kk:&&Eg .A*y-
2
i k
- ;-< y;; &
UITGAVEN VAN HET NEDERLANDS HISTORISCH-ARCHAEOLOGISCH INSTITUUT TE ISTANBUL Publications de 1'Institut historique-archCologique nCerlandais de Stamboul sous la direction de Machteld J. MELLINK, J. de ROOS, J.J. ROODENBERG et K.R. VEENHOF
LXXXVIII
RAINFALL AND AGRICULTURE IN NORTHERN MESOPOTAMIA
.
*i"-q-@%+i.\ ~WPSLFY
RAINFALL AND AGRICULTURE IN NORTHERN MESOPOTAMIA (MOS STUDIES 3) PROCEEDINGS OF THE THIRD MOS SYMPOSIUM (LEIDEN 1999)
edited by
R.M. JAS
Cover illustration: Middle Assyrian cylinder seal. Drawing by F.A.M. Wiggermann after A. Moortgat, ZA 47 (1942) 81, Fig. 66.
NEDERLANDS HISTORISCH-ARCHAEOLOGISCH INSTITUUT TE ISTANBUL 2000
Copyright 2000 by Nederlands Instituut voor het Nabije Oosten Witte Singe1 25 Postbus 9515 2300 RA Leiden, Nederland
PREFACE
All rights reserved, including the rights to translate or to reproduce this book or parts thereof in any form
This volume collects most of the papers read during the third MOS symposium that took place in Leiden on May 21-22, 1999. The necessary swiftness of execution the contributors generously agreed upon in preparing their lectures for publication is something for which I owe them additional, and warmest thanks. The organizer wishes to acknowledge the following institutions for generously sponsoring this seminar: At Leiden University: The Research School CNWS (School of Asian, African and Amerindian Studies), the Department of Near Eastern Studies (TCNO), and the Leids Universiteits Fonds (LUF); The Netherlands Organization for Scientific Research (NWO); The Koninklijke Nederlandse Akadernie van Wetenschappen. An expanded version of J. G. Dercksen's contribution will appear in his Old Assyrian Trade and Society (forthcoming). I. MArquez Rowe's discussion paper on the status and role of villages in the Mediterranean fringe, unfortunately, could not be included.
Rainfall and Agriculture in Northern Mesopotamia; Proceedings of the Third MOS Symposium (Leiden 1999) / ed. by R.M. Jas. Istanbul: Nederlands Historisch-Archaeologisch Instituut; Leiden: Nederlands Instituut voor het Nabije Oosten [distr.]. (Uitgaven van het Nederlands Historisch-Archaeologisch Instituut te Istanbul; ISSN 0926-9568; 88) ISBN 90 6258 089 0
Printed in Belgium
Thanks are due to Linda McLarnan for her valuable comments on parts of the manuscript, and to Drs. W. Hovestreydt for scanning several of the illustrations. For kind and able bibliographical assistance, I am much indebted to Mrs. M. F. Monti, Library Assistant in the Forestry Department of the F A 0 Library in Rome. The International Center for Agricultural Research in the Dry Areas, Aleppo, on more than one occasion, generously provided me with hard to find articles and books.
R. M. Jas Amsterdam, February 2000
TABLE OF CONTENTS PREFACE INTRODUCTION ABBREVIATIONS I.
THE ANCIENT LANDSCAPE, PLANT AND ANIMAL REMAINS AS INDICATORS
T. J. Wilkinson Settlement and Land Use in the Zone of Uncertainty in Upper Mesopotamia S. Bottema and R. T. J. Cappers Palynological and Archaeobotanical Evidence from Bronze Age Northern Mesopotamia C. Cavallo The Role of Animals in a Neolithic Agricultural System in the Ancient Jazira J. D. Lyon Middle Assyrian Expansion and Settlement Development in the Syrian Jazira: The View from the Balikh Valley 11.
TEXTUAL EVIDENCE: FROM DETAILS TO GENERALIZATIONS
B. Lafont Irrigation Agriculture in Mari J. C. Fincke Transport of Agricultural Produce in Arrapke F. A. M. Wiggermann ~ ~ r i c u l t uin r ethe Northern Balikh Valley: The Case of Middle Assyrian Tell Sabi Abyad 171 K. Radner How Did the Neo-Assyrian King Perceive his Land and its Resources? 233 R. M. Jas Land Tenure in Northern Mesopotamia 247 G. van Driel The Mesopotamian North: Land Use, An Attempt 265
INTRODUCTION
Agriculture in Ancient Mesopotamia is commonly, and conveniently, divided into rainfed agriculture in the North and irrigation agriculture in the South. Considering the fact that wherever possible farmers irrigated in Ancient Mesopotamia, in the North as well as in the South, the aim of this meeting was to try to establish more characteristics that distinguish agriculture in Northern Mesopotamia from that in the South that could justify this dichotomy. At the same time we hope the papers collected in this volume show that the idea of Northern Mesopotamia as a whole is somewhat of a simplification. The area, instead, can better be divided into different regions with their own potentials and limitations for either arable farming, horticulture, the keeping and raising of livestock, or pastoral farming. In certain areas more stable yields could be expected while in other areas, the management of rainfed crops could be quite a risky undertaking. Relief, for example, is an important determinant for the forms of land use found in the various regions of the North. Luckily, we happen to have written sources from areas differing as widely as the mountainous terrain in North-Eastern Iraq and the river valleys in Northern Syria. Another important question is how far the steppe was regarded as forming part of the farming system. This topic has great ramifications for, among other things, a correct assessment of the integration of crop and livestock production in the Ancient Near East. Because of the different uses made of land in Northern Mesopotamia, the patterns of land ownership and the access rights to land can be expected to differ from the South as well, several overviews of the written evidence and the situation in present day Iraq and Syria are included. The specialists from various disciplines that were invited to speak on the above topics, archaeologists, a paleobotanist and a zoo-archaeologist, historians and several philologist, had the occasion to discuss the contributions their various fields. of study have to make to a better understanding of the ways in which ancient agricultural production was organized in Northern Mesopotamia. The findings presented in this volume will be of interest for a general model of agriculture in Northern Mesopotamia that will need to integrate the written sources with archaeological and archaeobotanical data, and with the data on the modern Middle-Eastern environment, leaving room for the differences that exist within the rainfed areas of Ancient Mesopotamia that are caused by factors other than rainfall.
ABBREVIATIONS
ABBREVIATIONS
A AAS AbB ABL AcAnHung ADB ADD
A~S AfO AHw Amurrru AOAT AoF ARET ARM ARMT ARU AS AS1 ASJ ASOR AT BaF BAH
BaM BAR BASOR BATSH BBVO BCSMS BiOr BM BPO BRM
siglum for texts from Mari Les Annales Archtologiques de Syrie (later: Annales ArchCologiques Arabes Syriennes) Altbabylonische Briefe in Umschrift und aersetzung R. F. Harper, Assyrian and Babylonian Letters Acta Antiqua Academiae Scientimm Hungaricae C. H. W. Johns, An Assyrian Doomsday Book C. H. W. Johns, Assyrian Deeds and Documents G. Wilhelm, Das Archiv des S i l w a - ~ e E u ~ Archiv fiir Orientforschung W. von Soden, Akkadisches Handworterbuch Vol. 1: J.-M. Durand ed., Mari, ~ b l et a les Hourrites - dix ans de travaux. Actes du colloque international (Paris, mai 1993) Alter Orient und Altes Testament Altorientalische Forschungen Archivi reali di Ebla - Testi Archives royales de Mari Archives royales de Mari (texts in transliteration and translation) J. Kohler and A. Ungnad, Assyrische Rechtsurkunden in Umschrift und Ubersetzung, Leipzig 1913 Assyriological Studies Advanced Science Institutes, Series I: Global Environmental Change Acta Sumerologica Japonica American Schools of Oriental Research D. J. Wiseman, The Alalakh Tablets Baghdader Forschungen Bibliothkque archtologique et historique (136 = B. Geyer ed., Techniques et pratiques hydro-agricoles traditionelles en domaine irriguC (Actes du Colloque de Damas, 27 juin - lerjuillet 1987), Institut Franqais du Proche Orient, Paris 1990. Baghdader Mitteilungen British Archaeological Reports Bulletin of the American Schools of Oriental Research ~ 1 Diir-Katlimmu Berichte der Ausgrabung Tall S E Hamad Berliner Beitrage zum Vorderen Orient Bulletin of the Canadian Society for Mesopotamian Studies, QuCbec Bibliotheca Orientalis siglum for texts in the British Museum Babylonian Planetary Omens Babylonian Records in the Library of J. Pierpont Morgan
BS BSA BSOAS BT CA CAD CM CMA CRRA CSSH CT CTN EN ETAP Flor. Mar. Fs. Finet HSAO HSS IM JAC JCS JEN JEOL JNES JNSL KAH KAJ LAP0 MADD MAL MAR1 MARV MASCA Mayer Cat MDOG MDP NABU NALK NARGD ND NL OA OBRT OIP OLP OLZ
University of Amsterdam Balikh Survey Bulletin on Sumerian Agriculture Bulletin of the School of Oriental and African Studies siglum for texts from Balawat Current Anthropology Chicago Assyrian Dictionary Cuneiform Monographs (Groningen) Corpus Medio-Assiro (Rome) Comptes rendus des Rencontres assyriologiques internationales Comparative Studies in Society and History Cuneiform Texts from Babylonian Tablets (British Museum) Cuneiform texts from Nimrud Excavations at Nuzi Expanded Technical Assistance Program Florilegium Marianum, Mtmoires de NABU Reflets des deux fleuves (Akkadica Supplementum 6) Heidelberger Studien zum Alten Orient Harvard Semitic Series siglum for texts in the Iraq Museum, Baghdad Journal of Ancient Civilizations Journal of Cuneiform Studies Joint Expedition with the Iraq Museum at Nuzi Jaarbericht van het Vooraziatisch-Egyptisch Genootschap "Ex Oriente Lux" Journal of Near Eastern Studies Journal of Northwest Semitic Languages Keilschrifttexte aus Assur historischen Inhalts E. Ebeling, Keilschrifttexte aus Assur juristischen Inhalts Litteratures anciennes du Proche Orient (16 and 17 = J.-M. Durand, Documents tpistolaires du palais de Mari, 1 and 2 J. N. Postgate, More "Assyrian Deeds and Documents" Iraq 32 (1970) 129164. Middle Assyrian Laws Mari, Annales de Recherches Interdisciplinaires H. Freydank, Mittelassyrische Rechtsurkunden und Verwaltungstexte Museum Applied Science Center for Archaeology Walter Mayer, Nuzi-Studien I: Die Archive des Palastes und die Prosopographie der Berufe (AOAT 20511) Mitteilungen der Deutschen Orientgesellschaft MCmoires de la DClegation en Perse Nouvelles assyriologiques brsves et utilitaires T. Kwasman, Neo-Assyrian Legal Documents in the Kouyunjik Collection of the British Museum Postgate, Neo-Assyrian Royal Grants and Decrees siglum for Nimrud texts Nimrud Letter (H. W. F. Saggs, Iraq 17 [I9551 and following) Oriens Antiquus S. Dalley, et al. Old Babylonian Texts from Tell Rimah Oriental Institute Publications Orientalia Lovaniensia Periodica Orientalistische Literaturzeitung
XI1 OrNS PRU QGS
RA RAI RBC RGTC
RHA RIMA RIME RLA RS SAA SAAB SAAS SANE SCCNH SGKAO SM SMN StAT T Talon TAVO TCL TCS TIM TM TMO UF VAT VS WdO WZKMIS YNER ZA
ABBREVIATIONS
Orientalia Nova Series Le Palais Royal d7Ugarit Quaderni di Geografia Storica (Universith di Roma - Istituto di Studi del Vicino Oriente) Revue d'assyriologie et d'archeologie orientale Rencontre Assyriologique Internationale Rosen Babylonian Collection (Yale) Repertoire gkographique des textes cuneiformes (Vol. 3 = B. Groneberg, Die Orts- und Gewassernamen der altbabylonischen Zeit; Vol. 10 = J. Fincke, Die Orts- und Gewhsernamen der Nuzi-Texte) Revue hittite et asianique Royal Inscriptions of Mesopotamia Assyrian Periods Royal Inscriptions of Mesopotamia, Early Periods Reallexikon der Assyriologie siglum for texts from Ras ShamralUgarit State Archives of Assyria State Archives of Assyria Bulletin State Archives of Assyria Studies Sources from the Ancient Near East Studies on the Civilization and Culture of Nuzi and the Hurrians Schriften zur Geschichte und Kultur des Alten Orients Series Maior (Studia Pohl) siglum of Nuzi texts in the Harvard Semitic Museum Studien zu den Assur-Texten siglum of texts from Sabi Abyad Ph. Talon, Old Babylonian Texts from Chagar Bazar (Akkadica Supplementum 10) Tiibinger Atlas des Vorderen Orients Textes cuneiformes du Louvre Texts from Cuneiform Sources Texts in the Iraq Museum siglum for texts from Tell Mardikh~Ebla Travaux de la Maison de L'Orient (Lyon) Ugarit-Forschungen siglum for texts in the Staatliche Museen, Berlin Vorderasiatische Schriftdenkmaler Die Welt des Orients Wiener Zeitschrift fiir die Kunde des MorgenlandesIStudien Yale Near Eastern Researches Zeitschrift f i r Assyriologie
I
THE ANCIENT LANDSCAPE. PLANT AND ANIMAL REMAINS AS INDICATORS
PIHANS LXXXVIII, 2000
Settlement and Land Use in the Zone of Uncertainty in Upper Mesopotamia T. J. Wilkinson (The Oriental Institute, Chicago)
INTRODUCTION
Virtually the entirety of Upper Mesopotamia - that is the zone of rainfed cultivation in northern Iraq, northern Syria, and southeastern Turkey - could be defined as a zone of agricultural uncertainty. Despite the considerable risk that is inherent in cropping an area with such a wide interannual fluctuation in rainfall, very large settlements have developed throughout the Syrian and Iraqi Jazira, virtually up to the limit of rainfed cultivation. This begs the question as to what degree such large settlements could have been supported by cultivation of the surrounding areas. This paper will examine some of the factors that may have contributed to the development and growth of settlements in the rainfed farming zone of the Jazira, and suggest some reasons why there appear to be anomalies such as large centers growing in marginal areas where there were apparently insufficient resources to support them. The paper will also deal with issues that arise in the literature on the ancient Near East, specifically where some authorities regard sites such as Ebla, and indeed many other sites in northern Syria, as being reliant largely upon sheep and goat husbandry and very little upon cultivated crops (e.g. Gelb 1986). This discussion is based primarily upon field evidence derived from archaeological and landscape surveys, but it will also build upon theoretical considerations inferred from traditional agricultural systems in the region, and more generally from the agronomy of dry-land agriculture. The main focus is upon settlement and land use in the fourth and third millennium B.C., but issues of agricultural uncertainty in the second and first millennia B.C. and later will be addressed as appropriate. First the basic components of the Bronze Age Jaziran economy as recognizable by the field survey are outlined. Then factors that can result in the amelioration of the staple subsistence economy of Jaziran settlements are examined. Emphasis is upon the interaction between different sectors of the economy rather than on any single sector. THE ZONE OF UNCERTAINTY DEFINED
Agronomists at ICARDA, the center for studies of dry land agriculture located near Aleppo, sub-divide the Syrian Jazira into five agro-ecological or "crop stability" zones as follows (Pabot 1956; Cocks et al. 1988; see also Fig. 1 based on Jones 1993, fig. 3): Zone la: With rainfall > 600 mrn, this area exhibits a wide range of crops. Within Turkey, zone l a is characterized predominantly by wheatlbarley and lentils grown in alternating years, with vineyards occupying marginal soils (Wilkinson 1990).
4
T. J. WILKINSON
Zone Ib: Rainfall between 350-600 mm and not less than 300 mm in two of every three years. Main crops comprise wheat, chickpeas, lentils, fruit, vegetables and nonirrigated summer crops. Zone 2: Mean rainfall of 250-300 mm, and above 250 mm in two of every three years. Crops are barley, wheat, some food legumes and summer crops; livestock, mainly sheep and goats are of increasing importance. Zone 3: Mean annual rainfall greater than 250 mrn and not less than 250 rnm in one year out of two. Barley and livestock (mainly sheep and goats) are the main products, wheat is also grown but food legumes are of little importance. Zone 4: With average rainfall between 200 and 250 mm, and not less than 200 mm during half the years, barley and sheep or goats remain the primary products, barley being mainly grown to feed flocks. Zone 5: Even less rainfall than Zone 4 and extending into true desert, this area is not used for cultivation but is used for grazing, primarily by nomadic or semi-nomadic pastoralists. Where mean annual rainfall exceeds 325 mm wheat predominates, but barley may be important on poorer soils (Jones 1993:131; Fig. 1). Further south barley increases until the pastoral zone 5 takes over. The present limit of rainfed farming has been defined by the Syrian government as close to the 250 mm rainfall isohyet (Wachholtz 1996:19-20), but in the past it has formed a rather ragged edge which depends on a range of cultural and economic factors as well as on local soil and hydrological conditions. Thus soils along wadis may remain moist for greater parts of the year with the result that lobes of rainfed farming extend southward into the semi arid steppe. In drier areas most land is under barley, usually with an alternating year of fallow, but increasingly it is being grown as an annual crop. Because crop-production is not viable every year, grazing of animals increases towards the southern limit of Zone 4, and in some years barley is not even harvested but is directly grazed by the sheep and goats (see below). Wachholtz defines a zone of marginal cultivation between the rainfall isohyets of ca. 250mm and 180 mm (Fig. 2). For the area south-east of Aleppo (not strictly within the Jazira, but relevant to agriculture within it) Wachholtz recognizes first a "zone of permanence," where settlement is mainly in the form of sedentary villages; here rainfall is 250-300 mm per annum (and above) and most arable land is used for cropping. This is equivalent to zone 3 above.1 Second is a "zone of transition," which is mainly a semisedentary zone occupied by bedu, and where land use consists of both cultivation and grazing. This zone receives some 250 to 200 mm of rainfall per annum and is roughly equivalent to crop stability zone 4. Even drier is the "steppe rangeland" zone where rainfall is less than 200 mm per annum, occupation is by migratory bedu, and land use according to the Syrian government mandate is exclusively for grazing. The steppe rangeland zone falls within stability zone 5. A key point to emerge from the study of Wachholtz is that in the sedentary and semi-sedentary areas stability of family income is enhanced by the integration of crop production into their livestock systems thereby
Wachholtz considers this belt to be within zones 3 and 4, but from the rainfall figures it seems that this "zone of permanence," is more likely to be equivalent to zone 3 only.
SETTLEMENT AND LAND USE
5
securing a feed base (ibid. p. 195). The relevance of such integrative approaches to the ancient cultivation regime is enlarged upon below. In terms of settlement pattern, it is evident from figure 2 that most of the large Bronze Age tells occur where rainfall is slightly higher than defined by Wachholtz's belt of marginal cultivation. Nevertheless, many noteworthy sites do fall within this belt: Umm a1 Mara, Hadidi, Selenkahiye, Sweyhat, Hammam, Sabi Abyad, Tell es-Samir, Emar, Bica, as well as a number of Kranzhiigel sites within the Jebel Abd al Aziz steppe (Hole 1997; Kouchoukos 1998). Although some of these sites probably relied upon irrigation,2 palaeobotanical evidence suggests that those of the above sites that have been excavated appear to have relied to a significant degree upon rainfed cultivation (van Zeist and Bakker-Heeres 1985:198; Miller 1997). In addition to these sites within the marginal belt, many larger and even more important sites occur very close to this zone, for example Ebla, Chuera, Beyda, Brak in Syria, as well as Tells Khoshi and Khadail in Iraq. To understand how such major settlements developed, grew, and apparently thrived in such a climatically marginal location requires some knowledge of the structure of the Bronze Age economy. The margin of permanent settlement The limit of permanent settlement or rainfed cultivation is variable through time, depending upon political, social, economic, and environmental conditions. Thus during the twentieth century there has been a significant shift southward of the southern limit of permanent settlement in the Iraqi Jazira (Thalen 1979, Lewis 1987, Wilkinson and Tucker 1995). In the Khabur basin in the 16th century the limit of sedentary settlement extended from a little north of Ras al-'Ayn to Chagar Bazar and then to the north of the Wadi Radd (Hutteroth 1990). This demonstrates that even during a period that is often regarded as one of settlement decline, there was a significant number of sedentary villages in the northern Khabur basin, even as far south as Tell Beydar (Goyunq and Hutteroth 1997 map 3a). At this time the limit of settlement was blurred by the tendency of linear distributions of settlement to extend down the north-south wadi systems that were tributary to the Khabur. Similar settlement alignments were characteristic of third millennium B.C. settlement as can be seen on figure 4. However, this pattern was violated in the Iron Age when around sites such as Tell Beydar there was a tendency for small dispersed settlements to be developed over most parts of the landscape. This pattern appears to have had its roots in the Late Bronze Age, a period of some settlement decline in the Jazira. Because there is no evidence to show that the Late Bronze or Iron Age were any wetter than the third millennium B.C.3 it appears that such a dispersed pattern of settlement was due not to more favourable climatic conditions, but more likely was because political or economic circumstances were more conducive to settlement. It is therefore likely that the development of a dispersed pattern of small rural settlements appeared either because the area was selected by the Neo-Assyrian administration as an Specifically Emar and Tell Bica. Irrigation was probably also employed at other sites when needed; see for example the evidence for possible irrigation at Middle Assyrian Sabi Abyad cited by Wiggermann (this volume, 5 2). If anything, these periods were slightly drier than the third millennium B.C. according to Lemke and Sturrn 1997: fig. 5.
6
T. J. WILKINSON
SETTLEMENT AND LAND USE
area for the resettlement of displaced populations (Wilkinson 1995), or, the area was simply settled spontaneously by, for example, Aramean nomads.
mountains, and Afghanistan to the Mediterranean Sea. The existence of such wideranging economic catchments of differing size would have provided various degrees of buffering for the Jaziran communities so that droughts in any one area could have been mitigated by inter-regional transfers of food, and if economic circumstances allowed them to be transferred.
THE JAZIRA AS A THREE SECTOR ECONOMY
Large parts of the Jazira within Turkey, Iraq and Syria were densely populated during the third millennium B.C. when settlement occurred primarily in the form of tells, or various types of walled settlements. Populations were distinctly nucleated, and even though "rural" settlement occurred, even these were usually in the form of tells such as Tell Raqa'i and Hajji Ibrahim (e.g. Schwartz and Curvers 1992, Danti 1997). In the mid to late third millennium B.C. settlements that grew to 100 hectares in area may have housed as many as 10,000 or more (Wilkinson 1994). To be sustainable in such a climatically marginal area such sites are unlikely to have relied upon agriculture alone, but rather probably drew their sustenance from at least three main sectors of the economy (Fig. 3): Cultivated food products, primarily in the form of staples such as wheat, barley, and lentils, drawn from the immediate catchment. This can be termed staple production, as has been described by D'Altroy and Earle (1985). Pastoral products and the wealth derived therefrom, from large flocks of sheep and goats that were either pastured within the territory of the central settlement, or were attached to wide ranging pastoral nomads (Gelb 1986, Matthews 1974). Wealth generated by overland trade along well-defined route systems that crossed the Jazira up and down valleys, in roughly east-west alignments across the Jazira, and also along more oblique trajectories from north-west to south-east. Such routes, which are in places visible on the ground, can be traced on air photographs and some satellite images to link many large settlements as is discussed below (van ~ i e r and e Lauffray 1954; Wilkinson 1993). Although a wider or narrower range of economic sectors could be employed, I have selected the above three because they are the components that can be most readily recognized by field survey. These three economic sectors capture notional catchments of increasing size: subsistence based on staple products, owing to the high frictional effect of overland transport, would have been primarily restricted to a limited area around the main site, usually to within 3 km of the central tell.4 In the case of larger centers up to 100 ha, however, agricultural territories could be either simple catchments of 5 km radius or compound catchments up to around 15 krn radius (Wilkinson 1994). Some pastoral production would have been based within this local area, but allowance must also be made for the use, by village-based flocks, of outlying pastures when available, as well as the flocks of pastoral nomadic groups. The last named could have been loosely tied with the sedentary centers and, by analogy with traditional practice in the region, must have been much more wide-ranging (see below). Finally, trade would tap both resources and wealth drawn from a very wide area, ultimately perhaps from Oman to the Taurus 4
Cultivation around small villages can extend further, albeit at a lower level of intensity (Hillman 1973), and around Middle Assyrian Sabi Abyad, the small central settlement of some 0.36 ha may have controlled a cultivated field area of some 3 km radius (Wiggermann this volume; for further discussion see below).
7
In the context of the urbanization of EB WIII Palestine Portugali and Gophner emphasize interactions between trade, urbanism and agriculture as follows: "the urban system, due to accumulation, planning, long-distance trade and other features of urbanism, enabled the regional agricultural subsystems to absorb, or overcome their local instabilities, previously fatal during the agricultural era. Only when several local instabilities coincided did the urban system as a whole, or major parts of it, collapse or undergo a major crisis. Thus, relative to the agricultural era, the urban system was characterized by relatively long periods of socio-spatial stability, but short, but very intensive environmental crises. Urbanism thus reduced the frequency of socio-environmentalcrises, but increased their intensity Portugali and Gophner, 1993 Similar issues are examined here. Emphasis is placed on the heterogeneity of the ancient Jaziran economy and this is illustrate using examples drawn from the following areas (from west to east): Tell es-Sweyhat, Kurban Hoyiik, Balikh Valley, Tell Beydar and the Khabur valley, Tell al-Hawa. Sector 1: Staple production in the Jazira
During the third millennium B.C. sites such as Tell al-Hawa (Iraq) and Tell es-Sweyhat (Syria) were probably surrounded by large areas of intensive rainfed cultivation (e.g. Wilkinson 1982, 1994). Cultivated areas were initially estimated by off-site sherd scatters or "field scatters." These are evident as subtle scatters of pottery on the surface of modem ploughsoils, and are inferred to result from the spreading of settlement refuse, ash, and other household waste (inevitably including some potsherds) on fields as fertilizer. The date of the potsherds, mainly Early Bronze Age (EBA) in the case of Sweyhat and Hawa, or a minority EBA in the case of Kurban and Titrish Hoyuks, suggests that in the two first areas, Bronze Age fertilizing was significant, whereas in the second two cases manuring with settlement-derivedrefuse was apparently less important (Wilkinson 1982).
In the Jazira a second type of evidence occurs in the form of linear hollows which often radiate from Bronze Age tells. These hollows were probably formed as a result of traffic of humans and their flocks concentrated along linear tracks; consequently compression of soil, and associated erosion from overland flow of water resulted in the development of hollows (Wilkinson 1993). Such hollows can then fade out usually some 3-5 km from the sites. This fade-out point is inferred to represent the zone where the cultivated fields of the Bronze Age ended and either traffic to fields was less, or that traffic then radiated out across the steppe. In either case such fade-out zones can be used to estimate the long term limit of maximum cultivation.
8
9
T. J. WILKINSON
SETTLEMENT AND LAND USE
Together these two strands of evidence can be combined to suggest that most Bronze Age tells that exceeded 10 ha in area would have been surrounded by a cultivated area of some 3-5 km radius, the inner zone of which was frequently intensively cultivated and manured. Applying this and other principles to the Khabur Valley of northern Syria, one is able to produce a rough estimate of Bronze Age settlement and land use as follows.
economy that focussed on a narrow range of cereal crops and livestock husbandry. Furthermore during the third millennium there was a major restructuring of the economy from one comprising diversified husbandry of sheep, goat, cattle, and pig, to one dominated by sheep and goats. By making inferences from the off-site resources within extended site territories, it is possible to place such changes within a geographical context. On figure 4 those areas that fell beyond the inferred cultivated zones, if they were without signs of permanent settlement, field scatters, and linear hollows, can be inferred to have been likely areas of steppe pasture. Such areas are usually situated in areas of low rainfall andlor mainly have shallow soils. Of the areas defined as probable pasture on figure 4, the only area checked in the field by the author is that to the west of Tell Beydar, where a large area of basalt forms a plateau covered by thin soils (Fig. 7). Unfortunately, because this area and others like it on figure 4 show little evidence of human activity they can only be defined on the basis of negative evidence; that is as the residual of land that is left over when other types of land use (i.e. sedentary settlement and cultivation) have been accounted for. Consequently, on most site distribution and land use maps of the original surveys, such areas only appear as open space. The contribution of the pastoral sector is best illustrated from the following examples, which are updated from original surveys conducted by the author.
Detailed survey around Tell Beydar demonstrated that third millennium settlement was primarily restricted to tells greater than 5 m in height (Wilkinson n.d.) Thus if tells can be recognized on 1: 200,000 maps and CORONA satellite images, and cross-checked where possible by published surveys, it is possible to produce a map of tells. These can be tentatively regarded as a proxy indicator of EBA settlement. Using the buffering mode of ArcView GIs5 the 3 and 5 km radius zones around each tell can be mapped to provide an estimate of probable cultivation (c3 lun). Probable pasture is then estimated as the land beyond 3 or 5 km. The intermediate zone being probably a combination of both pasture and cultivation; with a bias one suspects, towards the former. Figure 4 shows the overall distribution of tells and land use zones in the western Khabur basin according to this principle. It indicates that the basin would have fallen into: An eastern zone (i.e. the central Khabur basin) of dense, almost continuous occupation, with one or two north-south open areas that could have served as pasture. A western zone with linear distributions of tells and cultivation along wadis and some open space between. A southwestern area around the Jebel Abd al-Aziz, where Bronze Age settlement and cultivation was even more limited, and pastoral production must have made a significant contribution to the economy (Kouchoukos 1998; Hole and Kouchoukus n.d.). Although the above method provides only a rough estimate of settlement and land use in the Early Bronze Age Khabur, the reconstruction demonstrates the heterogeneous nature of settlement and land use which formed a mosaic that was sometimes virtually continuous, but also patchy or linear, and in the driest areas included extensive areas of open steppe with few, but often large, settlements. Sector 2: The contribution of pastoral resources ArchaeoIogicaI surveys tend to under-estimate the contribution of pastoral resources because they frequently focus more on the sedentary component of settlement. Faunal analysis, on the other hand, provides explicit information on the animal and pastoral economy, and in a number of recent papers Melinda Zeder (1998, 1999), has pointed out that in the third millennium B.C. the Khabur basin probably shared a region-wide That is Geographical Information Systems that use computerized mapping techniques to analyze complex arrays of spatial data.
I . Tell es-Sweyhat (Syria: ca. 250 rnm per annum) The original landscape studies around Tell es-Sweyhat focussed upon the immediate catchment of the site. These studies suggested that in the late third millennium B.C. the site was probably encircled by a 3 km radius zone of intensive cultivation (Wilkinson 1982). Later mapping of a wider area by Michael Danti (Danti and Zettler 1998) as well as bioarchaeological studies by Miller (1997) and Weber (1997) have demonstrated the significance of pastoral resources and hunting to the Sweyhat economy (Fig. 5a). This is supported by my own re-mapping of a wider area that indicates that the area of off-site sherd scatters of the inner intensively cultivated zone was itself surrounded by extensive areas of plateau steppe and hillslopes, all of which would have provided extensive areas of steppe pasture (Fig. 5a). Clearly, as in most of the above areas, available pasture land was well in excess of cultivated land. Furthermore, these areas formed an interlinked series of semi-continuous steppe pasture that extended to the east, probably as far as the B alikh. 2. Kurban Hoyiik (Turkey: ca. 450 mm per annum) This 6 ha tell site along the Turkish Euphrates is located in an area of moderately plentiful rainfall. It would have drawn its staple products mainly from the main terrace along the Euphrates (Wilkinson 1990), but in addition, upland steppe on terraces and limestone hills to the south would have provided a substantial area of pastoral upland that would have interconnected with pastures of neighboring tells as well (Fig. 5b). A similar situation would have prevailed around the large center of Titrish Hoyiik to the south-east where, in addition to a spacious agricultural lowland, there existed extensive areas of upland pasture. Detailed survey suggests that although those uplands adjacent to the site may have been cultivated and perhaps also used as threshing floors, the remainder were probably upland pasture.
10
11
T. J. WILKINSON
SETTLEMENT AND LAND USE
Therefore around both Kurban and Titrish Hoyuk, although there was a significant area of cultivation, pastoral resources were probably significantly more extensive than cultivated land.
supported over the long term by rainfed cultivation. Kouchoukos suggests that the growth of these major Kranzhiigel sites was stimulated by the "commodification of textiles" that is by trade in wool, animals and textiles generated by pastoral communities affiliated with the Kranzhiigel sites. Although recognizing the importance of the pastoral nomadic economy, I suggest that the growth of such centers was also linked to the development of trade routes across the steppe as will be discussed below.
3. Balikh Valley (Syria: 200-300 mm per annum) The Balikh Valley has rolling uplands to both east and west that are devoid of Bronze Age sites as well as occupation of most other periods. These open spaces would have provided extensive upland steppe pasture. As indicated on Figure 6a, site territories estimated from Thiessen Polygons suggest that every EBA site along the River Balikh would have had ready access to steppe pasture. Limited mapping of the lowland along the Balikh suggests that when sherd scatters occur (i.e. proxy indicators of early manuring), they are Hellenistic through early Islamic in date; there is little evidence of Bronze Age sherd scatters (Wilkinson 1998). Therefore cultivation was either not intensive, or farmers were able to utilize plentiful animal manures that would have been generated by the numerous animals that grazed the steppe.
4. The Khabur Basin The above analysis of settlement in the Khabur basin suggests that pasture land was present in varying amounts and was mainly limited to occasional north-south corridors, usually along the lines of watersheds, as well on geologically determined uplands. The Tell Beydar area provides a small window into the settlement distribution within the Khabur basin. TELL BEYDAR (SYRIA. CA. 300 MM PER ANNUM)
Situated on the minor north-south Wadi Awaidj in the western Khabur, Tell Beydar was located between an area of low uplands to the east, (which were probably uncultivated during the third millennium B.C.), and a basalt plateau to the west (Fig. 7). The latter area had a cover of thin soils and was without evidence of archaeological sites apart from rare traces of post-Iron Age occupation (Wilkinson n.d.; Barbanes 1999). Frequent references in the Tell Beydar texts to shepherds and their sheep implies that the basalt area formed a major pastoral resource during the third millennium B.C. Certainly this has traditionally been the case, and the basalt uplands are still used in summer by bedu who have traveled up from the winter pastures in the deserthteppe to the south. During the third millennium B.C. cultivation was probably mainly confined to the lowlands of the Wadi Aweidj, as has been estimated on figure 4. THE JEBEL ABD AL-AZIZ STEPPE
To the southwest of the Khabur on the semi-arid steppe that surrounds the Jebel Abd alAziz a number of Kranzhiigel sites have long been recognized, and have recently been surveyed by Frank Hole and Nicholas Kouchoukos (Hole 1997; Kouchoukos 1998). These sites, although situated within areas of extremely low agricultural potential appear to have been surrounded by relatively limited areas of cultivable land. Such land, which would have been watered by the rather unreliable annual rainfall perhaps supplemented by runoff from the Jebel Abd al-Aziz, is estimated to have been less than was required to support the estimated population of those sites (Kouchoukos 1998). Whether the cultivation base was sufficient to support the estimated population of the sites or not, it is clear that such sites were climatically extremely marginal, and could not have been
5. Tell al-Hawa area {Northern Iraq: ca. 350 mm per annum) Landscape surveys around this 55 ha site in the Iraqi north Jazira demonstrate that in the third millennium B.C. most of the area was probably cultivated land (Wilkinson and Tucker 1995). However, an extensive zone of open space can be inferred to have developed in the western part of the survey area (Fig. 8) in the late fourth millennium B.C.6 This occurred as a result of the abandonment of Late Chalcolithic and Uruk settlements in the second half of the fourth millennium B.C. Thus the newly opened up pastoral area can be inferred to have been analogous to an area of enclosed pasture or nomadism (cf. Rowton 1973) that was virtually surrounded by settled areas. Nevertheless, in this relatively densely occupied area it appears that cultivation (much of it intensive) was as abundant (or more so) than pasture, and the pastoral areas remained as isolated enclaves rather than interlinked units with free passage between. Most of the Bronze Age catchments discussed above exhibited a viable base of rainfed cultivation. With the exception of Tell al-Hawa, located in the area of highest cultivation potential, all had access to large areas of pasture as well. The relative degree of reliance on pasture can be seen to have increased towards the south and the desert. This presumably is not because these formed the best pasture lands, but because areas to the north provided greater potential for cropping and eventually became primarily used for cultivation. Pasture lands appear to have extended freely away from most sites, and in most cases would have been seamlessly linked with other areas of steppe pasture accessible from other sites. In the Khabur linear zones of cultivation would have been separated by corridors of pasture that could have allowed free passage of pastoralists from north to south or vise versa. However, such freedom of movement would have depended upon ownership arrangements or rights to common pasture along these corridors. To judge by the availabiIity of open areas with high potentia1 for pasture, most Bronze Age settlements were therefore within convenient distance of large areas of potential pastures (Danti and Zettler 1998, for Tell es-Sweyhat). This provides support for the evidence from cuneiform texts that refer to a large pastoral sector (Gelb 1986). It is also noteworthy that around Tell al-Hawa, what can be interpreted as a pastoral zone opened up during the late fourthlearly third millennium B.C. This implies that pasture and sheep may have been sufficiently important to the economy for enclosed pastures to have been deliberately created, perhaps at the expense of settlement in agriculturally marginal areas.
In addition there were small areas of possible pasture on rolling upland to the north of the site.
12
T. J. WILKINSON
I
Sector 3: Trade and the wealth economy Archaeological survey does not provide direct evidence for trade or those sectors of the economy that might have generated wealth or its accumulation. Nevertheless some indication of these sectors must be sought in order to deal with larger issues of the economy such as are alluded to in the following statement: "One may judge the Ebla situation by scanning Pettinato (1979). One feels overwhelmed by the number of the texts dealing with wool, textiles, metals and metal products." Gelb 1986:158. From this Gelb goes on to conclude that agriculture (i.e. cultivated agriculture) played a secondary role in comparison to the manufacture of textiles and metal products (Gelb 1986:160). In the context of trade, the movement of luxury goods, and administration, field survey can provide two types of indirect evidence. These are inter-site linear hollows and the presence of expanded settlements, often with outer towns. Inter-site linear hollows These form broad shallow hollows that radiate from Bronze Age tells, but in contrast to linear hollows that fade out at the boundary of cultivation (see above), certain features continue across the landscape to link up with other sites in the settlement network. Within many parts of the Jazira these hollows provide a tantalizing glimpse of what may have been a complex network of Bronze Age cross-country routes. Because they link third millennium B.C. sites, these linear hollows appear to have been in use at this time. Although it is difficult to assign a precise function to such routes it can be suggested that in addition to providing general purpose lines of communication, they were used for administrative communications and by merchants, traders and their caravans. Expanded settlements The larger Bronze Age settlements in the Jazira fall in the size range of 50 -100 ha and rarely exceed 100 ha (Wilkinson 1994: table 2 and fig. 3). Because the agricultural potential of the Jazira increases to the north as rainfall increases, one might expect the largest sites to be found in the north where agricultural potential was greatest. This is not the case however; although some large sites do occur in the wetter part of the Jazira (Kazana, Leilan, Girnabaz, and Mozan spring to mind), a significant number occur much further south, close to or even beyond the limit of viable rainfed cultivation. For example a noteworthy series of large sites occur along an east-west alignment as indicated on Figure 9: Nineveh, Taya, Hadhail, Khoshi, Mashnaqa, Tell Mu'azar, Tell es-Seman, Sweyhat and Hadidi. Of these Taya, Hadhail and Khoshi are around 100 ha area or more, and the others, although more modest in size (Tell Mu7azar is ca. 11 ha: Kouchoukas 1998: table 7.4) are nevertheless much larger than would be expected from their climatically marginal position. Although it may well be correct that these sites benefited from a slightly higher rainfall during the early third millennium B.C. (Hole 1997), their growth probably resulted from additional factors that also stimulated the accumulation of wealth7 which in turn encouraged growth which resulted in an increase in the size of settlements.
See also Meijer (in press) who makes a similar case for the formation of Kranzhiigel sites. I am very grateful to Dr. Meijer for making this interesting paper available to me in advance of publication.
I
I
SETTLEMENT AND LAND USE
13
Two processes can be suggested for the generation of such wealth and associated settlement growth. First settlement expansion was stimulated by trade along selected major routes across the Jazira. For example it is possible that these sites fell upon a major east-west route system linking Nineveh on the Tigris to Tell Hadidi on the Euphrates, and onwards towards the Mediterranean coast (Fig. 9). As a result of the growth of interregional trade in the mid to late third millennium the passage of large numbers of merchants across the Jazira would have increased the demand for food at various stopping points on the route. This demand would mainly have been for basic staples, but there would also have been increased demand for meat, other animal products, and feed for pack animals. In addition, greater administrative links with Akkad would have required administrative contacts which again would have placed greater demands on local food production systems, especially for entertaining and f e a ~ t i n gSecond, .~ exchange of commodities between nomadic and sedentary communities in different parts of the Jazira could have increased demand for food and generated additional wealth (Meijer n.d.). Being on the boundary between steppe pastures to the south and mainly cultivated areas to the north, the above-mentioned sites could have been the focus of exchange between sedentary cultivators and more mobile groups.9 Trade and exchange at the boundary between geographical regions leads to the growth of "gateway cities" (Burghardt 1971), and such processes can plausibly be used to account for the growth of many large centers in the Jazira. Such interactions would probably have led to increased sedentary settlement of nomads, as well as increased demand for grain for the nomads themselves, and fodder for their animals. This would have resulted in both the growth in size of the settlement and an increase in the scale of agricultural activity. Expansion as a result of the above mechanisms of trade and exchange may account for the anomalous expansion of a site like Tell Taya, which exhibited a massive and extensive lower town that has always been difficult to explain (Reade 1982:77). Although Reade considers that this extensive spread of suburbs may be characteristic of many similar small tells in the Jazira, according to intensive survey around many sites in the Jazira, this does not frequently seem to be the case.1•‹Increased trade or exchange with nomads may also account for the expansion of Tell es-Seman, the largest EBA town on the Syrian Balikh (ca. 12-15 ha in area; Curvers 1991), as well as the ca. 100 ha walled sites of Tells Khoshi and Hadhail to the south of Jebel Sinjar. Despite their climatically marginal location, during periods of economic or agricultural stress these communities were perhaps able to survive by "buying in" grain and other staple crops as necessary. Whether such an alignment of sites is explained by the existence of a major east-west route through the Jazira or was a result of interaction and exchange between sedentary and nomadic communities, is difficult to say. If a major east west route once existed, its Sites such as Tells Khadhail and Khoshi in northern Iraq may have developed where the postulated east-west route was intersected by routes trending north-west up the Wadi Tharthar from southern Iraq (Fig. 9). See Meijer in press. An additional wealth generating factor, the primary extraction of minerals, is not considered here because the Jazira has few mineral resources. However basalt for grinding stones (e.g. around Tell Beydar) and salt or salt pans were both probably exploited in the third millennium B.C. Rather intensive surveys around tells suggest that extensive lower towns in the Khabur are more likely to be Late Bronze Age or later in date. Only in limited cases such as around Titrish Hoyiik, Tell Sweyhat (the southern extension Zettler 1997) and Daudpasha (in the Amuq) much further to the west and north did EBA lower towns develop. None exhibit the spectacular extent of Tell Taya.
14
15
T. J. WILKINSON
SETTLEMENT AND LAND USE
trajectory, to the south of the Jebel Sinjar and Abd al-Aziz, would have taken it toward a major crossing of the Euphrates. This crossing is implied by two pairs of towns: Tells Hadidi and Sweyhat (to the north), and Halawa and Selenkihiya (to the south: Fig. 9). Such paired settlements often grow up where major routes cross rivers, and are a characteristic feature of river crossing points in the USA (Burghardt 1959). In the Tabqa area paired settlements had disappeared by the Late Bronze Age, when they were replaced by a series of fortresses distributed in alternating positions along the river. In the Jazira it can therefore be argued that sites did not expand in proportion to the agricultural potential of the area. Rather, the largest sites probably developed in areas that were favorable for the generation of wealth, which in addition to agricultural resources, would have included long distance trade and exchange with mobile pastoralists. For example major settlements like Tell al-Hawa (Iraq) developed on inter-regional hollow way routes (Wilkinson and Tucker 1995). Such routes probably formed only one of many such systems, so that travelers may have shifted from route to route depending upon political circumstances. However, not all significant tells appear to have fallen on major routes, and access to the wealth-generating mechanisms of trade and exchange, would therefore have been unequal across the Jazira.
highland Anatolia, Afghanistan or the Gulf. Such variations in economic scale must have provided a considerable buffer for the local staple economies. Not only was the Jaziran economy heterogeneous, but also it probably exhibited complex links between the different sectors. The staple economy should not be regarded as simply the production of grains for local consumption by local sedentary communities; rather the pastoral and trade components would also place demands on the supply of staples as follows. First population growth and increased trade must have increased the demand for staple products through time (see above p. 12). Any substantial increase in the number of domestic animals would have increased the demand on barley for feed, especially for winter feed. Although it is true that the Bronze Age flocks would have utilized the natural forage resources during most seasons11 it seems from Old Akkadian texts from Tell Beydar that grain was used as animal feed (Meijer in press; Van Lerberghe 1997: 120-21). As local forage became reduced in areas of overgrazing, there would have been an increased need for animal feed. In addition surprise events like wars or the movement of armies through the area would increase the demand for food.12 Such rising demand could have.been met by increasing the intensity of production; nevertheless in dry years it is quite likely that production systems may have suffered a crisis. Because the Jazira receives most of its weather from similar air masses and weather systems, a dry year in one part of the Jazira is likely to have been dry throughout the Jazira (Wilkinson 1997, fig. 2). However, this is not always the case (Butzer 1997). Thus as indicated on figure 10a, when rainfall at Kamishli and Mosul is compared with that at Aleppo we can see that there is a high visual correlation between Aleppo and Kamishli but a much lower correlation between Aleppo and Mosul. In years when Kamishli received more rainfall than Aleppo, it would have been potentially feasible to export grain from Karnishli to Aleppo. When we consider Aleppo and Mosul, where the correlation was even lower (Fig. lob) there would have been even more opportunities (and reason) to export across the entire extent of the Jazira. Also those parts of the Jazira that were located in close proximity to much moister areas (for example the eastern Khabur) would have been able to bring in food from adjacent areas that remained reasonably well watered and supplied. Although such shifts of produce mav have taken place, they would have been difficult, in part because it is very expensive to move sufficient grain to feed a town of some 2000-10,000 people across the land.13 Furthermore, during major droughts and
THE BRONZE AGE JAZIRA ECONOMY
The ancient Jaziran economy may have operated as a patchwork of staple-based economies in which one main engine was the distribution of staple products, while the other comprised wealth generating economies in which light, high value items such as metals, animals or textiles were exchanged or accumulated. As a result of the spatial differentiation of economies into various mixes of staple cropping, pastoral production, and wealth-generating components, it is possible to suggest how settlements may have responded to conditions of stress. During dry years, settlements that were over-reliant on staples would not have the resources to buy in grain whereas others, perhaps in more marginal areas, would have had such potential. Therefore sites located in drier climatically marginal areas could have been able to override a drought simply by virtue of their ability to draw upon their capacity to generate some wealth. More likely than this straightforward picture is the more general statement that the Bronze Age Jaziran economy was rather heterogeneous and therefore could respond in complex ways to perturbations, be they environmental or socioeconomic. Furthermore the presence of a nomadic sector, the members of which had their own methods of evading dry years by the use of mobility to graze areas of optimum rainfall or pasture, would have supplied another component of flexibility to the Jaziran economies. Simply by providing a way of mobilizing items of high value (that is flocks and their products), pastoralists would have provided a flexible way of moving economic assets from place to place. Although sites located along the east-west alignment noted above may have benefited significantly from the exchange and sale of animals and their produce, this was not always the case. As Reade has argued (1982), there were times when such groups would have been in opposition to the occupants of the sedentary zone. Each part of the three sector economy related to a different economy of scale. The staple products would have been (usually) produced and consumed locally; flocks and the products of pastoral nomads would have ranged more widely, perhaps penetrating into the northern Arabian desert; finally, the trade economy had its roots even further afield, in
1 There has been a massive increase in the hand feeding of flocks over the last forty years, especially as cultivation has encroached on the steppe and the steppe vegetation has become further degraded (Wachholtz 1996:78). Butzer 1997. See also Fales 1990a for examples of the massive demand for food by the Assyrian -Y. 1 3 That transport on land (by pack animals either alone or drawing wagons) is more expensive than water transport is well known (Weiss 1986:94-95). For example, Foster notes the allocation of 5% of the harvest at Ugarit towards transportation costs (1986:115). However the "frictional effect of distance" should not be over stated because long-distance trade in grain clearly existed back to the third millennium B.C. (Silver 198562). Such movements were probably undertaken only in times of extreme stress, or if the resources (economic or political) were available to pay for it. Thus if the economy was mainly founded upon grain, paying for the transport of the grain would have been difficult, if not impossible, unless there was plentiful grazing on route, a situation that would not have prevailed during drought seasons or in times of overgrazing and degradation. On the other hand, if there were abundant non-staples such as metals available, payment for grain would have been easier, especially in the first
16
17
T. J. WILKINSON
SETTLEMENT AND LAND USE
famines, such movements would have been vulnerable to interception by equally starving communities en route. As an alternative, however, it is well known that domestic animals provide an excellent means of storage on the hoof and they can be exchanged or bartered when required. Consequently domestic flocks can be fed with surplus grain when it exists14 and can be moved more freely from place to place during times of food stress. Thus during periods of famine owners of flocks would almost certainly try to sell or exchange some animals, presumably in exchange for grain. For example during the recent Sahelian droughts some inhabitants benefited from the fall in the price of domestic livestock, (which fell in price because most people were selling their assets), as well as the rise in the price of grain (as there was competitive buying by many of grain that was owned by few: Mortimore 1989). During such times of crisis it might therefore have been more viable to transport the flocks through existing pastoral corridors to areas where feedstuffs were more readily available and the relative "price" of animals was more favorable. Such movements would have been easier to undertake than transporting cereals from place to place, but on the other hand this process would have been beneficial to a different sector of the population. The Bronze Age economy was probably sufficiently heterogeneous for there to be economic disparities across the Jazira that could have been met by some commodity flows from place to place. Metals, textiles and other high wealth items could have moved via gift exchange and trade, and some would have been exchanged for either animals or staple crops. Fundamental inequalities between the wet staple-providing areas in the north, and drier areas with more animals to the south, could therefore have been met by such commodity flows. As to whether such flows formed a significant part of the economy and whether they were administered predominantly by the temple, palace, or household/"private" sector is beyond the scope of this paper. Nevertheless, the economy was not static and controlled, but more likely, was heterogeneous and flexible.
long periods of time. This is especially pertinent to the Bronze Age when archaeological records suggest that many tells were occupied for hundreds or even thousands of years. In addition, fallowed land often develops a surface crust, which, by decreasing the infiltration capacity of the soil can lead to increased runoff and erosion, especially on sloping terrain. Therefore although biennial fallow appears to have been practiced since at least Late Assyrian times (Fales l99Ob: 119-21), the practice clearly exhibits both short term advantages and long term disadvantages. On the other hand annual cropping, by violating the water preserving qualities of fallowing results in lower crop yields per field. However because every field can be cropped every year the practice can result in higher gross yields. This is illustrated by figure 11 which is the result of a preliminary simulation of crop yield for the Tell al-Hawa area for one hundred years.15 Diamonds indicate crop yields for biennially cropped barley; these plunge to 0 during each fallow year. Grey squares indicate yields from annually cropped barley. The two trend lines show that yield in annual barley declines through time (solid line) whereas barley production increased very slightly. The latter increase is in line with a very slight increase in the simulated rainfall. In the long term it can therefore be argued that the fallowed fields, despite a slightly greater decrease in organic carbon provide increased yields per field and greater cropping stability than annual cropping. One way of combating both the loss of organic carbon from the soil, as well as the loss of nutrients, is the practice of manuring. This can be accomplished by the application of composted animal dung, by encouraging animals to graze on stubble fields, or to remain overnight on fields so that their dung provides a natural manure. In addition, the application of household refuse, ash, and organic waste will supplement the degradation of soil fertility through time (Wilkinson 1982, Miller and Gleason 1994). This practice is well illustrated for a traditional Middle Eastern society in highland Yemen. In small villages, household waste, usually consisting of ash cleared from hearths and ovens, is dumped outside the house in small oval stone-lined storage units. This ash, often being derived from burnt dung, forms a pale grey dust-like deposit which inevitably incorporates inorganic materials from the household. The accumulated waste is then carried out to the fields on a donkey and applied as fertilizer. As a result of such applications those Yemeni fields that are fertilized accumulate not only ash, but also a small but significant accumulation of potsherds and other inorganic inclusions. When applied to biennially fallowed land, such organic fertilizers have the advantage of arresting the decline of organic matter, adding nutrients to the soil, and in general allowing soil moisture retention to be maintained. Ideally therefore combining fertilizing with biennial fallow should provide a more sustainable system of cultivation than fallowing alone (Wilkinson 1997:79-86). There is often more overlap between cropping and pasture systems than is generally assumed. For example in particularly dry years it is often the practice to abandon land
Amelioration practices in the agricultural economy At the scale of the local economy similar flexibility probably allowed the Bronze Age Jazira farmers to change their production strategies and agricultural output to offset adverse conditions or to increase production. Some methods could have ameliorated conditions (e.g. fallowing by increasing soil moisture) whereas others would have amplified certain problems and could have rendered the agricultural system potentially unstable. The following are just some examples of such amelioration practices. Fallow is normally practiced biennially. In addition to allowing the soil to recover, it enables the soil to store moisture from one wet season to the next. As a result some 1020% of the previous year's rainfall is carried over to the next year to supplement that year's rainfall. This carry-over acts as an insurance against dry years (although not major droughts: Wilkinson 1997 for references). Unfortunately, this practice also results in greater loss of organic carbon from the soil so that the quality of the soil can progressively deteriorate. Ironically, the loss of humus also diminishes the water holding capacity so that theoretically, the advantage of the fallow year may have been lost over millennium and later when coinage became widespread. It remains likely therefore that during the Bronze Age and earlier most transport of grain would have been mainly confined to within the local region. Often in the case of old grain see Wiggermann's paper, this volume.
l 5 This was developed as part of a programme of modelling crop production around ancient Mesopotamian settlements being conducted jointly between the Oriental Institute, University of Chicago and the Decision and Information Sciences Division of the Argonne National Laboratory, Illinois. I am extremely grateful to John Christiansen for running the simulation. The model output is based upon modern barley strains, therefore is significantly higher than would be expected with ancient barley. Barley yields are generated by the EPIC crop model using modern data for soils, rainfall, and other environmental inputs for the Iraqi north Jazira.
