mangote and La Mula was poor, with collagen yields from whole bone in the range of 0—2 percent. Sitio Sierra skeletal remains produced greater yields of collagen, around 3 percent, but would have been heavily contaminated with soil organics without the NaOH soak during the collagen preparations. Apatite yield from whole bone from all sites was generally in the range of 40-45 percent. Methods. Collagen was prepared using 1 to 2 gm of clean, dry bone crushed to between 0.25 and 0.50 mm and placed into a 50 ml fritted disk funnel fitted with a teflon stopcock (Ambrose 1990; Norr 1990). Samples prepared prior to 1990 (Norr 1990) were demineralized in 1 m HCl for 20 minutes (following DeNiro and Epstein 1978); samples prepared since 1990 were demineralized in 0.2 m HCl for 24 to 72 hours. Collagen recovery was improved in poorly preserved samples by using the weaker acid. Neutral samples were soaked in 0.125 m NaOH for 10 to 20 hours (samples with especially low yields of collagen were soaked in a weaker solution of 0.0625 m NaOH for the same amount of time with favorable results; that is, improved yields and acceptable C:N ratios). Neutral samples were then put into i o 3 m HCl (pH 3) at 95°C for10 hours, with 100 ^tl 1 m HCl added after 5 hours. The hot solution was filtered, evaporated, and freeze-dried. Then 6 to 8 mg of collagen were sealed in quartz tubes under vacuum with copper granules, copper oxide wire, and silver foil. Combustion occurred at 88o°C, producing CO 2 , N 2 , and H 2 O. Slow cooling to 6oo°C over a 12-hour period allowed the formation of N 2 without the formation of any of the oxides of nitrogen. CO 2 and N 2 were distilled cryogenically and measured manometrically on a glass vacuum line. Isotope ratios were determined on Nuclide and Finnegan Delta-E mass spectrometers. Apatite was prepared from 1 gm of ^0.25 mm bone powder placed in 50 ml centrifuge tubes using methods developed by Lee-Thorp (1986). Organics were removed with 50 percent Clorox bleach, centrifuged and refreshed twice daily until effervescence ceased (usually 36—48 hours). Neutral samples were put into 1 m acetic acid and again centrifuged and refreshed twice daily until effervescence ceased (usually 48-60 hours). Neutral samples were freeze-dried and 100 mg was reacted under vacuum at 25°C with 100 percent phosphoric acid until bubbling
Table 9.2.. Human bone collagen and apatite carbonate stable isotope results from central Pacific Panama. Site Burial
Sex
Cerro Mangote M 69 child 68E infant 69I F ph3ex M 15B M 17 M 2-7 26 M M 23A 20A M 32A M 31E F F 31F F 31G 22A F M 5
% N/wt in collagen 9-9 11.1
7-7 1.6
% C/wt in collagen 29.1
32.6 22.2
4-7 8.9
3-5 4.0
11.1
3-5
10.2
12.0
9-3 3-4 3-4 6.1 3.8
12.3
34-7
26.6
C:N ratio 3-4 3-4 3.4 3-4 3-4 3-7 3.6 3-4 3-4 3-3
4.6
3-i
9-9
3-4 3-5
10.1
S15N%o 7-4
-12.8
7.6
-13.7
7.8
-12.9
7-4 7-7
-14.4 -13.7
8.1
-13.2
7-7 7.8
-13.8 -13.8 -13.7 -13.5
7-3 7-7 a
-12.5
6.6 6.6
-14.1
3.2
a
10.4
3.2
7-7
15-5
3.6
a
3.0
9.1
3-5
5-9
16.3
11.7 12.6
16.7
S13Cco%o
—14.6 -13.9 -14.2 —14.0
S13Cca%o
818O%0
J13Cca-co%o
a
a
a
-4.8 -3-7
9.2
-4.2
-6.7 -6.7 -6.9
-5.0
-7.2
~4-7 -5.6
-6.4 -5.6
a
a
a
a
a
a
-4-7 -5.8 -3.6 -3.6 -3.3 -3.8 -4.9
-7.8 ~7-3
6.7
8.9 10.2
8.7 8.5 8.2
Q Q O.O
-7.1
10.5
-7.2
11.0
-7.8 -6.6
10.6 10.4
-7.0
9.1
-4.6 -4-9 -6.4
~7-i
5.8
-7-1
-5-5 -6.0
-6.7
5-4 7-3
-7.0
7.0
-4.8
-6.6
6.1
-6.4 -6.3
-7.2
4-3
La Mula 14 626—11 829 832 930 11022
Sitio Sierra, Period IV F F 12 F 13 F 14 16 F M 17 M 2-3 6
Sitio Sierra, Period VI -i M M g-2F g-3 F g-4 M g-7 g
Note: a Sample iin progress
4.9
3.2
11.0
— 10.4 -10.3 -13.7 -11.7 -13.0 — 10.9
11.6
3-3 3-5 3-3 3-3
5-9 6.1
16.0
3.2
8.2
17.2
3-3
7-9
3-4 7-3 5-4
9-i 20.4
3-i
9.1
-10.7 -10.3 -10.8
3-3
14.1
3.0
8.4 7.6
-11.6
5-9
29.9 16.9
3-3 3-4
14.7
41.8
1.4 3.2 4.0
11.9
4.2
8.9
6.0
17.2
3-3 3-3 3-3
6-3 14.4
17-3
3.2
35.6
3-2-
15.0
38.3
12.5 IO.I 12.2
9.0
-9.9 -9-5
-6.2 -6.2
-6.3 -6.0
8.6
—10.4
-6.4
-6.3
-6.6 -6.5 -6.4 -6.6 -6.8 -6.9
6.2
4.0
4-4 3-7 5-3 3-5 4.0
8.6
-11.8
-6.6
~7-i
5-2
9.0
-11.6
-7.2
7-3 8.7
-12.4
-13.7
-5.6 -8.3
-6.4 -6.9
-12.5
-8.0
4.4 6.8 5-4 4-5
IO.I
-7-2
-6.9
2l8
LYNETTE NORR
Table 9.3. Summary statistics for central Pacific Panama sites n
Cerro Mangote La Mula Sitio Sierra, IV Sitio Sierra, VI
Minimum
Maximum
Mean
s.d.
13
6.6
8.1
6
749
10.1
12.6
11.68
19s
74 7-3
9-5
8.38
10.1
8.61
0.45 0.97 0.62 0.82
—14.6 -13.7
-12.5 -10.3
-13.68
0.58
6
—11.10
1.41
2Oa
-11.6
0.63
a
-13.7
~9-5 -9.6
-10.73 -12.25
1.19
-3.3 -4.6
-444 ~5-37
0.80
7 5
-5.8 -6.4 -6.4 -8.3
13
-7.8
6
-7.1
7 5
-7.2
i
3
a
13
S Cco%o Cerro Mangote La Mula Sitio Sierra, IV Sitio Sierra, VI S13Cca%o Cerro Mangote La Mula Sitio Sierra, IV Sitio Sierra, VI
16
I
3
13
6
-6.0
0.72
-6.26 -7.14
0.14
-6.98
0.63
-6.80 -6.71 -6.90
0.32
-7.2
-5.6 -6.3 -6.4 -6.4
6.7
11.0
54 3-5 44
7-3 5-3 6.8
9.29 6.30
0.72
4^7 5.26
0.59 0.96
-5.6
1.09
18
S O%<>
Cerro Mangote La Mula Sitio Sierra, IV Sitio Sierra, VI A13Cca-co%o Cerro Mangote La Mula Sitio Sierra, IV Sitio Sierra, VI
13
6 7 5
0.27 0.31
1.21
Note: a
See Norr 1990 for additional individuals.
ceased and the sample was dissolved (usually 3-4 days). CO 2 was distilled cryogenically, measured manometrically, and analyzed by mass spectrometry in the same manner as the CO 2 from the collagen samples. Results. The Cerro Mangote skeletal remains have a mean bone collagen S13C value of — i3.7%o±o.6%oandameanS 15N value of + 7.5%o±o.5%o (Norr 1990). The S13C values are intermediate between C3 and C4 foodchains and are in the range of marine diet. The S15N values, however, are low and indicate that very little of the dietary protein was of a marine origin. The bone collagen stable isotope results from Cerro Mangote indicate a predominantly terrestrial diet with a 13C/12C input greater than that of a purely C3 diet (Figure 9.3; bone collagen
Interpreting dietary maize from bone stable isotopes
2.19
isotope data from human groups with isotopically distinct diets, published by Schoeninger et al. 1983, are illustrated in the figure for comparison). Additional isotopic evidence for paleodiet at Cerro Mangote comes from the carbon isotope ratios in the bone apatite carbonate. Twelve individuals from Cerro Mangote (Tables 9.2 and 9.3) have a mean bone apatite carbonate §13C value of — 4.5±o.8%o. Again, the S13C values represent a diet intermediate between C3 and C4 foodchains and are in the range of a marine diet (the carbonate value of — 4.5 less the average A13Cd-ca of 9-5%o gives a diet S13C of - i4.o%o). The mean difference between human bone carbonate and collagen S13C values (A13Cca.co) is large, 9.2%o± i.3%o (n = 11), which suggests a diet with a protein source that was C3 and an energy source that was C4, analogous to a diet comprised of predominantly terrestrial fauna and maize. How do the results from Cerro Mangote compare to later sites from central Pacific Panama? No isotope data are available for the subsequent Early Ceramic Period, which is represented by the type site of Monagrillo (Willey and McGimsey 1954) and by occupations at Aguadulce Shelter (Ranere and Hansell 1978) and Cueva de los Ladrones (Bird and Cooke 1978) further inland. The skeletal sample from the subsequent village of La Mula, which was obtained during mapping, shovel tests, and test excavations across the site by Hansell (1988), includes nine samples that may represent seven individuals from two temporal components. The high S15N and the intermediate difference between human bone carbonate and collagen S13C values at La Mula (from 5 to y%o; Table 9.2) suggest a marine protein (more C4-like in its carbon isotope composition when compared to the terrestrial fauna) and both C3 and C4 energy. The skeletal sample from later periods was excavated by Cooke (1984) at Sitio Sierra. The two components at Sitio Sierra, roughly 200 BC and AD 1100, have isotopic values essentially identical to one another, but with the mean S13CCO value of the later component more negative by i-5%o. The S15N, S13CCO, S13Cca, and the A13Cca.co values all suggest the diets of these individuals were highly mixed, with both terrestrial and marine protein and both C3 and C4 plant foods (Tables 9.2 and 9.3, Figures 9.3-9.5). Discussion A reconstruction of the isotopic composition of these prehistoric diets is compared to the isotopic composition of the food resources in central Pacific Panama (Figure 9.5). The first step of this interpretive process involves isotopic determinations of the potential food resources. In the case of plants, modern samples were used, and the S13C values were adjusted by + i.5%o to compensate for 12C enrichment in the modern atmosphere from the burning of fossil fuels (Keeling et al. 1979; Tieszen 1991). For animals, mammal bone collagen from archaeological specimens was analyzed. The S15N values were adjusted by + o.6%o and S13C values were adjusted by — 2.4%o to represent the isotopic composition of the flesh, the major edible portion of the animal (DeNiro and
LYNETTE NORR
22O
21
Marine
19 Central Panama Sites
17
8
O A • •
15
jg 13 00
11
Cerro Mangote LaMula Sitio Sierra, IV Sitio Sierra, VI
Terrestrial C 3
9 7
00
-25 -23 -21 -19 -17 -15 -13 -11 -9
-7 -5
Figure 9.3. Isotopic composition of human bone collagen from sites in central Pacific Panama compared to those of human bone collagen from populations with isotopically distinct diets (data from Schoeninger et al. 1983).
Central Panama Sites 0 Cerro Mangote A La Mula
• Sitio Sierra IV •
Sitio Sierra VI
Experimental Diets (Ambrose and Norr 1993) 0
C4 Protein C3 Energy
• Monoisotopic •
9
C3 Protein C4 Energy
10 11 12
u
A C c a _ C 0 %* Figure 9.4. A13Cca.co of human bone from sites in central Pacific Panama compared to that of laboratory rats fed isotopically controlled diets (data from Ambrose and Norr 1993).
Interpreting dietary maize from bone stable isotopes
2,21
1U
14
Central Panama Sites
12
-
10
Z To
O Cerro Mangote A La Mula • Sitio Sierra, IV B Sitio Sierra, VI
6 4
2
cP
0
A
-2
Food Resources
-V.
D n
q?" -
A
Terrestrial Fauna Marine Catfish, Jack Maize Legumes Other C3 Plants
-4 -30 -28 -26 -24 -22 -20 -18 -16 -14 -12 -10 -8 Q 13/^
ryf
Figure 9.5. Isotopic composition of prehistoric human diets in central Pacific Panama based on bone apatite carbonate S13C and bone collagen 815N compared to the isotopic composition of food resources. Epstein 1978, 1981). Fish bone collagen S15N was adjusted by + i.7±o.7%o, and S13C by - 34%o, to represent fish flesh (Keegan and DeNiro 1988). The second step is to reconstruct the isotopic composition of the diet from that of the human bone. Since the apatite carbonate S13Cca values represent the isotopic composition of the whole diet with the most consistent diet-to-tissue fractionation factor, 9.5%o, regardless of the isotopic composition of dietary components (Ambrose and Norr 1993), the mean carbonate (tissue) values from human bone were adjusted by — ^.$%o to obtain a S13C value for the prehistoric diet. The S15N dietto-tissue spacing of 2.5%o (DeNiro and Epstein 1981) was used to determine diet S15N from human bone collagen S15N (collagen S15N less 2.5%o). The error bars in Figure 9.5 represent one sigma standard deviation from the means. The chronological scheme presented earlier, representing time periods during which maize may have been adopted as a dietary component or a subsistence staple, included the Late Preceramic (5000-2500 BC), the Early Ceramic (25001200 BC), and the Formative (1200-300 BC). Based on the isotopic results from Panamanian samples, the inclusion of dietary maize in central Pacific Panama was established by 2500 BC. As early as the Late Preceramic, there may have been small agricultural groups with an inland home-base and seasonal camps on the coast, giving their diets a larger terrestrial component than the faunal remains from the coastal shell midden suggest. The faunal remains from the Cerro Mangote midden indicate that catfish and other nearshore fish species predominate only slightly over White-tailed deer and other terrestrial fauna (Cooke and
222
LYNETTE NORR
Ranere 1989a, 1989b: Figure 26.3, 1993). The mean S15N value for the edible portion of the Parita Bay catfish and Jack (Ariidae from Period IV Sitio Sierra and modern Caranjidae, n = 5), is 4-12.7 ± I.I%O, while the S15N mean for thefleshof local terrestrial fauna is + 4.2 ± i.o%o. The Cerro Mangote diet S15N is estimated to be + 5.3 ± o.8%o. If the catfish and other near-shore and estuarine fish at Cerro Mangote (Cooke and Ranere 1989a, 1993) contributed a little more than 50 percent of the protein intake, then the S15N values of the Cerro Mangote vertebrate diet (the bulk of the protein) would have been + 8.8 ± 2.2%o and the bone collagen + n.4%0, values considerably higher than those obtained. To have the observed bone collagen and apatite isotopic composition, the Cerro Mangote people could have spent about two months per year consuming the protein sources represented in the midden. Thus, it is likely that the subsistence strategy of the Cerro Mangote population was primarily terrestrial, with a seasonal or periodic occupation of the Cerro Mangote site for the exploitation of estuarine, mangrove, and near-shore marine protein resources. Is this interpretation compatible with current archaeological data on settlement and subsistence in central Pacific Panama? Information on the locations of sites collected by Cooke and Ranere from survey transects across central Pacific Panama indicates a ten- to twenty-fold increase in the number of archaeological sites from 5000 to 2500 BC, when compared to the preceding period (Cooke and Ranere 1992b; Ranere 1993a; Weiland 1993). Rockshelters appear to have been occupied repeatedly. Other low-density lithic scatters may be settlement sites up to 10 ha in size. Pollen and cross-shaped phytoliths from a primitive race of maize {Zea mays L.) have been identified by Piperno from Late Preceramic deposits at Cueva de los Ladrones (Piperno et al. 1985; Piperno, this volume), a rockshelter 25 km inland (Bird and Cooke 1978). Sediment cores from Lake La Yeguada in the continental divide have maize and squash (Cucurbita) pollen and phytoliths in the upper end of a segment dated to between 5650 and 4850 BC (Piperno et al. 1988; Piperno et al. 1990; Piperno, this volume). Cooke and Ranere (Cooke 1984; Cooke and Ranere 1984) have suggested that low density slash and burn farming communities, much like those of the modern Guaymi of western Panama, may have inhabited the inland hills of central Panama during the Late Preceramic. The isotopic results from the human burials at Cerro Mangote suggest that the people once thought to be coastal foragers may actually have lived in inland farming communities where maize was planted, with only a limited seasonal exploitation of coastal resources. The Proyecto Santa Maria survey located hamlet-sized sites that could have been such an occupation (Cooke and Ranere 1984). Conclusions Paleodietary reconstruction using the stable isotopes of carbon and nitrogen in bone collagen and apatite can be a complicated, time consuming, and expensive exercise, especially when applying the technique to poorly preserved samples
Interpreting dietary maize from bone stable isotopes
2.23
from isotopically complex environments like those found in the American tropics. Archaeologists and geochemists involved in collaborative efforts of paleodietary reconstruction from archaeological bone need to do much more than pass samples and results back and forth between them. The archaeologist must be familiar enough with the analytical technique to know when it is appropriate to submit samples for analysis, what kinds of samples should be submitted, and the limitations and potential of the analyses. The geochemist must be familiar with the questions that the archaeologist is trying to answer (for example, paleodiet), as well as with the complications that can arise when interpretations are attempted following inappropriate or inadequate methodologies (see Ambrose and Norr 1992). It is the archaeologist who has the ultimate responsibility to be informed and to take responsibility for the interpretation of archaeometric data. An informed archaeologist, who is familiar with the problems associated with dietary reconstruction from stable isotope geochemistry, and who collaborates with a geochemist who is sensitive to diagenetic problems associated with both the organic and inorganic components of buried bone, need not fear or distrust the people in white coats (cf. Flannery 1991). Note Archaeological bone samples from Panama were generously provided by Richard G. Cooke, Smithsonian Tropical Research Institute; Anthony J. Ranere, Temple University; and Charles R. McGimsey III, University of Arkansas, Fayetteville, with the permission of Patrimonio Historico, Panama. Collagen isotopic analyses were conducted under predoctoral grants from the National Science Foundation (BNS 84-07181), the Wenner Gren Foundation (4651), and a Sigma Xi Grantin-Aid; apatite carbonate analyses were conducted as a third-year Postdoctoral Research Associate under NSF grant DBS 92—12466 awarded to Stanley H. Ambrose, University of Illinois, Urbana. I am grateful to Jack Liu, Illinois State Geological Survey, and to Stanley H. Ambrose, Anthropology, University of Illinois, Urbana, for mass spectrometry. I thank Stanley Ambrose, Nancy Sikes, Theresa Schober, and Valerie Williams for helpful comments on an earlier version of this chapter. While I am indebted to many people for the successes in my career, I wish to acknowledge here, as part of my contribution to this book dedicated to Donald Lathrap, the methodological rigor and interpretive insights that Don instilled in me as a student of Neotropical archaeology. As always, loving gratitude goes to my husband, Paul Garber, and to Sara and Jenni Garber for their support in my academic endeavors.
10
From potsherds to pots: a first step in constructing cultural context from tropical forest archaeology J. SCOTT RAYMOND
In this chapter I do not intend to plow new ground, but take a retrospective stance through explicating a particular methodology which has been applied to the classification of pottery in the South American tropical lowlands. I shall refer to this methodology as structural classification, since it is modeled after the methodology of descriptive linguistics. It is not to be confused with structuralism or structuralist analysis (Leone 1982) as those terms are currently used in a LeviStraussian sense in the archaeological literature. It is more akin to what has been called componential analysis, or ethnoscience in anthropological literature of the past four decades. Donald Lathrap was a strong advocate of this methodology, and his unpublished thesis (Lathrap 1962) is the earliest example of a thorough application of this methodology to an archaeological data set. I write this with a slight feeling of apprehension, knowing that classification, particularly as it applies to ceramics, has historically been a hotly debated subject among archaeologists. As Spaulding (1982: 1) astutely observed, "Basic concepts and their implications are often controversial simply because they are basic." If I touch some raw nerves, it is not by intention. My objective is to elucidate a particular methodology which has been used and developed by archaeologists working in the South American tropics over the past three decades. Pottery is hardly the most striking feature of an indigenous Amazonian village. Even before the advent of European metal pots, pottery is unlikely to have been the first thing to catch the eye of a visitor. Much more striking were the thatched long-houses, large canoes and balsa rafts, gardens, ceremonial plazas, and painted bodies of the villagers. Among the smaller objects competing with pottery for the visitor's attention were varieties of baskets, colorful featherwork, wood carvings, painted textiles, weapons, and fishing gear made from cane, wood, and fiber. Far more interesting from an anthropological perspective were the social, political, economic, and religious customs and activities which made use of these objects. However, an observant visitor who wandered over to the garbage heaps and poked around would have found that most of the discarded material remains decayed quickly, and that surviving refuse consisted mainly of bits of broken pottery vessels. 224
From potsherds to pots
225
Because of this bias of preservation, analysis of pottery has been the overriding concern of most investigations of prehistoric settlements in the South American tropical lowlands. For an archaeologist visiting a village a few centuries after it has been abandoned, sherds of pottery are the most noticeable, and often the only, indication of human settlement. Although modern archaeological technology is making it possible to extract more and more clues from ancient tropical forest settlements (see, for example, other chapters in this volume and Roosevelt 1991), ceramic remains have and probably will continue to constitute the main source of interpretive data. It seems audacious to think that much can be learned about a society from the remnants of its pottery, and since throughout much of the Amazon Basin there were native societies for anthropologists to study, it should be no surprise that archaeology had a slow start in lowland South America. Reporting of prehistoric ceramics began as early as the last century (for example, Hartt 1871), but through the first half of this century most of the reported ceramics consisted of small samples for which there were brief descriptions of diagnostic characteristics and rarely specific provenience data. Historical synthesis consisted of plotting the geographical distributions of ceramic traits and attempting to correlate these with presumed dispersions of linguistic groups (for example, Joyce 1912; Nordenskiold 1930). In 1947 Howard presented a correlation of the ceramic distributional data for the whole of the eastern lowlands of South America, from the mouth of the Rio de La Plata to the Greater Antilles. He was frank about the deficiencies of the data and explicit about his methodology. For the purpose of determining "the spatial and temporal relationships of the prehistoric cultural entities" (Howard 1947: 11), the ceramics were classified according to style, trait, and complex. A ceramic style he defined as "the aggregate of the pottery traits." A trait, the minimal component of the ceramic style, was any characteristic which proved meaningful in comparing and contrasting artifacts. Complexes were a "characteristic grouping of traits within a style." This classificatory framework served Howard's purpose, given the poor quality of much of the data, but it did little to advance the methodology of ceramic classification in the South American lowlands. In the 1950s, following trends in North America, it became fashionable among some South American archaeologists to divide assemblages of pottery into sherd types. The objective was to define historical trends within a regional tradition in order to establish temporal relationships among sites. This method has been thoroughly explained (Ford 1952, 1954) and critically reviewed by others (cf. Lathrap 1962; Rowe 1961; Spaulding 1953; 1954a, 1954b). Here I will only point out some particular problems which are accentuated when the method is applied to ceramic assemblages in the Tropical Lowlands. Given poor conditions of preservation in moist tropical sites, the majority of sherds in an assemblage may exhibit few distinctive attributes. Sherd typologies, then, are often defined on only one or two nominal variables. Some 80 to 90 percent of many assemblages, for example, may be sorted according to distinc-
226
J. SCOTT RAYMOND
tions in temper and core color. For some assemblages, it may be that only attributes of temper can be used. Variables of shape and decoration may be represented, but in such low frequencies that they are meaningless in typefrequency charts constructed to show how types increased or decreased in popularity along a presumed time vector. Under such conditions it becomes questionable whether the chart is measuring chronological variation, spatial variation, or either (see Evans and Meggers i960; Lathrap 1964, 1966), and there is very likely a paucity or complete lack of independent evidence to test the question. A further problem has been a tendency for the frequency charts to represent all culture change as gradual, a function of the inevitable mechanical displacement of sherds between occupation layers (Lathrap 1962:40—45). This is accentuated in a moist tropical setting where there is often no clear stratigraphic separation between assemblages. In order to meet the statistical requirement for large numbers of associated sherds, assemblages are defined by areally extensive arbitrary excavation levels, which increases the probability of interassemblage mixture. Unfortunately, though they are created for chronological purposes, such typologies are often required to do double duty as the definitive classification of the ceramic complex, style, or phase; a purpose for which they are decidedly unsuited. The completion in 1962 of Lathrap's classification of ceramics from his excavations in the Peruvian Amazon charted a new path in ceramic analysis. His basic methodological principles have since been followed by many South American archaeologists. Lathrap did not invent his methodology out of thin air but owed a huge intellectual debt to several individuals who had wrestled with classificatory problems before him. He was particularly influenced by the work of Rouse (1939), Spaulding (1953, i960), and Rowe (1959, 1961). At a more fundamental level, his thinking about cultural systematics was shaped by Kroeber (1944, 1957), Sapir (1951), and Kluckhohn (i960), his mentor at Harvard. A further influence, as strong as any of these great intellects, was the experience of carrying out excavations while living in a Shipibo village, a traditional Tropical Forest community with ceramic production fully integrated into the fabric of the society. The distinction between ethnography and archaeology becomes blurred in such a setting, and Lathrap saw his task as that of an ethnographer dealing with the prehistoric past of Tropical Forest Culture. In that capacity it is not surprising that he took effective methodology from ethnography and descriptive linguistics, and adapted it to the study of archaeological, that is principally ceramic, data. Objectives and principles
In its most general application, a classification groups artifacts according to shared attributes. However, in order to be more than a mindless task in which
From potsherds to pots
227
artifacts could be grouped and regrouped endlessly according to different sets of attributes, a classification must adhere to a specific set of rules which are devised to solve a particular problem. In the first place, the rules must determine how the set of artifacts is selected, as well as the variables relevant to the problem. In the approach adopted here, the systematic task facing archaeologists in the tropical lowlands is seen as fundamentally equivalent to that which faces ethnographers and descriptive linguists. To quote Taylor (1967: 32—34): "cultural contexts have to be constructed before a comparative study of culture and the history of culture may be begun," and Glassie (1975: 8): "a method of inquiry must include a synchronic statement as a prelude to diachronic explanation . . . How can you study change before you know what is changing?" In order to proceed, it is first necessary to define and describe the units which are to be analyzed and compared in constructing the contexts. For the ethnographer the study unit is often defined by the boundaries of common residence and interaction among a group or groups of individuals within a given geographical area. Families, communities, lineages, villages, tribes, chiefdoms, and so on are inferred, or constructed, by the ethnographer in writing the ethnography; that is, the historiographic task (Deetz 1988; Taylor 1967). For the linguist, analysis is focused on a community of individuals who communicate in a single language. Although the boundaries are in a sense arbitrary in that the criteria are imposed by the observer/analyst, they are not determined randomly or capriciously. They proceed from an explicit methodology which rests on the principle that the unit of analysis consists of groups of individuals who share a common language, a common set of cultural rules, or who ascribe themselves to a particular social grouping. The same principle must apply to the classification of an archaeological ceramic assemblage; that is, the assemblage must be drawn from an archaeological context which can be ascribed to a group which shared ideas and rules about making and using pottery. Obviously, this ideal is impossible to determine with certainty given the taphonomic variables involved in the formation of archaeological deposits. A principal objective of formal analysis in structural linguistics is to construct a model of a language which is capable of encoding and decoding utterances in that language; that is, a model which has the linguistic competence of a native speaker and which also explains that competence (Chomsky 1966). In applying this to the formal analysis of artifacts the objective is to construct a model which approaches the competence of a native user and/or artisan of a particular style. Such a model ought to consistently distinguish pottery which belongs to the style from that which does not, and specify why. Furthermore, in a predictive way, the model ought to be able to generate stylistically valid specimens which do not occur in the assemblage, that is, it should have a creative capacity. It also ought to distinguish stylistic variation from free variation; that is, variation which has cultural significance from variations in the way an individual or different individuals make the same formal category.
228
J. SCOTT RAYMOND
The use of structural linguistic methodology to analyze cultural variability is not original to archaeologists. Pike (1964) suggested that distinguishing etic from emic variability provided a heuristic structure for describing and comparing cultures. Componential analysis, applied by Goodenough (1957), Wallace and Atkins (i960), and others was rooted in linguistic methodology. Lounsbury (1964) applied the principles of transformational linguistics to define the functioning of Crow/Omaha kinship terminology, and the more recent stochastic kinship algorithms (Buchler and Selby 1968) are also inspired by the methodology of linguists. Two points need clarification. First, although the goal of such an analysis is to construct a model which can achieve what humans do cognitively, there is neither the claim nor reason to claim that it is a cognitive model or what some have called a "mental template." The goal is not to understand, explain, or replicate the workings of the human mind. Second, although recognizing cultural meaning from archaeological data is problematic, the ambiguities are only different in degree and not in kind from ethnographic or linguistic research. Thus if etic and emic variability can be distinguished ethnographically, it should be distinguishable archaeologically. For both cases, the rules defining structure among variables are produced via emic hypotheses constructed through repeated observations of values along dimensions of variability, and through observed patterning. Methodology The key steps in structural analysis are: (1) to define those units which exhibit structure; (2) to determine the dimensions of variability; (3) to identify and describe those values of a variable which affect "meaning"; and then (4) to construct the rules which structure the relationships among the dimensions and generate the units which carry "meaning." "Meaning" is to be understood as how a category of artifacts is evaluated or interpreted in either a functional or symbolic sense by the group which makes and uses it. Since meaning can only be inferred indirectly through the comparative analysis of archaeological contexts, and since archaeological contexts are often ambiguous, it is inevitable that some variability will be spuriously assumed to affect meaning, thus creating some irrelevant "noise" in the classification. The amount of noise should diminish through contextual testing of the classification. In applying this methodology to the classification of ceramic artifacts (excluding figurines, icons, and sculptures), the units which exhibit structure (that is, artifacts) are ceramic vessels, thus analysis must proceed through the comparison of variability among vessels. This seems obvious, but since pots are almost always found as sherds in the archaeological record, its methodological importance is often ignored in the structure of typologies. Many typologies begin by dividing assemblages according to the variables which are most frequently represented in
From potsherds to pots
229
the assemblage (for example, clay/temper, firing, thickness). Types so constructed are unlikely to bear any close relation to the ceramic categories that carried meaning in the cultural context in which the pottery formerly participated, thus further contextual analysis is problematic. The importance of this methodological concern can, perhaps, be made clearer by considering an analogous problem in faunal analysis. If the bones from an archaeological assemblage were to be sorted and counted according to weight, length, width, density without first determining which species they came from and which parts of the respective animal skeletons they represented, most of the useful information would be inaccessible. Potsherds, then, must be analyzed as parts of pots. Therefore, the first analytical task is to infer, as completely as possible, the formal characteristics of the vessels. For some assemblages which are particularly badly fragmented, the inference of structure may be very incomplete. In such instances any classification is unlikely to be very informative. But by using a kind of set theory and studying variability in rims associated with necks, necks associated with body segments, body segments associated with bottom forms, and other combinations it is possible to deduce most of the vessel shapes even in a fairly fragmented assemblage. Once the range of formal variation in vessel shapes is known, the variables which define the shapes are determined, and specimens are then compared across these variables. If formal variation in shape is meaningful, such a comparison will lead to the determination of specific values which distinguish the different vessel shapes. Such values are analogous to linguistic phonemes. Deetz (1967: 88) has proposed that they be called "factemes," but in the South American literature they are referred to as "modes," a term which Lathrap (1962: 218) appropriated from Rouse (1939). I shall use the term "mode" hereafter. Modes, then, are values ranged along dimensions of variability, and are assumed to be minimal units of formal variation which affect meaning. Modes are mutually exclusive properties of a nominal variable. They may be defined as discrete attributes (for example, an everted rim or a vertical rim), or as values along a continuous scale (such as mouth diameters). In either case they will exhibit a range of variation which can be narrow or broad. Distinct modes should be statistically separable (for a discussion of this see Spaulding i960, 1982). If distinct modes cannot be defined along a dimension of variability, it suggests that the dimension was not culturally significant (Lathrap 1962: 236); however, modal variation may be obscured by taphonomic processes. If, for example, modal sizes of vessels increased gradually over a period of time, and if the analyzed assemblage was deposited over a long period of time, then three size modes might look statistically like two or only one. Also, as DeBoer (1974) has shown, larger vessels tend to have longer life spans than smaller vessels and, therefore, accumulate more slowly in an archaeological assemblage. If each size mode of smaller vessels had a broad range of variation, this could distort the
230
J. SCOTT RAYMOND
statistical expression of the modes and obscure their cultural significance. These and other cautionary examples mean that modal structures inferred from archaeological assemblages will always suffer a degree of imperfection. Once the modes of shape have been determined and their structural relations defined, other variables are analyzed for modal variation. The dimensions of clay, temper, or firing, for example, may prove to be multi-modal and structurally related to modes of shape. Similarly the modes of surface finish and decorative techniques must be determined and structurally incorporated into the generation of vessel categories. The modal analysis of decoration and decorative techniques on ceramics has been something of a stumbling point. As stated above, modes are mutually exclusive (that is, two modes from a dimension of variability cannot cooccur in the same structural environment). Yet two modes of decoration may occur on the same pot (for example, more than one kind of incision, painting, punctation). Lathrap (1962: 233) was aware of this problem and suggested that such occurrences were indicative of a system with "minimum structuring." The problem, however, is methodological. Design and decorative technique must be analyzed as a structural system separate from the shape and fabric of the pot. Decoration, then, is the set of rules which structures the application of designs to a vessel shape. Just as it is necessary to construct (at least on paper) complete vessels to analyze vessel form, it is necessary to construct complete design statements in order to understand design structure. Again, more often than not only tiny fragments of design are preserved for analysis in an assemblage of potsherds. In such circumstances it is tempting to describe decoration in very simple terms by creating decorative types (for example, zoned-painted ware, white-on-red ware, fine-line incised), or by producing a laundry list of variables of decorative technique and design elements. It may be possible to squeeze more information out of the data by the application and interpretation of multivariate statistical tests, but much important structural information will be inaccessible for further contextual analysis. Unless the potsherds have been broken up into small bits, it is usually possible to analyze part of the design structure. And if different sherds have preserved different, but overlapping, segments of similar design statements, it may be possible to construct nearly complete statements. The task is similar to that of trying to decipher a code from many scraps of different written messages. You may not understand the messages but you will begin to identify similar elements and discover how the elements are structurally related to each other. In cases of high redundancy among the design statements (or messages) the task of decipherment is easier than in cases of low redundancy. Therefore, the chances of successfully constructing complete design statements varies not just according to the degree to which designs have been fragmented, but also according to a culturally determined variable (that is, the diversity of the designs). Furthermore, decorating pottery is commonly not just a craft, but an art or an aesthetically
From potsherds to pots
231
creative medium. This means that among some communities of potters, whose fragmented efforts we might be faced with in an archaeological assemblage, there was an emphasis on generating aesthetically new designs. In other communities, decorating pottery may simply mean selecting one from a set of designs to copy. In these cases variability among designs in the assemblage reflects individual differences in skills of craftsmanship and differences in favorite designs. If complete design statements can be constructed, it becomes analytically possible to separate the minimal elements of design, and to specify the rules which structure the generation of designs. Using such a design grammar, it is possible to describe individual design layouts as algorithms and to create new designs which would be "grammatically correct" but not necessarily represented in the assemblage. The grammar should also incorporate the modes of decorative techniques, defining their structural relationships. For example, if for the dimension of incision there is more than one modal line width, the rules must define their contextual relationship within the design structure. If such rules cannot be specified, the analysis has failed to demonstrate that there is more than one mode, and the observed variation in line width must be assumed to be free variation (that is, of no cultural significance). By analyzing decorative technique as part of the design structure, the ambiguity of more than one mode occurring on a single pot (or potsherd) is removed, and the criteria for determining modes are clarified. However, with small fragmented assemblages, it may not be possible to carry the analysis of structure this far. A final set of rules directs design execution within specific decorative fields defined according to the set of vessel forms. For some styles it will be necessary to specify structural relationships among decorative fields if more than one occurs on a single vessel form; for other styles the decorative fields are independent of other fields, even on the same vessel. Design elements may transform as they are accommodated to a particular decorative field, and a complete grammar should have a loop which feeds this information back to the generative rules of design. Application It is beyond the scope of this paper to present a full structural analysis of a ceramic style. Here, for the purpose of illustration, I present a brief summary of a ceramic assemblage classification from the Upper Amazon of Peru. Excavations carried out in 1970 at the Granja de Sivia in the Lower Apurimac Valley (Figure 10.1) recovered an exceptionally rich assemblage of ceramics representing a previously unrecorded ceramic style (Raymond 1972). Although there were no complete vessels, a very high percentage was reconstructed through painstaking comparison and refitting. Thus it was possible to define the culturally significant vessel shapes, identify the dimensions of variability and modes which distinguished them, and construct the rules which specified the structural relationships. These are given in schematic form in Figure 10.2.