18
T. J. WILKINSON
SETTLEMENT AND LAND USE
under growing cereals and allow sheep and goats to graze on the failing crop (Nordlom 1983). Rather than suffering a total loss of crop, the farmer will have therefore have produced an enhanced grazing resource for his own flocks or even nomadic pastoralists. The latter, if they graze those fields can develop a reciprocal obligation or debt, which then requires repayment at a later date. As a result crop loss on the one hand will be balanced by increased social investment on the other. A related practice is that of harvesting the green cereal early. This might occur if it is anticipated that the year will be too dry to produce good well filled grain. The premature harvesting of the green cereal results in a dish called friki, a naturally sweetened and nutritious porridge. Not only does this practice allow farmers to salvage a poor crop, but also the harvesting of cereals early has the advantage of avoiding the collection of tax on the harvested crop (Abd al-Amir al-Azm, personal communication August 1998). Irrigation is often perceived as being practiced exclusively in arid areas where there is too little rainfall for rainfed cultivation. But because irrigation is also employed for the intensification of agriculture, it can be introduced into areas of rainfed cultivation where it is necessary to increase production. This apparently took place in the Balikh Valley in Turkey where Islamic canals were dug in an area that formerly was almost certainly not irrigated (le Strange 1905). The construction of large canals along modest-size rivers like the Balikh and Khabur would therefore decrease the overall base flow of the river so that downstream communities, who relied entirely upon irrigation for their agriculture, would be deprived of some of their flow. Such practices may have resulted in the water disputes recorded upstream of Tutu1 in the Mari records (Villard 1987; Wilkinson 1998). Thus intensification upstream, where irrigation was less important, could result in a loss of waters downstream where irrigation was crucial to survival. In a similar manner climatic drying would encourage the increased withdrawal of irrigation water thereby exacerbating the effect of the drought.
settlements that consistently exported grain to the growing centers, unless, again steps were taken to redress such losses. Even pastures are affected, because if sheep graze on outlying pastures and are then stalled around the settlement or on neighboring fields, there will be a loss of nitrogen and various trace elements from the pastures (White et al. 1997). Conversely the stalled area would receive a net gain because livestock contribute nitrogen from urine and feces. In general there would be a net flow of nutrients and trace elements depending upon the practice of animal husbandry and the duration and location of stalling. If there is a net loss from the pastures then it can be assumed that they too would become degraded. Today the Jaziran steppe appears extremely degraded (Thalen 1979), and has become even more degraded in recent years as a result of the extension of cultivation into progressively drier areas and the increased use of pastures. The above mentioned nutrient loss is therefore in addition to the more conspicuous and better attested processes of degradation.
Otherfactors that in$uence traditional and/or ancientfanning practice The above factors provide buffers against some of the fluctuations in agricultural production, but in addition a number of other factors can have an impact on settlement and land use in the zone of uncertainty. Nutrient recycling and transfers It should not be assumed that the soils of the semi arid Jazira have remained the same since initial cultivation took place, probably in the ceramic Neolithic (i.e. the Hassuna or Halaf periods). Unfortunately there is little evidence which allows us to describe the original soil resources in details (but see Courty 1994 and Blackburn 1998 for varying interpretations), therefore the following observations are theoretical rather than based on observation. Many Bronze Age tells were probably occupied for several millennia, and the annual or biennial cropping of nearby fields would have progressively withdrawn nutrients from the soil unless there was a deliberate attempt to return them, for example by manuring. Even though manuring was probably carried out in many areas, the more distant fields would probably not have received manure, therefore we may surmise that such areas were steadily depleted of nutrients and organic matter to the point when they were significantly degraded. Similar losses could have occurred around outlying smaller
19
Anticipation of total annual rainfallfiom autumn rainfall In principle, because total annual rainfall may be correlated with autumn rainfall, it should be possible for the farmer to anticipate whether the year will be a good or bad year from the autumn rainfall. However, data from Mosul from 1923 until 1955 suggests that there is only a weak correlation between rainfall from September to December and for the entire year so that such predictions will be unreliable. The correlation from other stations is higher, however, so that for Urfa especially dry years can sometimes be forecast from the September to December rainfall (Fig. 12). For example in the hydrological years16 1972173, 1990191, and 199-5196 a dry autumn would have been a good predictor of an overall dry year. Nevertheless, in spite of the reasonable visual correlation on this graph, it is evident that the farmer would have to use such a relationship with caution. Similarly when fallowing is employed to carry over rainfall from the fallow year to the year of cropping it is possible that a good year will result if the previous year was wet (see above). Unfortunately although this factor does apply in the Jazira, it does not appear to provide a reliable rule for the farmer. THE JAZIRA ECONOMY THROUGH TIME
Although quite a lot is known about the details of the Jazira economy in the fourth, third and second millennia B.C., the overall structure and dynamics of the economy remain less understood. It is likely however that the economy changed considerably over this long interval. During the Chalcolithic the agricultural economy was probably a mixed farming system characterized by reliance upon diverse and flexible practices that included husbandry of sheeplgoats, pigs and cattle, hunting (Zeder 1998), together with low intensity cultivation. By the period of state formation toward the middle of the third millennium, agricultural systems appear to have been much more specialized with more evidence for intensive cultivation, at least around major settlements (Wilkinson 1994), and some separation of pasture into distinct reserves (such as at Tell al-Hawa). There was also a specialized focus upon a narrower range of domesticated animals than formerly (Zeder 1998). Spatially, as has been suggested above, the economy was probably rather That is corresponding to a year starting in September and finishing in August. This keeps an entire winter's rainfall within the same year.
20
21
T. J. WILKINSON
SETTLEMENT AND LAND USE
heterogeneous with an underpinning of staple production as well as varying degrees of input from animal husbandry (both nomadic and sedentary) and wealth derived from trade and exchange. Both animal husbandry and tradelexchange would have provided a buffer against times of adversity, especially in climatically marginal locations. Thus from the fourth millennium to the late third millennium B.C., the structure of the economy probably changed from more localized mixed agricultural economies as noted above towards, by the mid-late third millennium, more specialized but much larger scale economies that drew upon a wider range of buffering mechanisms.
that all communities would have suffered from a critical food shortage, but some more than others (Weiss et al 1993).
CONCLUSIONS
The above summarizes only some of the complexities of the agroecological and economic systems of the Jazira. It appears that numerous buffering mechanisms were in operation to ameliorate crop production at the level of the farm as well as at the regional level. The political and subsistence economy was heterogeneous (see also Zeder 1999) so that different parts of the settlementleconomic system would have shown a different response to any perturbations of the environment or economy. The various buffering mechanisms would have provided a degree of flexibility to the economy so that during times of climatic stress, areas that were more economically robust may have been able to ride out some years of drought. On the other hand, areas that were economically marginal would have suffered more from economic changes such as shifting trade routes. Even if region-wide drought prevailed, some communities would have been better equipped in terms of resources to "exchange" their way out of the crisis, at least for the short term. During drought years there was probably increased exchange across the region, or conversely there would have been a flow of people from one area to another. Furthermore certain wealthier communities were in a position to benefit significantly from climatic crises. This is not to suggest that a market economy was necessarily fully developed during the Early Bronze Age, but rather that there was significant spatial variation in the economy and that economic mechanisms were in operation. This should be taken into account when we examine widespread or long-term changes in settlement. By the mid to late third millennium B.C. there is compelling evidence to suggest that systems of local economies had become transformed into a region-wide economic system as urbanization increased (Zeder 1999). The spatial mosaic of alternations of pasture and cultivation with occasional north to south corridors of steppelpasture, would have contributed to this region-wide integration. Potentially these steppelpasture zones could have provided additional conduits for interaction within a complex regional economy which would have provided the framework for response to occasional shortages of grain or straw in the south, or sheep and goats in the north. It is also necessary to distinguish between settlements that were marginal in terms of climate and those that were economically marginal. For example a site like Tell esSweyhat was climatically marginal, but during times when cross-river trade was vigorous, the town would have been in an economically viable position. If however the site ceased to become a locus of economic activity then it would have been rendered economically marginal in addition to being marginal climatically. This change would therefore have been critical to its long-term survival. In extreme cases major catastrophes, such as an intense drought, could have over-ridden buffering mechanisms with the result
Acknowledgments
First I must thank Remko Jas for hosting an excellent meeting in Leiden in May 1999. Thanks also go to my colleagues at the Oriental Institute, University of Chicago and the Decision and Information Sciences Division of the Argonne National Laboratory, Illinois who have contributed many ideas on the subject of the agricultural economy of Upper Mesopotamia. I particularly wish to thank John Christiansen (DISIANL), McGuire Gibson, Steven Cole, Colleen Coyle, Carrie Hritz, and Jason Ur (all of the Oriental Institute), for their help and advice during the modeling programme. In addition I must thank Prof. Dr. Sultan Muhesen, Director General of Antiquities in Damascus, Dr. Mu3ayyad Sa'id Damerji, Director General of Antiquities and Heritage, Baghdad, and Manhal Jabr, Director of the Mosul office, for encouraging the various stages of fieldwork that have contributed to this study.
22
23
T. J. WILKINSON
SETTLEMENT AND LAND USE
Barbanes, E. R. Heartland and Province: Urban and Rural Settlement in the Neo-Assyrian 1999 Empire. Ph.D. dissertation, University of California, Berkeley. Blackburn, M. Paltosols du 111~milltnaire et sols contemporains de tell 'Atij et de tell 1998 Gudeda dans la moyenne vallte du Khabour, en Syrie du Nord, Canadian Society for Mesopotamian Studies Qukbec, Bulletin 33 (= M. Fortin and 0 . Aurenche eds., Natural Space, Znhabted Space in Northern Syria [loth- 2nd millenium B.C.]), 69-82. Burghardt, A. F. The Location of River Towns in the Central Lowlands of the United States, 1959 Annals of the Association of American Geographers 49, 305-23. A Hypothesis about Gateway Cities, Annals of the Association of American 197 1 Geographers 61, 269-85. Butzer, K. Sociopolitical Discontinuity in the Near East C 2200 B.C.E.: Scenarios from 1997 palestine and Egypt, 245-96 in H. Niizhet Dalfes et al. eds., Third Millennium BC Climate Change and Old World Collapse. NATO AS1 Series I: Global Environmental Change, Vol. 49. Berlin: Springer Verlag. Cocks P. S. et al. Degradation and Rehabilitation of Agricultural Land in North Syria. 1988 ICARDA, Aleppo, Syria. Courty, M.-A. Le cadre paltogtographique des occupations humaines dans le bassin du 1994 Haut-Khabour (Syrie du Nord-est): Premiers rtsultats, Palkorient 2011: 2159. The Soil Record of an Exceptional Event at 4000 B.P. in the Middle East, 1998 93-108 in B. J. Peiser er al. eds., Natural Catastrophes During Bronze Age Civilizations. Archaeological, Geological, Astronomical, and Cultural Perspectives. Oxford: B.A.R. International Series 728. Curvers, H. H. The Balikh Drainage in the Bronze Age. Ph.D. dissertation, University of 1991 Amsterdam. D' Altroy, T. and T. Earle Staple Finance, Wealth Finance, and Storage in the Inka Political Economy, 1985 Current Anthropology 2612, 187-206. Danti, M. D. and R. L. Zettler The Evolution of the Tell es-Sweyhat (Syria) Settlement System in the Third 1998 Millennium B.C., Canadian Society for Mesopotamian Studies Qukbec, Bulletin 33 (= M. Fortin and 0 . Aurenche eds., Natural Space, Znhabted Space in Northern Syria [loth- 2nd millenium B.C.]), 209-28. Fales F. M. 1990a Grain Reserves, Daily Rations, and the Size of the Assyrian Army: A Quantitative Study, State Archives of Assyria Bulletin 411, 23-34. 1990b The Rural Landscape of the Neo-Assyrian Empire: A Survey, State Archives of Assyria Bulletin 412, 81-142.
Foster, B. R. 1986 Agriculture and Accountability in Ancient Mesopotamia, 109-28 in H. Weiss ed., The Origins of Cities in Dry-farming Mesopotamia in the Third Millennium B.C. Guilford, Conn.: Four Quarters Publishing Co. Gelb, I. J. Ebla and Lagash: Environmental Contrast, 157-67 in H. Weiss ed., The 1986 Origins of Cities in Dry-farming Mesopotamia in the Third Millennium B.C. Guilford, Conn.: Four Quarters Publishing Co. Goyunq, N. and W. Hiitteroth Land an der Grenze. Osmanische Venvaltung im heutigen tiirkisch-syrisch1997 irakischen Grenzgebiet im 16. Jahrhundert. Istanbul: Eren. Hillman, G. Agricultural Productivity and Past Population Potential at Agvan, Anatolian 1973 Studies 23, 225-40. Hole, F. Evidence for Mid-Holocene Environmental Change in the Western Khabur 1997 Drainage, Northeastern Syria, 39-66 in H. Nuzhet Dalfes et al., eds., Third Millennium BC Climate Change and Old World Collapse. NATO AS1 Series I: Global Environmental Change, Vol. 49. Berlin: Springer-Verlag. Hutteroth, W. Villages and tribes of the Gezira under early Ottoman administration (16th 1990 century). A preliminary report, Berytus 38, 179-84. Jones, M. J. Barley-based farming systems of the Mediterranean, 129-44 in M. Jones et 1993 al. eds., The Agrometeorology of Rainfed Barley-based Farming Systems. ICARDA, Aleppo, Syria. Kouchoukos N. Landscape and Social Change in Late Prehistoric Mesopotamia, Ph.D. 1998 dissertation, Yale University. Lemcke, G. and M. Sturm dl80 and Trace Element Measurements as Proxy for the Reconstruction of 1997 Climate Changes at Lake Van (Turkey): Preliminary Results, 653-78 in H. Nuzhet Dalfes et al eds., Third Millennium BC Climate Change and Old World Collapse. NATO AS1 Series I: Global Environmental Change, Vol. 49. Berlin: Springer-Verlag. 1,e Strange, G. Lands of the Eastern Caliphate. London: Frank Cass. 1905 Liere, W. J. van and J. Lauffray Nouvelle prospection archtologique dans la Haute Jezireh Syrienne, AAS 41954 5, 129-48. I'ewis, N. Nomads and Settlers in Syria and Jordan, 1800-1980. Cambridge University 1987 Press. Matthews, V. H. 1978 Pastoral Nomadism in the Mari Kingdom (ca. 1830-1760 BC). ASOR Series 3, Cambridge, Mass. Meijer, D. J. W. (in press) Ecology and Archaeology: Perceptions and Questions. Paper presented at the Rencontre Assyriologique Internationale (Venice 1997).
REFERENCES
24
T. J. WILKINSON
Miller, N. F. Farming and Herding Along the Euphrates: Environmental Constraint and 1997 Cultural Choice (Fourth to Second Millennia B.C.), 123-32 in R. L. Zettler ed., Subsistence and Settlement in a Marginal Environment: Tell es-Sweyhat, 1989-1995 Preliminary Report, MASCA Research Papers in Science and Archaeology 14, University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia. Miller, N. F. and K. L. Gleason Fertilizer in the Identification and Analysis of Cultivated Soil, 25-43 in N. F. 1994 Miller and K. L. Gleason eds., The Archaeology of Garden and Field. Philadelphia: University of Pennsylvania Press. Nordblom, T.L. Livestock-Crop Interactions: The Decision to Harvest or to Graze Mature 1983 Grain Crops, Discussion Paper No. 10, Farming Systems Program, ICARDA, Aleppo, Syria. Pabot, H. Rapport au gouvernement de Syrie sur l'ecologie vCgCtale et ses applications, 1956 Rome F A 0 (ETAP-Rapport No. 663). Portugali, J. and R. Gophna Crisis, Progress and Urbanization: The Transition from Early Bronze I to 1993 Early Bronze I1 in Palestine, Tel Aviv 2012, 164-86. Reade, J. E. Tell Taya, 72-78 in J. Curtis ed., F i . Years of Mesopotamian Discovery. 1982 British School of Archaeology in Iraq, London. Rowton, M. B. Urban Autonomy in a Nomadic Environment, Journal of Near Eastern 1973 Studies 32, 201-15. Silver, M. Economic Structures of the Ancient Near East. London: Croom Helm. 1985 Thalen D. C. P. Ecology and Utilization of Desert Shrub Rangelands in Iraq. Ph.D. 1979 dissertation, Rijksuniversiteit Groningen. UNESCOIFAO Soil Map of the World. Vols. I and VII. Paris. 1977 Van Lerberghe, K. The Beydar Tablets and the History of the Northern Jazirah, 119-22 in F. 1996 Ismail et al. eds., Administrative Documents from Tell Beydar (Seasons 19931995). Subartu 2, Turnhout: Brepols. Villard, P. Un conflit d'autoritb 2 propos des eaux du Balih, MAR1 5, 59-96. 1987 Wachholtz, R. Socio-Economics of Bedouin Farming Systems in Dry Areas of Northern 1996 Syria. Kiel: Wiss.-Verl. Vauk. Weber, J. A. Faunal Remains from Tell es-Sweyhat and Tell Hajji Ibrahim, 133-68 in R. 1997 L. Zettler ed., Subsistence and Settlement in a Marginal Environment: Tell es-Sweyhat, 1989-1995 Preliminary Report, MASCA Research Papers in Science and Archaeology 14, University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia.
SETTLEMENT AND LAND USE
Weiss, H. 1986
25
The Origins of Tell Leilan and the Conquest of Space in 3rd Millennium Mesopotamia, 71-108 in H. Weiss ed., The origins of cities in dry-farming Syria. Guilford, Conn.: Four Quarters Publishing Co. Weiss, H. et al. The Genesis and Collapse of Third Millennium North Mesopotamian 1993 Civilization, Science 261, 995-1004. White, P. F. et al. Nitrogen Cycling in Semi-Arid Mediterranean Zones: Removal and Return 1997 of Nitrogen to Pastures by Grazing Sheep, Australian Journal of Agricultural Research 48, 317-22. Wilkinson, T. J. The Definition of Ancient Manured Zones by Extensive Sherd Sampling Techniques, Journal of Field Archaeology 9, 323-333. Extensive Sherd Scatters and Land-Use Intensity: Some Recent Results, Journal of Field Archaeology 1611, 31-46. Town and Country in Southeastern Anatolia vol.1: Settlement and Land-use at Kurban Hoyuk and Other Sites in the Lower Karaba Basin. OIP 109. Chicago. Linear hollows in the Jazira, Upper Mesopotamia, Antiquity 67, 548-562. The Structure and Dynamics of Dry-Farming States in Upper Mesopotamia, Current Anthropology 3515, 483-520. Late-Assyrian Settlement Geography in Upper Mesopotamia, 139-59 in M. Liverani ed., Neo-Assyrian Geography. QGS 5. Rome. Environmental Fluctuations, Agricultural Production and Collapse: A View from Bronze Age Upper Mesopotamia, 67-106 in H. Niizhet Dalfes et al. eds., Third Millennium BC Climate Change and Old World Collapse. NATO AS1 Series I: Global Environmental Change, Vol. 49. Berlin: SpringerVerlag . Water and Human Settlement in the Balikh Valley, Syria: Investigations from 1992-1995, Journal of Field Archaeology 25, 63-87. (in press) Archaeological Survey of the Tell ~ e ~ Region, d & Syria, 1997: A Preliminary Report, in K. Van Lerberghe ed., Subartu 6. Turnhout: Brepols. Wilkinson, T. J. and Tucker, D. J. Settlement Development in the North Jazira, Iraq. A Study of the 1995 Archaeological Landscape. Iraq Archaeological Reports 3, Warminster, Wiltshire, England: Aris & Phillips. Zeder, M. A. 1998 Environment, Economy and Subsistence on the Threshold of Urban Emergence in Northern Mesopotamia, Canadian Society for Mesopotamian Studies Que'bec, Bulletin 33 (= M. Fortin and 0. Aurenche eds., Natural Space, Inhabted Space in Northern Syria [lofh- 2nd millenium B.C.]), 55-67. 1999 The Role of Pastoralism in Developing Specialized Urban Economies in the Ancient Near East. Paper presented at the Annual Meeting of the Society for American Archaeology, Chicago, March 1999. Zeist, W. van and J. A. H. Bakker-Heeres 1985 (1988) Archaeobotanical Studies in the Levant 4. Bronze Age Sites on the North Syrian Euphrates. Palaeohistoria 27, 247-3 16.
26
Zettler, R. L. Introduction, 1-10 in R. L. Zettler ed., Subsistence and Settlement in a 1997 Marginal Environment: Tell es-Sweyhat, 1989-1995 Preliminary Report, MASCA Research Papers in Science and Archaeology 14, University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia.
27
SETTLEMENT AND LAND USE
T. J. WILKINSON
Wheat-based systems
Barley-based systems
Summer crops Tree crops
-
Fallow
I
I
I .. :. . . . . . . . .................................. .I
I
.
I
A=--&.
:~:;..;.:..~:;<~.::;:,.:\l
Nomadic
I pastoralism
I
Fodder crops.
I
!
-_-
300
200
Rainfall (mm) Fig. 1. Crop types and agricultural zones near the margin of rainfed cultivation in the Jazira (re-drawn from Jones 1993).
Fig. 2. Zone of marginal cultivation (stippled) in nothem Syria as defined by Wachholtz (1996) with selected major Bronze Age sites close to or within it. 1= Elba, 2= Umm al-Mara, 3= Tell Hadidi, 4= Tell es-Sweyhat, 5= Hammam et-Turkman, 6= Chuera, 7= Mozan, 8= Chagar Bazar, 9= Brak, 10= Leilan, 11= Hawa (Iraq), 12=Khoshi (Iraq), 13= Rimah (Iraq).
T. J. WILKINSON
29
SETTLEMENTAND LAND USE
Network economy: Systems
Trade and Exchange
Fig. 3. Tringular diagram of three-sector economy. Any given town would be expected to occupy different locations within this space trough time as indicated by arrows.
Upper Khabur
<
Steppe pasture (4)
- Rlver~ne - land (5) -
E-flowmg wadls
Fig. 5a. Cultivated and pastoral zones inferred for area around later third millenium B.C. Tell es Sweyhat, Syria.
Tempsitesshp.sh~ & Wadis Perennial Rivers Large Wadis Small Wadis Agricultural Zones
R#s
Fig. 4. Bronze Age tells, inferred cultivated land (plotted at 3 km per tell), and pasture estimated for the western Khabur basin, Syria (GIs map compiled by Jason Ur and Carrie Hritz).
\
Lmear hollow
30
SETTLEMENT AND LAND USE
T. J. WILKINSON
Later 3rd Millennium
Modern villaga /
/'
Dirt road
...:=: ...<
> .;.:.. .;..:.:,. :.:;, .+.
A
Cultivation
\
I
117 Arikok
0
.
Euphrates flood plain
Period F
*, 11
G H-l
1
Fig. 5b. Kurban Hoyiik area, Turkey showing inferred area of steppe pasture. Period F= early third millennium, Period G= mid-late third millennium, and Period H/I = late thirdlearly second millennium B.C.
+
Center or large site Minor site
Fig. 6. Major Bronze Age sites in the Balikh Valley of Syria showing inferred areas of pasture to the east and west.
T. J. WILKINSON
SETTLEMENT AND LAND USE
-
y 1111,,,
-
I A site
0 tell
..'
Limit of h e a r hollows 200 persons per hax -
Pasture?
,--.
-- -
,/-, /
Most mtensive cultivation; fleld scatters > 60 sherds 1100 m2
' -.
100 persons per ha,
Moderately Intensive cultivation; field scatters 20-60 sherdsi100 m i Least intensive cultivation;
,-' marginal land with fallow; field
'\,
scatters < 20 sherdsl100 rnZ
Fig. 8. The area of Tell al-Hawa in the Iraqi North Jazira showing the "empty zone" in the western part of the survey area in relation to areas of inferred cultivation. This empty area probably developed as a pastoral enclave towards the end of the fourth millennium B.C.
5 km
basail plateau
-
Bronze Age tracks ('hear hollows')
watershed
Fig. 7a. Area of Tell Beydar survey showing surveyed sites in relation to inferred long-term pasture resource of basalt plateau (stipple). I.A. - Iron Age sites; tell were mainly occupied in the Bronze Age and late Chalcolithic (by Eleanor Barbanes).
Appendix A: 1 Appendix A: 3
Appendix A: 2
/
soil 3
Undulating steppe
Wadi 'Awaidj
Tmmmm_/rmn Fig. 7b. Schematic cross-section from east to west across Beydar area showing potential pasture areas to east on steppe and to west on basalt.
Fig. 9. East-west alignment of large sites across the Jazira. The hatched line represents either the potential east-west route or the zone of interaction between mobile groups to the south and more sedentary communities to the north. The hatching south-east of Hadhail represents a possible route leading south towards southern Mesopotamia.
SETTLEMENT AND LAND USE
T. J. WILKINSON
Urfa (1959-1998)
Annual Rainfall 1959-1980
Annual Rainfall 1959-1998
Aleppo (Rainfall i n rnm)
Aleppo (Rainfall i n mm)
Fig. 10. Rainfall correlation between (left) Aleppo and Kamishli, and (right) Aleppo and Mosul (by Colleen Coyle).
Hawa Aug 99 Annual & Biennial barley
Hydrological Years
Fig. 12. Autumn rainfall as a predictor of the total year's rainfall for Urfa (Turkey) 1960 to 1996197. Encircled points appear to be those years when it would have been possible for a farmer to predict a drought year.
I
I
Years
- - - -Linear (Biennial barley) -Linear (Annual barley) --- -r--Annual
barley
--t Biennial
barley
Fig. 11. Barley production for the Tell al-Hawa area simulated for an arbitrary one hundred year period. Note the gradual decline in the yield of annually cropped barley compared to the stability of the biennial barley yield (data for this preliminary simulation was provided by John Christiansen, DIS division, Argonne National Laboratory.
PIHANS LXXXVIII, 2000
Palynological and Archaeobotanical Evidence from Bronze Age Northern Mesopotamia S. Bottema and R. T. J. Cappers (Groningen) INTRODUCTION
Various sources of evidence inform us about the botanical history of Mesopotamia. Translations from texts on clay tablets, preserved seeds, fruits or wood found in former settlements or pollen in sediments, all give information on the past vegetation, each in their own way and with their own possibilities and restrictions. The relationship between the textual evidence and human activity is obvious. Of special interest are the differences in farming practices for various parts of Mesopotamia, as suggested by philologists. The main problem is the proper translation of the text and solid proof which plant or product is meant. The main advantage of palynological investigations is that they are based on the study of sediments not directly related to human settlements, which does not exclude that pollen analysis reflects the impact of man on its environment. In fact, both human activity as well as climatic changes will change the natural vegetation, which in turn is documented in the shifts of the individual pollen curves. Unfortunately, the identification level of the pollen data seldom is detailed enough to indicate former ecological conditions and local anthropogenic exploitation of plant species. Although Mesopotamia can be treated as one unit, exposed to a continental derivative of the Mediterranean climate, the area clearly demonstrates an ecological gradient that will control to a certain extent the possibilities of living for plants, animals and, not in the least, for human habitation. This diversity is caused by a variation in abiotic conditions, such as climate, salinity, moisture regime, nutrient availability and farming. Farming is divided into two main systems, one system in which rainfed agriculture is practised, and a system in which farming is only possible with irrigation. It goes without saying that in between there is a transitional zone where irrigation is a welcome addition to rainfed farming or where the unpredictability of the annual precipitation stresses the need to use supplementary water. In this paper only those archaeobotanical data from Mesopotamia will be discussed that are provided by large scale, systematic excavations and sampling. These data mostly originate from the extensive studies by Willem van Zeist and collaborators on charred macrofossils from archaeological excavations in the Near East. Unfortunately, the information for Mesopotamia is restricted to the northern part and no comparable palaeobotanical studies are available for the southern part. Van Zeist generally published those studies per site or per group of sites, paying attention to the diversity caused by the opposition rainfed versus irrigation farming (Van Zeist, 1999). In addition to the information obtained from studies which have been published, Van Zeist kindly supplied data from a series of studies on the Balikh and the Khabur area which are in press.
38
S. BOTTEMA AND R.T.J. CAPPERS
The aim of this study is twofold. It presents an up-to-date list of cultivated and wild plants identified to the level of genus or species from Bronze Age northern Mesopotamian sites. This record may be helpful in translating texts. Furthermore, an analysis is made of wild plant species on the basis of available pollen and macro remains. This analysis is focused primarily on the ecological interpretation of these remains. The data sets have been screened qualitatively for the period of roughly 3000-1500 B.C. and in part of the sites this time is subdivided into Early, Middle and Late Bronze Age. ENVIRONMENTAL SETTING
When we define Mesopotamia as the lands delimited by the course of the Euphrates and the Tigris, it is obvious that the area is not uniform. On an average, most of the area receives less than 100 mm of rainfall, but in the north a gradient towards precipitation of 500 mm can be observed. The archaeological sites from which macrofossil evidence is used in this paper are displayed in figure 1 together with the rainfall reliability to give an idea of their moisture conditions. Figure 1 is re-drawn from the TAVO rainfall reliability map (Alex, 1985) and gives the amount of annual precipitation that can be predicted in 80% of the time. This reliability map gives a better idea of the possibilities of agriculture than the map with the average annual precipitation, which can hide the occurrence of wide fluctuations. The boundary of dry farming can be set at the isohyet of the average annual precipitation of 300 mm, practically the predictable, guaranteed annual level of 200-250 mm (Oates and Oates, 1976). The average January temperature ranges from 10-15•‹Cin the south to 0-(-)5"C in the north. The averages for July run from over 35•‹Cnear the Gulf to about 27•‹Cin the north. In figure 2 the soil variety after Straub (1988) is given. Soil conditions differ in the area from the half-desert soils rich in lime in the north to gypsum containing desert soils in the middle. In the south, solonchaks, saline soils, prevail. Alluvial soils are predominantly formed by river sediments, which vary from chestnut brown soils to crusted limestone or gypsum cover, saline clay or stony soil. Fig. 3 is a simplified map of the vegetation re-drawn from Frey and Kiirschner (1989). The northern part of Mesopotamia is characterised by very open dwarf-shrub vegetation, whereas the southern part mostly carries annuals in a dense steppeldesert vegetation. Under natural conditions, the river valleys would carry riparian vegetation including poplar (Populus euphratica) and extensive reed marsh. THE INFORMATION LEVEL OF PALEOBOTANICAL REMAINS IN MESOPOTAMIA
Palynological evidence Palynological investigations provide the palaeo-precipitation of pollen produced by past vegetation. Pollen, representing plant taxa and vegetation, has an advantage over macrofossils in excavations by its distribution, which is anthropogenically unintentional. The distribution of pollen is fairly homogenous and numbers of pollen are very large. Some can be dispersed over long distances, but the majority of pollen precipitates within a few kilometres from the source. A disadvantage, compared to macro-remains, is the low identification level of many pollen taxa which is mostly restricted to the genus or to family
level. To obtain palynological information, sediment is sampled in which pollen has been preserved. Conditions for the preservation of pollen and spores are rather special and require low oxygen. A guarantee for low oxygen is permanent waterlogging of the sediment. Such a situation is very rare in the Near East where sediments are generally formed during the winter half-year in basins that mostly dry up during the summer. Consequently, the pollen that precipitated in a basin is oxidised and no information on the past vegetation can be obtained from such deposits. Sediment suitable for palynological investigation is almost completely absent from most of northern Mesopotamia, while the southern part has not been investigated for this purpose. Apart from the corrosive conditions, the deposit in a meandering river is also not the first choice for the reconstruction of past conditions because a river, which shifts in its bed, constantly clears out older sediments. Re-deposition repeatedly takes place, and in case pollen was preserved, the result would be a greatly disturbed archive. Sediment deposits suitable for the preservation of pollen preferably are collected in basins or lakes, which receive little or no discharge of rivers. Such situations are common in areas with sufficient annual rainfall but almost absent below the 300 mm annual precipitation isohyet, which means that almost the complete Jazira is too dry for a suitable pollen sediment. In the Syrian Jazira, cores were collected in several places, for instance in the meanders of the Balikh and the Khabur (Grernrnen and Bottema, 1991). The Balikh core was taken near Tell Hammam et Turkrnan and the cores in the Khabur valley near Tell Schech Hamad and near Tell Bderi. Invariably the results were negative because organic matter including pollen had disappeared from the deposits. A location that turned out to be moist enough for pollen preservation is the salt flat of Bouara on the border of Syria and Iraq, southeast of Tell Schech Hamad (Gremmen and Bottema, 1991). A negative characteristic of Bouara is the nature of the salt flat preventing the growth of vegetation. Thus, most of the pollen present in the Bouara deposit originates from further away and little more can be said about the area itself than that it was either barren or densely covered with halophytic chenopods and Tamarix. Furthermore, information on the vegetation history of the Zagros Mountains, east of Mesopotamia, based upon the cores from Zeribar and Mirabad (Van Zeist and Bottema, 1977) is instructive for the reconstruction of environmental conditions in Mesopotamia during the younger part of the Holocene. The information from the Iranian pollen diagrams may indicate a development trend that is valid for the Mesopotamian lowland as well, and will be used here as such. Archaeobotanical evidence Archaeobotanical evidence available for the interpretation of past environmental conditions consists of botanical macro-remains, mainly seeds, fruits and threshing remains originating from excavated settlements. In this respect, they differ basically from palynological data as these are obtained from deposits not fundamentally changed by human activity. Therefore, insight in taphonomical processes is a precondition for a reliable interpretation of these macrofossil remains. Taphonomic processes are the various processes undergone by such plant remains in which both natural and anthropogenic agents are active, explaining the ultimate conditions of the subfossil record as it is sampled. Important processes are connected with dispersal, harvesting, and preservation
40
41
S. BOTTEMA AND R.T.J. CAPPERS
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
(Hillman, 1984; Cappers, 1995). Dispersal and preservation conditions will be discussed in more detail. In the Near East plant macro-remains can be preserved in several ways: by charring, by desiccation, by mineralisation or as imprints. Imprints are found on pottery, and can be identified as cereal grains or threshing remains. Desiccated remains are rare as they only survive in permanent arid conditions. A small category of seeds found in excavations may be the result of mineralising. Seeds of Boraginaceae are often found mineralised for unknown reasons. Anatolian excavations show the presence of mineralised seeds of Celtis, the hackberry tree. Possibly the berries were cooked for mash. Cooked fruit stones, found in large quantities, are killed that way and do not germinate or rot, but rnineralise. Charring has preserved most botanical macro-remains found in the Near East. The background of charred remains in archaeological sites differs widely from the deposition of pollen. How and why seeds get charred in settlement sites is a matter of debate. One can easily understand that it is not the intention of farmers to burn harvests. The question remains what kind of processes lie behind the presence of numerous charred seeds, farm crops as well as herbs, weeds or other plants in prehistoric settlement accumulations. The presence of large concentrations concerns domesticated crops, sometimes intermixed with remains of wild plant species (weeds), and is among the easiest to understand. Very likely this presence is the result of a large fire that hit the building or even the entire settlement as could be evidenced for Tell Sabi Abyad where complete storerooms filled with charred grain have been unearthed. In case complete buildings and storage facilities for cereals burned down, the outer part of a grain storage might have burned completely, whereas the inside of the structure may not have been touched by the fire at all. This latter part will have disappeared subsequently by rotting away. Only the part sufficiently charred to prevent germinating or rotting was thus preserved and it is this part that is potentially discovered by an excavator. However, the majority of samples collected especially for palaeobotanical studies are not the irregular, scarce bulk amounts resulting from a catastrophe and easily discovered with the naked eye, but thoses retrieved by simple flotation methods applied to the bulk fill of the accumulation of the excavated settlement. The problem of dispersal is of particular interest when we are dealing with the remains of wild plants. How did they enter the site and why did they become charred? At first, these charred remains were simply accepted as a source of information. The finds were considered as a reflection of crop farming and the accompanying weeds. Plants not belonging to these categories were considered as part of the wild flora that gave additional information, although their origin was not questioned. Subsequently, however, investigators started to wonder how such seeds got charred and whether the mechanism behind this charring caused a representative picture of the vegetation around the site or not. As far as remains of domesticated crops are concerned, there are several possibilities. The origin of the scatter of charred seeds and small pieces of charcoal that is present as "noise" in accumulation debris is very convincingly explained by Miller and Smart (1984) as the result of the use of animal dung as fuel. Dung contains numerous seeds which are not digested as long as the seed coat is not damaged by chewing. A part of these undamaged seeds becomes charred during the use of dung cakes for fuel (Neef and Bottema, 1991). Seeds of grazed plants would thus dominate the charred seed
assemblages, the daily menu eaten by cattle and sheep. This may be even true for cultivated species, such as barley, if they are used as animal fodder. Especially beasts of burden, that needed to be in good condition, would have been additionally fed with such, more exclusive, food. Deadldry plants are also gathered as fuel for baking bread or for cooking and the plants may still contain seeds. A certain part, lying at the edge of the fire, escapes from total burning and is swept out of the bread oven during cleaning and thrown on refuse dumps. Such remains may concern cultivated species like uprooted sesame (Sesame indicum L.) stems that still are used this way in northern Syria, as well as wild species like Alhagi spp., Prosopis farcta and Artemisia herba-alba, which remain ungrazed because of their taste or spiny appearance. Bread ovens and fireplaces for burghul-cooking produce hundreds of litres of ashes from animal dung, which are either swept away or thrown on rubbish heaps, or which remain on the spot proper from where they are subsequently dispersed by wind. This dispersal by wind may be an important reason why the "noise" in an excavation can be rather uniform. Thus, seeds of various plant species with a very different background may arrive in settlements where they get charred and where they are preserved, either at the place of the fire or in a secondary place where refuse was dumped, or even dispersed from there by wind. To obtain a good picture of the presence of seeds in a site, an accurate sampling should be done. This also includes the registration of the parts of the site which yielded no seeds. It is needless to say that such sampling is very labour intensive and requires a lot of analysis afterwards as well. The sampling for botanical matter in the Near East, including Mesopotamia, ranges from a few imprints on pottery (Helbaek, 1966a, 1966b) to intensively sampled excavations. A continuous and intensive sampling of a Near Eastern site, to obtain as much botanical information as possible, was done for the first time in 1965 by Van Zeist and Bottema during the excavation of Pre-Pottery Neolithic Tell Ramad, south of Damascus, near Qatana (Van Zeist and Bakker-Heeres, 198211985). The site was sampled vertically (in time). Extensive sampling in horizontal and vertical sense has been done in Mesopotamia for a restricted number of sites only. The results prove that such sampling produces a large number of charred plant remains belonging to many species. The simple flotation of buckets of soil provides us with floating seeds and small pieces of charcoal which clearly originate from other events than large-scale catastrophic fire. It is obvious that the results of such sites are difficult to compare if the sampling method is not the same and that from this varied evidence no conclusive picture about Mesopotamian farming can be made. It should be stressed that the demand for a complete (or as complete as possible) sampling is the result of questions about the nature of farming and farming economy in the Near East which subsequently developed through time. At first, the investigators were quite content with the demonstration of wheat and barley, but soon it was made clear that such finds could be expected in a farming society that was already defined from other evidence. Especially the discovery of texts in excavations contributes to the interpretation of aspects of agriculture and trade as well.
42
S. BOTTEMA AND R.T.J. CAPPERS
RECONSTRUCTION OF VEGETATION AND CLIMATE
The palynological information To obtain information about the general vegetation during past times, palynological studies are the most suitable tool. Such information leads to conclusions that, although broad, offer the possibility to evaluate human exploitation, either by crop cultivation or animal husbandry. The application of palynology is altogether scarce in this area and restricted to one site for the period under discussion, that of Bouara (fig. 4). The pollen diagram of Bouara (Gremmen and Bottema, 1991) has been prepared from a sediment collected at the side of the spring that provides the Bouara salt flat with water. The sediment consists mainly of clay, sometimes mixed with sand or organics. The spring is located at the edge of a white crystalline salt deposit. At a depth of 210-230 cm and at 340-360 cm, organic clay and peat is found, indicating that locally marsh vegetation occurred. The lower organic deposit was dated 5730 +I20 BP. The whole section is about 6500 years old. The large salt flat is devoid of vegetation. At the edges, halophytic vegetations are found formed by species of the goose foot-family (Chenopodiaceae), and near the spring a few specimens of Tamarix grow. Surface samples, displaying the modern pollen rain, were collected at the edge of the salt flat and in the neighbouring sites of ar-Roda at some 15 km and Wadi 'Agig at about 10 km distance. These samples show that pollen percentages of the salt-indicating Chenopodiaceae fluctuate between 15 and 80 %. Even near the marsh such values range from 20 to 80%, depending upon the distance of chenopod vegetation to the sample. Several metres from a dense vegetation of Halocnemum, 80 % Chenopodiaceae pollen was present, whereas about 50 metres into the salt flat the share of this type dropped below 20%. Therefore it is questionable how important the behaviour of the dominant salt plants is, because changes in pollen percentages may have been caused by shifts in large scale environmental conditions but also by very local changes in the hydrology, for instance shifts in run-off of the well. The catchment basin the core was taken from received pollen from vegetation on moist parts in the vicinity, mainly members of the Chenopodiaceae. Further away, the Artemisialgrass and herb-steppe attributed to the pollen precipitation, and, as a third group, 21 tree pollen types came from outside the Jazira. For a general reconstruction of the environment in this part of the Jazira and the changes that took place, the three groups of pollen have to be separated. The first indication that informs us about the environment is the fact that around Bouara at about 6500 BP moist surroundings developed upon a hard, dried-up underground of clay. If we look at the Holocene vegetation development in the Zagros Mountains, a striking increase in deciduous-oak pollen took place from about 7000 BP. Maximum oak-pollen values had developed at about 5500 BP. The rise in deciduous-oak pollen must have been caused by an increase in average annual precipitation that enabled the spread of these trees over a zone from about 700 m up to over 2000 m. Part of this rainfall will have run down the slopes, elevating the groundwater level in the lowlands, creating wells here and there, which produced water. The spring of Bouara received its water supply from the Gabal Singar. The supply was not huge but persistent, creating wet surroundings that built up sediment, mostly in shallow open water. The evaporation of the water finally formed the large salt flat. Locally, a lush swamp existed that was at times less saline than on other occasions and in
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
43
which peat was formed. Both the Bouara diagram and the Zeribar diagrams (van Zeist and Bottema, 1977) indicate moisture conditions for the Jazira and the bordering Zagros Mountains which lasted rather unaltered for the past 6000 years. It is not clear to what extent fluctuations in the average precipitation occurred, but conditions never returned to the much drier environment of the first half of the Holocene. The effect of local pollen producers, mostly grasses, is concluded from the fluctuation of percentages of extraneous tree pollen which decrease in value. In case of peaty sediment there may have been an increase in local vegetation. There still is natural (non-peat) vegetation which produces a considerable amount of grass pollen, as was found in a surface sample of Wadi 'Agig. This border area showed important vegetation regeneration owing to military control and, consequently, the absence of herds and included several grass species (such as Hordeum cf. murinum) and numerous, up to one and a-half-metre high Artemisia scoparia. The latter only makes up 5% of the pollen precipitation, whereas the grasses number 28%. In other cases, Artemisia is overrepresented in the pollen rain (Iglesias, 1998: p. 116). Gremmen and Bottema (1991) concluded from the rising values for several plantain (Plantago) species that grazing pressure must have increased. This is also concluded from the appearance of pollen of the genus Haplophyllum, which members are reported to be "pungently and odorous herbs" (Townsend and Guest, 1980: p. 456). Pollen of inedible plants may increase in a relative way, because the edible plants are grazed away, but it can also increase in absolute numbers when pollen-producing plants occupy the empty space where other plants have been eaten. Finally, an increase in grazing pressure may be indicated by the occurrence of sand in the clay sediment as well. Reducing of the vegetation cover may have caused more erosion, either by aeolian transport or by water. It is concluded that during the past 6000 years no fundamental environmental changes took place in the Mesopotamian lowlands or in the Zagros Mountains apart from increasing human impact. Eventual changes in animal and crop husbandry in the area were related to cultural, economic and socio-political developments. Increasing exploitation pressure to complete over-exploitation resulted in the replacement of specific vegetation types. Correcting measures might have been taken to control the changes, but this was often impossible. The archaeobotanical information In this section it will be seen to what extent information can be obtained from the plant taxa of Bronze Age sites in northern Mesopotamia. These sites can be clustered into four groups. Two groups of settlements are located in the main valley of the Euphrates. Korucutepe and Tepecik belong to the Turkish part of the valley, whereas es-Sweyhat, Hadidi and Selenkahiye are located in the more southern Syrian part. The other sites are located in the valleys of two tributaries. Tell Harnmam and Tell Sabi Abyad are located in the upper part of the Balikh valley. The last group of sites is located in the Khabur valley: Tell Bderi and Tell al-Raqa'i in the middle part, and Tell Schech Hamad in the lower part. Especially those plant species which are known to be unsuitable for human consumption but which may be indirectly connected with arable farming, for instance in the form of weeds, will be tested for their ecological information. Their ecological validity may be of interest for the definition of the surrounding territory and may also be indicative of the way the seed arrived in the archaeological site. The distribution of the
44
S. BOTTEMA AND R.T.J. CAPPERS
various plant species over northern Mesopotamia will be followed to see whether differences appear. The distribution may be connected with the difference in ecological conditions such as precipitation, temperature or soil. Characteristics have been collected and briefly listed in appendix 1. On the basis of the properties described in appendix 1, plants are grouped in appendix 2 according to the indication of farming activity. The sites of Selenkahiye, Tell Hammam et Turlunan and Hadidi were intensively sampled and analysed (Van Zeist and Bakker-Heeres, 198511988). During the excavation of es-Sweyhat, samples for botanical examination were collected with the naked eye by the director of the excavation, Dr. T. A. Holland, (Van Zeist and Bakker-Heeres, 198511988). These samples provided sufficient information on cultivated plants, but wild plant species proved to be under-represented. Fortunately, new data from this site became recently available in a study by Miller (1997). With respect to the cultivated plants, records from both Van Zeist and Miller are presented, whereas for the wild plant species only Miller's results are used. Additional information on the taxa found in the sites of Tell Bderi, Tell Schech Hamad and Tell al-Raqa'i is given by Van Zeist (in press). The Turkish sites of Korucutepe and Tepecik were excavated in connection with the Keban Dam project. Obviously, the informative potential of the latter sites is lower compared to the Syrian sites mentioned first. The Mesopotamian agro-botanical results were presented in a paper on "Plant Cultivation in Ancient Mesopotamia: the Palynological and Archaeological Approach" during the Berlin meeting of the 41e RAI (1994) (Van Zeist and Bottema, 1999). In that paper the ratios of barley and wheat in 33 prehistoric sites are discussed in order to trace the effect of salinisation. Unfortunately, the way the southern Mesopotamian sites are sampled is altogether inadequate for the retrieval of suitable amounts of non-agricultural herbs, so they have no value for a proper ecological comparison attempted in this study. CHARACTERISTICS OF PLANTS FOUND IN NORTH MESOPOTAMIAN BRONZE AGE SITES
The finds from the Bronze Age sites from northern Mesopotamia number 208 taxa. They include a wide variety of cultivated plants that is summarised in table 1. The results from Van Zeist and Bakker-Heeres (1988) and Miller (1997) with respect to Sweyhat have been combined as the former study almost exclusively deals with pure crop samples, with only a few wild species present. It is shown that both barley (Hordeum vulgare) and wheat (Triticum monococcum, T. dicoccum andor T. aestivum/durum) are found in most sites. Millet (Panicum miliaceum) has only been evidenced in Schech Hamad. Pulses are well represented and include grass pea (Lathyrus sativus), lentil (Lens culinaris), pea (Pisum sativum), bitter vetch (Vicia ewilia) and chickpea (Cicer arietinum). The impressive number of fruit trees shows that in addition to field crops, gathering food from the wild or from cultivated trees nearby the settlements supplemented the diet with a wide variety of edible plant products. The following fruit-bearing trees and shrubs have been recorded: fig (Ficus), olive (Olea europaea), pine tree (Pinus sp.), pistachio (Pistacia), mahaleb cherry (Prunus mahaleb), blackthorn (Prunus spinosa), pomegranate (Punica granatum), hawthorn (Crataegus), brambles (Rubus), caper bush (Capparis spinosa), melon (Cucumis melo), and grape (Vitis). Oil and fibre plants are represented by gold-of-pleasure (Camelina sativa), safflower (Carthamus tinctorius), flax (Linum usitatissimum) and sesame (Sesamum indicum).
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
45
A last category concerns spices and condiments and is only represented by Hammam et-Turkman: coriander (Coriandrum sativum), fennel (Foenicuhm vulgare), hyssop (Hyssopus) and fenugreek (Trigonella foenum-graecum). About 85% of the taxa are considered as wild plant species, although it cannot be ruled out that they still include some plant species that might have been purposely collected or grown. Table 2 shows the distribution of the wild plant taxa over the investigated sites per period in relation to the level of identification. The number of taxa per site and period varies from 10 to 87 with an average number of 53. Families most frequently recorded are the pea family (Leguminosae: 13x) and the grass family (Gramineae: 1lx). The most frequently recorded genera are: bedstraw (Galium: 13x), mayweed (Medicago: 1lx), canary grass (Phalaris: 1lx), Eremopyrum (I lx), goat's-face grass (Aegilops: 1lx), catchfly (Silene: 1lx) and milk vetch (Astragalus: 1lx). On average, less than one third of the taxa could be identified to the level of species, including some taxa that consist of a well-defined cluster of species. These taxa are specified in table 3 and include possible progenitors of domesticated plants. From this overview it is obvious that some sites are well represented whereas others are represented by a few species only, or, in the case of Korucutepe, even lack identifications on a species level. The small number of species might be the result of the limited sample volumes that were available for research, as was the case for Bderi and Tepecik. Another reason might be that samples proved to be severely cleaned from weed species, such as the almost pure samples from Bronze Age Sabi Abyad. Samples from Selenkahiye, on the other hand, proved to be rich in identifiable plant remains offering a detailed basis for an ecological interpretation (Van Zeist and Bakker-Heeres, 1988). CLUSTERING OF INFORMATION
To explore the composition of the wild species of the sites, a correspondence analysis was carried out. With this kind of analysis it is possible to reduce the complexity of the data matrix. The analysis is based on presencelabsence, primarily for practical reasons as no quantitative data have been published for Bronze Age Sabi Abyad. A quantitative analysis can be more informative, although large differences in seed production of wild plant species may obscure the outcome. In such a case a data transformation should be applied. Because the record of Korucutepe yielded no identifications on a species level, this site is not included in the correspondence analysis. The results have been plotted in two separate figures. Fig. 7 shows the ordination diagram of the nine sites, in which the composition of Hadidi and Hammam et-Turkman are presented separately for the Early, Middle or Late Bronze Age. Each dot represents the mean species scores of all species from that particular site and period. Obviously, Tepecik and Hadidi (Late Bronze) have the most individual species composition. Tepecik takes up an isolated position, which can be attributed to the presence of Solanum nigrum. The same is true for Late Bronze Hadidi, which in turn is determined by Chenopodium murale and Malcolmia africana. The ordination diagram of the species is shown in fig. 8. For reasons of visibility, not all the species are shown in the cluster beneath the origin of the axes. Each species is abbreviated with eight letters, the first three referring to the genus and the next four to the specific epithets. If this pattern is compared with the ecological information of these species, it can be concluded that there is no clear clustering of ecologically related species.
46
S. BOTTEMA AND R.T.J. CAPPERS
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
Several explanations can be put forward for the absence of a more distinct clustering. First of all, it should be kept in mind that this analysis is based only on the presence or absence of wild plant species identified to the level of species. Although a quantitative analysis might produce a more distinct clustering, it should be realised that differences still are based on a limited number of species and, moreover, that differences in seed numbers of wild plant species are by definition not correlated with species abundance. Wild species differ basically in this respect from, for example, cereals. In the case of cereals proportions can be based on either seed counts or the number of rachis internodes, although in the latter case a correction with respect to the average number of flowers per internode has to be taken into account. Both the charred mode of preservation and the incompleteness of reference collections hamper the identification of wild species. A reliable identification to the level of species is only possible if all potential candidates can be considered. Extensive Near Eastern reference collections, such as available in the Groningen Institute of Archaeology, nevertheless still lack many rare species. This seriously hampers identification beyond genus level and, therefore, the need for even more reference material from the Near East is felt. Especially well preserved plant remains that are only identified to the level of genus or family will represent species that belong to the Near Eastern vegetation. Revealing their identity may give expression to the varied exploitation of Mesopotamia. Besides identification problems, the cleaning of cereal samples is partly responsible for the low number of weeds, as could be demonstrated for samples from both Tell Hammam and Tell Sabi Abyad (Van Zeist, in press). The weed composition of the samples from Selenkahiye in particular contrasts sharply with this low frequency. The location of all sites in the same drainage area may also obscure potential differences between the clusters of sites. Scirpus maritimus and Polygonum corrigioloides, for example, are plants with a riverine distribution in these areas. S. maritimus is a common species native to Eurasia and Africa, and is present in the subfossil records of Tepecik (Turkish part of the Euphrates), Selenkahiye and Hadidi (Syrian part of the Euphrates), Tell Harnrnam and Tell Sabi Abyad (Balikh Valley) and Tell al-Raqa'i (Khabur Valley). P. corrigioloides, on the other hand, is considered as an endemic species recorded from the environment of Meyadine, where the Balikh river meets the Euphrates, and the surroundings of Baghdad (Mouterde, 1966). Obviously, this plant species has been attested from both Tell Bderi and Tell al-Raqa3i in the middle part of the Khabur Valley. Another problem is the indicative value of species in an archaeobotanical context. As plant remains often become mixed up in an archaeological context, it is often difficult to correlate weed species with a particular crop species. As far as species indicative of irrigation are considered, it has to be realised that irrigation might have been limited to the valleys proper, as was and still is the case in for example the upper part of the Balikh Valley. Rainfall enables dry farming, but would only have been sufficient for the cultivation of more drought resistant crops such as barley. Evidence for irrigation in the Balikh Valley during the Bronze Age stems from written sources and is also inferred from a shift to naked wheat (Van Zeist, in press). Even plants that do not grow alongside streams or are not primarily adapted to water dispersal are subjected to riverine transport, which may result in a certain homogeneity. Flat riverbanks may provide suitable threshing areas, as was recently observed, for example, in Demirci (Central Anatolia). Huge quantities of cereals are brought from all
over the area to grassland along the river, which is flooded during winter times. In this way, many seeds of field weeds will be transported downstream. Finally, part of the wild plant species cannot be considered as arable weeds but originate from the grass steppe and are probably brought to the settlement via animal dung. Examples of such species are Alyssum alyssoides, Bromus sterilis and Trifolium pratense.
47
Appendix 1 DESCRIPTION
In this section a short description of the wild plant species is given which is presented in table 2. The ecology of each species is briefly described, with special reference to the habitat. Alkanna strigosalcf. hirsutissima: shrub-steppe and cornfields (Davis, 1965-1988, vol. 6:423). Alyssum alyssoides: grazed by domestic animals in open steppe. Ammi majus: perennial in fields, waste places and along ditches (Davis, 1965-1988, vol. 4:427). Androsace maxima: (Davis, 1965-1988, vol. 6:111) cultivated or fallow fields: gravel sands, clay steppe, annual. Amebia decumbens: on stony, volcanic hillocks (Davis, 1965-1988, vol. 4:313). Amebia linearifolia: fields, waste places (Davis, 1965-1988, vol. 4:3 13). Bromus danthoniae: dry grassy mountain slopes, hillsides and fields in the steppe, silty depressions and sandy or gravelly places in wadis, weed in irrigated fields in the subdesert (Townsend and Guest, 1966-1985, vol. 9:136-141). Bromus sterilis: annual (biennial) grass which is considered good grazing, in oak forest on limestone, by stream in shade of Juglans, wet waste land, in irrigated orchard. Buglossoides awensis (Lithospermum awense): ruderal, segetal weed, not related to ground water (Davis, 1965-1988, vol. 6). Buglossoides tenuifZora (Lithospermum tenuiflorum): limestone slopes, stony places (Davis, 1965-1988, vol. 4:316). Bupleurum subovatum = lancifolium: open dry habitat, secondarily as segetal weed. (Davis, 1965-1988, vol. 4:399).
48
49
S. BOTTEMA AND R.T.J. CAPPERS
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
Carex divisa: perennial, along irrigation canals and streams, under trees, red marley banks in valleys, grazed by animals (Townsend and Guest, 1966-1985, vol. 8:396), hayed before flowering.
Peganum harmala: along desert tracks, around encampments near wells, on rubbish heaps in villages and disturbed soil near ancient ruins and graveyards, indicator of human settlements, also along wadis in the desert, weed in fields in the steppe (Townsend and Guest, 1966-1985, vol. 4/1:303). Unpalatable herb and nitratophyt (Evenari et al., 1986). A weed or ruderal, nearly always on disturbed ground (Migahid, 1978).
Cephalaria syriaca: fields, waste places (Davis, 1965-1988, vol. 4590). Chenopodium murale: eaten by man and animals, not related to ground water. Roadside, waste places, refuse heaps and irrigated fields (Post and Dinsmore, 193211933). Weed of garden, occasionally seen in desert, especially around old camp sites (Migahid, 1978; Edgecombe, 1970).
Plantago lagopus/ovatdpsyllium: P. lagopus: steppic vegetation; P. ovata: sandy places and deserts; P. psyllium: steppic vegetation, also in fields.
Chenopodium rubrum: often not related to groundwater, in sandy areas connected with groundwater; waste places, cultivated ground.
Polygonum corrigioloides: endemic of the banks of the Euphrates, "moist meadows" along the Balikh (see also P. venantianum, Mouterde, 1996, vol. 1:400).
Euphorbia falcata: large range, also in cultivated and abandoned fields.
Polygonum lapath$olium: disturbed habitats, such as fields, waste places and roadsides.
Glaucium aleppicum/corniculatum: Townsend and Guest, 1980, vol. 412:785) desert plains and dry steppic hills, rather sandy places. G. comiculatum: indicative of silty soils (Evenari et al., 1986). Batha (Mediterranean dwarf-shrub formation) and steppe, rocky ground (Post and Dinsmore, 193211933).
Ranunculus sardous: waste places and grasslands. Indicative of moist soils, related to groundwater.
Hordeum vulgare ssp. distichum: the natural habitat is the steppe, but the plant also occurs as an admixture in barley fields. Lolium perennehigidum: (Townsend and Guest, 1966-1985, vol. 9:93) L. perenne only above 900 m in forest zone. L. rigidum near irrigation canalslseepage, open places in forest/scrub, silty depressions. Not related to groundwater level. L. rigidum a weed of agricultural land, particularly around grain fields; also occasionally found at roadsides or on other disturbed ground (Migahid, 1978).
Polygonum aviculare: disturbed habitats, such as fields, waste places and roadsides.
Reseda alba: steppic plains and sub-desert plains, usually on gravelly or slightly sandy soils (Townsend and Guest, 1980-1985, vol. 9:200-202). Disturbed habitats, such as fields, waste places and roadsides. Rumex pulcher: dunes, roadsides, ditches (Davis, 1965-1988, vol. 2:291). Roadsides and damp places (Post and Dinsmore, 193211933). Salsola laricina: Salsola species are indicative of dry and/or saline habitats. This type may include Noaea species (Van Zeist, in press).
Lolium temulentum: obnoxious weed, young plants edible but generally poisonous by fungus and fatal when consumed. On hillsides and in valleys, in the mountains, a segetal weed in dry foothills and plains. Not related to groundwater.
Scirpus maritimus: near streams, along ditches, edge of swamps, shallow pools from irrigation canal overflow, sometimes in saline places, also weed in rice fields (Townsend and Guest, 1985, vol. 8), indicator of fresh and salt water. Grazed. Requires groundwater table on or over surface level for part of the year. Swampy ground around fresh water in oases (Townsend and Guest, 1966-1985, vol. 8).
Malcolmia (Strigosella) africana: waste places on limestone slopes, dry steppic hills, sandy gravel in wadis, irrigated alluvium along ditches and channels, sandy river banks, weed in fields, among ruins (Townsend and Guest, 1966-1985, vol. 4/2:1032).
Scirpus setaceus: moist surroundings not containing calcium (Mouterde, 1966).
Medicago radiata: hillsides and steppic grasslands, fields on lower mountain slope, valley pastures, silty depressions, clay and gravel (Townsend and Guest, 1966-1985, vol. 3). Nepeta cataria: herb used for flavouring. Not related to groundwater. Neslia paniculata: disturbed habitats, such as fields, waste places and roadsides.
Setaria verticillatdviridis: S. verticillata: waste places, including fields. Moist ground, also along irrigation channels. S. viridis: waste places including fields. Moist ground. Sherardia awensis: meadows, grassy slopes, fields and disturbed soil, sometimes among scrub oak, occasionally on steppic plains, spread with agriculture (Townsend and Guest, 1966-1985, vol. 412). Not connected to groundwater. Solanum nigrum: fields and waste places (Edgecombe, 1970). Weed of farms and gardens (Migahid, 1978). Not connected to groundwater.
50
S. BOTTEMA AND R.T.J. CAPPERS
Trachynia distachya: steppe, fields and waste places all over the area. Trifoliumpratense: damp sub-alpine meadows, by mountain streams (at lower altitudes), weed in hill fields and orchards, grazed (Townsend and Guest, 1966-1985, vol. 3). Not connected to groundwater. Trigonella astroites: Artemisia herba-alba steppe, bare Iimestone slopes below tree line, barren degraded steppe (Stipetum capense), sand dunes, in sub-desert (Townsend and Guest, 1966-1985, vol. 3:102). Steppe (Post and Dinsmore, 193211933). Also on cultivated land. Triticum boeoticum: grassland on lower limestone mountain slopes, degraded oak forest and oak shrub (Townsend and Guest, 1966-1985, vol. 9:200-202). Vaccaria pyramidata: indicative of silt soils (Evenari et al., 1986). Weed of croplands. Not connected to groundwater. Valerianella coronata: rocky slopes, open woodlands, fields and roadsides (Davis, 19651988, vol. 4576). Valerianella vesicaria: rocky slopes, fields (Davis, 1965-1988,vol. 4578). Veronica persica: meadows, not connected to groundwater.
Appendix 2 GROUPING OF PLANT SPECIES ACCORDING TO THE INDICATION OF FARMING ACTIVITY
If possible, the plant species were attributed to categories including grazing, weeds in arable land, waste places and irrigation. Certain plant species may be found in more than one group. Plant species which are connected with grazing, fodder: Alyssum alyssoides - Bromus sterilis - Carex divisia - Chenopodium murale - Chenopodium rubrum - Scirpus maritimus - Trifolium pratense - Veronica persica? Plant species attested as weeds in meadows, segetal fields: Alyssum alyssoides - Ammi majus - Androsace maxima - Arnebia linearifolia - Buglossoides awensis - Bupleurum subovatum - Cephalaria syriaca - Chenopodium murale - Lolium rigidum - Lolium temulentum - Malcolmia africana - Medicago radiata - Neslia paniculata - Sherardia awensis - Solanum nigrum - Vaccaria pyramidata - Veronica persica. Plant species connected with waste lands, ruderal places: Buglossoides arvensis Cephalaria syriaca - Chenopodium murale - Chenopodium rubrum - Solanum nigrum Ammi majus - Peganum harmala - Rumex pulcher.
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
51
Plant species connected with irrigation: Bromus sterilis - Carex divisa - Chenopodium murale - Cynodon (dactylon)-Digitaria (sanguinalis) - Malcolmia africana. The ecological grouping has been based upon general information from the literature. One has to keep in mind that the various plant species all are of Near Eastern origin, where they grow in a variation of treeless vegetation. In some way or other human action, farming and other activity did extend or even optimalise the biotope demanded by this selection of the Mesopotamian flora.
Table 1. Presence of cultivated plant species. Abbreviations: SE = Selenkahiye, HA = Hadidi, SW = es-Sweyhat, HT = Harnrnam et-Turkman, SA = Sabi Abyad, RA = al-Raqa'i, SH = Schech Hamad, BE = Bderi, TE = Tepecik, KO = Korucutepe, E = Early Bronze Age, M =Middle Bronze Age, L = Late Bronze Age. [ ] = identification uncertain on genus level; ( ) = identification uncertain on species level. SE E Camelina sativa Capparis spinosa Carthamus tinctorius
HA HA M
L
SW E
HT E
(01
HT M
.
(01
HT L
SA L
RA E
SH L
BE E
TE E
KO E Gold-of-pleasure Caper bush Safflower
Table 2. Number of wild plant species identified to the level of family, genus or species. The category 'others' refers to a combination of genera or to taxa with an uncertain identification to the level of species. In the right-hand column the absolute number for each taxonomic level is given. Abbreviations: SE = Selenkahiye, HA = Hadidi, SW = es-Sweyhat, HT = Harnrnam et-Turkman, SA = Sabi Abyad, RA = al-Raqa'i, SH = Schech Hamad, BE = Bderi, TE = Tepecik, KO = Korucutepe, E = Early Bronze Age, M = Middle Bronze Age, L = Late Bronze Age.
Total number of taxa
I 83
149
134
( 57 187 176 168 126 174 151 141 134
1 10 1 176
Table 3. Presence of wild plant species identified to the level of species or to group of related species. Abbreviations: SE = Selenkahiye, HA = Hadidi, SW = es-Sweyhat, HT = Harnmam et-Turkman, SA = Sabi Abyad, RA = al-Raqa'i, SH = Schech Hamad, BE = Bderi, TE = Tepecik, KO = Korucutepe, E = Early Bronze Age, M = Middle Bronze Age, L = Late Bronze Age.
Peganum harmala Plantago lagopus/ovata~psyllium Polygonum aviculare(-type) Polygonum corrigioloides
. . .
.
.
Table 4: Presence of taxa identified on the level of genus. Abbreviations: SE = Selenkahiye, HA = Hadidi, SW = es-Sweyhat, HT = Hammam et-Turkman, SA = Sabi Abyad, RA = al-Raqa'i, SH = Schech Hamad, BE = Bderi, TE = Tepecik, KO = Korucutepe, E = Early Bronze Age, M = Middle Bronze Age, L = Late Bronze Age.
S. BOTTEMA AND R.T.J. CAPPERS
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
REFERENCES
Alex, M. 198 5
Vorderer Orient. Niederschlagsverla&keit 1:I2000 000, Karte A IV 5. Tubinger Atlas des Vorderen Orients. Ludwig Reichert Verlag, Wiesbaden. Cappers, R. T. J. A Palaeoecological Model of the Interpretation of Wild Plant Species, 1995 Vegetation History and Archaeobotany 4, 249-257. Davis, P. H. (ed.) 1965-1988 Flora of Turkey and the East Aegean Islands, Vol. 1-10. Edinburgh University Press, Edinburgh. Edgecombe, W. S. Weeds of Lebanon. Third edition, American University of Beirut, Lebanon. 1970 Evenari, M., I. Noy-Meir, and D. W. Goodall (eds.) Hot Deserts and Arid Shrubland (= Ecosystems of the World 12B). Elsevier, 1986 Amsterdam. Frey, W. and H. Kiirschner Vorderer Orient. Vegetation 1:8 000 000, Karte A VI 1 Tubinger Atlas des 1989 Vorderen Orients. Ludwig Reichert Verlag, Wiesbaden. Gremmen, W. H. E. and S. Bottema Palynological Investigation in the Syrian eazira, 105-117 in H. Kiihne ed., 199 1 Die rezente Umwelt von Tall &h Hamad und Daten zur Umweltrekonstruktion der assyrischen Stadt Dur-katlimmu. Dietrich Reimer Verlag, Berlin. Helbaek, H. 1966a Commentary on the Phylogenesis of Triticum and Hordeum, Economic Botany 20, 350-360. 1966b Pre-pottery Neolithic Farming at Beidha. A Preliminary Report, Palestine Exploration Quaterly 98, 61-66. Hillman, G. C. Interpretation of Archaeological Plant Remains: The Application of 1984 Ethnographic Models from Turkey, 1-41 in W. van Zeist and W. A. Casparie eds., Plants and Ancient Man. Studies in Palaeoethnobotany. Balkema, Rotterdam. Iglesias, M. 1998 Relation ve'ge'tation - pluie pollinique actuelle phytomasse epige'e pe'renne duns les steppes du Sud-Est de 1'Espagne et du Nord-Est du Maroc. Thkse, Toulouse. Migahid, A. M. Flora of Saudi Arabia. Vol. 1-2. Riyadh University Publication. 1978 Milller, N. F. 1997 Sweyhat and Hajji Ibrahim: Some Archaeobotanical Samples from the 1991 and 1993 Seasons, 95-122 in R. L. Zettler, Subsistence and Settlement in a Marginal Environment: Tell es-Sweyhat, 1989-1995, Preliminary Report. MASCA Research Papers in Science and Archaeology 14. Miller, N. F. and T. L. Smart 1984 Intentional Burning of Dung as Fuel: A Mechanism for the Incorporation of Charred Seeds into the Archeological Record, Journal of Ethnobiology 4, 15-28.
62
S. BOTTEMA AND R.T.J. CAPPERS
Mouterde, P. NouvellefZore du Liban et de la Syrie. ~ditionsde l'imprimerie catholique, 1966 Beyrouth. Neef, R. and S. Bottema Mest als Bron voor Verkoold Plantaardig Materiaal uit Opgravingen in het 199 1 Nabije Oosten. Waarnemingen en Experimenten, Paleo-Aktueel 2, 72-76. Oates, D. and J. Oates Early Irrigation Agriculture in Mesopotamia, 109-135 in G. de Sieveking et 1976 al. eds., Problems in Economic and Social Archaeology. Duckworth, London. Post, G. E. and J. E. Dinsmore (eds.) 193211933 Flora of Syria, Palestine and Sinai. Vol. 1-2. American Press, Beirut. Straub, R. Vorderer Orient. Boden 1:8 000 000. Karte A I1 6. Tiibinger Atlas des 1988 Vorderen Orients. Ludwig Reichert Verlag, Wiesbaden. Townsend, C. C. and E. Guest 1966-1985 Flora of Iraq. Ministry of Agriculture and Agrarian Reform, Republic of Iraq, Baghdad. Zeist, W. van Evidence of Agricultural Change in the Balikh Basin, Northern Syria, 3501999 373 in C. Gosden and J. Hather (eds.), The Prehistory of Food. Appetites for Change. Routledge, London. Comments on Plant Cultivation at Two Sites on the Khabur, Northeastern in press Syria. The Plant Husbandry of Tell al-Raqa'i. in press Zeist, W. van and J. A. H. Bakker-Heeres Prehistoric and Early Historic Plant Husbandry in the Altinova Plain, 1975 Southeastern Turkey, 224-257 in M. N. van Loon ed., Korucutepe. Final Report on the Excavation of the Universities of Chicago, California (L.A.) and Amsterdam in the Keban Reservoir, Eastern Anatolia, 1968-1970, vol. 1. Amsterdam/Oxford. 1982185 Archaeobotanical Studies in the Levant 1. Neolithic Sites in the Damascus Basin: Aswad, Ghoraifk, Ramad, Palaeohistoria 24, 165-256. 1985188 Archaeobotanical Studies in the Levant 4. Bronze Age Sites on the North Syrian Euphrates, Palaeohistoria 27, 247-3 16. Zeist, W. van and S. Bottema Palynological Investigations in Western Iran, Palaeohistoria 19, 19-95. 1977 Plant Cultivation in Ancient Mesopotamia: the Palynological and 1999 Archaeological Approach, 25-41 in H. Klengel and J. Renger eds., Landwirtschaft im alten Orient; ausgewahlte Vortrage der XLI. Rencontre Assyriologique Internationale, Berlin, 4.-8.7.1994 (= BBVO 18), Dietrich Reimer Verlag, Berlin.
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
Fig. 1. 80% Rainfall reliability for Mesopotamia (Alex, 1985). Archaeological site numbers: 1. Tepecik. 2. Korucutepe. 3. Sabi Abyad. 4. Tell Hamman et-Turkman. 5. es-Sweyhat. 6. Hadidi. 7. Selenkahiye. 8. Bderi. 9. al-Raqa'i. 10. Schech Hamad. 11. Bouara.
S. BOTTEMA AND R.T.J. CAPPERS
Fig. 2. Simplified soil map of Mesopotamia after Straub (1988). 1. Solonchaks, saline soils. 2. Calcaric fluvisols, limerich riversediment. 3. Luvisols, clay accumulation. 4. Water. 5. Xerosols, halfdesert soil. 6. Yennosol, desert soil.
PALYNOLOGICALAND ARCHAEOBOTANICAL EVIDENCE
Fig. 3. Simplified vegetation map of Mesopotamia after Frey and Kiirschner (1989). 1. Annual vegetation. 2. Riparian vegetation . 3. Montane woodland. 4. Water. 5. Dwarf-shrubland.
I
-
1 w u r m m d
~
2
~
-
-
~
-
-
-
-
-
N
-
-
~
1 -
~
~
~
~
~
~
~
Z
B
Y
K
B
M SAMflPNUMBERS Y Z B I
Y
~
AICARPINUS ORIENTALIS
NTAGO MAG0 NTAGO NTAGO NTAGO
MAJOR-TYPE MEDIA-TYPE TENUIFLORA-TYR CYLINDRIC&-TYPE MARITIMA-TYPE
NTAGO CORONOWS-TYPE
Z
~
%
K
Y
D
PALYNOLOGICAL AND ARCHAEOBOTANICAL EVIDENCE
S. BO'ITEMA AND R.T.J. CAPPERS
I */
Species
TEPECIK
Fig. 7. Correspondence ordination diagram of the nine sites, which contained botanical macro-remains identified to the level of species.
Fig. 6. Main part of the pollen diagram Ib from Zeribar, westem Iran (van Zeist and Bottema, 1977).
S. BOTTEMA AND R.T.J. CAPPERS
PIHANS LXXXVIII, 2000
The Role of Animals in a Neolithic Agricultural System in the Ancient Jazira Chiara Cavallo (Amsterdam) SETUIUE
SOLNIGR
CEPSVRI
.I
POLLAPR TRIBOEO. SCINARI
BUPSUBO POLCORR
TRIASTRO S~LLIIRI
ARNDECU
OURCPYRI GLmLEP
'muGTmu
UERPERS UEDRRDI
CHERUBR OTRIPRRT ALYIlLVS
SCISETR BRODENT.PLAPSK RESALBA NESPANI
Fig. 8. Correspondence ordination diagram of the species that are indicative of irrigation, grazing and disturbance.
The study of faunal remains is a useful means for reconstructing the patterns of animal exploitation by ancient communities. In recent years Near Eastern archaeology to an increasing degree has shown interest in the animal bones found during archaeological excavations, resulting in the publication of more and more zooarchaeological studies.1 This paper aims to approach the topic of reconstructing ancient agricultural systems from the viewpoint of a zooarchaeologist, emphasizing therefore the information this specialization has to offer in comparison to other specialists, such as archaeologists, philologists, botanists, and geologists, whose contributions were presented during the meeting.2 The case study of Tell Sabi Abyad's Neolithic occupational levels will be used for a reconstruction of the relationship between man and environment, highlighting the potentialities of zooarchaeological research. First, the major research results will be presented, after which, in line with the meeting's main theme, a model of the use of both river valley and steppe by an early agricultural community during the late Neolithic will be suggested. "Agriculture" is intended here as a term including the manipulation and exploitation of plants, that is cultivation, as well as of animals, that is husbandry. ZOOARCHAEOLOGICAL INVESTIGATIONS
Animal bones represent one of the largest categories of remains found during archaeological excavations. They provide information on the ways of exploitation and consumption of animals by humans. Usually, the deposition of animal bones in archaeological contexts is not accidental. Bones undergo a series of processes of complex origin before they reach the stage in which they are found. The branch of zooarchaeology studying these processes is called taphonomy, a term indicating the passage of organic material from the biosphere to the lithosphere (Efremov 1940:85). The transformational processes involved can be divided into two major groups, those of a diagenetic, or chemical, kind, and those of a anthropogenic nature. Undergoing these processes, the original living animal assemblage is affected by a series of losses until it reaches the stage of the death animal assemblage as it is found during excavations. Zooarchaeologists attempt to reconstruct the original assemblage. See the various volumes on the Proceedings of the international symposia on the archaeozoology of the Near east and adjacent areas (ASWA) (Buitenhuis and Clason 1993; Buitenhuis and Uerpmann 1995; Buitenhuis et al. 1998). I am grateful to Dr. Remko Jas for asking me to participate in the meeting and for his stimulating remarks.
72
THE ROLE OF ANIMALS
C. CAVALLO
Comparatively strong diagenetic processes are typical of Near Eastern faunal samples. Environmental conditions in arid zones impoverish the organic content of the bone, making it rather brittle and fragile, which results, as a consequence, in a rather high fragmentation of the material. Usually it is impossible to identify about 70 percent of the Near Eastern bone assemblages because it consists to a large extent of fragments or splinters. The Neolithic assemblage of Sabi Abyad has shown that the identifiability of the material correlates with the depth of the level in which the bones were found. For the upper levels a very low identifiability of circa 15 percent was ascertained, which steadily increased towards the lowest levels, where about 30 percent of the faunal assemblage could be identified. This indicates that assemblages from deeper levels are better preserved and protected, most likely as a consequence of moister soil conditions and less affecting erosion. Due to a considerable aeolic and alluvial soil deposition, the position of the lowest levels is collocated at circa 4 112 meters under the modem field level, the depth at which natural soil was reached (Wilkinson 1996). When bones are englobed into the soil, they have already been, subjected to other processes of loss and fragmentation caused by biotic destructive agents. After being killed, animals are butchered. Traces of butchering, such as cut- and chop marks are left on the bone surface. Cut marks are the traces most frequently encountered in the Near Eastern assemblages. They are usually localized on the metapodial as a result of skinning, and on the articulation joints where they are caused by dismembering (fig. 1). On articulation joints the cut marks usually are repeated since these joints are covered by numerous ligaments and tendons connecting different bones. Chop marks are rare and related to thick bones, like horn-cores. Traces of chop marks, for example, were found at the base of caprine and gazelle horn-cores (van Wijngaarden - Bakker 1989: 308, fig. x.2.; Cavallo 1996: fig. 9.1 1). In some cases it is possible to observe the result of a heavy blow inflicted mainly on the shaft of long bones for marrow extraction, or, in one case, on the lower part of lower jaw of a pig (fig. 2). These blows result in a semi-rounded flaked breakage (fig. 2, detail). Refuse disposal of bones is the result of the last types of activity to which the animal has been subjected, namely butchering, food preparation or consumption, and, possibly, a ritual use. The rarity of primary (in situ) contexts, however, often hampers the reconstruction of these activities. In most cases the material comes from secondary deposits consisting of refuse that has been discarded outside the area where it has been generated, or from tertiary deposits, in which original processes are mixed up together (Shiffer 1972). Animals themselves may contribute to the destruction of bones before they are incorporated into the soil. The animals most active in this sense are carnivores, especially dogs. They leave signs on the bone surface in the form of irregular pitting and furrowing that are mainly localized on the extremities where soft and more spongy bone is present (fig. 3). The percentage of bones chewed by dogs is often in contrast with that of the actual number of dog bones found on a site. It is even possible that gnawed bones represent the only evidence for the presence of this animal. In the case of Neolithic Sabi Abyad dog remains amounted to only 0.3 percent of the identified bones, while the percentage of gnawing marks observed on the bones of other species varies from to 2 to 23 percent. This could result from a dispositional pattern, in which complete carcasses are left outside the village or inhabited areas. Dogs themselves, however, contrary to modem perception, were the object of exploitation as dead animals as well. Traces of cut marks
73
were found on the shaft of a dog's radius at Neolithic Sabi Abyad (Cavallo 1996: fig. 9.10). Other taphonomic evidence traceable on the bones is evidence due to burning. Burnt bones, however, are not always connected with a human activity such as cooking, or better, roasting. This is the case with partially burnt bones, the flesh of which protected them to some extent from the fire. Calcinated or completely dark brown, small fragments could derive from material accidentally included in fires and in the soil. Homogeneous red-colored bones are common in the Near East as a result of fires taking place in villages that affected the bones already partly or completely incorporated in the soil. Finally, the utilization of bones for instruments or tools cannot be excluded. The taphonornic processes described above were recognized in the Neolithic sample of Sabi Abyad as 2.2 percent of butchering marks, 5.5 percent of burning traces and 4.3 percent of gnawing marks on the most common species (ovicaprids, bovids, suids, gazelle and equids). In addition, worked bones were found. They were made especially from the long bones of small herbivores. Awls are the most abundant tools within the Neolithic assemblage and are usually shaped from metapodias of sheep, goat, or gazelle (fig. 4). PRESENTATION OF THE SITE
The site of Sabi Abyad is situated in the Balikh valley, in northern Syria, at circa 30 kilometers from the modem Syro-Turkish border, in the vicinity of the modern village of Hamrnam et-Turkman. Tell Sabi Abyad, which means 'Mound of the White Boy,' is a mound of circa five hectares large at its base and five to ten metres high, consisting of a group of four smaller mounds that have merged into one. The excavation seasons concentrating on the Neolithic levels were conducted under the direction of Peter M. M. G. Akkermans in the years 1986-88 by the University of Amsterdam (UvA) and in the years 1991-93 by the Netherlands National Museum of Antiquities (RMO) (Akkermans 1996). During these excavations a series of several superimposed levels belonging to a late phase of the Neolithic dated to the second half of the sixth millennium and containing architectural remains of rectangular houses, tholoi, and courtyards from subsequent periods was unearthed. To be exact, 11 levels of occupation have been discovered, which were grouped into three major phases (pre-Halaf, Transitional and Early Halaf) (Verhoeven and Kranendonk 1996). The site, therefore, offered an excellent opportunity for a diachronic study on human-animal relationship within one community and one area. The earliest levels (11-7) of Sabi Abyad belong to an early stage of the Pottery Neolithic and antedate the appearance of the fine, painted ceramics typical of the Halaf culture. They are called therefore pre-Halaf (P) levels and date from 6000-5700 to about 5200 B.C., not calibrated. A transitional stage between the pre-Halaf levels and the earliest Halaf levels is represented by three levels (4, 5 and 6), called Transitional (T). They are dated between 5200-5150 and 5 100 B.C. The latest levels (1,2 and 3) represent the full appearance of Halaf-culture, characterized by a majority of the pottery consisting of carefully manufactured and painted ceramics. They are called Early Halaf (E), because they represent an earlier stadium in comparison with the Halaf periods known from other areas, like Iraq. These levels are dated to 5100-5000 B.C.
74
C. CAVALLO
PRESENTATION OF THE RESULTS
The faunal assemblage, amounting to circa 7,000 identified fragments, is largely dominated by domestic animals (Cavallo 1997a, b). They amount to 94.8 percent (P), 92.5 percent (T) and 96.6 percent (E) of the identified sample in each subsequent phase. The most abundant categories of animals are the ovicaprids (sheep, Ovis aries, and goat, Capra hircus) which cover a percentage of 73.7 percent (P), 68.5 percent (T), 70.6 percent (E), followed by cattle (Bos taurus) with 16.3 percent (P), 15.9 percent (T), 9.8 percent (E) and pigs (Sus domesticus) with 4.3 percent (P), 7.1 percent (T), 15.4 percent (E) in terms of the number of fragments. The faunal assemblage is completed by a rather small percentage of wild animals (5.2 percent P, 7.5 percent T, 3.4 percent E) but with a quite variegated spectrum of species including wild sheep (Ovis orientalis), wild goat (Capra aegagrus), onager (Equus hemionus), aurochs (Bos primigemius), gazelle (Gazella subgutturosa), fallow deer (Duma mesopotarnica), red deer (Cewus elaphus), wild boar (Sus scrofa), striped hyena (Hyaena hyaena), brown bear (Ursus arctos), fox (Vulpes vulpes), and hare (Lepus capensis). In addition we find birds and a few bones of rodents, tortoises, frogs, fishes and mollusks. One of the major results of the analysis of the faunal remains from Sabi Abyad was the evidence for a clear transformation through time of the exploitation of animals, in particular of domestic animals. The transformation of the ovicaprids is mainly visible in the change in their mortality patterns (fig. 5). The pre-Halaf phase shows a concentration of animals slaughtered at the age of three to four years, followed by a gradual decline of older animals that were probably kept for reproduction purposes. This pattern can be interpreted as exploitation of these animals essentially for their meat because they are killed at the moment of a maximum return in terms of meat weight and reproductivity. The age between two-and-a-half and four years represents the moment when rapid growth has ceased and the gain in weight no longer increases proportionally to the fodder input. A change in this pattern occurred in the subsequent later Transitional phase, in which a clear increase of animals slaughtered at a young age, between one and two years, can be detected. This mortality pattern indicates a shift towards an exploitation of these animals for their secondary products such as milk. When milk exploitation is the main goal of ovicaprid husbandry, the mortality pattern is usually characterized by a group of very young animals and by a group of mainly female adults (Payne 1973). Starting from their second winter, ewes have a reproductive life of five to six years, with an average of three years (Bates 1973:147; Cribb 1991). The data suggests that in the Transitional phase sheep were kept until a higher age than in the Early phase. The last phase (the Early Halaf phase) again shows a different pattern, with a concentration of animals killed in a few age categories, namely in the first and the second year and, then, in the third to fourth year. In a certain sense the two previous patterns are fused together. On the one hand the meat exploitation pattern is recognizable in the high percentage of the three to four year age category and in the few older animals. On the other hand the high percentage of young animals of the previous Transitional phase continues as well as the near absence of the two to three year age category, representing animals which would have been killed only one year later, after at least one lambing. The selective killing of two categories, however, could also be interpreted differently, namely that this pattern may represent only part of the original herd, that is, the part exploited and
THE ROLE OF ANIMALS
75
eaten on the site. The remaining herds would have been brought outside, far away from the site, at least for part of the year. For this reason, only certain categories are retrieved from the faunal assemblage from Sabi Abyad. These changes in the ovicaprid mortality patterns could therefore be interpreted as a transformation towards a possible beginning of a more mobile pastoralism. The transformation of cattle in the first instance became visible in a change of size. Neolithic cattle in general are large, which sometimes makes the distinction from the original wild form, the aurochs, difficult. At present, there is little doubt about the domestic status of the bovid remains from (late) Neolithic sites. Full domestic status of the four main husbandry animals as early as PPNB has recently been advocated for the site of Tell Halula, on the Middle Euphrates (Safia Segui 1999). An early stage of cattle domestication is proposed for the mid-sixth millennium site of Tell Aray 2 in the El-Rouj Basin in northwestern Syria (Hongo 1996). A decrease in size was observable on various skeletal elements of the Neolithic bovid material of Sabi Abyad. It is evident in particular in the dimensions (greatest length of the peripheral half and the smallest breadth of the shaft) of the first phalanxes (fig. 6). Large animals belong to the pre-Halaf phase. The Transitional phase shows the largest range of measurements, including large animals of the size of the previous (pre-Halaf) phase as well as some small ones, which could be compared to the small size of the subsequent (Early Halaf) phase. The Early Halaf phase shows the smallest animals with a range similar to that of the pre-Halaf phase. Another transformation of cattle is visible in the change in their mortality patterns towards a general increase of young animals. Most of the information could be gained from the stage of epiphyseal fusion of long bones. In the pre-Halaf phase less than 10 percent of the animal died at a juvenile stage and more than 50 percent survived the subadult and adult age. The presence of one fused and three unfused vertebrae suggests that killing at a high age rarely occurred. The Transitional phase shows an increase of circa 10 percent of individuals killed at a young age combined with a corresponding increase of individuals in the Middle and Late age categories. The Early Halaf phase shows a percentage of young animals killed in the first (youngljuvenile) category similar to that of the pre-Halaf phase, but, in contrast, shows a clear increase of individuals killed before they reached the Middle and Late categories (increase of 43 percent and 23 percent respectively) so that 25 percent of the cattle could reach a fully adult age. In none of the three periods very old animals are present. The transformation of pigs consists of their increase in number during the course of time. They are steadily incorporated in the husbandry economy of the site, especially as a form of meat supply. Pigs are omnivorous and can be fed an extremely large variety of food as well as human refuse. Their constraints are more related to the availability of water and humid environmental and climatic conditions. Their wide range of food resources in combination with high reproduction rates makes pigs suitable for a faster food return in terms of labor and food imput, more advantageous than ovicaprids and even larger herbivores. Pigs, therefore, are almost exclusively exploited for their meat. Fat can be exploited as well; it is mentioned in the Tell Beydar archive where it is used for cleaning wool and skin (Van Lerberghe 1996:112). Summing up the basic results reached on the basis of the data on domestic animals: 1) the increased control on domestic animals shown by a gradual transformation of their
76
77
c . CAVALLO
THE ROLE OF ANIMALS
exploitation and 2) a differentiation in the use of the animals leading to a dichotomy in the exploitation of their products: on the site in the valley animals were exploited mainly for their meat and some herds of sheep and goat were kept, while the upland steppe witnessed an extensive exploitation as grazing grounds for the ovicaprids.
which was dominated by poplar (Populus euphratica), tamarisk (Tamarix sp.), willow shrubs (Salix sp.) and dense reed beds (van Zeist et al. 1988; Gremmen and Bottema 1991:130). The ratio between the two categories of wild animals found at Sabi Abyad varies in the course of the site's occupation, providing evidence for a steady, increasing exploitation of the steppe, as was already suggested in the case of the ovicaprids (fig. 7).
Wild animals, although found at Sabi Abyad in small percentages only, provided relevant information for the reconstruction of the environmental exploitation. According to their preference of habitat, the wild animals found at Sabi Abyad were grouped into two categories: 1) those representative of a dry environment (steppe) and 2) those representative of a moister environment, the more covered area (forest) of willow shrubs, poplar and tamarisk along the river. Onager and gazelle were the main hunted species. Both are typical steppe animals. Today onagers, like other wild equids, live in open, flat and rather dry areas. They can withstand drought and high temperatures. In Iran kulans can tolerate daytime temperatures up to 58' (Groves 1974). Although they usually drink daily, they can survive without water for at least two to three days (Heptner et al. 1966). Their diet varies according to the season and can include 110 taxa, although they prefer to feed on juicy low-growing plants, such as Carex sp., Poa sp. and Stipa sp. (Heptner et al. 1966:849; Groves 1974:102). They can also feed on salty herbaceous plants (Anabasis and Salsola) and shrubs such as Artemisia (ibid.). Gazelles have a similar habitat. Today the goitred gazelles live on gravel plains and limestone plateaus, like the onager with whose herds they have been observed to run (Groves 1974). They feed on 35 to 40 different species of grass and a large variety of shrubs, and on leaves of trees like the pistachio tree (Heptner et al. 1966:537-8). The dwarf scrub, Rhanterium eppaposum, is one of the most important food plants of this species in central Arabia (Harrison 1968:364). Gazelles are also extremely drought and heat resistant and can survive without water for three to seven days (Heptner et al. 1966:538). Besides onager and gazelle, two other species belonging to a dry environment were identified,3 namely the wild sheep -the Asiatic mouflon- and the wild goat. Judging from the numerous remains found on archaeological sites, their ancient geographical dispersion must have been much more southwards than the modern limit of the Taurus mountains in south Anatolia (Uerpmann 1987). Apart from these steppe species a large variety of other game species are indicators of an exploitation of moister environments. Within these species we find red deer, fallow deer and the brown bear as typical forest species, together with the wild boar and aurochs. Today the Duma, or fallow deer, lives in dense impenetrable jungles of white poplar and tamarisk that border the rivers of southwestern Persia and south Mesopotamia (Haltenorth 1959). They are browsers, feeding on young shoots of poplar bushes (Harrison 1968:368). The wild boar, whose omnivorous diet includes rhizomes of aquatic and marsh plants, today mainly lives in dense thickets and reed jungles (Harrison 1968:375-376). Aurochs require a less restricted habitat and are more adaptable to various environments. However, they also demand moister environments compared to the ovicaprids, onager and gazelle, and a habitat with dense continuous vegetation. These species could have found their typical habitat in the riverine forest along the Balikh valley The identification of wild sheep and goat was made according to the morphological characteristics of the horn-cores and to the size of the bones, clearly larger than the rest of the domestic ovicaprids sample (Cavallo 199754-58).
INTERPRETATION
The data presented above could indicate that the Sabi Abyad community towards the last phase of occupation of the site turned more and more into a twofold economy, with part of the population involved in a more mobile pastoralism, favored by the possibility of developing secondary products such as milk, as has been observed in the Transitional phase. Milk products are considered to be a prerequisite for the development of a pastoral economy which usually depends on this type of product (Bar-Yosef and Khazanov 1992:16). The permanent character of the village of Sabi Abyad, however, remains, which is testified by the constant high percentage of domestic animals. The sedentary character of the site furthermore follows from the continuing evidence of crop cultivation and from the architectural features, as well as from the increase of pigs' remains, animals usually associated with sedentary communities. The incidence of cattle remains however decreases, indicating a scarce or more likely absent exploitation of these animals for agricultural purposes. The young age of the cattle and the absence of stress-related pathological deformations or anomalies on the bones point in this direction as well (Bartosiewicz et al. 1997). Other archaeological evidence seems to suggest that already during the Transitional phase part of the Sabi Abyad population was not permanently present at the village. The numerous seals concentrated in a few rooms of a storage building from level 6 (Transitional phase) provide evidence of storage of goods and of devices of property control (Akkermans and Duistermaat 1997). The size of the possible containers, the number and type of impressions, and the concentration in a limited area of the village led the excavators to suppose that these goods did not belong to people living in the village, for whom this kind of control would not be necessary, but to people not permanently present at the village, these people being part of the nomadic population. The detailed analysis of the architectural space and of the distribution of the finds reflects a constant formal and functional duality, which would be the reflection of two distinct social and economic groups, the sedentarian agriculturists and the mobile (nomadic) pastoralists (Verhoeven 1996). The twofold economy that would have characterized Sabi Abyad becomes even more evident from the occupation of the valley during the subsequent late Halafian period, in the first half of the fifth millennium. This period is distinguished by larger, permanently occupied sites with a scattered number of small, briefly occupied sites. Sites like Khirbet esh-Shenef, with two phases of occupation divided by a modest erosion level, or Damishliyya, with a very small area of occupation devoid of archaeological features, show evidence of small-scale and repeated occupation (Akkermans 1993). These sites might have functioned as small-scale agricultural sites or as purely pastoral camps (Damishlliyya) that were more specialized in different degrees in pastoral activities than
78
C. CAVALLO
THE ROLE OF ANIMALS
the larger, permanently occupied sites. The location of some of these sites on the edge of the steppe-valley in combination with a more mobile character of the population would lead to a major exploitation of wild steppe fauna. The high percentage of wild animals retrieved would also indicate this: at Khirbet esh-Shenef wild animals make up 36 percent of the identified assemblage while at Shams ed-Din more than half of the assemblage (54 percent) consists of wild animals (Hendrichs 1990; Uerpmann 1982).
of animal husbandry for the ancient population of Sabi Abyad. According to the risk reduction model developed by Halstead and O'Shea (1989) there are four basic responses to food scarcity deriving from risk and uncertainty: diversification, mobility, storage, and exchange. Concerning the animals, diversification and mobility are the most applicable forms to the case of Sabi Abyad. Diversification would consist in the first instance of different types of products derived from the domestic animals, such as milk in addition to meat. Pastoralism can be considered a form of diversification as well, especially if we consider pastoralism as a form of exploitation of a wider environment including different types of environments, as seems to have been especially the case during the last phases of occupation of Tell Sabi Abyad. A third form of diversification is represented by hunting, which, like pastoralism, is characterized by a rather high degree of mobility. Hunting represents diversification not only in enlarging the variety of species, but also as another form of environmental exploitation in addition to husbandry. Therefore, both hunting and husbandry are indications of (increasing) use of the steppe as an integrated part of the economy of the site. In addition, the study of the wild species would suggest that hunting was concentrated in a certain part of the year. Results of the analysis of the dentition of the gazelle would indicate that they were killed mainly in fall and winter. Most of the species of birds identified are winter visitors. At this point it can be tried to reconstruct a kind of economic calendar in which the steppe can be regarded as an integrated part of the agricultural system, explaining the nature and the degree of the integration of crop and livestock. The analysis of the botanical remains indicates that at Sabi Abyad dry farming was practiced (Van Zeist and Waterbolk-van Rijn 1989, 1996). It was based on cereal cultivation, mainly of emmer wheat, followed by einkorn and, to a lesser extent, by barley. These are winter crops, which are nowadays harvested in early spring while the fields are sown in autumn as soon as the rains start (Lewis 1988:688). Spring would represent the period of most activities around the village, like harvesting, lambing and consequent exploitation of dairy products as well as the slaughtering of the young animals. Hunting would have taken place in autumn and winter, when, on the one hand more prey is available, and, on the other hand, people are free from land labor and pastoralism. In this scenario it is easy to understand that cultivation and animal husbandry, though taking different directions within the landscape and the society, were integrated parts of a common agricultural system.