232.
J. SCOTT RAYMOND
A
Quimpiri Complex
•
Sivia Complex
Figure IO.I. The lower Apurimac and Upper Ene Valleys, showing locations of archaeological sites of the Sivia and Quimpiri complexes. There were three classes of decoration based on the techniques used to execute the designs: zoned-painting; resist painting; and applique. The applique designs were simply redundant representations of the same theme, an anthropomorphic face, and did not warrant creation of a structural grammar. The resist-painted designs were so poorly preserved that it was not possible to construct the
From potsherds to pots
2.33
BASAL ANGLE VESSEL FORM
1
Figure 10.2. Diagrammatic illustration of rules which generate the most common vessel forms of the Sivia style. The following key is used: X, modes which are most commonly represented; O, modes which are frequently represented; R, modes which are rarely represented. structural relations of design elements. However, the partial representation which could be discerned suggested that the underlying rules were the same as those for the zoned-painted designs, with differences arising from constraints of differently shaped decorative fields and different qualities of the decorative techniques. A very large number of complete zoned-painted design statements was recorded. Despite a great deal of variability among the designs, there was also a high consistency and redundancy among the several thousand specimens to compare. Therefore, this was an ideal assemblage for separating the minimal structural elements of design, and for constructing the rules of design generation. These rules, together with complete design statements, are graphically presented in Figures 10.3 and 10.4. Finally, Figure 10.5 illustrates the rules for applying designs to vessel forms and the generation of vessel categories. Interpretation Although the 1973 excavations at the Granja de Sivia greatly expanded the ceramic inventory, the structural classification proposed for the smaller 1970 assemblage (Raymond 1972) encompassed all the variability of shape, except for
234
I- SCOTT RAYMOND
VI. Deletional Rules
III. Relational Rules
VII. Merging Rules
IV. Conformational Rules
VIII. Substitutional Rules
X. Rules of Contrast
V. Contractional Rules
Figure 10.3. Diagrammatic illustration of design rules which define the generation of incised and zone-painted design statements of the Sivia style (modified from Raymond et al. 1975: Figure 55(a), 55(b)). an extra-large size mode which functioned as a burial urn. I also predicted design statements that occurred in the new, but not the former, assemblage (Raymond et al. 1975). In other words, the structural classification successfully defined the Sivia style, as represented at the Granja de Sivia. Therefore, any deviations in the modal values and/or structural rules become obvious and can be evaluated for their possible cultural significance. I will first examine some hypothetical possibilities. We may suppose that Sivia potters may have occasionally copied designs from other styles. This should be apparent from the presence of a "foreign" design on a local vessel shape, that is, vessel decoration which cannot be generated (or predicted) from the structural rules. If such a specimen were found in a Sivia assemblage, it would be reasonable to infer that it was made by a Sivia potter who had seen some pottery from style X, or perhaps the design on a woodcarving or textile, and decided to copy it onto one of her pots. But it could, of course, have been made by an X potter who copied the shape of a Sivia vessel and applied his/ her own design. In the latter case, the association of the vessel with a Sivia assemblage could be explained in several ways. For example, the "foreign" potter may have been visiting or living in a Sivia settlement, or perhaps the vessel was a gift. To go further in specifying the cultural significance we would need to find additional clues on the specimen, know the structure and grammar of style X, and/or have independent evidence with which to test the possibilities.
From potsherds to pots
Figure 10.4. A sample of design statements of the Sivia style.
2.35
J. SCOTT RAYMOND
EXTERIOR BODY ZONE NECK ZONE VESSEL FORM
VESSEL CATEGORY
Figure 10.5. Diagrammatic illustration of rules which define the decoration of the most commonly decorated vessel forms of the Sivia style and of the generation of vessel categories. Note: when zoned painting occurs, it is always in association with incised designs. Resist designs are always over a red slip. It is easy to think of specimens which could be interpreted as borrowings, imitations, or fanciful whims of Sivia potters: (1) a "foreign" vessel shape with correct Sivia designs; (2) Sivia design elements incorporated into an unrecognizable design statement; (3) a design statement from a known other style executed according to Sivia decorative techniques, and so on. These are simply interesting curiosities, however, unless they can be related to a larger socio-cultural context. With that in mind let me turn to some specific examples from the Granja de Sivia assemblage and consider them in light of the probable regional social-cultural environment at the time of occupation. Quimpiri is one of two ceramic styles coeval with the Sivia style in the Lower Apurimac Valley. Quimpiri and Sivia occur separately at several sites (Figure 10.1). Their exact temporal boundaries are uncertain, but they co-existed in the Lower Apurimac region at least during the tenth to thirteenth centuries. I have suggested elsewhere (Raymond 1983,1985) that Sivia was associated with larger settlements oriented to the river, and Quimpiri with small settlements situated on
From potsherds to pots
2.37
10c cm. Figure 10.6. Illustration of a Quimpiri style vessel with decoration (A) and two sherds from two pseudo-Sivia bowls which display Quimpiri style decoration (B andC). steep hills above and back from the river/Furthermore, based mainly on the settlement data, I proposed that they had different political/social/economic organizations. Superficially the two styles are very similar. Both assemblages exhibit comparable frequencies of decorated sherds and share several other decorative techniques. Fine-line incision, used to outline contrasting bands of color, is the most common form of decoration in each complex. Even some design elements, such as step frets and scrolls, are shared. Indeed from a listing of traits, the differences between these two complexes do not emerge clearly. It is only as analysis proceeds from potsherds to pots, and as the structure and generative rules are defined, that their distinctiveness is clearly revealed. Excavations in 1970 and 1973 at the Granja de Sivia recovered tens of thousands of potsherds (Raymond 1972; Raymond et al. 1975). Ninety-nine percent of the sherds conform to the rules of the Sivia style. The remaining one percent is represented by a few sherds from the Quimpiri style and a third contemporaneous style which I have associated with neighboring Andean populations. Two sherds, however, do not conform to any known style, yet they share characteristics with both Sivia and Quimpiri (Figure 10.6). In a quantitative typology these sherds would almost certainly be statistically insignificant. In a typology which sorted according to decoration, they would probably have been lumped in a type which could have been called "Sivia fine-line incised" and which
238
J. SCOTT RAYMOND
would have included nearly a third of the assemblage. In a type-variety classification they would have been sorted, perhaps, into an infrequent variant of a type. None of these typologies, however, would have informed us about the social or cultural context. A structural analysis of Sivia and Quimpiri, however, allows one to specify exactly how these two sherds deviate from the two styles and possibly infer something of social significance. The vessel shape does not occur in Quimpiri but is closely similar to a carinated bowl form of Sivia. However, the sidewalls are too convex; the bottom below the carination is too shallow; the everted lip is too exaggerated to qualify as a "proper" Sivia bowl; and the respective values are outside the range of free variation permitted in Sivia. Modes relating to variables of vessel construction reveal other meaningful differences. The proportionate thickness of the vessel wall is less than that of the thinnest walls of Sivia vessels, and both sherds are oxidized beyond the moderate firing standard of Sivia. The combination of subtle deviations from Sivia shape and manufacturing modes suggests that the potter was not completely familiar with Sivia modal values. The even thickness of the sidewalls, the even surfaces, and the consistent arch of the horizontal profile, however, indicate that the potter was an experienced craftsperson. It is possible, that these sherds resulted from the experiments of a creative Sivia potter; however, analysis of the incised decoration suggests otherwise. First, although at a glance the designs look as if they could be Sivia designs, analysis shows that the monotonous Sivia rules for design generation were not followed and that the rules of symmetry conform to those of Quimpiri. Second, one of the sherds (Figure io.6(c)) exhibits a distinctive Quimpiri design element. Finally, the vessel sidewall on the other sherd (Figure io.6(b)) has been divided horizontally into two design fields, which would be a violation of an inviolable rule in the Sivia style, yet acceptable in the canons of Quimpiri ceramic art. The two sherds, then, would appear to be examples of Quimpiri designs applied to imitation Sivia vessels. The most probable social implication is that they were made by a Quimpiri potter. The highly oxidized firing, which conforms to the Quimpiri firing standards, is further support for the inference. Had it been a Sivia potter, one would expect the vessel form, wall thickness, and firing to be true to the Sivia standards. The structural analyses defining the Quimpiri and Sivia ceramic styles, then, allow us to go further in constructing the social context at the Granja de Sivia. The association of "foreign" pottery from highland neighbors is the first and most obvious evidence of inter-ethnic relations, and one which would probably have been noticed without a structural analysis because of the very apparent differences in the ceramic styles. The presence of Quimpiri pottery is a second indicator of inter- ethnic contact or exchange, and one that would not have emerged as clearly without comparisons through structural analysis. Finally, the identification of imitation Sivia vessels made by a Quimpiri potter not only
From potsherds to pots
239
indicates inter-ethnic relations but suggests the nature of those relations, that is, that a Quimpiri potter was in residence at the Granja de Sivia. This further suggests that there was a flow (or trickle) of adult personnel from Quimpiri settlements and not just contact or exchange between the two groups. Comparable analyses of Quimpiri assemblages should indicate whether the flow was in both directions. A second example of the potential application of structural analysis concerns the comparison of Sivia with Cumancaya, a style broadly contemporary with Sivia but occurring at sites along the wide flood plain of the Ucayali about 300 to 400 km down river from the Granja de Sivia. Comparisons (Raymond et al. 1975) reveal cognate vessel forms and similar structural rules between the two complexes, but distinct differences in the values of modes and a much broader "vocabulary" of vessel forms in Cumancaya. Decoration in Cumancaya is much more varied and design statements are more elaborate. But again, though there are some variations in the basic elements, the rules of design generation are closely similar. The Cumancaya potters simply seem to have been motivated to be more creative and elaborate in using those rules. Cumancaya also used a wider range of decorative techniques. Although there is disagreement about the ethnic/linguistic implications of the similarities and differences between Sivia and Cumancaya (DeBoer and Raymond 1987; Lathrap et al. 1985,1987), it is plausible that Sivia potters were participating in a stylistic tradition which associated them with other peoples who controlled and dominated the main river systems of the Upper Amazon in Peru. This hypothesis emerges from the comparisons of structures of pottery styles and can be examined through further similar investigations and comparisons from other sites. My third and last example illustrates the value of including the technology of production in a structural classification. Logistic constraints and laws prohibiting the removal of cultural materials from the home country, as well as the need to complete reports within a reasonable time frame, have meant that much of the technical analysis of pottery from the South American lowlands has been carried out in the field, under poor lighting conditions with the assistance of only a handlens. Analyses were typically limited to visual, often inaccurate, assessments of tempering and firing. As I mentioned earlier, tempering and core colors have been used as diagnostics in defining types, but rarely have these been structurally integrated with the production and decoration of ceramic vessels. This example concerns the question of whether pottery making was invented independently or whether this technology diffused from a small number of centers. This is a question about the transfer or transmission of technological knowledge of pottery making; yet, it has almost always been examined through trait list comparisons of decoration and vessel shape. New World archaeologists are familiar with the controversial proposal in the 1960s that pottery-making technology in the Americas began in coastal Ecuador
240
J. SCOTT RAYMOND
during the fourth millennium BC, having been brought there from the Japanese Islands. From Ecuador it was supposed to have spread to other early potterymaking cultures in South and North America (Meggers et al. 1965; Ford 1969). The relationship between Valdivia in Ecuador and Puerto Hormiga in Colombia was a critical first link in the diffusionary chain. Support for the hypothesis consisted of the early absolute dates for the ceramic complexes and trait lists which were supposed to demonstrate the similarities between them. The dates became an immediate source of debate, and a tug of war ensued over whether ceramics were earlier in Ecuador or Colombia, a question which is further complicated by the recent discovery of pottery on the Lower Amazon which may be earlier still (Roosevelt et al. 1991). Dating, however, serves only to determine in which directions the diffusionary arrows should point. It does not resolve the more important questions about whether a transfer of information or technology occurred, and if so what was in fact transferred and what were the cultural mechanisms of transferral. Comparisons of structural analyses of ceramic assemblages take us much further toward an answer than do trait lists. Although a full structural analysis of the earliest Colombian ceramics has yet to be completed, preliminary comparisons with the morphological and design structures of the Early Valdivia complex indicate that the structures are very different. It is possible to transform one into the other, but a similar scale of transformation could transform any early ceramic complex into another. If this analysis is correct, it tells us at least that ideas about what pottery should look like were not transmitted between the two regions, but it does not rule out the possibility that the basic technology of pottery making was transferred. To examine that question, it is necessary to include variables of ceramic technology in the structural analysis. With the technological advances in laboratory equipment of the past two decades (Rice 1987; Rye 1981), it is often possible to learn a great deal about how clay was prepared and how vessels were made. To that end, samples of the Early Valdivia pottery and the earliest Colombian pottery (San Jacinto) were recently subjected to radiographic analysis, petrographic examination, and scrutiny under powerful binocular microscopes (Raymond et al. 1994). Such analysis has allowed technological variables to be meaningfully included in the structural analysis. And as a result we can say that the technology of ceramic production was also distinctly different in these early ceramic complexes. The differences include how the clay was prepared, how the vessels were constructed, and the firing technology. A comparison of ceramic structures, then, eliminates much of the ambiguity of the earlier studies and does not support the hypothesis that there were close or direct historical connections between the early Formative ceramic technologies of Ecuador and Colombia. It becomes necessary, then, to consider other explanations for the close timing of ceramic development in the tropical lowlands, and
From potsherds to pots
241
that necessitates an examination of the social and economic context of ceramic production in the prehistoric settlements.
Conclusion Structural classification as a method for classifying ceramic remains is predicated on the assumption that the first step in the synthesis of archaeological data must be historiographic, that is, construction of "cultural contexts." A structural classification creates a "descriptive grid" for inter- and intra-assemblage comparisons. It is important to remember that a classification is only a beginning step toward constructing cultural contexts of the past, not an end in itself. Classes of artifacts generated by structural classifications, such as the vessel categories illustrated above, should not be confused with functional classes or functional types. Function, whether it be utilitarian, symbolic, social, or economic, can only be inferred from additional contextual information about specimens representative of the classes. Such information may include such things as associations with other artifacts or features, with food residues, or with usewear patterns. The version of structural classification presented here is explicitly analogous to the methodology of descriptive linguistics. Modes, like phonemes, are treated as the minimal variants which affect meaning. Vessel categories, like morphemes, are the units which enter into a relationship with the content of cultural behavior. There seems to have been the understanding on the part of some archaeologists (for example, Hayden 1984; Dunnell 1971) that categories generated by a structural classification are intended to be congruent with those held in the minds of the makers and users. No such direct correspondence between the structural classification and cognition or cognitive categories is intended. It is the behavioral context which assigns meaning to categories. A large decorated jar may be a door stop, a flower pot, or a beer storage vessel. At the same time it may signal social affiliation and/or demonstrate the artistic prowess of its maker. Conversely, however, meaning cannot be inferred from context without a classification which identifies categories which potentially carry meaning. I have not addressed issues and problems related to quantification and sampling. In conclusion, however, I should emphasize again that a relatively unmixed assemblage is a pre-condition for a successful structural classification. Excavation strategies, then, must be sensitive to potential mechanical disturbances and must have tight horizontal and vertical spatial control. In tropical forest sites in which the physical evidence of stratigraphic separation has often faded, comparisons of artifact distributions among small contiguous excavation units may reveal mixing between components (see Lathrap 1962: 62-202, for a good illustration). Assemblages drawn from small, single component sites,
242
J. SCOTT RAYMOND
however, may be the best data sets for the initial descriptive analysis of a style (Rowe 1961). Finally, a structural classification is not necessarily a substitute for all other kinds of classifications or typologies. Comparisons among assemblages, for example, may show that certain variables are more sensitive measures of interassemblage differences than others. Structural classifications, then, may be only a foundation for derivative typologies used to scale temporal, spatial, or functional differences among sites and assemblages.
11
Returning to Pueblo Viejo: history and archaeology of the Chachi (Ecuador) WARREN R. DEBOER
Prologue An earlier version of this paper was presented to the 1987 Bennington Conference on Lowland South America, an event that Ken Kensinger has been hosting for a number of years. Don Lathrap was not attending the conference, so I sent him a copy of the paper. A few weeks later, Don mailed the manuscript back, his only comment scrawled across the title page: "I hope you get to Pueblo Viejo, for I have been there." I never found out what Don meant by this comment. As far as I know, he was never physically present in the Santiago-Cayapas region of northern Esmeraldas Province (Ecuador), much less at Pueblo Viejo, the legendary home of the Chachi. Perhaps Don's message was just one of those simultaneously reprimanding and encouraging prods. Or perhaps Don was speaking in a shamanic voice: "I have taken flight-medicine to Pueblo Viejo." Maybe it was just some lingering fight with Berkeley where, contra Roosevelt (1989: 58), I completed my graduate training. Who knows? Anyway, Don, I'm still trying to get to Pueblo Viejo. The following paper falls under the rubric of "ethnohistory," one of those hybrid terms in which anthropology asserts a cross-disciplinary status. A main issue is the origin of the 3,000 or so Chachi who currently occupy the tropical forest lowlands centering on the Cayapas Basin of the Esmeraldas coast. Three classes of evidence (or "texts") are brought to bear on this issue: (1) oral traditions that the Chachi have concerning their own history; (2) the imported testimony of Spanish chronicles; and (3) the more anonymous and thereby less tendentious record of archaeological residues. To what extent do these texts reinforce or contradict one another, or otherwise reveal or conceal the Chachi past? A motivation in this endeavor was expressed well by the late Bob Marley; the Chachi, and I suspect Don as well, would concur: "If you know your history, then you'd know where you're coming from."
243
244
WARREN R. DEBOER
Archaeology and history in the tropical forest: problems and promises As many contributors to this volume emphasize, there are severe problems besetting the practice of archaeology in the tropical forest. Preservation of organic remains is often poor and the toll that decomposition takes on material cultures dominated by perishable artifacts can approach totality. In large parts of the Amazon Basin, workable stone is a rare resource, must be traded or transported into stone-poor areas, and was consequently curated, in the form of ground axes or nether stones, nearly to the point of archaeological invisibility. As a consequence, a heavy interpretive burden is regularly placed upon durable ceramics. As Lathrap (1970: 63) put it: The archaeologist has been justly criticized for his preoccupation with pottery, but in the Amazon Basin the minutiae of ceramic style must carry the full burden of our attempts to study old populations, old trade routes, and the boundaries of now extinct political units. But these precious sherds are less important than their contexts. In this regard, tropical forests also present daunting challenges. Burrowing life abounds: leaf cutting ants, excavating crayfish, bushmasters, armadillos, among a host of other fossorial critters, blend stratigraphy (see Siegel, Zeidler, this volume) into a mix that is bound to suggest gradual change even when none exists. Even the earth itself is hardly solid. Rivers cut, shuffle, redeposit, and regularly remake landscapes such that archaeological survey by canoe along contemporary waterways is likely to miss, in a systematic way, the ancient stage of human action (Lathrap 1968a; see Erickson, Zeidler, this volume). This dismal verdict, however, should not be overstressed. I know archaeologists working in eastern North America who cite similar and equally unfriendly forces at work. Furthermore, one might ask, what is it that is charming about tropical forest archaeology? Do only masochists venture into this world where archaeology is particularly difficult? Is this some sort of macho cult within a profession that feminist critics chide as "the cowboy of the social sciences?" Other factors are at work. "Real Tropical Forest Archaeologists" (to borrow and modify Flannery's 1976 caricature) live among native Americans while pursuing their archaeological work. These tropical forest peoples, whose house the archaeologist lives in and whose food he/she eats (perhaps supplemented by imported canned tuna, bottled ketchup, or a tin of Australian butter), are likely to be the workmen who excavate and who, after draughts of manioc beer, contemplate the meaning of what is dug up. Imagine Cahokia, that famous late prehistoric settlement and dirt pile of Illinois, being excavated by the Illini with whom the archaeologist then discusses the symbolism of the Birger figurine (Prentice 1986)! As Lathrap knew, the position of Real Tropical Forest Archaeologists is privileged. Ceramics are still made and can be observed in utilitarian and traditional contexts, mythologies are
History and archaeology of the Chachi (Ecuador)
245
alive, and oral traditions often pointing to specific archaeological sites are recounted. Thus the Shipibo of the Peruvian Amazon still talk about the ancestral settlement of Cumancaya (c.AD 900, see Raymond, this volume) much like the Chachi speak of Pueblo Viejo. No archaeologist who has gone through this ethnographic experience is likely to be the same again. After such an experience, the archaeologist knows full well that "his people" know how to live in a tropical forest setting better than he does and, as a consequence, is less likely to dismiss entirely the historicity of their oral traditions. This credence, perhaps even gullibility, is further buttressed by the often spotty availability of more conventional histories furnished by explorers, missionaries, and other colonial foreigners during their hit-and-run sorties into environments they often misperceived to be forested wildernesses. Let us now turn to this jungle of multiple and differently informed histories. Chachi oral traditions The Chachi maintain strong and specific traditions that their former homeland was situated in the Andes to the east. In a 1570s account of the indigenes of Esmeraldas, Cabello de Balboa (1945: 14) was voicing a Chachi point of view when he remarked: "sus originarios bajaron de la sierra" (their ancestors came down from the sierra). Beginning with Barrett (1925), most ethnographers have accepted the Chachi claim to a highland origin. Recently, however, Moreno Navarro (1979) and Palop Martinez (1986) question the historicity of Chachi oral traditions and argue that the Chachi have always lived in their present territory centering on the Cayapas Basin. In particular, Moreno suggests that these traditions of a former homeland in the Andes are merely fabrications in an ongoing nativist "mythology" of a past Golden Age that, in fact, never existed. Palop accepts this revisionist appraisal on the basis of her reading of early maps and accounts provided by the Spanish invaders. One purpose of this paper is to find out what "really existed" and what did not. Let us begin by looking at the substance of Chachi traditions about their own origin. Although differing in detail, all six of the Chachi traditions listed in Table 11.1 agree in general outline; version I stands alone in referring to an ancient arrival from the sea but otherwise is concordant with the others. The former Chachi homeland was located in the Andes near the present-day and colonial city of Ibarra. The Chachi refer to Ibarra with the Spanish loan word villa or "town" (Lindskoog and Lindskoog 1964). Other toponymic evidence for highland connections can be cited. For instance, Tumbaviro, a settlement located northwest of Ibarra (see Figure 11.1), is a common place name and surname among today's Chachi. In contrast to the dispersed single-house settlements and plantain-based diet of today, the highland Chachi are said to have lived in villages and to have grown maize as a staple.
246
WARREN R. DEBOER
Figure 11.1. Upper left: late sixteenth-century map of northern Esmeraldas (based on Palop Martinez 1986: Figure 2). Lower left: location of sixteenth-century place names as plotted by Palop Martinez (1986: Figure 3) on a contemporary map of northwestern Ecuador. Right: portion of the "Mapa de la Zona Ecuatoriana de Integracion Fronteriza con Colombia" (SIFCE-DE 1973) showing major rivers and the location of Pueblo Viejo favored in the text.
From their home in the Ibarra area, the Chachi migrated westward down the Andes. Four (I-IV) of the six accounts specify that this abandonment of the Ibarra area was prompted by the Spanish invasion of the 1530s. As the Chachi describe the invaders: "hombres blancos y barbudos, armados de truenos y relampagos, cabalgando monstruos endemoniados que comian hierro" (Costales and Costales 1983: 78) (bearded white men, armed with thunder and lighting, and riding demonic monsters who ate iron). All but one account (IV-H) specify that resettlement on the western slopes took place at Pueblo Viejo, or Tusac in Chachi, a site located in the rugged and unoccupied upper reaches of the Santiago River near the Cordillera de Lachas (VI-H). The siting of Pueblo Viejo was determined by the spot where a large jaguar enlisted by Chachi shamans pawed the ground (it should be remembered that the Chachi, like many South American Indians, believe that their shamans can turn into jaguars, usually through the assistance of various psychoactive drugs). The Chachi told Barrett (1925: 32) that in Pueblo Viejo they "lived for a considerable length of time in much of the same manner as before - typical mountaineers, knowing very little of river life, and with maize as their staple food."
History and archaeology of the Chachi (Ecuador)
247
Tranquil life at Pueblo Viejo, however, was not to continue forever. Depredations, especially the stealing of Chachi women, by cannibalistic (II-I) "Indios Bravos," "Malabas" (VI-L), or uyala (the Chachi term for cannibals) made life at Pueblo Viejo increasingly precarious. In response to these attacks, Chachi shamans sent two spies (III- J, V-J) downriver to infiltrate the villages of the Indios Bravos and to steal the magical lance-wands that gave their enemies such power. After a series of adventures, these spies successfully completed their mission and returned to Pueblo Viejo. There Chachi shamans learned to control the magic of the captured lance-wands. Armed with these wands, the Chachi then descended to the country of the Indios Bravos and wiped out their settlements one by one. One version (VI-L) suggests that this victory was not accomplished without difficulty. The "country" of the Indios Bravos is clearly the Cayapas Basin. In Barrett's (1925: 35) full text, the attack against the Indios Bravos is said to have been initiated on the Zapallo Grande, a major eastern tributary of the Cayapas (see Figure 11.1), to have proceeded to the Cayapas proper, and then to have descended downstream to the sea. Some of the Indios Bravos are thought to have escaped the Chachi campaign of extermination by retreating far upstream to the headwaters of the Cayapas Basin. Even today, Chachi of the San Miguel and Rio Grande, headwater affluents of the Cayapas, tell of distant wisps of smoke which are thought to mark encampments of their lingering foes (in this regard, it is interesting to note that one of the tributaries of the Rio Grande is called "Estero de Malabas"). Once on the Cayapas, the Chachi founded Punta Venado (II-M, IV-M). Punta Venado continues to be the premier ceremonial center among today's Chachi, although it has been augmented by the establishment of three other centers (Zapallo Grande, San Miguel, and Tzejpi; see DeBoer and Blitz 1991). These centers are vacant throughout most of the year but host large gatherings of Chachi during the ceremonies of Christmas and Easter (the two ceremonies alluded to in IV-M). It should be noted that according to the Chachi the exodus from Pueblo Viejo to their present territory on the Cayapas was neither instantaneous nor en masse. Rather it was a process that continued over considerable time. If we are to trust the traditions of Table 11.1, one factor that played a role in this abandonment process was disgruntlement over demands by the Spanish priest (V-N) or by the Chachi gobernador (governor) himself (II-N) that males cut their hair short. In Barrett's day (c. 1909), a few families were said to still live at Pueblo Viejo. Today a few elderly Chachi claim to remember having visited Pueblo Viejo during their youth and express an interest in returning, although the trails are now overgrown and forgotten. Altschuler (1964: 7) reported a 1930s attempt by a faction of Chachi to return to Pueblo Viejo where "the Saints were stronger," but this return effort got stalled on the upper Zapallo Grande where the center of Tzejpi was established. For the Chachi, Pueblo Viejo is living history, a place that existed just yesterday and a place that can be returned to, if necessary. Let's now see what the Western chronicles indicate.
Table I I . I . Comparison of six (I—VI) Chachi oral traditions pertaining to their history. (The sequence of events is from A to N. Bracketed dates in column headings refer to time when data were collected.) I
II
Costales and Costales Vittadello (1988: II, (1983: 77-7%), citing 5-31; [1978]) Arcecio Ortiz
III
IV
V
VI
Barrett (1925: I, 3 1 - Illanes L. (1984: 89- Carrasco (1983: 37; [1909]) 9i; [1980-82]) 140-143; [1981])
Estupifian Tello (1983: 39-41)
Homeland near Ibarra
Homeland near Ibarra
Homeland in North Sierra
Inca invasion recognized
Descended from Inca
Paid tribute to Inca
Entrusted with Inca treasure
Keep Atahualpa's ransom
Arrive from the sea, guided by large jaguar B During trial-laden trek, ascend the Andes to spot where jaguar paws the ground C Homeland near Ibarra Homeland near Quito
Homeland near Ibarra
D
Flee Spanish invasion
Shamans invoke large jaguar to find new home on western Andean slopes H Pueblo Viejo founded on the upper Santiago where the jaguar paws the ground
Flee Spanish invasion
Flee Spanish invasion
Shamans invoke large jaguar to find new home on western Andean slopes
Shamans invoke large jaguar to find new home on western Andean slopes
Pueblo Viejo founded on the upper Santiago where the jaguar paws the ground
Pueblo Viejo founded on the upper Santiago; site picked by jaguar
Forest giants eating Chachi
Depredations by cannibal Indios Bravos
Flee Spanish invasion
Flee to avoid conflict
Flee civil strife
Shamans invoke large jaguar to find new home on western Andean slopes A church and small village founded at Pueblo Viejo
Pueblo Viejo founded on the upper Santiago near the Cordillera de Lachas
Depredations by Indios Bravos and Negros
Indios Bravos stealing Chachi women
Depredations by Indios Bravos or Malabas
Two women sent downriver to infiltrate settlement of Indios Bravos
—
Two men disguised as women infiltrate settlement of Indios Bravos
—
Magic lance-wands stolen
—
Magic lance-wands stolen
Chachi shamans steal Malabas magic
K
Table I I . I . (cont.) I L
II
III
IV
V
VI
Indios Bravos defeated by Chachi shamans; Cayapas occupied to the Pacific
Indios Bravos defeated by Chachi and Cayapas occupied
In those days, Chachi were fierce warriors
Indios Bravos defeated by Chachi and Cayapas occupied
After initial setbacks, Chachi defeat Malabas
M Punta Venado founded at the "mitad" where the "paj-ala" bird sang N
Governor prompts male haircutting; dissidents leave Pueblo Viejo
Punta Venado founded where two ceremonies held each year Padre requires male haircutting; dissidents leave Pueblo Viejo
History and archaeology of the Chachi (Ecuador)
251
Western history
The Chachi (their own term for themselves meaning "true people") enter Western history in the late 1580s or early 1590s under the rubric Cariapa ("hermano menor del padre," or father's younger brother in Chachi kinship terminology; Carrasco 1983: 93), the name of a cacique (or chief) who visited the Spanish outpost of Lita in order to request aid against coastal mulatos who were pillaging Chachi settlements (Monroy 1938: 314). The mulatos in question would seem to be the rapidly expanding population descendent from Negro slaves marooned on the Esmeraldas coast in the mid-sixteenth century (Cabello de Balboa 1945: 18—22; Phelan 1967). Other Chachi testimony from 1597 indicates (translation from Monroy 1938: 327): ". . . the Cayapa chief formerly lived downriver at a settlement called Campi, but because of the mulatos from Esmeraldas and Indians called Soncon and Ceronda, he retreated (upstream) to where he is now found." In a contemporary map, recently published by Palop Martinez and rendered in schematic form in Figure 11.1, Cambi (sic) is placed at or near what would appear to be the confluence of the Santiago and Cayapas Rivers while the Ceronda are situated nearby. Unfortunately the Ceronda disappear from subsequent ethnonymy. The location of Campi and of the sixteenth-century Chachi favored by Palop Martinez will be evaluated later. The first detailed account of the Chachi dates to 1597. In that year, the Mercedarian friar Gaspar de Torres, in the company of the cacique of Lita and an entourage of thirty cargadores, descended the western slopes to the "country of the Cayapas." As in the case of Cariapa, Cayapa is the name of a principal leader, and according to Ortiz's etymology can be glossed as "little father" (Costales and Costales 1983: 77). It is by the term Cayapa that the Chachi are still known to most ethnographers. The following comments and Table 11.2 are abstracted from the documents presented by Monroy (1938: 314—364) concerning Gaspar de Torres's visit to the Chachi. The journey from Lita to the first Chachi settlement took four days. The trail passed over rough terrain with many snakes and an occasional maize chacra. On July 11, the party reached the house of Don Natinquilla and was well received with maize cakes, plantains, and fish. Chontaduro palm fruits and avocados, among other foods, were also reported. Natinquilla was identified as the leader of the "asiento de Cunaha" (Cundoha in the map of Figure 11.1?) which was situated on the upper Tumbibucho River and which, according to baptismal counts taken at a later date (Table 11.2), numbered minimally ninety-eight individuals. Four leagues downriver from Cunaha (better treated as a day's travel time than as an absolute distance), Gaspar de Torres reached the house of Don Francisco Cayapa at a site named Singobucho where a suspension bridge crossed the river. Don Francisco was the major cacique and his house was said to be located at the center of Chachi country. From Singobucho, one could look up and downriver and see the scattered houses of the Chachi. At Singobucho, a church
Table 11.2. Baptismal counts for the Chachi in 1597 (Monroy 1938: 334—348) Date (1597)
Leader of "aillo"
August 20
Don Francisco Cayapa and his son Don Felipe
August 30
Sept. 5 Sept. 8 Sept. 10 Sept. 12 Sept. 14 Sept. 16 Sept. 18 Sept. 20 Sept. 22 Sept. 24
Total
Adult males
n
Don Gaspar Unaatapa (brother of Don Felipe) Don Diego Natinquilla Don Pedro Chilmiso Don Juan Pifique (father-in-law of Cariapa)
Boys
2-3
20
23
2
20
26
7
2-9 n
Adult females
34 33
Girls J
Total
Sex ratio
9
85
1.2
6
79
2.0
80
2.1
87
2.1
2-3
2-5
7 5
12
14
1
60
8
3-7
2
30
2.0
19
2-5
12
8
34
2-3
21
8
86
1.8
37
18
2-5
18
98
i-7
2
1
0
1
4
1.0
4*
2-5
26
14
107
i-7
11
4
8
1
2-4
12
1
2
0
15
3.8 14.0
174
203
82
755
i-9
z9 6
History and archaeology of the Chachi (Ecuador)
253
Juan Pifique Francisco Cayapa = Cariapa?