Different reasons can underlie the economical transformations within the Halaf community. The nature of the analyzed material leads to an emphasis on the role of a possible degradation of the environment. This argument is supported by the culling pattern of the ovicarpid sample of the Transitional phase, which could also be partly interpreted as the result of a herd-security strategy. Redding (1981:204) points out that the age of six months to two years is the optima1 age for slaughter in case the security of the herd is more important than the maximization of its energy (that is, exploitation mainly for its meat). In the same Transitional period there is a slight increase in hunting as well as in the number of species hunted. This, combined with an increase of goats in the later Early Halaf phase, might be a response to a crisis period of low animal feeding resources which might have been resolved by moving part of the ovicaprid herds away from the site. On the other hand, an increase in human population, which took place in the valley towards the end of the sixth and beginning of the fifth millennium and which is reflected in a development of settlements, is elsewhere emphasized (Akkermans 1993:l86ff.). The two explanations, an environmental one and a demographic one, are, however, closely linked. A degradation of the areas around the site due to overexploitation, both through farming and grazing, could have been worsened by the increased pressure on agricultural yields due to a larger population demand. Intensification of agricultural production without an adequate repletion of the soil moisture and organic nutrients would lead to a rapid exhaustion of the soil (see Wilkinson, this volume). In addition, we cannot avoid taking the geographical location of the site into consideration when trying to define the characteristics of the region. The Balikh valley is located in a semi-arid area of steppe. The present climatic conditions of the area are characterized by a pronounced aridity with a low annual precipitation averaging little more than 300 mm. Precipitation is characterized by a large variation not only during the year (summer-winter), but also in the course of different years. The isohyet of 250mm, which is considered the limit of rain-fed agriculture, may vary extremely as a consequence of a few subsequent dry years from the middle of the Balikh valley northwards, until the entire region is outside the range of rain-fed agriculture (Wirth 1971: maps 3 and 4). According to pedological research the limit of this isohyet was approximately five kilometers south of Sabi Abyad (Wilkinson 1996). The location of the line still was reflected in the pattern of the settlement distribution of the valley in the nineteenth century. The majority of the population was concentrated in the north-western part of the valley, while the southern, more arid part where agriculture necessitated irrigation was virtually devoid of permanent sites and was occupied by Bedouins living in tent villages (Lewis 1988:688). The characteristic of an extreme sensitivity to climatic fluctuations in combination with the crucial position of the site in this marginal area on the border of the dry farming zone could therefore have caused a high degree of risk in crop failures or in the intensity
79
80
c. CAVALLO
REFERENCES
Akkermans P. M. M. G. Villages in the Steppe: Later Neolithic Settlement and Subsistence in the 1993 Balikh Valley, Northern Syria, Ann Arbor. Tell Sabi Abyad - The Late Neolithic Settlement, Istanbul. 1996 Akkermans, P. M. M. G. and K. Duistermaat Of Storage and Nomads: The Sealings from the Late Neolithic Sabi Abyad, 1997 Syria, Pale'orient: 2212, 17-44. Akkermans P. M. M. G and M. Verhoeven An Image of Complexity: The Burnt Village at Late Neolithic Sabi Abyad, 1995 Syria, American Journal of Archaeology 9911, 5-32. Bartosiewicz L., W. van Neer and A. Lentacker Draught Cattle: Their Osteological Identification and History, Tervuren. 1997 Bar-Yosef 0 . and A. Khazanov Pastoralism in the Levant: Archaeological Materials iiz Anthropological 1992 Perspectives, Madison. Buitenhuis H. and A. T. Clason Archaeozoology of the Near East, Leiden. 1993 Buitenhuis H. and H.-P. Uerpmann Archaeozoology of the Near East 11, Leiden. 1995 Buitenhuis H., L. Bartosiewicz and A. M. Choyke Archaeozoology of the Near East 111, Groningen. 1998 Cavallo C. The Animal Remains - A Preliminary Account, 475-520 in P. M. M. G. 1996 Akkermansed., Tell Sabi Abyad: The Late Neolithic Settlement, Istanbul. 1997a Animals in the Steppe - A Zooarchaeological Analysis of Later Neolithic Tell Sabi Abyad, Syria, Ph.D. dissertation, University of Amsterdam. 1997b Animal Remains Enclosed in Oval Clay Objects from the "Burnt Village" of Tell Sabi Abyad, Northern Syria, Anthropozoologica 25-26, 663-670. Efremov I. A. Taphonomy: A New Branch of Palaeontology, Pan-American Geologist 74, 1940 81-93. Gremmen W. H. E. and S. Bottema Palynological Investigations in the Syrian Gazira, 105-116 in H. Kiihne ed., 1991 Die rezente Umwelt von Tall St?b Hamad und Daten zur Umweltrekonstruktion der assyrischen Stadt Dur-katlimmu, Berlin. Groves C. P. Horses, Asses and Zebras in the Wild, Hollywood, Fla. 1974 Halstead P. and J. O'Shea. Introduction, 1-7 in P. Halstead and J. O'Shea eds., Bad year economics: 1989 cultural responses to risk and uncertainty, Cambridge. Haltenorth T. Beitrag zur Kenntnis des Mesopotamischen Damhirsches-Cewus (Duma) 1959 mesopotamicus Brooke, 1875- und zur Stammes- und Verbreitungsgeschichte der Darnhirsche allgemein, Saugetierkundliche Mitteilungen 7, 1-89. Harrison D. L. The mammals of Arabia, vol. 11: Carnivora - Artiodactyla - Hyracoidea, 1968 London.
THE ROLE OF ANIMALS
81
Hendrichs U. 1990 De dierresten van Khirbet esh-SheneJ een Laat-Halaf vindplaats in NoordSyrie, MA Thesis IPP, University of Amsterdam. Heptner V. G., A. A. Nasimovi and A. G. Bannicov 1966 Die Saugetiere der Sowjetunion, Band I: Paarhufer und Unpaarhufer, Jena. Hongo H. 1996 Faunal Remains from Tell Aray 2, Northwestern Syria, PalLorient 2211, 125144. Lewis N. 1988 The Ballkh Valley and its People, 683-695 in M. N. van Loon ed., Hammam et-Turkman I, vol. 2, Istanbul. Redding R. W. 198 1 Decision Making in Subsistence Herding of Sheep and Goats in the Middle East, Ph.D. dissertation, University of Michigan. Saiia Segui M. Arqueologia de la domesticacidn animal - La gestidn de 10s recursos 1999 animales en Tell Halula (Valle del ~ufrates- Siria) del 8.800 a1 7.000 BP, Treballs d'Arqueologia del Pr6xim Orient, I, Barcelona. Shiffer M. B. Archaeological Context and Systemic Context, American Antiquity 37, 1561972 165. Uerpmann H. P. Faunal Remains from Shams ed-Din Tannira, a Halafian Site in Northern 1982 Syria, Berytus 30, 3-52. The Ancient Distribution of Ungulate Mammals in the Middle East, Beihefte 1987 zum Tiibinger Atlas des Vorderen Orients A, 27, Wiesbaden. Van Lerberghe K. The Livestock, 107-117 in F. Ismail et al., Administrative Documents from 1996 Tell Beydar (Subartu 2), Turnhout. Verhoeven M. Excavation at Tell Sabi Abyad, a Later Pre-Pottery Neolithic B Village in the 1994 Balikh Valley, Northern Syria, Orient Express 1, 9-12. 1999 An Archaeological Ethnography of a Neolithic Community: Space, Place and Social Relations in the Burnt Village at Tell Sabi Abyad, Syria, Istanbul. Verhoeven M. and P. Kranendonk The Excavations: Stratigraphy and Architecture, 25-118 in P. M. M. G. 1996 Akkermans ed., Tell Sabi Abyad - The late Neolithic settlement, Istanbul. Wijngaarden-Bakker L. H. van 1989 The Animal Remains from Tell Sabi Abyad -Square P14, 301-323 in P. M. M. G. Akkermans ed., Excavations at Tell Sabi Abyad - Prehistoric Investigations in the Balikh Valley, Northern Syria, B.A.R. International Series 468, Oxford. Wilkinson T. J. 1996 Sabi Abyad: The Geoarchaeology of a Complex Landscape, 1-24 in P. M. M. G. Akkermans ed., Tell Sabi Abyad - The Late Neolithic Settlement, Istanbul. Wirth, E. 1 Syrien. Eine geographische Landeskunde, Darmstadt.
82
c. CAVALLO
THE ROLE OF ANIMALS
Zeist, W. van and W. Waterbolk-van Rooijen Plant Remains from Tell Sabi Abyad, 325-335 in P. M. M. G . Akkermans 1989 ed., Excavations at Tell Sabi Abyad - Prehistoric Investigations in the Balikh Valley, Northern Syria, B.A.R. International Series 468, Oxford. The Cultivated and Wild Plants, 521-550 in: P. M. M. G. Akkermans ed., 1996 Tell Sabi Abyad - The Late Neolithic Settlement, Istanbul. Zeist W. van, W. Waterbolk-van Rooijen and S. Bottema Some Notes on the Plant Husbandry of Tell Hammam et-Turkman, 705-715 1988 in M. N. van Loon ed., Hammam et-Turkman I, Istanbul, Appendix 2.
Fig. 1. Humerus of sheep with cutmark (photo: M. IJdo, IPPJAAC University of Amsterdam).
83
84
c . CAVALLO
THE ROLE OF ANIMALS
Fig. 2a. Suid lower jaw with a blow inflicted on its lower part (photo: M. IJdo, IPPIAAC University of Amsterdam).
Fig. 4. Awl made from an ovicaprid metapodial (photo: M. IJdo, IPP/AAC University of Amsterdam). Fig. 3. Femur of onager with gnawing marks (photo: M. IJdo, IPPIAAC University of Amsterdam).
Fig. 2b. Detail.
THE ROLE OF ANIMALS
C. CAVALLO
Early Halaf
30 25 20 %15
10 5 0
3-4 4-5
5-6
6-7
7-8 8-9 9-10
years
Transitional
GLpe Fig. 6. Measurements of the greatest length of the peripheral half (GLpe) versus the smallest breadth of the diaphysis (SD) of the first bovid phalanges. Pre-Halaf: ; Transitional: W ; Early Halaf: 0
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9 9-10
years
Transitional
CI 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10
years
dry
environment
Fig. 5. Ovicaprid mortality based on mandibular wear according to the Payne's wear stages (Payne 1973). Fig. 7. Percentage of wild animals representative of dry moist environment.
PlHANS LXXXVIII, 2000
Middle Assyrian Expansion and Settlement Development in the Syrian Jazira: The View from the Balikh Valley* Jerry D. Lyon (University of Chicago) Most investigations of regional settlement development are founded upon the assumption that settlement distribution is not a random process, and that the locations of settlement are selected or arise based on a combination of environmental and historical factors. Local ecological, economic, and political variables, such as topography, agricultural potential, the historical trajectory of agricultural land use, and the spacing of pre-existing permanent villages must be considered in the reconstruction of any agricultural settlement system. Due to their potential for exerting influence and managing information over increasingly large areas, however, state level societies may be more fully understood when viewed within a larger context. Such a view emphasizes the role of large-scale factors and explicitly external forces, or historical contingencies, in regional settlement development. Large-scale factors may include the distribution of necessary non-local resources, the degree of interdependence of foreign economies, as well as the status, power, and orientation of foreign political elites. Such factors may be manifested as a continuum of interregional or global processes, from trade and collaboration at one end, to imperial expansion, conquest, colonization, and incorporation at the other. Such large-scale processes are particularly well-documented in the historical and archaeological record of greater Mesopotamia, which chronicles the episodic expansion and contraction of complex political systems from the rule of Sargon of Agade to that of Sargon 11 and beyond. This paper presents preliminary information on the expansion or intrusion of Middle Assyrian settlement into one portion of the Syrian Jazira previously incorporated in the MitanniIHanigalbat state, the Balikh valley. Analysis of historic settlement development may draw on multiple lines of textual and archaeological evidence. Contemporary textual sources provide important contextual information on the function and ethnic or political affiliation and economic orientation of specific settlements, information on specific historical events that have effected the development of settlement, and, in some cases, insight into the motivations behind events and processes. Recent studies that incorporate archaeological and documentary evidence (e.g. Knapp 1992; Smith 1992) point out the necessity of considering the two lines of evidence separately and critically in terms of the logic inherent within each discipline (Feinman 1997; Kepecs 1997). In this respect, it is hoped that research presented here complements the work of Wiggermann (this volume), by viewing Middle Assyrian expansion within a regional context and within the broader temporal context of late second millennium settlement development.
90
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
91
MIDDLE ASSYRIAN EXPANSION
THE BALIKH VALLEY DATASET
Most historical reconstructions of late second millennium B.C. northern Mesopotamia suggest that following the decline of the Mitanni state, northern Syria was divided between the HittiteKarchemish and Middle Assyrian realms with the muchreduced Mitanni state of Hanigalbat in an intermediate but subordinate position. Texts from central Anatolia and the Assyrian heartland indicate extensive interaction and competition in the border region between the Euphrates and BalikhIKhabur rivers, especially in the vicinity of Harran. Assyrian royal inscriptions (Grayson 1972) report that, following the resurgence of Assur under ASSur-uballit I, Adad-nirari I (1295-1264, regnal years after Freydank 1991:188) led campaigns to the west. At this time the Assyrian army defeated Hanigalbat, put down subsequent revolts, and instituted a system of tribute extraction. Adad-nirari's successors Shalmaneser 1 (1263-1234) and TukultiNinurta I (1233-1197) continued the wars in Syria and increased activity in the Hittitecontrolled territories along the Euphrates in the vicinity of Carchemish (Grayson 197223283). Tukulti-Ninurta I even reports crossing the Euphrates and deporting thousands of "Hittites" to Assur (Grayson 1972:118). These royal accounts, details of which are supported by collaborative economic and administrative texts, e.g., the ethnic affiliation and designations for slaves or laborers, describe a pattern of repeated military conflict throughout the 13th century, often upon succession to the throne in Assur, which resulted in continued Assyrian expansion into and control of the former Hanigalbat territory (Harrak 1987).
Due to the relative intensity of recent research in the Balikh valley in northern Syria, we have a substantial body of data relevant to the discussion of late second millennium B.C. settlement processes. The research presented here is primarily a reassessment of survey data collected by Peter M. M. G. Akkermans in 1983 (Akkermans 1984) in light of recenf finds from the Middle Assyrian settlement at Tell Sabi Abyad. The Bronze Age material from the Balikh was originally treated by Hans Curvers in an unpublished dissertation at the University of Amsterdam. To supplement this I collected additional survey data in 1995. The study also draws on previous research in the area: for example, excavations of relevant sites in the valley, including Mallowan's (1946) work, recent excavations at Tell Sabi Abyad (Akkermans and Rossmeisl 1990; Akkermans et al. 1992), Tell Hammam et-Turlunan (van Loon 1982), and Tell BiCa(preliminary reports in MDOG), as well as recent regional landscape studies conducted by Tony Wilkinson (1996, 1998).
These processes have been seen as part of a more general political development from the initial Middle Assyrian annexation of Hanigalbat in the late 13th century to the full development of the Late Assyrian imperial system of the 8th century (Liverani 1988; Postgate 1992). This development would have involved 1) an influx of Assyrian settlers, 2) efforts at agricultural colonization, and 3) the further development of networks of communications and transportation of goods (trade and tribute). Liverani (1988) has described this as the transition from "network-empire'' to "territorial-empire." Archaeological research in northern Iraq and Syria has documented the process of Assyrian territorial expansion and political and economic reorientation in the 13th through 1lth centuries B.C. at a number of sites from the Assyrian heartland proper and the Syrian Jazira. In general, the process is demonstrated archaeologically by the succession of "Mitannian" or "Nuzi" and "Middle Assyrian" material culture assemblages, e.g., at Tell Billa (Sibaniba) (Speiser 1932-33), Tell Mohammed Arab (Roaf 1984), Tell a1 Rimah (Postgate et al. 1997), Tell al-Hawa (Wilkinson and Tucker 1995), Tell Brak (Oates et al. 1997), Tell Hamidiya (Eichler et al. 1985; Eichler 1990), Tell Mohammed Diyab (Faivre 1992; Lyonnet 1990), and Tell Fakhariyah (McEwan 1958). Evidence of changes in material culture is drawn primarily from ceramic assemblages (Pfdzner 1995) and cylinder seal style preferences (Kantor 1958; Matthews 1990; Nissen 1967). More specific correlation of changes in material culture with political affiliation is corroborated most securely with reference to textual materials from the sites themselves; for example, by explicit reference to Assyrian administration, implicit use of Assyrian administrative conventions (e.g. limu dates), and to a lesser extent, the occurrence of linguistically Assyrian personal names and terminology (e.g. see commentary in Jas 1990).
The Balikh is one of two perennial rivers bisecting the Syrian Jazira and connecting various climatic microzones and a pronounced precipitatian gradient from north to south (Figure 1). Important for the distribution of settlement in the area is the theoretical limit for rainfed agriculture at the 250 mm isohyet (Brichambaut and WallCn 1963; van Zeist 1994; Wilkinson 1994, 1998). Recent analysis of satellite imagery of the area (Beaumont 1996) shows a very close correlation between vegetation density and the theoretical limits calculated by Brichambaut and WallCn. A map of population distribution in 1945 shows significant differences between the northern and southern portions of the valley, which corresponds to the distribution of rainfall (Lewis 1988: P1.207). In the Balikh valley, the theoretical limit for rainfed agriculture occurs near the center of the valley, in the vicinity of Tell Zkero (BS-152), 6 km south of Tell Hammam et-Turkman (Figure 2). Although irrigation canals were important throughout the valley, in the southern Balikh, agricultural settlement would have been largely dependent on irrigation. Dry-farming would have been more feasible in the northern portion of the valley, where it attains some 300+ mm per annum. Although the Balikh represents a relatively minor tributary of the larger Euphrates river system, its more or less perennial flow attracted settlement from a very early date. Stemming from the Jullab river, 'Ain al-'Ark, and Wadi Qaramoukh, the river connects the Harran plain in the north to Raqqa in the Euphrates valley in the south. At the same time, the river is a virtual oasis in the midst of the steppic zones to the east and west. Despite the relative or local importance of the river system and valley, Mallowan (1946:114) noted that "it would be exaggerating the importance of the district to view it as anything more than a provincial backwater." However, due to its peripheral nature and the attraction for settlement it held, the Balikh valley may be seen as a significant part of the ecological and cultural boundary that connects and separates the Euphrates and Tigris regions to the east and west, as well as the lowland and highland zones to the north and south. The region also represents a zone of intersecting trade routes that extend north to southeastern and central Anatolia, west to the Euphrates valley and on to the Mediterranean coast, south to the Middle Euphrates and Babylonia, and east to the Assyrian heartland. If this "backwater" is important at all, it is due to its dual nature as
92
93
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
both an edge of physical, cultural, and possibly political landscapes and a special type of "functional center" (Crumley 1976:67; Marquardt and Crumley 1987:8-9) connecting such landscapes. In geo-political terms, for example, the Balikh valley may also be seen as a Late Bronze Age territorial frontier at the limit of one expanding polity (Middle Assyrian) and a political frontier or border both delimiting and connecting the territorial limit of another polity (Hatti).
systematic or targeted exploration, especially given the likelihood of destruction by modem ploughing.
Previous research Study of the history of settlement in the Balikh valley was begun in earnest by Sir Max Mallowan in 1938. Following work at Tell Brak, he conducted soundings and limited excavations at five mounded sites in the middle and northern stretches of the Balikh valley: Tell Mefesh, Tell Sahlan, Tell Aswad, Tell Jidle, and Tell (Hammam) Ibn esh-Shahab (Mallowan 1946). Based on this, Mallowan presented a provisional sequence of settlement in the Balikh from the fourth millennium to the mid-second millennium B.C. Of relevance to the present discussion of settlement trajectories in the second millennium B.C are the limited soundings at Tell Jidle (BS-276), Tell (Harnrnam) Ibn esh-Shahab (BS-275), and Tell Sahlan (BS-247). Mallowan's pioneering work provides a general picture of regional decline in the 14th century B.C. that was frequently marked by violent conflagration. Although the chronology and historical causes of decline cited by Mallowan may be open to revision today, this information must be considered in any reconstruction of settlement development in the Balikh valley. In 1983, Peter M. M. G . Akkermans of the National Museum of Antiquities of the Netherlands, Leiden, conducted the University of Amsterdam Balikh Survey (Akkermans 1984, 1993). The goal of this survey was to inventory ancient settlements in the valley and to study settlement organization and development through time. As the survey methodology has been described in detail by Akkermans (1993), general procedures are only summarized here. Potential sites were first identified by Akkermans with the aid of satellite images, aerial photographs, topographic maps, and pedological reports from the University of Utrecht, and, based on features visible in these media, a survey map of possible site locations was produced. Field-checking of suspected sites led to the discovery of 210 archeological sites with occupations ranging from the 8th-6th millennia B.C. (Akkermans 1993) through the medieval Islamic periods (Bart1 1994). Although the 1983 survey is not considered a full-coverage survey of the valley (sensu Fish and Kowalewski 1990), it stands as the most extensive and systematically collected database of settlement in the Balikh valley. Akkermans (1993) notes that as the survey was restricted to that portion of the valley extending from the Syro-Turkish border in the north to the Euphrates River in the south, it reflects only a portion of a more extensive drainage system. Within this general area, the survey was further restricted to the flood plain and immediate edges of bordering plateaus. Because aerial photographs yielded very few traces of cultural features outside the valley proper (Akkermans 1993: 139), we can be fairly confident that few substantial ancient population centers would be present there. On the other hand, numerous small seasonal habitations or camp sites of nomadic or semi-nomadic (Bedouin farmers) would be difficult to detect without
More recently, additional investigation of the Balikh valley has occurred under the aegis of the Western Jazira Archaeological Landscape Project directed by Tony Wilkinson (1996, 1998). This project was conducted for the Oriental Institute, University of Chicago, in collaboration with Akkermans' Tell Sabi Abyad excavations (Akkermans and le Miere 1992; Akkermans et al. 1993; Akkermans and Verhoeven 1995). This research program involved both an intensive study of the archaeological landscape around Tell Sabi Abyad, including pedological, geomorphological, and cultural landscape studies, and an evaluation of the geomorphological contexts of other sites in the valley. Surveys conducted during the 1993, 1994, and 1995 field seasons led to the identification of a number of low-mounded sites along the banks of the Balikh River, sites that were not recognizable on aerial photographs (Wilkinson 1996:14). The Western Jazira Project has thus increased our coverage of the valley, while at the same time indicating the potential limitations of the overall regional database. Other surveys in the Balikh and adjacent valleys have documented archaeological sites, although in each case the emphasis was on prehistoric periods (e.g. Copeland 1979, 1982) or the investigation of the most prominent mounds for future excavations (C6rdoba 1988). We await additional survey and full publication of previous survey results in the Qaramokh valley and in the steppe areas east and west of the Balikh valley (e.g., Einweg 1993; but see Danti 1998). Middle Assyrian expansion re-survey In order to study late second millennium settlement development and Middle Assyrian expansion in the Balikh valley, reassessment of data collected during Akkerrnans' 1983 survey and retrieval and interpretation of new surface collections was proposed as part of the collaborative effort of the Oriental Institute's Western Jazira Archaeological Landscape Project and Peter Akkermans' Tell Sabi Abyad Excavation Project. The goal of this study was to confirm late second millennium B.C. settlement components and to attempt to identify Middle Assyrian occupations with reference to a growing corpus of Middle Assyrian ceramics from Tell Sabi Abyad. Fieldwork conducted by the author in 1995 involved the investigation of 30 sites with occupations that were originally attributed to the late second millennium B.C. Nonprobabilistic sampling of site surfaces was as thorough as time allowed. The probability of identifying Middle Assyrian diagnostics was expected to be proportionate to the total amount of time spent examining site surfaces. In most cases, collections of diagnostic sherds were made and sketch maps were drawn for comparison with the 1983 survey record. Sampling areas included entire mounds if they were less than 1 ha in area, and subdivisions of larger sites according to topographic units, as was the case during the original Balikh valley survey (Akkermans 1993). Diagnostic sherds, including rims, bases, and decorated sherds, as well as occasional miscellaneous finds such as figurines, beads, and stone vessels were drawn or collected.
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
Within the Balikh valley, our principle ceramic comparanda come from the Middle Bronze (Hammam VII) and Late Bronze (Hammam VIII) levels at Tell Hammam etTurkman (Curvers 1988; Smits 1988) and the Mitanni and Middle Assyrian levels at Tell Sabi Abyad (Akkermans and Rossmeisl1990; Akkermans et al. 1992). The ceramic type series used here follows period designations for the late second millennium B.C. originally published by Smits (1988): Balikh Period VIIIA and VIIIB (Table 1). These two phases roughly correspond to the Mitanni and Middle Assyrian phases at Tell Sabi Abyad, respectively. Period V I W settlements have been identified based on the presence of diagnostic Middle Assyrian ceramics. Middle Assyrian ceramic types are, admittedly, part of a restricted assemblage that may have served specific functions in Assyrian administration and business. Therefore, the absence of such forms at a site does not preclude the presence of contemporary, non-Assyrian occupations. This is especially the case in the extreme southern portion of the Balikh valley, where the majority of late second millennium ceramics are sand-tempered wares with forms more similar to those from the upper Euphrates (Tabqa) sites than those of the upper Balikh and Khabur drainages. Similar problems of distinguishing between chronological and cultural or political variation in ceramics have been treated by Parker (1997) for the Neo-Assyrian period.
Three general problems with utilizing spatial analytical techniques in the reconstruction of past settlement systems must be addressed: 1) use of a potentially incomplete record of past settlements due to uncontrollable biasing processes; 2) correspondence problems between estimated site size, population density and the functional size of sites; and 3) the correspondence between area under investigation and the areas or scale of past settlement systems.
94
The number of diagnostic sherds found on a site may provide one indication of the significance or duration of the occupation. I have attempted to follow a methodology similar to that laid out by Wilkinson and Tucker (1995: 17) for the North Jazira Survey. Accordingly, sites are classified as either "minor occupations" (c3 diagnostic sherds), "significant occupations" (3-5 diagnostic sherds), and "major occupations" (>5 sherds) (Table 2). These designations may be related to the size and occupational stability (longterm vs. short-term occupation) of settlements (Horne 1993, 1994). Site size estimates are based on paced measurements and site sketches. Most site areas were calculated using a simple formula for the area of an oval: (length/width/2)1. On large, multi-component sites, settlement areas for each period are estimated based on the extent of diagnostic ceramics relative to mound subdivisions. For sites measuring less than 1 ha, the entire site area was used for each period.
Problems and limitations of survey data For analysis of the spatial patterning of archaeological remains there exist a number of analytical techniques borrowed and adapted from human geography. Common methods employed here for identifying and interpreting settlement patterns include analyses that emphasize synchronic patterning, such as the gravity model of interaction (Alden 1979; Haggett 1965:35-40; Johnson 1977:481-487; Plog 1976), reconstruction of hypothetical territories based on Thiessen (or Voronoi) polygons (Boots 1986, Grant 1986), and rank-size analysis (e.g., Crumley 1976; Hodder 1979; Johnson 1980, 1981; Pearson 1980). Methods that highlight diachronic patterning include the comparison of these patterns through time and the determination of demographic trends based on site size and population estimates and continuity of settlement location.
95
The most fundamental source of potential error in the analysis of settlement patterns is the coverage and reliability of the survey dataset. This is directly related to variable preservation and visibility of sites and components on sites, as well as to survey methodology. In the present case, because the survey procedure started with "threedimensional" landscape features recognized through remote sensing, we may expect that some low-lying or "two-dimensional" artifact scatters have been missed. As noted above, Wilkinson's intensive pedestrian survey along stretches of the Balikh River in the microregion of Tell Sabi Abyad and Tell Hammam et-Turkman has revealed several lowmounded sites that were not visible on the aerial photographs. Likewise, according to Wilkinson's (1996, 1998) geomorphological and pedological information, low site visibility due to aggradation along the Balikh River and its tributaries suggests that sites in these areas may be missing from the regional dataset. Sites potentially missing from this dataset are probably rather small in size and may be characterized by low occupational stability, i.e. short-term or seasonal habitation. All but one or two sites discovered along the Balikh River north of Tell Hammam et-Turkman in the area intensively surveyed by Wilkinson are smaller than one hectare in area. That all such sites are not missing from the survey record, however, is indicated by site BS-86, a very low site that was identified in 1983 in the cotton fields surrounding Tell es-Seman (BS-83). Discovery of this site suggests that survey coverage around large sites was more intensive than in empty areas or areas of low site density, another biasing process. The geomorphological study of Wilkinson (1996, 1998) allows us to divide the survey area into zones of varying degrees of visibility. Generally, good visibility is restricted to the terraces to the east and west of the river valley. Here, smaller sites and landscape features such as hollow ways, canals, areas of upcast, and off-site sherd scatters are visible on the surface. Areas of low or poor archaeological visibility are restricted to the flood plain of the Balikh and its tributaries, which, for the most part, occur on the east side of the valley. Moderate visibility characterizes the valley north of Tell Hammam et-Turkman (BS-175). To the south, however, in the broad plain around Tell es-Seman (BS-83), the entire width of the valley is characterized by low visibility. South of there, where the valley narrows before approaching the Euphrates, visibility improves and small sites should be visible if present. The potential under-representation of small sites (i.e., less than one hectare) should not effect the reconstruction of general settlement patterns, such as the division of the valley into territories around larger village sites with Thiessen polygons. Likewise, the accuracy of gravity model analysis might not be severely effected because small sites "weigh" less and frequently occur within interaction "plateaus" around larger sites. According to Johnson (1987), rank-size curves excluding sites measuring less than one hectare may be relatively accurate reflections of settlement systems. Rank-size curves
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
typically fall-off steeply below one hectare, which Johnson (1987: 109) suggests may reflect the size below which "settlement economic viability decreases rapidly" or the size below which the scale of agricultural productivity or cooperative requirements are not relevant to the regional economic system, security, or administrative requirements. The under-representation of small sites, however, will bias our view of microregional patterning within site catchment areas, and the size and distribution of dispersed or rural settlement.
spring of 'Ain al-'Aris 4 km south of the modem Syro-Turkish border. In this sense, "the Balikh valley forms a distinct geographical and ecological unit located entirely within our area of survey" (Akkermans 1993:139). In such an arid environment, hydrological and climatic conditions will have determined to a large extent the distribution of prehistoric settlements in the valley. As noted above, rainfall data and landscape characteristics suggest that rainfed cultivation was only possible in the northern portion of the valley, and that irrigation was necessary for the cultivation of crops further south (e.g., Wilkinson 1998; Beaumont 1996). That being the case, we might expect distinct settlement patterns in the two portions of the valley (cf. Beaumont 1996; Lewis 1988). For example, settlements dependent on irrigation systems should tend to be located along the river or canals tapping the river, whereas settlement in the dry-farming zone may be more dispersed, perhaps depending more on access to and the quality of soils. Likewise, dependence on irrigation systems may dictate a certain level of cooperation between settlements concerning maintenance of canals and allocation of water supplies. These possibilities were already recognized by Akkermans, who suggested that within the Balikh valley, "social units or polities may divide the valley in coherently organized regional units" (Akkermans 1993:7).
96
Another difficulty that has featured prominently in the spatial literature is the problem of using site size to estimate population levels. Related to this is the relation between population levels and functional size, or the number and range of goods and services available at a specific site. Ethnoarchaeological research generally confirms a direct relationship between site size and population, although with a few caveats (e.g., Kramer 1980, 1982; Sumner 1989). Such research in Iran suggests that population density may be significantly different in small and large settlements (Kramer 1980), local environmental conditions must be considered, and the age of a settlement may also be a significant determinant of variations in population density. The correlation between population and range of available functions or functional size has been generally confirmed by research in human geography (Haggett 1965:101-107), but pre-capitalist and special-purpose sites may not be so categorized. Joan Oates (1977:106), for example, notes that for Assyrian capitals, documentary evidence shows that a correlation between size, importance, and the number of "service functions" cannot be made. Similarly, frontier settlement systems may contain special-purpose frontier towns where far more goods and services would be concentrated than would be expected from size correlations (Lewis 1984). Finally, we must address problems with the correlation between survey regions and past cultural, political, or economic regions and territories. Related to this are the determination of an appropriate scale for regional investigation (determining system boundaries) and the fit between the scale of inquiry and the variable scales of past political, economic, and social systems. Due to the fact that the size of an area of investigation is largely determined by logistics and modern political boundaries that have very little to do with past systems, it is unlikely that survey units will contain a complete settlement system (Paynter 1983:253-255). To address these kinds of concerns Crurnley (1979; and Marquardt and Crumley 1987) introduced the concept of effective scale, which refers to any scale at which pattern may be recognized and meaning inferred. Recently, LuAnn Wandsnider (1998) has applied the related concept of "landscape elements" (Forman and Godron 1986) in the identification of regions and the investigation of regional scale processes. Landscape elements are spatial units that contain distinct amounts and distributions of energy, materials, and species. Wandsnider suggests partitioning the landscape into "landscape elements that are sensitive to the prehistoric processes and conditions we hope to understand," and then "examining archaeological materials with respect to these landscape elements" (Wandsnider 1998:97). As for the survey data at hand, Akkermans (1993:7) noted that, even though the Balikh Survey was restricted to a portion of a more extensive drainage network that extends as far north as Urfa in Turkey, the actual Balikh river floodplain stems from the
97
In the analysis of settlement patterns that follows, we are interested in three things: 1) distinctive settlement patterns that may reflect adaptations to local climatic and environmental conditions; 2) the degree of continuity of settlement locations through time; and 3) local conditions and historical trajectories that may have motivated and structured Middle Assyrian settlement in the region. LATE SECOND MILLENNIUM SETTLEMENT PATTERNS IN THE BALIKH VALLEY
Based on recent re-survey and assessment of previous survey records, occupations of Period VII (early second millennium), VIIIA (14th-13th centuries), and VmB (13th1lth centuries), and M (early first millennium) were identified. Descriptions of each site and recorded ceramic assemblages are summarized in Table 2. Early second millennium B. C.settlement patterns: Balikh VII The distribution of early second millennium B.C. settlement in the Balikh valley provides a baseline, of both demographic and hierarchical information, for comparison with succeeding periods. Two sites in the Balikh valley figure prominently in relevant documents from the early second millennium: Tuttul identified with Tell BiCa(BS-1) in the south, at the confluence with the Euphrates, and Harran in the north, in the plain that bears its name, just north of the present Syrian-Turkish border. These two sites are the most prominent mounds in the area and may have been centers of two principle settlement systems centered on the Balikh valley throughout much of the Bronze Age. The present study does not include Harran or any sites in the vicinity because our survey data are limited to the Syrian side of the international border. Other sites mentioned in the Emar itinerary (Goetze 1953, 1964; Hallo 1964) and presumably located within the Balikh valley include: Apqu la Baliha, the source of the Balikh River at 'Ain al-'Ariis or a site in the vicinity; Sahlala, probably Tell Sahlan (BS-247); Zalpah, possibly identified with Tell
98
J. D. LYON
Hammam et-Turkman (BS-175) (van Loon and Meijer 1988:xxv-xxvi); and SerqiAhuna, possibly Tell es-Seman (BS-83/84) (van Liere 1963:Fig.2). I have identified 50 sites with components dated to the early second millennium (Figure 2). Average site size during this period is 2.78 ha and ranges from 25 ha, the largest site, Tell Bica, to less than half a hectare. Together they account for about 143 ha of aggregate site area. Because this period is not the main focus of my research and I have not dealt with the early second millennium ceramic sequence to the extent that I have the late second millennium, these data are preliminary. They will no doubt be subject to subsequent revision and correction. Thiessen polygons constructed around higher order sites reveal eight hypothetical territorial divisions of the valley, with areal constriction increasing from south to north. These territories correspond to seven interaction clusters defined by the gravity model that form roughly around the seven larger settlements of the valley. These include Tell BiCaat the Balikh-Euphrates confluence, Tell Seman in the Seman plain, Site BS-139 on the Wadi 'Ain 'Isa near the Qaramokh-Balikh confluence, Hammam et-Turkrnan, Tell Jittal, Tell Sahlan, and Tell Breri-Tell Abyad at the modern international border (Figure 2). With the exception of Tell Bica, each higher order site includes from 4 to 10 smaller nearest neighbors. The most extensive interaction cluster is that focused on Tell esSeman, but the settlement interaction becomes more intense with a greater degree of boundary overlap north of Tell Hammam et-Turkman. This is a clear indication of the greater density of settlement in the north, i.e. above the 250 mm isohyet, where rainfed agriculture was less risky. There also appears to be a diminution of settlement near the 250 rnm isohyet, south of Tell Zkero (BS-152). Rank-size curves provide further insight into the organization of settlement throughout the valley. The rank-size curve for Period VII sites from the entire Balikh valley is nearly log-linear (rsi=0.05). This is frequently taken as an indication of a wellintegrated settlement system (Johnson 1977, 1980, 1987; Paynter 1978; Falconer and Savage 1995). When we divide the valley into north and south land-use zones according to the distribution of settlement relative to variability in annual rainfall, quite different rank-size curves emerge (Figure 3). The curve for settlement in the north is convex, with a greater number of larger sites than predicted by the rank-size rule (rsi=0.21). This suggests that there may be two or more pooled systems or, more likely, in light of the sites of Harran and Sultantepe looming in the north, that the northern part of the study area is located on the periphery of a larger settlement system and political economy (Johnson 1980, 1981; Paynter 1983). The southern rank-size curve is best described as primo-convex (rsi=-0.3I), which may suggest the existence of separate systems (Falconer and Savage 1995) or a settlement system composed of relatively independent or weakly integrated subsystems (Johnson 1981). In light of the location and size of Tell Bica, it seems likely that Tell Bica may have been more closely integrated with additional settlements and possibly other settlement clusters along the Euphrates.
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
99
Mid-second millennium B. C. settlement patterns: Balikh VZZZA Very few toponyms or textual information of any sort related to the Balikh valley are known from the mid-to-late second millennium until the period of Middle Assyrian ascendancy. This is undoubtedly a result of the fact that archives of the Mitanni royal city, Washshukanni, have not been found ti1 now, and that no substantial textual remains, like those from Munbaqa on the Euphrates or those from Nuzi, have been found in contemporary sites in northern and northeastern Syria. That Harran was still an important city and region is implied by its prominent mention, for example, in the Deeds of Suppiluliuma, and in later treaties and land disputes between the Hittites and Assyrians. For the time being, we must rely entirely on archaeological evidence for the reconstruction of Mitanni and Hanigalbat settlement histories in the Balikh valley. Forty-one sites in the Balikh valley contain evidence of mid-second millennium B.C. occupations (Figure 4). Average site size for this period is 2.37 ha, and the range extends from 11 ha for the largest site, still Tell Bica, to less than 0.5 ha. As in the preceding period, average site size in the south (3.16 ha) is greater than that in the north (2.03 ha). Mitanni period sites account for a total of 94.6 ha of aggregate settled area, and represent a slight decrease from the preceding period. Thiessen polygons constructed for the higher order mid-second millennium B.C. (Period VIIIA) settlements reveal essentially the same eight territorial divisions of the Balikh valley identified above for the early second millennium (Figure 5). The resulting areas are of more or less equal length throughout the southern three-quarters of the valley but become much more confined in the north, toward the Turkish-Syrian border. The gravity model analysis generally confirms these hypothetical divisions of the valley by defining as many as six interaction clusters, with the greatest amount of congestion occurring in the north between Tell Hammam et-Turkman and Tell Sahlan (Figure 5). As was recognized for the preceding period, the gravity model analysis also highlights the division of the valley into north and south zones which correspond to the theoretical dryfarming-irrigation zones. Two connected site clusters contain most of the settlement of the southern portion of the Balikh: one centered on Tell es-Seman (BS-83/84) and one on sites BS-117 and BS139. These site clusters correspond to broad expanses of Balikh valley alluvium. The Tell es-Seman cluster is located in the plain that bears its name, and the BS-117 cluster is situated at the confluence of the Qaramokh and Balikh Rivers, where the valley becomes slightly constricted. The northern Balikh valley is characterized by a higher density of settlement, as it was in preceding periods (Akkermans 1993) and in more recent times (Lewis 1988). Here, average settlement size is about 2 ha. Based on the gravity model analysis three interaction clusters are identified in the north: one centered on Tell Hammam et-Turkman and Tell Jittal, one at Tell Sahlan, and one including Tell Breri and Tell Abyad. Rank-size curves depicting the vertical distribution of Mitanni settlement in the valley are not very revealing. When combined, sites recorded in the valley form a convex curve. Rank-size curves for separate north and south portions of the valley are also convex,
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
although the rank-size curve of settlements in the south is nearly log-linear (Figure 6). The pattern in the south reflects merely the dwindling of the size and possibly the influence and power of Tell Bica, the highest ranked site. It seems likely that settlements in the north were still part of a larger settlement system that included Harran.
the preeminent site in the valley, nonetheless, and was, at some time, home of a be1 pahete.
100
The relative isolation of Tell Bica is further emphasized by the distinct types of 14th century (Spatbronzezeitliche) ceramics found there. Of the recently excavated Spatbronzezeitliche sherds, which Prof. Stromrnenger allowed me to examine in 1995, most are more like types published from Tell Munbaqa and Tell Hadidi (Dornemann 1979, 1981) in the Upper Euphrates than those identified at Tell Harnrnam et-Turkman and throughout much of the Balikh valley. Further investigation of stylistic and technological variability in the Balikh Valley must await full publication of these materials from Tell Bica.
Late second millennium (Middle Assyrian) settlement: Balikh VIlIB For the historical geography of Middle Assyrian settlement in the Balikh valley we have documents from two sites on the western periphery of the Middle Assyrian state, Tell Chuera (Harbe) and Tell Sabi Abyad, and texts from Tell Sheikh Hamad (Durkatlimmu), whose officials received correspondence concerning this area (see Wiggermann this volume).
101
BS-327 located in the south may be situated along a transport route leading northsouth along the Balikh valley, or east towards the Lower Khabur, although its significance is not yet known because of the possibility of another Middle Assyrian settlement cluster at Tell es-Seman. Similarly, the nature of settlement in the northern part of the valley is not completely understood because of the slight possibility of a Middle Assyrian settlement at and in the vicinity of Tell Abyad. Due to the prominence of Harran in contemporary documents we might expect a substantial population there. It may be significant that ~ i d d l eAssyrian settlement was concentrated near the southern limit of feasible dry-farming. This may suggest an attempt at agricultural colonization in an under-utilized or previously abandoned marginal area. It may also reflect the colonization of an area within a safe distance from surviving power centers or without significant land ownership issues. Reference to grain subsidies from Khabur in the Dur-katlimmu texts, as well as the supply of horses and wagons, indicate that Middle Assyrian settlement in the Balikh valley was subsidized to some extent, perhaps only during seasons of low yields. Clues to understanding this include the degree of settlement continuity and the nature and location of settlement abandonment and decline. SETTLEMENT AND DEMOGRAPHIC TRENDS
Six sites with Middle Assyrian (Period VIIIB) components have been identified in the Balikh survey dataset (Figure 7). These are Tell Sahlan (BS-246), Tell Jittal (BS-21I), Tell Hammam et-Turkman (BS-175), Tell Sabi Abyad (BS-189), Khirbet esh-Shenef (BS- l7O), and Tell Abbara (BS-327). These sites range in size from as much as 8 ha (Tell Sahlan) to 0.3 ha (Khirbet esh-Shenef). Six other sites have been included as possible, although minor Period VIIIB occupations due to the presence of one or two possible Middle Assyrian forms. These include Tell es-Seman (BS-83), BS-106, BS-161, BS199, BS-200, and BS-296, all of which would have been quite small settlements. From what we know about the size and compact nature of excavated Middle Asyrian settlements (e.g. at Tell Sabi Abyad and Tell Chuera) it is possible that Middle Assyrian remains may have been destroyed or capped by later and more extensive remains at Tell es-Seman, Tell Abyad, or Tell Bica. That Middle Assyrian ceramics could be found in considerable quantities on smaller sites is suggested by the assemblage present in surface collections at Khirbet esh-Shenef. Excavations at this small site confirmed the attribution to the Middle Assyrian period (Bart1 1990).
Demographic trends are typically represented by number of sites and aggregate settled area for each recognized archaeological phase, with the tacit assumption that site area is generally proportional to human population and range of economic or political functions present. Related to this are patterns of continuity in settlement location and occupational stability, which can be represented by summations of the number of specific locations reoccupied, abandoned, or newly established in consecutive periods.
Most of the sites with definite Middle Assyrian components are concentrated along the Balikh River between Tell Hammam et-Turkman and Tell Sahlan, i.e., north of the present-day 250 mm isohyet. Average nearest neighbor distance for these sites is just over 4 km. This spacing fits well with the image of interdependent Middle Assyrian settlements with Tell Sabi Abyad out-fitting neighboring settlements, possibly itinerant potters and brewers or other specialists who appear to have been attached to the state or family business collectives (Wiggerman this volume). The rank-size curve of Middle Assyrian sites is nearly log-normal, but the relevance of this is uncertain primarily because of the uncertainty of the size of Middle Assyrian Tell Sahlan. It would have been
Survey data from the Balikh valley depict a general decline in population during the second millennium, both in the total number of settlements occupied and in the aggregate settled area (Figure 8). Similar trends have been noted for most regions of greater Mesopotamia and throughout much of Southwest Asia. For the Balikh valley, we are interested in the rate of decline as it may have effected Middle Assyrian expansion and settlement, and in the nature and location of decline. The mid-second millennium is marked by a slight decline from Period VII to VIIIA, and an even larger drop in the Middle Assyrian period (Period VIIIB). Only 12 Period VIW sites have been identified to date, and as many as six of these are minor or questionable occupations.
General settlement trends in the Balikh valley have been discussed by Wilkinson (1996). He notes a general shift from aggregated settlement centered on large sites in the early second millennium, for example Tell Hammam et-Turkman, to a more dispersed pattern of smaller rural sites of Iron Age, Hellenistic, and Roman date. This defines the long-term demographic trends within which Middle Assyrian settlement and agricultural expansion played a pivotal or transitional role.
102
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
Many researchers have noted a high degree of cultural continuity in the Jazira and northern Mesopotamia between the first and second half of the second millennium (e.g., Dornemann 1978, 1979; Mallowan 1947; Wilkinson and Tucker 1995). The high degree of locational stability or continuity from the early to mid-second millennium B.C. settlement confirm these trends for the Balikh valley (Figure 9). Settlement continuity measurements include the total number of sites occupied during two consecutive periods, the number of sites abandoned between those periods, and the number of newly established sites of the successive period. A relatively high degree of continuity is suggested between the early and mid-second millennia (Period VII and VIIIA), with about 70% of all early second millennium sites also occupied in the Mitanni period (Period VIIIA). Only about 10% of all Period VIIIA sites are new establishments. The trend is reversed, however, in the subsequent Middle Assyrian period, with the abandonment of over 80% of Period VIIIA sites (n=34). By the first few centuries of the first millennium the low continuity trend continues, but overall the number of settlements increases, with the establishment of at least 37 new site locations (90%). The number of sites increases to 41 during Period JX, a number equal to that of Period VIIIA. These sites, however, are generally smaller and occupy almost one-third less aggregate area than did the Period VIIIA sites.
in the south, but a slight increase in the north between the end of Period VII (VII+) and Period VIIIA. This lack of continuity suggests that most sites were abandoned between Period VIIIA and VIIIB, that is, before the establishment of Period VIIIB settlements. This hiatus has been recognized at Tell Hammam et-Turkman and Tell Sabi Abyad, for example, but the duration can only be estimated at present. Estimated population levels further support this trend. Because of the potential variability in population density through time and between functionally distinct site types, population levels may be more effectively displayed as possible ranges of population density, in this case from a high of 200 people per hectare and low of 100 people per hectare. Figure 13 presents possible population levels for the north and south portions of the Balikh valley. Here, we see a great deal of overlap for Periods VII and VIIIA, but no overlap in possible population ranges between Periods VIIIA and VIIIB or between Periods VIIlB and IX. Changes in settlement patterns between the early and mid-second millennium include the disappearance of a number of small settlements and the reduction in the size of large higher order settlements (Figures 14 and 15). For example, Tell Bica apparently declined from a total area of about 25 ha to about half of that, according to the distribution of LBA ceramics recorded during the years of excavation there. Likewise, settlement at Tell esSeman and Tell Harnmam et-Turkrnan was also apparently significantly reduced during the first half of the second millennium. This decline is more pronounced in the south than in the north. For example, at Tell Abyad (BS-280), Period VIIIA (Mitanni) ceramics were identified on a low terrace some 70 m from the main mound, suggesting that it may have even increased in extent. Decline in the number and size of settlements during the Middle Assyrian period (13th-12th century) is even more dramatic. Archaeological evidence from excavations in the valley, e.g. Tell (Hammam) Ibn esh-Shehab, Tell Jidle, Tell Hammam et-Turkman, and Tell Sabi Abyad provide evidence of abandonment or hiatus between the two phases marked by conflagration in some places and packing up and moving in others.