Don Felipe
/\
^L
Gaspar Unaatapa
Figure 11.2. Possible relationships among Chachi caciques (solid triangles). was built and Don Francisco erected a new and larger house. The settlement was christened Espiritu Santo on July 22. Here Gaspar de Torres remained until September 25, baptizing and preaching in Spanish as Quechua was not understood (Spanish missionaries regularly used Quechua as a lingua franca). The baptismal records from Espiritu Santo contain interesting information (Table 11.2). The Chachi of the time are divided into a number of groups (the imported Quechua term "aillo" is used by Gaspar de Torres). In addition to the group headed by Don Francisco Cayapa, four other groups are identified after their leaders: (1) Natinquilla, mentioned previously and located upriver from Espiritu Santo; (2) Gaspar Unaatapa; (3) Pedro Chilmiso from the settlement of Democsilli on the Bilpi River; and (4) Juan Pifique on the Tubolompi River. The Bilpi is shown on contemporary maps (Figure 11.1), but neither the Bilpi nor Tubolompi survive as later place names, and neither can be readily equated with any present-day river. It is significant, however, that pi is the Chachi suffix meaning "river." The baptismal record suggests other matters of note. First, Anapa — the most common surname among today's Chachi — is already attested in 1597. Furthermore, many of the surnames listed by Gaspar de Torres are clearly Chachi kinship terms, a matter that deserves additional study. Second, the sex ratios given in Table 11.2 are improbably biased toward males. Only for August 20 does the relative number of males and females appear demographically realistic. Significantly it is on this date that Don Francisco Cayapa and those Chachi resident at, or in the nearby vicinity of, Espiritu Santo were baptized. It is likely that on later occasions Chachi coming to Espiritu Santo from more distant areas left women, and particularly daughters, at home. Clearly the 755 individuals baptized at Espiritu Santo do not represent the total Chachi population of 1597. One final point of interest is that the subtotal of adult males baptized between August 20 and September 12, all possibly pertaining to the "aillo" of Don Francisco Cayapa, is close to the 150 "Indios de guerra" (warriors) claimed by the cacique Cariapa in his previously mentioned visit to Lita. It is possible, therefore, that Cariapa and Cayapa were one and the same individual. If this were so, then three of the five Chachi caciques listed for 1597 would be related as shown in Figure 11.2. Among
254
WARREN R. DEBOER
today's Chachi, the position of gobernador (uni in Chachi) is ideally inherited from father to son or, if a son is lacking, to a son-in-law. In contrast to the 1597 case, however, fathers, sons, or son-in-laws do not serve simultaneously as gobernadores of separate centers. Where is the Chachi country visited by Gaspar de Torres? This question raises a debate in which interpretations of the chronicles diverge more markedly than the multiple versions of Chachi oral tradition arrayed in Table 11.1. The debate stems from arguments advanced by Moreno Navarro (1979) and Palop Martinez (1986) that the Chachi have been resident in the Cayapas Basin, their current homeland, throughout the historic period and probably well back into prehistory. This in situ hypothesis is argued forcefully by Palop on the basis of her reading of sixteenth-century texts and cartography, the latter summarized in Figure 11.1. Palop's position rests on two major pieces of evidence. First she notes that the Chachi of 1597 asserted that their former homeland centered at Campi well downstream from the area visited by Gaspar de Torres. As a singular testimony (one perhaps cognate with I-B of Table 11.1?), I would tend to regard this claim as a politically astute creation on the part of the Chachi, one designed to pander to the explicit Spanish goal of constructing a road to the coast and perhaps to enlist Spanish assistance against their mulatto enemies. Palop's second piece of evidence is a contemporary map plotting the itinerary of Gaspar de Torres's 1597 visit. As Palop notes, except for the Mira and Lita Rivers, none of the names on this map can be securely equated with modern designations, making a geographical reading very difficult. As shown in Figure 11.1, Palop argues that the settlements of Cayapa, Espiritu Santo, the Pueblo Viejo of Chachi tradition, and the contemporary ceremonial center of Punta Venado are all one and the same, an identification that indeed would place the sixteenth-century Chachi on the Cayapas River. This identification, however, is implausible for several reasons. First recall that Espiritu Santo was located at Singobucho, the site of a suspension bridge spanning the river below. One could hardly imagine a less likely place for a suspension bridge than Punta Venado on the lower Cayapas. The river here is over a hundred meters wide, sluggish, and flanked by low clay banks. In contrast, the upper Santiago, through much of its course, flows through narrow, steep walled gorges. Today's Chachi still talk about the suspension bridge that used to cross the Santiago at Pueblo Viejo. Second Palop errs by misplacing the Lachas River on her modern map (Figure 11.1, lower left). She incorrectly shows the Lachas as a tributary of the Cayapas while, in fact, it is a tributary of the upper Santiago. This distinction is critical with respect to the location of Espiritu Santo. On September 26,1597, Gaspar de Torres left Espiritu Santo for the country of the Aguatene located northward toward the Mira River. That night, after passing through rough and uninhabited terrain, the party camped at the Lachapi or Lacha River (Monroy 1938: 329). If we accept the identity of the Lachapi and the modern Lachas, this would place
History and archaeology of the Chachi (Ecuador)
255
Espiritu Santo on the upper Santiago. Finally Palop's argument, by relying exclusively on sixteenth-century documents, fails to address abundant later testimony that continues to place Espiritu Santo and Pueblo Viejo on the upper Santiago while identifying Punta Venado as a new settlement founded around 1800. Thus both the eighteenth-century map of Maldonado (Rumazo 1948-50:1) and the early twentieth-century map of Barrett (1925) situate "Cayapas" or "Pueblo Viejo" on the upper Santiago, precisely the area indicated in Chachi oral traditions. It is highly likely that the Espiritu Santo of Gaspar de Torres and the Pueblo Viejo of Chachi tradition are one and the same settlement. Further documents shed some light on tracking the Chachi from the upper Santiago to their present home on the Cayapas. Seventeenth-century sources for the Chachi are scant, in large part because of native rebellions which closed Esmeraldas to the Spaniards. In 1607, as part of a plan to open a route from Ibarra to the coast, Captain Cristobal de Troya entered the mouth of the "rio de Cayapas" without, however, furnishing any specific information about its inhabitants (Rumazo 1948-50: IV, 55). Four years later, Diego de Ugarte founded the ill-fated port of San Ignacio de Montesclaros which, according to Phelan (1967: 7), was located at the confluence of the Santiago and Cachabi Rivers. Between 1616 and 1619, Pablo Durango del Gadillo opened a trail from Ibarra to San Ignacio, but all of these efforts came to an end in the socalled Malaba rebellion of 1619. According to Moreno's brief account (1987), the Malabas, tired of cutting timber, transporting salt, and maintaining bridges and roads for the Spaniards, attacked all settlements from the Lachas to San Ignacio. The latter settlement was razed and more than thirty Spaniards, mulattos, and negroes were killed. I have no information on any role that the Chachi may have had during this period of turmoil. Years later (1687), when Spanish efforts to build a road to the coast shifted southward to the Esmeraldas-Guayllabamba drainage, a scheme was announced to recruit Chachi laborers, but the outcome of this plan is left unclear (Rumazo 1948-50: IV, 348). A question that is opened by these brief glimpses of the seventeenth century is the identity of the Malabas, a group equated with the Indios Bravos of Chachi oral tradition (Table 11.1, VI-I). In his travels of 1597, Gaspar de Torres encountered a small group of Malabas but never visited their homeland on the Mira River. In 1611, the Malabas are again placed on the Mira (Alcina Franch and Peiia 1976: 76). In his map of the mid-eighteenth century, Maldonado (Rumazo 1948-50: I) placed the following caption on the headwaters of the Mataje, between the Mira and Santiago drainages: "Por aqui vive la Nacion de los Malaguas que se rebelo antiguamente" (Here live the Malaguas who formerly rebelled). The last reference to the Malabas, before they drop out of history, dates to 1809 when they unexpectedly appeared on the San Miguel River, an upstream affluent of the Cayapas (Figure 11.1). By this time, as shall be seen below, the Chachi were well established on the Cayapas proper. According to Stevenson's (1825: II, 410-
Table 11.3. A Chachi chronicle. Date
Source
Comments
May 1736
Rumazo 1948-50: I, 119-123
1744?
Rumazo 1948-50: II, 128, 142
1749
Ruamzo 1948-50: II, 332
1803
Rumazo 1948-50: VIII, 87
1809
Stevenson 1825: II, 410-418
1852
Onffroy de Thoron 1983: 167
1850s
Villavicencio 1984: 341
early 1870s
Wolf 1879: 13-15
December 1891
Basurco 1894
Don Geronimo Uiiapa, governor of the Pueblo de Cayapas, protests to Quito that "Espanoles, meztizos, negros, mulatos, y indios forasteros" are raiding Chachi fields - these predators are based on the lower Santiago. El Pueblo de Cayapas, "situado en las partes altas del Rio de Santiago," is composed of sixty families with fifty "habiles," ages 18-50, who pay tribute; Chachi visit Quito to sell pita. Don Geronimo Uiiapa (see entry for 1736) still governor of Pueblo del Espiritu Santo de Cayapas. Ciudad de Cayapas said to be a short distance up the Cayapas River from its confluence with the Santiago. Punta Venado called the "new village of the Cayapas;" Malabas visited on the Rio San Miguel. "Abandoned" ceremonial center of San Miguel on the upper Cayapas located on an affluent of the Rio Cayapas. Major Chachi center is San Miguel on the upper Cayapas; lower Cayapas previously vacated because of the depredations by "blancos" and "negros." Chachi extend from the mouth of the Onzole to the Barbudo; ceremonial centers of Punta Venado (twenty houses) and San Miguel (six houses). Mentions ceremonial centers of Espiritu Santo (Punta Venado) and San Miguel; on the upper Zapallo Grande, there is a trail leading to Imbabura.
History and archaeology of the Chachi (Ecuador)
257
418) chatty account, relations between the two groups were hostile. The Malabas employed "lookouts" to guard against Chachi incursions, spoke a language said to resemble Quechua, and were governed by an alcalde named Cushicagua. Their dress and physical appearance were said to resemble those of the Colorado, whom Stevenson was to visit on a later occasion. On the other hand, Cushicagua used the stocks to punish wrong-doers, a Spanish-derived feature shared with the Chachi. It is difficult to evaluate this mixed bag of observations. However, it is tempting to equate these Malabas with the Indios Bravos of Chachi tradition and, furthermore, to view the group visited by Stevenson as an enclave that had survived the Chachi takeover of the Cayapas. But this surmise jumps ahead of the story. The track of the Chachi picks up during the eighteenth and nineteenth centuries. Some of the more useful information is summarized in Table 11.3. Of particular interest is that until the mid-eighteenth century, the Chachi, under their governor Don Geronimo Unapa, were still centered at Espiritu Santo far up the Santiago River. By the beginning of the nineteenth century, however, Chachi distribution appears to change in dramatic ways. Thus, in 1803 "Ciudad de Cayapas" is located a short distance up the Cayapas River. Stevenson's visit of 1809 leaves little doubt that "Ciudad de Cayapas" is Punta Venado, the major Chachi ceremonial center of Barrett's time (1925, [1909]) as well as of today. It is significant that Stevenson referred to Punta Venado as the "new village of Cayapas," suggesting that this center was recently established and stood in opposition to an "old village," namely Espiritu Santo or Pueblo Viejo on the upper Santiago. Thus between 1749 and 1803, the Chachi began their exodus from Pueblo Viejo to their present homeland on the Cayapas. The geography of this expansion from the upper Santiago to the Cayapas warrants comment. It should be noted that the upper Santiago is a white-water river unsuitable to canoe navigation. As Wolf (1879: 16) observed for the Chachi still living on the upper Santiago in the late nineteenth century: "they communicate with the rest of their tribe by land and by the Zapallo Grande and Barbudo Rivers; they never descend the Santiago" (my translation). Wolf's observation is fully confirmed by the Maldonado map of the mid-eighteenth century where the interfluves separating the upper Santiago from the headwaters of the Zapallo Grande and Barbudo are labeled "embarcadero de Cayapas" (portage of the Chachi). It should not escape our attention that in one Chachi oral tradition the invasion of the Cayapas begins precisely on the Zapallo Grande. The archaeological record Evidence pertinent to Chachi history has been acquired during the ongoing project of archaeological survey and excavation carried out in the Cayapas and Santiago Basins (Tolstoy and DeBoer 1989). The Chachi occupation of the Cayapas is but the most recent episode in a cultural sequence that can now be
258
WARREN R. DEBOER
traced back to at least the first millennium BC. This occupation can be discussed in terms of several issues: (1) the archaeological identity of the Chachi; (2) the distribution and form of Chachi sites; (3) the age of such sites; and perhaps most germane to the present paper, (4) the question of Chachi origins. Chachi sites can be recognized with some confidence. They all produce a distinctive ceramic ware that matches with great specificity the pottery described by Barrett (1925: 173-181). In Figure 11.3, archaeological ceramics from Chachi sites are arrayed against the Barrett collections housed in the Museum of the American Indian in New York City. The two sets are nearly identical. They share: (1) simple flat-bottomed plates, occasionally with nicked lips, that served as food bowls (see Figure 11.3 for the Chachi terms); (2) carinated vessels, often with a band of shell stamping on the carination, used as drinking bowls; (3) variously sized jars, often with nicked or incised lips, serving as all-purpose containers and cooking pots; and (4) large cooking and brewing urns, occasionally adorned with an applied clay strip at the neck—body juncture. This distinctive pottery cannot be confused with any other ceramic complex known from the Santiago—Cayapas and acts as an effective signature for the Chachi. This signature is tentative in only one respect. At four sites, Chachi pottery is mixed with a different ceramic ware not recorded in Barrett's descriptions or collections. This ware, called Cantarana after the site where it was first identified in 1989, consists of a poorly finished, thick-walled, and coarsely tempered (rock inclusions reach 7 mm in diameter!) pottery that in its technical and aesthetic dreadfulness stands completely apart from the generally well-made and gracile vessels of the Chachi. At only three sites is Cantarana material found without Chachi admixture. The distribution of sites yielding Cantarana, mixed Cantarana-Chachi, and Chachi ceramics is plotted in Figure 11.4. Of twenty-one unmixed Chachi sites, all but two are in the Cayapas Basin and these two (Ri and R45 in Figure 11.4) are found on the lower Santiago. In contrast, the three unmixed Cantarana sites are all on the Santiago, one (R42) well upstream. The four mixed Cantarana-Chachi sites all occur on the Cayapas, and three of these are situated on the lower Cayapas below the mouth of the Onzole. Overall, the Cantarana material would appear to be more Santiago-based than the Chachi material with which it is occasionally associated on the Cayapas. This differential distribution accords well with the historic and recent distributions of AfroEcuadorian and Chachi populations, the former dominating the Santiago, the latter, of course, based on the Cayapas. In this regard, it is interesting to recall West's (1957: 183) remark that "old Negroes have told me that their grandmothers made clay pots in the days of slavery." It is possible, therefore, that the Cantarana pottery is of Afro-Ecuadorian manufacture, although this suggestion remains only an informed guess without additional information. The Chachi settlement pattern is as distinctive as Chachi ceramics. Except for Punta Venado (C52 in Figure 11.4), where limited excavations indicate that a deep sherd- and shell-laden midden underlies the contemporary ceremonial
History and archaeology of the Chachi (Ecuador)
Plate _ (To'palato)
Drinking bowl (PTa'ma)
Jar (Kanda'ro)
TTT
Brewing urn (Bu'lum BTama)
0 |
10 L
20 cm I
Figure 11.3. Chachi ceramics collected by Barrett in 1909 (left) and ceramics from Chachi archaeological sites (right).
259
z6o
WARREN R. DEBOER
^^ 4 "
.'•A- PACIFIC &<;.:&
$'&/:.;..... •.-*£u^y
WiWw
\
is*
J > if
}
BORBONDhJ^^
/!« R1
In \ #C74 «(Sky/ » \s
ct
\ 4\
" C 4, '^6/-
S
(fc=rfb)C77 C78 XySTT 75
9o
/\}
C64^
\
-l^Pa /y o Grande
Onzo/e
• X ) OC124 A C72
^a
^ ^
c
KM 10
i ^
Figure 11.4. The Santiago and Cayapas Basins showing the distribution of Chachi sites (solid circles), Cantarana sites (open circles), and mixed Chachi-Cantarana sites (solid with open circles).
History and archaeology of the Chachi (Ecuador)
261
center (DeBoer and Blitz 1991: 58-59), all Chachi middens are relatively shallow and of limited extent. Most cover less than 0.025 ha and suggest a pattern of single-house settlements dispersed along rivers, exactly the residential pattern attested to by Stevenson (1825: II, 409), Wolf (1879: 52), Basurco (1894), Barrett (1925), and all more recent investigators. Most Chachi archaeological sites are marked by shell middens. Barrett mentioned these middens as characteristic domestic refuse, and in the last century Wolf (1879: 54) noted: At certain times they descend to the seacoasts in order to get provisions of fish, oysters, clams, and other shellfish; I was often surprised by the large shell mounds found at the sides of their houses, far up the Cayapas and so distant from the sea (my translation). In summary, the Chachi presence is highly diagnostic in terms of both ceramics and settlement type. Determining the age of Chachi archaeological sites is, at present, difficult, but a few relevant data are available. Pottery is no longer produced by the Chachi. The ceramic art was effectively extinct by at least 1959 (Altschuler 1964:19). All of our Chachi archaeological sites, therefore, precede the 1950s when motor boats, Protestant missionaries, and aluminum pots were introduced. In a few cases, Chachi informants identified sites that they remember being occupied during the 1930s and 1940s, but, as has been pointed out elsewhere (DeBoer 1989), such testimony is only of limited use as Chachi sites are often palimpsests produced by multiple reoccupations spanning several generations (see Siegel, this volume). Several Chachi sites produce small samples of imported European ceramics, glass, and metal implements such as scissors and knife blades. If we use Punta Venado as a gauge, and if we trust the historical data that this ceremonial center was founded at the start of the nineteenth century, then there is at present no definitive artifactual evidence that any of our Chachi sites predate 1800. The only radiocarbon date for Punta Venado, run on a buried guayacan post, is "statistically modern" (Beta-20638). A second date on charcoal from a shallow midden at C37 is 150 BP±7
262
WARREN R. DEBOER
December of 1986, for instance, I observed a Tumbaviro pedestal base being reemployed as a candelabrum by Chachi at the ceremonial center of San Miguel. Conclusion In the preceding review of varying kinds of evidence concerning Chachi origins, one might be tempted to discount the historicity of oral traditions phrased in an idiom of pawing jaguars, magical wands, and singing paj-ala birds. As arrayed in Table 11.1, these Chachi traditions do resemble the gap-laden matrix of the kind that Levi-Strauss (1967:210) has used to reveal the structure of myth. In this sense, Chachi oral tradition might indeed belong to the realm of "historia mitica" as so claimed by Carrasco (1983: 140). Yet Western history is not immune to a similar critique, with the mythical version favored by Palop Martinez differing radically from my own reading of the same texts. At this juncture, the independent testimony of archaeology plays an important role and, as has been seen, this testimony tends to favor the Chachi view of their own history. They indeed are relative newcomers to the Cayapas, displacing the Indios Bravos. Whether the Indios Bravos and the Malabas are one and the same remains unresolved, and whether either or both can be equated with the archaeologically defined Tumbaviro complex remains an important research issue. Finally, to further corroborate the Chachi account, the settlement of Pueblo Viejo should be archaeologically identified. In November of 1986, Pedro Tapuyo, governor of Zapallo Grande, asked the visiting archaeologists to finance a return expedition to Pueblo Viejo where the "Saints are stronger." Such a return was beyond our immediate budget and agenda, but when it takes place, the Chachi version of their own history will be fully confirmed.
References
Absy, Maria Lucia. 1979. A Palynological Study of Holocene Sediments in the Amazon Basin. University of Amsterdam, Amsterdam. 1982. Quaternary Palynological Studies in the Amazon Basin. In Biological Diversification in the Tropics, edited by G. T. Prance, pp. 67-92. Columbia University Press, New York. Adams, R. E. W., W. E. Brown, Jr., and T. P. Culbert. 1981. Radar Mapping, Archaeology, and Ancient Maya Land Use. Science, 213: 1457-1463. Alchon, S.A. 1991. Native Society and Disease in Colonial Ecuador. Cambridge University Press, Cambridge. Alcina Franch, Jose and Remedios de la Pefia. 1976. Textos para la Etnohistoria de Esmeraldas. Trabajos Preparatories 4. Proyecto "Arqueologia de Esmeraldas," Madrid. Alegria, Ricardo E. 1965. On Puerto Rican Archaeology. American Antiquity, 31: 246-249. Altschul, Jeffrey H. and Christopher L. Nagle. 1988. Collecting New Data for the Purpose of Model Development. In Quantifying the Present and Predicting the Past: Theory, Method, and Application of Archaeological Predictive Modeling, edited by W. J. Judge and L. Sebastian, pp. 257-299. United States Department of the Interior, Bureau of Land Management, Denver. Altschuler, Milton. 1964. The Cayapa: A Study in Legal Behavior. Unpublished Ph.D. dissertation, Department of Anthropology, University of Minnesota, Minneapolis. Ambrose, S. H. 1990. Preparation and Characterization of Bone and Tooth Collagen for Isotopic Analysis. Journal of Archaeological Science, 17: 431—451. 1993. Isotopic Analysis: Methodological and Interpretive Considerations. In Investigations of Ancient Human Tissues: Chemical Analysis in Anthropology, edited by M. K. Sandford, pp. 59—130. Gordon and Breach, Langhorne, Pennsylvania. Ambrose, S. H. and M. J. DeNiro. 1986a. Isotopic Ecology of East African Mammals. Oecologia, 69: 395-4O6. 1986b. Reconstruction of African Human Diet using Bone Collagen Carbon and Nitrogen Isotope Ratios. Nature, 319: 321—324. Ambrose, S. H. and L. Norr. 1992. On Prehistoric Subsistence in the Soconusco Region. Current Anthropology, 33: 401-404. 1993. Experimental Evidence for the Relationship of the Carbon Isotope Ratios of Whole Diet and Dietary Protein to Those of Bone Collagen and Carbonate. In Molecular Archaeology of Prehistoric Human Bone, edited by J. Lambert and G. Grupe. Springer, Berlin, in press. Ambrose, S. H. and N. E. Sikes. 1991. Soil Carbon Isotope Evidence for Holocene Habitat Change in the Kenya Rift Valley. Science, 253: 1402-1405. Andrews, Peter. 1990. Owls, Caves and Fossils. University of Chicago Press, Chicago. Andrews, Peter and E. M. N. Evans. 1983. Small Bone Accumulations Produced by Mammalian Carnivores. Paleobiology, 9: 289-307.
264
References
Angel, J. L. 1966. Porotic Hyperostosis, Anemias, Malarias, and Marshes in the Prehistoric Eastern Mediterranean. Science, 153: 760-763. Ardila Calderdon, Gerardo. 1983. Proyecto Carbonifero de el Cerrejon 'Zona Norte'. Arqueologia de Rescate: Area de Palmar. EPAM, Bogota. 1984. Proyecto Carbonifero de el Cerrejon 'Zona Central'. Arqueologia de Rescate: Areas de Patilla y El Paredon. EPAM, Bogota. 1986. Arqueologia de la Guajira. Boletin de Arqueologia, 1(3): 64-74. Bogota. 1990. Acercamiento de la Historia Prehispanica de la Guajira. In La Guajira, edited by G. Ardila C., pp. 59-80. Fondo FEN Colombia, Universidad Nacional de Colombia, Bogota. Armillas, Pedro. 1971. Gardens on Swamps. Science, 174: 653-661. Arnold, Dean and Kenneth A. Prettol. 1989. Aboriginal Earthworks near the Mouth of the Beni, Bolivia. Journal of Field Archaeology, 15: 457-465. Aufderheide, A. C , L. L. Tieszen, M. J. Allison, J. Wallgren, and G. Rapp Jr. 1988. Chemical Reconstruction of Components in Complex Diets: A Pilot Study. In Diet and Subsistence: Current Archaeological Perspectives, edited by B. E. Kennedy and G. M. LeMoine, pp. 301306. Archaeology Association, University of Calgary, Calgary. August, Peter V. 1983. The Role of Habitat Complexity and Heterogeneity in Structuring Tropical Mammal Communities. Ecology, 64: 1495-1507. Baksh, Michael. 1991. Applications of a Hand-Held GPS Receiver in South American Rain Forests. In Applications of Space Age Technology in Anthropology, edited by C. A. Behrens and T. L. Sever, pp. 227-236. National Aeronautics and Space Administration, John C. Stennis Space Center, Mississippi. Barker, Graeme. 1975. To Sieve or not to Sieve. Antiquity, 49: 61-63. Barrett, S. A. 1925. The Cayapa Indians of Ecuador. 2 vols. Museum of the American Indian Notes and Monographs No. 40, Parts 1 and 2. Heye Foundation, New York. Bartlett, A. S. and E. S. Barghoorn. 1973. Phytogeographic History of the Isthmus of Panama During the Past 12,000 years (A History of Vegetation, Climate, and Sea-level Change). In Vegetation and Vegetational History of Northern Latin America, edited by A. Graham, pp. 203-299. Elsevier, Amsterdam. Basurco, M. Santiago. 1894. Trois Semaines chez les Indiens Cayapas. Tour du Monde, 67: 401—416.
Becker-Donner, Etta. 1956. Archaologische Funde am Mittleren Guapore (Brazilien). Archiv Fur Volkerkunde, 11: 202-249. Beckerman, Stephen. 1979. The Abundance of Protein in Amazonia: A Reply to Gross. American Anthropologist, 81: 533-560. 1980. Fishing and Hunting by the Bari of Colombia. Working Papers on South American Indians, 2: 67-109. Beckerman, Stephen and Tom Sussenbach. 1983. A Qualitative Assessment of Dietary Contribution of Game Species to the Subsistence of South American Tropical Forest Tribal Peoples. In Animals and Archaeology, I: Hunters and Their Prey, edited by Juliet Clutton-Brock and Caroline Grigson, pp. 337-350. British Archaeological Reports, International Series 163, Oxford. Behrens, Clifford A. 1991. Applications of Satellite Image Processing to the Analysis of Amazonian Cultural Ecology. In Applications of Space Age Technology in Anthropology, edited by C. A. Behrens and T. L. Sever, pp. 9-33. National Aeronautics and Space Administration, John C. Stennis Space Center, Mississippi. Behrens, Clifford A., and Thomas L. Sever (eds.). 1991. Applications of Space Age Technology in Anthropology. National Aeronautics and Space Administration, John C. Stennis Space Center, Mississippi. Behrensmeyer, Anna K. 1975. Paleoecology of Plio-Pleistocene Vertebrate Assemblages East of Lake Rudolf, Kenya. Bulletin of the Museum of Comparative Zoology, 146: 473-578.
References
265
1978. Taphonomic and Ecologic Information from Bone Weathering. Paleobiology, 4: 150-162. Beinroth, Friedrich H. 1969. An Outline of the Geology of Puerto Rico. Bulletin 213. Agricultural Experiment Station, University of Puerto Rico, Mayagiiez Campus. Bender, M. M. 1968. Mass Spectrometric Studies of Carbon-13 in Corn and Other Grasses. Radiocarbon, 10: 468—472. Bennett, Charles F. 1962. The Bayuno Cuna Indians, Panama: An Ecological Study of Livelihood and Diet. Annals of the Association of American Geographers, 52: 32-50. Bennett, Wendell C. 1936. Excavations in Bolivia. Anthropological Papers of the American Museum of Natural History, 35(4): 329—507. Benz, B. F., and H. H. Iltis. 1990. Studies in Archaeological Maize I. The "Wild" Maize from San Marcos Cave Reexamined. American Antiquity, 55: 500-511. Bergman, Roland W. 1980. Amazon Economics. The Simplicity of Shipibo Indian Wealth. Dellplain Latin American Studies, 6. University Microfilms, Ann Arbor. Bianchi, Cesar. 1988. El Shuar y el Ambiente. Conocimiento del Medio y Caceria no Destructiva. Abya-Yala, Quito. Bibby, B. G. 1943. The Relationship Between Microscopic Hypoplasia (Mellanby) and Dental Caries. Journal of Dental Research, 22: 218. Binford, Lewis R. 1964. A Consideration of Archaeological Research Design. American Antiquity, 29: 42-5-4411981. Bones. Ancient Men and Modern Myths. Academic Press, New York. Binford, Lewis R., Sally R. Binford, Robert Whallon, and Margaret Ann Hardin. 1970. Archaeology at Hatchery West. Society for American Archaeology Memoirs No. 24, Washington D.C. Binford, Michael W., Mark Brenner, Thomas J. Whitmore, Antonia Higuera Gundy, E. S. Deevey, and Barbara Leyden. 1987. Ecosystems, Paleoecology and Human Disturbance in Subtropical and Tropical America. Quaternary Science Review, 6: 115-128. Bird, J. B. and R. G. Cooke. 1978. La Cueva de los Ladrones, Datos Preliminares Sobre la Ocupacion Formativa. Adas del V Simposium Nacional de Antropologia, Arqueologia, y Etnohistoria, pp. 282-322. Universidad de Panama, Instituto Nacional de Cultura, Panama City. Bird, J. B., J. Hyslop, and M. D. Skinner. 1985. The Preceramic Excavations at the Huaca Prieta, Chicama Valley, Peru. American Museum of Natural History, New York. Birks, H. J. B. and A. D. Gordon. 1985. Numerical Methods in Quaternary Pollen Analysis. Academic Press, London. Bligh, E. G. and W. J. Dyer. 1959. A Rapid Method of Total Lipid Extraction and Purification. Canadian journal of Biochemistry, 37: 911-917. Boaz, N. T. and A. K. Behrensmeyer. 1976. Hominid Taphonomy: Transport of Human Skeletal Material in an Artificial Fluviatile Environment. American Journal of Physical Anthropology, 45: 53-60. Boomert, Arie. 1987. Gifts of the Amazon: "Green Stone" Pendants and Beads as Items of Ceremonial Exchange in Amazonia and the Caribbean. Antropologica, 67: 33-54. Bozarth, S. R. 1992. Classification of Opal Phytoliths in Selected Dicotyledons Native to the Great Plains. In Phytolith Systematics, edited by G. R. Rapp Jr. and S. C. Mulholland, pp. 193-214. Plenum Press, New York. Brain, C. K. 1981. The Hunters or the Hunted? An Introduction to African Cave Taphonomy. University of Chicago Press, Chicago. Bray, Warwick. 1984. Across the Darien Gap: A Colombian View of Isthmian Archaeology. In The Archaeology of Lower Central America, edited by F. W. Lange and D. Z. Stone, pp. 305-338. University of New Mexico Press, Albuquerque. Bray, Warwick, Leonor Herrera, and Cardale Schrimpff. 1988. Report on the 1984 Field Season in Calima. Pro Calima, 5: 2-42. Bray, W., L. Herrera, M. C. Schrimpff, P. Botero, and J. G. Monsalve. 1987. The Ancient
z66
References
Agricultural Landscape of Calima, Colombia. In Pre-Hispanic Agricultural Fields in the Andean Region, edited by W. M. Denevan, K. Mathewson, and G. Knapp, pp. 443—481. British Archaeological Reports, International Series 359 (ii), Oxford . Brochado, Jose P. 1984. An Ecological Model of the Spread of Pottery and Agriculture into Eastern South America. Unpublished Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana-Champaign. Brochado, Jose P. and Donald W. Lathrap. 1982. Amazonia. Ms. on file, Department of Anthropology, University of Illinois, Urbana-Champaign. Broida, M. 1984. An Estimation of the Percentage of Maize in the Diets of Two Kentucky Fort Ancient Villages. In Late Prehistoric Research in Kentucky, edited by D. L. Pollack, C. B. Hockensmith, and T. N. Sanders, pp. 68-82. Kentucky Heritage Council, Frankfort. Brown, James A. 1975. Deep-Site Excavation Strategy as a Sampling Problem. In Sampling in Archaeology, edited by James W. Mueller, pp. 155-169. The University of Arizona Press, Tucson. Bryant, V. M., Jr. 1993. Phytolith Research: A Look Toward the Future. In Current Research in Phytolith Analysis: Applications in Archaeology and Paleoecology, edited by D. M. Pearsall and D. R. Piperno, pp. 175-181. MASCA Research Papers in Science and Archaeology 10, University of Pennsylvania Museum, Philadelphia. Bryant, V. M., Jr. and R. G. Holloway. 1983. The Role of Palynology in Archaeology. In Advances in Archaeological Method and Theory, VI, edited by M. B. Schiffer, pp. 191-223. Academic Press, New York. Buchler, Ira R. and Henry A. Selby. 1968. Kinship and Social Organization: An Introduction to Theory and Method. MacMillan, New York. Buikstra, J. and G. R. Milner. 1991. Isotopic and Archaeological Interpretations of Diet in the Central Mississippi Valley. Journal of Archaeological Science, 18: 319-329. Buikstra, Jane, Susan Frankenberg, Joseph B. Lambert and Liang Xue. 1989. Multiple Elements: Multiple Expectations. In The Chemistry of Human Bone, edited by T. Douglas Price, pp. 155210. Cambridge University Press, Cambridge. Buikstra, J., W. Autry, E. Breitburg, L. Eisenberg, and N. J. van der Merwe. 1988. Diet and Health in the Nashville Basin: Human Adaptation and Maize Agriculture in Middle Tennessee. In Diet and Subsistence: Current Archaeological Perspectives, edited by B. E. Kennedy and G. M. LeMoine, pp. 243-259. Archaeology Association, University of Calgary, Calgary. Bullen, Ripley P. and Adelaide K. Bullen. 1974. Tests at Hacienda Grande, Puerto Rico. Boletin Informativa 1. Fundacion Arqueologica, Antropologica e Historica de Puerto Rico, Santurce. Burger, R. L. and N. van der Merwe. 1990. Maize and the Origin of Highland Chavin Civilization: An Isotopic Perspective. American Anthropologist, 49: 334-349. Bush, M. B. and P. A. Colinvaux. 1988. A 7000-year Pollen Record from the Amazon Lowlands. Ecuador. Vegetatio, 76: 141-154. Bush, M. B., D. R. Piperno, and P. A. Colinvaux. 1989. A 6,000 Year History of Amazonian Maize Cultivation. Nature, 340: 303-305. Bush, M. B., D. R. Piperno, P. A. Colinvaux, P. E. De Oliveira, L. A. Krissek, M. C. Miller, and W. E. Rowe. 1992. A 14,300-yr Paleoecological Profile of a Lowland Tropical Lake in Panama. Ecological Monographs, 62: 251-275. Bustos, Victor. 1976a. Investigaciones Arqueologicas en el Departamento del Beni. Proyecto: Excavacion de Loma Cachipere en Puerto Almacen. Documentos Internos del INAR, La Paz. 1976b. Proyecto de Reconocimiento Arqueologica en Trinidad. Documentos Internos del INAR, 23/76, La Paz. 1976c. Investigaciones Arqueologicas en las Tierras Bajas de Bolivia. Documentos Internos del INAR, La Paz. 1978a. La Arqueologia en los Llanos del Beni, Bolivia. Documentos Internos del INAR, 32/78, La Paz.
References
267
1978b. Proyecto de Excavaciones Arqueologicas en el Departamento del Beni, en Lomas de la Region de Cocharcas y Villa Banzar. Documentos Internos del INAR, 10/78, La Paz. 1978c. Una Hipotesis de Relaciones Culturales entre el Altiplano y la Vertiente Oriental de los Andes. Pumapunku, 12: 115—126, La Paz. i978d. Investigaciones Arqueologicas en Trinidad, Departamento del Beni. Instituto National de Arqueologia, 22, La Paz. Bustos S., Victor y Juan Faldin. 1978. Prospeccion Arqueologica entre las Localidades de San Ignacio y San Borja (Sitio 1023011) del Departamento del Beni. Documentos Internos del INAR, 14/78, La Paz. Buys, J. and V. Dominguez. 1988a. Un Cementerio de Hace 2000 Anos: Jardin del Este. In Quito Antes de Benalcazar, edited by I. C. Cevallos, pp. 31-50. Centro Cultural Artes, Quito. 1988b. Hace Dos Mil Anos en Cumbayd. Proyecto Arqueologico "Jardin del Este," Cumbaya, Provincia de Pichincha, Ecuador, Instituto Nacional de Patrimonio Cultural, Quito. 1990. El Proyecto Arqueologico "Jardin del Este." La Preservation y Promotion del Patrimonio Cultural del Ecuador, pp. 6-21. Cooperacion Tecnica Ecuatoriana, Belga 4. Instituto Nacional de Patrimonio Cultural, Quito. Cabello de Balboa, Miguel. 1945. Obras, I. Editorial Ecuatoriana, Quito. Cadavid, Gilberto and Luisa Fernanda Herrera de Turbay. 1985. Manifestaciones Culturales en el Area Tairona. Informes Antropologicas, 1: 5-54. Instituto Colombiano de Antropologia, Bogota. Canadas, Luis. 1983. El Mapa Bioclimdtico y Ecologico del Ecuador. Banco Central del Ecuador, Quito. Capone, D. G., and E. J. Carpenter. 1982. Nitrogen Fixation in the Marine Environment. Science, 217:
1140-1142.
Carneiro, Robert L. 1961. Slash-and-Burn Cultivation among the Kuikuru and its Implications for Cultural Development in the Amazon Basin. In The Evolution of Horticultural Systems in Native South America: Causes and Consequences, edited by Johannes Wilbert, pp. 47-67. Antropologica Supplement Publication No. 2. Sociedad de Ciencias Naturales la Salle, Caracas. 1970. Hunting and Hunting Magic Among the Amahuaca of the Peruvian Montana. Ethnology, 9- 33i-34i-
Carr, Christopher. 1982. Handbook on Soil Resistivity Surveying: Interpretation of Data from Earthen Archeological Sites. Center for American Archeology Press, Evanston, Illinois. Carrasco, Eulalia. 1983. El Pueblo Chachi. Abya-Yala, Quito. Casteel, Richard W. 1972. Some Biases in the Recovery of Archaeological Faunal Remains. Proceedings of the Royal Prehistoric Society, 38: 383-388. 1976. Comparison of Column and Whole Unit Samples for Recovering Fish Remains. World Archaeology, 8: 192-196. Cerling, T. E., J. Quade, S. H. Ambrose, and N. E. Sikes. 1991. Miocene Fossil Soils, Grasses and Carbon Isotopes from Fort Ternan, Kenya: Grassland or Woodland? Journal of Human Evolution, 21: 295-306. Ceron, Benhur. 1986. Los Awa-Kwaiker. Abya-Yala, Quito. Chace, Paul G. 1969. Biological Archaeology of Some Coastal Middens, Orange County, California. Pacific Coast Archaeological Society Quarterly, 5: 65-77. Chagnon, Napoleon A. 1991. GIS, GPS, Political History and Geo-demography of the Aramamisi Yanomamo Expansion. In Applications of Space Age Technology in Anthropology, edited by C. A. Behrens and T. L. Sever, pp. 35—62. National Aeronautics and Space Administration, John C. Stennis Space Center, Mississippi. Chagnon, Napoleon A. and Raymond B. Hames. 1979. Protein Deficiency and Tribal Warfare in Amazonia: New Data. Science, 203: 910—913.
268
References
Chanlatte Baik, Luis A. 1979. Excavaciones Arqueologicas en Vieques. Revista del Museo de Antropologia, Historia y Arte de la Universidad de Puerto Rico, 1(1): 55-59. Chisholm, B. S. 1986. Reconstruction of Prehistoric Diet in British Columbia using Stable Carbon Isotopic Analysis. Unpublished Ph.D. dissertation, Department of Archaeology, Simon Fraser University, British Columbia. 1989. Variation in Diet Reconstructions Based on Stable Isotopic Evidence. In The Chemistry of Prehistoric Human Bone, edited by T. D. Price, pp. 10—37. Cambridge University Press, Cambridge. Chisholm, B. S., D. E. Nelson, and H. P. Schwarcz. 1982. Stable-Carbon Isotope Ratios as a Measure of Marine Versus Terrestrial Protein in Ancient Diets. Science, 216: 1131-1132. Chomsky, Noam. 1966. Syntactic Structures. Mouton, The Hague. Clark, J. G. D. 1954. Excavations at Star Carr: An Early Mesolithic Site at Seamer, Near Scarborough, Yorkshire. Cambridge University Press, Cambridge. Clason, A. T., and W. Prummel. 1977. Collecting, Sieving and Archaeological Research. Journal of Archaeological Science, 4: 171-175. Cody, Annie. 1993. Distribution of Exotic Stone Artifacts through the Lesser Antilles: Their Implications for Prehistoric Interaction and Exchange. Proceedings of the International Association for Caribbean Archaeology, 14: 204-226. Cohen, M. N. and G. J. Armelagos (eds.). 1984. Paleopathology at the Origins of Agriculture. Academic Press, Orlando. Colinvaux, P. A., M. C. Miller, K.-b. Liu, M. Steinitz-Kannan, and I. Frost. 1985. Discovery of Permanent Amazon Lakes and Hydraulic Disturbance in the Upper Amazon Basin. Nature, 131: 42-45. Conant, Francis Paine. 1990.1990 and Beyond: Satellite Remote Sensing and Ecological Anthropology. In The Ecosystem Approach in Anthropology: From Concept to Practice, edited by Emilio F. Moran, pp. 357-388. University of Michigan Press, Ann Arbor. Conniff, Richard. 1991. RAP: On the Fast Track in Ecuador's Tropical Forests. Smithsonian, 22: 36-49. Conrad, Geoffrey W. 1981. Cultural Materialism, Split Inheritance, and the Expansion of Ancient Peruvian Empires. American Antiquity, 46: 3—28. Cooke, Richard G. 1984. Archaeological Research in Central and Eastern Panama: A Review of Some of the Problems. In The Archaeology of Lower Central America, edited by F. W. Lange and D. Z. Stone, pp. 263—302. School of American Research Advanced Seminar Series. University of New Mexico Press, Albuquerque. 1992. Prehistoric Nearshore and Littoral Fishing in the Eastern Tropical Pacific: An Ichthyological Evaluation, journal of World Prehistory, 6: 1-49. Cooke, Richard G. and Warwick Bray. 1985. The Goldwork of Panama: An Iconographic and Chronological Perspective. In The Art of Precolumbian Gold: The Jan Mitchell Collection, edited by J. Jones, pp. 34-45. Weidenfeld and Nicolson, London. Cooke, Richard G. and A. J. Ranere. 1984. The "Proyecto Santa Maria": A Multidisciplinary Analysis of Prehistoric Adaptations to a Tropical Watershed in Panama. In Recent Developments in Isthmian Archaeology: Advances in the Prehistory of Lower Central America, edited by F. W. Lange, pp. 3—30. International Series 212, British Archaeological Reports, Oxford. 1989a. Changing Human Responses to Lagoon-Estuary Systems in Prehistoric Central Panama. Paper presented at the Circum-Pacific Prehistory Conference, August 1-6, 1989, Seattle, Washington. 1989b. Hunting in Precolumbian Panama: A Diachronic Perspective. In The Walking Larder, edited by J. Clutton-Brock, pp. 295-315. Unwin Hyman, London. 1992a. The Origins of Wealth and Hierarchy in the Central Region of Panama: A Historical and Ecological Approach. In Wealth and Hierarchy in the Intermediate Area, edited by F. W. Lange, pp. 243-316. Dumbarton Oaks, Washington D.C.
References
269
1992b. Prehistoric Human Adaptations to the Seasonally Dry Forests of Panama. World Archaeology, 24(1): 114-133. 1993. The Relation of Fish Resources to the Location, Diet Breadth, and Procurement Technology of a Preceramic and a Ceramic Site in an Estuarine Embayment on the Pacific Coast of Panama. In Proceedings of the Circum-Pacific Prehistory Conference, edited by M. Blake, in press. Cordero Miranda, Gregorio. 1984. Reconocimiento Arqueologico en las Margines del Rio Beni. Arqueologia Boliviana, 1: 15-38, La Paz. Correal Urrego, Gonzalo and Maria Pinto Nolla. 1983. Investigacion Arqueologica en el Municipio de TLipacon, Cundinamarca. Banco de la Republica, Fundacion de Investigaciones Arqueologicas Nacionales, Bogota. Costales, Alfredo. 1956. Los Indios Colorados. Llacta, 1: 1-112. Costales, Piedad and Alfredo Costales. 1983. Amazonia. Mundo Shuar, Quito. Courty, Marie Agnes, Paul Goldberg, and Richard Macphail. 1989. Soils and Micromorphology in Archaeology. Cambridge University Press, Cambridge. Craig, H. 1957. Isotopic Standards for Carbon and Oxygen and Correction Factors for Mass Spectrometric Analysis of Carbon Dioxide. Geochimica et Cosmochimica Ada, 12: 133-149. Crane, C. J. 1986. Late Preclassic Maya Agriculture, Wild Plant Utilization, and Land-use Practices. In Archaeology at Cerros, Belize, Central America, I, edited by R. A. Robertson and D. A. Freidel, pp. 147-165. Southern Methodist University Press, Tulane. Crawford, G. W. 1983. Paleoethnobotany of the Kameda Peninsula Jomon. Museum of Anthropology, University of Michigan, Ann Arbor. Crocker, Jon C. 1985. Vital Souls: Bororo Cosmology, Natural Symbolism, and Shamanism. The University of Arizona Press, Tucson. Culbert, T. Patrick, Laura Levi, and Luis Cruz. 1990. Lowland Maya Wetland Agriculture: The Rio Azul Agronomy. In Vision and Revision in Maya Studies, edited by Flora Clancy and Peter Harrison, pp. 115-124. University of New Mexico Press, Albuquerque. Dahlin, Bruce and Jeffrey Pope. 1989. Ancient Maya Wetland Agriculture: New Insights from Ecological and Remote Sensing Research. Journal of Field Archaeology, 16: 87—106. Damp, Jonathan E. 1984. Architecture of the Early Valdivia Village. American Antiquity, 49: 573-585. Dansie, Amy J. 1984. Human and Carnivore Modification of Small Mammals in the Great Basin. First International Conference on Bone Modification, Carson City Nevada, Abstracts, 8-9. Darch, J. P. (ed.). 1983. Drained Fields of the Americas. British Archaeological Reports, International Series, 189, Oxford. DeBoer, Warren R. 1974. Ceramic Longevity and Archaeological Interpretation: an Example from the Upper Ucayali, Peru. American Antiquity, 39: 335-343. 1989. The House that Jesusito Built. Proceedings of the Annual Chacmool Conference, 21: 478489. Department of Archaeology, University of Calgary. DeBoer, Warren R. and John H. Blitz. 1991. Ceremonial Centers of the Chachi. Expedition, 33(1): 53-62. DeBoer, Warren R. and Donald W. Lathrap. 1979. The Making and Breaking of Shipibo—Conibo Ceramics. In Ethnoarchaeology: Implications of Ethnography for Archaeology, edited by C. Kramer, pp. 102-138. Columbia University Press, New York. DeBoer, Warren. R. and J. Scott Raymond. 1987. Roots Revisited: The Origins of the Shipibo Art Style. Journal of Latin American Lore, 13(1): 115-132. Deetz, James. 1967. Invitation to Archaeology. The Natural History Press, Garden City. 1988. History and Archaeological Theory: Walter Taylor Revisited. American Antiquity, 53: 12-23.
Deevey, E. S., Jr., M. S. Gross, G. E. Hutchinson, and H. I. Kraybill. 1954. The Natural C-14 Content of Materials from Hard Water Lakes. Proceedings of the National Academy of Sciences, 40: 285-288.
270
References
Delcourt, P. A., H. R. Delcourt, P. A. Cridlebaugh, and J. Chapman. 1986. Holocene Ethnobotanical and Paleoecological Record of Human Impact on Vegetation in the Little Tennessee River Valley, Tennessee. Quaternary Research, 25: 330-349. DeMarcay, Gary B. and D. Gentry Steele. 1986. The Value of Fine Screening on Inland Based Hunter-Gatherer Habitation Sites. In Archaeological Investigations at 41 LK zoi, Choke Canyon Reservoir, Southern Texas, edited by C. L. Highley, pp. 250-264. Center for Archaeological Research Choke Canyon Series 11. University of Texas, San Antonio. Denevan, William M. 1963. Additional Comments on the Earthworks of Mojos in Northeastern Bolivia. American Antiquity, 28: 540-545. 1966a. The Aboriginal Cultures of the Llanos de Mojos, Bolivia. University of California Press, Berkeley. 1966b. A Cultural-ecological View of Former Aboriginal Settlement in the Amazon Basin. The Professional Geographer, 18: 346-351. 1970. Aboriginal Drained-Field Cultivation in the Americas. Science, 169: 647—654. 1976. The Aboriginal Population of Amazonia. In The Native Population of the Americas in 1492, edited by William Denevan, pp. 205-234. University of Wisconsin Press, Madison. 1980. Tipologia de Configuraciones Agricolas Prehispanicas. America Indigena 40: 610-652. 1982. Hydraulic Agriculture in the American Tropics: Forms, Measures, and Recent Research. In Maya Subsistence, edited by Kent V. Flannery, pp. 181-203. Academic Press, New York. 1991. Prehistoric Roads and Causeways of Lowland Tropical America. In Ancient Road Networks and Settlement Hierarchies in the New World, edited by Charles D. Trombold, pp. 230—242. Cambridge University Press, Cambridge. 1992. The Pristine Myth: The Landscape of the Americas in 1492. Annals of the Association of American Geographers, 82: 369-385. Denevan, William M. and B. L. Turner II. 1974. Forms, Functions, and Associations of Raised Fields in the Old World Tropics. Journal of Tropical Geography, 39: 24-33. Denevan, William M., Kent Mathewson, and Gregory Knapp (eds.). 1987. Pre-Hispanic Agricultural Fields in the Andean Region. British Archaeological Reports, International Series 359(0 and (ii), Oxford. DeNiro, M. J. 1985. Postmortem Preservation and Alteration of In Vivo Bone Collagen Isotope Ratios in Relation to Paleodietary Reconstruction. Nature, 317: 806—809. 1987. Stable Isotopy and Archaeology. American Scientist, 75: 182-191. DeNiro, M. J. and S. Epstein. 1978. Influence of Diet on the Distribution of Carbon Isotopes in Animals. Geochimica et Cosmochimica Ada, 42: 495-506. 1981. Influence of Diet on the Distribution of Nitrogen Isotopes in Animals. Geochimica et Cosmochimica Ada, 45: 341-351. DeNiro, M. J. and C. Hastorf. 1985. Alteration of 15N/14N and 13C/12C Ratios of Plant Matter during the Initial Stages of Diagenesis: Studies Using Archaeological Specimens from Peru. Geochimica et Cosmochimica Ada, 49: 97-115. DeNiro, M. J. and M. J. Schoeninger. 1983. Stable Carbon and Nitrogen Isotope Ratios of Bone Collagen: Variations within Individuals, between Sexes, and within Populations Raised on Monotonous Diets. Journal of Archaeological Science, 10: 199—203. Dimbleby, G.W. 1985. The Palynology of Archaeological Sites. Academic Press, London. Dodson, Peter. 1973. The Significance of Small Bones in Paleoecological Interpretation. University of Wyoming Contributions to Geology, 12: 15—19. Dodson, Peter and Diane Wexlar. 1979. Taphonomic Investigation of Owl Pellets. Paleobiology 5: 275-284. Doebley, J. 1990. Molecular Evidence and the Evolution of Maize. Economic Botany, 44: 6-27. Doolittle, William. 1984. Agricultural Change as Incremental Process. Annals of the Association of American Geographers, 74: 124-137. Dougherty, Bernardo. 1985. Archaeological Research in Northeastern Beni, Bolivia. National Geographic Research Reports, 21: 129-136.
References
271
Dougherty, Bernardo and Horacio Calandra. 1981. Nota Preliminar Sobre Investigaciones Arqueologicas en los Llanos de Moxos, Departamento del Beni, Republica de Bolivia. Revista del Museo de la Plata, 8 (Seccion Antropologica no. 53): 87-106, La Plata. 1981-82. Excavaciones Arqueologicas en la Loma Alta de Casarabe, Llanos de Moxos, Departamento del Beni, Bolivia. Relaciones de la Sociedad Argentina de Antropologia, 14(2): 9-48, Buenos Aires. 1984. Prehispanic Human Settlement in the Llanos de Mojos, Bolivia. In Quaternary of South America and Antarctic Peninsula, II, edited by Jorge Rabassa, pp. 163-199. A. D. Balkema, Rotterdam. 1985. Ambiente y Arqueologia en el Oriente Boliviano. La Provincia Itenez del Departamento del Beni. Relaciones de la Sociedad Argentina de Antropologia, Buenos Aires. Doyon, L. G. 1988. Tumbas de la Nobleza en La Florida. In Quito Antes de Benalcazar, edited by I.C. Cevallos, pp. 51-66. Centro Cultural Artes, Quito. Duckworth, W. L. H. 1951. Human Skulls from Ecuador. In The Archaeology of the Santa Elena Peninsula in South-west Ecuador, by G. H. S. Bushnell, pp. 144-148. Cambridge University Press, Cambridge. Duke, J. 1968. Darien Ethnobotanical Dictionary. MS on file, Battele Memorial Institute, Columbus. Dumont, Jean-Paul. 1972. Under the Rainbow. Nature and Supernature Among the Panare Indians. University of Texas Press, Austin. Dunnell, Robert. 1971. Systematics in Prehistory. Free Press, Glencoe. Eden, Michael J., Warwick Bray, Leonor Herrera, and Colin McEwan. 1984. Terra Preta Soils and Their Archaeological Context in the Caqueta Basin of Southeast Colombia. American Antiquity, 49: 125-140. Eidt, Robert C. 1984. Advances in Abandoned Settlement Analysis: Applications of Prehistoric Anthrosols in Colombia, South America. The Center for Latin America, University of Wisconsin, Milwaukee. Eisenberg, John F. 1978. The Evolution of Arboreal Herbivores in the Class Mammalia. In The Ecology of Arboreal Folivores, edited by G. G. Montgomery, pp, 135—152. Smithsonian Institution Press, Washington D.C. 1980. The Density and Biomass of Tropical Mammals. In Conservation Biology. An Evolutionary Perspective, edited by Michael E. Soule and Bruce A. Wilcox, pp. 35-55. Sinauer Associates, Sunderland, Massachusetts. 1983. Behavioral Adaptations of Higher Vertebrates to Tropical Forests. In Tropical Rain Forest Ecosystems, edited by F. B. Golley, pp. 267—278. Ecosystems of the World, XIVA. Elsevier Scientific Publishing, Amsterdam. 1989. Mammals of the Neotropics, I. University of Chicago Press, Chicago. 1990. Neotropical Mammal Communities. In Four Neotropical Rainforests, edited by Alwyn H. Gentry, pp. 358-368. Yale University Press, New Haven. Eisenberg, John F. and Richard W. Thorington Jr. 1973. A Preliminary Analysis of a Neotropical Mammal Fauna. Biotropica, 5: 150—161. Eisenberg, John F., M.A. O'Connell, and Peter V. August. 1979. Density, Productivity, and Distribution of Mammals in Two Venezualan Habitats. In Vertebrate Ecology in the Northern Neotropics, edited by John F. Eisenberg, pp. 187-207. Smithsonian Institution Press, Washington D.C. El-Najjar, M. Y., D. J. Ryan, C. G. Turner II and B. Lozoff. 1976. The Etiology of Porotic Hyperostosis Among the Prehistoric and Historic Anasazi Indians of Southwestern United States. American Journal of Physical Anthropology, 44: 477—487. Ellen, Roy F. 1986. What Black Elk Left Unsaid. Anthropology Today, 2(6): 8-12. Emmons, Louise H. 1984. Geographic Variation in Densities and Diversities of Non-flying Mammals in Amazonia. Biotropica, 16: 210-222.
272
References
Emmons, Louise H. and Francois Feer. 1990. Neotropical Rainforest Mammals. University of Chicago Press, Chicago. Emslie, Steven D. 1988. Vertebrate Paleontology and Taphonomy of Caves in Grand Canyon, Arizona. National Geographic Research, 4: 128-142. Engwall, Evan C , Margaret van de Guchte, and Ari Zighelboim (eds.). 1992. Gifts to the Cayman: Essays in Honor of Donald W. Lathrap. Journal of the Steward Anthropological Society, 20 (1 and 2). EPAM Ltda. 1982. Estudio de Efecto Ambiental de la Explotacion Inicial de Carbon en la Zona Central del Proyecto (El Cerrejon. CARBOCOL, Bogota. Erickson, Clark L. 1980. Sistemas Agricolas Prehispanicos en los Llanos de Mojos. America Indigena, 40: 731-755. 1981. Nuevas Investigaciones en el Estudio de Campos Elevados: El Uso de LANDS AT. Paper presented in the Segunda Reunion Internacional de las Jornadas de Arqueologia, Anthropologia, y Historia Boliviano-Peruana, Copacabana, Bolivia. 1985. Applications of Prehistoric Andean Technology: Experiments in Raised Field Agriculture, Huatta, Lake Titicaca, Peru, 1981-1983. In Prehistoric Intensive Agriculture in the Tropics, edited by Ian Farrington, pp. 209—232. British Archaeological Reports, International Series 232, Oxford. 1986a. Agricultura en Camellones en la Cuenca del Lago Titicaca: Aspectos Tecnicos y su Futura. In Andenes y Camellones en el Peru Andino: Historia Presente y Futuro, edited by Carlos de la Torre and Manuel Burga, pp. 331—350. Consejo Nacional de Ciencia y Tecnologia, Lima. 1986b. Waru Waru: Una Tecnologia Agricola del Altiplano Prehistorico. In Andenes y Camellones en el Peru Andino: Historia Presente y Futuro, edited by Carlos de la Torre and Manuel Burga, pp. 59-84. Consejo Nacional de Ciencia y Tecnologia, Lima. 1987. The Dating of Raised Field Agriculture in the Lake Titicaca Basin of Peru. In Pre-Hispanic Agricultural Fields in the Andean Region, edited by W. M. Denevan, K. Mathewson, and G. Knapp, pp. 373-383. British Archaeological Reports, International Series 359, Oxford. 1988a. An Archaeological Investigation of Raised Field Agriculture in the Lake Titicaca Basin of Peru. Unpublished Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana—Champaign. 1988b. Raised Field Agriculture in the Lake Titicaca Basin: Putting Ancient Andean Agriculture Back to Work. Expedition, 30(2): 8-16. 1991. The Organization of Prehispanic Intensive Agriculture in the Lake Titicaca Basin. Paper presented at the 47th International Congress of the Americanists, New Orleans. 1992a. Prehistoric Landscape Management in the Andean Highlands: Raised Field Agriculture and Its Environmental Impact. Population and Environment, 13: 285-300. 1992b. Applied Archaeology and Rural Development: Archaeology's Potential Contribution to the Future, journal of the Steward Anthropological Society, 20: 1-16. 1993. Methodological Considerations in the Study of Ancient Andean Field Systems. In The Archaeology of Garden and Field, edited by Naomi Miller and Kathryn L. Gleason, University of Pennsylvania Press, Philadelphia, in press. Erickson, Clark L. and Daniel A. Brinkmeier. 1991. Raised Field Rehabilitation Projects in the Northern Lake Titicaca Basin. MS on file, Department of Anthropology, University of Pennsylvania, Philadelphia. Erickson, Clark L. and Kay L. Candler. 1989. Raised Fields and Sustainable Agriculture in the Lake Titicaca Basin. In Fragile Lands of Latin America: Strategies for Sustainable Development, edited by John Browder, pp. 230-248. Westview Press, Boulder. Erickson, Clark L. and Juan Faldin. 1979. Informe Preliminar sobre un Reconocimiento Arqueologico en los Llanos de Mojos, Bolivia: San Ignacio a San Borja. Documentos Internos del INAR, 36/78, La Paz.
References
273
Erickson, Clark L., Jose Esteves, Wilma Winkler, Marcos Michel. 1991. Estudio Preliminar de los Sistemas Agricolas Precolombinos en el Departamento del Beni, Bolivia. MS on file, Department of Anthropology, University of Pennsylvania, Philadelphia, and the Instituto Nacional de Arqueologia, La Paz. Estupinan Tello, Julio. 1983. Historia de Esmeraldas. Imprenta Los Colorados, Santo Domingo de los Colorados. Evans, Clifford and Betty J. Meggers, i960. Archaeological Investigations in British Guiana. Smithsonian Institution, Bureau of American Ethnology, Bulletin 117. Washington D.C. 1968. Archaeological Investigations on the Rio Napo, Eastern Ecuador. Smithsonian Contributions to Anthropology No. 6. Washington D.C. Faegri, K. 1966. Some Problems of Representivity in Pollen Analysis. Paleobotanist, 15: 135—140. Faegri, K. and J. Iversen. 1975. A Textbook of Pollen Analysis. 3rd edn. Haffer, New York. Faldin, Juan. 1984. La Arqueologia Beniana y su Panorama Interpretive Arqueologia Boliviana, 1: 83-90. Falesi, Italo C. 1974. Soils of the Brazilian Amazon. In Man in the Amazon, edited by C. Wagley, pp. 201-229. University Presses of Florida, Gainesville. Falk, Carl R. and Holmes A. Semken Jr. 1990. Vertebrate Paleoecology and Procurement at the Rainbow Site. In Woodland Cultures on the Western Prairies: The Rainbow Site Investigations, edited by David W. Benn, pp. 145-164. Report No. 18. Office of the State Archaeologist, University of Iowa, Iowa City. Farrington, Ian (ed.). 1985. Prehistoric Intensive Agriculture in the Tropics. British Archaeological Reports, International Series 232 (i) and (ii), Oxford. Fassig, Oliver L. 1909. Average Annual Rainfall of Porto Rico, W.I. Monthly Weather Review, 37: 982-986. 1911. The Normal Temperature of Porto Rico, West Indies. Monthly Weather Review, 39: 299-302. Fejos, Paul. 1943. Ethnography of the Yagua. Publications in Anthropology 1. Viking Fund, New York. Fernandez-Jalvo, Y. and P. Andrews. 1992. Small Mammal Taphonomy of Gran Dolina, Atapuerca (Burgos), Spain. Journal of Archaeological Science, 19: 407-428. Fisher, D.C. 1981. Crocodilian Scatology, Microvertebrate Concentrations, and Enamel-less Teeth. Paleobiology, 7: 262-275. Flannery, Kent V. 1991. Discussant's Comments in the Symposium "The Transition to Agriculture: New World." Paper Presented at the 56th Annual Meeting of the Society for American Archaeology, New Orleans. Flannery, Kent V. (ed.). 1976. The Early Mesoamerican Village. Academic Press, New York. Fleming, Theodore H. 1973. Numbers of Mammal Species in North and Central American Forest Communities. Ecology 54: 555—563. 1975. The Role of Small Mammals in Tropical Ecosystems. In Small Mammals: Their Productivity and Population Dynamics, edited by F. B. Golley, K. Petrusewicz, and L. Ryszkowski, pp. 269-298. Cambridge University Press, Cambridge. Fogel, M. L., N. Tuross, and D. Owsley. 1989. Nitrogen Isotope Tracers of Human Lactation in Modern and Archaeological Populations. 1988-1989 Annual Report of the Director, pp. 111116. Geophysical Laboratory, Carnegie Institution of Washington. Ford, J. A. 1952. Measurements of Some Prehistoric Design Developments in the Southeastern States. Anthropological Papers of the American Museum of Natural History Vol. 44: Part 1, New York. 1954. The Type Concept Revisited. American Anthropologist, 56(1): 42-54. 1969. A Comparison of Formative Cultures in the Americas. Smithsonian Contributions to Anthropology No. 11, Washington D.C.
274
References
Foster, R. 1982. Structure and History of the Vegetation of Barro Colorado Island. In The Ecology of a Tropical Forest: Seasonal Rhythms and Long-term Changes, edited by E. G. Leigh Jr., A. S. Rand, and D. M. Windsor, pp. 67-81. Smithsonian Institution Press, Washington D.C. Gallagher, Patrick. 1976. La Pitta: An Archaeological Site in Northwestern Venezuala. Publications in Anthropology 76. Yale University, New Haven. Garaycochea Z., Ignacio. 1986a. Rehabilitacion de Camellones en la Comunidad Campesina de Huatta, Puno. Unpublished Agricultural Engineering thesis, Department of Agronomy, Universidad Nacional del Altiplano, Puno, Peru. 1986b. Potencial Agricola de los Camellones en el Altiplano Puneno. In Andenes y Camellones en el Peru Andino: Historia Presente y Futuro, edited by Carlos de la Torre and Manuel Burga, pp. 241-251. Consejo Nacional de Ciencia y Tecnologia, Lima. 1987. Agricultural Experiments in Raised Fields in the Titicaca Basin Peru: Preliminary Considerations. In Pre-Hispanic Agricultural Fields in the Andean Region, edited by W. M. Denevan, K. Mathewson, and G. Knapp, pp. 385—398. British Archaeological Reports, International Series 359, Oxford. Garland, A.N. 1987. A Histological Study of Archaeological Bone Decomposition. In Death, Decay and Reconstruction, edited by A. Boddington, A. N. Garland and R. C. Janaway, pp. 109-126. University of Manchester Press, Manchester. Garlick, J. D. 1970. Buried Bone: The Experimental Approach in the Study of Nitrogen Content and Blood Group Activity. In Science in Archaeology, edited by Don Brothwell and Eric Higgs, pp. 503-512. Praeger Publishers, New York. Garson, Adam. 1980. Prehistory, Settlement, and Food Production in the Savanna Region of La Calzada de Paez, Venezuela. Unpublished Ph.D. dissertation, Department of Anthropology, Yale University, New Haven. Genoves, S. 1967. Proportionality of the Long Bones and Their Relation to Stature Among Mesoamericans. American journal of Physical Anthropology, 26: 67-77. Gentry, A. H. and L. H. Emmons. 1987. Geographical Variation in Fertility, Phenology, and Composition of the Understory of Neotropical Forests. Biotropica, 19: 216-227. Gibbons, Ann. 1990. New View of Early Amazonia. Science, 248: 1488—1490. Gifford, Diane P. 1981. Taphonomy and Paleoecology: A Critical Review of Archaeology's Sister Disciplines. In Advances in Archaeological Method and Theory, IV, edited by Michael B. Schiffer, pp. 365-438. Academic Press, New York. Gillespie, William B. 1989. Vertebrate Faunal Remains. In Hohokam Archaeology Along Project. Phase B of the Tucson Aqueduct Central Arizona Project, IV, edited by Jon S. Czaplicki and John C. Ravesloot, pp. 183-195. Arizona State Museum Cultural Resource Management Division Archaeological Series 78. University of Arizona, Tucson. Gillin, J. 1941. Quichua-speaking Indians of the Province of Imbabura (Ecuador) and Their Anthropometric Relations with the Living Populations of the Andean Area. Bureau of American Ethnology Bulletin, 128: 167-228. Glassie, Henry. 1975. Folk Housing in Middle Virginia. University of Tennessee Press, Knoxville. Gleason, Katherine and Naomi Miller (eds.). 1993. The Archaeology of Garden and Field. University of Pennsylvania Press, Philadelphia, in press. Goldman, Irving. 1963. The Cubeo. Indians of the Northwest Amazon. Illinois Studies in Anthropology 2. University of Illinois, Urbana. Gomez-Pompa, Arturo, Hector Luis Morales, Epifanio Jimenez Avilla, and Julio Jimenez Avilla. 1982. Experiences in Traditional Hydraulic Agriculture. In Maya Subsistence, edited by Kent V. Flannery, pp. 327—342. Academic Press, New York. Goodenough, Ward. 1957. Cultural Anthropology and Linguistics. In Report of the Seventh Annual Round Table Meeting on Linguistics and Language Study, edited by P. Gavin, pp. 167—173. Monograph Series on Language and Linguistics, No. 9. Georgetown University Press, Washington D.C.
References
275
Goodman, A. H., D. L. Martin, G. J. Armelagos, and G. Clark. 1984. Indications of Stress from Bone and Teeth. In Paleopathology at the Origins of Agriculture, edited by M. N. Cohen and G. J. Armelagos, pp. 13-49. Academic Press, New York. Goodman, A. H., R. B. Thomas, A. C. Swedlund, and G. J. Armelagos. 1988. Biocultural Perspectives on Stress in Prehistoric, Historical, and Contemporary Population Research. Yearbook of Physical Anthropology, 31: 169-202. Gordon, Burton Le Roy. 1957. Human Geography and Ecology in the Sinu Country of Colombia. Ibero-Americana 39. University of California, Berkeley. 1982. A Panama Forest and Shore. Natural History and Amerindian Culture in Bocas del Toro. Boxwood Press, Pacific Grove. Grace, Barry. 1983. The Climate of the Altiplano, Department ofPuno. Canadian International Development Agency, Puno. Graffam, Gray. 1989. Back across the Great Divide: The Pakaq Senorio and Raised Field Agriculture. In Multidisciplinary Studies in Andean Anthropology, edited by Virginia J. Vitzthum, pp. 33-50. Michigan Discussions in Anthropology, 8 (Fall 1988), Ann Arbor. 1990. Raised Fields without Bureaucracy: An Archaeological Examination of Intensive Wetland Cultivation in the Pampa Koani Zone, Lake Titicaca, Bolivia. Unpublished Ph.D. dissertation, Department of Anthropology, University of Toronto, Toronto. Grayson, Donald K. 1981a. The Effects of Sample Size on Some Derived Measures in Vertebrate Faunal Analysis, journal of Archaeological Science, 8: 77-88. 1981b. A Critical View of the Use of Archaeological Vertebrates in Paleoenvironmental Reconstruction, journal of Ethnobiology, 1: 28-38. 1984. Quantitative Zoo archaeology. Topics in the Analysis of Archaeological Faunas. Academic Press, New York. 1988. Danger Cave, Last Supper Cave and Hanging Rock Shelter: The Faunas. Anthropological Papers of the American Museum of Natural History 66: Part 1, New York. 1989. Bone Transport, Bone Destruction, and Reverse Utility Curves. Journal of Archaeological Science, 16: 643-652. 1991. The Small Mammals of Gatecliff Shelter: Did People Make a Difference? In Beamers, Bobwhites and Blue-Points. Tributes to the Career of Paul W. Parmalee, edited by James R. Purdue, Walter E. Klippel and Bonnie W. Styles, pp. 99-109. Scientific Papers, vol. 23. Illinois State Museum, Springfield. Green, Roger. 1979. Sampling Design and Statistical Methods for Environmental Biologists. John Wiley and Sons, New York. Gregor, Thomas. 1977. Mehinaku: The Drama of Daily Life in a Brazilian Indian Village. University of Chicago Press, Chicago. Gross, Daniel R. 1975. Protein Capture and Cultural Development in the Amazon Basin. American Anthropologist, 77: 526-549. Gumerman, G. IV, and B. S. Umemoto. 1987. The Siphon Technique: An Addition to the Flotation Process. American Antiquity, 52: 330-336. Hames, Raymond B. 1979. A Comparison of the Efficiencies of the Shotgun and the Bow in Neotropical Forest Hunting. Human Ecology, 7: 219—252. 1980. Game Depletion and Hunting Zone Rotation Among the Ye'kwana and Yanomamo of Amazonas, Venezuala. Working Papers on South American Indians, 2: 31-66. Hames, Raymond B. and William T. Vickers. 1982. Optimal Diet Breadth Theory as a Model to Explain Variability in Amazonian Hunting. American Ethnologist, 9: 358-378 Hammond, N., S. Donaghey, C. Gleason, J. Staneko, D. Tuerenhout, and L. Kosakowsky. 1987. Excavations at Nohmul, Belize, 1985. journal of Field Archaeology, 14: 257-282. Hanke, Wanda. 1957. Einige Funde im Beni-Gebiet, Ostbolivien. Archiv fur Volkerkunde, 12: 136-143.