If we investigate patterns in the northern and southern portions of the valley separately we may highlight settlement shifts that reflect adaptations to local conditions and distinct landscape units. Figure 10 presents sites and aggregate area for the northern and southern portions of the valley (as distinguished by the gravity model and 250 mm isohyet), and Figure 11 shows continuity measurements for the same. These data indicate that the rate of decline was greater in the south, both in the number of sites and aggregate area. Likewise, settlement continuity between Periods VII and VIIIA was greater in the north than in the south, as were the number of newly established sites. Forty-three percent of sites in the south were abandoned, whereas less than one-quarter were abandoned in the north. Conversely, over 75% of sites in the north have both Period VII and VIIIA components, and just over half (57%) of the sites in the south contain both periods. Almost no newly established sites occur in the south until the first millennium. It is important to point out that the number of sites and aggregate area for Period VII and IX may be somewhat inflated because these numerical data assume that all sites were occupied during the entire period. Due to the limitations of survey data and our knowledge of the ceramic sequence, these periods represent large spans of time that may be further subdivided upon more intensive dating. The use of standardized site area per unit of time has been suggested as one way of presenting more accurate estimates (Schacht 1984; Weiss 1977:358-361), but as Sumner (1990:9) points out, this kind of standardization makes it impossible to estimate absolute population levels or rates of change. Using methods proposed by Sumner (1990) and Dewar (1991), it is possible to incorporate continuity or occupational stability and occupation span in settlement pattern analysis. Sumner (1990:9), for example, assumes that most sites occupied in two successive phases (e.g. Balikh Periods VII and VIIIA) represent a contemporaneous group of occupations inhabited before, during, and after the transition between the two phases. Figure 12, which in effect combines occupation span and continuity measures by adding the theoretical transitional phase to the demographic trend, shows a steady decline
103
This pattern is further documented at Tell Jittal, a medium-sized three-part low mound along the Balikh about 7 krn north of Tell Hammam et-Turkman. Intensive mapping and systematic surface sampling of Tell Jittal was suggested by Tony Wilkinson because of its proximity to Tell Sabi Abyad and because of the potential it holds for the identification of subtle changes in the size and distribution of occupations for periods relevant to the present study. Sampling was conducted by means of 505-m squares positioned at 5 m intervals along six transects placed across the site. All visible artifacts within these sample units were collected, sorted into chaff-tempered and sand-tempered ware groups, and counted. Figure 16 presents sherd density measurements for Period VIIIA (14th cent.), Period VIIIB (13th-12th cent.), and Period IX (first millennium) phases on the main mound. The adjoining low mound to the northeast contained primarily 14th century ceramics. This points to a decline in settled area from more than 5 ha for the Mitanni phase to about 2 ha for the Middle Assyrian phase. The confined Middle Assyrian settlement at Tell Jittal located at the highest point of the mound is similar to patterns seen at Tell Sabi Abyad and Tell Chuera. This suggests that Middle Assyrian settlements were newly built, smaller, and more concentrated than previous settlements.
104
J. D. LYON
CONCLUSIONS: THE CONTEXTS OF MIDDLE ASSYRIAN AGRICULTURAL EXPANSION
The goal of this brief discussion of settlement development in the Balikh valley was to view Middle Assyrian expansion in a regional and broader historical context. Climate, rainfall, and the potential for feasible agriculture surely played an important role in the development of settlement in the region. However, as the Middle Assyrian example highlights, the development of settlement in the Balikh valley of the late second millennium (and thereafter), was largely determined by the priorities and potentialities of non-local (Assyrian) elites. Such non-local actors may introduce unique methods of organizing agriculture and settlement with distinct economic orientations.
In the Balikh valley, Middle Assyrian expansion seems to have followed a period of decline marked by the decline and eventual abandonment of large settlements. Abandonment of most of the valley is suggested by the general demographic trends presented above and further supported by the distinct structure and orientation of the Middle Assyrian settlement system. Abandonment of population centers in the valley appears to have been complete prior to Middle Assyrian colonization, in which case Assyrian agricultural settlement was located in a vacuum of sorts without sedentary power structures. Because of the limitations of survey data, however, we cannot rule out the possibility that indigenous (Hurrian) inhabitants of the Balikh continued to occupy portions of the valley. For example, we know that nomadic peoples, primarily the SutQ, frequented in the area, and may have settled periodically along the river or at the edge of the steppe. At present we can only speculate about the destructive effects of Middle Assyrian aggression in creating the apparent power vacuum. In this case, abandonment of the region could have been hastened by the policy of deportation, the destruction of villages, pestilence, and the general ravages of war. These processes, however, may have followed other "ecologically or socially induced disasters" (Adams 1981:ll). The information presented here suggests that decline began earlier in the southern portion of the valley where subsistence was highly dependent on irrigation agriculture. Decline in the south may have resulted from the abandonment or failure to maintain irrigation systems. This may have involved a shift from settled agriculture to nomadic or semi-nomadic modes of subsistence. At the same time, agriculturalists could have focused on more reliable rainfed areas to the north. Wilkinson's (1998) investigation of a soil profile near Tell es-Seman suggests that the river had silted up or that flow had stopped reaching this far south by the late second or early first millennium B.C. Lewis (1988) notes that "semi-nomadism" in what Rowton (1973) has labeled "the dimorphic zone" was well adapted to the environmental structure of the Balikh valley. He refers to the historical setting of the late 19th- early 20th century AD, when there was no central sedentary authority and power was held by large Beduin groups. The social and political circumstances of settlement development in the Balikh valley may have been quite similar in the late second and early first millennia B .C.
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
105
These processes seem to echo what Liverani describes for the collapse of the regional system of the late second millennium as a whole: the degradation of basic, domestic culture, which was reflected in depopulation of rural areas, and in a marked concentration of settlement in coastal or valley areas better provided with land and water [Postgate 19741. The coincidence of a mass of runaways, and the desertion of non-irrigated lands, produces a growth in nomadic (or semi-nomadic) groups which in some regions supplied the political and ideological pole of attraction for the rural villages who no longer felt any solidarity with the royal palace... (Liverani l987:69) Finally, it is important to remember that the Balikh valley was only one of multiple contexts of Middle Assyrian expansion. Other portions of northern Mesopotamia incorporated into the Assyrian state, each potentially having different agricultural potential, population densities, different tactical uses, proxirnities to other resources, and even increasingly greater or shorter distances from the Assyrian core, may have distinct settlement histories and structures. Future research could contrast source areas (Assyrian heartland) and target areas, but also the different environmental and historical structures of various target areas (Balikh, Khabur Basin, Lower Khabur, North Jazira, Middle Euphrates, Tigris, etc.) incorporated in the Middle Assyrian state in order to achieve a fuller picture of the late second millennium in particular and the interplay of environmental and historical factors in general. Due to recent fieldwork on Middle Assyrian sites in the northern Jazira and the immediacy of the results, this picture is coming together.
106
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
J. D. LYON
Table 2. Second and early first millennium B.C. archaeological sites in the Balikh valley and size estimates (arranged north to south) .
Table 1. Correspondance of Balikh temporal phases with previous studies and sites Present study
Tell Hammam etTurkrnan (van Loon 1988) Harnmarn VlIA
Balikh VII
Tell Sabi Abyad
BS-
Hassan (Pfalzner 1995)
I Hammarn VIIIA
T. Abyad T. Hammam Ibn esh-Shehab
early 2nd
T. Brer~ T. Jidle T. Shrey'an T. &Abed
I
I
Balikh VIIIA Balikh VIIIB
Name
330
Mitanni
1 1450-
0 1 mitannische
mittelassyrische I
Hammam WEE3 Middle Assvrian
1
mittelassyrische It mittelassyrische III
1200
I
1100
1
1
T. Sahlan
T. Eftaim T. Jittal
T. Qardana T. S a b ~Abyad V T. Sab! Abyad 111 T. Sabl Abyad I T. Hammam et-Turkman Kh. esh-Shenef Kh. al-Hajaje T. Breilat
T. ez-Zkero T. Mafraq Slouk
T, as-Sawwan T. Khadriya
T. as-Saman Kh. al8assal
1 T. Bi'a Note. cell shading reflects significance of occupation
Vn
WIIA
i*@i@g&$$e
VIW
D<
J. D. LYON
REFERENCES
Adams, R. McC. Heartland of Cities, Chicago. 198 1 Akkermans, P. M. M. G. Archaologische Gelandebegehung im Balih-Tal, AfO 3 1, 188-90. 1984 Villages in the Steppe: Later Neolithic Settlement and Subsistence in the 1993 Balikh Valley, Northern Syria, Ann Arbor. Akkermans, P. M. M. G., J. Limpens, and R. H. Spoor On the Frontier of Assyria: Excavations at Tell Sabi Abyad, 1991, Akkadica 1992 84-85, 1-52. Akkermans, P. M. M. G. and I. Rossmeisl Excavations at Tell Sabi Abyad, Northern Syria: A Regional Centre on the 1990 Assyrian Frontier, Akkadica 66, 13-60. Akkermans, P. M. M. G. and M. Verhoeven An Image of Complexity: The Burnt Village at Late Neolithic Sabi Abyad, 1995 Syria, American Journal of Archaeology 9911, 5-32. Alden, J. R. A Reconstruction of Toltec Period Political Units in the Valley of Mexico, 1979 169-200 in C. Renfrew and K. L. Cooke, Transformations: Mathematical Approaches to Culture Change, New York. Bartl, K. Friihislamische Besiedlung im Balih-Tal / Nordsyrien, Berlin. 1994 Khirbet esh-Shenef: A Late Bronze Settlement in the Balikh Valley, Northern 1990 Syria, Akkadica 67, 10-32. Beaumont, P. Agricultural and Environmental Changes in the Upper Euphrates Catchment 1996 of Turkey and Syria and their Political and Economic Implications, Applied Geography 1612, 137-157. Boots, B. N. Voronoi (Thiessen) Polygons, Norwich, England. 1986 Brichambaut, G. Perrin de, and C. C. Wallin A Study of Agroclimatology in Semi-Arid and Arid Zones of the Near East, 1963 World Meteorological Organization, Technical Note No. 56, Geneva. Cbrdoba, J. M. Prospecci6n en el valle de rio Balih (Siria): Informe provisional, Aula 1988 Orientalis 6, 149-88. Cmmley, C. L. Toward a Locational Definition of State Systems of Settlement, American 1976 Anthropologist 78, 59-73. 1979 Three Locational Models: An Epistemological Assessment for Anthropology and Archaeology, Advances in Archaeological Method and Theory 2, 14374. Curvers, H. H. The Period VII Pottery, 397-456 in M. N. van Loon ed., Hammam et1988 Turkman I: Report on the University of Amsterdam's 1981-1984 Excavations in Syria, Uitgaven van het Nederlands Historisch-Archaeologisch Instituut te Istanbul 63, Leiden.
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
109
Danti, M. D. 1998 Regional Surveys and Excavations, 85-94 in R. Zettler ed., Subsistence and Settlement in a Marginal Environment: Tell es-Sweyhat, 1989-1995 Preliminary Report, MASCA Research Paper 14, Philadelphia. Dewar, R. E. 199 1 Incorporating Variation in Occupation Span into Settlement-Pattern Analysis, American Antiquity 56, 604-620. Dornemann, R. H. 1979 Tell Hadidi: A Millennium of Bronze Age City Occupation, AASOR 44, 112-41. 198 1 The Late Bronze Age Pottery Tradition at Tell Hadidi, Syria, BASOR 241, 29-47. Eichler, S., V. Haas, D. Steudler, M. Wafler, and D. Warburton 1985 Tall al-Hamidiya 1. Vorbericht 1984, Series Archaeologica 4, Gottingen. Eichler, S., M. Wdler, and D. Warburton Tall al-Hamidiya 2. Vorbericht 1985-1987, Series Archaeologica 6, 1990 Gottingen. Einwag, B. 1993 Vorbericht iiber die Archaologische Geliindebegehung in der Westgazira, Damaszener Mitteilungen 7, 23-43. Faivre, X. 1992 La Cirarnique de Mohammed Diyab, 1990-1991, 55-89 in J.-M. Durand ed., Recherches en Haute M6sopotamie. Tell Mohammed Diyab, Campagnes 1990 et 1991, MCmoires de N.A.B.U. 2, Paris. Falconer, S. E. and S. H. Savage Heartlands and Hinterlands: Alternative Trajectories of Early Urbanization in 1995 Mesopotamia and the Southern Levant, American Antiquity 6011, 37-58. Feinman, G. M. 1997 Thoughts on New Approaches to Combining the Archaeological and Historical Records, Journal of Archaeological Method and Theory 4, 367377. Fish, S. K. and S. A. Kowalewski, eds. 1990 The Archaeology of Regions, Washington, D.C. Forman, R. T. T. and M. Godron 1986 Landscape Ecology, New York. Freydank, H. 1991 Beitrage zur mittelassyrischen Chronologie und Geschichte, Berlin. Goetze, A. 1953 An Old Babylonian Itinerary, JCS 7, 51-74. 1964 Remarks on the Old Babylonian Itinerary, JCS 18, 114-119. Grant, E. 1986 Hill Forts, Central Places and Territories, 13-26 in E. Grant ed., Central Places, Archaeology and History, Sheffield. Grayson, A. K. 1972 Assyrian Royal Inscriptions, Vol. I, Wiesbaden. Hallo, W. W. 1964 The Road to Emar, JCS 18, 57-88. Haggett, P. 1965 Locational Analysis in Human Geography, London.
110 Harrak, A. 1987
Hodder, I. 1979
Home, L. 1993
1994 Jas, R. 1990
J. D. LYON
Assyria and Hanigalbat: A Historical Reconstruction of Bilateral Relations from the Middle of the Fourteenth to the End of the Twelfh Centuries B.C., Texte und Studien zur Orientalistik 4, Hildesheim. Simulating the Growth of Hierarchies, 117-143 in C. Renfrew and K. L. Cooke eds., Transformations: Mathematical Approaches to Culture Change, New York. Occupational and Locational Instability in Arid Land Settlement, 43-53 in C. M. Cameron and S. A. Tomka eds., Abandonment of Settlements and Regions, Cambridge. Village Spaces, Washington, D.C.
Two Middle-Assyrian Lists of Personal Names from Sabi Abyad, Akkadica 67, 33-39. Johnson, G. A. Aspects of Regional Analysis in Archaeology, Annual Review of 1977 Anthropology 6, 479-508. Rank-Size Convexity and Systems Integration: A View from Archaeology, 1980 Economic Geography 6, 324-47. Monitoring Complex System Integration and Boundary Phenomena with 1981 Settlement Size Data, 143-188 in S. E. van der Leeuw ed., Archaeological Approaches to the Study of Complexity, Amsterdam. The Changing Organization of Uruk Administration on the Susiana Plain, 1987 107-140 in F. Hole ed., The Archaeology of Western Iran, Washington, D.C. Kantor, H. J. The Glyptic, 69-85 in C. W. McEwan et al., Soundings at Tell Fakhariyah, 1958 Chicago. Kepecs, S. Introduction to New Approaches to Combining the Archaeological and 1997 Historical Records, Journal of Archaeological Method and Theory 413-4, 193-198. Knapp, - - A. B. Archaeolow 1992 -- and Annales: Time, Space, and Change, 1-21 in A. B. Knapp ed., Archaeology, Annales, and Ethnohistory, Cambridge. Kramer, C . Estimating Prehistoric Populations: An Ethnoarchaeological Approach, 3151980 334 in L'Archiologie de l'lraq: Perspectives et Limites de l'lnterprktation Anthropologique de Documents, Colloques Intemationaux du C.N.R.S. 580, Paris. Village Ethnoarchaeology: Rural Iran in Archaeological Perspective, New 1982 York. Kuhne, H. The Urbanization of the Assyrian Provinces, 55-84 in S. Mazzoni ed., Nuove 1994 Fondazioni nel Vicino Oriente Antico: Realtii e Ideologia, Pisa. The Assyrians on the Middle Euphrates and the Habur, 69-85 in M. Liverani 1995 ed., Neo-Assyrian Geography, Rome.
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
111
Lewis, K. E. 1984 The American Frontier: An Archaeological Study of Settlement Pattern and Process, Orlando. Lewis, N. 1988 The Balikh Valley and its People, 683-696 in M. N. van Loon ed., Hammam et Turkman I: Report on the University of Amsterdam's 1981-1984 Excavations in Syria, Uitgaven van het Nederlands HistorischArchaeologisch Instituut te Istanbul 63, Leiden. Liere, W. J. van 1963 Capitals and citadels of Bronze-Iron Age Syria in their relationship to land and water, AAS 13, 109-22. Liverani, M. 1987 The Collapse of the Near Eastern Regional System at the End of the Bronze Age: The Case of Syria, 66-73 in M. Rowlands, M. T. Larsen, and K. Kristiansen eds., Centre and Periphery in the Ancient World, Cambridge. 1988 The Growth of the Assyrian Empire in the HaburIMiddle Euphrates Area: A New Paradigm., State Archives of Assyria Bulletin 212, 81-98. Loon, M. N. van ed. Hammam et-Turkman I: Report on the University of Amsterdam's 1981-1984 1988 Excavations in Syria. Uitgaven van het Nederlands HistorischArchaeologisch Instituut te Istanbul 63. Leiden: Nederlands Instituut voor het Nabije Oosten. Loon, M. N. van and D. J. W. Meijer 1988 Introduction, xxv-xxvi in M. N. van Loon ed., Hammam et-Turkman I: Report on the University of Amsterdam's 1981-1984 Excavations in Syria, Uitgaven van het Nederlands Historisch-Archaeologisch Instituut te Istanbul 63, Leiden. Lyonnet, B. 1990 Prospection archiologique de Tell Mohammed Diyab, 71-115 in J.-M. Durand ed., Tell Mohammed Diyab, Campagnes 1987 et 1988, Cahiers de N.A.B.U 1, Paris. Mallowan, M. E. L. 1946 Excavations in the Balikh Valley, 1938, Iraq 8, 111-159. 1947 Excavations at Brak and Chagar Bazar, Iraq 9, 1-266. Marquardt, W. H. and C. L. Crumley 1987 Theoretical Issues in the Analysis of Spatial Patterning, 1-18 in C. L. Crumley and W. H. Marquardt eds., Regional Dynamics, Burgundian Landscapes in Historical Perspective, New York. Matthews, D. M. 1990 Principles of Composition in Near Eastern Glyptic of the Later Second Millennium B. C., Gottingen. McEwan, C. W. 1958 Soundings at Tell Fakhariyah, Chicago. Nissen, H. J. 1967 Aus dem Geschaftsleben assyrischer Kaufleute im 14. Jhdt. v. Chr., 111-120 in Heidelberger Studien zum Alten Orient: Festschrift Adam Falkenstein, Wiesbaden. Oates, D., J. Oates, and H. McDonald 1997 Excavations at Tell Brak. Vol. 1: The Mitanni and Old Babylonian Periods, London.
112 Oates, J. 1977
J. D. LYON
Archaeology and Geography in Mesopotamia, in J. L. Bintliff ed., Mycenaean Geography, Cambridge.
Parker, B. J. 1997 The Northern Frontier of Assyria: An Archaeological Perspective, 217-244 in S. Parpola and R. M. Whiting eds., Assyria 1995, Helsinki. Paynter, R. W. Expanding the Scope of Settlement Analysis, 233-275 in J. A. Moore and A. 1983 S. Keene eds., Archaeological Hammers and Theories, New York. Pearson, C. E. Rank-Size Distributions and the Analysis of Prehistoric Settlement Systems, 1980 Journal of Anthropological Research 3614, 453-462. Pfdzner, P. Mitannische und Mittelassyrische Keramik: Eine Chronologische, 1995 Funktionale und Produktionsokonomische Analyse. Berlin. Plog, S. Measurements of Prehistoric Interaction between Communities, 255-272 in 1976 K. V. Flannery ed., The Early Mesoamerican Village,'New York. Postgate, C., D. Oates, and J. Oates The Excavations at Tell al-Rimah: The Pottery, Iraq Archaeological Reports 1997 4, Wiltshire, England. Postgate, J. N. Some Remarks on Conditions in the Assyrian Countryside, JESHO 1713, 1974 225-243. 1992 The Land of Assur and the Yoke of Assur, World Archaeology 2313, 24763. Rod, M. Tell Mohammad 'Arab, Iraq 46, 141-56. 1984 Rowton, M. B. Autonomy and Nomadism in Western Asia, Orientalia 42, 247-258. 1973 Schacht, R. The Contemporaneity Problem, American Antiquity 49, 678-695. 1984 Smit, F. The Period VIII Pottery, 457-498 in M. N. van Loon ed., Hammam et1988 Turkman I: Report on the University of Amsterdam's 1981-1984 Excavations in Syria, Uitgaven van het Nederlands Historisch-Archaeologisch Instituut te Istanbul 63, Leiden. Smith, M. Braudel's Temporal Rhythms and Chronological Theory in Archaeology, 1992 23-34 in A. B. Knapp ed., Archaeology, Annales, and Ethnohistory, Cambridge. Speiser, E. A. 1932-33 The pottery of Tell Billa, Museum Journal 23, 249-83. Surnner, W. M. Population and Settlement Area: An Example from Iran, American 1989 Anthropologist 9 1, 63 1-640. An Archaeological Estimate of Population Trends since 6000 BC in the Kur 1990 River Basin, Fars Province, Iran, 3-16, in M. Taddei ed., Southwest Asian Archaeology 1987, Rome.
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
113
Wandsnider, L. 1998 Regional Scale Processes and Archaeological Landscape Units, 87-102 in A. F. Ramenofsky and A. Steffen eds., Unit Issues in Archaeology: Measuring Time, Space, and Material, Salt Lake City. Weiss, H. Periodization, Population and Early State Formation in Khuzistan, in L. D. 1977 Levine and T. Cuyler Young, Jr. eds., Mountains and Lowlands: Essays in the Archaeology of Greater Mesopotamia, Bibliotheca Mesopotamia 7. Wilkinson, T. J. 1994 The Structure and Dynamics of Dry-Farming States in Upper Mesopotamia, Current Anthropology 3515, 483-520. Sabi Abyad: The Geoarchaeology of a Complex Landscape, 1-24 in P. M. 1996 M. G. Akkermans ed., Tell Sabi Abyad. The Late Neolithic Settlement I, Leiden. Water and Human Settlement in the Balikh Valley, Syria: Investigations from 1998 1992-1995, Journal of Field Archaeology 25, 63-87. Wilkinson, T. J. and D. J. Tucker Settlement Development in the North Jazira, Iraq: A Study of the 1995 Archaeological Landscape. Iraq Archaeological Reports 3, Warminster, Wiltshire, England. Zeist, W. van Some Notes on Second Millennium B.C. Plant Cultivation in the Syrian 1994 Jazira, 541-553 in H. Gasche et al. eds., Cinquante-deux reflexions sur le Proche-Orient ancien: offertes en hommage ~2LJon De Meyer. Mesopotamian History and Environment, Occasional Publications 11, Louvain. *I wish to thank Dr. Remko Jas for the invitation to contribute and for his generous hospitality. This paper stems from dissertation research supported by the Ryerson and Helen Rich Funds of The University of Chicago.
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
Balikh Valley Perlod Vll Settlements (early 2nd mlllennltlm B C )
Figure 2. Map of Balikh VII (early 2nd millennium) sites.
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
J. D. LYON
10 Rank
, , No-
South
Figure 3. Rank-size curves for northern and southern Balikh VII (early 2nd millennium) sites.
Balikh Valley Perlod VlllA Settlements (14th century M~tannf)
Figure 4. Map of Balikh VIIIA (14th century) sites.
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
10 Rank
Figure 6 . Rank-size curves for northern and southern Balikh VIIIA ( 1 4 ~ century) settlement.
Figure 5. Theoretical subdivisions and interaction clusters of Balikh VIIIA settlement.
121
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
J. D. LYON
Tell Sabl Abyad (BS-189) Hammam et-Turkman (6s-175) hlrbet esh-Shenef (6s-170)
I -
Sites
Agg. area (ha.)
Figure 8. Sites and aggregate area from the late 2nd through 1st millennia B.C.
Balikh Valley ir
Period VlllA Settlements (14th centuw-Mltannl) Perlod VlllB Settlements (13 12th century Mlddle Assyrian)
0
Pose~bleVlllB Settlements
Period 8A-8B
Figure 7. Map of Middle Assyrian sites. Continuity
Abandon
Period 8
0
New
Figure 9. Settlement continuity measurements for 2nd through 1st millennia B.C.
122
123
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
J. D. LYON
-
Balikh Period
-
North sites
sites
North aggregate area (ha)
+ North-aggregate area
0
South-sks
4 South-aggregate area
South aggregate area (ha)
Figure 10. Sites and aggregate area for the northern and southern portions of the Balikh Valley.
Figure 12. Sites and aggregate area with continuity variable.
Continue
Aband
Balikh Period
Figure 13. Hi-low plot of estimated population for the Balikh Valley, 2nd through 1st millennia B.C. Penod 7-8A
Period 8A-8B
Penod 8B-9
Figure 11. Continuity measurements for the northern and southern portions of the Balikh Valley. Fig. 1.
I
124
J. D. LYON
MIDDLE ASSYRIAN EXPANSION AND SETTLEMENT
125
M o d IX
Figure 1 4 Site size frequency for Balikh sites, 2nd through 1st millennia B.C
Figure 15. Schematic maps of Balikh settlement from the early 2nd through the 1st millennia B.C
126
J. D. LYON 8.1,k"VIIII
,M,"."l,
II
Figure 16. Sherd frequency plot of Mittani (top), Middle Assyrian (middle), and Late Assyrian (bottom) diagnostics at Tell Jittal (BS-211).
TEXTUAL EVIDENCE: FROM DETAILS TO GENERALIZATIONS
PIHANS LXXXVIII, 2000
Irrigation Agriculture in Mari* Bertrand Lafont (CNRS, Nanterre)
At first sight, one could wonder why Mari is included in the present symposium: its location is not in Northern Mesopotamia, nor in a zone of rainfed cultivation. Nevertheless, it is easy to find several reasons legitimating an interest in it: the written documentation from Mari is full of exceptional information about agricultural techniques and practices used outside of Babylonia during the first third of the second millennium B.C., and, moreover, shows many connections linking the Middle Euphrates valley (Mari) to areas in Northern Mesopotamia (Upper Euphrates, Balih and fjabur valleys, Upper Tigris). This explains, I suppose, why I have been asked to present an overview of what we know about Mari agriculture. On such a theme, the aim of this paper is to propose a short synthesis of recent itrchaeological and philological research that has renewed the topic to a considerable extent. These works are mainly the result of two series of surveys that have been carried out during the past fifteen years. These investigations do not necessarily tally, since the initial questions asked by those who started them were not always the same: 1) field surveys, joint enterprises of archaeologists and geographers: the entire Middle Euphrates valley between Deir ez-Zor and Abu Kemal has been covered by a large survey and a regional study that aimed at specifying the relation between Mari and the river.' Another ambition of this kind of research was to gather information on spatial organization and the possible existence of canal networks and hydraulic installations in ancient times. organized * ~ r a t e f u acknowledgment l is made to L. McLarnan (Chicago) and R. M. Jas (Amsterdam) for their comments on this revised, English version of the paper I read in Leiden. 1 P. Sanlaville, "L'espace gkographique de Mari," MARI 4 (1985) 15-26; B. Geyer, "GComorphologie et occupation du sol de la moyenne vallte de 1'Euphrate dans la rCgion de Mari," MARI 4 (1985) 27-39; B. Geyer, "AmCnagements hydrauliques et terroir agricole dam la moyenne vallCe de l'Euphrate," BAH 136 (1990) 63-85; B. Geyer and J.-Y. Monchambert, "Prospection de la moyenne vallCe de 1'Euphrate: rapport prCliminaire (1982-1985)," MARI 5 (1987) 293-344; J.-C. Margueron, "8tat prksent des recherches sur l'urbanisme de Mari, I," MARI 5 (1987) 483-498; J.-C. Margueron, "Espace agricole et amknagement regional B Mari au dkbut du IIIe milltnaire," BSA 4 (1988) 49-60; J.-C. Margueron, "L'amCnagement de la rCgion de Mari: quelques consid6ations historiques BAH 136 (1990) 171-191; J.-C. Margueron, "Problkmes de transports au dCbut de l'iige du Bronze," 119-126 in M. Lebeau and Ph. Talon eds., Reflets des deuxfleuves, Volume de Me'langes offert d Andre' Finet, Louvain 1989; J.-C. Margueron, "Mari, 1'Euphrate et le Khabur au milieu du IIIe millCnaire," BCSMS 21 (1991) 79-100; J.-C. Margueron, "Mari: une clC des relations internationales entre Syrie et MCsopotamie au IIIe millCnaire," 9-26 in E. Frezouls and A. Jacquemin eds., Les Relations internationales, Strasbourg and Paris, 1995; J.-C. Margueron, "AmCnagement du territoire et organisation de l'espace en Syrie du nord h 19Agedu Bronze: lirnites et possibilitCs d'une recherche," 167-178 in M. Fortin and 0 . Aurenche, Espace naturel, espace habite' en Syrie du Nord (10e-2e mille'naires av. J.-C.), BCSMS 33 = TMO 28, Qutbec and Lyon, 1998.
130
131
B. LAFONT
IRRIGATION AGRICULTURE IN MARI
2) philological studies, made possible by a surge of new text publications and the reassessment of earlier text editions; the entire technical vocabulary dealing with agriculture and irrigation has been reevaluated in detail and put back into a historical and environmental context that is now much better known.2 These different inquiries have brought about several debates, in particular about the role of the Euphrates as a "large thoroughfare for commercial penetration," the "key" position Mari occupies in the traffic between the northwest and the southeast, the existence of a significant "canal system required by an intense trade by river" during the third and the beginning of the second millennia B.C., and about the date of the canal traces which were found.3 At the same time, several other works on agriculture and irrigation in Mesopotamia have allowed significant breakthroughs in the comprehension of the studied phenomena. The two 1987 conferences in Leiden and Damascus should be mentioned first.4 All the following data directly stem from these works to which the reader is referred.
1) the triangular area between DGr-Yakdun-L?m (on the right bank of the Euphrates, close to modern Deir ez-Zor),5 Saggagturn (on the left bank of the gabur),6 and modern Meyadin (downstream on the Euphrates); this important area at the confluence of the Habur and the Euphrates falls under the authority of the governor of Saggaraturn; 2) the area of Terqa, southeast of the preceding area; it should be noticed that only the right bank of the river seems to fall under the authority of the governor of Terqa; 3) tde right bank area (ahariitum) of the Mari alveolus, from Doura Europos down to Abu Kemal (Der); this area is particularly marked by the outflow of significant western wadis (especially the wadi es Souab); 4) the left bank area (aqdamiitum) of the river, located between tell Darnaj (facing AshardTerqa) and the Baghouz cliff (facing Abu KemalJDer); it is worth noting that these last two areas fall under the sole authority of the governor of Mari.7
1.
NATURAL ENVIRONMENT AND CLIMATIC CONDITIONS
Between Deir ez-Zor and Abu Kemal, the Euphrates meanders at the bottom of a U-shaped valley framed by a calcareous plateau. There is a difference of about 40 meters in height between the plateau and the bottom of the valley (fig. I). Today, the river runs down on the so-called "historical terrace," recently formed and located approximately 2 meters downwards of the low terrace (from Holocene times) on which most of the archaeological sites are located. The width of the valley varies from 0 to approximately 15 krn on the right bank, with rather marked escarpments between the various terraces and the plateau. It is much narrower and softer inclined on the left bank of the river. Tightenings, or on the contrary widenings of the valley constitute as many natural "alveoli" (basins). This is the reason we often speak, for example, of the "Mari alveolus," a unit extending between the cliff of Doura Europos in the north and the Baghouz cliff (facing Abu Kemal) in the south (fig. 2). The alveolus structure is more pronounced on the right bank of the river than on the left bank. Within this environmental context, the Mari texts show that in the heart of the kingdom four principal natural as well as administrative units can be distinguished (fig. 2). From upstream to downstream:
In the entire area, climatic conditions constitute a major constraint for any agricultural activity. With less than 150 mm of rainfall (far removed from the famous isohyet of 250 rnrn so often evoked to appreciate the agricultural possibilities within the Fertile Crescent) and less than 40 days of rain per year, rainfed agriculture is impossible. The maximum of rains occurs in February and March. Rains in autumn and spring, most significant for agriculture, happen to be also the most irregular. The length of the dry season, the strong insolation and high temperatures during the summer (over 35•‹Cfor more than 100 days a year in Abu Kemal) are other important obstacles. To these difficulties a strong salinity of the soils (partly compensated for by the annual washing of the flood) and the harmful role of the desiccant winds (Hamsin) in spring and summer must be added.8 As pointed out by B. Geyer, at less than one kilometer from the river it is possible to water the soils directly by drawing water from the river. A few slight devices are sufficient, but this system does not enable the cultivation of large areas.9 On the contrary, feeding the soils of the low terrace that are not in direct contact with the river requires the setting up of a large irrigation system. Such a system, however, can only be conceived by the use of running water from the Euphrates, the Habur and (possibly) the wadis, by means of canals. In fact, these canals are now clearly documented, perhaps in archaeology, and surely in the texts (see below).
5
J.-R. Kupper, "L'irrigation B Mari," BSA 4 (1988) 93-103; J.-M. Durand, "Problkmes d'eau et d'irrigation au royaume de Mari: l'apport des textes anciens," BAH 136 (1990) 101-142, supplemented by LAPO 17 (1998) 513-676. See, for example, J.-M. Durand, LAPO 17 (1998) 573-578, and B. Lyonnet, Subartu 411 (1997) 179-194, contra J.-C. Margueron (in several of the references given above, n. 1). Damascus colloquium published in BAH 136 (1990),and Leiden colloquium published in BSA 4 and 5 (1988). See also M. Stol, "Kanalisation" in RLA 5 (1976-80) 355-365; M. Civil, The Farmer's Instructions, Aula Orientalis Supplementa 5, Barcelona 1994; F. Joannks, ARM 23, 105-132; M. Birot, ARM 27, especially 9-13; B. Lafont, "Une nuit dramatique B Mari," Flor. Mar. 1 (1992) 93-105, and "Techniques arboricoles B l'tpoque amorrite. Transport et acclimatation de figuiers 21 Mari," Flor. Mar. 3 (1997) 263-268. See now also H. Klengel and J. Renger eds., Landwirtschaft im alten Orient, CRRA 41 (BBVO 18), Berlin 1999.
Tell Mohasan, about 100 km downstream from Deir ez-Zor, has been proposed by B. Geyer and J.Y. Monchambert as the location of DBr-Yabdun-Lim (MARI 5 [I9871 325). J.-M. Durand, on the contrary, prefers to locate DBr-Yahdun-Lim closer to Deir es-Zor (BAH 136 [I9901 109 and 123). ti J.-M. Durand, BAH 136 (1990) 123 n. 84. 7 Because the crossing of a major river like the Euphrates must have been difficult, the distinction between the right bank (aharhtum) and the left bank (aqdamhtum)of the river was probably an important feature. For references to crossing the river, curiously enough not that frequent in the Mari texts, see, for example, ARM 3 5 = LAPO 17 726. It is worth noting, however, that these winds are rarely mentioned in our texts, even though Prof. St01 kindly reminds me of the reference to the 5hr haribtim (warki ebcrim) in B. Lion, Flor. Mar. 1 (1992) 110, n. 13. B. Geyer, BAH 136 (1990) 68. The techniques and devices for drawing water in Mari were most likely similar to the ones used in Syria until the beginning of the present century (gharraf, doulab, nasba, norias, etc.: see for example A. Hamidt, BAH 136 [I9901 25-31).
132
B. LAFONT
At this point though, another difficulty occurs, related to the significant and sudden rises of the water level. The flood period starts with the autumn rains and culminates in spring. After this, a decrease sets in until the low water level of the beginning of autumn is reached. At the most intense time of the flood, meanders can move very quickly and destroy everything in the low terrace. The western wadis (nahlum) that emerge from the plateau can also threaten the valley; the swelling of their waters starts earlier (in FebruaryIMarch) and can last for several weeks. Ultimately, the Euphrates and Habur waters are essential everywhere in the area. In the Mari texts, there is no mention of awaiting the rain, or of complaints about a lack of rain: for agriculture, only the river is important. Still, the Mari kingdom should not be considered as merely three or four isolated and small alveoli irrigated by the river and encircled by the desert. The so-called "desert" areas west and east of Mari, in fact, were not "true" deserts, nor insuperable obstacles. They were crossed, in particular, by many roads, and were occupied by Bedouins. In addition, water and cultivated areas were perhaps more abundant in this part of the country than it is seen and thought off today. In the Mari kingdom, therefore, the "desert" and the inhabited areas of the four "alveoli" do not neatly fall into two opposite categories. From an economic point of view, the former is characterized by its livestock, and the latter by its crops. Even though we do not have as much information on stock raising as on agriculture, the Mari texts suggest that cultivation and animal husbandry, although marked differently in both landscape and society, were very well integrated. 2.
THE sons AND THE CONTROL OF WATER
The governor letters from Mari (ARM 6), Terqa (ARM 3), Saggarltum (ARM 14), and QaQunln (ARM 27), as well as the administrative texts from the Mari palace (especially ARM 23) contain much information on the way agriculture was organized. It should be noted that these texts use a technical vocabulary of mostly West Semitic stock for matters of irrigation and cultivation. It is not only difficult to establish the exact meaning of the terms, but they often provide only sketchy information, without a general context. Nevertheless, they allow us to distinguish three areas within the Mari kingdom (fig. 1 and 2): 1) the zone of the river's major bed, and the low terraces: this is the sedentary zone in which cities, canals and agricultural fields are located; this area is called rib nririm "banks of the river"; 2) the zone of the middle and high terraces: this is the sphere of the grazing lands (nawGm); 3) the zone of the plateau: from here, the steppe and the arid lands stretch as far as the desert border which is called qasGm; of course, this area cannot be irrigated or cultivated. One thus sees that the different zones are arranged according to the river (rib gabur, rib Purattim),lo a situation which differentiates this area from the more septentrional dry-
lo
See J.-M. Durand, BAH 136 [I9901 112 n. 37.
IRRIGATION AGRICULTURE IN MARI
133
farming zones which use different denominations, generally following the mountainous axis (Maras, Palbum, Saggar, etc.). Agricultural exploitation is only possible in the first of these three zones (rib nririm, or mdtum). Within such a restricted area, one wonders if the Mari, Terqa and Saggarfitum alveoli were densely populated? According to estimates based on ARM 3 3 = LAPO 17 798, a city like Terqa, for example, had probably less than 2.000 inhabitants. It has also been proposed that Mari itself was rather low on inhabitants. Most of the population, therefore, lived in villages. l On the right bank, the alluvial terrace is rather broad between the cliff of Doura Europos (= Hum-Muluk) and modern Abu Kemal (= Der). On this bank, the zone of MiSlln (= Tell Ramadi) was perhaps the real "granary" of the whole area. Opposite to MiSlh, the large borough of Suprum was located. Since the alveolus is much narrower on this left bank side, it was less suitable for large irrigation works. Nevertheless, the population density seems to have been high in this area, which was perhaps easier to cultivate with the help of small installations drawing water directly from the river. The main cities were provided with water by special canals, as the one the remains of which were discovered in Mari: with a length of 4 km, its watering point in the Euphrates was located 2.5 krn upstream from the city (perhaps towards Appfin).l2 From an agricultural point of view, two further zones within the rib nririm area can be distinguished: 1) the immediate surroundings of the river: the lowest part of the valley is indicated by the term hamqum. Very close to the river, and requiring no large irrigation systems, the fields located more downwards (5uprilGm) are distinct from the lands which must be furrowed and fed by small irrigation channels (m2re5tum). Another category of lowlands ~ ~ natural depressions, always is formed by the grassy lands for grazing ( u s a l l ~ r n ) .Some on the low terrace, retain water of the river's old meanders, and can be cultivated without artificial irrigation: such a zone is called a balitum (fig. 1). Regularly fed by the flood, it can be cultivated in the same manner as the lands close to the river. Another term to This observation is important from a geographical-historical point of view, as adopted, for example, by M. Liverani in his studies of the ancient agricultural landscape. Compare the following statement by M. Liverani, which is just as valid for the Mari kingdom as for the country of Akkad and the Diyala valley: "urban centers always remained inserted within a network of numerous and vital villages [with a] rurality rate continuously remaining high." (M. Liverani, "The Role of the Village in shaping the Ancient Near Eastern Landscape," 45 in L. Milano et al., eds., Landscapes, Territories, Frontiers and Horizons in the Ancient Near East, CRRA 44, Padua 1999). On this topic, and on the opposition between an "Akkadian" landscape with "square" fields scattered with villages, and a "Sumerian" landscape with "long" fields highlighted by large cities, see also the following articles by M. Liverani: "Reconstructing the Rural Landscape of the Ancient Near East," JESHO 39 (1996) 1-49, and "Lower Mesopotamian Fields: South vs. North," 219-227 in B. Pongratz-Leisten et al. eds., Ana tad2 Labniini lu allik (AOAT 247 [I9971 Fs. W. Rollig). For the question of the ancient landscape and the Middle-Euphrates valley, compare also G. Buccellati, "The Rural Landscape of the Ancient Zor: the Terqa evidence," BAH 136 (1990) 155-169. B. Geyer and J.-Y. Monchambert, MARI 5 (1987) 494-498. See also B. Lafont, Flor. Mar. 1 (1992) 98-105, and J.-M. Durand, LAPO 17 (1998) 625. l 3 For usallurn, cf. K. R. Veenhof, 373-374 in M. A. Beek et al. eds., Symbolae Biblicae et Mesopotamicae Francisco Mario Theodoro de Liagre Bohl dedicatae, Leiden 1973. For the three category distinction (tupdlum I m2reitum I usallum) in Mari, see especially ARM 23 467.
134
B. LAFONT
describe a natural depression retaining water is miggum (perhaps a type of drainage basin").14 2) on the remainder of the low terrace, moving away from the river, land can be cultivated only by irrigation. In this area, the most important irrigation device is the rdkibum. It is worth emphasizing that an important step in reconstructing the irrigation techniques used in Mari was made with the correct explanation of the meaning of this word15 and the simultaneous discovery of remains of irrigation canals in the Deir ez-Zor and Mari alveoli.16 What, exactly now, is a rdkibum? In the Mari texts, it seems to be a generic term for a large irrigation canal crossing the valley. This canal owes the meaning of "rider" to the fact that it is not dug, but built in overhang between two levees on the valley. A water supply point, rather high upstream in the river, allows a water distribution by gravity and renders elevatory devices needless. The rdkibum also makes it possible to recover and control water from the western wadis flowing out from the plateau, thus preventing the damage they could do. At the end of its course, the rdkibum returns to the river. This canal can receive different proper names inspired by the regions it goes across. Lands irrigated by a rdkibum are known as maSqitum (from SQ', "to drink"), and the surrounding lands which, after irrigation, are furrowed are known as me^reStum. According to their location, various appellations are used to describe the different cultivated lands stretched out in the low part of the valley, not directly at the river: salhum (from SLH, "to sprinkle"): land immediately around the cities where the abundance of water allows direct watering; this zone undoubtedly is made up mainly by gardens and small parcels devoted to horticulture; the salhum possibly corresponds to the areas of cultivation surrounding archaeological sites as described by T. J. Wilkinson;l7 ugdrum: country land, irrigated by canal networks (see also above under me^reStum); daluwdtum (from DL', "to draw"): lands too high to be reached by irrigation canals, but which can be cultivated by drawing water in areas with wells. It must be noticed that arable land can also be distinguished according to natural qualities: damqum ("good"), lemnum ("bad"), or eqel idrdni "salty field," in reference to the saline groundwater (idrdnum). l Ultimately, two natural units are radically opposed within the Mari kingdom: the ah ndrim, embracing the entire low, cultivable terrace, and the nawam, representing the uncultivable steppe that stretches as far as the border of the desert (qap2m).
I I
IRRIGATION AGRICULTURE IN MARI
3. REQUIREMENTS FOR IRRIGATION: CANALS AND PROCEDURES FOR THE CONTROL OF WATER
Within the Mari kingdom during the Amorite period (18th century B.C.), it appears that three large rdkibum-canals irrigating the lands were interdependent and constituted a real unit. This situation, besides illustrating the coherence in the heart of the Mari kingdom, permits us to speak about "canal networks" (ARM 14 13 = LAPO 17 804) as well: 1) the IEm-Yahdun-Lim canal: it is now quite certain this is the name of a canal going from Deir ez-Zor (= perhaps DQr-Yahdun-Lim)lgdown to Ashara (= Terqa). The main question is whether this I3m-Y&dun-Lim canal is identical to the canal nowadays known as Nahr Said, the remains of which can be observed along a distance of around 30 kilometers south of Deir ez-Zor. This long rdkibum-canal covers the two alveoli of DQrYabdun-Lim and Terqa (fig. 2); 2) the Mari canal, which goes through the Mari alveolus on the right bank of the river; its watering point in the Euphrates is probably located in the proximity of the cliff of Doura Europos, and it irrigates the entire low terrace down to Abu Kemal (= D&). Known mainly as "Mari canal," this rdkibum possibly assumes different names inspired by the area it goes through, such as, for example, "Hubur (= d l ~ I . canal," ~ ~ ~a )name used perhaps in reference to one of the wadis flowing down from the western plateau that is associated with this canal.20 The possible remains of this rdkibum have been discovered along a distance of around 15 kilometers (see the maps in MARI 5 119871 308, Flor. Mar. 1 [I9921 105, and our fig. 2); 3) the Habur canal; supplied with water from the Habur river on the left bank: its watering point is probably located at the level of Saggartit~m,~~ and it joins the Euphrates perhaps near Suprum, or just before the Baghouz cliff (facing Abu Kemal = Der). Remains of such a rdkibum-canal, however, have not been discovered on the left bank alveolus, as it seems that this canal needs to be distinguished from the Nahr Dawrin canal, the remains of which can be observed today along the cliff of the Eastern plateau for over a 100 km. Because of its position and its typical features, archaeologists think that the Nahr Dawrin is used for navigation only, allowing boats to avoid the Euphrates' meanders.22 The problem is that navigation canals are not mentioned in the Mari texts, For the location of DGr-Y&dun-Lim, see above n. 5. J.-M. Durand, BAH 136 (1990) 114 and 129, and B. Lafont, Flor. Mar. 1 (1992) 99-100. In fact, three main wadis flow out from the west of the Mari alveolus: 1) the wadi es Souab (= gaqat, or wadi of MiSl2n) in the northwest of the alveolus (the most important one); 2) the wadi Bir el Ahmar (= Hubur) in the middle of the alveolus; 3) the wadi er Radqa (= Balih, or wadi of Dgr) in the southwest of the alveolus. As pointed out by J.-M. Durand, the names of the two main tributaries of the Euphrates in Syria (Balib and Habur) seem to recur as names of wadis in the Mari alveolus (cf. the well-known phenomenon of "mirror toponymy" in Mari). Therefore, we must not forget that when the Habur is mentioned in our texts, the waterway in question can either be the tributary river in the Syrian Jazira, the rdkiburn-canal on the left bank of the Euphrates, or the wadi in the Mari alveolus on the right bank of the Euphrates. Only as a convention, and in order to distinguish homonymous river and wadi, we call the first one gabur and the second one Hubur. J.-M. Durand, LAPO 17 (1998) 631. 22 J.-C. Margueron, BSA 4 (1988) 52-54, and BAH 136 (1990) 171-183; B. Geyer, BAH 136 (1990) 75-76. It is true that the Habur and the swelling of its waters are not as violent as the Euphrates, allowing a better regulation of the water flow into a derivative canal. Compare, nevertheless, the letter 20
l4
J.-M. Durand, BAH 136 (1990) 121-122, and LAPO 17 (1998) 515. J.-M. Durand, BAH 136 (1990) 126-127. B. Geyer and J.-Y. Monchambert, MARI 5 (1987) 293-344. l 7 T. J. Wilkinson, this volume. For the word salhurn, cf. also the important observations of J. Eidem, RA 85 (1991) 133-134. As pointed out by J. Eidem, it is worth noting that words like salhurn and nawLim are not only used for a kind of soil or area, but also for the people and animals living there (cf. also J.-M. Durand, LAPO 17 [I9981 475). Cf. J.-M. Durand, BAH 136 (1990) 139. The unpublished letter A.392 quoted there is the only reference made in the Mari texts to both the potential and the real risk of soil salinization. l5
135
136
137
B. LAFONT
IRRIGATION AGRICULTURE IN MARI
which, besides, show that during the Amorite period river traffic on its way to Mari never avoids Terqa. The date of construction and use of the Nahr Dawrin canal (as of the other canals the remains of which have been discovered) remains, anyhow, a problem: third millennium according to J.-C. Margueron, or Islamic times according to J.-M. Durand? We really do not know.
charge of a highly vulnerable locality: the rdkibum-canal's watering point, at the junction of the Terqa and Mari alveoli (facing the Doura Europos cliff). Most of the (new) explanations of the technical vocabulary concerning irrigation, found mainly in these letters, are the direct results of J.-M. Durand's research presented in 1987 during the Damascus symposium.26They are briefly summed up here: Annual works are necessary at the "mouth" (KA/pfim) of the canals to remove clay deposits and reed wisps blocking the input (ARM 2 83 = LAPO 17 786, ARM 3 5 = LAPO 17 799, ARM 14 14 = LAPO 17 802). In this context, the vocabulary we mainly encounter consists of ki~rum,"clay plug," s'iJdqum, "contracting of the canal bed," ba@fum, "to (re)dig," etc. (ARM 3 4 and 5 = LAPO 17 787 and 799, etc.). At the mouth (KA/piim), or watering point, of the canal, the fundamental structure is the muballittum (not to be confused with the balitum mentioned above [formed on BL']), a technical device allowing the diversion of water from the river and control of the canal's water level (ARM 6 4 and ARM 14 13 = LAPO 17 803 and 804). This structure consists of a fence made of poles (tarqullum), reinforced by bundles of sticks and reeds, and makes it possible to divert part of the course of the river towards the canal. Other particular devices were used to control the flow, such as the complex system of sluice gates called errztum. This e r r h m , in the bank of the canal, enables the evacuation of excess water. The texts show that high moisture prevails around such installations, with water oozing permanently. Several technical terms are used to describe the different components of an e r r h m , in analogy with facial parts: thus, the errgtum has a "nose" ( a p p ~ m made ) ~ ~ of stone (abnum), "cheekbones" (usukkum), and "eyebrows" (s'firum) made of wooden cradles (dtum). Some other pieces of technical vocabulary mentioned in the texts are: atappum = rivulet or ditch, at the last level of irrigation canals @AH 136 [I9901 137); kisirtum = dam to stop the water (LAPO 17 [I9981 613); maiallum = terra-cotta tube or pipe inserted in the bank of the canal and supplying a basin before dispersing in secondary drains (BAH 136 [I9901 137, and LAPO 17 [I9981 624); na&a)lum = wadi (BAH 136 [I9901 114 n. 44, and 141); idditum = brook or torrent tumbling down from the plateau (LAPO 17 [I9981 598); takkirum = diversion canal (BAH 136 [I9901 131-132; ycibiltum = small canal (LAPO 17 [I9981 624).
If the remains which have been discovered during the regional surveys are remains of rdkibum-canals from Amorite times (which, as we have seen, is still debated), they show that these devices, overhanging the terrace and allowing the irrigation of the surrounding lands, have a width of between 50 and 100 m (ditches included), and a height of up to 2.5 m. At the same time, it has been calculated that the water depth usually could not have exceeded half a meter, which seems little compared to the dimensions of the device. Such a disproportion between the size of the entire installation and the actual water depth can be explained by the need to absorb every year the sudden violence of the flood. Another explanation is also possible, as an astonishing letter shows that, exceptionally, these rdkibum-canals could be used not only for irrigation, but for navigation as well: in this case it was necessary to increase the canal's water level and reinforce the banks to allow navigation by boats (Flor. Mar. 1 [I9921 93-105).23 Originating from this small number of rdkibum-canals, smaller, derivational canals, the remains of which perhaps have been discovered especially along the Nahr Sdid (KimYabdun-Lim canal?), irrigated the surrounding lands. Two solutions can be advanced to explain why so few Bronze Age archaeological sites have been discovered along rdkibum-canals, if we accept that the remains visible today are theirs: either at the time there was no settlement along the canals in order to preserve a maximum of cultivable land,24or the remains of the sites have disappeared due to an exceptional rise in the water table and alluviation. Now another problem arises: what happened with all the water diverted this way from the river by a rcikibum-canal? Drainage networks were absolutely necessary to evacuate water in order to avoid complete salinization of the soil. The problem is that no remains of drainage installations have been discovered, but it is possible that annual alluviations erased those remains. Another possibility is that the small canals which have been discovered in the Mari alveolus along the cliff of the Western plateau (canals which cannot have been useful for irrigation), were used for draining water from both the canals and the wadis (fig. 2) .25 With regard to irrigation works, the letters written by Kibri-Dagan (ARM 3) are especially informative because KibrSDagan, the governor of the Terqa district, was in ARM 14 13 = LAPO 17 804: 41-42, which shows a sudden rise of 2 m in the Habur's water level within a few hours. 23 This letter shows that no difference was made, at the time, between irrigation and navigation canals. Several other documents show that the situation was the same, for example, in the kingdom of Eshnunna (cf. D. Charpin, AfO 29-30 [I9831 103-108). Therefore, the drastic opposition of these two types of canals as advoiated by archaeologists is, perhaps, too radical. 24 If this explanation is correct, quite the contrary is documented for the Islamic period, during which many settlements were located along the Nahr Sai'd canal (see the survey by B. Geyer and J.-Y. Monchambert in MARI 5 [I9871 293-344). 25 See B. Geyer, MARI 5 (1987) 312.
The mere maintenance of these canal devices is, in the end, a considerable effort: 2.000 mobilized people are insufficient for this work according to the letter ARM 6 7 = LAPO 17 796. Furthermore, as one can imagine, managing canal networks with many complicated devices that are vulnerable to the strength of the river course, requires great technical skills. The men responsible for this maintenance work are the szkirum. Their presence is essential, and in order to obtain the optimum water level in the canals they have to know, at the same time, how to let the water go (wus'iurum), how to direct it (ifiiurum) and how to stop it (sekihm). Their goal is to feed correctly and on time the 26 27
BAH 136 (1990) 101-142. For appum, cf. also M. Powell, BSA 4 (1988) 166-168, and R. Dwight, NABU 1998185.
138
B. LAFONT
many rivulets that cross the fields (atappum). In addition to calculating the water pressure on the canal banks in order to prevent damage caused by bad weather and swelling of the waters, the sgkirum must permanently monitor the river course. Their main problem is the swelling of the waters which is feared and hoped for simultaneously, because, occurring in spring, the flood reaches its maximum at the very moment of the harvest. The difficulty, therefore, is to have sufficient time to gather the crops before the water causes inevitable damage (see for example M. Birot, ARM 27, 9-10). By far the best way to control the swelling of the waters is to foresee and anticipate it. A text like ARM 2611 14: 25-27, seems to show, unfortunately without explaining how, that certain methods allowing such forecasts were known at that time. Ultimately, we can see that several very large irrigation works have been carried out this way during the Amorite period, which must have required, of course, the presence of a strong and centralized power ensuring their achievement. Without entering again some ancient debates about the hypothetical presence of centralized "hydraulic institutions" in Ancient Mesopotamia, there is no choice but to accept that it was necessary to firmly control a very large area in the Middle Euphrates valley, and to be able to mobilize the entire population in order to exploit its potential agricultural resources. We absolutely do not know who was originally responsible for these irrigation works and when they were first executed. The canal remains which have been discovered cannot be dated,28 and most of our documentation (mainly from the Zimri-Lim period) speaks only about the upkeep of these canal systems, not about their construction. On the basis of the many natural constraints and our current knowledge of the regional history we can guess that only short and discontinued periods of regional exploitation could have existed during the third and the beginning of the second millennium in the Middle Euphrates area. Thanks to archaeological and textual data, we know of at least three periods in which the Mari area was predominant: the so-called Ebla Age, the Ur In and Sakkanakku period, and the Amorite period (from Yahdun-Lim to Zimri-Lim). The last of these three periods is also the best documented, but what is certain is that, of the known Mari kings, only Yaljdun-Lim was a real "canal digger" or builder (see his inscriptions and year names),29 as opposed to Yasmah-Addu or ZirnriLim who were satisfied with maintaining and repairing the existing canal networks. In the documentation of these last two kings one never encounters large irrigation projects on the scale of the entire kingdom.30 The best documented period is thus characterized by worries about "maintenance," with ceaseless works of redigging canals (ha&um) and fighting against a sudden rise of the water level. It is therefore possible that the knowledge and technique which prevailed during the previous periods were partly lost at that time.
28 29
J.-Y. Monchambert, BAH 136 (1990) 87-100, and the present paper, above. See for example the inscription in which YGdun-Lim states: "I opened canals and did away with the drawing of water in my land" (inscription Y&dun-Lim E4.6.8.1 in D. R. Frayne, RIME 4 [I9901 603). 30 The famous "setting up in order" of the banks of the Euphrates, evoked by the year name ZL It, must be interpreted politically instead of technically.
I i
IRRIGATION AGRICULTURE IN MARI
4.
139
THE SETTLEMENT SYSTEM AND THE DIFFERENT FORMS OF LAND EXPLOITATION
This is a difficult topic because no explicit information on settlement and exploitation systems is given by the texts, nor on the way public and private properties, both of which are documented in Mari, were combined. We have seen that the lowest part of the valley is called the kamqum. This hamqum, where fields do not require special irrigation devices, seems to be entirely state-owned, excluding private fields. Letters show, however, that not necessarily the entire hamqum is cultivated: pastures (called usallum) can be located there as well. Barley is grown on hamqum fields, which represent, without doubt, a significant share of the entire public arable. The surface of the hamqum fields around Terqa, for example, necessitated the mobilization of the entire city (ARM 3 30 = LAPO 17 841). This situation can be compared to the one in south Mesopotamia, where lands situated directly along the river banks are excluded from private property. Because of the divagations of the river, and because they are rich in grass they are kept as grazings instead of being cultivated.31 Like the hamqum, the nawlim ("steppe") excludes private property. It represents the most significant part of the kingdom, stretching out as far as the desert border (qa~fim), and essentially serves as a pasture area devoted to stock-raising. Private agricultural property is thus limited to those parts of the land located outside the bamqum, nawlim and qa~lim(fig. 1 and 2). It is clear the palace keeps the best lands for itself, pushing back other landowners towards the more or less irrigable highlands. The palace may grant large portions of arable land in its possession to certain communities. The Mari king did this in favor of several Benjaminite communities: in such a case, the palace "frees" or "forgoes" (wuSSurum) these lands (ARM 2 55 = LAPO 17 705). For the state owned lands, a classification can be proposed according to three different systems of exploitation, the same we find in south Mesopotamia: 1) Lands exploited directly by the palace. These fields are cultivated according to a system called "plough" (epinnum). A Mari "plough" actually corresponds to an agricultural unit, including the human and technical means necessary for its e ~ p l o i t a t i o nA. ~"plough" ~ represents a surface from 50 to 80 ikti (perhaps 18 to 30 ha)33 and a team of 10 to 15 people. The chief in charge of such a unit is called guzallim, or more simply ikkarum "farmer." The amount of harvest is fixed in advance by the palace authorities, based on the expected productivity of each of these units. Coming from the capital, official experts always ask for the m a ~ i m u m , 3inducing ~ frequent protests from people under their jurisdiction (ARM 27 1). CorvCe takes place on this category of fields, the people liable to this work being called dlik eqlim. Their task corresponds to the ilkum
See D. Charpin and J.-M. Durand, "Remarques sur 17&levage intensif dans 1'Iraq ancien," 150 in M.-Th. Barrelet ed., L'Archkologie de l'lraq: perspectives et limites de l'interpre'tation anthropologique des documents, Paris 1980, and the remarks of D. Charpin, RA 87 (1993) 89 on the letter AbB 12 166. 32 M. Birot, ARM 9, 332-344, and F. Joannb, ARM 23, 105-132. If we accept the ika in Mari is the same as in southern Mesopotamia (which is M. Powell's opinion in RLA 7 (1990) 486, but see the doubts expressed by M. Birot, ARM 27, 11 and the problem raised by the letter ARM 2 61 [for this text see below, on rented fields and calculated yields]). 34 The expected output is called iSkGrum: see J.-M. Durand, LAPO 17 (1998) 540 note g.
140
141
B. LAFONT
IRRIGATION AGRICULTURE IN MARI
(service due to the king).35 This category of royal lands seems to be given to a voluntary and planned agricultural economy. 2) Tenures granted to royal civil and military servants. Many letters attest the foundation of "houses" (bitum) by state servants (ARM 1 32, 41 and 56 = LAPO 17 750, 75 1 and 756), including buildings and fields. These fields are granted with all their equipment (oxen, plough, and so on). A governor, for example, might receive about 60 ikii (about 21 ha?) of arable land, and a soldier only 5 ikii (about 1,75 ha, ARM 27 107). When a king's servant receives a subsistence allotment like this, he has to offer a gift to the king in return (ARM 2 32 = LAPO 17 768), and then annually pay a sum related to his function (Jcipitiitum, sugcigiitum, etc.). The surface areas of these fields of civil servants are regularly readjusted by the authorities (ARM 5 57 and 73 = LAPO 17 762 and 763), showing that the various agricultural public lands are constantly reorganized. Upon the death or discharge of a civil servant, all the components of his "house" return to the palace (see the so-called baJitum inventory texts), as if the king regarded himself as sole legitimate heir of his servants.36As for the fields granted to regular soldiers of the army, a letter like ARM 1 6 = LAPO 17 641 gives some important information about the way they were managed (see the commentary by J.-M. Durand in LAPO 17 [I9981 335336). 3) Rented fields: the remaining royal lands, not cultivated by the palace, are granted to notables or to ordinary people (wedfiturn or muJk2num), probably for a rent. Only one letter (ARM 2 61 = LAPO 17 703) speaks about such a rent, though, and the amount is difficult to calculate: lines 18-19 show it suddenly rose from 10 to 20 gur per ikii, that is from 1200 to 2400 liters per ikii (on the assumption that one ikii in Mari equals 0,35 ha, as in south Mesopotamia, but see above n. 33). These numbers are amazing: average yields known in Ur 111Sumer, for example, fluctuate around 30 gur per bbr, that is to say 500 liters per ikii. A comparison like this is sufficient to realize that the Mari numbers of ARM 2 61 do not seem to indicate the rent, in case we accept the usual equivalencies. The exact amount of the rents of public agricultural lands thus remains difficult to evaluate.
A final, but significant, point concerning ownership of land is that in Mari, land belonging to temples is not attested, contrary to the situation in south Mesopotamia.
Since textual information comes exclusively from the palace, instead as in the south, where information from private archives is available as well, we know very little about private lands. The most explicit texts are the sale contracts from ARM 8. In addition, we see that a person as important as Asqudum owns a private property of 1000 ikii (= perhaps 350 ha?) along the banks of the Euphrates (ARM 2 28 = LAPO 17 830). Ordinary people, independent from the palace (muJk2num), can have their own fields as well, as shown by a letter from ARM 10 (ARM 10 151, to be read with MARI 3 [I9841 137-138). Another interesting letter is Flor. Mar. 2 45, in which a soldier must fight to keep his own personal property of 10 ikii separate from the public fields managed by his general.37
5.
CULTIVATED PRODUCTS, YIELDS, AGRICULTURAL CALENDAR
Cereal agriculture in the Mari kingdom provides primarily barley (SE), but also a little emrner ( Z ~ Z= kunciium) and wheat (burrum). Production, however, is characterized by instability, and never seems to be sufficient to feed the entire population, which is shown by the fact that authorities often are obliged to seek grain elsewhere (Aleppo and the upper Jazira, for example). Floodings, bad weather, locust swarms, war, in addition to the other constraints we have described, are mainly responsible for this instability. With the current state of documentation it is almost impossible to calculate the average yields in the Mari area. All the more so, since we do not really know whether the ikii in Mari has the same value as in southern Babylonia. Once gathered, the available data for such yield calculations are very disparate, as shown by the following table:
-r Text reference and date
Surface
812.7 ikzl
77 ikO I
1C
ARM 23 426: 8-9 (20-i-ZL 7')
330 ikli
ARM 23 464: 1-2 (ZL 7')
37890 ikli 488 ikli
224,5 ikli
39 ikli
?rpm-
17 ikli
(no date)
47 ikli
ARM 24 2: 7-9 (no date)
535 ikli
I
5B 35 36
For the dlik eqlim, see F. Joannbs, ARM 23, 109-113. On this question see J.-M. Durand, LAPO 17 (1998) 533, and, most recently, the conclusions of my contribution to the volume ~ t u d e smisopotamiennes. Me'langes offerts & J.-L. Huot (C. Breniquet and C. Kepinski eds., in press). 37 A situation comparable to that in Larsa during the reign of Harnmurabi of Babylon, a period known from the Samag-bbzir letters published by F. R. Kraus in AbB 4. Cf. for example AbB 4 16.
c q(no date)c m 7
433 ikli
474 ikli
Se output (in gur of Man) (then converted in the Babylonian and current measure systems, on the hypothetical basis that 1 ikli in Mari is the same as 1 ikli in I 1 Babylon 1 904.9 gur 1 1.1 gur of Mari per ikli = 7.9 ~ a b s o n i a ngur pe;bhr = 231 kg per ha [in calculating 1 qa of Se = 1 liter = 620 g] 134,3 gur 1,7 gur of Mari per ikii = 12,2 Babylonian gur per bhr (+ GN) = 357 kg per ha 233 gur 0,7 gur of Mari per i k l i = 5 Babylonian gur per bhr (+ GN) I I = 52 kg per ha 1 357241 gur 19,4 gur of Mari per ikli = 67,7 I I ~abvlonianeur ~ e bhr r I = 1683 kg hi 2421 gur 4.9 , .eur , of Mari Der ikli = 35.3 Babylonian gur per bhr ' 1 (+ GN) = 1034 kg ha 1188,s gur 5,3 gur of Mari per ikli = 38,2 ~abiloniangur pe; bhr I (+ GN) = 1119 kg ha 228,s gur 5.9 gur of Man per ikli = 42,5 Babylonian gur per bhr (+ GN) = 1245 kg per ha 128 gur 7,5 gur of Mari per ikli = 54 Babylonian gur per bhr (+ GN) = 1582 kg per ha 827,4 gur 17,6 gur of Mari per ikli = 126,7 (+ si-18 GN) Babylonian gur per bhr = 3711 kg per ha 6570 gur 12,3 gur of Mari per ikli = 88,s (+ si-18 GN) Babylonian gur per bhr = 2592 kg hi 3446,5 gur 8,O gur of Mari per ikli = 57,6 ~abyloniangur pe; bhr (+ GN) = 1687 kg per ha 2830 gur 6,O gur of Mari per ikli = 43,2 Babylonian gur per bhr (+ GN)
142
143
B. LAFONT
IRRIGATION AGRICULTURE IN MARI
Several observations can be made as a comment on this table (in order to facilitate comparisons, the following discussion speaks about Babylonian gur per bhr [= 300 liters per 18 ikfi], arbitrarily accepting a standard value for the Mari ikfi): 1) Most of the few available references date from the first month (time of the harvest, as observed by van Driel during the colloquium) of the year ZL 7', and thus seem to belong to the same group of texts. 2) Two entries sharply contrast with the others: the results of 5A and 5B are much higher (126 and 88 gur per bar) than the other yields. It should be kept in mind, however, that these texts are constructed according to the scheme "numbers + si-18 + G N (others only have "numbers + GN," simply indicating the origin of the field): we do not know the meaning of the term si-18 + GN, but it is certain that texts of this kind do not calculate yields (see also ARM 23 69: 3-5, where the given numbers followed by si-18 + GN are even more amazing and impossible to interpret as yields). 3) Other entries show differences in yields on a scale from 1 to 13 among their results: from 5 gur per bhr in entry lC, up to 67 gur per bhr in entry 2. If we accept the average yield in south Mesopotamia during the Ur 111 period (30 gur per bhr = about 860 kg per ha) as a reference point, entries lA, lB, 1C give low results (between 5 and 12 gur), entries 3A, 3B, 3C give a standard result (between 35 and 42 gur), and entries 2,4, 6A, 6B give excellent results (between 43 and 67 gur). 4) Entry nr. 2 is especially interesting because of the huge area and amount of produce involved, and because it calculates the surface and barley produce as "cultivated by the mu$k&num between GN1 and GN2, excluding fallow land,"38 with excellent results: 67 gur per bhr, more than twice as high as the average yield in southern Mesopotamia (30 gur per bhr). Does this mean that non-institutional or independent farmers (mu$k&num)usually obtained better results than the farmers working palace lands? So, within just one year, results seem to vary considerably within a small area as the Mari alveolus. In fact, we do not have sufficient data to understand the actual situation. The main problem is that we do not know the exact status of these texts: do they calculate the expected harvest (i$kirum, see n. 34), real yields after gathering the crop, or yet something else? We do not know, and finally have to remain in uncertainty.
Tree cultivation in Mari is more important that one would be1ie~e:~O fig trees, pomegranates, poplars, junipers and grapes41 are cultivated in specific areas (ARM 22 329). Olives, dates and (sometimes) wine are imported products. A recently published letter shows the know-how of mariote horticulturists to import and acclimatize several specifically foreign vegetable species around Mari42 Just a few words remain to be said, before finishing, about the agricultural calendar: it in no way differs from elsewhere in Mesopotamia: the autumn is devoted to preparing the fields and sowing, as well as to culling ~ e s a m e ; ~winter 3 is irrigation time, but this season is used for overseeing the growth of future harvests as well. The main business of the year, falling in spring, is the harvest. It takes place in AprilMay, at the time of the maximum flood. Three dangers form threats: enemy attacks, flood damages, and locust swarms.44 In order to harvest the palace fields, the entire local work force is mobilized. Grain is stored locally, first on the threshing floors (maikanum), before being delivered to, or gathered in, village lofts about a month later. During the summer (from July until autumn) principal works of preparation, maintenance, and repair of the canal networks are undertaken (clearing out the bed, the valves, the banks, installing and maintenance of technical devices, etc.). Ultimately, the general impression is that grain shortage forms a permanent threat to the Mari kingdom. Agriculture alone is often insufficient to feed the entire population. It is always necessary to keep breaking up and irrigating new fields that can carry harvests, and to import agricultural products from outside. The larger part of the information provided by the Mari letters relates to this problem and to problems with providing the fields with water, and mobilizing a work force large enough to get the crops in on time (ARM 3 1 = LAPO 17 797).
Besides cereal agriculture, sesame (Se-$4) is another main product cultivated in the Mari kingdom (ARM 2611 94). Other products grown are pulses (g6-gal, g6-tur, Se-lG, hallfiru, pulilu, iablitum [ARM 23 123]), and various aromatic and medicinal plants (Flor. Mar. 2 4), as well as, perhaps, flax (kitfim).39 These various horticultural plants are perhaps cultivated around the cities in fields and gardens called salhum, or in the middle terrace area, with water supplied from wells (daluwdtum). 40
Not much information is available on fallowing practices and crop rotation. What we translate here as "fallow land" is called A.SA wa-bi-tum. 39 There is no direct reference to flax cultivation in our texts, mention is only made of the use of linen clothes in Mari.
See J. Bottkro, ARM 7, 254, and the texts ARM 22 329 and Flor. Mar. 3 129. See also J.-M. Durand, LAPO 16 (1997) 333-337. 41 Some vineyards were cultivated within the Mari kingdom: ARM 22 329, and B. Lion, Flor. Mar. l(1992) 108-109. 4 2 Flor. Mar. 3 129. 4 3 J.-R. Kupper, 269 in M. A. Beek et al. eds., Symbolae Biblicae et Mesopotamicae Francisco Mario Theodoro de Liagre Bohl dedicatae, Leiden 1973. 44 On locust swarms, see ARM 27, esp. p. 10.
PIHANS LXXXVIII, 2000
Transport of Agricultural Produce in Arrapbe* Jeanette C . Fincke - Wiirzburg INTRODUCTION
The kingdom of Arraphe, situated east of the 6ebal Hamrin and west of the Zagros mountains with the Lower Zab crossing its northern part, was named after its capital city ~ r r a p h e l ~ l - i l a nmodern i, Kirkiik. The population was mainly of Hurrian origin, but the documents they left behind were written in Akkadian. The general term for these documents is 'Nuzi texts' because Nuzi, modern Yorgan Tepe, was the first Arraphean city that was identified with an ancient mound. The Nuzi texts span about 100 years,l beginning at around 1450 B.C.2 Most of the Nuzi texts dealing with the agriculture and topography of the Arraphean kingdom refer to the middle and mid-western section of the kingdom, which is due to the places the excavated tablets come from.3 The incidental tablet finds in ArrapbeIKirkuk4 originate from private archive^.^ The tablets from Tell al-FabhirlKurrul~anniwere unearthed in different rooms of the so-called 'Green Palace.'6 Most of the known *My thanks are due to the organizers of the 'Third MOS Symposium' on 'Rainfall and Agriculture in Northern Mesopotamia' for inviting me to give a paper on this topic. Especially, the arrangements of Dr. Remko M. Jas, Amsterdam, made this symposium unforgettable. I would like to thank Dr. Sabina Franke, Schwerin, for our discussions about the Nuzian transportation - many of her suggestions enriched this paper. Nicole Pfeifer, Wiirzburg, kindly read the final version of this article. 1 See G. Wilhelm, Untersuchungen zum @n-ro-Akadischenvon Nuzi, AOAT 9, Neukirchen-Vluyn 1970, 10, and A. H. Friedman, "Toward a Relative Chronology at Nuzi," SCCNH 2, Winona Lake, Indiana 1987, 109-129. 2 For absolute dates of the Nuzi texts see D. Stein, "A Reappraisal of the 'SauStatar Letter' from Nuzi," ZA 79 (1989) 36-60: Nuzi was destroyed around 1348 B.C. 3 'Nuzi tablets' with obscure origin are published by M. Krebernik, "Eine 'Memorandum'-Tafel ([uppi tabsilti)," SCCNH 8 (1996) 305-308, and J. Fincke, "Nuzi Fragments from the Estate of R. F. S. Starr," SCCNH 9 (1998) 63-70. 4 For the publication of these texts see J. Fincke, Die Orts- und Gewassernamen der Nuzi-Texte, RGTC 10, Wiesbaden 1993, xii footnote 1, and J. Fincke, "More Joins Among the texts from Arraphe (Kirkiik)," SCCNH 9 (1998) 49f. For further publications of British Museum Nuzi texts see M. P. Maidman, "Some Late Bronze Age Legal Tablets from the British Museum: Problems of Context and Meaning," in: B. Halpern and D. W. Hobson (editors), Law, Politics and Society in the Ancient Mediterranean World, 1993, 42-89, and by G. G. W. Miiller, Studien zur Siedlungsgeographie und Bevolkerung des mittleren Osttigrisgebietes, HSAO 7, Heidelberg 1994, 235-271, and G. G. W. Miiller, Londoner Nuzi-Texte, SANTAG 4, Wiesbaden 1998. 5 The information about these accidental finds is summarized by J. Fincke, SCCNH 9 (1998) 49-51. See also 0. PedersCn, Archives and Libraries in the Ancient Near East 1500-300 B.C., Bethesda, Maryland, 1998, 30-32. 6 See the catalogue of the texts published by F. N. H. Al-Rawi, Studies in the Commercial Life of an Administrative Area of Eastern Assyria in the Fifteenth Century B.C., Based on Published and
148
149
J. C. FINCKE
TRANSPORT OF AGRICULTURAL PRODUCE
Kumhanni-tablets7 originate from private archives as well and only a few belong to the administration of the palace. The tablets excavated at Yorgin TepeLNuzi8 belong to both private archives and the palace archive. These three sites are located in the middle of the kingdom of Arraphe: Kurruhanni located in the mid-west, Nuzi in the center, and Arraphe in the mid-eastern part (see the map added to this paper). To study the structure of agriculture in Arraphe and the transport of agricultural produce we need the palace archives or archives of landowners with large properties. Private archives usually refer to a rather limited area. The consequence is that nearly all tablets referring to overland transport originate from the palace archive of Nuzi, a principle city of h a p h e .
The Nuzi texts distinguish between two ANSE measures (see HSS 9, 66)lO: A 'large' ANSE consisting of 10 BAN and a 'small' ANSE of 8 BAN. Assuming that HSS 14,205 refers to the large ANSE of 10 BAN, the amount of 1 ANSE equals about 67.4 litres of flour.
THE RECORDS OF THE ADMINISTRATIVE PALACE ARCHIVE REFERRING TO TRANSPORT
All records referring to transport are composed according to the type of good delivered. In most cases the transported agricultural product was one of different kinds of cereal, flour (mundu), or wood. There are a few references to the transport of malt ( B ~ L U G ) , sesame seeds (SE.~.GISana NUMUN), and garden-cress seeds (sahlu ana NUMUN). The records refer to a certain amount of the good that is either delivered to another village or delivered by the village. Very rarely, it is noted that the delivered goods are brought by a certain person. Sometimes the records are sealed. Because of the character of these short notes, in some cases, it is not clear whether the good is received 'by' a village or by the Arrapbean administration9 and the goods were coming 'from' the village.
Most of the records cannot be dated, not even according to the relative chronology, because of the lack of personal names that could be used for prosopographical research.11 Therefore, the different reports of transport cannot be related to each other, and investigations concerning an arrangement of periodical transport cannot be done. According to other texts some documents referring to transports might have been written during the latest period of the Nuzi text corpus.12 This was a period of increased military presence in the country because of violent incidents at the frontiers. The records concerning the transport of military equipment and supplies to the soldiers fighting at the frontiers13 certainly refer to the latest Nuzian period. But the transport of agricultural produce to the villages in the border area might refer either to an increased need caused by the presence of soldiers, or have other reasons; e.g. crop failure. A study about transport of agricultural produce in Arraphe has to investigate the inner Arrapbean transport14 and agriculture according to the following topics:
General aspects of transport in Arraphe. How far away was the destination of transports from Nuzi? How many days were needed to cover these distances? What can be said about the means of transport and the transportation routes? Who supervised the transport? Who equipped the caravans? Who was liable for the transported goods?
E.g. HSS 14,205 (SMN 966), room L 14 (palace). obv. 1 [XANS]E mu-un-du [x ANS]E of flour 2 '$a1 mah-rii is received 3 '$a1 URU Ni-WA-ia bylfrom the village Niwaja. 1 ANSE of flour 4 1 ANSE mu-un-du 5 Sa URU Su-li-i-l[i] bylfrom the village Su-il[i] 6 Sa mab-rii is received. rev. (not inscribed not sealed)
Unpublished Cuneiform Texts, Ph. D. Thesis, University of Wales, 1977, 545-549. See also 0 . PedersCn, Archives and Libraries, 1998, 28-30. 7 For the Kurrubanni texts see RGTC 10, 1993, xiii footnote 7. 8 For the publication of these tablets see RGTC 10, 1993, xiii footnote 5 (HSS XI11 is missing in this list). Furthermore the publication of Excavations at Nuzi: EN 1011, 1-65, SCCNH 8 (1996) 379468, and EN 1012, 66-174, SCCNH 9 (1998) 219-373, as well as the publication of additional Nuzi texts and joins of Nuzi texts collected in part I11 (Nuzi Notes) of the SCCNH volumes. The Nuzi texts of the P. A. Hearst Museum of Anthropology at the University of California at Berkely, originate from Nuzi, too. For these texts see J. W. Carnahan, K. G. Hillard and A. D. Killmer, "Nuzi Texts," JCS 46 (1994) 105-122, and the review of this publication by G. G. W. Muller in SCCNH 8 (1996) 309-318. The Nuzi text from the former collection of Lord Amherst of Hackney, published by S. M. Maul, AoF 22 (1995) 117-124, most likely originates from Nuzi. 9 See e.g. G. Dosch, Zur Struktur der Gesellschaft des Konigreichs Arraphe, HSAO 5, Heidelberg 1993, 40-41.
Terms of agricultural organization in Arraphe. How was the agriculture organized, in general? Who owned the agricultural production areas? Was there any relation between the size of the farm and the social status of the owner? Who was able to produce a surplus? Where have the different agricultural production areas been located? What can be said about irrigation? What can be said about the organization and structure of the canal system? What was produced in the agricultural areas of the kingdom? How high or low was the average yield of the fields? 1 ANSE = 10 BAN= 80 S ~ L A(about 67.4 1). 1 (PI) = 6 BAN = 48 S ~ L A(about 40.4 1). 1 BAN = 8 6.7 1); see G. Wilhelm, Das Archiv des Silwa-teEup, Heft 2: Rationenlisten I, Wiesbaden 1980, 27. 1 1 For a study of the prosopographical data from Nuzi texts see G. G. W. Muller, HSAO 7, 1.994. 12 Most of the documents referring to the latest period were found in room N 120 of the Nuzi palace and in rooms C 19 and C 28 belonging to a building north of the temple; see G. Dosch, HSAO 5, 1993, 12. 1 3 See e.g. T. Kendall, Warfare and Military Matters in the Nuzi Tablets, Brandeis University, Ph. D., 1975, 31, 34, 36. 14 The foreign trade is investigated by C. Zaccagnini, "The Merchant at Nuzi," Iraq 39 (1977) 171189. 10
S ~ L A(about
150
J. C. FINCKE
3. 3.1-5.
1.
Aspects of the transported goods. What are the limits in interpreting the records of transporting cereals and other agricultural products, in general? The transported goods barley (3.1.), flour (3.2.), malt (3.3.), vegetables (3.4.), and wood (3.5.) are to be investigated according to the following aspects: What was transported from where to where? How much was transported? What was the reason for the transport?
GENERAL ASPECTS OF TRANSPORT WITHIN THE ARRAPHEAN KINGDOM
The records concerning the transport of agricultural produce were found in the administrative palace archives. Therefore, the palace of Nuzi was one of the administrative centers for organizing overland transports on a large scale.
I .l.
Distances within the kingdom
The main problem in dealing with the geography of Arraphe is the identification of sites. Only the locations Kunuhanni (Tell al-Fa&$-),Nuzi (Yorgin Tepe), and Arrapbel~lilani (Kirkuk) are known. The location of three further villages has been proposed: TurSa (Tall Mahiis), Natmane (Tall 'Ali), and Lubdi (Tauq) (see the map added to this paper). The valuable information the Nuzi texts contain on transportation and agriculture can only be used for interpretation if it can be related to one of these six villages. Nuzi (Yorgiin Tepe), respectively the palace of Nuzi, was either the starting-point or the destination of most of the attested transports. Therefore, the distances between Nuzi and the other known Arraphean villages are of interest. The given period of time for travelling from one village to the other represents the minimum number of days based on the calculated air-line distance and an average marching rate of about 18-20 km daily.15 The northern frontier of the kingdom was located in the region north of the Lower Zib. In TurSa, which is supposed to be modern Tall Mahus, the territory of the village extended to the northern bank of the river (ina eberta) including a forest ( G I ~ T I R :JEN 525; 670) and part of the dimtu Tebip-tilla (JEN 284). This part of Tur6a is close to Assyrian territory (HSS 16, 393). The air-line distance between Tall Mahus and Yorgin Tepe (Nuzi) is about 42 krn, between Tall Mahus and K i r h k (Arraphe) about 52 km. To travel from TurSa to Nuzi one would need more than two days, and approximately three days from Tur6a to the capital city Arraphe. Natmane is the north-westmost Arraphean village at the border with Assyria (HSS 16, 393) known to us. Natmane is supposed to be modern Tell 'Ali,16 which is about 16 km south-west of Tall Mahus (TurSa). If this identification is correct, the air-line distance 15 A caravan was able to cover an average distance of about 20 km a day, if the country was plain, see D. 0 . Edzard, "Karawane," RLA 5, 1976-80, 416a. The transport of goods within Arrapbean territory certainly was done by a smaller group of individuals; this group must have been able to travel faster than an overland caravan. 16 See A. Fadhil, Studien zur Topographie und Prosopographie der Provinzstadte des Konigreichs Arrapbe, BaF 6, 1983, 360b.
TRANSPORT OF AGRICULTURAL PRODUCE
151
from Natmane to Nuzi would be about 52 km, a march of approximately three days, and from Natmane to Arraphe about 65 km,17 a march of about three and a half days. Kunuhanni (Tall al-Fahhir) is the westmost identified village. The air-line distance to Nuzi is about 35 km, a two days march, and to Arraphe it is about 45 krn,*8a march of two and a half days. A transport to Tur6a at the northern border would have taken about two days (about 40 km).l9 The villages known to be located at the southern frontier with the Kassites (Ka6Su) are Kapra, Dur-ubla, and Apil-sin(we), but we do not even know their approximate location. Another southern village is Lubdi, which presumably can be identified with modern Tauq.20 The air-line distance to Nuzi is about 33 km, which would mean a two days march. The capital city ~ r r a ~ h e l ~ l - i l(Kirkuk) ini is the eastmost town of the kingdom. There is no known road leading from Arraphe to the east.21 The distance between Arraphe and Nuzi is about 13 km,22a one day march. According to these data, the longest air-line distance within the kingdom between Nuzi and a village in the border area was about 52 km, a journey of approximately three days. Even assuming that the roads covered a longer distance, according to the relief of the landscape,23 most of the transports starting at Nuzi must have reached their destiny within four or five days. These short distances might have made organizing the transport of goods easy and fast because the demand for equipment was small and there was no need for its documentation. This matter might of course explain why only few documents referring to transport and transportation have been preserved.
17
B. Kh. Ismail, "Informationen iiber Tontafeln aus Tell-Ali," in H. Klengel (ed.), Gesellschaft und Kultur im alten Vorderasien. SGKAO 15, Berlin 1982, 117, describes Tall 'Ali to be about 42 km west of KirkGk (should the number refer to miles instead of kilometers?). 1 8 Y. Mahmoud, "Tell al-Fakhar. Report on the First Season's Excavations," Sumer 26 (1970) 109. 19 A. Fadhil, BaF 6, 1983,208b. 20 This identification is mainly based on Sargonic, Middle Babylonian, Middle Assyrian and NeoAssyrian texts. For a short summary of the history of this village see A. Fadhil, BaF 6, 1983, 105b126a, esp. 126a. 21 It is hard to believe that the capital city of Arrapbe was the eastmost city of the country with no other settlement in the border region to protect the city against the enemies of the Zagros mountains (e.g. the people of KUR N/Lullu). Therefore, the lack of known villages east of Arrapbe (Kirkuk) is either due to the sources or has other reasons. One reason might have been the nature of the region east of Arrapbe and west of the 'badlands' of the Zagros: Nowadays, this countryside is characterized by sporadic date orchards located on the western rim of the Zagros foothills watered by the larger rivers that break through the Zagros mountains; see E. Wirth, Agrargeographie des Irak, 1962, pp. 166-169. According to the Nuzi texts the only settlement known to be located in this area was Azubnni (~hemchema~6amgamal or GSWKuk Tepe), which is the last Arrapbean town on the route to the country Nkullu. On that account, we must assume this area to have been a more or less wide district of uncultivated and deserted land. This situation could have been able to protect the agricultural areas of the kingdom against mountaineers. 22 R. F. S. Starr, Nuzi. Report on the Excavations at Yorgan Tepa near Kirkuk, Iraq, Conducted by Haward University in conjunction with the American Schools of Oriental Research and the University Museum of Philadelphia 1927-1931. Vol. I, Text, Cambridge, Mass. 1939, xxix. 23 See e.g. the map of the Arrapbaen roads G. G. W. Miiller added to his volume HSAO 7, 1994.
152 1.2.
TRANSPORT OF AGRICULTURAL PRODUCE
J. C . FINCKE
Transportation routes
Vehicles must have been used for every transport between two cities, although loadcarrying vehicles (GI&MAR.G~D.DA,ereqqu) are known only for transporting different kinds of ~ o o d . AS 2 ~ for animals, donkeys (ANSE, imiru) and oxen (GUD, alpu) were used in caravans.25 The system of roads linking the different villages26was well organized, and those sections exposed to danger were guarded by chariot-drivers (e.g. HSS 15, 126).27 If necessary, the roads crossed watercourses and canals either by bridges28 or fords." There is no evidence that the natural Arraphean watercourses were used for transportation. Both rivers attested in the Nuzi texts run parallel in northeast to southwest direction. The Lower Zsb, which is referred to with the expression 'on the opposite bank' (?a ebertan),30 might have been too close to Assyria to use it for tran~portation.3~ The second river attested in the Nuzi texts is called $u&. This river links the territory of Zizza (JEN 270)' a village northwest of Nuzi, with the territory nearby Unap-Se(we) (JEN 98)'" a village that must be located between Kurmbanni (Tall al-Fa&ir) and TurSa (Tall Mahiis ?).33 Maybe the $u&, like the wadis (nahlu),34 was too small to use it for navigation.
24 In particular fire-wood (~1S.ambanni)produced in the areas near the forests (see below 3.5. Transport of wood) was brought to Nuzi by load-carrying vehicles. 25 See e.g. HSS 15, 43B (SMN 3562 - not identical with EN 913, 3 16; room F 25) 4-1 8: 10 c1S.h~ar-wa-ra-ah-bu (5) ti-ib-nu (6) a-na mAr-zi-iz-z
(7) d a+na ml-pal-lu (8) a+na GUD.MES (9) Sa i+na KUR Nu-ul-lu-a-i-ti(10) Sa D U - ~ U . M(lower E S edge 11) 10 (over erasure) ~ ~ S . h a - a r - w a - ~ r a - a h(12) ~-hu IN.NU.MES (rev. 13) a+na ANSE.ME[S] (14) a+na m@-er-r[u] (15) i-din (16) Sa i+na KUR Nu-ul-lu-a-i (17) So D U - ~ U . M E S (18) i-din (4-10) Give 10 pitchforks of straw to Ar-zizza and to Ipallu for the oxen that go to the Nulluaian country. (11-18) Give 10 pitchforks of straw to Herr[u]. For the donkeys that go to the Nulluaian country give (the straw)." 26 See --- e.g. C. Zaccaenini. " . The Rural Landscape of the Land of Arraphe, Rome 1979, 52-57; RGTC 10, 1993, 419-420; G. G. W. Miiller, HSAO 7, 1994, 187-205. 27 See G. Dosch, HSAO 5, 1993, 38. 28 In Nuzi the NiraSSe canal is crossed by a bridge. The road leading to this bridge is the road to Tarkulli (JEN 13 and JEN 699). 29 The description of one section of a field as located 'at the other side' of a river (ina ebertan) might point to the presence of a ford; see G. G. W. Miiller, HSAO 7, 1994, 156, concerning the ford of the river Su& in the territory of Zizza (JEN 270 obv. 7-8): 2 ANSE A.SA.MESi+na e-be-er-ti (8) in Su-ti-a-ah a-Sar URU Zi-iz-za. 30 E.g. the dimtu Sa TeMp-tilla which is divided into two parts; see A. Fadhil, BaF 6, 1983, 204b205a, RGTC 10, 287, and G. G. W. Miiller, HSAO 7, 1994, 182. Within the Silwa-teSSup archive eberta also refers to the Lower Ziib, see G. Wilhelm, Das Archiv des Silwa-teSSup, Heft 2, 1980, 22. 3 1 Another possible explanation is that those texts referring to the river were archived in TurSa or in another northern village not yet excavated. 32 The $u& flows into the MalaSu-canall-wadi which is crossing the territory of Unap-Se(we). A first reconstruction of the territory of UnapSe(we) is offered by C. Zaccagnini, The Rural Landscape, 1979, between pages 160 and 161. For corrections see A. Fadhil, BaF 6, 1983,277a-328b, G.G. W. Miiller, HSAO 7, 1994, 138-139. 33 See A. Fadhil, BaF 6 , 1983, 328b. 34 For the attestation of wadis in the Nuzi texts see e.g. RGTC 10, 394-395: A wadi is known to run through the region west of Nuzi crossing the dimtu SumaSSawalli (EN 911, 219), and through the "
--Q-
-
153
The only hint at a possible use of Arraphean watercourses for navigation is the name of two dimtu-settlements: dimtu Mahazi (KAR = hum mahazi = akk. karu 'quay, harbor')35 located in Artihe, and dimtu K%ri (akk. kiru) located either in Nuzi or in the neighbouring Anzukalli. The dimtu Mabazi was located close to the tilu Papante (JEN 30) which borders the 'swamp' or 'lake' (jarru) (JEN 483)' and was crossed by a canal (JEN 771). This description of the dimtu as well as the location of dimtu Kiri in N ~ z (or i Anzukalli ~ ~ ?) led A. FadhiP7 to the assumption that both settlements were linked by the same canal presumably the NiragSe-canal - which must have been navigable, and both settlements therefore were named after a real quay. If Fadhil is right 'quay' should rather be understood as a quay for a ferry crossing the canal instead of a quay for boats transporting goods between two villages unless there is further evidence for the use of navigation in Arrapbe.
1.3.
Supemison of the transport
Overland expeditions like the transport of cereals to the villages in the border area can be ~ expeditions such guided by a 'herold' (~rj.niigiru)~~ or a 'governor' ( G A R . K U R ) . ~Other as the transport of cereals and other supplies to the soldiers fighting at the frontiers were guarded by chariot-drivers (riikib narkab~ite).4~ In most cases, it is not known who was to transport or to watch over the goods.
territory of Surini(we) (JENu 730). The wadi attested for the territory of UnapSe(we) (JEN 58) and the MalaSu-wadi/-canal (see nablu MalaSu in JEN 91:6; 399:6,28,39) might be identical. 35 Sa vocabulary texts from Ugarit; see e.g. J. Huehnergard, Ugaritic Vocabulaly in Syllabic Transcription, HSS 32, Atlanta, Georgia 1987, entry no. 183.5 (pp. 38/39, 83). 36 The Nuzi territory is crossed by several canals although none of them is attested for the dimtu K5ri. 37 A. Fadhil, BaF 6, 1983,44a. 38 E.g. the nagiru Arnia (HSS 16, 92) and the nagiru Qip-dalili (HSS 16, 88), who worked either seperately or together as a team (HSS 15, 264; HSS 16, 89; 103; EN 911, 286). For the duties of a nagiru in Nuzi see W. Mayer, Nuzi Studien I: Die Archive des Palastes und die Prosopographie der Berufe, AOAT 20511, Neukirchen-Vluyn 1978, 164-165. UR (HSS 14, 92; HSS 15, 264). Another duty of a GAR.KuR was 39 E.g. the ( L ~ . ) G A R . K TieS-urbe guiding delegations to foreign countries; e.g. to the country NLullu (HSS 13, 36). For the other duties of this official in Nuzi see C. Zaccagnini, Iraq 39 (1977) 171-172; W. Mayer, AOAT 20511, 1978, 121123. As for the reading of GAR.KUR: Although there are similarities between the Nuzian GAR.KUR and the (Neo-)Assyrian Saknu, as pointed out by Dr. K. Radner during the symposium, Nuzian GAR.KuR can not be read phonetically as M-kin. In Nuzi the Saknu seems to be an representative of another individual and not a governor; see e.g. HSS 9, 150 rev. 1-2: mPa-i-te-Sup Sa-ak-[nu]Sa 1 mgi-il-wa-te-Sup; according to other texts Pai-teSSup is slave OR), "man" @IS) or Sellintannu of Silwa-teSSup. Three officials written GAR.x are attested in the Nuzi texts: GAR.& Sakin biti, "administrator I manager / keeper of the house" (e.g. EN 913, 475:5-6: ... a-Sar mEl-hi-ip-til-laGAR.E I il-te-q2 "he received from Elhip-tilla, the GAR&'), usually associated with a village (e.g. HSS 13, 156:3-4 [palace archive]: mEl-hi-ip-til-la I GAR.^^ b URU Nu-zi "Elbip-tilla, the GAR.^ from Nuzi"), GAR.E.GAL,Sakin ekalli, "administrator I manager of the palace" (e.g. HSS 14, 136:16-17: mAl-wi-Su-ub z2 1 mrEP-hi-ip-tilla GAR.~~.GAL-A), and the GAR.KUR (e.g. HSS 14, 92: 6-8: ... *Ti-e-eS-ur-hi/ Lo.GAR.KUR / il-q2 "... Tie$-urbe, the GAR.KUR, received"). GAR.KUR might be read Sakin mati but it is also possible a Hurrian term existed for this official. 40 G. Dosch, HSAO 5, 1993, 38.
154
1.4.
Other transport facilities
The administrators of the Nuzi palace kept records concerning the receipt or the delivery of the goods, but none concerning the transportation. This might indicate that the people who delivered or received goods from the palace had to organize the transportation themselves. In case of a village, the mayor or another official4' of that town might have organized the transport to or from Nuzi. In addition, the merchants (LCJ.DAM.GAR) who are attested leading caravans to other countries42might have been involved in the innerArrapbean transport as we11.43 There is no hint at an individual or organization liable for the transported goods. The liabilty could either have rested on the supervisor or on the organizer of the transport, e.g. the palace. 2.
TRANSPORT OF AGRICULTURAL PRODUCE
J. C. FINCKE
TERMS OF AGRICULTURAL ORGANIZATION IN ARRAPHE
Most of the country's acreage was in the possession of the palace. Relatives of the king and a few private individuals unrelated to him also owned landed estates on a large scale.44 The real estate of the palace was spread out over the entire country. Despite the assumption that every palace in the kingdom45 must have had its own fields to support the household, fields owned by either the palace46 or the queen47 are attested in villages without a palace. Members of the king's family had their own fields. The archives of two of these rich landowners give an insight into the accumulation of landed property (archive of Tebip-
tilla, son of Surki-tilla, who might have been related to the king by marriage)48 and the organization of the administration (Silwa-teSSup, son of the king49).50 Apart from the members of the royal familiy, free landowners existed.51 Most of the fields carried a duty or tax with them called ilku, but property without a duty owned by a family is Bttested as we11.52 The ilku-fields are thought to have been 'crown-land' distributed by the king? the term ilku describes a labor tax based on ownership of real estate to be given to the king through the local government, the mayor. The other kind of tax, the izkaru, consisting either of finished goods or performing public labor was to be given to the palace.s4 Farms produced their food supply (cereals, vegetables, lifestock) themselves. Regardless of its size, every farm was organized for self-sufficiency. Even the processing of agricultural produce took place on the farms. The palace and the big landowners were able to produce a surplus. This surplus was partially stored in storerooms or granaries (8.qaritu) and partially used for commercial enterprises; e.g. lending grain to other farmers that was to be paid back after the harvest.55 Although these grain loans point to the need of farmers for cereals, almost no contracts documenting the sale of grain are attested in N ~ z i . ~ ~ The surplus was also used for accumulating fields to build up landed estates on a large-scale.57 This could be done by entering two different kinds of contract. The surplus could have been lent out as a long-term credit to the debtor in return for real estate as a pledge ($uppi titennuti).58 Or the surplus could have been given as a 'gift' to an adopted son to receive real estate as 'share' (see the contracts about the so-called 'sale adoption,' tuppi r n ~ i r u t i ) . ~ ~
48
41
See e.g. the order of Wantia in his letter EN 911, 99: Tunilenna was to go to the village hrini(we) and take 12 ANSE of barley and give it to Tannia. If there was not enough barley, the remaining amount should be taken from the village Apz*uluSSe. The order indicates Tunilenna himself had to organize the transportation. 42 For some of the goods that were sold to other countries or imported from them, see e.g. C. Zaccagnini, Iraq 39 (1977) 171-189, and RGTC 10, 1993,36. 4 3 Some merchants served as agents of the palace or as agents of private individuals; see C. Zaccagnini, Iraq 39 (1977) 171-189, W. Mayer, AOAT 205/1, 1978, 158-161. But some of them even acted on their own risk, see e.g. JEN 195 referring to a decree of the king concerning merchants who had bought Arrapheans in the country of NILullu and brought them back to Arraphe. For an archive of an Arraphean merchant, Pul&ali, see M. A. Morrison, "The Eastern Archives of Nuzi," SCCNH 4, 1993, 95-1 14. 44 See G. Wilhelm, "Zur Rolle des GroRgrundbesitzes in der hurritischen Gesellschaft," RHA 36 (1978) 205-213. 45 There are palaces attested for e.g. Al-ilanil~rra~he, fjurazina rabG, Kurrubanni, Lubti, Nuzi, and Ulamme. 46 Fields of the palace are attested for e.g. Atakkal, Nuzi, Hulumeni(we), Sinina, TilpaSte, and TurSa. 47 Fields of the queen are located in e.g. Atakkal, Nuzi, Ulamme, and Unap-Se(we).