276
References
Hansell, P. 1988. La Mula-Sarigua: A Formative Settlement on the Coast ofParita Bay, Central Pacific Panama. Unpublished Ph.D. dissertation, Department of Anthropology, Temple University, Philadelphia. Hansen, B. C. S., H. E. Wright, and J. P. Bradbury. 1984. Pollen Studies in the Junin area, Central Peruvian Andes. Geological Society of America Bulletin, 95: 1454-1465. Hare, P. E. 1980. Organic Geochemistry of Bone and its Relation to the Survival of Bone in the Natural Environment. In Fossils in the Making. Vertebrate Taphonomy and Paleoecology, edited by Anna K. Behrensmeyer and Andrew P. Hill, pp. 208-219. University of Chicago Press, Chicago. Harrison, Peter D. and B. L. Turner II (eds.). 1978. Pre-Hispanic Maya Agriculture. University of New Mexico Press, Albuquerque. Hartt, Charles F. 1871. The Ancient Indian Pottery of Marajo. American Naturalist, 5: 259-271. Hasenstab, Robert J. 1986. Windows on the World Beneath: Shovel Test Pits and Site Assessment in the Northeast. Unpublished Ph.D. dissertation, Department of Anthropology, University of Massachusetts, Amherst. Hayden, Brian. 1984. Are Emic Types Relevant to Archaeology? Ethnohistory, 31: 79-92. Hayden, Brian and Aubrey Cannon. 1983. Where the Garbage Goes: Refuse Disposal in the Maya Highlands. Journal of Anthropological Archaeology, 2: 117-163. Heer, O. 1866. Treatise on the Plants of the Lake Dwellings. In The Lake Dwellings of Switzerland and Other Parts of Europe, edited by F. Keller (cited in Renfrew 1973). Longmans, Green, London. 1878. Abstract of the Treatise on the "Plants of the Lake Dwellings." In The Lake Dwellings of Switzerland and Other Parts of Europe, I, edited by F. Keller, pp. 518-544. Longmans, Green, London. Henderson, Janet. 1987. Factors Determining the State of Preservation of Human Remains. In Death, Decay and Reconstruction, edited by A. Boddington, A. N. Garland and R. C. Janaway, pp. 43-54. University of Manchester Press, Manchester. Henley, Paul. 1982. The Panare. Tradition and Change on the Amazonian Frontier. Yale University Press, New Haven. Henry, D.O. 1989. From Foraging to Agriculture. University of Pennsylvania Press, Philadelphia. Herrera de Turbay, Luisa Fernanda. 1985. Agricultura Aborigen y Cambios de Vegetacion en la Sierra Nevada de Santa Marta. Banco de la Republica, Fundacion de Investigaciones Arqueologicas Nacionales, Bogota. Herrera, Luisa Fernanda, Ines Cavelier, Camilo Rodriguez, and Santiago Mora. 1992. The Technical Transformation of an Agricultural System in the Colombian Amazon. World Archaeology, 24: 98-113. Hershkovitz, Philip. 1972. The Recent Mammals of the Neotropical Region: A Zoogeographic and Ecological Review. In Evolution, Mammals, and Southern Continents, edited by Allen Keast, Frank C. Erk, and Bendy Glass, pp. 311-431. State University of New York Press, Albany. Hesse, B. 1985. Archaic Exploitation of Small Mammals and Birds in Northern Chile. Estudios Atacamehos, 7: 42-61. 1986. Buffer Resources and Animal Domestication in Prehistoric Northern Chile. Archaeozoologia Melanges, 1986: 73-85. Hilbert, Peter P. 1968. Archa'ologische Untersuchungen am Mittleren Amazonas: Beitrage zur Vorgeschichte des Sudamerikanischen Teiflandes. Dietrich Reimer Verlag, Berlin. Hockett, Bryan Scott. 1989. Burned Bones in Woodrat Nests from Northwestern Nevada. Current Research in the Pleistocene, 6: 41-43. 1991. Toward Distinguishing Human and Raptor Patterning on Leporid Bones. American Antiquity, 56: 667-679. Hoffman, Rob. 1988. The Contribution of Raptorial Birds to Patterning in Small Mammal Assemblages. Paleobiology, 14: 81-90.
References
277
Hofman, Jack L. 1986. Vertical Movement of Artifacts in Alluvial and Stratified Deposits. Current Anthropology, 27: 163-171. Holloway, R. G. and V. M. Bryant Jr. 1986. New Directions of Palynology in Ethnobiology. Journal of Ethnobiology, 6: 47-65. Hooghiemstra, Henry. 1984. Vegetational and Climatic History of the High Plain of Bogatd, Colombia: A Continuous Record of the Last 3.5 Million Years. Dissertationes Botanicae, Band 79. J. Cramer, Vaduz. Horowitz, S. and G. Armelagos, with K. Wachter. 1988. On Generating Birth Rates from Skeletal Populations. American Journal of Physical Anthropology, j6: 189-196. Howard, George D. 1947. Prehistoric Ceramic Styles of Lowland South America, Their Distribution and History. Publications in Anthropology No. 37. Yale University, New Haven. Hutchinson, D. L. and L. Norr. 1993. Late Prehistoric and Early Historic Diet in Gulf Coast Florida. In The Frontiers of Biological Adaptation, edited by C. S. Larson and G. R. Milner. WileyLiss, New York, in press. Illanes L., Lewis. 1984. Realidad Chachi. Editorial Porvenir, Quito. Isbell, William H. 1978. Environmental Perturbations and the Origins of the Andean State. In Social Archaeology: Beyond Subsistence and Dating, edited by C. L. Redman, E. V. Curtin, W. T. Langhorne Jr., N. M. Versaggi, and J. C. Wanser, pp. 303-313. Academic Press, New York. Jacob, John G. 1991a. El Villar Site, Llanos de Mojos, Bolivia: Preliminary Report on Soil Analysis. MS on file, Department of Soil and Plant Sciences, Texas A. &c M. University, College Station, Texas. Jacob, John G. (ed.). 1991b. The Agroecological Evolution of Cobweb Swamp, Belize. The Palynology Laboratory, Texas A. & M. University. Janson, Charles H. and Louise H. Emmons. 1990. Ecological Structure of the Nonflying Mammal Community at Cocha Cashu Biological Station, Manu National Park, Peru. In Four Neotropical Rainforests, edited by Alwyn H. Gentry, pp. 314-338. Yale University Press, New Haven. Jarman, H. N., A. J. Legge, and J. A. Charles. 1972. Retrieval of Plant Remains from Archaeological Sites by Froth Flotation. In Papers in Economic Prehistory, edited by E. S. Higgs, pp. 39—48. Cambridge University Press, Cambridge. Jimenez, Juan and Arturo Gomez-Pompa. 1987. Las Chinampas Mexicanas. Pensamiento lberoamericano,
12: 210—215.
Johansen, O. S., S. Gulliksen, and R. Nydal. 1986. 13C and Diet: Analysis of Norwegian Human Skeletons. Radiocarbon, 28: 754-761. Johnson, Allen. 1982. Reductionism in Cultural Ecology: The Amazon Case. Current Anthropology, 23: 413-428. Jones, J. G. 1988. Middle to Late Preceramic (6000-3000 B.P.) Subsistence Patterns on the Central Coast of Peru: The Coprolite Evidence. Unpublished Masters Thesis, Department of Anthropology, Texas A. & M. University, College Station, Texas. 1991a. Palynology. In The Agroecological Evolution of Cobweb Swamp, Belize, edited by J. S. Jacob, pp. 109-131. Preliminary Report to the National Geographic Society, Palynology Laboratory, Texas A. & M. University. 1991b. A Palynological Study of Raised Fields in the Llanos de Mojos, Bolivia. MS on file, Department of Anthropology, Texas A. & M. University, College Station, Texas. Jones, Kevin T. 1983. Forager Archaeology: The Ache of Eastern Paraguay. In Carnivores, Human Scavengers and Predators: A Question of Bone Technology, edited by G. M. LeMoine and A. S. MacEachern, pp. 171-191. The Archaeological Association of the University of Calgary, Calgary. 1984. Flunting and Scavenging by Early Hominids: A Study in Archaeological Method and Theory. Unpublished Ph.D. dissertation, Department of Anthropology, University of Utah, Provo.
278
References
Joyce, Thomas A. 1912. South American Archaeology. G. P. Putnam's Sons, New York. Jungerius, P. D. 1976. Quaternary Landscape Development of the Rio Magdalena Basin Between Neiva and Bogota (Colombia). A Reconstruction Based on Evidence Derived from Paleosols and Slope Deposits. Palaeogeography, P alaeo climatology, P alaeoecology, 19: 89—137. Kadane, J. B. 1988. Possible Statistical Contributions to Paleoethnobotany. In Current Paleoethnobotany. Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, pp. 206—214. University of Chicago Press, Chicago. Kaplan, Joanna O. 1975. The Piaroa. A People of the Orinoco Basin. Clarendon Press, Oxford. Karasov, W. H. and J. M. Diamond. 1988. Interplay between Physiology and Ecology in Digestion. BioScience, 38: 602—611. Karsten, Rafael. 1935. The Head-Hunters of Western Amazonas: The Life and Culture ofthejibaro Indians of Eastern Ecuador and Peru. Centraltryckeriet, Helsingfors. Katzenberg, M. A. 1984. Chemical Analysis of Prehistoric Human Bone from Five Temporally Distinct Populations in Southern Ontario. National Museum of Man, Mercury Series, Toronto. 1988. Stable Isotope Analysis of Animal Bone and the Reconstruction of Human Palaeodiet. In Diet and Subsistence: Current Archaeological Perspectives, edited by B. V. Kennedy and G. M. LeMoine, pp. 307-314. Proceedings of the Nineteenth Annual Chacmool Conference. Archaeological Association of the University of Calgary, Calgary. 1989. Stable Isotope Analysis of Archaeological Faunal Remains from Southern Ontario. Journal of Archaeological Science, 16: 319-329. Kauffman, J. Boone, C. Uhl, and D. L. Cummings. 1988. Fire in the Amazon 1: Fuel Biomass and Fire Chemistry in the Evergreen Rainforest of Venezuela. Oikos, 53: 167-175. Kay, Richard F. and William L. Hylander. 1978. The Dental Structure of Mammalian Folivores with Special Reference to Primates and Phalangeroidea (Marsupialia). In The Ecology of Arboreal Folivores, edited by G. G. Montgomery, pp. 173—191. Smithsonian Institution Press, Washington D.C. Kaye, Clifford A. 1959. Shoreline Features and Quaternary Shoreline Changes, Puerto Rico. Geological Survey Professional Paper 317—B. Washington D.C. Kealhofer, L. and D. R. Piperno. n.d. Phytoliths in the Southeast Asia Flora. Review ofPaleobotany and Palynology, MS under review. Keegan, W. and M. J. DeNiro. 1988. Stable Carbon- and Nitrogen-Isotope Ratios of Bone Collagen Used to Study Coral-Reef and Terrestrial Components of Prehistoric Bahamian Diet. American Antiquity, 53: 320-336. Keeling, C. D., W. G. Mook, and P. P. Tans. 1979. Recent Trends in the 13C/12C Ratio of Atmospheric Carbon Dioxide. Nature 277: 121-123. Kelm, Heinz. 1953. Archaologische Fundstiicke aus Ostbolivien. Baessler-Archivr, Neue Folge, 11: 65-92. Kennedy, B. V. E. 1988. Variation in 13C Values of Post-Medieval Europeans. Unpublished Ph.D. dissertation, Department of Archaeology, University of Calgary, Calgary. Kent, S. 1986. The Influence of Sedentism and Aggregation on Porotic Hyperostosis and Anaemia: A Case Study. Man, 21: 605-636. Kerley, E. R. and D. H. Ubelaker. 1978. Revisions in the Microscopic Method of Estimating Age at Death in Human Cortical Bone. American Journal of Physical Anthropology, 49: 545-546. Killion, Thomas W. (ed.). 1992. Gardens of Prehistory: The Archaeology of Settlement Agriculture in Greater Mesoamerica. University of Alabama Press, Tuscoloosa. Kiltie, Richard A. 1980. More on Amazon Cultural Ecology. Current Anthropology, 21: 541—544. Kintigh, Keith W. 1988a. The Effectiveness of Subsurface Testing: A Simulation Approach. American Antiquity, 53: 686-707. 1988b. The Archaeologist's Analytical Toolkit. Software and documentation commercially distributed by the author, Tempe Arizona.
References
279
Kintigh, Keith W. and Albert J. Ammerman. 1982. Heuristic Approaches to Spatial Analysis in Archaeology. American Antiquity, 47: 31-63. Klein, Richard G. 1989. Why does Skeletal Part Representation Differ Between Smaller and Larger Bovids at Klasies River Mouth and Other Archaeological Sites? journal of Archaeological Science, 16: 363-381. Klepinger, L. L. and R. Mintel. 1986. Metabolic Considerations in Reconstructing Past Diet from Stable Carbon Isotope Ratios of Bone Collagen. In Proceedings of the 24th Archaeometry Symposium, edited by J. S. Olin and M. J. Blackman, pp. 43-48. Smithsonian Institution Press, Washington D.C. Kluckhohn, Clyde, i960. The Use of Typology in Anthropological Theory. In Selected Papers of the Fifth International Congress of Anthropological and Ethnological Sciences, pp. 134-140. Philadelphia. Knapp, Gregory. 1991. Andean Ecology: Adaptive Dynamics in Ecuador. Westview Press, Boulder, CO. Knapp, Gregory and Roy Ryder. 1983. Aspects of the Origin, Morphology, and Function of Ridged Fields in the Quito Altiplano, Ecuador. In Drained Fields of the Americas, edited by J. P. Darch, pp. 201-220. British Archaeological Reports, International Series 189, Oxford. Kobori, Larry S. 1979. Differential Bone Recovery Experiment. In Ezra's Retreat: A Rockshelter/ Cave Occupation Site in the North Central Great Basin, by James C. Bard, Colin Busby, and Larry S. Kobori, pp. 228-229. Center for Archaeological Research at Davis Publication 6. University of California, Davis. Koch, P., A. K. Behrensmeyer, N. Tuross, and M. L. Fogel. 1990. Isotopic Fidelity during Bone Weathering and Burial. 1989-1990 Annual Report of the Director, pp. 105-110. Geophysical Laboratory, Carnegie Institution, Washington. Kolata, Alan L. 1986. The Agricultural Foundations of the Tiwanaku State: A View from the Heartland. American Antiquity, 51: 748-762. 1991. The Technology and Organization of Agricultural Production in the Tiwanaku State. Latin American Antiquity, 2: 99—125. Kolata, Alan (ed.). 1989. Arqueologia de Luqurmata, II: La Tecnologia y Organizacion de la Produccion Agricola en el Estado de Tiwanaku. Instituto Nacional de Arqueologia, La Paz. Kolata, Alan and Gray Graffam. 1989. Los Campos Elevados de Lukurmata, Bolivia. In Arqueologia de Lukurmata, edited by Alan Kolata, pp. 173-212. Instituto Nacional de Arqueologia, La Paz, Bolivia. Kolata, Alan L. and Charles Ortloff. 1989. Thermal Analysis of Tiwanaku Raised Field Systems in the Lake Titicaca Basin of Bolivia. Journal of Archaeological Science, 16: 233-262. Korth, William W. 1979. Taphonomy of Microvertebrate Fossil Assemblages. Annals of the Carnegie Museum, 48: 235-285. Krakker, James J., Michael J. Shott, and Paul D. Welch. 1983. Design and Evaluation of ShovelTest Sampling in Regional Archaeological Survey, journal of Field Archaeology, 10: 469—480. Kreshover, S. J. i960. Metabolic Disturbances in Tooth Formation. Annals of the New York Academy of Sciences, 85: 161-167. Kroeber, Alfred L. 1944. Configurations of Culture Growth. University of California Press, Berkeley. 1957. Style and Civilizations. Cornell University Press, Ithaca. Krueger, H. W. 1991. Exchange of Carbon and Strontium with Hydroxyapatite. Journal of Archaeological Science, 18: 355-361. Krueger, H. W., and C. H. Sullivan. 1984. Models for Carbon Isotope Fractionation between Diet and Bone. In Stable Isotopes in Nutrition, edited by J. E. Turnlund and P. E. Johnson, pp. 205222. Symposium Series 258, American Chemical Society. Kuhry, Peter. 1988. Palaeobotanical—palaeoecological Studies of Tropical High Andean Peatbog Sections (Cordillera Oriental, Colombia). Dissertationes Botanicae, Band 116. J. Cramer, Berlin-Stuttgart.
280
References
Kuljis N., Danilo, Victor Bustos S., and Bernardo Dougherty. 1977. Prospeccion del Rio Mamore, Departamento del Beni. Documentos Internos del INAR 4/77, La Paz. Kunth, C. 1826. Examen Botanique. In Catalogue Raisonne et Historique des Antiquites Decouvertes en Egypte, edited by J. Passalacqua. Paris (cited in Renfrew 1973). Kusmer, Karla D. 1990. Taphonomy of Owl Pellet Deposition. Journal of Paleontology, 64: 629-637. Lallo, J. W., G. J. Armelagos, and R. P. Mensforth. 1977. The Role of Diet, Disease, and Physiology in the Origin of Porotic Hyperostosis.. Human Biology, 49: 471-483. Lambert, Joseph B., Carole B. Szpunar and Jane E. Buikstra. 1979. Chemical Analysis of Excavated Human Bone from Middle and Late Woodland Sites. Archaeometry, 21: 115-129. Lambert, Joseph B., Sharon M. Vlasak, Anne C. Thometz, and Jane E. Buikstra. 1982. A Comparative Study of the Chemical Analysis of Ribs and Femurs in Woodland Populations. American Journal of Physical Anthropology, 59: 289-294. Lambert, Joseph B., Sharon Vlasak Simpson, Carole B. Szpunar, and Jane E. Buikstra. 1985a. Bone Diagenesis and Dietary Analysis. Journal of Human Evolution, 14: 477-482. Lambert, Joseph B., Sharon Vlasak Simpson, Susan Gorell Weiner and Jane E. Buikstra. 1985b. Induced Metal-Ion Exchange in Excavated Human Bone. Journal of Archaeological Science, 12: 85-92.
Land, L. S., E. L. Lundelius, and S. Valastro. 1980. Isotopic Ecology of Deer Bones. Palaeogeography, Valaeoclimatology, Palaeoecology, 32: 143-151. Langebaek Rueda, Carl Henrik. 1987a. La Cronologia de la Region Arqueologica Tairona Vista desde Papare, Municipio de Cienega. Boletin de Arqueologia, 2(1): 83-101. 1987b. Relaciones de los Desarollos del Area Tairona y el Intercambio. Boletin de Arqueologia, 2(2): 32-41. Lanning, E. 1964. Peru Before the Incas. Prentice-Hall, Englewood Cliffs, NJ. Larsen, C. S., R. Shavit, and M. C. Griffin. 1991. Dental Caries Evidence for Dietary Change: An Archaeological Context. In Advances in Dental Anthropology, edited by M. A. Kelley and C. S. Larsen, pp. 179-202. Wiley-Liss, New York. Larsen, C. S., Larsen, M. J. Schoeninger, N. J. van der Merwe, K. M. Moore, and J. A. Lee-Thorp. 1992. Carbon and Nitrogen Stable Isotopic Signatures of Human Dietary Change in the Georgia Bight. American Journal of Physical Anthropology, 89: 197—214. Lathrap, Donald W. 1962. Yarinacocha: Stratigraphic Excavations in the Peruvian Montana. Unpublished Ph.D. dissertation, Department of Anthropology, Harvard University, Cambridge, MA. 1964. An Alternative Seriation of the Mabaruma Phase, Northwestern British Guiana. American Antiquity, 29(3): 353-359. 1966. The Mabaruma Phase: A Return to the More Probable Interpretation. American Antiquity, 31(4): 558-566. 1968a. Aboriginal Occupation and Changes in River Channel on the Central Ucayali, Peru. American Antiquity, 33: 62—79. 1968b. The "Hunting" Economies of the Tropical Forest Zone of South America: An Attempt at Historical Perspective. In Man the Hunter, edited by R. B. Lee and I. DeVore, pp. 23-29. Aldine, Chicago. 1970. The Upper Amazon. Thames and Hudson, Southampton. 1973a. Gifts of the Cayman: Some Thoughts on the Subsistence Basis of Chavin. In Variation in Anthropology, edited by Donald W. Lathrap and J. Douglas, pp. 91-105. Illinois Archaeological Survey, Urbana. 1973b. The Antiquity and Importance of Long-Distance Trade Relationships in the Moist Tropics of Pre-Columbian South America. World Archaeology, 5: 170-186. 1974. The Moist Tropics, the Arid Lands, and the Appearance of Great Art Styles in the New World. Art and Environment in Native America, edited by M. E. King and I. R. Traylor Jr., pp. 115-158. Special Publications 7. The Museum, Texas Tech LIniversity, Lubbock.
References
281
1977. Our Father the Cayman, Our Mother the Gourd: Spinden Revisited, or a Unitary Model for the Emergence of Agriculture in the New World. In Origins of Agriculture, edited by Charles A. Reed, pp. 713-751. Mouton, The Hague. Lathrap, Donald W., Angelika Gebhart-Sayer, and Ann M. Mester. 1985. The Roots of the Shipibo Art Style: Three Waves on Imiriacocha or There Were "Incas" Before the Incas. Journal of Latin American Lore, 11: 31-119. Lathrap, Donald W., Jorge G. Marcos, and James Zeidler. 1977. Real Alto: An Ancient Ceremonial Center. Archaeology, 30(1): 2-13. Lathrap, Donald W., A. Gebhart-Sayer, T. P. Myers, and A. M. Mester. 1987. Further Discussion of the Roots of the Shipibo Art Style: A Rejoinder to DeBoer and Raymond. Journal of Latin American Lore 13: 225-271. Lean, J., and D. A. Warrilow. 1989. Simulation of the Regional Climatic Impact of Amazonian Deforestation. Nature 242: 411-413. Leavitt, S. W., and A. Long. 1986. Trends of 13C/12C Ratios in Pinyon Pine Tree Rings of the American Southwest and the Global Carbon Cycle. Radiocarbon, 28: 376-382. Lee, Kenneth. 1979. 7.000 Anos de Historia del Hombre de Mojos: Agricultura en Pampas Esteriles: Informe Preliminar. Universidad Beni, pp. 23-26. Universidad Tecnica del Beni, Trinidad. Lee-Thorp, J. A. 1986. Stable Carbon Isotope Analysis of Bone Apatite: The Implications for Dietary and Paleoenvironmental Reconstruction. Paleoecology of Africa, 17: 133—135. 1989. Stable Carbon Isotopes in Deep Time: The Diets of Fossil Fauna and Hominids. Unpublished Ph.D. dissertation, Department of Archaeology, University of Cape Town. Lee-Thorp, J. A. and N. J. van der Merwe. 1987. Carbon Isotope Analysis of Fossil Bone Apatite. South African Journal of Science, 83: 712-715. 1991. Aspects of the Chemistry of Modern and Fossil Biological Apatites. Journal of Archaeological Science, 18: 343-354. Lee-Thorp, J. A., J. C. Sealy, and N. J. van der Merwe. 1989. Stable Carbon Isotope Ratio Differences between Bone Collagen and Bone Apatite, and Their Relationship to Diet. Journal of Archaeological Science, 16: 585-599. Lennon, Thomas J. 1982. Raised Fields of Lake Titicaca, Peru: A Pre-Hispanic Water Management System. Unpublished Ph.D. dissertation, Department of Anthropology, University of Colorado, Boulder. 1983. Pattern Analysis of Prehispanic Raised Fields of Lake Titicaca, Peru. In Drained Fields of the Americas, edited by J. P. Darch, pp. 183-200. British Archaeological Reports, International Series 189, Oxford. Lentz, D. L. 1991. Maya Diets of the Rich and Poor: Paleoethnobotanical Evidence from Copan. Latin American Antiquity, 2: 269-287. Leonard, Robert D. and George T. Jones (eds.). 1989. Quantifying Diversity in Archaeology. Cambridge University Press, Cambridge. Leone, Mark P. 1982. Recovering Mind. American Antiquity, 47: 742-760. Letolle, R. 1980. Nitrogen-15 in the Natural Environment. In Handbook of Environmental Isotope Geochemistry, I: The Terrestrial Environment, edited by A. P. Fritz and J. C. Fohtes, pp. 407— 433. Elsevier, Amsterdam. Levi-Strauss, Claude. 1948. The Tribes of the Upper Xingu River. In Handbook of South American Indians, III, edited by Julian H. Steward, pp. 321-348. Bureau of American Ethnology, Bulletin No.143. Smithsonian Institution, Washington D.C. 1967. Structural Anthropology. Anchor Books, New York. Levinson, M. 1982. Taphonomy of Microvertebrates - From Owl Pellets to Cive Breccia. Annals of the Transvaal Museum, 33: 115—121. Lightfoot, Kent G. 1986. Regional Surveys in the Eastern United States: The Strengths and Weaknesses of Implementing Subsurface Testing Programs. American Antiquity, 51: 484-504. Lightfoot, Kent G., R. Kalin, and James Moore. 1987. Prehistoric Hunter-gatherers of Shelter
282
References
Island, New York: An Archaeological Study of the Mashomack Preserve. Contributions of the University of California Archaeological Research Facility 46, Berkeley. Limp, W. Fredrick. 1989. The Use of Multispectral Digital Imagery in Archaeological Investigations. Arkansas Archaeological Survey Research Series No. 34, Fayetteville. Linares, Olga F. 1976. "Garden Hunting" in the American Tropics. Human Ecology, 4: 331-349. Lindskoog, John N. and Carrie A. Lindskoog. 1964. Cayapa. Serie de Vocabularios Indigenas 9. I. L. V., Quito. Liu, K.-b. and P. A. Colinvaux. 1988. A 5200-year history of Amazon Rain Forest. Journal of Biogeography, 15: 231-248. Livingstone, D. A. and T. van der Hammen. 1978. Palaeogeography and Palaeoclimatology. In Tropical Forest Ecosystems: A State-of-knowledge Report Prepared by Unesco/Unep/FAO. UNESCO, Paris. Long, A., B. F. Benz, D. J. Donahue. A. J. T. Jull, and L. J. Toolin. 1989. First Direct AMS Dates on Early Maize from Tehuacan, Mexico. Radiocarbon, 31: 1035—1040. Lounsbury, Floyd G. 1964. A Formal Account of the Crow- and Omaha-Type Kinship Terminologies. In Explorations in Cultural Anthropology, edited by W. H. Goodenough, pp. 351—393McGraw-Hill, New York. Lovell, N., D. E. Nelson, and H. P. Schwarcz. 1986. Carbon Isotope Ratios in Paleodiet: Lack of Age or Sex Effect. Archaeometry, 28: 51-55. Lovis, William A. 1976. Quarter Sections and Forests: An Example of Probability Sampling in the Northeastern Woodlands. American Antiquity, 41: 364-372. Lyman, R. Lee. 1984. Bone Density and Differential Survivorship of Fossil Classes. Journal of Anthropological Archaeology, 3: 259-299. 1985. Bone Frequencies: Differential Transport, In Situ Destruction, and the MGUI. Journal of Archaeological Science, 12: 221-236. 1986. On the Analysis and Interpretation of Species List Data in Zooarchaeology. Journal of Ethnobiology, 6: 67-81. 1991. Taphonomic Problems with Archaeological Analyses of Animal Carcass Utilization and Transport. In Beamers, Bobwhites and Blue-Points: Tributes to the Career of Paul W. Parmalee, edited by James R. Purdue, Walter E. Klippel and Bonnie W. Styles, pp. 125-138. Scientific Papers, vol. 23. Illinois State Museum, Springfield. 1992. Anatomical Considerations of Utility Curves in Zooarchaeology. Journal of Archaeological Science, 19: 7-22. Lyman, R. Lee and G. L. Fox. 1989. A Critical Evaluation of Bone Weathering as an Indication of Bone Assemblage Formation. Journal of Archaeological Science, 16: 293-317. Lyman, R. Lee and Stephanie D. Livingston. 1983. Late Quaternary Mammalian Zoogeography of Eastern Washington. Quaternary Research, 20: 360-373. Lyman, R. Lee and M. J. O'Brien. 1987. Plow-zone Zooarchaeology: Fragmentation and Identifiability. Journal of Field Archaeology, 14: 493-498. Lyman, R. Lee, Lori E. Houghton, and Annell L. Chambers. 1992. The Effect of Structural Density on Marmot Skeletal Part Representation in Archaeological Sites. Journal of Archaeological Science, 19: 557"573Lynott, M. J., T. W. Bouton, J. E. Price, and D. E. Nelson. 1986. Stable Carbon Isotopic Evidence for Maize Agriculture in Southeast Missouri and Northeast Arkansas. American Antiquity, 51: 51-65. Lyon, Patricia J. 1970. Differential Bone Destruction: An Ethnographic Example. American Antiquity, 35: 213-215. Maas, Mary C. 1985. Taphonomy of a Late Eocene Microvertebrate Locality, Wind River Basin, Wyoming (U.S.A.). Palaeogeography, Palaeoclimatology, Palaeoecology, 52: 123-142. MacArthur, Robert H. 1972. Geographical Ecology. Patterns in the Distribution of Species. Harper and Row, New York.
References
283
McGimsey III, C. R. 1956. Cerro Mangote: A Preceramic Site in Panama. American Antiquity, 22: 151-161.
1957. Further Data and a Date from Cerro Mangote, Panama. American Antiquity, 23: 434—435. McGimsey III, C. R., M. B. Collins, and T. W. McKern. 1966. Cerro Mangote and Its Population. Paper presented at the 37th International Congress of Americanists, Mar del Plata, Argentina. McKenna, Malcolm C. 1962. Collecting Small Fossils by Washing and Screening. Curator, 3: McManamon, Francis P. 1984. Discovering Sites Unseen. In Advances in Archaeological Method and Theory, VII, edited by M. B. Schiffer, pp. 223-292. Academic Press, New York. 1994. Discovering and Estimating the Frequencies and Distribution of Archaeological Sites in the Northeast. In Cultural Resource Management: Archaeological Research, Preservation Planning, and Public Education in the Northeastern United States, edited by J. K. Kerber, pp. 99114. Bergin and Garvey, Westport, CT. McNab, Brian K. 1978. Energetics of Arboreal Folivores: Physiological Problems and Ecological Consequences of Feeding on an Ubiquitous Food Supply. In The Ecology of Arboreal Folivores, edited by G. G. Montgomery, pp. 153-162. Smithsonian Institution Press, Washington D.C. McPherron, Shannon P. and Harold L. Dibble. 1989. On the Computerization of Archaeological Projects. Journal of Field Archaeology, 15: 431-440. Mangelsdorf, P. C. 1953. Review of: "Agricultural Origins and Dispersals," by Carl O. Sauer. American Antiquity, 19: 86-90. Marcos, Jorge G. 1978. The Ceremonial Precinct at Real Alto: Organization of Time and Space in Valdivia Society. Unpublished Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana. Marean, Curtis W. 1991. Measuring Post-depositional Destruction of Bone in Archaeological Assemblages. Journal of Archaeological Science, 18: 677—694. Marean, Curtis W. and Lillian M. Spencer. 1991. Impact of Carnivore Ravaging on Zooarchaeological Measures of Element Abundance. American Antiquity, 56: 645-658. Marean, Curtis W., Lillian M. Spencer, Robert J. Blumenschine, and Salvatore D. Capaldo. 1992. Captive Hyaena Bone Choice and Destruction, the Schlepp Effect and Olduvai Archaeofaunas. Journal of Archaeological Science, 19: 101-121. Mariotti, A. 1983. Atmospheric Nitrogen is a Reliable Standard for Natural 15N Abundance Measurements. Nature, 303: 685-687. Marshall, Fiona and Tom Pilgram. 1991. Meat Versus Within-Bone Nutrients: Another Look at the Meaning of Body Part Representation in Archaeological Sites. Journal of Archaeological Science, 18: 149-163. Mason, J. Alden. 1931-39. Archaeology of Santa Marta, Colombia: The Tairona Culture. Anthropological Series 20, Parts 1, 2, 3. Field Museum of Natural History, Chicago. Masters, P. M. 1987. Preferential Preservation of Non-collagenous Protein During Diagenesis: Implications for Chronometric and Stable Isotopic Measurements. Geochimica et Cosmochimica Ada, 51: 3209-3214. Mathewson, Kent. 1987. Estimating Labor Inputs for the Guayas Raised Fields: Interim Considerations. In Pre-Hispanic Agricultural Fields in the Andean Region, edited by W. M. Denevan, K. Mathewson, and G. Knapp, pp. 321-326. British Archaeological Reports, International Series 359, Oxford. Matson, R. G. and B. S. Chisholm. 1991. Basketmaker II Subsistence: Carbon Isotopes and Other Dietary Indicators from Cedar Mesa, Utah. American Antiquity, 56: 444-459. Matsutani, A. 1972. Spodographic Analysis of Ash from the Kotosh Site. In Andes 4: Excavations at Kotosh, Peru, 1963 and 1966, edited by S. Izumi and K. Terada, pp. 319—326. University of Tokyo Press, Tokyo. Maybury-Lewis, David. 1985. Foreword. In Vital Souls: Bororo Cosmology, Natural Symbolism, and Shamanism, by Jon Christopher Crocker, pp. xi-xiii. University of Arizona Press, Tucson.
284
References
Medina, E. and P. Minchin. 1980. Stratification of Leaf S13C Values in Amazonian Rain Forests. Oecologia, 45: 377-378. Medina, E., G. Montes, E. Cuevas, and Z. Rokzandic. 1986. Profiles of COZ Concentration and Delta 13C Values in a Tropical Rainforest of the Upper Rio Negro Basin, Venezuela. Journal of Tropical Ecology, 2: 207-217. Meggers, Betty J. 1966. Ecuador. Praeger Publishers, New York. 1971. Amazonia. Man and Culture in a Counterfeit Paradise. Aldine, Chicago. 1991. Cultural Evolution in Amazonia. In Profiles in Cultural Evolution: Papers from a Conference in Honor ofElman R. Service, edited by A. T. Rambo and K. Gollogly, pp. 191216. University of Michigan, Museum of Anthropology, Anthropological Papers No. 85, Ann Arbor. 1992. Prehistoric Population Density in the Amazon Basin. In Disease and Demography in the Americas, edited by J. W. Verano and D. H. Ubelaker, pp. 197-205. Smithsonian Institution Press, Washington D.C. Meggers, Betty J. and Clifford Evans. 1957. Archaeological Investigations at the Mouth of the Amazon. Bureau of American Ethnology, Bulletin 167, Smithsonian Institute, Washington D.C. 1983. Lowland South America and the Antilles. In Ancient South Americans, edited by Jesse Jennings, pp. 543-591. W. H. Freeman and Co., San Francisco. Meggers, Betty. J., Clifford Evans, and Emilio Estrada. 1965. Early Formative Period of Coastal Ecuador: The Valdivia and Machalilla Phases. Smithsonian Institution Contributions to Anthropology, vol. 1. Washington D.C. Meggers, Betty J., Ondemar F. Dias, Eurico Th. Miller, and Celso Perota. 1988. Implications of Archeological Distribution in Amazonia. In Proceedings of a Workshop on Neotropical Distribution Patterns, edited by P. E. Vanzolini and W. R. Heyer, pp. 275-294. Academia Brasileira de Ciencias, Rio de Janeiro. Mellett, James S. 1974. Scatological Origin of Microvertebrate Fossil Accumulations. Science, 185: 349-350. Mengoni Gonalons, Guillermo L. 1983. Prehistoric Utilization of Faunal Resources in Arid Argentina. In Animals and Archaeology, I: Hunters and Their Prey, edited by J. Clutton-Brock and C. Grigson, pp. 325-335. British Archaeological Reports, International Series 163, Oxford. Metraux, Alfred. 1942. The Native Tribes of Eastern Bolivia and Western Matto Grosso. Bureau of American Ethnology, Bulletin 134, Washington D.C. Michel, Marcos and Carlos Lemuz. 1992. Influencia Barrancoide en el Bajo Maniqui. Nuevos Aportes (Revista de Antropologia), Enero-Abril: 51-65. Michie, James L. 1990. Bioturbation and Gravity as a Potential Site Formation Process: The Open Area Site, 36GE261, Georgetown County, South Carolina. South Carolina Antiquities, 22(1 and 2): 27-46. Miksicek, C. H. 1983. Macrofloral Remains of the Pulltrouser Area: Settlements and Fields. In Pulltrouser Swamp. Ancient Maya Habitat, Agriculture, and Settlement in Northern Belize, edited by B. L. Turner II and P. D. Harrison, pp. 94-104. University of Texas Press, Austin. Miksicek, C , R. Mck. Bird, B. Pickersgill, S. Donagney, J. Cartwright, and N. Hammond. 1981. Preclassic Lowland Maize from Cuello, Belize. Nature, 289: 56-59. Minagawa, M. and E. Wada. 1984. Stepwise Enrichment of 15N along Food Chains: Further Evidence and the Relationship between S15N and Animal Age. Geochimica et Cosmochimica Ada, 48: 1135-1140. Monroy, Joel L. 1938. El Convento de la Merced de Quito de 1534-1617. 2nd edn. Editorial Labor, Quito. Monsalve, J. 1985. A Pollen Core from the Hacienda Lusitania. Pro Calima, 4: 40-44.