155
For this archive see e.g. M. P. Maidman, A Socio-Economic Analysis of a Nuzi Family Archive, Ph.D. dissertation, University of Pennsylvania, 1976. 49 This archive is to be analysed by G. Wilhelm in the series Das Archiv des Silwa-teGup (see: Das Archiv des Silwa-teiiup, Heft 2: Rationenlisten I, Wiesbaden 1980; Das Archiv des Silwa-teEup, Heft 3: Rationenlisten 11, Wiesbaden 1985; Das Archiv des Silwa-teEup, Heft 4: Darlehensurkunden und venvandte Texte, Wiesbaden 1992). 50 See G. Wilhelm, The Hurrians. Translated from German by Jennifer Barnes with a chapter by Diana L. Stein, (Aris & Phillips Ltd) Warminster 1989, 46-47. 51 For the part of the military in the structure of the Arraphean society, the Arrapbean riikib narkab(ii)te and the Hanigalbatean marijannu, see recently G. Dosch, HSAO 5, 1993. 52 See G. Dosch, HSAO 5, 1993, 68-69. 53 See recently G. Dosch, HSAO 5, 1993, 69-71. 54 See E. Cassin, "Quelques remarques ?I propos des archives administratives de Nuzi," RA 52 (1958) 24-27, G. Dosch, HSAO 5, 1993, 39, 40, and M. P. Maidman, "'Privatization' and Private Property at Nuzi: The Limits of Evidence," in: M. Hudson and B. A. Levine, eds. Privatization in the Ancient Near East and Classical World, Cambridge, Mass. 1996, 155. 55 See D. I. Owen, The Loan Documents from Nuzu. Brandeis University, Ph. D., 1969. 56 There are only a few exceptions: e.g. HSS 16, 231 and HSS 19, 126, two documents belonging to the archive of the merchant Pul&ali; see M. Morrison, SCCNH 4 (1993) 104-105. 57 See G. Wilhelm, RHA 36 (1978) 207-208 with further literature. 58 For these documents see e.g. B. L. Eichler, Indenture at Nuzi: The Personal Tidennitu Contract and its Mesopotamian Analogues, YNER 5, New Haven 1973. 59 See recently G. Dosch, HSAO 5, 1993, Chapter 2, with further literature.
156
157
J. C. FINCKE
TRANSPORT OF AGRICULTURAL PRODUCE
Contrary to this situation, the real estate of the free landowners who had not been able to produce a surplus to enlarge their property became smaller with every generation. This is due to the Arraphean inheritance practice:6O the eldest son and chief heir inherited a double share of the testator's property while each of his younger brothers received a single share. Consequently, some farms became too small for a family to live of the produce of the farm alone. Additionally, farmers with smaller estates depended on a good harvest. A year with crop-failure could lead them into debt. There are a lot of farmers attested in the Nuzi texts who lost their real estate to the big landowners by an mariitu-arrangement. According to the terms of this arrangement the farmers still cultivated the fields they formerly owned and had usufruct of (part of ?) the crop. In some cases, these indebted farmers must have even sold their family property to big landowners.61
rain.66 Within the territory of the same settlement 'irrigated' as well as 'non irrigated' fields67 occur; see e.g. Anzukalli, Nuzi, Turs'a, (dimtu) Sillia(we), and Zizza. Gardens and orchards are also said to be 'irrigated.'68 Because irrigation was vital to protect the farmers against crop failure the organization and structure of the Arraphean canal system should be summarized:
2.1.
2.1.1.
The canal network of the kingdom
The system of the canals crossing the territory of the different settlements was well organized.@ The canal inspector (gugallu)70 was responsible for the canal(s) of a village.71 His task was to watch the condition of the canals?2 to organize the maintenance of the canals,73 to open further canals,74 and to give the rights for using the water to the owners of the adjoining fields.75
Excursus: Agricultural production areas
The acreage is located within the territory of a village ( U R u , a h ) or a dimtu (AN.zA.GAR).~~In most cases, a dimtu is associated with a village and can be located within the territory closer to it (ugaru) or within its ~ e r i t u(originaly 'uncultivated area, steppe'63). The open space between the dimtus of a village, obviously, is property of the village. The limit of a village's ~ e r i t uis equal to the limit of the village itself (mi~ru).A dimtu can also have its own ugaru and ~eritu.Sometimes, a dimtu-settlement extended into a village; see e.g. AN.ZA.GAR1 URU Enna-mati(we) or AN.ZA.GAR/ URU Selwuhu. Gardens and orchards can be located within a city (ina kerhi) as well as in the ugaru or jeritu of a village or dimtu. Threshings floors are located in the inner city, in a dimtu, or, adjoining roads, in the ugiru of a village.64 Sometimes the texts differ between 'irrigated' fields (A.SA Siqu), fields bordering on a watercourse ( f 'river,' ~ hiritu 'ditches,' atappu 'canal'), and 'non irrigated' fields (A.SA la fiqu / ul5qu / ul Sa iiqi).65 Additionally, fields that are not specified to be 'irrigated' or bordering on a watercourse must be understood to be 'non irrigated' fields watered by 60 See e.g. J. S. Paradise, Nuzi Inheritance Practices, Ph.D. Dissertation, University of Pennsylvania, 1972. 6 1 Although no contracts documenting real estate sales have been found in the Arrapbean archives, there is evidence for the possibility to sell property; see e.g. HSS 9, 22 + EN 1011, 27:18-19 (A.SAii E.MESa-na Si-mi / la i+na-an-din 'do not give fields or houses for sale'); HSS 16, 3135-6 ([x] UDU anu R-mi a-na 'A.SA' ... / [nudnu] 'x sheep [given] as price for the field'); HSS 19, 44:lO ( r B 1 . u ~ . ~Sa. ~ S a-na Si-mi el-qli-8 'houses I obtained by purchase'); HSS 19,74:7-8 (ii mi-nu-um-me-e A.SA.MESS[a ...I / Sa a-na Si-mi el-qli-d 'whatever fields o[f ...I I obtained by purchase'); EN 9/1,26:20-23 (mi-nu-um-me-e / ma-ag-ra-at-ti ii DUMU.MES [...I I a-nu Si-mi a-na PN 1 Sa SUM-nu 'whatever threshing floors the sons [of ...I had given for sale to PN'); HSS 19, 5:40-43 (US-tuA.SA.MES-ia us'-tu E.MES-ia / '2' US-tu mi-immu rSu-unl-[Su-i]a / a-na Si-mi a-na ~ 0 . n ~ - k a 4 -...r i/ 'la i+na-an-din1 'from my fields, my houses, or whatever the name of my (property) might be, he is not allowed to give (anything) for sale to an unrelated individual'). 6 2 For the distinction between dimtu and alu see e.g. C. Zaccagnini, The Rural Landscape, 1979, 1424; RGTC 10, 1993, xx-xxv. 63 See G. Wilhelm, OA 22 (1983) 312-313 (review of C. Zaccagnini, The Rural Landscape, 1979). 64 See T. Richter, "Die Tenne in den Texten aus Nuzi," SCCNH 7 (1995) 69-79. 6 5 See e.g. C. Zaccagnini, The Rural Landscape, 1979, 107-113.
The canals were fed by the few rivers and wadis known to have crossed the countryside (see above 1.2. Transportation routes). Due to the number of places the Nuzi texts come from (Nuzflorgan Tepe, ArrapheKirkuk, KunuhannilTell al-Falyhir) the evidence for the canal system refers to a rather limited number of cities and villages. But nevertheless, the number of at least 28 canals76 known by name confirms the importance of irrigation for the yield of the fields in Arrapbe. See e.g. the six different canals attested for the 66 67
See C. Zaccagnini, The Rural Landscape, 1979, 107. The combination of 'imgated' (A.SASiqu) and 'non irrigated' fields (A.SA la Siqu / ul Siqu / ul Sa Siqi) in Arraphe reflects the present-day situation in this region ('transition area'); see E. Wirth, Agrargeographie des Irak, 1962, pp. 169-170 ("Ubergangssaum"). In general, this region is characterized by enough annual rainfall for dryfarming but the bad soil condition (gypsum and conglomerate in large quantities) as well as the relief (bare range of hills and rough strata) prohibit cultivation to a full extent. To avoid crop failure, the regulary cultivated area is divided up almost equally into fields watered by rain and fields watered by small canals. 68 E.g. HSS 9, 32 and HSS 14, 33. For both texts see below footnotes 71 and 74. 69 See e.g. the list of settlements that received water, HSS 15, 128: ... (32) '301 [UR]U [an-n]u-ti s'a A.MESTUK-Su-ti (33) 'kil-[me] '@-em1 [LIUGAL (34) 10 'ANSE~ A.TA.AN AM-Su-nu (35) ri+nal-ak-ki-slima ii (upper edge 36) 'a1-[nu] 'I?~.GAL'-~~! i-leq-'qlil-li ... (32-33) (These are the) 30 settlements that received water according to the king's instruction. (34-35) In each case they (the farmers) (are to) separate 10 ANSEof their field(s), and (36) they (the officials ?) take it for the palace.' 70 See W. Mayer, AOAT 20511, 1978, 129-131; A. Fadhil, BaF 6, 1983, 87a-88a. 71 See e.g. HSS 9, 32:4: ~ 0 . g ~ ~ - g a lSa - l URU u An-zu-gal-li, and HSS 14, 22:6-10: (6) mse-ha-al-t[e]rSupl (7) DUMU Te-hu-up-'Sel-e[n-n]i (8) ana gu,-gal-l[u]-ti (9) i+'nal A.MESSa URU Hu-(10) irte-pul-[u]S-mi 'He assigned Sebal-teSSup, the son of Tehup-Senni, to the office of gugallu over the waters of HuSri.' 7 2 The letter EN 911, 113 is adressed to the gugallu Sa atappi u Sa tabriu. The 'chair' of the water was raised (1. 4: ku-us-sli Sa A.MESi-te-li-ma), and the canal did not let the water reach the fields (1. 8: [ti] 'A~.MESa-tap-pu la li-Se-re-eb). 73 See e.g. JEN 370. 7 4 See e.g. the advice given in the letter HSS 14, 33:6-8: (6) 2 GIS.KIRI~ (lower edge 7) Sa LMES B.GAL-~~ (rev. 8) A.MESli-is'-qli-li ... 'the gardens possessed by the palace-buildings, should be irrigated by water...' 7 5 See e.g. the court case documented in AASOR 16, 41, and the list of 'irrigated' and 'non irrigated' fields and gardens in HSS 9, 32; for the list see C. Zaccagnini, "The Yield of the Fields at Nuzi," OA 14 (1975) 210-212. 76 See e.g. RGTC 10, 1993, 373-388; G. G. W. Miiller, HSAO 7, 1994, 204-208.
158
territory of Nuzi itself:77 atappu NiraSSe, atappu ArtamaHSe, atappu Kalalu, atappu i a Kelia, atappu Sara(e), and atappu Sa TabriSti. 2.2 The produce of the agricultural area On the main part of the fields different kinds of grain are cultivated. Seeds (e.g. sesame~ ~ to have been cultivated in a rather limited area seeds, SE.LGIS ana N U M U N )seem (ApenaS, Lubti, TemtenaH, and Ulamrne). Records concerning vegetables and spices are very rare. They must have been planted in the gardens of every house to supply the individuals living on the farms with them. In addition, there are gardens in which trees were planted ( ~ 1 3 . ~ i1a ~~ 11~ 3 . / ~GIS~ ...3 i a ~ 1 5 . ~ 1 ~ 1 ~ ) . There are different kinds of barley (SE, ieJu), of emmer (Z~Z.AN.NA,kuniu), and of wheat (GIG, kibtu) known from the Nuzi texts. The terms of the different qualities of these cereals are: ialana kalteniwe79 and Salana GUD.MES.~O As for barley, the analysis of the seeds excavated in Nuzi does not help to identify the different kinds of barley named in the Nuzi texts.81 There are two additional terms which might refer to different types of barley32 SE kalpurbe83 and SE samminni.84 77 78 79
TRANSPORT OF AGRICULTURAL PRODUCE
J. C. FINCKE
See RGTC 10, 1993, 203; G. G. W. Miiller, HSAO 7, 1983, 82. HSS 14, 72; see A. Fadhil, BaF 6, 1983, 275b (Ulamme). The term kalteniwe does not denote a certain kind of preparation of cereals, but rather a quality of cereal; See G. Wilhelm, Das Archiv des Silwa-teSSup. Heft 3, 1985, pp. 190-192, commentary to A ~ S 170. The different kinds of kalteniwe-cereals are wheat (GIG, kibtu), emrner (Z~Z.AN.NA, kunaiu), barley (SE.MES,&'a), and seeds in general (NUMUN.MES). 80 These cereals are not used to feed livestock (for fodder zerika a m GUD.MES or SE.MEStabrci is used) but rather are a special kind of cereal, cultivated, tansported, and stored seperately; see G. Wilhelm, Das S Archiv des Silwa-teSSup. Heft 3, 1985, pp. 190-192, commentary to A ~ 170. 8 1 See. e.g. C. Zaccagnini, OA 14 (1975) 215-216. 8 2 There are other terms attached to SE in the Nuzi texts which might refer to different types, qualities or products of barley, but they also could have been added to explain the reason for giving the barley. To give an impression of the variety of these terms some of them are listed here (for anzannu see CAD A I1 152b; AHw 56a): E.g. the terms agarinna and SuSSurinna (for the other references see CAD S I11 385 'a barley byproduct'): HSS 16, 29:8 [x ANSE]3 BAN SE.MESa-gas-ri-in-[nu] '[x ANSE]3 BAN of barley (made) beer mash,' and HSS 16, 43:ll: 1 ANSESE Su-US-Su-ri-in-na '1 ANSEof SuSSurinna-barley' compared with HSS 16, 236: 10-11: [x ... SE.MBS] a-na a-gas-[ri-in-nu] (1 1) a-[nu] fu-US-Su-ri-in-[nu]'[ ... of barley] to (produce) bear mash and SuSSurinna.' E.g. the terms zarae and hurae: HSS 14,36: 1-6: [x A]NSESE.MESza-ra-e ci bu-ra-e (2) iS-tu SE.MES Sa tab-re-e ... (6) il-rte-qu'-u'l 'they received [x A]NSEof zarae and hurae barley from the barley to feed animals;' for both Hurrian adjectives see also HSS 14, 34:l-2 (10 ANSESE.MESza-ra-e / ci hu-ra-e), and 35:l-2 (5 ANSESE.MES/ [zla-ra-e i hu-ra-e). E.g. the term Sa l a n a Sibae: HSS 16, 357:21-23: 3 BAN SE.MESSa Si-pa-i 1 a-Sar PN 1 li-qu'4 '... ANSEof barley Sa Hbae is received from PN' (see also HSS 13, 172:9-10: iS-tu SE.MES/ Sa Si-pa-i). The Human term Sibae is also used in lists referring to tin (?), see e.g. HSS 13, 3:8-10: 7 MA.NA annaku / a nu H-ba-a / [Sa] nu-ad-nu '7 MA.N.4 of tin is given for (to produce ?) Sibae' (see also HSS 15, 160:2021: ... AN.NA.MES / [a-n]a Si-pa-i a-'nu1 [PN]), and in the ration list HSS 16, 348:36: (L&MES)... Si-baa Sa i-leq-qu'-u' '(people) who received Sibae.' E.g. the expression Sa purri(we): For this term and the Nuzi references see G. G. W. Miiller, SANTAG 4, 1998, 115 (SE.MESbpu-ur-ri-we; SE.MESSa pu-ri-we; SE.MESi+m GIS.BAN-Su'Sa pu-ur-riwe). Compare purri, purrena describing a kind of property (house, field) in Alal& (AHw 881a), and burru, which is a cereal in Mari (CAD B 330b; AHw 140a).
2.3.
159
Excursus: The average yield of the fields
As indicated before, there were years with crop failure; see e.g. the year formula: MU Sa in-nk-er-iu-ma 2 la-a i-~i-dzi-~u?'year in which they planted but did not harvest' (JEN 289:33-34). In his study of 'The Yield of the Fields at Nuzi' (OA 14 [I9751 181-225) C. Zaccagnini draws attention to the need for irrigation to increase the yield of the Arrapbean fields (pp. 213-214). This can be demonstrated by the calculated rates of crop : seed for barley (p. 213): The worst rate of 1:1 is attested for the fields of Ulamme, a village that must have been located southeast of Nuzi. There is no canal attested for the territory of Ulamrne. Artibe located in the region near the lake or swamp (jarru) crossed by at least one canal, shows the high ratios from 10:1 and 8: 1 to 7.5: 1. The fields of ApenaS, a village west of Nuzi with several 'irrigated' fields,85 show the typical ratio 8: 1. The ratios from the fields of Unap-Se(we), where a wadi and several canals are attested, vary from 20: 1 to 7.1: 1 to 6.2: 1, but the typical rate seems to have been 7.5: 1. As for the cereals barley, emmer, and wheat in the kingdom Arraphe, C. Zaccagnini figured out that the proportion of cultivating these different grains was 4.5:l between barley and 'ernrner+wheat' (p. 217). This ratio can be explained by the fact that wheat requires more water and has a longer growing season with a greater risk of deseases, locust, etc. (pp. 216-217). Another reason might have been that wheat is less tolerant to saline soils than barley and emmer (p. 216), whereas the progressive salinization was intensified by irrigation (p. 217). 3.
THE TRANSPORTED GOODS
Most of the texts refer to the transport, respectively the delivery, of cereals in general. Keeping in mind that the evidence for the given rates of crop : seed (between 20: 1 and 1:l; see above 2.3.) comes from a limited number of texts referring to different years and different villages, the average ratio might have been slightly higher, in some cases.86 According to ration lists SE.MESSa URUDU.MES is given as ration to women of the palace (HSS 13, or to various individuals (HSS 13, 159:15: [SU.N~GIN 11 104:8: 3 BAN SE.MESSa URUDU.MES) ANSESE.MESSa U R U D U . ~SE.MES ~ ) . b ZABAR is given to the children of the palace (HSS 13, 80:l-2: [5 B]ANSE ii-tu / SE.MESSa zABAR), and SE.MES;a K ~ J . s I G is ~given ~ to women (HSS 13, 168:12-13: SU.N~GIN 1 ANSE1 (PI) 3 BAN SE.MES/ $a K ~ . s I G ~ ~ ) . 83 According to the ration lists the quality of barley called kalpurbe was given to soldiers (HSS 14, 217), to farmers for seed (ana NUMUN; HSS 14, 86), and to other men (e.g. HSS 14, 82; 119), as well as to feed mules (HSS 15, 273), or to produce flour (HSS 14, 54). 8 4 This type of barley (CAD S 118b) is known from three lists of grain only. HSS 14, 86:l-2: 43 ANSESE.MES(2) za-am-mi-in-ni; HSS 15, 273:13-14: SU.N~GIN 27 ANSE(14)l (PI) 2 BAN SE za-am-minu; HSS 16, 124:12: 4 ANSEza-am-mi-[in-nil. 8 5 See RGTC 10, 31. 86 One could ask why the fields of Ulamme were cultivated at all if the ratio of 1:l was the regular rate.
160
J. C. FINCKE
Nevertheless, there were areas which were less productive than others, and years with crop failure, as well. This situation could have been a reason for the transport of cereals to these regions. Villages that delivered cereals and other agricultural products to the Nuzi palace must have been located within the main acreages for the different cereals and products. But the analysis of the lists in this context is complicated by our limited knowledge of the location of the Arraphean villages. In addition, the villages Al-ilini and TaSenni(we), Arwa and dimtu Selwuhu, Sumanni(we) and Zuja(we), Pa?jl$iraS(we), Sillia(we), and Zizza are closely associated with the large-scale estates of Silwa-teSup, son of the king.87 We do not know anything about the role of Silwa-teSSup in the general accumulation and distribution network of the palace. With regard to these villages it is difficult to distinguish between transport initiated by Silwa-te%up himself on account of his own estate and transport initiated by the Nuzi palace to support individuals.
3.1.
Transport of barley
The records concerning the delivery of a certain good by the palace in Nuzi can be compared with regular ration lists that refer to the ration given to a person. The difference between them is that the records concerning the delivery by or to villages usually refer neither to a certain length of time88 nor to the number of persons of that village. E.g. HSS 16, 205 (SMN 1551), no room number (possibly palace archive room R 76)89
1' 2' 3' 4' 5' 6' 7'
(upper part of the tablet destroyed) [x] ~ A N S [SE.MES E~ flu URU I-k[i-ru] [x ANISE3 BAN SE.M[ESd a URU Qa-za-[ [x ANS]E2 BAN SE.MES'$a1 URU SIMUG!(MURUB~).[MES] [XANS]E3 BAN SE.MES$a URU Zu-[ [XANS]E5 BAN SE.MES3a URU "A1-[ '3' ANSE6 BAN SE.MES3a URU [ 13 ANSE4 BAN SE.MES3a URU Te-[
8' 9'
57 ANSE5 BAN SE.MES 3a URU A-ri-a-we
obv.
lower edge
rev.
(erasure) (rest of rev. destroyed)
87
See G. Wilhelm, Das Archiv des Silwa-teEup. Heft 2: Rationenlisten I, 1980, and Heft 3: Rationenlisten 11, 1985. 88 See e.g. G. Wilhelm, Das Archiv des Silwa-teiiup. Heft 2, 1980, 22: Most of the rations are given monthly, but there are lists of rations referring to two, three, or five months, to half a year or even to a year. The rations differ according to age and sex of the receiver. 89 Transliteration of the tablet according to my collations made in 1990.
TRANSPORT OF AGRICULTURAL PRODUCE
161
(1') [XI ANSEof [barley o]f the village Ik[iru]. (2') [x ANISE(and) 3 BAN of barle[y o]f the village Qaza[ ...I. (3') [x AN$]E(and) 2 BAN of barley of the village Nappi[hi]. (4') [x ANISE (and) 3 BAN of barley of the village Zu[ ...I. (5') [x ANS]E(and) 5 BAN of barley of the village A[. ..I. (6') 3 ANSE(and) 6 BAN of barley of the village [...I. (7') 13 ANSE(and) 4 BAN of barley of the village Te-x[...I. (rev. 8') 57 ANSE(and) 5 BAN of barley (9') of the village Aria(we). The monthly ration given to a male slave is about 3 BAN of barley. In this text the high amount of 57 ANSE(and) 5 BAN of barley is given which is 575 BAN. 575 BAN of barley could be used as ration for about 191 individuals. The reason for this transport is unknown. It is also unknown whether barley was delivered by the villages to the Nuzi palace or given by the palace as ration to the villages. But it can be stated that either in Aria(we) or in Nuzi there was an additional need to feed 191 individuals. And the palace of Nuzi was responsible to take care of these rations. This might indicate that the individuals were instructed to work for the palace either in the village Aria(we) or in Nuzi. Most of the ration lists or the records concerning the transport of cereals refer to barley. Therefore, barley was the cereal the kingdom Arraphe needed most. There is a record about 4 ANSE (and) 5 BAN of barley (about 405 BAN)that was brought in from the country of the Zagros-mountaineers called N/Lullu (HSS 16, 37).90 Compared to 575 BANof barley given to or by the village Aria(we) in HSS 16, 205 or to the total of 11000 BAN of barley to be transported to Al-qanniti (HSS 14, 124)g1the amount of 405 BAN of barley brought down from the mountains is rather small. This amount is not appropriate to compensate an increased need for barley e.g. because of crop-failure in certain areas of the country. There must have been other reasons for this 'transport' of barley - maybe this transport refers to an enterprise initiated by a private individual or merchant? Therefore, HSS 16, 37 should not be taken as evidence for regular import of grain organized by the administration of the palace. Because barley was the cereal Arraphe needed most it also was the one cultivated most. Since we do not know the location of the villages associated with barley, we can only give these results: The transport of seed-barley to the villages east of Nuzi, to Tarkulli and Turzanzi (HSS 15,235), might refer to a year with crop failure.
90
HSS 16, 37 (room R 76 - palace): 4 ANSE5 BAN SE.MES(2) ii-tu KUR Nu-ul-lu-a-i-d (3) mNi-nu-atal(4) DUMU Ha-lu-la-a+a (5) d-bi-la (6) ... '4 ANSE(and) 5 BAN of barley Ninu-atal, son of ualulaja, brought from the Nulluian country...' 9 1 HSS 14, 124 (room R 49 - palace): (1) 7 'ma-at1 ANSESE.MES(2)mWa-an-ti-ia (3) 3 ma-at ANSE SE.MES(4) mHu-ti-pa-pu ci (5) mMu-US-ie-ia(6) il-q2 (7) 1 ma-at ANSESE.MES(lower edge 8) mA-ri-ia (9) il-qd (rev. 10) SU.N~GIN 1 li-im (1 1) 1 ma-at ANSESE.MES(12) ii-tu SE.MESi a (13) MUNUS.LUGAL iitu URU (14) A-be-na a-na (15) URU URU-qa-an-na-din(16) na-ad-nu '(1-6) 700 ANSEof barley Wantija (and) 300 ANSEof barley @tip-apu and Mug-Feja took (text: singular). (8-9) 100 ANSEof barley Arija took. (10-16) The total of 1100 ANSEof barley from the barley of the queen from (the granary in) the village Apenag to the village .&l-qannsti(or: to the villages in the border area) is given.'
162
163
J. C. FINCKE
TRANSPORT OF AGRICULTURAL PRODUCE
The delivery of seed-barley by the two villages in the western border area, Halmani(we) and dimtu Marduku(e) (HSS 16,237: SU.N~GIN80 ANSE SE.MES), attests a surplus of barley. The delivery of barley to the villages at the southern frontier might reflect an increased need for cereals because of the military: Irbahe and Teliperra (e.g. HSS 16, 88: 10 ANSE), and Temtena(5) (e.g. HSS 16, 154: 50 ANSE). This might also be the case for the delivery of barley to feed animals (SE.MES tabrfi): Irbabbe and Teliperra (e.g. HSS 16,89: 3 ANSE). The villages at the northern frontier are attested in connection with the delivery of barley: Hurazina (HSS 14, 177: 1 ANSE) and Natmane ( e g HSS 15, 243: 20 ANSE). ZipuSa is the only northern village that receives seed-barley (HSS 16, 196: 9 ANSE).
analysis of all the texts from the Nuzi palace archives93 referring to flour gives the following result: The amount of flour to be delivered by each village to the Nuzi palace and 397 to 4 seems to be fixed. This quota varies94from 1 A N S E ~to~2 ANSE of flou~-96
3.2.
Transport offlour (mundu)
There are several receipts of flour by the palace. The regular formula for these receipts is a list of certain amounts of the delivered good by a village (s'a URU ... or is'tu URU ...) that is summed up with '(x) is received' (s'a maly-zi). E.g. HSS 14, 208B (SMN 1047); room R 76 (palace). Receipt of flour delivered by different villages: obv.
7
[2 ANSE] mu-un-du s'a URU rTul-ul-tu-ul-li-we $a mah-r2 2 ANSE KI.MIN s'a URU Ku-uz-za-ri-we 3 ANSE KI.MIN s'a URU Zu-ar-rril-mi-na 4 ANSE KI.MIN $a URU Qa-ra-[t]a-a+a 2 ANSE KI.MIN s'a URU Si-i-bu 2 ANSE KI.MIN s'a rURU Ku-bi-ial-we
8
SU.N~GIN15 ANSE 'mu1-[uln-du
1 2 2a 3 4 5
6
rev.
(seal impression)
(1-2a) [2 ANSE] of flour of the village Tultulliwe which is received. (3) 2 ANSE of the same (sc. flour) of the village Kuzzari(we). (4) 3 ANSE of the same (sc. flour) of the village Zarrimena. (5) 4 ANSE of the same of the village Kara[t]aja. (6) 2 ANSE of the same of the village Sibu. (7) 2 ANSE of the same of the village KGbia(we). (8) All together: 15 ANSE of flour (rev. seal impression). This list might indicate a certain quota of flour the different villages had to deliver. One memorandum points to an annual quota of flour that was to be given to the palace.92 An
92
See e.g. HSS 14,209 (room R 76 - palace) with no reference to a village: (1) 8 ANSEmu-un-du (2) 10 ANSEbu-uq-lu (3) Sa Sa-ad-dd-aq-rddl (4) Sa la mah-rii '8 ANSEof flour (and) 10 ANSEof malt from (the delivery of) the last year which is not received.'
ANSE.~~ There are two overlaps, only: According to HSS 14, 211 the village called Tu-xruSSe was to deliver 3 ANSE of flour, according to HSS 14,207 the amount was 4 ANSE. According to HSS 14, 207 SamaS-Sarri(we) was to give 2 ANSE of flour, according to HSS 14, 204 it was 1 ANSE only. Because the records concerning the delivery of flour cannot be dated there is no hint at whether the amount was increased or reduced for any reason. The total amount of flour to be given by the two villages Sii-ili rab0 and Su-ili ~ e h r u is 4 ANSE (HSS 14,207). One of the villages was to deliver 1 ANSE (HSS 14,205) and the other 3 ANSE (HSS 14, 199) but the texts do not distinguish between the two villages. Most of the villages listed with flour are otherwise unknown; all villages that had to deliver 1 ANSE of flour only belong to this group.99 Four villages are known to have delivered malt (BOLUG; see below 3.3.) additionally (KGbia(we), Barbi, Sep-adad, Su-ili). One village in each case is otherwise listed in association with barley (Karantaja), barley and malt (Diiri-appi(we)), and barley and livestock (Sibu(e)). Two villages delivered additional fire-wood (G~karnbanni;see below 3.5.) (Apil-sin(we), and Bd-ahhESu). The other four villages are known from texts without reference to the delivery of agricultural products.1O0 These references indicate a specialization in processing agricultural produce in certain villages. Most of the flour producing villages are specialized in processing barley, in principle: they produce flour, malt, or both products at the same time.
9 3 HSS 16, 413 (no room number) might come from the palace archives as well: (1) [x]+4 ANSEmuun-du (2) Sa mah-r& (3) NA~(.KI$IB) mZi-il(3a) :-te-ia '[x]+4 ANSEof flour which are received. Seal of Zil-teja.' 94 Unknown amount: HSS 14, 202 (2? ANSE)from the village Sin-iSmlnni(we); HSS 14, 205 ([I? ANSIE)from Niwija; HSS 14, 210 ( I ? ANSE)from A l l i w w ; HSS 14, 21 1 (4? ANSE)from Awinni(na); HSS 15,334 (too fragmentary for any result). 95 1 ANSEof flour: HSS 14, 199 from the villages IStaraSSe(we), and Am-urbe(we); HSS 14,204 from the villages Samag-Sarri(we), and Arik-kanari(we); HSS 14, 205 from the village Su-ili; HSS 14, 207 from the villages Arik-kanari(we), and hkri-te%up; HSS 14,211 from the village of Sukri-teSSup. 96 2 ANSE of flour: HSS 14, 204 from the villages Sibu(e), Tur-Senni(we), Bel-&beSu, and Barbi; HSS 14, 207 from Sep-adad, SamaS-Sarri, Duri-appi(we), Sibu(e), Barbi, Bel-a&eSu, K6bia(we), TurSenni, WApiti, Sin-iSmlnni, and Bawurni(we); HSS 14, 210 from Sin-iSmlnni; HSS 15, 77 from Bela 5 S u (2 ANSE(and) 5 BANof flour), and As'uri. 97 3 ANSEof flour: HSS 13, 102 from the village Zarrimena; HSS 14, 199 from Su-ili; HSS 14, 211 from Hawurni(we), and Tu-x-ruSSu. 98 4 ANSE of flour: HSS 14, 206 from the village Awinni(na); HSS 14, 207 from $6-ili ~ e andb rabfi (together), Tu-x-ruSSe, Apil-sin(we), Alkera; HSS 14,210 from Kulattu(e); HSS 15,77 from Alkera ([2]?+ '2' ANSE(and) 2 BAN of flour). 99 A l l i w a b (?), Arik-kanari(we), Am-urbe(we), IStaraSSe(we), Niwia (?), SarnaS-Sarri(we) (1 and 2 ANSEof flour), and Su-ili rab6. Kuzzari(we), Sin-iSmlnni, Tultulli(we), and WApiti were to deliver 2 ANSEof flour. Tu-x-ruSSu was to deliver 3 and 4 ANSEof flour, and Alkera 4 ANSEof flour. 100 Hawurni(we), Kulattu(e), Tur-Senni(we), and Zarrimena.
164
165
J. C. FINCKE
TRANSPORT OF AGRICULTURAL PRODUCE
Contrary to the clear results concerning the different amounts of flour delivered by each village, our knowledge of the location of these settlements is poor: Sibu(e) is said to be located in the vicinity of Zizza,lol or between Sillia(we) and Zizzal02 (northwest of Nuzi). It has been proposed that Karantaja(we) is located nearby Zuja(we) and TannimuHa(we)l03 - both villages are known from the archive of Silwa-teEup.
various individuals living on the farms.1•‹6The palace received malt from unnamed individuals,l07 from a group of professional harSultannu,l08 or from various villages.
3.3.
Transport of malt (BULUG)
There are not enough records to calculate an average annual quota of malt to be delivered to the Nuzi palace by each village. The given amounts vary from 1 ANSE to 4 ANSE of malt,'@ with two higher rates of 8 ANSE~~O and 11 ANSE of malt.ll1 The location of the villages producing malt is not known. Some of them were already named with regard to producing flour (see above 3.2.).
A group of documents refers to the delivery of malt ( B ~ U G ) . E.g. HSS 14, 190 (SMN 986); room L 14 (palace). Record concerning the outstanding delivery of malt by two villages: obv.
rev.
1 2 3 4 5 6
10 ANSE BULUG.MES Sa re-bu Sa URu.Bar-bi 7 ANSE BULUG.MES Sa re-@ Sa URU A-wi-in-ni (seal impression)
10 ANSE of malt which are outstanding of the village Barbi 7 ANSE of malt which are outstanding of the village Awinni (seal impression)
The Nuzi texts give some hints at the production and usage of malt: First of all, malt is a product of processing (regular) barley.104 It is used as food rations as well as barley-corn and other barley products like beer-bread and flour.105 The big farms were able to produce enough malt for themselves; the production was done by
G. G. W. Miiller, HSAO 7, 1994, 100. A. Fadhil, BaF 6, 1983, 158a. A. Fadhil, BaF 6, 1983, 102a. Differently: G. G. W. Miiller, HSAO 7, 1994, 59. See e.g. HSS 14, 193 (no room number - palace ?): (1) 1 SE a+na (2) BULUG d ba-up-pf ri (3) i+na URU Nu-zi ... (8) SUM-din'1 of barley to (produce) malt and beer-bread in Nuzi ... is given.' 105 See e.g. HSS 13, 32 (room A 14 - Silwa-te%up, A ~ 166) S ... (8) 1 ANSESE.MESiu-ku-na.ME$ (9) 5 BAN 1 (over erasure) S ~ bap-pi-ra A (10) 4 BAN 4 S ~ L ASE.MES(over erasure) B~JLUG.MES (lower edge 2 ANSE4 BAN 3(0ver erasure: SE) s~A.SE.[M]ES (rev. 13) 11) [5 B]ANSE.MESZ~D.DA SUK (12) [S]U.N~GIN [flu-ku-nu a-na ... (15) nu-din '1 ANSEof barley for food, 5 BAN (and) 1 S ~ L Aof beer-bread, 4 BAN (and) 4 S ~ ofA malt, 5 BAN of flour. The total amount of 2 ANSE,4 BAN (and) 3 S ~ L Aof barley(-products) for food ... is given.' HSS 13, 82 (room C 14 = L 14 - palace) (1) 80 ANSEmu-un-du4 (2) Sa nu-ad-nu (3) 84 BULUG.MES (4) Sa nu-ad-nu (5) i-na SU (6) mEh-li-ip-LuGAL '80 ANSEof flour which is given. 84 of malt which is given in the hand of Eblip-Sarri.' HSS 13,412 (room A 23 = SilwateSSup, A ~ 112): S ... (22) 7 ANSE2 BAN SE a+na bu-uq-li (23) u! a!+na ba-pi-ri a+na mTa-ti-ip-til-l[a] (24) Q a-na fi-ti-ra-an-ti (25) ki-ma ba-pi-rril-Su<-nu> d ki-ma BULUG (26) a-di-i ITU ie-eb-li (27) a+na SU-tif~i-in-zi(28) nu-ad-nu (29) 4 ANSE (erasure) <SE> a+na BURU14 a+na BULUG.MES (30) li a+na bapi-ri a+na (31) SU-timPa-i-te-iup nu-ad-nu '(22-28) 7 ANSE(and) 2 BAN of barley to (produce) malt and beer-bread to Tatip-tilla and to fIttirianti as their (ration on) beer-bread and malt until the month Sehli is given in the hand of fKinzi. (29-31) 4 ANSEof for the harvest, (and) to (produce) malt and beerbread is given in the hand of Pai-teSSup.' 101 102 103 104
Compared to the ration of barley given to produce malt for 7 months to the palace inhabitants of ~ l - i l a n i(4 ANSE (and) 3 BAN) and to Nuzi (2 ANSE, 1 BAN (and) 4 s I L A ) , the ~ ~ given ~ amounts of malt delivered to the palace are rather low. The annual delivery the Nuzi palace receivedl13 must just have been enough for the demand of the
106 See e.g. HSS 13, 347 (room A 23 - Silwa-te~Su~, A ~ 17) S ... (44) 1 ANSE1 BAN SE a-na B ~ J L U G a-na fWA-Si-ip-ki nu-din ... (48) 2 ANSE4 BAN SE.MESa-na BOLUG.MES 2 bap-pi-ra.mS (49) a-na fitti-ri-a-an-ti nu-din ... '(44) 1 ANSE(and)l BAN of barley to (produce) malt is given to fWaSipki ... (4849) 2 ANSE(and) 4 BAN of barley to (produce) malt and beer-bread is given to fIttirianti ...' 107 HSS 13, 111 (room C 14 = L 14 - palace): (1) 32 ANSE 7 (PI) 1 BAN (2) bu-uq-lu Sa SU (3) LIIMES mab-r2 '32 ANSE(and) 7 (PI) 1 BAN of malt from the hand of the people is received.' HSS 14, 189 ~ l S]U . ~(3) ~'LCJ'.MES S 'A1.TA.'A1.[AN] (room R 76 - palace): (1) 33 ANSE SE (2) a-'nu1 r ~ U ~ ~'a1-n[a (4) mZi-li-pu-g[ur] (5) nu-ad-n[u] '33 ANSEof barley to (produce) malt is given by Zilip-ukur in the hand of each individual.' 108 HSS 14, 188 (room R 76 - palace): (1) 25 ANSEBZTLUG.MES (2) i a mab-r2 (3) Sa SU.MESL ~ . M E S ha-ar (4)-Su-ul-ta-an-nu (5) 14 ANSE BOLUGi+na UGU-@&-nu (6) ir-t[e-e]h '25 ANSEof malt is received from the hand of the hars'ultannu-men. 14 ANSEof malt is outstanding on their account.' The Hurrian term ~ar3ultannuli(barZ=o/ul=dann=i) can be explained as a profession based on the root bar?-. This root might be identical to the one quoted in the (h)iSuwa festival as bar-fa-(a-)iS, bar;a-e-es'; see I. Wegner and M. Salvini, Die hethitisch-hurritischen Ritualtafeln des (b)iSuwa-Festes, Corpus der hurritischen Sprachdenkmaler I. Abteilung: Die Texte aus Bog'azkoy, Band 4, Rome 1991, 236. There might also be an association between Hurrian bar% and Hittite NINDA.harsi (see the Hittite dictionaries J. Friedrich, KurzgefaJtes Hethitisches Worterbuch, Heidelberg 1990 (unveranderter Nachdruck der Ausgabe 1952-66), 60; J. Tischler, Hethitisches etyrnologisches Glossar, Innsbruck 1983, 186-187; J. Puhvel, Hittite etymological Dictionary Vol. 3: Words beginning with H, Berlin-New York 1991, 190-198), and the Middle- and Neo-Assyrian barSu (AHw 328b 'eine Frucht,' CAD 116a 'a food, probably a fruit'). 109 1 ANSEof malt: HSS 13, 197 from the villages Tukkeni(we), Sep-adad(we), and Diiri-appi(we); HSS 13, 87 given to mTajjuki. 1 ANSE(and) 1 (PI) 2 BAN of malt: HSS 14, 75 from Uri rabQ. 2 ANSE (and) 5 BAN of malt: HSS 14, 201 from DCiri-appi(we). 3 ANSEof malt: HSS 14, 192 from Barbi(we); HSS 14, 195 from Uri rabk 4 ANSEof malt: HSS 14, 194 from Suntarak(we). 110 8 ANSEof malt: HSS 13, 432 from 36-ili. 111 11 ANSEof malt: e.g. HSS 14, 191 from Klibia(we). 112 HSS 14, 172 (room L 17 - palace): (1) 4 ANSE3 BAN SE a+na BOLUG... (3) 'il-na URU DINGIR.MES (4) 2 ANSE1 BAN 4 S ~ L ASE.MESa+na BULUG .... (6) i+na URU Nu-zi (7) SU.N~GIN... SE.MES(8) a+na BULUG u ... (lower edge 9) a+na 7 ITU.MES ... (13) nu-ad-nu '(1-3) 4 ANSE(and) 3 BAN of barley to produce malt ... in 81-ilini. (4-6) 2 ANSE,1 BAN (and) 4 S ~ L Aof barley to (produce) malt ... in Nuzi. (713) The total of ... barley to (produce) malt and ... for 7 months ... is given.' 1 13 See e.g. HSS 14, 209 (room R 76 - palace) with no reference to a village: (1) 8 ANSEmu-un-du (2) 10 ANSEbu-uq-lu (3) Sa Sa-ad-dd-aq-rdal (4) Sa la mab-rir '8 ANSEof flour (and) 10 ANSEof malt from (the delivery of) the last year which is not received.'
166
167
J. C. FINCKE
TRANSPORT OF AGRICULTURAL PRODUCE
palace itself, but, obviously, not enough to use it as rations for individuals working for palaces in other villages. In this case the rations given consisted of barley to produce malt.
of a product made of wood (e.g. HSS 13, 315: ~1S.barwarahhu'a pitchfork,' GIS.SUSSU~U Sa NINDA 'a kind of box for bread,' ~1S.nawena'part of a vehicle,' and others). The delivered wood can also be fire-wood.
3.4.
Transport of vegetables and vegetable seeds
There are very few records concerning the transport of vegetables. Usually, the personal supply of vegetables was produced in the gardens of the different farms for immediate consumption. Transporting vegetables over a large distance would destroy them.
E.g. HSS 15,72 (SMN 853+1291); room M 79 (palace), a record concerning the arrival and the outstanding arrival of load-carrying vehicles (G&MAR.G~D.DA, ereqqu) with firewood (~1S.arnbanni): obv.
E.g. HSS 14, 73 (SMN 3410); room D 6 (a building north of the temple). Seeds of garden-cress are given to various individuals or villages: obv.
rev.
3 BAN sd-ab-rlul [a-na NUMUN] mrFjal-Su-ar i[BSa-ak-ku] 2 BAN KI.MIN! a-na NUMUN i+na URU DINGIR.MES [XBA]NKI.MIN a-na NUMUN [i+n]a URU Ta-Se-ni-[we] [x x X] 'KI.MIN1 a-na [NUMUN] (rest of obv. destroyed) (beginning is destroyed) 8' [xx] 'xl [ x x x ] ' 9' 2 BAN '5 S ~ A K[I.MIN] 10' a-na &GAL-1I 11' mFju-ta-am-m[u-US-nil 12' 'il-qdl
1 2 3 4 5 6 7
3 BAN of garden-cress [as seeds] HaSuar, the i[SSakku] 2 BAN of seed-cress idto the village Al-iliini [x BA]Nof seeds of the same (cress) [idtlo the village TaSenni[(we)] [X....I [seeds] of the same [...]...[...I 2 BAN (and) 5 S ~ L Aof the s[ame] for the palace Hutarn-mu[Sni] received. rev.
Al-i16ni/Arrapbe and TaSenni(we) are located close to each other. The distance between Nuzi and Al-ilanil~rra~be and between Nuzi and TaSenni(we) was short enough to be covered within one day. Therefore, this transport was of no danger for the transported goods, although HSS 14, 73 refers to the transport of seeds and not of the vegetable plants.114 This record might refer to an area sown with cress for the first time. This would indicate that the owner of the cress seed had the authority to choose the agricultural seeds for the fields of different villages. The owner of the seed could have been the palace, although the tablet was not excavated in the palace. The fact that the palace receives cressseed as well might indicate that the owner or provider of the cress-seeds is one of the other wealthy landowners owning real estate in different villages. Nevertheless, the organization of a landed estate concerning the sown plants was centralized, even if the fields were located in different villages and the plant in question was a vegetable.
3.5.
Transport of wood
There are records referring to wood delivered to the palace. The delivery can consist of a certain kind of wood (e.g. HSS 15, 76: (GIS)ba-kul-l[u]-x-Se; 15, 141: GIS.iai&gu) or 1 14 For garden-cress given as ration to various individuals see e.g. HSS 14, 69 and 184.
upper edge
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
... ...
[?up-puGIS.MAR.G~D.D]A.MES am-ba-an-nu GIS.MAR.G~D.D]A.ME KI.MIN s'a URU DUBSAR-ri-ni-we [x KI.MIN Sa URU Ha]-al-ma-ni-e [x KI.MIN Sa UR]U Za-al-mu-e [x KI].'MIN ?a1 URU Ar-ta-u-ta [x K]I.MIN Sa URU Na-at-ma-ni [x K]I.MIN ;a URU ~al-maini-we 2 KI.MIN Sa URU Pu-bi-Se-en-ni-wes'a 'bar1-ri-ni-we 1 KI.MIN '$a URU1 EN.SES-$U-~ 2 KI.MlN Sa URU Pu-bi-ia-we 2 KI.MIN s'a URU Ma-a's'-ka-ni-we 1 KI.MIN Sa URU En-na-ma-ti-we [x] KI.MIN $a URU Kap-ra GAL ~ S U ~ . N ~'27' G I GIS.~MAR~.G~D.DA N Sa GIkam-pa-an-nu Sa i-il-li-rkul 1 KI.MIN Sa URU ga-al-ma-ni-we 1 KI.MIN ?a URU PU-hi-~e-'en1-ni-we [X
37 38 39 40
(1) [Tablet (00 load-carrying vehiclles with fire-[wolod (2) [x load-carrying vehiclels of the same (sc. of fire-wood) from the village TupSarrini(we). (3) [x of the same (sc. loadcarrying vehicles with fire-wood) from the village Ha]lmani(we) (4) .... (14) All together: 27 load-carrying vehicles with fire-wood (15) that departed. (16) One of the same (sc. load-carrying vehicles with fire-wood) from the village Halmani(we) (17) One of the same (sc. load-carrying vehicles with fire-wood) from the village Pubi-Senni(we) (18) ... (37) All together: 35 load-carrying vehicles (38) (with) fire-[wlood which (39) [did not] depart (40) [....].we. Although there are quite a lot of gardens and groves attested in the Nuzi texts there is very little evidence for woods or forests ( ~ 1 3 . ~ in 1 ~the ) kingdom of Arrapbe: Additionally to the forest that seems to have formed the northern frontier, located between Assyria and the Arraphean dimtu Ekalli (JEN 525), we know about a 'forest-road' or a ~ ) territories of Artibe 1 ~ $a ~ 1 3 . ~in1 the 'road through a forest' (KASKAL~ 1 3 .1~KASKAL
168
J. C. FINCKE
(JEN 274), of the dimtu Arik-kani(we) (JEN 258), and of Tupgarri(we) (IM 70956 = TF1 354, Al-Rawi, Diss., 462). If A. Fadhil is correct to locate the dimtu Arik-kani(we) nearby Lubdi,ll5 all of these three forest-roads must have been in the southern or southwestern part of the kingdom: Lubdi east of the swamp, Artihe west of the swamp, and Kunubanni in the western part of the kingdom. The distance between these forests and Nuzi was short enough to be covered within two days. The forest in the north could have been reached within two to three days. Because the small gardens and groves might not have produced enough wood for the remaining part of the country, the usufruct of the forest north of the Lower Ziib should have been very important for the Arrapbeans. One might assume that the fire-wood (~1S.ambanni)ll~ was collected in the steppe or produced in the gardens or groves. But the villages delivering the fire-wood were located nearby these forests. And only two of them have gardens attested: Natmane and Pub-Senni(we). This might indicate an intentional forestry in Arrapbe that produced firewood, as well. SUMMARY
Most of the tablets referring to transport of agricultural produce in Arraphe cannot be dated. It is even impossible to date them according to the relative chronology of the Nuzi corpus spanning about 100 years. Other texts that are related to the agricultural transportation records had been written during the latest period of the text corpus. This period was characterized by violent incidents at the frontiers. According to the origin of the tablets most records concerning transport in the kingdom Arrapbe refer to Nuzi. The distances within the kingdom were so short that a transport from Nuzi to any other Arrapbean village could be done within four or five days. The equipment for such a short journey certainly did not include too many items their documentation was not necessary. This might explain the fact that there are only a few tablets referring to transport and transportation. The Arrapbean transport can be characterized as overland transport - none of the natural watercourses was used for navigation. The well organized road system linking the different villages allowed the goods to be transported by vehicles. Most of the texts concerning transport of agricultural products refer to the delivery of barley. Barley was given by the palace as ration to villages or to individuals. There is one record about 575 BAN of barley given by or to a village. This amount of barley is equal to a montly ration given to 191 male slaves. These individuals must have been agents or workmen of the palace. 115 A. Fadhil, BaF 6, 1983, 115b-1l7a. 116 Most of the firewood certainly was used for warming houses during the cold months; see e.g. G.
Wilhelm, Das Archiv des Silwa-teSup, Heft 4, 1992, 49-50. But also the Hurrian rituals needed firewood for the burnt offerings (ambassi); see D. Schwemer, "Das alttestamentliche Doppelritual 'lwt wslrnyrn im Horizont der hurritischen Opfertermini arnbas's'i und keldi," SCCNH 7 (1995) 81-116.