References
285
Mooney, H. A., J. H. Troughton, and J. A. Berry. 1977. Carbon Isotope Ratio Measurements of Succulent Plants in Southern Africa. Oecologia, 30: 295-305. Moore, A. M. T. and G. C. Hillman. 1992. The Pleistocene to Holocene Transition and Human Economy in Southwest Asia: The Impact of the Younger Dryas. American Antiquity, 57: 482-494. Moore, K. M., M. L. Murray, and M. J. Schoeninger. 1989. Dietary Reconstruction from Bones Treated with Preservatives. Journal of Archaeological Science, 16: 437-466. Mora C , Santiago, Luisa Fernanda Herrera, Ines Cavelier F., and Camilo Rodriguez. 1991. Cultivars, Anthropic Soils and Stability: A Preliminary Report of Archaeological Research in Araracuara, Colombian Amazonia. University of Pittsburgh Latin American Archaeology Reports No. 2. Pittsburgh. Moran, Emilio F. 1990. Levels of Analysis and Analytical Level Shifting: Examples from Amazonian Ecosystem Research. In The Ecosystem Approach in Anthropology: From Concept to Practice, edited by E. F. Moran, pp. 279—308. University of Michigan Press, Ann Arbor. Moreno, Secundo. 1987. Alzamientos Indigenas en la Audiencia de Quito, 1534-1803. Abya-Yala, Quito. Moreno Navarro, Isidoro. 1979. Ecologia y Sociedad de los Cayapas de Esmeraldas: Los Patrones de Asentamiento. Actes XLII Congres International des Americanistes, 9—A: 319—333. Paris. Morlan, Richard E. 1980. Taphonomy and Archaeology in the Upper Pleistocene of the Northern Yukon Territory: A Glimpse of the Peopling of the New World. Archaeological Survey of Canada Paper 94, Ottawa. Munizaga, J. R. 1965. Skeletal Remains from Sites of Valdivia and Machalilla Phases. In Early Formative Period of Coastal Ecuador: The Valdivia and Machalilla Phases, by B. J. Meggers, C. Evans, and E. Estrada, pp. 219—234. Smithsonian Contributions to Anthropology, vol. 1. Smithsonian Institution Press, Washington D.C. Murray, M. D. and M. J. Schoeninger. 1988. Diet, Status, and Complex Social Structure in Iron Age Central Europe: Some Contributions from Bone Chemistry. In Tribe and Polity in Late Prehistoric Europe, edited by D. B. Gibson and M. D. Geselowitz, pp. 155-178. Plenum, New York. Muse, Michael and Fausto Quintero. 1987. Experimentos de Reactivacion de Campos Elevados del Penon del Rio, Guayas, Ecuador. In Pre-Hispanic Agricultural Fields in the Andean Region, edited by W. M. Denevan, K. Mathewson, and G. Knapp, pp. 249-266. British Archaeological Reports, International Series 359, Oxford. Muto, Guy R. and Joel Gunn. 1981. A Study of Late-Quaternary Environments and Early Man Along the Tombigbee River, Alabama and Mississippi, Phase I Final Report. Submitted to the Heritage, Conservation, and Recreation Service, Washington D.C. and the Interagency Archeological Services, Atlanta; and the United States Army Corps of Engineers, South Atlantic District, Mobile. Myers, Thomas P. 1972. A Seasonal Campsite in the Peruvian Montana: Ethnography and Ecology in Archaeological Interpretation. American Antiquity, 37: 540-545. 1973. Toward the Reconstruction of Prehistoric Community Patterns in the Amazon Basin. In Variation in Anthropology: Essays in Honor of John C. McGregor, edited by Donald W. Lathrap and Jody Douglas, pp. 233-259. Illinois Archaeological Survey, Urbana. 1992. Agricultural Limitations of the Amazon in Theory and Practice. World Archaeology, 24: 82-97. Nance, Jack D. 1983. Regional Sampling in Archaeological Survey: The Statistical Perspective. In Advances in Archaeological Method and Theory, VI, edited by M. B. Schiffer, pp. 289-356. Academic Press, New York. 1990. Statistical Sampling in Archaeology. In Mathematics and Information Science in Archaeology: A Flexible Framework, edited by A. Voorrips, pp. 135-163. Holos, Bonn.
286
References
Nance, Jack D. and Bruce F. Ball. 1986. No Surprises? The Reliability and Validity of Test Pit Sampling. American Antiquity, 51: 457-483. Nelson, B. K., M. J. DeNiro, M. J. Schoeninger, D. P. DePaolo, and P. E. Hare. 1986. Effects of Diagenesis on Strontium, Carbon, Nitrogen and Oxygen Concentration and Isotopic Composition of Bone. Geochimica et Cosmochimica Acta, 50: 1941-1949. Nelson, D. G. A. and J. D. B. Featherstone. 1982. Preparation, Analysis and Characterization of Carbonated Apatites. Calcified Tissue International, 34: 569—581. Nietschmann, Bernard. 1972. Hunting and Fishing Focus Among the Miskito Indians, Eastern Nicaragua. Human Ecology, 1: 41-67. 1973. Between Land and Water. The Subsistence Ecology of the Miskito Indians, Eastern Nicaragua. Seminar Press, New York. Nimuendaju, Curt. 1946. The Eastern Timbira. University of California Publications in American Archaeology and Ethnology, Vol. 41, Berkeley. 1952. The Tukuna. University of California Publications in American Archaeology and Ethnology, vol. 45, Berkeley. Nordenskiold, Erland von. 1910. Archaologische Forschungen im Bolivianischen Flachland. Zeitschrift fur Ethnologie, 42: 806-822. 1913. Urnengraber und Mounds im Bolivianischen Flachlande. Baessler Archiv, 3: 205-255. 1924a. The Ethnography of South America as seen from the Mojos in Bolivia. Comparative Ethnographic Studies, No. 3, Goteborg. 1924b. Finds of Graves and Old Dwelling-Places on the Rio Beni, Bolivia. Ymer, 44: 229-237. 1930. L'Archeologie du Bassin de UAmazone. Ars Americana 1. Editions G. van Oest, Paris. 1938. An Historical and Ethnological Survey of the Cuna Indians. Goteborgs Museum, Goteborg. Norr, L. 1984. Prehistoric Subsistence and Health Status of Coastal Peoples of Lower Central America. In Paleopathology at the Origins of Agriculture, edited by M. N. Cohen and G. J. Armelagos, pp. 463-490. Academic Press, New York. 1990. Nutritional Consequences of Prehistoric Subsistence Strategies in Lower Central America. Unpublished Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana-Champaign. 1991. Dietary Diversity along the Northwestern Coast of Costa Rica. Paper Presented at the 48th International Congress of Americanists, New Orleans. 1992. Reassessment of the Relationship between a Maize-based Diet and Porotic Hyperostosis in Prehistoric Central Panama. Paper Presented at the 61st Annual Meeting of the American Association of Physical Anthropologists, Las Vegas. Norton, P. 1987. The Collaborative Approach to Rescue and Conservation: The Case of Salango. In Rescue Archeology: Proceedings of the Second New World Conference on Rescue Archeology, edited by R. L. Wilson, pp. 201-210. Southern Methodist University Press, Dallas. Oberem, Udo. 1971. Los Quijos. Historia de la Transculturacion de un Grupo Indigena en el Oriente Ecuatoriano (1538-1956). Memorias del Departamento de Antropologia y Etnologia de America, vol. 1. Universidad de Madrid, Madrid. Odell, George H. and Frank Cowan. 1987. Estimating Tillage Effects on Artifact Distributions. American Antiquity, 52: 456-484. O'Leary, M. 1981. Carbon Isotope Fractionation in Plants. Phy to chemistry, 20: 553-567. Oliver, Jose R. 1992. Donald W. Lathrap: Approaches and Contributions in New World Archaeology. Journal of the Steward Anthropological Society, 20: 283-345 (also published in Antropologica, 75-76, 1991). Ollendorf, A. L. 1992. Toward a Classification of Sedge (Cyperaceae) Phytoliths. In Phytolith Systematics, edited by G. R. Rapp Jr. and S. C. Mulholland, pp. 91-111. Plenum, New York. Onffroy de Thoron, Enrique. 1983. America Ecuatorial. 2 vols. Coleccion Ecuador, Quito.
References
287
Ortner, D. J. and W. G. J. Putschar. 1981. Identification of Pathological Conditions in Human Skeletal Remains. Smithsonian Contributions to Anthropology, vol. 28. Smithsonian Institution Press, Washington D.C. Ortner, Donald J., David W. Von Endt, and Mary S. Robinson. 1972. The Effect of Temperature on Protein Decay in Bone: Its Significance in Nitrogen Dating of Archaeological Specimens. American Antiquity, 37: 514-520. Oviedo, G. F. de. 1959. Historia General y Natural de las Indias. Biblioteca de Autores Espanoles No. 97, Madrid. Oyuela Caycedo, Augusto. 1986. De los Taironas a los Kogi: Una Interpretacion del Cambio Cultural. Boletin del Museo del Oro, 17: 32-43. 1987a. Implicaciones de las Secuencias Locales y Regionales en los Aspectos Culturales de los Tairona. In Chiefdoms in the Americas, edited by R. D. Drennan and C. A. Uribe, pp. 213—229. University Press of America, Lanham. 1987b. Gaira: Una Introduccion a la Ecologia y Arqueologia del Litoral de la Sierra Nevada de Santa Marca. Boletin del Museo del Oro, 19: 35-55. Palop Marrnez, Josefina. 1986. Los Cayapas en el Siglo XVI. Misceldnea Antropologica Ecuatoriana, 6: 231-252. Parker, Ronald B. and Heinrich Toots. 1980. Trace Elements in Bones as Paleobiological Indicators. In Fossils in the Making. Vertebrate Taphonomy and Paleoecology, edited by Anna K. Bchrensmeyer and Andrew P. Hill, pp. 197-207. University of Chicago Press, Chicago. Parkington, J. 1991. Approaches to Dietary Reconstruction in the Western Cape: Are You What You Have Eaten? Journal of Archaeological Science, 18: 331—342. Parra, R. 1978. Comparison of Foregut and Hindgut Fermentation in Herbivores. In The Ecology of Arboreal Folivores, edited by G. G. Montgomery, pp. 205-229. Smithsonian Institution Press, Washington D.C. Parsons, James J. 1977. Pre-Columbian Ridged Fields in Colombia and Ecuador. University of California, Berkeley. 1980. Europeanization of the Savana Lands of Northern South America. In Human Ecology in Savana Environments, edited by D. R. Harris, pp. 267-289. Academic Press, London. Parsons, James J. and William A. Bo wen. 1966. Ancient Ridged Fields in the San Jorge Floodplain, Colombia. The Geographical Review, 56: 317—343. Parsons, James and William Denevan. 1967. Pre-Columbian Ridged Fields in New World Archaeology. Scientific American, 217(1): 92—100. Parsons, J., E. Brumfiel, M. Parsons, V. Popper, and M. Taft. 1985. Chinampa Agriculture and Aztec Urbanism in the Valley of Mexico. In Prehistoric Intensive Agriculture in the Tropics, edited by Ian Farrington, pp. 49-96. British Archaeological Reports, International Series 232(1) and (ii), Oxford. Paulsen, Allison C. 1976. Environment and Empire: Climatic Factors in Prehistoric Andean Culture Change. World Archaeology, 8: 121-132. 1981. The Archaeology of the Absurd: Comments on "Cultural Materialism, Split Inheritance and the Expansion of Ancient Peruvian Empires". American Antiquity, 46: 31-37. Payne, Sebastian. 1972. Partial Recovery and Sample Bias: The Results of Some Sieving Experiments. In Papers in Economic Prehistory, edited by E. S. Higgs, pp. 49—64. Cambridge University Press, Cambridge. 1975. Partial Recovery and Sample Bias. In Archaeozoological Studies, edited by A. T. Clason, pp. 7-17. American Elsevier Publishing, New York. 1983. Bones from Cave Sites: Who Ate What? Problems and a Case Study. In Animals and Archaeology, 1: Hunters and Their Prey, edited by Juliet Clutton-Brock and Caroline Grigson, pp. 149-162. British Archaeological Reports, International Series 163, Oxford. Pearsall, D. M. 1978. Phytolith Analysis of Archaeological Soils: Evidence for Maize Cultivation in Formative Ecuador. Science, 199: 177—178.
288
References
1979. The Application of Ethnobotanical Techniques to the Problem of Subsistence in the Ecuadorian Formative. Unpublished Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana. 1985. Analysis of Soil Phytoliths and Botanical Macroremains from El Bronce Archaeological Site, Ponce, Puerto Rico. In Archaeological Data Recovery at El Bronce, Puerto Rico, Final Report, Phase 2, edited by L. S. Robinson, E. R. Lundberg, and J. B. Walker, pp. 131—146. Submitted to United States Army Corps of Engineers, Washington D.C. 1987a. Evidence for Prehistoric Maize Cultivation on Raised Fields at Penon del Rio, Guayas, Ecuador. In Pre-Hispanic Agricultural Fields in the Andean Region, edited by W. M. Denevan, K. Mathewson, and G. Knapp, pp. 279-295. British Archaeological Reports, International Series 359 (i), Oxford. 1987b. Phytolith Analysis of Soil Samples from Raised Fields in the Yumes area, Daule River Floodplain, Ecuador. Report on file, American Archaeology Division, University of Missouri, Columbia. 1988. An Overview of Formative Period Subsistence in Ecuador: Palaeoethnobotanical Data and Perspectives. In Diet and Subsistence: Current Archaeological Perspectives, edited by B. V. Kennedy and G. M. LeMoine, pp. 149—164. Proceedings of the Nineteenth Annual Chacmool Conference. Archaeological Association of the University of Calgary, Calgary. 1989. Paleoethnobotany. A Handbook of Procedures. Academic Press, San Diego. 1991. Prehistoric Subsistence and Agricultural Evolution in the Jama River Valley, Manabi Province, Ecuador. Paper Presented at the 56th Annual Meeting of the Society for American Archaeology, New Orleans. 1992. The Origins of Plant Cultivation in South America. In Origins of Agriculture. An International Perspective, edited by C. W. Cowan and P. J. Watson, pp. 173-205. Smithsonian Institution Press, Washington D.C 1993. Agricultural Evolution and the Emergence of Formative Societies in Ecuador. In Development of Agriculture and Emergence of Formative Civilizations in Pacific Central and South America, edited by M. Blake. Washington State University Press, Seattle, under review. Pearsall, D. M. and B. Ojeda. 1988. Final Report of Analysis of Botanical Materials from the El Paraiso Site, Peru. With a Study of Cucurbitaceae, by Andrea Hunter. MS in possession of the authors, American Archaeology Division, University of Missouri, Columbia. Pearsall, D. M. and D. R. Piperno (eds.). 1992. Current Research in Phytolith Analysis: Applications in Archaeology and Paleoecology. MASCA Research Papers in Science and Archaeology 10, University of Pennsylvania Museum, Philadelphia. Phelan, John. 1967. The Kingdom of Quito in the Seventeenth Century. University of Wisconsin Press, Quito. Pike, K. 1964. Towards a Theory of the Structure of Human Behavior. In Language in Culture and Society, edited by D. Hymes, pp. 54-62. Harper and Row, New York. Pinto Parada, Rodolfo. 1987. Pueblo de Leyenda. Editorial Tiempo del Beni, Trinidad. Piperno, D.R. 1985. Phytolith Analysis and Tropical Paleoecology: Production and Taxonomic Significance of Siliceous Forms in New World Plant Domesticates and Wild Species. Review of Paleobotany and Palynology, 45: 185-228. 1988a. Phytolith Analysis: An Archaeological and Geological Perspective. Academic Press, San Diego. 1988b. Primer Informe Sobre los Fitolitos de las Plantas del OGSE-80 y la Evidencia del Cultivo de Maiz en el Ecuador. In Prehistoria Temprana de la Peninsula de Santa Elena, Ecuador: Cultura Las Vegas, edited by K. E. Stothert, pp. 203—214. Serie Monografica 10. Museos del Banco Central del Ecuador, Guayaquil. 1989. The Occurrence of Phytoliths in the Reproductive Structures of Selected Tropical Angiosperms and their Significance in Tropical Paleoecology, Paleo-ethnobotany and Systematics. Review of Paleobotany and Palynology, 61: 147-173.
References
289
1990. Aboriginal Agriculture and Land Usage in the Amazon Basin, Ecuador, journal of Archaeological Science, 17: 665—677. 1991. The Status of Phytolith Analysis in the American Tropics. Journal of World Prehistory, 5: 155-191. 1992. The Origins and Development of Food Production in Pacific Panama. In The Development of Agriculture in Central and South America, edited by M. Blake. Washington State University Press, Olympia, under review. 1993a. Modern Phytolith Assemblages and Their Role in Paleobotanical Interpretation. MS in possession of author. 1993b. Phytolith Assemblages from Modern and Terrestrial Soil Profiles in the Central Amazon, Brazil. MS in possession of author, n.d. Phytolith and Charcoal Evidence for Prehistoric Slash and Burn Agriculture in the Darien Rain Forest of Panama. The Holocene, MS under review. Piperno, D. R. and K. Husum-Clary. 1984. Early Plant Use and Cultivation in the Santa Maria Basin, Panama: Data from Phytoliths and Pollen. In Recent Advances in Isthmian Archaeology, edited by F. Lange, pp. 85-121. British Archaeological Reports, International Series 212, Oxford. Piperno, D. R. and D. M. Pearsall. 1990. Antiquity of Maize Cultivation in Ecuador: Survey and Reevaluation of the Evidence. American Antiquity, 55: 324-337. 1993. The Silica Bodies of Tropical American Grasses: Morphology, Phylogeny, and Implications for the Systematics of the Gramineae. MS in possession of authors. Piperno, D. R., M. B. Bush, and P. A. Colinvaux. 1990. Paleoenvironments and Human Occupation in Late-Glacial Panama. Quaternary Research, 33: 108-116. 1991a. Paleoecological Perspectives on Human Adaptation in Central Panama. I. The Pleistocene. Geoarchaeology, 6: 201-226. 1991b. Paleoecological Perspectives on Human Adaptation in Central Panama. II. The Holocene. Geoarchaeology, 6: 227-250. Piperno, D. R., P. A. Colinvaux and M. B. Bush. 1992. The First Environmental, Cultural, and Forest History from the Darien, Panama. MS in possession of authors. Piperno, D. R., M. B. Bush, P. A. Colinvaux, and R. G. Cooke. 1988. Man-Land Relationships in the Humid Forests of Panama: The Last 13,000 Years. Paper presented at the 45th International Congress of Americanists, Amsterdam. Piperno, D. R., K. Husum-Clary, R. G. Cooke, A. J. Ranere, and D. Weiland. 1985. Preceramic Maize in Central Panama: Phytolith and Pollen Evidence. American Anthropologist, 87: 871-878. Plafker, George. 1963. Observations on Archaeological Remains in Northeastern Bolivia. American Antiquity, 28: 372-378. Plazas, Clemencia and Ana Maria Falchetti de Saenz. 1981. Asentamientos Prehispdnicos en el Bajo Rio San Jorge. Banco de la Repiiblica, Fundacion de Investigaciones Arqueologicas Nacionales, Bogota. 1986. La Cultura del Oro y el Agua. Un Proyecto de Reconstruccion. Boletin Cultural y Bibliogrdfico, 23 (6): 57-72. Biblioteca Luis Angel Arango, Bogota. 1987. Poblamiento y Adecuacion Hidraulica en el Bajo Rio San Jorge, Costa Atlantica, Colombia. In Pre-Hispanic Agricultural Fields in the Andean Region, edited by W. M. Denevan, K. Mathewson and G. Knapp, pp. 483-503. British Archaeological Reports International Series 359, Oxford. Plazas, Clemencia, Ana Maria Falchetti, Thomas van der Hammen, and Pedro Botero. 1988. Cambios Ambientales y Desarollo Cultural en el Bajo Rio San Jorge. Boletin del Museo del Oro, 20: 55-88. 1992. Environmental Changes and Cultural Development in the Lower San Jorge River, Colombia. In Archaeology and Environment in Latin America, edited by O. R. Ortiz-
290
References
Troncoso and T. van der Hammen, pp. 183—205. Instituut Voor Pre- en Protohistorische Archeologie 'Albert Egges van Giffen', Universiteit van Amsterdam, Amsterdam. Plog, Stephen. 1976. Relative Efficiencies of Sampling Techniques for Archaeological Surveys. In The Early Mesoamerican Village, edited by K. V. Flannery, pp. 136-158. Academic Press, New York. Plog, Stephen, Fred Plog, and W. Wait. 1978. Decision Making in Modern Surveys. In Advances in Archaeological Method and Theory, I, edited by M. B. Schiffer, pp. 383-421. Academic Press, New York. Plowman, T. 1984. The Origin, Evolution, and Diffusion of Coca, Erythroxylum spp., in South and Central America. In Pre-Columbian Plant Migration, edited by D. Stone, pp. 125-163. Harvard University Press, Cambridge, MA. Pohl, Mary (ed.). 1989. Ancient Maya Agriculture: Excavations on Albion Island, Northern Belize. Westview Press, Boulder, CO. Portugal Ortiz, Max. 1978. La Arqueologia de la Region del Beni. Editorial Casa Municipal de la Cultura "Franz Tamayo," La Paz. Posey, D. A. and W. Balee (eds.). 1989. Resource Management in Amazonia: Indigenous and Folk Strategies. Advances in Economic Botany, 7. Poyart, C. F., E. Burseaux, and A. Freminet. 1975. The Bone CO2 Compartment: Evidence for a Bicarbonate Pool. Respiratory Physiology, 25: 88-99. Prance, G. and J. Kallunki (eds.). 1984. Ethnobotany in the Neotropics. The New York Botanical Garden, Bronx, New York. Pratt, Ann E. 1989. Taphonomy of Microvertebrate Fauna From the Early Miocene Thomas Farm Locality, Florida (U.S.A). Palaeogeography, Palaeo climatology, Palaeoecology, 76: 125—151. Prentice, Guy. 1986. An Analysis of the Symbolism Expressed by the Birger Figurine. American Antiquity, 51: 239-266. PRONAREG (Ecuador). 1978. Mapa Ecologico. Programa Nacional de Regionalizacion Agraria. Ministerio de Agricultura y Ganaderia, Quito. Puleston, Dennis E. 1977. Experiments in Prehistoric Raised Field Agriculture: Learning from the Past. Journal ofBelizan Affairs, 5: 36-43. Quilter, J., B. Ojeda E., D. M. Pearsall, D. H. Sandweiss, J. G. Jones, and E. S. Wing. 1991. Subsistence Economy of El Paraiso, an Early Peruvian site. Science, 251: 277-283. Rainey, Froelich G. 1940. Porto Rican Archaeology. Scientific Survey of Porto Rico and the Virgin Islands, vol. 18 (1). New York Academy of Sciences, New York. 1941. Excavations in the Ft. Liberte Region, Haiti. Yale University Publications in Anthropology No. 23, New Haven. Ramos, Claudio. 1986. Evaluacion y Rehabilitacion de Camellones o "Kurus" en Asillo. Allpanchis, 27: 239-284. 1990. Rehabilitacion, Uso y Manejo de Camellones: Propuesta Tecnica. Programa de Rehabilitacion y Uso de Waru-Waru, Centro Canadense de Estudios y Cooperation InternacionalCentro de Communicacion Audiovisual para la Education Popular, Lima. Ranere, A. J. 1989. Implements of Change in the Holocene Environments of Panama. Paper Presented at the 45th Annual Meeting of the Society for American Archaeology, Pittsburgh. 1993a. The Manufacture and Use of Stone Tools during the Preceramic in the Santa Maria River Basin of Central Panama. In Cazadores y Recolectores in Centro- y Sudamerica, edited by G. Correal and R. G. Cooke. British Archaeological Reports, Oxford, in press. 1993b. New Excavation at and Re-evaluation of Cerro Mangote, a Late Preceramic Shellmound on the Pacific Coast of Central Panama. In Cazadores y Recolectores in Centro- y Sudamerica, edited by G. Correal and R. G. Cooke. British Archaeological Reports, Oxford, in press. Ranere, A. J. and R. G. Cooke. 1991. Paleoindian Occupation in the Central American Tropics. In Clovis: Origins and Human Adaptation, edited by R. Bonnichsen and K. Fladmark, pp. 237253. Peopling of the Americas. Center for the Study of the First Americans, Department of Anthropology, Oregon State University, Corvallis.
References
291
Ranere, A. J. and P. Hansell. 1978. Early Subsistence Patterns along the Pacific Coast of Panama. In Prehistoric Coastal Adaptations: The Economy and Ecology of Maritime Middle America, edited by B. Stark and B. Voorhies, pp. 43-59. Academic Press, New York. Ranere, A. J., R. G. Cooke, and P. Hansell. 1980. Food Procurement in the Parita Bay Region of Panama, 5000B.C—A.D.500. Paper Presented at the 45th Annual Meeting of the Society for American Archaeology, Philadelphia. Rapp, G., Jr. and S. Mulholland (eds.). 1992. Phytolith Systematics. Emerging Issues. Plenum Press, New York. Raymond, J. Scott. 1972. The Cultural Remains from the Granja de Sivia, Peru: An Archaeological Study of Tropical Forest Culture in the Montana. Unpublished Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana-Champaign. 1983. Quimpiri, a Ceramic Style from the Peruvian Montana. Nawpa Pacha, 20: 121-146. 1985. Quechuas y Chunchos: Ethnic Boundaries in Eastern Peru. In Status, Structure and Stratification: Current Archaeological Reconstructions, edited by M. Thompson, M. T. Garcia, and F. Kense, pp. 35-50. Archaeological Association of the University of Calgary, Calgary. Raymond, J. Scott, Warren R. DeBoer, and Peter G. Roe. 1975. Cumancaya: A Peruvian Ceramic Tradition. Occasional Papers No. 2, Department of Archaeology, University of Calgary, Calgary. Raymond, J. Scott, Augusto Oyuela Caycedo, and Patrick Carmichael. 1944. Una Comparacion de las Tecnologias de la Ceramica Temprana de Ecuador y Colombia. In Tecnologia y Organizacion de la Produccion de Ceramica Prehispanica en los Andes, edited by Izumi Shimada, pp. 33-52. Fondo Editorial de la Pontificia Universidad Catolica, Lima. Read, Dwight W. 1975. Regional Sampling. In Sampling in Archaeology, edited by J. W. Mueller, pp. 45-60. University of Arizona Press, Tucson. 1986. Sampling Procedures for Regional Surveys: A Problem of Representativeness and Effectiveness. Journal of Field Archaeology, 13: 477-491. Redman, Charles L. 1973. Multi-Stage Fieldwork and Analytical Techniques. American Antiquity, 38: 61-79. 1975. Productive Sampling Strategies for Archaeological Sites. In Sampling in Archaeology, edited by James W. Mueller, pp. 147-154. University of Arizona Press, Tucson. Redman, Charles L. and Patty Jo Watson. 1970. Systematic, Intensive Surface Collection. American Antiquity, 35: 279-291. Reichel-Dolmatoff, Gerardo. 1965. Colombia. Thames and Hudson, London. Reichel-Dolmatoff, Gerardo and Alicia Reichel-Dolmatoff. 1951. Investigaciones Arqueologicas en el Depto. del Magdalena, Colombia en 1946-1950. Parte I: Arqueologia del Rio Rancheria. Boletin de Arqueologia, 3(1-6): 19-208. 1957. Momil: A Formative Sequence from the Sinu Valley, Colombia. American Antiquity, 22: 226-234. Reid, J. Jefferson, Michael B. Schiffer, and Jeffrey M. Neff. 1975. Archaeological Considerations of Intrasite Sampling. In Sampling in Archaeology, edited by James W. Mueller, pp. 209-224. University of Arizona Press, Tucson. Reister, Jurgen. 1981. Arqueologia y Arte Rupestre en el Oriente Boliviano. Editorial Amigos del Libro, La Paz. Renfrew, J. M. 1973. Palaeoethnobotany. Columbia University Press, New York. Rice, Prudence M. 1987. Pottery Analysis, A Source Book. University of Chicago Press, Chicago. Riley, Thomas and Glen Freimuth. 1979. Field Systems and Frost Drainage in the Prehistoric Agriculture of the Upper Great Lakes. American Antiquity, 44: 271-285. Rindos, D. 1984. The Origins of Agriculture. An Evolutionary Perspective. Academic Press, Orlando. Riviere, Peter. 1969. Marriage Among the Trio. Clarendon Press, Oxford. Roe, Peter E. and Peter E. Siegel. 1982. The Life History of a Shipibo Compound: Ethnoarchaeology
292
References
in the Peruvian Montana. Journal of Archaeology and Anthropology, 5: 94—118. Roosevelt, Anna C. 1980. Parmana: Prehistoric Maize and Manioc Subsistence Along the Amazon and the Orinoco. Academic Press, New York. 1987. Chiefdoms in the Amazon and Orinoco. In Chiefdoms in the Americas, edited by R. D. Drennan and C. A. Uribe, pp. 153—184. University Press of America, Lanham, Maryland. 1989. Resource Management in Amazonia Before the Conquest: Beyond Ethnographic Projection. Advances in Economic Botany, 7: 30-62. 1991. Moundbuilders of the Amazon: Geophysical Archaeology on Marajo Island, Brazil. Academic Press, San Diego. Roosevelt, A. C , R. A. Housley, M. Imazio da Silveira, S. Maranca, and R. Johnson. 1991. Eighth Millennium Pottery from a Prehistoric Shell Midden in the Brazilian Amazon. Science 254: 1621—1624.
Roper, Donna C. 1976. Lateral Displacement of Artifacts Due to Plowing. American Antiquity, 41: 372.-375Ross, Eric Barry. 1978. Food Taboos, Diet, and Hunting Strategy: The Adaptation to Animals in Amazon Cultural Ecology. Current Anthropology, 19: 1-36 Rottlander, R. C. A. 1976. Variation in the Chemical Composition of Bone as an Indicator of Climatic Change. Journal of Archaeological Science, 3: 83-88. Roubik, D. W. and J. E. Moreno. 1991. Pollen and Spores ofBarro Colorado Island. Monographs in Systematic Botany, vol. 36. Missouri Botanical Garden, St. Louis. Rouse, Irving. 1939. Prehistory in Haiti. Yale University Publications in Anthropology No. 21. New Haven. 1952a. Porto Rican Prehistory: Introduction; Excavations in the West and North. Scientific Survey of Porto Rico and the Virgin Islands, vol. 18, Part 3. New York Academy of Sciences, New York. 1952b. Porto Rican Prehistory: Excavations in the Interior, South, and East; Chronological Implications. Scientific Survey of Porto Rico and the Virgin Islands, vol. 18, Part 4. New York Academy of Sciences, New York. 1974. The Indian Creek Excavations. Proceedings of the International Congress for the Study of Pre-Columbian Cultures of the Lesser Antilles, 5: 166-176. 1982. Ceramic and Religious Development in the Greater Antilles. Journal of New World Archaeology, 5(2): 45-55. 1989. Peoples and Cultures of the Saladoid Frontier in the Greater Antilles. In Early Ceramic Population Lifeways and Adaptive Strategies in the Caribbean, edited by Peter E. Siegel, pp. 383-403. British Archaeological Reports, International Series 506, Oxford. 1992. The Tainos: Rise and Decline of the People Who Greeted Columbus. Yale University Press, New Haven. Rouse, Irving and Ricardo E. Alegria. 1990. Excavations at Maria de la Cruz and Hacienda Grande Village Site, Loiza, Puerto Rico. Yale University Publications in Anthropology No. 80, New Haven. Rouse, Irving and Louis Allaire. 1978. Caribbean. In Chronologies in New World Archaeology, edited by R. E. Taylor and Clement W. Meighan, pp. 431-481. Academic Press, New York. Rouse, Irving and Jose M. Cruxent. 1963. Venezuelan Archaeology. Yale University Press, New Haven. Rowe, John H. 1959. Archaeological Dating and Cultural Process. Southwestern Journal of Anthropology, 5(4): 317-324. 1961. Stratigraphy and Seriation. American Antiquity, 26: 324-330. Ruddle, Kenneth. 1974. The Yukpa Cultivation System. A Study of Shifting Cultivation in Colombia and Venezuala. Ibero-Americana vol. 52. University of California Press, Berkeley. Rue, D. J. 1987. Early Agriculture and Early Postclassic Maya Occupation in Western Honduras. Nature, 326: 285-286.
References
293
1988. Archaic Middle American Agriculture and Settlement: Recent Pollen Data from Honduras. Journal of Field Archaeology, 16: 177-184. Rumazo, Jose (compilacion). 1948—50. Documentos para la Historia de la Audiencia de Quito. Pedro Vicente Maldonado: Tomos I—VII. Alfrodisio Aguado, Madrid. Ryden, Stig. 1941. A Study of the Siriono Indians. Elanders Boktryckeri Aktiebolag, Goteborg. Rye, O. S. 1981. Pottery Technology: Principles and Reconstruction. Taraxacum, Washington D.C. Saffray, D. 1876. Les Antiquites Peruviennes a l'Exposition de Philadelphia. La Nature (Paris) 4: 401-407 (cited in Renfrew 1973). Saldarriaga, J. G. and D. C. West. 1986. Holocene Fires in the Northern Amazon Basin. Quaternary Research, 26: 358-366. Salomons, J. B. 1986. Palaeoecology of Volcanic Soils in the Colombian Central Cordillera (Parque Nacional Natural de los Nevados). Dissertationes Botanicae, Band 95. J. Cramer, Berlin-Stuttgart. Sanchez, Julia L. and Brian S. Shaffer. 1991. Biases of|and |" Screening on Controlled Samples of Vertebrate Skeletons. Paper Presented at the 56th Annual Meeting of the Society for American Archaeology, New Orleans. Sanford, R. L., Jr., J. Saldarriaga, K. E. Clark, C. Uhl, and R. Herrera. 1985. Amazon Rain-forest Fires. Science, 227: 53-55. Sapir, Edward. 1951. The Psychological Reality of Phonemes. In Selected Writings ofEdward Sapir, edited by D. C. Mandelbaum, pp. 46-60. University of California Press, Berkeley. Sauer, C. O. 1952. Agricultural Origins and Dispersals. The American Geographical Society, New York. Saville, M. H. 1913. Pre-Columbian Decoration of the Teeth in Ecuador, with Some Occurrence of the Custom in Other Parts of North and South America. American Anthropologist, 15: 377-394Scheaffer, Richard L., William Mendenhall, and Lyman Ott. 1986. Elementary Survey Sampling. 3rd edn. Duxbury Press, Boston. Schiffer, Michael B. 1972. Archaeological Context and Systemic Context. American Antiquity, 37: 156-165. 1976. Behavioral Archaeology. Academic Press, New York. 1987. Formation Processes of the Archaeological Record. University of New Mexico Press, Albuquerque. Schiffer, Michael B., and Susan J. Wells. 1982. Archaeological Surveys: Past and Future. In Hohokam and Patayan: Prehistory of Southwestern Arizona, edited by R. H. McGuire and M. B. Schiffer, pp. 345-383. Academic Press, New York. Schiffer, Michael B., Alan P. Sullivan III, and Timothy C. Klinger. 1978. The Design of Archaeological Surveys. World Archaeology, 10: 1-28. Schoeninger, M. J. 1985. Trophic Level Effects on 15N/14N and 13C/12C Ratios in Bone Collagen and Strontium Levels in Bone Mineral. Journal of Human Evolution, 14: 515—525. 1989. Reconstructing Prehistoric Human Diet. In The Chemistry of Prehistoric Human Bone, edited by T. D. Price, pp. 38-67. Cambridge University Press, Cambridge. Schoeninger, M. J. and M. J. DeNiro. 1982. Carbon Isotope Ratios of Apatite from Fossil Bone Cannot be Used to Reconstruct Diets of Animals. Nature, 297: 577-578. 1983. Reply to: Carbon Isotope Ratios of Bone Apatite and Animal Diet Reconstruction. Nature, 301: 177-178. 1984. Nitrogen and Carbon Isotopic Composition of Bone Collagen from Marine and Terrestrial Animals. Geochimica et Cosmochimica Acta, 48: 625-639. Schoeninger, M. J., M. J. DeNiro, and H. Tauber. 1983. Stable Nitrogen Isotope Ratios of Bone Collagen Reflect Marine and Terrestrial Components of Prehistoric Human Diet. Science, 220: 1381-1383.
294
References
Schoeninger, M. J., K. M. Moore, M. L. Murray, and J. D. Kingston. 1989. Detection of Bone Preservation in Archaeological and Fossil Samples. Applied Geochemistry, 4: 281-292. Schreve-Brinkman, Elisabeth J. 1978. A Palynological Study of the Upper Quaternary Sequence in the El Abra Corridor and Rock Shelters (Colombia). Palaeogeography, Palaeo climatology, Palaeoecology, 25: 1-109. Schwarcz, H. P. 1991. Some Theoretical Aspects of Isotope Paleodiet Studies. Journal of Archaeological Science, 18: 261-275. Schwarcz, H. P., J. Melbye, M. A. Katzenberg, and M. Knyf. 1985. Stable Isotopes in Human Skeletons of Southern Ontario: Reconstructing Paleodiet. Journal of Archaeological Science, 12: 187-206. Sealy, J. 1986. Stable Carbon Isotopes and Prehistoric Diets in the South-Western Cape Province, South Africa. British Archaeological Reports, International Series 293, Oxford. 1989. The Use of Chemical Techniques for Reconstructing Prehistoric Diets: A Case Study in the South-Western Cape. South African Archaeological Society, Goodwin Series, 6: 69—76. Sealy, J. and N. J. van der Merwe. 1985. Isotope Assessment of Holocene Human Diets in the Southwestern Cape, South Africa. Nature, 315: 138-140. Sealy, J., N. J, van der Merwe, J. A. Lee-Thorp, and J. L. Lanham. 1987. Nitrogen Isotope Ecology in Southern Africa: Implications for Environmental and Dietary Tracing. Geochimica et Cosmochimica Ada, 51: 2707-2717. Semken, Holmes A., Jr. and Carl L. Falk. 1991. Micromammalian Taphonomy of Three Late Prehistoric Plains Village Tradition Refuse Pits, Walworth County, South Dakota. In Beamers, Bobwhites and Blue-Points: Tributes to the Career of Paul W. Parmalee, edited by James R. Purdue, Walter E. Klippel and Bonnie W. Styles, pp. 111-124. Scientific Papers vol. 23. Illinois State Museum, Springfield. Setz, Eleonore Zulnara Freire. 1991. Animals in the Nambiquara Diet: Methods of Collecting and Processing. Journal of Ethnobiology, 11: 1-22. Sever, T. and J. Wiseman. 1985. Remote Sensing and Archaeology: Potential for the Future. National Aeronautics and Space Administration, Earth Resources Laboratory, John C. Stennis Space Center, Mississippi. Shaffer, Brian S. 1992a. Quarter-Inch Screening: Understanding Biases in Recovery of Vertebrate Faunal Remains. American Antiquity, 57: 129-136. 1992b. Interpretation of Gopher Remains from Southwestern Archaeological Assemblages. American Antiquity, 57: 683-691. Shearer, G. and D. H. Kohl. 1986. N2-Fixation in Field Settings: Estimations Based on Natural 15N Abundance. Australian Journal of Plant Physiology, 13: 699-756. Shimada, Izumi, Crystal Barker-Schaaf, Lonnie G. Thompson and Ellen Moseley-Thompson. 1991. Cultural Impacts of Severe Droughts in the Prehistoric Andes: Application of a 1,500-year Ice Core Precipitation Record. World Archaeology, 22: 247-270. Shott, Michael J. 1985. Shovel-Test Sampling as a Site Discovery Technique: A Case Study from Michigan. Journal of Field Archaeology, 12: 457—468. 1987. Feature Discovery and the Sampling Requirements of Archaeological Evaluations. Journal of Field Archaeology, 14: 359-371. Siegel, Peter E. 1976. The Making of a Midden: Ethnoarchaeology in the Peruvian Montana. MS on file, Department of Anthropology, University of Delaware, Newark. 1985. Waiwai Mapping Project and Ethnoarchaeology. MS on file, Centro de Investigaciones Indigenas de Puerto Rico, Viejo San Juan, Puerto Rico. 1986. The Maisabel Project Excavation. Arqueologia, 2(1): 20-21. 1988. Occupational History of an Early Ceramic Age Settlement on the North Coast of Puerto Rico. Proposal submitted to the Anthropology Program, National Science Foundation, Washington D.C.
References
295
1989. Site Structure, Demography, and Social Organization in the Early Ceramic Age of the Caribbean. In Early Ceramic Population Lifeways and Adaptive Strategies in the Caribbean, edited by Peter E. Siegel, pp. 193-245. British Archaeological Reports International Series 506, Oxford. 1990. Ethnoarchaeological Study of a Waiwai Village in the Tropical Forest of Southern Guyana. Proceedings of the International Association for Caribbean Archaeology, 11: 400—417. 1991a. Migration Research in Saladoid Archaeology: A Review. The Florida Anthropologist, 44: 79-91. 1991b. On the Antillean Connection as a Potential Corridor for Cultigens into Eastern North America. Current Anthropology, 32: 332-334. 1992. Ideology, Power, and Social Complexity in Prehistoric Puerto Rico. Unpublished Ph.D. dissertation, Department of Anthropology, State University of New York at Binghamton, Binghamton. Siegel, Peter E. and Peter G. Roe. 1986. Shipibo Archaeo-Ethnography: Site Formation Processes and Archaeological Interpretation. World Archaeology, 18: 96-115. Siegel, Peter E. and Kenneth P. Severin. 1993. The First Documented Prehistoric Gold-Copper Alloy Artefact from the West Indies. Journal of Archaeological Science, 20: 67—79. Siemans, Alfred. 1989. Tierra Configurada. Consejo para la Cultura y las Artes, Mexico City. Siemans, Alfred and Dennis Puleston. 1972. Ridged Fields and Associated Features in Southern Campeche: New Perspectives on the Lowland Maya. American Antiquity, 37: 228-239. Sillen, Andrew. 1989. Diagenesis of the Inorganic Phase of Cortical Bone. In The Chemistry of Prehistoric Human Bone, edited by T. Douglas Price, pp. 211-229. Cambridge University Press, Cambridge. Sillen, A. and R. LeGros. 1991. Solubility Profiles of Synthetic Apatites and of Modern and Fossil Bones, journal of Archaeological Science, 18: 385-397. Sillen, A., J. C. Sealy, and N. J. van der Merwe. 1989. Chemistry and Paleodiet Research: No More Easy Answers. American Antiquity, 54: 504-512. Silva, Alcionilio Briizzi Alves da. 1962. A Civilizaqao Indigena do Vaupes. Centro de Pesquias de Iavarete, Sao Paulo. Simonetti, Javier A. and Kuis E. Cornejo. 1991. Archaeological Evidence of Rodent Consumption in Central Chile. Latin American Antiquity, 2: 92-96. Siskind, Janet. 1973. To Hunt in the Morning. Oxford University Press, Oxford. Smith, B. N. 1972. Natural Abundance of the Stable Isotopes of Carbon in Biological Systems. BioScience, 22: 226-231. Smith, B. N. and S. Epstein. 1971. Two Categories of 13C/12C Ratios for Higher Plants. Plant Physiology, 47: 380-384. Smith, B. N., J. Oliver, and C. McMillan. 1976. Influence of Carbon Source, Oxygen Concentration, Light Intensity, and Temperature on 13C/12C Ratios in Plants. Botanical Gazette, 137: 99-104. Smith, B. N., C. B. Otto, G. E. Martin, and T. W. Boutton. 1979. Photosynthetic Strategies of Plants. In Arid Land Plant Resources, edited by J. R. Goodin and D. K. Northington, pp. 474-481. Texas Technical University Press, Lubbock. Smith, C. E. 1980. Plant Remains from the Chiriqui Sites and Ancient Vegetational Patterns. In Adaptive Radiations in Prehistoric Panama, V, edited by Olga Linares and A. J. Ranere, pp. 151-174. Peabody Museum Monographs, Harvard University, Cambridge, MA. Smith, Clifford T., William M. Denevan, and Patrick Hamilton. 1968. Ancient Ridged Fields in the Region of Lake Titicaca. The Geographical journal, 134: 353-367. Smith, Nigel. 1980. Anthrosols and Human Carrying Capacity in Amazonia. Annals of the Association of American Geographers, 70: 553—566. Smole, William J. 1976. The Yanoama Indians. A Cultural Geography. University of Texas Press, Austin.
296
References
Sombroek, W. G. 1984. Soils of the Amazon Region. In The Amazon, edited by H. Sioli, pp. 521535. Dr. W. Junk Publishers, Dordrecht. Southgate, Douglas and Morris Whitaker. 1992. Promoting Resource Degradation in Latin America: Tropical Deforestation, Soil Erosion, and Coastal Ecosystem Disturbance in Ecuador. Economic Development and Cultural Change, 40: 787-807. Sowles, Lyle K. 1984. The Peccaries. University of Arizona Press, Tucson. Spaulding, Albert C. 1953. Statistical Techniques for the Discovery of Artifact Types. American Antiquity, 18: 305-313. 1954a. Reply. American Anthropologist, 56: 112-114. 1954b. Reply to Ford. American Antiquity, 19: 391-393. i960. The Dimensions of Archeology. In Essays in the Science of Culture, edited by G. E. Dole and R. L. Carneiro, pp. 437-456. Thomas Y. Crowell Co., New York. 1982. Structure in Archaeological Data: Nominal Variables. In Essays on Archaeological Typology, edited by R. Whallon and J. Brown, pp. 1-20. Center for American Archaeology Press, Evanston. Spielmann, K. A., M. J. Schoeninger, and K. Moore. 1990. Plains-Pueblo Interdependence and Human Diet at Pecos Pueblo, New Mexico. American Antiquity, 55: 745—765. Sponsel, Leslie E. 1986. Amazon Ecology and Adaptation. Annual Review of Anthropology, 15: 67-97. Spurling, Brian E. 1980. Site Discovery and Assessment Techniques for Mixed-Cover Survey Regimes. Saskatchewan Archaeology, 1: 25-56. Stafford, T. W., K. Brendel, and R. C. Duhamel. 1988. Radiocarbon, 13C, and 15N Analysis of Fossil Bone: Removal of Humates with XAD-2 Resin. Geochimica et Cosmochimica Acta, 52: 2257-2267. Stahl, Peter W. 1984. Tropical Forest Cosmology: The Cultural Context of the Early Valdivia Occupations at Loma Alta. Unpublished Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana. 1986. Hallucinatory Imagery and the Origin of Early South American Figurine Art. World Archaeology, 18: 134-150. 1991. Arid Landscapes and Environmental Transformations in Ancient Southwestern Ecuador. World Archaeology, 22: 346—359. 1992a. Encountering Paradise Lost: The Archaeological Recovery of Ancient Biodiversity in the Lowlands of Western Ecuador. Paper Presented at the 91st Annual Meeting of the American Anthropological Association, San Francisco. 1992b. Diversity, Body Size, and the Archaeological Recovery of Mammalian Faunas in the Neotropical Forests. Journal of the Steward Anthropological Society, 20: 209-233. 1994. Archaeologically Recovered Faunal Taxa from the Jama Valley. In Regional Archaeology in Northern Manabi, Ecuador, I: Environment, Cultural Chronology, and Prehistoric Subsistence in the jama River Valley, edited by James A. Zeidler and Deborah M. Pearsall. University of Pittsburgh, Memoirs in Latin American Archaeology No. 8, Pittsburgh, pp. 185-199. Stahl, Peter W. and James A. Zeidler. 1990. Differential Bone-Refuse Accumulation in FoodPreparation and Traffic Areas on an Early Ecuadorian House Floor. Latin American Antiquity, 1: 150-169. Stemper, David. 1987. Raised Fields and Agricultural Production, A.D. 1400-1600, Rio Daule, Guayas, Ecuador. In Pre-Hispanic Agricultural Fields in the Andean Region, edited by W. M. Denevan, K. Mathewson, and G. Knapp, pp. 297-319. British Archaeological Reports, International Series 359(1) and (ii), Oxford. Sternberg, L. S. L., S. S. Mulkey, and J. S. Wright. 1989. Ecological Interpretations of Leaf Carbon Isotope Ratios: Influence of Respired Carbon Dioxide. Ecology, 70: 1317-1324.
References
2.97
Stevenson, Marc G. 1982. Toward an Understanding of Site Abandonment Behavior: Evidence from Historic Mining Camps in the Southwest Yukon, journal of Anthropological Archaeology, 1: 237-265. Stevenson, W. B. 1825. A Historical and Descriptive Narrative of Twenty Years' Residence in South America. 3 vols. Hurst Robinson, London. Steward, J. H. and L. C. Faron. 1959. Native Peoples of South America. McGraw-Hill, New York. Stewart, Frances L. 1991. Floating for Fauna: Some Methodological Considerations Using the Keffer (AkGv-14) Midden 57 Faunal Sample. Canadian Journal of Archaeology, 15: 97-115. Stothert, K.E. 1985. The Preceramic Las Vegas Culture of Coastal Ecuador. American Antiquity, 50: 613-637. Stothert, K.E. (ed.). 1988. Prehistoria Temprana de la Peninsula de Santa Elena, Ecuador: Cultura Las Vegas. Miscelanea Antropologica Ecuatoriana. Serie Manografica 10. Museos del Banco Central del Ecuador, Guayaquil. Struever, Stuart. 1968. Flotation Techniques for the Recovery of Small-Scale Archaeological Remains. American Antiquity, 33: 353-362. Stuart-Macadam, P. 1985. Porotic Hyperostosis: Representative of a Childhood Condition. American Journal of Physical Anthropology, 66: 391-398. 1987a. A Radiographic Study of Porotic Hyperostosis. American Journal of Physical Anthropology, 74: 511-520. 1987b. Porotic Hyperostosis: New Evidence to Support the Anemia Theory. American Journal of Physical Anthropology, 74: 521-526. Sullivan, C. H. and H. W. Krueger. 1981. Carbon Isotope Analysis in Separate Chemical Phases in Modern and Fossil Bone. Nature, 292: 333-335. 1983. Carbon Isotope Ratios of Bone Apatite and Animal Diet Reconstruction. Nature, 301: 177. Szarek, S. R. and I. P. Ting. 1977. The Occurrence of Crassulacean Acid Metabolism among Plants. Photosynthetica, 11: 330-342. Szuter, Christine R. 1984. Faunal Exploitation and the Reliance on Small Animals Among the Hohokam. In Hohokam Archaeology Along the Salt-Gila Aqueduct Central Arizona Project, VII: Environment and Subsistence, edited by L. S. Teague and P. L. Crown, pp. 139-169. Arizona Archaeological Series vol. 150. Arizona State Museum, Tucson. 1989. Hunting by Prehistoric Horticulturists in the American Southwest. Unpublished Ph.D. dissertation, Department of Anthropology, University of Arizona, Tucson. Tartusi, Martha R. A., Antonio J. Nino, and Victor A. Nunez Regueiro. 1984. Relaciones entre el Area Occidental de la Cuenca del Lago de Maracaibo con las Areas Vecinas. In Relaciones Prehispdnicas de Venezuala, edited by E. Wagner, pp. 67-81. Fondo Editorial Acta Cientificas Venezolana, Caracas.. Tauber, H. 1981. 813C Evidence for Dietary Habits of Prehistoric Man in Denmark. Nature, 292: 332.-333Taylor, Walter W. 1967. A Study of Archaeology. Southern Illinois University Press, Carbondale. Teran de Rodriguez, P. 1988. Estudio de Investigacion Arquelogica, Convento de San Francisco de Quito, Sitio: OPQSF-2. Report Presented to Instituto de Cooperacion Ibero Americana de Espana. Quito. Thomas, David H. 1969. Great Basin Hunting Patterns: A Quantitative Method For Treating Faunal Remains. American Antiquity, 34: 393-401. 1971. On Distinguishing Natural From Cultural Bone in Archaeological Sites. American Antiquity, 36: 366-371. Thompson, L. G., E. Moseley-Thompson, J. F. Bodzan, and B. R. Koci. 1985. A 1500-year Record of Tropical Precipitation in the Ice Cores from the Quelccaya Ice Cap, Peru. Science, 229: 971-973.
References Tieszen, L. L. 1991. Natural Variations in the Carbon Isotope Values of Plants: Implications for Archaeology, Ecology and Paleoecology. Journal of Archaeological Science, 18: 227-248. Tieszen, L. L., T. W. Boutton, K. G. Tesdahl, and N. A. Slade. 1983. Fractionation and Turnover of Stable Carbon Isotopes in Animal Tissues: Implications for the S13C Analysis of Diet. Oecologia, 57: 32-37. Tieszen, L. L., M. M. Senyimba, S. K. Imbamba, and J. H. Troughton. 1979. The Distribution of C3 and C4 Grasses and Carbon Isotope Discrimination along an Altitudinal and Moisture Gradient in Kenya. Oecologia, 37: 337-350. Tolstoy, Paul and Warren R. DeBoer. 1989. An Archaeological Sequence for the Santiago-Cayapas River Basin, Esmeraldas, Ecuador. Journal of Field Archaeology, 16: 295—308. Tomlinson, P. B. 1961. Anatomy of the Monocotyledons. II, Palmae. Oxford University Press, Oxford. 1969. Anatomy of the Monocotyledons. Ill, Commelinales-Zingiberales. Oxford University Press, Oxford. Torres-Gonzalez, Arturo and Jose R. Diaz. 1984. Water Resources of the Sabana Seca to Vega Baja Area, Puerto Rico. Water Resources Investigations Report 82—4115. United States Geological Survey, San Juan, Puerto Rico. Towle, M. 1961. The Ethnobotany of Pre-Columbian Peru. Aldine, Chicago. Trotter, M. 1970. Estimation of Stature from Intact Long Limb Bones. In Personal Identification in Mass Disasters, edited by T. D. Stewart, pp. 71-83. Smithsonian Institution, Washington DC. Troughton, J. H., P. V. Wells, and H. A. Mooney. 1974. Photosynthetic Mechanisms and Paleoecology from Carbon Isotope Ratios in Ancient Specimens of C4 and CAM Plants. Science, 185: 610-612. Turner, B. L. II and Peter D. Harrison (eds.). 1983. Pulltrouser Swamp: Ancient Maya Habitat, Agriculture, and Settlement in Northern Belize. University of Texas Press, Austin. Turner, B.L. II and C. Miksicek. 1984. Economic Plant Species Associated with Prehistoric Agriculture in the Mayan Lowlands. Economic Botany, 38: 179-193. Turner, C.G. II. 1978. Dental Caries and Early Ecuadorian Agriculture. American Antiquity, 43: 694-697. Tuross, N., M. L. Fogel, and P. E. Hare. 1988. Variability in the Preservation of the Isotopic Composition of Collagen from Fossil Bone. Geochimica et Cosmochimica Ada, 52: 929-935. Ubelaker, D. H. 1974. Reconstruction of Demographic Profiles from Ossuary Skeletal Samples: A Case Study from the Tidewater Potomac. Smithsonian Contributions to Anthropology, vol. 18. Smithsonian Institution Press, Washington D.C. 1978. Microscopic Analysis of Archaeological Bone. In Ancient Disease in the Midwest, edited by D. Morse, pp. 66-72. Reports of Investigations, vol. 15. Illinois State Museum, Springfield. 1980a. Human Skeletal Remains from Site OGSE-80, a Pre-ceramic Site on the Sta. Elena Peninsula, Coastal Ecuador. Journal of the Washington Academy of Sciences, 70: 3-24. 1980b. Prehistoric Human Remains from the Cotocollao Site, Pichincha Province, Ecuador. Journal of the Washington Academy of Sciences, 70: 59—74. 1981a. The Ayaldn Cemetery: A Late Integration Period Burial Site on the South Coast of Ecuador. Smithsonian Contributions to Anthropology, vol. 29. Smithsonian Institution Press, Washington D.C. 1981b. Approaches to Demographic Problems in the Northeast. In Foundations of Northeast Archaeology, edited by D. R. Snow, pp. 175—194. Academic Press, New York. 1983a. Prehistoric Demography of Coastal Ecuador. National Geographic Society Research Reports, 15: 695-703. 1983b. Human Skeletal Remains from OGSE-MA-172, an Early Guangala Cemetery Site on the Coast of Ecuador. Journal of the Washington Academy of Sciences, jy. 16-26. 1984. Prehistoric Human Biology of Ecuador: Possible Temporal Trends and Cultural Correla-
References
299
tions. In Paleopathology at the Origins of Agriculture, edited by M. N. Cohen and G. J. Armelagos, pp. 491-513. Academic Press, New York. 1987a. Estimating Age at Death from Immature Human Skeletons: An Overview. Journal of Forensic Sciences, 32: 1254-1263. 1987b. Dental Alteration in Prehistoric Ecuador, A New Example from Jama-Coaque. Journal of the Washington Academy of Sciences, 77: 76-80. 1988a. Skeletal Biology of Prehistoric Ecuador: An Ongoing Research Program. Journal of the Washington Academy of Sciences, 78: 1-2. 1988b. Restos de Esqueletos Humanos del Sitio OGSE-80. In Prehistoria Temprana de la Peninsula de Santa Elena, Ecuador: Cultura Las Vegas, edited by K. E. Stothert, pp. 105-132. Museos del Banco Central del Ecuador, Guayaquil. 1988c. Human Remains from OGSE-46, La Libertad, Guayas Province, Ecuador. Journal of the Washington Academy of Sciences, 78: 3-16. i988d. Prehistoric Human Biology at La Tolita, Ecuador, a Preliminary Report. Journal of the Washington Academy of Sciences, 78: 23-37. 1988c Restos Humanos Prehistoricos del Sitio Cotocollao, Provincia del Pichincha, Ecuador. In Cotocollao: Una Aldea Formativa del Valle de Quito, edited by M. Villalba, pp. 557-571. Miscelania Antropologica Ecuatoriana, Serie Monografica 2. Museos del Banco Central del Ecuador, Quito. i988f. Porotic Hyperostosis in Prehistoric Ecuador (abstract). Collegium Antropologicum 12, Supplement 34. Paper Presented at the XII International Congress of Anthropological and Ethnological Sciences, Zagreb. 1989a. Human Skeletal Remains. Excavation, Analysis, Interpretation, 2nd edn. Taraxacum, Washington. 1989b. The Estimation of Age at Death from Immature Human Bone. In Age Markers in the Human Skeleton, edited by M. Y. Iscan, pp. 55-70. Charles C. Thomas, Springfield. 1990a. Human Skeletal Remains from "Jardin del Este," Cumbaya, Pichincha, Ecuador. In La Preservation y Promocion del Patrimonio Cultural del Ecuador, pp. 22—52. Cooperacion Tecnica Ecuatoriana-Belga No. 4. Instituto Nacional de Patrimonio Cultural, Quito. 1990b. Porotic Hyperostosis, Parasitism and Sedentism in Ancient Ecuador (abstract). American Journal of Physical Anthropology, 81: 309-310. 1991. Ecology of Porotic Hyperostosis in Ancient Ecuador. American Journal of Physical Anthropology, Supplement 12: 176. 1992. Porotic Hyperostosis in Prehistoric Ecuador. In Diet, Demography, and Disease: Changing Perspectives on Anemia, edited by P. Stuart-Macadam and S. Kent, pp. 201-217. Aldine de Gruyter, New York. 1993a. Human Remains from Santo Domingo, Quito, Ecuador. Submitted for publication to the "Proyecto de Cooperacion Tecnica Ecuatoriano-Belga 'Preservacion y Promocion del Patrimonio Cultural del Ecuador'." Quito. 1993b. Human Skeletal Biology of Convento de San Francisco de Quito. Report on file, Instituto Nacional de Patrimonio Cultural del Ecuador (INPC) and the Agencia Espaiiola de Cooperacion Iberoamerican (AECI), Quito. Ubelaker, D. H. and C. Ripley. 1993. The Ossuary of San Francisco Church, Quito, Ecuador: Human Skeletal Biology. MS in possession of the authors. Ugent, D., S. Pozorski, and T. Pozorski. 1986. Archaeological Manioc (Manihot) from Coastal Peru. Economic Botany 40: 78-102. Uhl, C , H. Clark, and K. Clark. 1982. Successional Patterns Associated with Slash-and-burn Agriculture in the Upper Rio Negro Region of the Amazon Basin. Biotropica, 14: 249—254. Uhl, C , J. Boone Kauffman, and D. L. Cummings. 1988. Fire in the Amazon 2: Environmental Conditions Necessary for Forest Fires in the Evergreen Rainforest of Venezuela. Oikos, 53: 176-184.
300
References
Uhl, C , K. Clark, H. Clark, and P. Murphy. 1982. Early Plant Succesion After Cutting and Burning in the Upper Rio Negro Region of the Amazon Basin. Journal of Ecology, 69: 631-649. Umlauf, M. 1988. Paleoethnobotanical Investigations at the Initial Period Site of Cardal, Peru. Unpublished Masters Thesis, Department of Anthropology, University of Missouri, Columbia. Valdez, F. 1987. Proyecto Arqueologico "La Tolita" (1983-1986). Museos del Banco Central, Quito, van der Hammen, T. 1963. A Palynological Study on the Quaternary of British Guiana. Leidse Geologische Mededelingen, 29: 125—180. 1966. The Pliocene and Quaternary of the Sabana de Bogota (The Tilata and Sabana Formations). Geologie en Mijnbouw, 45: 102-109. 1979. Historia y Tolerancia de Ecosistemas Parameros. In El Medio Ambiente Paramo, edited by M. L. Salgado-Labouriau, pp. 55-66. Ediciones Centro de Estudios Avanzados, Caracas. 1981. The Pleistocene Changes of Vegetation and Climate in the Northern Andes. In The Glaciation of the Ecuadorian Andes, edited by S. Hastenrath, pp. 125-145. A. A. Balkema, Rotterdam. 1986a. Fluctuaciones Holocenicas del Nivel de Inundaciones en la Cuenca del Bajo MagdalenaCauca—San Jorge (Colombia). Geologia Norandina, 10: 11—18. Bogota. 1986b. Datos Sobre la Historia de Clima, Vegetacion y Glaciacion de la Sierra Nevada de Santa Marta. In Studies in Tropical Andean Ecosystems, II, edited by T. van der Hammen and P. M. Ruiz, pp. 561-580. J. Cramer, Berlin-Stuttgart. van der Hammen, T. and E. Gonzalez, i960. Upper Pleistocene and Holocene Climate and Vegetation of the Sabana de Bogota. Leidse Geologische Mededelingen, 25: 261-315. 1964. A Pollen Diagram from the Quaternary of the Sabana de Bogota (Colombia) and its Significance for the Geology of the Northern Andes. Geologie en Mijnbouw, 43: 113—117. 1965. A Pollen Diagram for Laguna del la Herrera (Sabana de Bogota). Leidse Geologische Mededelingen, 32: 183-191. van der Hammen, T., J. H. Werner, and H. V. Dommelen. 1973. Palynological Record of the Upheaval of the Northern Andes: A Study of the Pliocene and Lower Quaternary of the Colombian Eastern Cordillera and the Early Evolution of its High-Andean Biota. Review of Palaeobotany and Palynology, 16: 1-122. van der Merwe, N. J. 1982. Carbon Isotopes, Photosynthesis, and Archaeology. American Scientist, 70: 596-606. 1989. Natural Variation in 13C Concentration and Its Effect on Environmental Reconstruction Using 13C/12C Ratios in Animal Bones. In The Chemistry of Prehistoric Human Bone, edited by T. D. Price, pp. 105-125. Cambridge University Press, Cambridge, van der Merwe, N. J. and E. Medina. 1989. Photosynthesis and 13C/12C Ratios in Amazonian Rain Forests. Geochimica et Cosmochimica Acta, 53: 1091—1094. 1991. The Canopy Effect, Carbon Isotope Ratios, and Foodwebs in Amazonia. Journal of Archaeological Science, 18: 249-259. van der Merwe, N. J. and J. C. Vogel. 1978. 13C Content of Human and Collagen as a Measure of Prehistoric Diet in Woodland North America. Nature, 276: 815—816. van der Merwe, N. J., J. Lee-Thorp, and R. H. V. Bell. 1988. Carbon Isotopes as Indicators of Elephant Diets and African Environments. African Journal of Ecology, 26: 163-172. van der Merwe, N. J., A. C. Roosevelt, and J. C. Vogel. 1981. Isotopic Evidence for Prehistoric Subsistence Change at Parmana, Venezuela. Nature, 292: 536-538. van Geel, B. and T. van der Hammen. 1973. Upper Quaternary Vegetational and Climatic Sequence of the Fuquene Area (E. Cordillera, Colombia). Palaeogeography, Palaeoclimatology', Palaeoecology, 14: 9-92. Vega, Bernardo. 1979. Los Metales y los Aborigenes de la Hispaniola. Investigaciones Antropologicas No. 12. Museo del Hombre Dominicano, Santo Domingo.
References
301
Veintimilla, C. I. and D. M. Pearsall. 1992. A Preliminary Analysis of Past Vegetation in the Jama River Valley, Manabi Province, Ecuador. Paper Presented at the 57th Annual Meeting of the Society for American Archaeology, Pittsburgh. Verano, J. W. and M. J. DeNiro. 1993. Locals or Foreigners? Morphological, Biometric, and Isotopic Approaches to the Question of Group Affinity in Human Skeletal Remains Recovered from Unusual Archaeological Contexts. In Investigations of Ancient Human Tissues: Chemical Analysis in Anthropology, edited by M. K. Sandford, pp. 361-386. Gordon and Breach, Langhorne, PA. Vescelius, Gary S. and Linda S. Robinson. 1979. Exotic Items in Archaeological Collections from St. Croix: Prehistoric Imports and their Implications. Paper Presented at the 8th International Congress for the Study of the Pre-Columbian Cultures of the Lesser Antilles, St. Kitts and Nevis. Vickers, William T. 1980. An Analysis of Amazonian Hunting Yields as a Function of Settlement Age. Working Papers on South American Indians, 2: 7-29. 1984. The Faunal Components of Lowland South American Hunting Kills. Interciencia, 9: 366-376. 1989. Patterns of Foraging and Gardening in a Semi-Sedentary Amazonian Community. In Farmers as Hunters. The Implications of Sedentism, edited by Susan Kent, pp. 46-59. Cambridge University Press, Cambridge. Vigne, Jean-Denis and Marie-Christine Marinval-Vigne. 1983. Methode Pour la Mise en Evidence de la Consommation du Petit Gibier. Animals and Archaeology, I: Hunters and Their Prey, edited by Juliet Clutton-Brock and Caroline Grigson, pp. 239-242. British Archaeological Reports, International Series 163, Oxford. Villa, Paola. 1982. Conjoinable Pieces and Site Formation Processes. American Antiquity, 47: 276-290. Villa, Paola and Jean Courtin. 1983. The Interpretation of Stratified Sites: A View from Underground. Journal of Archaeological Science, 10: 267—281. Villalba, M. 1988. Cotocollao: Un Aldea Formativa del Valle de Quito. Miscelanea Antropologica Ecuatoriana Serie Monografica 2. Museos del Banco Central, Quito. Villavicencio, Manuel. 1984. Geografia de la Republica del Ecuador. Corporacion Editora Nacional, Quito. Vittadello, P. Alberto. 1988. Cha'palaachi: El Idioma Cayapa. 2 vols. Miscelanea Antropologica Ecuatoriana, Serie Monografica 5. Museos del Banco Central del Ecuador, Guayaquil. Vogel, J. C. 1978a. Isotopic Assessment of the Dietary Habits of Ungulates. South African Journal of Science, 74: 298-301. 1978b. Recycling of Carbon in a Forest Environment. Oecologia Plantarum, 13: 89-94. Vogel, J. C. and N. J. van der Merwe. 1977. Isotopic Evidence for Early Maize Cultivation in New York State. American Antiquity, 42: 238—242. Vogel, J. C , A. Fuls, and R. P. Ellis. 1975. The Geographical Distribution of Kranz Grasses in South Africa. South African Journal of Science, 74: 209-215. Von Endt, David W. 1979. Techniques of Amino Acid Dating. In Pre-Llano Cultures of the Americas: Paradoxes and Possibilities, edited by Robert L. Humphrey and Dennis Stanford, pp. 71—100. Anthropological Society of Washington, Washington D.C. Von Endt, David W. and Donald J. Ortner. 1984. Experimental Effects of Bone Size and Temperature on Bone Diagenesis. Journal of Archaeological Science, 11: 247-253. Von Hagen, Victor W. 1939. The Tsdtchela Indians of Western Ecuador. Museum of the American Indian Notes and Monographs vol. 51. Heye Foundation, New York. Voorhies, Michael R. 1969. Taphonomy and Population Dynamics of Early Pliocene Vertebrate Fauna, Knox County, Nebraska. University of Wyoming Contributions to Geology Special Paper 1. Wada, E., T. Kadonaga, and S. Matsuo. 1975.15N Abundance in Nitrogen of Naturally Occurring
302
References
Substances and Global Assessment of Denitrification from Isotopic Viewpoint. Geochemical Journal 9: 139-148. Wafer, Lionel. 1934. (Original 1680-1) A New Voyage and Description of the Isthmus of America. Hakluyt Society, Oxford. Wagner, Erika (ed.). 1991. Homenaje al Dr. Donald W. Lathrap. Antropologica, 75-76. Wagner, G. 1976. 1DOT Flotation Procedure Manual. MS on file, Illinois Department of Transportation, District 8. 1977. The Dayton Museum of Natural History Flotation Procedure Manual. MS on file, Dayton Museum of Natural History, Dayton, Ohio. 1982. Testing Flotation Recovery Rates. American Antiquity, 47: 127-132. 1988. Comparability Among Recovery Techniques. In Current Paleoethnobotany. Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, pp. 17-35. University of Chicago Press, Chicago. Walker, P. L. 1985. Anemia Among Prehistoric Indians of the American Southwest. In Health and Disease in the Prehistoric Southwest, edited by C. F. Merbs and R. J. Miller, pp. 139-164. Anthropological Research Papers 34, Arizona State University, Tempe. 1986. Porotic Hyperostosis in a Marine-Dependent California Indian Population. American Journal of Physical Anthropology, 69: 345-354. Walker, P. L. and M. J. DeNiro. 1986. Stable Nitrogen and Carbon Isotope Ratios in Bone Collagen as Indices of Prehistoric Dietary Dependence on Marine and Terrestrial Resources in Southern California. American Journal of Physical Anthropology, 71: 51-61. Wallace, Anthony F. C. and John Atkins, i960. The Meaning of Kinship Terms. American Anthropologist, 62(1): 58-80. Watson, J. P. N. 1972. Fragmentation Analysis of Animal Bone Samples from Archaeological Sites. Archaeometry, 14: 221-228. Watson, P. J. 1976. In Pursuit of Prehistoric Subsistence: A Comparative Account of Some Contemporary Flotation Systems. Mid-Continental Journal of Archaeology, 1: 77-100. Weiland, D. 1984. Prehistoric Settlement Patterns in the Santa Maria Drainage of Panama: A Preliminary Analysis. In Recent Advances in Isthmian Archaeology, edited by F. Lange, pp. 31-53. British Archaeological Reports, International Series 212, Oxford. 1993. Prehistoric Settlement Patterns in the Santa Maria Drainage in Central Panama. In Cazadores y Recolectores in Centro- y Sudamerica, edited by G. Correal and R. G. Cooke. British Archaeological Reports, Oxford, in press. Weiss, Andrew and Loa Traxler. 1991. Compass Mapping Program at Copan. Paper Presented at the 56th Meeting of the Society for American Archaeology, New Orleans. West, Robert C. 1957. The Pacific Lowlands of Colombia. A Negroid Area of the American Tropics. Louisiana State University Press, Baton Rouge. Whallon, Robert. 1984. Unconstrained Clustering for the Analysis of Spatial Distributions in Archaeology. In Intrasite Spatial Analysis in Archaeology, edited by Harold Hietala with editorial contributions by Paul A. Larson Jr., pp. 242-277. Cambridge University Press, Cambridge. White, C. D. and H. P. Schwarcz. 1989. Ancient Maya Diet: As Inferred from Isotopic and Elemental Analysis of Human Bone. Journal of Archaeological Science, 16: 451—474. White, E. M. and L. A. Hannus. 1983. Chemical Weathering of Bone in Archaeological Soils. American Antiquity, 48: 316- 322. White, Theodore E. 1952. Observations on the Butchering Technique of Some Aboriginal Peoples: 1. American Antiquity, 17: 337-338. 1953a. A Method of Calculating the Dietary Percentage of Various Food Animals Utilized by Aboriginal Peoples. American Antiquity, 18: 396-398. 1953b. Observations on the Butchering Technique of Some Aboriginal Peoples: 2. American Antiquity, 19: 160-164.
References
303
1954. Observations on the Butchering Technique of Some Aboriginal Peoples: 3, 4, 5, and 6. American Antiquity, 19: 254-264. 1955. Observations on the Butchering Technics of Some Aboriginal Peoples: 7, 8, and 9. American Antiquity, 20: 170-178. 1956. The Study of Osteological materials in the Plains. American Antiquity, 21: 401-404. Whitten, Norman E. Jr. 1976. Sacha Runa. Ethnicity and Adaptation of Ecuadorian Jungle Quichua. University of Illinois Press, Urbana. Whyte, Thomas Richard. 1988. An Experimental Study of Small Mammal Remains in Archaeological Pit Features. Unpublished Ph.D. dissertation, Department of Anthropology, University of Tennessee, Knoxville. 1991. Small-Animal Remains in Archaeological Pit Features. In Beamers, Bobwhites and BluePoints: Tributes to the Career of Paul W. Parmalee, edited by James R. Purdue, Walter E. Klippel and Bonnie W. Styles, pp. 163-176. Scientific Papers vol. 23. Illinois State Museum, Springfield. Wijmstra, T. A. 1967. A Pollen Diagram from the Upper Holocene of the Lower Magdelena Valley. Leidse Geologische Mededelingen, 39: 261-267. Wijmstra, T. A. and T. van der Hammen. 1966. Palynological Data on the History of Tropical Savannas in Northern South America. Leidse Geologische Mededelingen, 38: 71-90. Wilbert, Johannes. 1972. Survivors of Eldorado: Four Indian Cultures of South America. Praeger, New York. Wilkie, David S. 1989. Performance of a Backpack GPS in a Tropical Rain Forest. Photogrammetric Engineering and Remote Sensing, 55: 1747-1749. Willey, G. and C. R. McGimsey III. 1954. The Monagrillo Culture of Panama. Papers of the Peabody Museum of Archaeology and Ethnology, Harvard University, vol. 49, No. 2, Peabody Museum, Cambridge, MA. Williams, Valerie. 1992. Bone Collagen Purification with Ether and NaOH for Stable Isotope Analysis. MA Paper in Lieu of Thesis. Department of Anthropology, University of Illinois, Chicago. Wiseman, F. M. 1983. Analysis of Pollen from the Fields at Pulltrouser Swamp. In Pulltrouser Swamp, edited by B. L. Turner and P. D. Harrison, pp. 105—119. University of Texas Press, Austin. Wobst, H. Martin. 1983. We Can't See the Forest for the Trees: Sampling and the Shapes of Archaeological Distributions. In Archaeological Hammers and Theories, edited by J. A. Moore and A. S. Keene, pp. 37-85. Academic Press, New York. Wolf, Teodoro. 1879. Memoria sobre la Geografia y Geologia de la Provincia de Esmeraldas. Imprenta del Comercia de Carlos Kaiser, Guayaquil. Wolff, Ronald G. 1973. Hydrodynamic Sorting and Ecology of a Pleistocene Mammalian Assemblage from California (U.S.A). Palaeogeography, Palaeoclimatology, Palaeoecology, 13: 91-101.
1975. Sampling and Sample Size in Ecological Analyses of Fossil Mammals. Paleobiology, 1: 195-204.
Wright, H. E. Jr. 1967. The Use of Surface Samples in Quaternary Pollen Analysis. Review of Paleobotany and Palynology, 2: 321-330. 1983. Late-Pleistocene Glaciation and Climate around the Junin Plain, Central Peruvian Highlands. Geographiska Annaler, 65 A(1-2): 35-43. Yesner, David R. 1988. Subsistence and Diet in North-Temperate Coastal Hunter-Gatherers: Evidence from the Moshier Island Burial Site, Southwestern Maine. In Diet and Subsistence: Current Archaeological Perspectives, edited by B. V. Kennedy and G. M. LeMoine, pp. 207226. University of Calgary Archaeological Association, Calgary. Yost, James A. and Patricia M. Kelley. 1983. Shotguns, Blowguns, and Spears: The Analysis of Technological Efficiency. In Adaptive Responses of Native Amazonians, edited by Raymond B. Hames and William T. Vickers, pp. 189-224. Academic Press, New York.
304
References
Zeidler, James A. 1984. Social Space in Valdivia Society: Community Patterning and Domestic Structure at Real Alto, 3000-2000 B.C. Unpublished Ph.D. dissertation, Department of Anthropology, University of Illinois, Urbana. 1991. Regional Environment in Northern Manabi (Ecuador): Sampling for Spatial Patterning. Paper Presented at the 56th Annual Meeting of the Society for American Archaeology, New Orleans. Zeidler, James A. and Robin Kennedy. 1994. Environmental Setting. In Regional Archaeology in Northern Manabi, Ecuador, I: Environment, Cultural Chronology, and Prehistoric Subsistence in the Jama River Valley, edited by James A. Zeidler and Deborah M. Pearsall, pp. 13-41. University of Pittsburgh, Memoirs in Latin American Archaeology No. 8, Pittsburgh. Zeidler, James A. and Deborah M. Pearsall. 1990. Tropical Human Ecology, Agricultural Production, and Prehistoric Settlement in the Jama River Drainage, Manabi Province, Ecuador. Research proposal submitted to the National Science Foundation, Washington D.C. 1994. Regional Archaeology in Northern Manabi, Ecuador, 1: Environment, Cultural Chronology, and Prehistoric Subsistence in the Jama River Valley. University of Pittsburgh, Memoirs in Latin American Archaeology No. 8, Pittsburgh. Zimmerman, L. S. 1994. Chapter 13. Palynological analysis. In Regional Archaeology in Northern Manabi, Ecuador, 1: Environment, Cultural Chronology, and Prehistoric Subsistence in the Jama River Valley, edited by J. A. Zeidler and D. M. Pearsall, pp. 175-183. University of Pittsburgh, Memoirs in Latin American Archaeology No. 8, Pittsburgh. Zimmerman, L. S. and V. M. Bryant Jr. 1991. Pollen Studies of Manabi Province, Ecuador. Paper Presented at the 56th Annual Meeting of the Society for American Archaeology, New Orleans. Zucchi, Alberta and William Denevan. 1979. Campos Elevados e Historia Cultural Prehispdnica en los Llanos Occidentales de Venezuela. Universidad Catolica Andres Bello, Caracas.
Index
accessibility, 2, 7, 9, 18 causes of low, 12 defined, 12 forms of low, 18 and site discovery probabilities, 11 Ache (Paraguay), 173 achira, 126, 128 aerial photography, 7, 19, 71, 74, 75, 91 computer enhanced, 80 digitization, 79, 87, 93 and Landsat imagery, 87 limitations of reliance on, 84 stereoscopic analysis, 75 see also, digital satellite imagery, satellite imagery, survey, topographic mapping agouti, 168, 169 Aguadulce (Site), Panama, 219 Aguatia (River), Peru, 14 alluvial soils, and magnet sites, 29 see also, anthropic soil, terra Firme, varzea Altschuler, M., 247 Amazonia, 7, 8, 13, 15, 17, 21, 67, 70, 153, 155, 203, 205, 244 Bolivian, 66, 68, 73 Colombian, 97 Ecuadorian, 40, 127 floodplain habitats, 8, 70 Peruvian, 4 anteater, 169, 171 anthropic soil, 13, 16, 18, 22 defined, 10 terra preta do indio, 16, 109, 133 anthropogenic landscape, 67, 99, 103, 106, 126, 133, 140, 142, 148, 151, 153, 156, 158 identification with microremains, 152 and succession, 152 see also, forest clearance, raised field agriculture, slash and burn agriculture Apurimac Valley, Peru, 232, 236 Archaic (Period), 214 Ardila, G., 109 armadillo, 169, 176, 244
arrowroot, 126, 148 phytoliths, 140 artifacts, abundance, 10 classification, 227, 229 clustering, 10 defined, 10 obtrusiveness, 10 preservation, 13 and site discovery, 11,33 surface densities, 13 auger test pits (ATP), 3, 48 see also, augering, shovel probes, shovel test pits augering, 11, 18, 22, 47, 61, 83 by machine, 47, 48 interval, 53 avocado,102,251 pollen, 112 Ayalan (Site), Ecuador, 181, 184, 189, 192, 193, 196 and skeletal representation, 187 Barghoorn, E., 121 Barrett, S., 245, 246, 247, 255, 258, 261 Bartlett, A., 121 bats, 168, 176, 179 bean, 128 Belize, 135 bias, in archaeofaunal remains, 156, 161, 169, 177 assessment, and test pit geometry, 31 in biological research, 181, 185, 192, 197 ceramic analysis, 226 in macrobotanical remains, 124 in phytolith production, 138 in pollen analysis, 122 in recovery, 116—18 in site preservation, 9 size, 116 in survey, 8, 23, 31, 39 blowguns, 171 Bolivia, 66 bone, apatite, 199, 200, 207, 208, 213, 215, 219, 222 3O5
Index
306
bone, (cont.) collagen, 199, 200, 203, 206-9, ZI1-> 2 I 2 >
2I
5>
2I
85
222
cultural retention, 173 density-mediated survivorship, 174 diagenesis, 165—7, 207, 209, 211, 212, 223 discard, 172, 173 hydroxyapatite, 166 marrow, 173 preservation, 4, 181, 185, 189, 192, 197, 200, 212, 2.15 protein degradation, 166, 167 scavenging, 173, 177 weathering, 174, 177 see also, faunal remains, human remains, mammals, stable isotope analysis, zooarchaeology bows and arrows, 171 Brazil, 43, 147 Bray, W., 3, 96 Brown, J., 48 budares, 111 burial, 53, 61, 210 bundles, 215 cemetery, 56, 61, 64, 183, 215 ossuaries, 192 primary and secondary, 183, 184, 215 shaft tombs, 184 tumulus, i n urn, 184, 187, 192, 235 see also, human remains Bush, M., 135 Buys, J., 183 Cabello de Balboa, M., 245 Calima, Colombia, 99, 103, 127, 135 Cannon, A., 60 Cantarana (Site), Ecuador, 258 capybara, 168, 169 carbon, particulate, 142, 146, 148, 149 particulate, and agricultural intensification, 150, 153 particulate, production, 152 see also, slash and burn agriculture carbon isotopes, 4, 129, 199, 200, 203, 204, 207, 209, 222
in animals, 205 and forest canopy effect, 205 incorporation into tissue, 206, 207 see also, nitrogen isotopes, stable isotope analysis Carrier (Site), Haiti, 62 Cauca (River), Colombia, 104 Cayapas (River), Ecuador, 243, 245, 247, 254, 255, 257, 258 ceramics, 244 Chachi, 258 classification, 228, 242 clay, 230, 231, 240 complex, 226 decoration, 227, 231-3, 235, 237, 239, 258 design statements, 232
design structure, 231, 232 durability, 4, 136, 226, 244 factemes, 230 firing, 230, 231, 238-40 fragmentation, 230-2 function, 241, 258, 262 heavy reliance upon, 67 history of research, 226 modes, 4, 230, 232, 235, 238 petrographic examination, 240 preservation, 4, 225 radiographic analysis, 240 shape, 227, 230-2, 234, 235, 238, 239 structural classification, 4, 224, 228, 229, 232, 234, 238-42 style, 226, 232 taphonomy, 228, 230 temper, 227, 230, 231, 239 trait, 226 typological classification, 4, 226, 238 typologies, problems with, 226, 227, 229 Cerro Mangote (Site), Panama, 140, 214, 218, 221, 222
Chachi (Ecuador), 4, 243, 245, 257, 261 archaeology, 258, 261 ethnohistoric documents, 251, 254 oral traditions, 4, 243, 245, 246, 254, 255, 257, 262 Chavin (Culture), Peru, 129 Chavin de Huantar (Site), Peru, 203 Chile, 201 chile pepper, pollen, 133 Chorrera (Culture), Ecuador, 126 Clark, J.G., 115 climatic change, 97, 100, 105, 108, 109, 111 Cob-III (Site), Belize, 135 Cogollo (Site), Colombia, 107 Colha (Site), Belize, 135 Colinvaux, P., 135 Colombia, 3, 97-9, 103, 112, 113, 127, 240 Colorado (Ecuador), 257 community organization, 3, 64 analysis of, 42, 62, 68 Saladoid, 43 community patterning, 46 Compositae, pollen, 112 Cooke, R., 222 Copan (Site), Honduras, 134 coprolite, 116 Corono (Site), Panama, 139 Cotocollao (Site), Ecuador, 183, 190, 193, 195, 196 cotton, 128 Cueva de los Ladrones (Site), Panama, 140, 141, 219,
222
Cueva de los Vampiros (Site), Panama, 140 Cuevas (Style), Puerto Rico, 61 cultural stress, 98 Cumancaya (Site), Peru, 8, 245 Cumancaya (Style), Peru, 239 Cumbaya (Site), Ecuador, 183 curare, 171
Index DeBoer, W., 4, 62, 230, 243 deer, 161, 168, 169, 171, 208 Brocket, 176 white-tailed, 161, 176, 221 Deetz, J., 230 Delcourt, P., 122 Denevan, W., 71, 73 digital elevation model (DEM), 19 digital satellite imagery, 19 Dimbleby, G., 123 dogs, as scavengers, 174 used in hunting, 170 Dominguez, V., 183 Dominican Republic, 43 Doyon, L., 184 ecology, paradigms, 67 Ecuador, 4, 5, 23, 40, 116, 117, 120, 127, 128, 135, 157, 173, 179, 181, 183, 184, 189, 197, 239, 240, 243 El Dorado (Valley), Colombia, 100 El Paraiso (Site), Peru, 116 El Valle (Lake), Panama, 144 El Villar (Site), Bolivia, 73, 83, 92 electromagnetic (EM) conductivity, 23 Erickson, C , 3, 66 Esmeraldas (Province), Ecuador, 243, 245, 251, 255 ethnographic analogy, 64, 66, 70 ethnohistoric documents, 4, 64, 66, 243, 245, 251, 2-55 ethnohistory, 4, 243 Evans, C , 39 Marajo Island, 13 excavation, "macroblock", 53, 57 stratification, 50 excavation techniques, 10 deep trenching, 10 horizontal stripping, 3, 10, 68 raised fields, 3, 87 stratification, 47 subsurface testing, 10 trench, 18, 53 see also, test pit, visibility experimental archaeology, 3, 66, 70, 74 raised fields, 87, 92 faunal remains, 156, 198, 221, 230 collagen extraction, 212 density-mediated survivorship, 163, 174, 176 evenness, 161 fragmentation, 161, 164, 167, 169, 173, 174, 176 hydrodynamic sorting, 174, 177 microvertebrate, 165, 174-6 nature of samples, 156, 158, 159 numerical abundance, 158 preservation, 4, 156, 163, 166 from raised fields, 87 recovery, 4, 158, 167, 175-7 size, 161, 164, 167
307
stable isotope analysis, 219 taxonomic richness, 159, 167 weight, 161 see also, bone, human remains, mammals, zooarchaeology features, abundance and diversity, 15 defined, 10 in excavation, 55 postmold, 53 preservation, 13 structure, 57 urn burials, 15 felines, 173 jaguars, 246 fine sieving, 116 effect on macroremain recovery, 116 preferable to flotation, 118 flotation, 9, 68, 116, 131 adapting to tropical lowlands, 117 IDOT system, 117, 118 in raised fields, 92 and recovery of archaeofaunas, 158, 160, 175, 176 SMAP system, 118 types of, 117 see also, macrobotanical remains forest clearance, 102-5, n o , 112, 127, 135, 141, 150 and agricultural intensification, 98, 150 and Heliconia, 146 and mammals, 169, 170 recent, 1, 112,158 and seed cultivation, 135 Yotoco (Period), 100 see also, anthropogenic landscape, carbon, slash and burn agriculture formation processes, 8, 57, 64, 228, 244 bioturbation, n deposition of botanical materials, 125 fluvial forces and site destruction, 5, 7, 244 fluvial forces and site discovery, 15, 244 leaf cutting ants, n in raised fields, 90 see also, horizontal stratigraphy, site Formative Period, Colombia, 240 Ecuador, 24, 127, 128, 136, 240 Panama, 214, 221 Gallagher, P., n o Gaspar de Torres (Fr.), 251, 253-5 Genoves, S., 190 Geographical Information Systems (GIS), 19, 94 and Global Positioning System (GPS), 19 Geral (Lake), Brazil, 153 Glassie, H., 228 Global Positioning System (GPS), 19, 80, 84 also geographical positioning system, 79 and topographic mapping, 84 gourd, 126, 128 Gramineae, pollen, 112, 127
308
Index
Granja de Sivia (Site), Peru, 232, 234, 238, 239 Granja de Sivia (Style), Peru, 236 Granulosa-Incisa pottery, 107 Guajira (Colombia), 108, 109 Guangala (Culture), Ecuador, 196 guanin, 61 Guaymi (Panama), 222 Hacienda El Dorado (Site), Colombia, 127 Hacienda Grande (Complex), Puerto Rico, 61 Hacienda Grande (Site), Puerto Rico, 46 Hansell, P., 219 Hayden, B., 60 Herrera de Turbay, L., 112 horizontal stratigraphy, river meanders, 7 Horno (Period), Colombia, 109, 111 Howard, G., 226 huaqueros, 186 see also, human remains, pothunters, saqueadores human remains, 181 age estimation, 192, 195, 197 demographic reconstruction, 187, 191, 192 disease, 189, 191, 193-6 hair, 211 and huaqueros, 184 maize consumption, 196 preservation, 185, 189, 192, 194 recovery, 181 sampling, 181, 187, 189, 191, 193-7, 210 and skeletal representation, 4, 183, 187 and stable isotope analysis, 203, 204, 208, 210, 211, 214, 219 stature estimation, 189, 190 taphonomic factors, 181, 190, 192, 194, 215 trauma, 193 see also, bone, burial, faunal remains, maize, mammals, stable isotope analysis, zooarchaeology Ibarra, Ecuador, 245, 255 IDRISI software, 79 Incisa-Alisada (Tradition), 108 Indian Creek (Site), Antigua, 62 Integration Period, 24 Jama (Valley), Ecuador, 9, 23, 117, 119, 122, 126, 156, 158, 176 Jama-Coaque (Culture), Ecuador, 24, 120, 126, 196 Jomon-Valdivia hypothesis, 240 Jones, J., 135 K-means cluster analysis, 58 Kensinger, K., 243 Kintigh, K., 33 Kluckhohn, C , 227 Kogi (Colombia), n o Koster (Site), Illinois, 46 Kotosh (Site), Peru, 136 Kroeber, A., 227
La Florida (Site), Ecuador, 184 La Libertad (Site), Ecuador, 183 La Mula (Site), Panama, 139, 140 La Mula-Sarigua (Site), Panama, 214 La Pitia (Site), Venezuela, 109 Siruma phase, 109 La Tolita (Site), Ecuador, 183, 196 La Yeguada (Lake), Panama, 142, 146, 148, 150, 153,222
lagomorphs, 169 Lake Ayauch1, Ecuador, 128 landscape archaeology, 3, 67, 74, 95 Las Vegas (Tradition), Ecuador, 127, 183, 190 Lathrap, D. W., 1, 5, 62, 68, 243 and agricultural origins, 114, 155 and archaeological research, 2, 5, 68 and bioturbation, n and botanical study in archaeology, 5, 131 ethnography and archaeology, 227, 244 and methodology of ceramic classification, 4, 5, 225, 227, 230, 231 and phytolith analysis, 5,136 and preservation bias, 2, 219, 244, 250 site discovery, 8 site location, 7 Levi-Strauss, C , 262 Llanos de Mojos (Bolivia), 3, 67, 68, 71, 74, 95 see also, raised field agriculture logistics, 67 constraints, 1, 7, 42, 46, 185, 239 midden depth, 47 vegetation, 47 see also, site, visibility Loma (Period), Colombia, 109 Long Island, New York, 22 Machalilla (Phase), Ecuador, 181 macrobotanical remains, 3, 113, 125, 198 blanket sampling, 119 comparative collections, 3, 120 fragmentation and identification, 120 opportunistic sampling, 119 from raised fields, 87, 91 recovery, 3, 115 sampling, 118 see also, flotation, individual plants Magdalena (River), Colombia, 103, 104 Magdalena (Valley), Colombia, 102, 103, 108 magnetometry, 23 Maisabel (Site), Puerto Rico, 3, 42, 43, 50 maize, 102, i n , 112, 116, 126-8, 135, 141, 153, 196, 198, 201, 210, 222, 245, 246, 251 antiquity, 128, 133, 136, 199, 214, 221 cultivation, 109 dietary importance, 199, 203, 204, 209, 219, 221 and human remains, 196, 203, 208 macroremains, 125 phytoliths, 125, 127, 136, 137, 140, 141, 222 pollen, 107, 112, 128, 133, 134, 140, 149, 222 Malamboid (Tradition), Colombia, i n Malibues (Colombia), 108
Index mammals, abundance, 168-70, 177 arboreal, 167-9, 171 availability, 155, 170 behavior, 168, 171 biomass, 168-70 consumption by humans, 172, 173 density, 168, 169 dismemberment by humans, 171—3 diversity, 169 herbivores, 169 large, 156, 161, 167, 168, 170-7 and llanos habitats, 169 neotropical forest community, 167-9 predation, 169 predation by humans, 169, 170 representation, 161 richness, 169, 177 size, 168, 170 small, 161, 163, 164, 167-76 terrestrial, 167-71 transport by humans, 169, 171, 172 see also, bone, faunal remains, human remains, individual animals, zooarchaeology Manabi (Province), Ecuador, 23, 157, 196 Mangelsdorf, P., 114 manioc, 107, 111, 112, 141, 198 antiquity, 133 pollen, 133 Marajo (Island), Brazil, 15, 68, 97 anthropic soil, 16 Marajoara, 13 marsupials, 168 Mason, A., 111 McManamon, F., 10 Meggers, B., 17, 39 Marajo Island, 13 Metraux, A., 71 microremains, 3, 115, 132, 151 analyses in tandem, 142 see also, individual plants, palynology, phytoliths, pollen middens, deposition rates, 9 "mounded", 52, 60 shell, 140 Mississippi, site discovery techniques, 7 Modelada-Incisa (Tradition), Colombia, 107 Mompos Depression, Colombia, 104-6, 108 Monagrillo (Pottery), Panama, 139, 214 Monagrillo (Site), Panama, 140, 214, 219 monkeys, 168, 176 blowgun hunting, 171 callitrichid, 169 cebid, 169 Howler, 168 Monsalve, J., 135 Monserrate (Period), Puerto Rico, 61 Monserrate (Site), Puerto Rico, 52, 62 Monte Carlo simulation, 37, 39 Moran, E., 21
309
Moreno Navarro, I., 245, 254 Munizaga, J., 181 Myers, T., 13, 64 Nahuange (Phase), Colombia, i n Nance, J., 20 negative evidence, 8 nitrogen isotopes, 4, 199, 200, 204, 205, 209, 222 in animals, 201, 204 incorporation into tissue, 207 and legumes, 204 see also, carbon isotopes, stable isotope analysis Nordenskiold, E., 71 Norton, P., 184 Norr, L., 4, 198 OGSE-80 (Site), Ecuador, 127, 190, 193, 195 Orinoco (River), 62, 129 paca, 168, 169 Paleo-Indian (Period), Panama, 213, 214 paleoethnobotany, 113, 115, 132 comparative collections, 120 palimpsest assemblage, 7, 57, 261 palm (Aracaceae), 128 macroremains, 125 phytoliths, 125 Palop Martinez, J., 245, 251, 254, 255, 262 palynology, 121, 133, 136, 137 modern analog method, 152 off-site sampling, 122, 123 see also, deep coring, microremains, pollen Panama, 4, 132, 135, 138, 140, 142, 147, 151, 200, 213, 214, 219, 221, 222 Pearsall, D., 3, 113, 136 peccary, 161, 171, 176 collared, 169 white-lipped, 168, 169 Pechichal (Site), Ecuador, 119, 124, 179 pedestrian survey, 11 and visibility, 18 Penon del Rio (Site), Ecuador, 126 Peru, 5, 173, 232, 239 phosphate fractionation, 107 phytoliths, 3, 68, 115, 132, 139, 149, 198 analysis, 123, 126, 129, 132, 136 comparative collections, 120, 137, 142 deposition, 137 identification, 137, 138 maize, 136, 141, 222 and multivariate statistics, 129 preservation, 124, 126, 136, 137, 140 production, 137, 138 from raised fields, 88, 92 sampling, 124 as sole source data, 126 squash, 141, 222 see also, individual plants, microremains Piperno, D., 3, 121, 122, 124, 125, 130, 222 Plafker, G., 71 pollen, 3, 68, 115, 132, 139, 149, 198 analysis, 123, 126, 129, 132, 136
3io
Index
pollen, (cont.) Chile pepper, 133 comparative collections, 120, 133, 142 diagram, 99, 102, 103, 112, 133, 142 maize, 100, 102, 134, 140, 141, 149, 222 manioc, 133 opportunistic sampling, 123 paleoecological techniques, 134 paramo, 101, 102 preservation, 121, 123, 133, 134, 140 production, 121, 133
from raised fields, 88, 123 recovery, 122, 132, 134 squash, 133, 140, 222 see also, individual plants, microremains, palynology Portacelli (Style), Colombia, 109, n o pothunters, 3, 5, 43 see also, human remains, huaqueros, saqueadores preservation, 1,3, 66, 6j, 131, 225 archaeofaunas, 4, 155, 156, 163, 176, 177, 179 bias, 2, $,7^9^ *55 ceramics, 226, 231 macrobotanical remains, 113, 116, 118, 131, 132 organic remains, 3, 155, 164, 166, 172, 181, 185, 244 phytoliths, 124, 136, 140 pollen, 121, 140 pollen analysis, 126 in raisedfields,88 of raised fields, 75, 78 site destruction, 81, 184, 244 stable isotope analysis, 200, 211, 212, 215, 222 stratigraphy, 3, 227, 241, 244 Puerto Hormiga (Site), Colombia, 240 Puerto Rico, 42, 43, 46, 55, 62 Pulltrouser Swamp, Belize, 123 quadrats, 2, 20, 26 effectiveness, 26, 27 Quimpiri (Style), Peru, 236-9 Quito, Ecuador, 183, 184, 192-4 raised field agriculture, 3, 66, 70, 100, 104 abandonment, 71 aquaculture, 73 benefits, 70, 106 camellones, 107, 108 causeways and canals, 73, 87, 106 dating, 71, 90 experimentation, 92, 108 flood control, 87, 106 occupation sites, 73 radiocarbon dating, 90 seriation, 91 thermoluminescence dating, 90 see also, experimental archaeology, Llanos de Mojos, slash and burn agriculture, sustainable agriculture Rancheria (Valley), Colombia, 108, 109 Ranere, A., 138, 222 Raymond, J. S., 4, 224
Real Alto (Site), Ecuador, 68, 128, 183 refitting, 9 Reichel-Dolmatoff, G. and A., 109, n o remote sensing, 3, 17, 19, 47, 64, 70, 74, 94 digital, 79 Landsat, 74, 80 Rindos, D., 129 rodents, 161, 168, 169, 175 Roosevelt, A., 134, 161, 168, 169, 175 Rouse, I., 64, 227, 230 Rowe, J., 227 Rue, D., 134, 135 Saladero (Site), Venezuela, 43 Saladoid (Series), Puerto Rico, 43, 46, 61 middens, 46 Saladoid/Ostionoid (Series), Puerto Rico, 42 Salango (Site), Ecuador, 118, 183, 184 sampling, and archaeofaunas, 177 deep site, 46 probabilistic, 2, 20 and STP, 27, 31 sampling design, 10, 21, 45, 46 deep site, 47 regional, 20 San Francisco (Site), Ecuador, 184, 192 San Isidro (Site), Ecuador, 126 San Jacinto (Pottery), Colombia, 240 San Jorge (River), Colombia, 104-6, n o San Lorenzo del Mate (Site), Ecuador, 181, 183 Santa Maria (River), Panama, 138 Santarem, Brazil, 15 Santiago-Cayapas, Ecuador, 4, 243, 246, 251, 254, 2-57, 2.58 Santo Domingo (Site), Ecuador, 184 Sapir, E., 227 saqueadores, 43 see also, human remains, huaqueros satellite imagery, 17, 19, 20, 74 see also, aerial photography Sauer, C , 114 scanning electron microscopy (SEM), 114 Schiffer, M., 60 settlement organization, 46, 73 shamans, 246 Shipibo (Peru), 227, 245 shotguns, 170 shovel probes, 2, 10, 13, 18, 22 compared to pedestrian survey, 22 comparison of five and eight probe, 37 effectiveness, 26, 27, 31 eight probe, 33 five probe, 33 shovel test pits (STP), 11, 18, 22 effectiveness, 30, 47 in Jama Valley, Ecuador, 23 Shuar (Ecuador, Peru), 173 Siegel, P., 2, 42 Sierra Nevada de Santa Marta, Colombia, n o , in
Sinu (River), Colombia, 104, 106
Index site, assessment, 31 boundary definition, 42 constituents, 10 densities, 8, 29 discovery, 7-9, 11, 42 discovery and anthropic soil, 17 discovery probability, defined, 10 discovery with quadrats, 26 formation, 9 magnet, 24 mounded, 71, 72, 81, 82 multicomponent, 47 nature of in forested neotropics, 8, 67 nature of in Jama Valley, Ecuador, 24 occupational history, 42 preservation, 9 scavenging from, 261 settlement layout, 42 size, 42, 46 spatial structure, 42 structure, 46 see also, accessibility, artifacts, auger test pits, augering, electromagnetic conductivity, excavation, excavation techniques, features, formation processes, logistics, magnetometry, middens, palimpsest assemblage, pedestrian survey, quadrats, remote sensing, sampling, sampling design, shovel probes, shovel test pits, subsurface testing, surface collection, survey, test pit, visibility Sitio Sierra (Site), Panama, 140, 141, 214, 219, 222 Sivia (Style), Peru, 235-39 slash and burn agriculture, 134, 151 antiquity, 3, 149, 150, 153 see also, anthropogenic landscape, forest clearance, raised field agriculture sloths, 168, 169 Smole, W., 62 Sonso (Period), Colombia, 100 Sorce (Site), Puerto Rico, 46 Spanish conquest, 106, 108, n o , 112, 214, 246 Spaulding, A., 225, 227 spears, 170 Spurling, B., 27 squash, 126, 135, 153 phytoliths, 127, 137, 140, 141, 222 pollen, 133, 140, 222 stable isotope analysis, 4, 199, 200, 209 CAM photosynthetic pathways, 201 climatic factors, 201, 203 fractionation factor, 208, 221 linear mixing model, 206 sample preparation, 211, 213, 215 see also, carbon isotopes, nitrogen isotopes Stahl, P. W., 4, 62, 154 Stevenson, M., 255, 257, 261 Stothert, K., 183 structural linguistics, 228 subsurface testing, 10, n , 18, 21, 47, 64, 68 and bias assessment, 31, 33 and geophysical survey, 16
311
and site boundaries, 47 and survey intensity, 12 surface collection, 13, 46, 47, 81 survey, 7, 8, 45, 70 with aircraft, 3, 74, 78 comparison of techniques, 22 deep coring, 18 designs in the tropics, 12, 45 discovery model sampling, 20 and disturbances, 80, 244 efficiency, 20 intensity and site discovery probabilities, n intensity, defined, 12 interviewing guides, 17, 23 longitudinal, 17, 18 pedestrian, 3, n , 22, 26, 31, 74, 80 probabilistic, 17, 23, 24 purposive, 20, 23 right-of-way transects, 17 statistical inference, 20 statistical precision sampling, 20 strata, 21 stratification, 24 transects, 21, 26, 75, 82 use of ethnographic and ethnohistoric sources, 19 see also, site, subsurface testing, surface collecting, visibility sustainable agriculture, 67, 95 sweet potatoes, 102 Taino (Puerto Rico), 61 Tairona (Colombia), n o , i n Taperinha (Site), Brazil, 15 tapir, 168, 169, 171 traps, 171 Taylor, W., 228 Tehuacan (Valley), Mexico, 136 terra firme, 16-18, 21 see also, alluvial soils, anthropic soil, vdrzea Teso do Bichos (Site), Brazil, 47 test-pit, 13, 22 Aguatia River, 14 Titicaca (Lake), Peru, 74 topographic mapping, 84 and Landsat imagery, 87 topographic maps, 19 traps, 171 tumbagua, 111 Tumbaviro (Complex), Ecuador, 261, 262 Tumbibucho (River), Ecuador, 251 Ubelaker, D., 4, 181 Ucayali (River), Peru, 7, 11, 15, 68, 239 Uraba (Gulf), Colombia, 103 Valdez, F., 183 Valdivia (Phase), Ecuador, 24, 126-8, 181, 240 van der Hammen, T., 121 vdrzea, 8, 16, 21, 134 see also, alluvial soils, anthropic soil, terra firme Venezuela, 43, 127
Index visibility, 1,5,9 causes of nonexistent, 11 defined, 11 hindering community-oriented archaeology, 42 and sampling design, 25 and site discovery probabilities, 2, 11, 39, 185 and STP subsampling, 27, 31 surface, 7, 18, 42, 46, 185 and survey, 22, 26 see also, accessibility, site, subsurface testing, surface collecting Waiwai (Guyana), 62 Wapisiana (Guyana), 62 West, R., 258 West Indies, 43 White, T., 177
Wiseman, F., 123 Wolf, T., 257, 261 Wright, H., 121 Yanoama (Brazil), 62 Yotoco (Period), Colombia, 100 Yumes (Site), Ecuador, 126 Zapallo Grande (River), Ecuador, 247, 261 Zeidler, J., 2, 7, 119 Zeniies (Colombia), 107 Zimmerman, L., 122 Zipacon I (Rockshelter), Colombia, 102 zooarchaeology, 155, 177 MNI, 161, 163 NISP,161 see also, bone, faunal remains, mammals