TRANSPORT OF AGRICULTURAL PRODUCE
169
A lot of documents refer to the transport of flour (mundu), some to malt ( B ~ L U G ) . There is a fixed annual quota for the delivery of flour by each village. The records about malt are not sufficient enough to calculate an average annual quota. The texts indicate a specialization in processing barley, in general: some villages produce flour and malt, and sometimes also deliver barley. The transport of vegetables is not recorded, probably because of the nature of the plants. But seeds of vegetables were transported. The delivery of vegetable seeds to different villages and individuals proves that the organization of agriculture was centralized even in landed estates with fields located in different villages. The fire-wood (~1S.ambanni)was not collected in the steppe or produced in the known gardens or groves. The production of fire-wood took place in the villages close to one of the three forest in the territory of Arrapbe. This might indicate a well organized forestry. The amount of wood delivered to the Nuzi palace was measured in wagonloads.
PIHANS LXXXVIII, 2000
J. C. FINCKE
The kingdom Arrapbe Agriculture in the Northern Balikh Valley The Case of Middle Assyrian Tell Sabi Abyad* F. A. M. Wiggermann (Amsterdam) 1. Introduction: Sabi Abyad as an Agricultural Production Centre 2. Rainfed Cultivation and Irrigation in the Balikh Valley 3. Arable Land and Catchment Area of the dunnu Sabi Abyad 4. Producers and Consumers 5. Yield and Surplus 6. Major and Minor Crops 7. Animal Husbandry 8. Native Views on the Cosmic Foundations of Agriculture 9. Texts
1
Starting-point
Destination
Air-line distance
Days needed for travelling (1 day = about 18-20 km)
Nuzi (Yoririn Tepe) Nuzi t ~ o r g i n~ e i e ) Nuzi (Yorgin Tepe) Nuzi (YorgSm Tepe) Nuzi (Yorgin Tepe) TurSa (Tall MahEs)
~rrapbel&ilini (Kirktik) ~ u b d(Taiiq) i Kurmbanni (Tall al-FabbTir) TurSa (Tall Mahfis) Natmane (Tall 'Ali) Natmane (Tall 'Ali)
about 13 km about 33 km about 35 km about 42 km about 52 km about 16 km
march of march of march of march of march of march of
one day two days two days more than two days three days one day
1. Surrounding (?) the 'lake' I 'swamp' (jarru) west of Lubdi and east of Artibe 2. South(east ?) of Kurmbbani in the territory of TupSarri(we) 3. North of TurSa (dzmtu Ekalli) in the area north of the Lower Zib
Fig. 1. Distances within the kingdom of Arrapbe.
1.
INTRODUCTION: SABI ABYAD AS AN AGRICULTURAL PRODUCTION CENTRE
In the time of Tukulti-Ninurta I (1233-1197),1 the end of the late Bronze Age (LBA), the Middle Assyrian (MA) Empire reaches its apex. It stretches from the Zagros in the east to the Euphrates in the west. In the north it includes the Hurrian mountain states beyond the KaSijari, and in the south, for a short time, Kassite Babylonia. The empire is divided into two parts, the east with ASSur as its capital, and the west, Hanigalbat, where a branch of the royal family rules as "grand vizier" (sukallu rabti) and "king of Hanigalbat." Hanigalbat comprises at least the Balikh Valley and its northern extension, the land Barran, where perhaps its capital was located. In the west it borders on the Hittite Empire beyond the Euphrates, in the east it includes Harbu (Tell Chuera), and possibly the whole of the Habur region with Dur-katlimmu (Tell Seh Hamad) as its centre (Fig. 1). The family that rules the western part of the empire descends from IbaSHi-ili, a son of Adad-nirari I (1295-1264), and a brother of Shalmaneser I (1263-1234). His grandson ASSur-iddin is active during the first part of the reign of Tukulti-Ninurta I, and his greatgrandson Ili-pada at the end of this king's reign, and during that of his two successors ASSur-nadin-apli (1 196-1193) and ASSur-nirari 111 (1192-1187). In between father and son (ASSur-iddin and Ili-pada) a figure of unknown descent (but perhaps a member of the family), Sulmanu-rnu~ab~i, takes the office of "grand vizier." Tell Sabi Abyad, presently being excavated by Dr. P. M. M. G. Akkermans of the RMO Leiden, lies in the Balikh valley, more or less equidistant from Tuttul on the
h hanks are due to Prof. Dr. W. Rollig, who let me use his unpublished manuscript "Das Archiv von Dur-Katlimmu," to the RMO in Leiden, who makes the excavation financially possible, to Dr. P. M. M. G. Akkermans for his indefatigable directorship, and many hours of discussion, and to Dr. R. M. Jas for sharing his agricultural and cuneiform knowledge. 1 Regnal years after Freydank 1991:188.
172
173
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
Euphrates and the land Harran in the north. South of Sabi Abyad the valley is practically deserted (Lyon, this volume); the texts mention Gilma, Serda, and Tuttul (CancikKirschbaum 1996 index, with references and discussion), but the survey did not reveal clear MA presence. Tuttul was administrated by an Assyrian governor (T 97-3), but at Tell Bi'a, the site of earlier Tuttul, not a single MA sherd came to light. To the north there is evidence for habitation both from the survey (Lyon, this volume, with map) and from the texts: Tell Jittal (Dunni-Dagal?), Tell Sahlan (Sablalu), and Tell Abyad (DunniASSur?). To the north-east lies Tell Chuera, ancient Harbu, a small MA town where contemporary texts have been found. To the east and west stretches the steppe, uninhabited except by Sutaean tribes; as pasture for their sheep this part of the Jazira is of mediocre quality (Wirth 1971: map 10 after p. 252). A desert route with road stations connects Karkemish, Sabi Abyad, Dur-katlimmu, and ASSur (Kuhne 1994, Wiggermann 1994: 13). Tell Sabi Abyad is located in the middle of a broad part of the valley, in between wadis that once carried more water than today; three km to the south in between the same wadis lies Khirbet esh-Shenef, a Middle Assyrian farmstead that as yet has only been briefly investigated (Bart1 1993). At Tell Hammam there is some evidence for MA presence (Lyon, this volume), but not enough to identify it with Amimu, the seat of an Assyrian governor (Kiihne, C. 1995:217, 2); it is certainly no longer the regional population centre it was earlier in the Bronze Age. The ancient name of Sabi Abyad is still not known, but the texts (especially those of 1998) reveal something more important, the nature of the settlement. Sabi Abyad was a dunnu, a fortified agricultural production centre, owned by the "grand vizier" and "king of Hanigalbat" Ili-pada, and probably by his father ASSur-iddin before him. During the owner's absence, which may have been most of the time, the dunnu was administrated by the "chief steward" (abarakku), first Mannu-ki-Adad, then briefly Buria, and finally Tammitte. The texts express the relation between owner and chief steward by calling items of property i a Ill-pada i a qiti Tammitte abarakke i a Ill-pada, "of Ili-pada, in charge of Tammitte, the chief steward of Ili-~ada."~ The phrase "of Ili-pada" here takes the place of "ia eekalli," "of the palace," in texts from ASSur and Dur-katlimmu. As a unit of rural habitation the MA dunnu is the successor of the Mitannian dimtu, "tower," which denotes both the (fortified) farmstead and the territory belonging to it (Jankowska 1969, 1986, Zaccagnini 1979:47-53; for the Ugaritic gt see Helzer 1979, Liverani 1989). Like the dimtu a MA dunnu is often named after a person, presumably the founder, and seems to be a family affair (Aynard, Durand 1980:42f., Nashef 1982:83ff.). Apparently this type of agricultural enterprise takes place outside of the communally controlled village lands (Biagov 1976; for the juridical status of the village
community see Postgate 1982), but remains under the administration of a town (KAJ 110:llf. = Postgate 1988 no. 36; Finkelstein 1953:132 no. 47). A few land sale documents concerning rural estates (dunnu) "across the SiSSar" (Wadi Tharthar) owned by city families resident at ASSur (Postgate 1982:308, 31 1) inform us about the size of such properties: 10 ikii (3.6 ha, KAJ 162), 60 iko (21.6 ha, KAJ 160), and 100 ikii (36 ha, KAJ 177). Usually a dunnu estate includes a well (MAL B 5 10, Roth 1995:179), a threshing floor, and a garden (omitted only in KAJ 177). That a dunnu was administrated by a "chief steward" is confirmed by evidence from ASHur (KAJ 101:7, concerning the "dunnu of the royal city"). As the property of the viceroy Ili-pada and his family, the dunnu Sabi Abyad takes a special place among the dunnu's. Its "chief steward" is responsible to Ili-pada only, whose authority overrules that of the cities. As a royal "dunnu" it has a size and producing capacity far exceeding that of a regular family dunnu (below 3) and is almost that of a small town. From the fact that Sabi Abyad is a family owned agricultural production centre (dunnu) three important points follow: The owner has a city residence somewhere. This might be in the capital of Banigalbat (Harran or Dur-katlimrnu), in the national capital ASSur, or in both. It is to that unidentified location that Tammitte sends the things Ili-pada asks for in his letters: 100 homer of cress (T 93-6), oil and spice plants for the perfume makers (T 96-1 1, T 97-34), clothes, and good linen for his bed (T 97-34). Especially the last item gives the impression that Ili-pada's residence is in the r e g i ~ n . ~ Like the dunnu's across the SiSSar (KAJ 160, 162, 177), and all property, the estate must have been granted by the crown, the ultimate landowner (Postgate 1982). Whether this would be the state of Hanigalbat or the state of AgSur depends on the reality of the title "king of Hanigalbat," but the question has little bearing on the present subject, and needs not concern us here. Granted by whatever state, the property must have a description which defines its boundaries, and sets off its rights against those of the neigbouring cities and other private landowners. This description may well be the source of the (nearly) round figures which will be encountered below.
In MA Tell Fakhariyah the relation between the owner (ASSur-iddin) and his administrator (the scribe Buli, see Giiterbock 1958 no. 5:24, 6:25) is expressed in the same way (no. 9). That the ASSuriddin of the Tell Fakhariyah texts is in fact the well known (grand) vizier ASSur-iddin appears from the fact that an impression of his seal was found among the texts (Cancik-Kirschbaum 1993:22f.). Besides ASSur-iddin also a certain Ninuajti gives orders to Buli (no. 2). He is probably identical with ASSuriddin's son (and Ili-pada's brother) Ninuajti, and as such has the right to give orders to the family's personnel. It has been suggested that ASSur-iddin resided in Tell FakhariyahIASSukanni (identification not completely certain) while still only vizier, and that he was succeeded there by Sin-mudammeq, who appears in the Uarbu letters as a regional official (possibly vizier) under ASSur-iddin (then grand vizier) (Kiihne 1995:208, Cancik-Kirschbaum 1993:22f.).
3
The relation dunnu - city residence is exemplified by the texts from Tell Fakhariyah. The excavated structure where the texts were found (Sounding VI, McEwan 1958:Pl. 87) was judged by the excavators to be "the house of a wealthy merchant in a provincial city" (ibid. 19). Although no further MA structures were excavated, the house in question was certainly not an isolated farmstead, as appears from the spread of MA material on the Tell (tablet no. 11 found out of context in Sounding IX; MA pottery from Moortgat's sounding on two opposing points on the Tell, see Hrouda 1961). In a letter to Buli (no. 2) NinuajQ announces "I will come to the dunnu," and orders beer to be brewn. This passage led Giiterbock 1958:87f. to suggest that Dunnu is the name of Tell Fakhariyah, but it is much more likely that dunnu refers to the family's farmstead out of town. In that case the letter would originally have been sent to the farmstead (where at the time Buli resided), and have been moved only later to the city residence. Another text that points to agricultural activities and presumably originates from the dunnu is no. 7, where a small quantity of pig's fat is in charge of a [rub] ikkariite. As in Sabi Abyad those directly involved would refer to the farmstead simply as "the dunnu," and use the full name only to outsiders. The full name of the dunnu belonging to the Fakhariyah mansion is revealed by Sinmudammeq in a letter found at Dur-katlimmu: Dunni-ASSur (which he calls "my dunnu," CancikKirschbaum 1993 no. 2:6). This dunnu then supports the activities of the viziers (note 2) residing in Tell FakhariyahIASSukanni in the same way as Sabi Abyad supports those of the grand vizier residing in an as yet unknown locality.
174
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
In the Neo-Assyrian period the king grants land to officials like the sukkallu, "vizier," who, unlike other officials, do not have a province @ahutu) to cover their expenses; like a province, such land can be called apahutu (Postgate 1989:146ff.). Since it is highly unlikely, and in fact unattested, that the MA sukkallu raba at the same time could be a provincial governor, this must be the bearing of the pahutu $a Zli-pada in the gina'u archive from AESur (Freydank 1992:287f. nos 5 and 6): crown land administrated (and in practice owned) by Ili-pada to cover his expenses as a state official. It is not known whether Ili-pada exploited other dunnu's besides Sabi Abyad. Farming is the sole purpose of a dunnu enterprise, and farmland an inseparable part of a dunnu settlement. The owner of the dunnu owns the lands that go with it, there are no other private landowners on his property. This means that all farmers working under the administration of the dunnu are not private landowners, but dependents of some sort. This tallies with the fact that at Sabi Abyad there is no evidence for fields being bought or sold, or put up as security for loans. More in general, there are very few private contracts (witnessed and sealed), and the debtors are always either from elsewhere (Suadikanni, Sahlalu, Dunni-Kidin-ilani), or merchants. On the contrary, the farmers are employees, who, like other employees, yearly settle their accounts with the administration: "100 (homer) of barley, (measured) by the sutu-measure of the @burnu(-House), of Tammitte the chief steward, as fodder for the horses of Ili-pada, (which are the responsibility) of Nergal-aSared the ten-man; Resiatri the chief farmer has received and will distribute them. (When) his accounts are settled, he may break his tablet. Month Sippu, day 12, eponym Sulmanu-aha-iddina" (T 98-33, Text 1). The exact legal status of the dependents is a matter of dispute. It is clear that about half of them are s'iluhlu (below 3), a class of dependents that is known from other MA sources (Fincke 1994, with reference to unpublished texts from Dur-katlimmu and comparison with the s'ellu~luin Nuzi). They are unfree workers, and at least in part originally prisoners of war; mostly they are agricultural workers, rarely professionals. Since the texts are silent about it, the legal status of the other dependents, farmers and professionals, is more difficult to establish. It is a reasonable guess that they are alaju, free born dependent "villagers" liable to ilku obligations originating in the tenure of a sustenance field (Postgate 1971:496ff., 1982:307f.). In the case of Sabi Abyad the "villagers" would be "of Ili-pada" (comparable to those "of the palace" in MAL A 5 45) and be granted fields in return for which they would owe Ili-pada ilku-service being usually agricultural (tilling the fields of the dunnu), but sometimes perhaps military.4 Whatever the exact legal status of the dunnu's dependents, in one respect they are all alike: they need to be fed and clothed. Except for the rations of the s'iluhlu workers, information on how this was achieved is scarce. In the discussion that follows (below 3) we will take our clue from what is known about another group of dependents with the same needs, the deported Hurrians resettled around Kar-Tukulti-Ninurta (Freydank 1980). The realization that the settlement excavated at Tell Sabi Abyad represents a dunnu also enlightens its internal organization (Akkermans, Wiggermann, forthcoming, summary report of the campaigns 1988-1999). The dunnu consists of a walled precinct of
60x60 meters, enclosing an area of 0.36 ha, undoubtedly not by accident equaling 1 ika. Inside there are the owner's residence (a large Mittelsaalhaus), a tower (storage, treasury, jail), the residence of the chief steward, a reception court with a small office of the latter, and at the back (not yet completely excavated)5 servant quarters among which, next to a second entry, the office of the lower members (scribes) of the administration is found. Outside the walls there are further buildings, which at the most double the available roofed space (Fig. 2). Stables for oxen, horses and sheep have not been identified, there is insufficient room to store large quantities of grain (below 5), and there is no way the about 900 dependents could be housed on the Tell. This has a bearing on the interpretation of the survey results, and we will return to the housing question below. Groups of texts or archives were found in undisturbed contexts at three points: in the small office of the chief steward at the reception court (ca 135, mostly 1998), in the residence annex office of the chief steward (ca 93, mostly 1997, also 1996, 1993), and in the office of the lower administration at the back of the fortress (ca 17, and 6 in the house next door, all 1993). A few texts were found in the owner's residence (7, all 1996), and single texts appear here and there without obvious reason. Two groups of texts, a larger (26, all 1996) and a smaller (2, both 1993) one, were found in pits dug down from a higher level, and should consequently be younger (Wiggermann 1994). The careers of the functionaries and the growing number of datable limu's, however, indicate that they are early contemporaries of the bulk of the texts. They belong to the time that Ili-pada's predecessor, Sulmanu-mu~ab~i, was grand vizier, and a certain Mannu-ki-Adad chief steward of the estate. Two short lists of workers (and a third illegible one) stem from the industrial (?) building outside the walls (Jas 1990). The total number of texts (including envelopes) is now ca 285.6 The bulk of the texts belongs to the final years of MA level 2, the level of Ili-pada and his predecessors Sulmanu-mu~ab~i and ASSur-iddin. One text stems from MA level 1, which has been reached only at a few points (T 93-1). MA level 3 has been completely excavated, but yielded no texts that certainly belong to it. Compared to the floruit of the Ili-padflamrnitte period, the 12th century level 3 is a poor affair; the walls were still intact, but inside part of the buildings laid in ruins. The 12th century remains are of great importance for the history of Assyria in the dark age, and will be discussed elsewhere (Akkermans, Wiggermann, forthcoming). It is clear that the excavated texts do not constitute the complete administration of the dunnu. Thus there is no record of the yearly census of for instance oxen and donkeys (Deller, Tsukimoto 1985, Rollig 1984:192); and although there is some sort of textile industry (T 97-34), there is not a word about wool or textile workers. Certain administrative activities are represented by one or two texts that make sense only in the context of unattested monthly and yearly accounts. For instance, there are two small texts (T 97-23, T 97-24) from the same find spot, which record each 4 szitu of beer for the meal of Ili-pada, delivered by the brewer on two different days. Such texts imply that the brewer kept daily records, and that periodically his expenses were added up and checked by the administration. That the accounts of employees were settled yearly is known from
A badly broken text listing personnel allocated to (among others) farmers seems to refer to the latter as a-li-ku-'tu Sa il-kil (T 99-5+). The list of farmers is in part identical with that of T 98-50 and T 98-54 treated below.
The SW corner was excavated in 1999; there was a bakery, and an officelstoreroom of the baker (alahinnu) with a small archive. The excavations of 1999 brought 30 further texts to light, which brings the total to ca 315.
175
176
177
F, A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
other MA texts (provisionally CAD S sabiitu mng. 8, N nikkassu rnng. lh), and the Sabi Abyad texts occasionally refer to this event (T 98-33, cited above). A MA text from Leiden shows that nikkassu "accounts" are physical objects (tablets) that could be stored in boxes (quppu; van Driel, Jas 1989:63), but not a single one turned up at Sabi Abyad. Since only a few squares at the back of the dunnu remain to be excavated, it is highly unlikely that the haphazard nature of the documentation is accidental. If not due to incomplete excavation, it must somehow be a consequence of MA administrative proceedings, of their purpose and modus operandi. As to the purpose of the administration at the dunnu, we may safely conclude that it served first and foremost its owner, Ili-pada. Since he must have had a residence somewhere else in the region, it is likely that he kept the records of his agricultural interests there. Thus at least some of the harvest accounts of Duara were recovered in the nearby administrative centre of Durkatlimmu (Rijllig 1987 unp., Arnaud 1991 no. 106; cf. note 2 for the situation in Tell Fakhariyah). It is even possible that the administration of the dunnu was ultimately shipped off to ASSur, where in fact the harvest accounts of certain provincial centres were found (Freydank 1994), as well as the rural records of the ~rad-Seruafamily (Postgate 1988). Any judgment on the completeness of a MA archive is seriously hampered by the existence of tokens and wax tables. Tokens were found at two locations in the dunnu, in both cases associated with tablets; they probably played a part in day to day recording, but exactly how and of what remains unclear. Wax tables (li'u) were used to record the comings and goings of large groups of personnel; due to their nature they would be very well suited to keep track of day to day matters, such as the deliveries of the brewer. However MA administrative proceedings are assessed, it remains a fact that Sabi Abyad is relatively short on text types that are well represented elsewhere, especially in Dur-katlimmu: harvest accounts and personnel lists. Fortunately the available evidence just suffices to establish some basic facts on farmland, yield, and population. These, in addition to a (nearly) completely excavated and well preserved settlement contained in a small and thoroughly surveyed valley, put Sabi Abyad in an excellent position to contribute to the history of agriculture in Northern Mesopotamia. Due to its atypical location, however, the results cannot simply be transposed to the whole of the rainfed cultivation belt. A correct comparative assessment can be achieved only on a broader factual basis, in which the unpublished harvest accounts from Dur-katlimmu and Harbu will play a decisive part.
northwards into the rainfed belt, extending perhaps as far north as Tell Sahlan (Wilkinson 1996:22, l988a:77). Northwards the conditions for rainfed agriculture gradually improve, from an annual average of about 300 at Tell Abyad on the Turkish border, to about 500 mm north of Urfa (Lewis 1988:686) (Fig. 3). The Balikh is a small perennial river with an average flow of 6 m3/s, which contrasts with that of the Euphrates' 840 m3/s and of the Habur's 50 m3/s. Its main source is the spring 'Ain al-'Ariis about 4 km south of the Turkish border. The difference in height between this spring and the outfall below Raqqa is 120 m, roughly a fall of 1 m per km. When the river and its tributary wadis are used for irrigation, the stream peters out far above its confluence with the Euphrates (Lewis 1988:684f.). Wilkinson calculates that the base flow of the Balikh from 'Ain al-'Arus could irrigate 3600-6000 ha, an area capable of supporting 2400-6000 people depending on land and water use (Wilkinson 1998a:Sl). Until the introduction of motor pumps nearly all irrigation water was taken from the river in channels or distributaries each one of which originated at a bifurcation of the river, where water was diverted by a simple dam or weir. The channel then gradually diverged from the river, its gradient slightly more gentle than that of the river itself. Thus fields between the channel and the river could be gravity-irrigated. Such channels take off from the river all along its course, especially in the northern part where there are innumerable ramifications (description taken from Lewis 1988:685). The MA dunnu Sabi Abyad (BS 189) and its subsidiary at Khirbet esh-Shenef (BS 170) lie between the Nahr Slouq and the Wadi el-Kheder, of which at least the former is a possibly perennial source of water, ultimately deriving from springs (now dry) at Slouq a short distance to the north (Wilkinson 1996322f.). It is a low lying and moist part of the valley, with good farmland. The moisture may compensate somewhat for the low annual precipitation. Thus the climatic situation indicates that in order to feed its dependents and generate a surplus (below 5) the lands of the dunnu required at least supplementary irrigation. By itself this circumstantial consideration is insufficient to prove that irrigation actually was practiced, but there are a number of further indications that point in the same direction: During the preceding Middle Bronze Age irrigation was practiced around Zalpab (presumably Tell Hammam) and at Tuttul (Villard 1987; Old Babylonian letter from Mari). It appears that, as in the recent past, no water reaches Tuttul when the river is used for irrigation upstream. A MA letter from Dur-katlimmu (Cancik-Kirschbaum 1996 no. 2) describes a military action of Sin-mudammeq, who seals off the Balikh in order to prevent Hittite fugitives from crossing it and returning home. Troops from Dunni-ASSur (Tell Abyad) occupy the swampy river bank to Dunni-Dagal (Tell Jittal), and troops from Dunni-Dagal occupy Serda (some 30 km upstream from Tuttul) and the erretu of Tuttul from Gilma to Dunni-Dagal. Although the letter is not clearly formulated and there remain grammatical and geographical problems, it seems safe to conclude that upstream from Tuttul, in the general area of Sabi Abyad there was an erretu. An erretu is some sort of installation, not a dike, which regulates excess water (cf. Lafont, this volume, for its role in Mari); its presence unequivocally points to the use of the Balikh for irrigation, even large scale irrigation in view of its length. The regulation of water upstream has the added advantage that a supply of irrigation water for Tuttul is safeguarded. Tuttul can only be watered from the Balikh, not from the Euphrates (Schirmer 1987:60ff., Villard 1987), and must
2.
RAINFED CULTIVATION AND IRRIGATION IN THE BALIKH VALLEY
The climate of the Sabi Abyad area is characterized by a precipitation of ca 250 rnm per annum, with virtually no precipitation in summer and a large variation in mean monthly as well as mean annual amounts. Mean annual temperature is around 20•‹, mean winter temperature about So, and the mean summer temperature reaches 28". This shows the position of Sabi Abyad on the margin of rainfed agricultural land-use, and the high hazards of dry-farming in the area (Boennan 1988:lf.). The marginal position of the site is confirmed by the archaeological history of the valley, which shows that it is located in an ill defined zone stretching from around Tell Sahlan to the latitude of Tell Zkero (BS 152; Fig. 2), which corresponds roughly to the limit of rainfed cultivation. The irrigation zone appears to have been mainly downstream from Tell Zkero, but also penetrated
178
179
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
have had a large enough population to be the seat of an Assyrian governor (T 97-3; Cancik-Kirschbaum 1996:100, with further MA references). This presumably is the reason why the erretu was called "of Tuttul." The whole installation is certainly more ambitious than the simple dam or weir that made irrigation possible in the recent past. Among the botanical samples that have been analyzed and published (several from the tower and from the building outside the walls) besides barley, the usual low moisture crop, there were large quantities of wheat, the cultivation of which requires better conditions than the local precipitation affords (cf Wilkinson 1998a:72f.); moreover, the quality of the grain was such that it aroused archaeobotanical doubts as to their production on the fields of the dunnu. After a comparative discussion of grain sizes it was concluded that "the cereal crop stored ... could have been produced locally, on irrigated fields in the Balikh valley; in that case the Assyrian occupants must have demanded for their own use the best quality grain available in the area" (van Zeist 19943545f.1. Thus the type (wheat) and the quality of the crop imply at least well watered conditions, if not simply irrigation. Irrigation would give more dependable and on average higher yields than rainfed agriculture. With a relation seed cordyield of 1:7.35 the result of agriculture at Sabi Abyad is among the best of the MA period (below 5); from the amount of grain in storage it can be concluded that the yield was dependable at least for a sequence of years. Gardens and summer crops such as sesame and cress (below 6) require additional water, which implies at least small scale irrigation. Contrary to in Dur-katlimmu, where a major canal has to be dug, irrigation along this part of the Balikh can be achieved with a minimum of effort (dam or weirs, and gravity irrigation), while labor is plentifully available (below 4). The low effort, the availability of labor, the quality and type of the crops (wheat and summer crops), the high and apparently dependable yields, and the textually attested erretu, make it safe to conclude that the irrigation suggested by the climatically marginal position of the dunnu was actually practiced.7 The textual evidence for MA irrigation practices was discussed by Simonet 1977. Especially relevant are the irrigation activities in the area "across the SiSsar" (Wadi Tharthar), south of Hatra (Nissen 1967:115f.) and outside the dry-farming belt. Without irrigation the permanent settlement of dunnu's (above 1) and villages would probably be impossible. According to the place names irrigation is practiced in this area with the help of wells or cisterns: Bur-r&te "WellICistern of the irrigation ditches" and Gubbe-ekalle "Cistern of the Palace" (Simonet 1977:165f., Nashef 1982:317). The water is brought to the fields by the use of shadoofs (zaruqqu). Inside the rainfed belt in the area of ASSur, wells and cisterns are used for supplementary irrigation (MAL B $5 17f., 0 $ 5). The evidence for MA irrigation along the Habur is insufficient. Kiihne argues that at Dur-katlimmu precipitation is definitively too low to sustain rainfed cultivation. Confirmation for the use of irrigation is found in the fact that habitation centres on the river, while the surrounding steppe is largely uninhabited. An exception is Tell Umm 'Aqrebe some 40 km east of Dur-katlimmu, where, according to Kiihne, there is an abundance of ground water, easily tapped by means of pits and sufficient for drinking and small scale irrigation (Kiihne 1990a, 1990b, Ergenzinger, Kiihne 1991). Along the
Habur a canal datable to this period could not be identified, however, and the evidence remains circumstantial. Strong indications against irrigation can be found in the poor quality of the grain (van Zeist 19943547f.;no wheat!), and in the relatively low and highly variable yields (below 5), which are typical for rainfed agriculture. In the marginal zone of rainfed cultivation Sabi Abyad has the advantage over Durkatlimmu in that the little water there is, is easily available for additional irrigation; and it has the advantage over the SiSSar towns as well as Dur-katlimmu in that precipitation is more often sufficient. In this sense the position of Sabi Abyad can be called atypical. The proceeds of farming at Sabi Abyad are supplemented by what the natural surroundings have to offer. Since only a start has been made with the investigation of the local LBA plant and animal remains (van Zeist 1994; Bottema and Cappers, this volume), there is as yet no basis for firm conclusions regarding the ecology of the region and human use of its potential. In outline, however, the Balikh does not differ much from the Habur, the vegetation of which has been thoroughly studied (Kiihne 1991): "in der (...) FluBaue (...) mu13 eine dichte Vegetation bestanden haben, die sich aus Unterholz, Wasser- und Sumpfplanzen sowie Galeriewald (Populus Euphratica) zusammensetzte." (Kiihne 1990a: 198). This landscape follows from the presence of fallow-deer and wild boar, who need this kind of environment. Marshes (GIS.GI/U~U) along the Balikh between Dunni-ASSur and Dunni-Dagal are inscriptionally attested (Cancik-Kirschbaum 1996 no. 2, cited above), and still form a prominent feature of the landscape around Tell Sahlan. The wild boar, called "marsh boar" in Assyrian (5ahapu mentioned by Sennacherib as living in the marshes beside herons and water buffaloes, cf. CAD 311 79), shows its presence in the human use of its fat and skin (T 93-10, T 98-56; T 98-7).8 Further away, in the "wasteland" (huribtu) of Mitanni there were wild bulls, and on the Habur and in the land Harran there were elephants (Miller 1986 for the ecological implications); along with lions they were hunted by Assyrian kings from Tiglath-pileser I onwards (Heimpel 1976-1980, Sterks 1996), and a lion hunt on the Balikh is illustrated by ASSurna~irpalon a bronze band from a gate at Balawat (Barnett 1973; Fig. 4). The Habur team concluded that generally speaking the vegetation (and the climate) did not change the last 5000 years, but that in Assyrian times all formations were richer in characteristic species.
There are remains of ancient canals in the valley, but none can be proved to belong to the Late Bronze Age (Wilkinson 1998b).
8
3.
ARABLE LAND AND CATCHMENT AREA OF THE DUNNU SABI ABYAD
The evidence for the amount of arable land and the size of the property of the dunnu stems basically from one text, T 98-1 15. That this text concerns all arable land and the complete harvest of one year follows from its nature as a harvest account of crown lands. Some essential figures deduced from this text can be checked against other evidence: the number of agricultural workers against the numbers of the personnel lists and of a distribution of sickles; and the size of the dunnu's farmland against the results of the survey. The figures concerning work load per person and grain consumption are held in check by what is humanly possible. The text in question, although it does not explicitly say so, is based on an administrative procedure called the "dissolution of the grain heap" (piSerti karu'e): after A preliminary investigation of the faunal remains revealed the presence of fallow deer (Dr. C. Cavallo, personal communication).
180
181
E A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
the harvest, and after threshing, the "result of the harvest" (tilit eburi) is heaped in piles (karu 'u) and "weighed" (madadu), then "dissolved" (paGru). The "dissolution of the grain heap" consists of subtracting from the harvest next year's production costs, under official supervision since it concerns crown lands (a qZpu of the king may be present); whatever remains of the harvest after this operation goes into storage (ina bit karme tabaku), where the fresh barley (SE.GIBIL)is added to the barley left over from previous years (SE.SUMUN). The essentials of this procedure are cleared up by three texts from ASSur concerning crown lands in the provinces (Freydank 1994; below 5 with tabulation), and by as yet unpublished similar texts from Dur-katlimrnu concerning local (including Duara) harvests (Rollig 1987, unpub., but used by Freydank 1994, cf. Kiihne 1990b:20 with essential information; Arnaud 1991 no. 106). In these texts the costs subtracted from the harvest results are the following: seed corn (SE.NUMUN) rations for iilublu workers for a whole year fodder for the oxen used in agricultural work for six months B 1 112 qii a day per ox (0.93 kg). The same costs reappear in a different context in KAJ 109 (Postgate 1988 no. 34, cf. Freydank 1997: 137f., also on KAJ 113), where, under supervision of a representative (qZpu) of the king, deportees receive barley "as seed corn, rations for their cattle, and rations for themselves" (a-na NUMUN SUKU-atGuD.MES-~U-~U 2 SUKU-3u-n~).~ Obviously last year's production costs are already spent. The costs that are deducted here cannot be those of last year, but must be those of next year. What happens is that this year's costs are projected onto next year, as if nothing would change: the quantity of seed corn is based on this year's cultivated area, the price of labor (oxen and men) depends on the area under cultivation. The precise amounts would be established only after they were spent, and then appear in the next year's harvest account. Thus the precise amounts are explained; they are not based on estimates, but on last year's actual costs. This means that the subtracted amounts are not yet spent, and go into storage too. The question of storage is of archaeological interest, and we will return to it below (5). Knowledge of the underlying "dissolution of the grain heap" solves the problems of the Sabi Abyad texts T 98-115 (Text 2), which lacks the explicit terms. The "fresh barley" (2) that is stored together with the "old barley" is what is left of the harvest after the "seed corn" (5) and the Silu~lurations (6-16) are subtracted (talpittu, "written off," apparently covering both amounts) from the unmentioned "result of the harvest." The result of the harvest thus can be obtained by adding the subtracted costs to the fresh barley:
The fodder for the oxen is skipped in this text, which means that it will have to be subtracted from the stored barley at some point. It is a relatively small amount (270 homer), on which we will return below. A vital figure that is omitted in this text is the number of ikii (nearly 0.36 ha) under cultivation (meres'u or eqel araje in the ASSur texts, Freydank 1994), but it can be deduced on the basis of the nearly universal MA sowing rate of 0.30 homerlikii (Freydank 1994:29, with discussion; exceptionally 0.35 homerlikii): 663 homer seed corn x 0.30 homer seed codiki2 results in 1989 ikii land under cultivation. Assuming biennial fallow (for Nuzi: Zaccagnini 1979:31; for NA: Postgate 1989:144, Fales 1990:119) this results in an arable area of ca 4000 ikii (1440 ha) under direct control of the dunnu. This amount of land fits exactly the area between the Nahr Slouq and the Wadi el-Kheder inside the circle of the catchment area discussed below. The natural boundaries would be the location of Medinet el-Fir and the confluence of the two wadis; Sabi Abyad would lie in the middle of a stretch of good well watered farmland, which could be irrigated from a LBA forerunner of the Hammam et-Turkman canal (Fig. 5). The land worked under direct control of the dunnu is not the only land that it owns. A smaller amount of probably not the best farmland must be added, that given out to and worked by the dependents (above 1) to supplement their rations. Since the produce of this land is meant to be consumed by its producers, it remains outside the horizon of the texts, and neither land, nor produce are documented at Sabi Abyad. An inventory from KarTukulti-Ninurta, which lists Subaraean deportees with their possessions, must serve as a model for all dependents. This text indicates that we should count with between five ikii (1.8 ha) and 20 ikii (14.4 ha) per family (Freydank 1980). With Freydank one may speculate that the difference in size is determined by the type (irrigated or rainfed) and quality of the land, and that the larger farms on bad lands are just as productive as the smaller farms on good lands. In the case of the dunnu Sabi Abyad all land is of comparable good quality and can be irrigated, which indicates small farms (five ikii) for families of dependents (here averaged at four members). The lands of the dunnu feed two types of dependents (above I), the Siluhlu workers, of which there are about 400 (below 4; not more than ca 100 family units), and the Zlaju farmers, of which there are about 440 (below 4; 110 families). The former receive rations, the latter do not. The 100 3ilu& family units would be granted 500 ikG (180 ha), half of which under cultivation each year. If a family of four used all its land for the production of grain, it would dispose of 6.12 homer yearly (2.5x2.45; 2.45 is the yield per ikii), that is 0.13 monthly per person, nearly as much as an average monthly ration (0.18 homer; below 4). For their sustenance the ca 440 farmers (including families) would need as much as the s'ilublu workers, who consume the produce of their land and receive additional rations. The farmers must produce their "rations" themselves; if in a mixed population of farmers each family member consumes two homers yearly (rounded off), a total results of 880 homer yearly, which can be produced on 359 ikii of cultivated land (880:2.45; 2.45 is the yield per ikii), to be doubled on account of fallow: 718 ikii (258 ha). To this figure must be added 110x5 = 550 ikii (198 ha) corresponding to the fields of the 3ilublu. Thus the land worked by the farmers as their own sustenance fields amounts to 718+550 = 1268 ikii (456 ha), half of which under cultivation yearly. Each of the 110 families of four
fresh barley stored seed corn for next year labor costs (jilublu) for next year
3366 homer 663 homer 844 homer
result of the harvest (tilit eburi)
4873 homer
For Kassite Babylonia see Sassmannhausen 1999:158.
182
E A. M. WIGGERMANN
would dispose of 11.5 ikli (4.14 ha), and produce 14 homers of grain. In a mixed population this amounts to 0.29 homer monthly per average person. The combined fields of the dependents, 500+1268 = 1768 ikli (636 ha), would produce yearly 2078 homer of grain (1768:2, on account of fallow, x2.45). This brings the harvest of all farmland owned by the dunnu to: 4873 (produced under direct control of the dunnu) + 2078 (produced by dependents for their own use) = 6951 homer of grain (430,962 kg). We will return to the yields below (5). The land-use of the dunnu is not restricted to grain production; space must be reserved for gardens and pasture. A garden forms part of the one family dunnu's owned by residents of AHSur (above I), and there is evidence for gardening in the dunnu's territory in the form of two s'iluhlu gardeners (T 98 45 Side B vi 14', 18')1•‹and small quantities of garden products (below 6). Gardens require regular attention, and are usually found near the settlement from which they are cultivated; those cultivated by employees of the dunnu should be near the tell, where they could be watered from the wadi or from wells. Other gardens may be associated with the rural habitations of s'iluhlu and farmers. In comparison to grain land the size of gardens is negligible; Sabi Abyad does not give any measures, but two gardens in Dur-katlimmu are somewhat over 1 ikli and 2 ikli respectively (Fales 1989; Rollig, Tsukimoto l999:43 If., with further evidence on the size of gardens in Dur-katlimmu). In connection with the gina'u tax grain was sent from the pahutu of Ili-pada instead of fruit (Freydank 1992:287f.), which may mean that the territories of Ili-pada do not produce fruit at all, or simply that in the particular year of the texts no fruit was available. The gardens in Dur-katlimmu produce a variety of fruit (azamru sammu~u). The amount of pasture needed can only be guessed at roughly. It depends on the number of animals (mainly oxen and donkeys) and on the way the land is used for grazing. Minimal numbers can be established for oxen (below 4,7) and donkeys (below 7), but the further evidence required is not available. One model (Hunt 1987:166) assumes for "pasture, waste, woods etc." about thrice the amount of land under cultivation. On account of the "dimorphic" structure of Mesopotamian land-use, a large part of the "waste" would not form part of the settled community, but would be exploited by more or less independent pastoralists. This means that in the specific Mesopotamian situation the share of "pasture, waste, woods etc." must be considerably lower than elsewhere. For this reason Hunt's estimated amount is halved (and rounded off). Thus, on the basis of texts from the dunnu itself (land under cultivation, number of dependents), and an estimate (pasture, waste, woods) the catchment area of the dunnu can be calculated:
T 99-5+ rev. 9ff. lists three gardeners with ten workmen.
Arable land in well wateredlirrigated area under control of the dunnu under cultivation {sustenance fields iilublu workers sustenance fields alaju farmers
+
total fallow (biennial)
+ total amount of arable land
ikli
ha
2000 250 634
720 90 228
2884
1038
2884
1038
5768
2076
= 20.76 km2
10.000
3600
=36km2
pasture, waste, woods etc.
CATCHMENT AREA
The round figure (10.000 ikli) is of course due to manipulation of the only estimated figure, that for pasture, waste, woods. It is quite well possible, however, that the round figure is realistic, and not just a consequence of estimation. Above (1) it was argued that besides being an agricultural production centre, the dunnu is also a juridical construct, which implies fixed boundaries and a fixed amount of land. Along the Balikh, in conquered territory and unhampered by pre-existing rights (this part of the valley is as good as empty), any agricultural project can be implemented as planned, and the result would be close to the model-farm in the mind of an Assyrian official. Other state farms show evidence for planning by having their chief farmers work plots of exactly 100,200, or 300 ikG (below 4); such evidence for Sabi Abyad consists in the exactly 100 farmers who till its fields (below 4), and in the exactly one ikli enclosed space on the tell. The deduced catchment area of 36 km2 can be checked against reality by plotting it on the map as a circle with a radius of approximately 3.5 km. This circle has consequences both for the internal organization of the dunnu, and for the external organization of the region. To begin with, it defines the catchment area of the next MA settlement to the north west, Tell Jittal, as a circle with a radius of 3 km. From the point where the two circles meet a boundary line can be drawn to the Balikh, the natural border in the west. This boundary line in its turn defines a square of 6x6 km, an alternative for the circular catchment area (Fig. 5). The square catchment area has the advantage over the circular one in that it suggests the features that would have played a part in the legal description of the property: the Balikh in the south west, the non-arable areas in the north and east, and the rights of Tell Jittal in the north west. There is no MA habitation west of the Balikh. The location of the two identified subcentres, Khirbet esh-Shenef and Tell Hammam Gust on the border of the square) suggests a regular spread of subcentres worked out in Fig. 6. Each subcentre would have its own catchment area, the size of which is defined by the position of Khirbet esh-Shenef in relation to the border of the square. If each subcentre has a radius of ca. 1.4 km, the square is filled with 7 subcentres (including Sabi
184
F, A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
Abyad) at regular distances (2.5 to 3 km to neighbor) from each other; four of them remain to be identified. A recent re-examination of the survey indicates that BS 161 may have been settled in the MA period. If this is indeed the case, the site falls outside of the square, and outside of the pattern of regular relations suggested by the identified subcentres. If its area can be estimated on the basis of its distance to Kirbet esh-Shenef, it would be of about the same size (radius of 1.4 km, ca. 6 km2) and probably function as an independent private dunnu. It can be considered a fact that the dunnu's dependents numbered at least some 900 souls, who cannot all have lived on the tell. It is a reasonable assumption that they lived on or near their sustenance fields, probably flocking together in subcentres more or less evenly distributed over the dunnu's property. For dependent deportees such rural habitation is proved for Kar-Tukulti-Ninurta (Freydank 1980:91), for the farmers and their families it is a natural choice. Thus archaeology (location of sites) and philology (number of dependents) converge in their expectation of rural subcentres. The fact that they did not show up in the survey remains to be discussed by the specialists (Lyon, this volume). Generally speaking, the invisibility of small and short-lived sites has a bearing on the use of surveys to establish population densities, especially in periods (like LBA) of declining cities and increasing rural habitation (Wilkinson 1998a:72). On average each of the seven (sub)centres (including Sabi Abyad) would be associated with an area of 1428 ikB (5.14 kmz), and accommodate 128 people (s'ilu& and farmers). The stables and stores missing at Sabi Abyad would lie in the country, and be operated from the (sub)centres. Intensified surveying, and excavation of Khirbet eshShenef could prove (or disprove) these deductions. On the tell the walled precinct of the dunnu houses the administrative and domestic staff. On the basis of the available space their number can be set at ca 60. The other 68 people would live outside the walls. In respect to the previous Mitanni period there is continuity and change. At Sabi Abyad, underneath the Assyrian "tower," a fortified mansion with the same ground plan was excavated, dated on the basis of pottery to the late 14thlearly 13th century, and in any case pre-Assyrian (Akkermans 1993a). This building is undoubtedly a dimtu, the Mitannian forerunner of the Assyrian dunnu. With its ca 484 m2 (the walled precinct was not yet in existence) it is small in comparison to the dimtu's in the Nuzi area (Jankowska 1969:243f.). This dimtu cannot have been the rural centre Sabi Abyad was to become under Assyrian domination, and must have been dependent on some nearby town. The obvious candidate is Tell Hamrnam, but the date of the Mitannian palace there is problematic. The excavator dates it to the earlier part of the LBA (1550-1350), but at the time all datable evidence that could be adduced for comparison stemmed from areas where Mitannian rule ended earlier than in Hanigalbat, its last stronghold. The catchment area of Sabi Abyad immediately defines the catchment area of the next MA settlement to themorth west, Tell Jittal, at a distance of 6.5 km; and from the catchment area of Tell Jittal that of Tell Sahlan can be deduced. The political status of Tell Jittal has not yet been established, but its size, form (Lyon, this volume), and presumed name (Dunni-Dagal) suggest that it was another dunnu, controlling less than half the area of Sabi Abyad, if the territory west of the Balikh did not belong to it. Since dunnu's normally fall under the administration of the local city, Tell Jittal probably forms a unit
with Tell Sahlan (S@lalu), 12,5 km upstream on the other side of the Balikh. Still further to the north, probably at or near Tell Abyad, lies another dunnu, Dunni-ASSur, which the important (military) functionary Sin-mudammeq calls "my dunnu" (Cancik-Kirschbaum 1996 no. 2:6). As a dunnu supporting an important state functionary it may have been independent of a town, or, as its name suggests, dependent on ASSur itself.ll In the same vein it can be proposed that Dunni-Dagal is dependent not on Sablalu, but on Tuttul, whose tutelary god appears in the name. In that case the location of the erretu of Tuttul (above 1) in the north would get a natural explanation. The geography of the region will be discussed elsewhere (Lyon, this volume, for the survey results). The size of the catchment area (36 km2) is partly based on deductions depending on a yield of 2.45 homerlika. As we will see below (below 5) this yield is relatively high, and it could be objected that a lower (average) yield would lead to the assumption of more land for the dependent d a j u farmers; the size of the farmland of the s'ilublu is based on a text from Kar-Tukulti-Ninurta, and independent of the local yield. Since extending the radius with only half a kilometer (to 4 km) would increase the available catchment area with over 1300 ha, the doubling (additional 912 ha) or even tripling (additional 1368 ha) of the sustenance fields of the farmers would be of no consequence for the geographical deductions based on the size of the catchment area. 4.
185
PRODUCERS AND CONSUMERS
The population of the dunnu consists of two types of dependents, the unfree s'iluhlu workers who receive rations, and the free-born Zlaju farmers who do not (above 1). The personnel lists do not always explicitly mark groups of ration receivers as 3lu@u, which they probably are, but if they are not, they are at least identical to them from the viewpoint of labor and consumption. The small number of administrative and domestic staff probably belongs to either of these two types; even if some of them were free Assyrian land owners, this would have little effect on the overall population, its production and consumption. The composition of the s'ilu@u workforce and its grain demand can be established from two different personnel lists written a few days apart in limu KaStiliaSu, which belongs to the Mannu-ki-Adad period of the dunnu. One of them concerns Kassites and a group of hostages with Hurrian names (T 96-21), of the other the subscript is damaged (T 96-20). polulation yearly ration T 96-20 Male adults Female adults Children Male adult hostages Totals without hostages:
T 96-21 Total
homer
kg
20 38 28 0
5 10 9 21
25 48 37 21
90 115.2 45.1 75.6
-
-
-
- -
86
+
45 = I 3 1 110
325.9 250.3
The owners seem to have had a residence in Tell Fakhariyah (see note 3).
5850 7488 2931.5 4914 21183.5 16269.5
186
F. A. M. WIGGERMANN
In the personnel lists the rations go by month, for adult men always 0.3 homer, for adult women always 0.2 homer, and for children usually 0.1 homer, but sometimes a little more.12 In Sabi Abyad one homer equals ten sutu, and one s i t u equals ten qii. The qii is set at 0.84, the traditional value, but it may be ca 19% larger (Powell 1984:33, 41-42, approximately one litre). The weight of one litre of barley is set at 0.74 kg (van Driel, this volume), assuming that it was cleaned (as in fact it was when stored, cf. van Zeist 1994:545f.);uncleaned (as it was when stored in Dur-katlimmu, van Zeist 1994547) it would weigh much less (0.587 kg). The problem was recently discussed by van der Spek 1998:249f., who sets the value at 0.62 kg. In the calculations that follow 1 homer of barley equals 62.16 kg (rounded off to 62), based on a qii of 0.84, and a weight of 0.74 kg per litre. With van der Spek's value of 0.62 kg per litre all figures in kg would be 16% lower.
In a population mix as the one above the average monthly ration is 0.18 homer per person, and the average yearly ration 2.26 homer per person. The existence of two practically contemporary texts containing two completely different sets of personnel is a reminder of the fact that as long as we do not understand the administrative purpose of these lists, all uncovered extrapolation from single bits of evidence remains hazardous. Since all our other evidence stems from the later (a decade or so) Ili-pada/Tammitte period, we will use these texts only to supply the missing facts on the composition of the population of dependents. The few occasional hostages will be ignored. Population Use of grain persodyear Male adult Female adult Child
22% 43.5% 34.5%
36% 46% 18%
B 3.6 homer = 234 kg B 2.4 homer = 149 kg B 1.2 homer = 74 kg (rarely more)
Among the texts from Tammitte's office at the reception court there is a large number of fragmentary personnel lists, which belong to at least three and possibly more tablets. The largest joined text (T 98-45) contains six columns of about 60 lines, but the upper and lower parts ire still missing, and the total number can be estimated at somewhat over 400. Beginning and end are not preserved, but the subscript is supplied by a 'duplicate' (T 98-45+): PAP 422 BRIN.MESs'i-l[uh-lu], "altogether 422 iiluhlu workers." In view of the fragmentary state of these texts, the composition of the population is less easy to establish, but looks about the same as that of the Mannu-ki-Adad personnel lists. On the basis of the Mannu-ki-Adad figures on population and consumption, the following can be deduced for the Jli-pada/Tamrnitteperiod: share Male adult 22% Female adult 43.5% Child 34.5% Totals l2
number
yearly barley consumption
92.84 183.57 145.59
334.8 homer 44 1.6 homer 175.2 homer
422
For I-_Iarbusee Kiihne, C. 1999.
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
187
When we now return to thepiierti karu'e text (T 98-115, above 3), we get a grip on the 844 homers of grain reserved for the (unmentioned) iiluhlu workers. This figure implies a mixed iiluhlu population of ca 375 persons, quite in the range of the 422 of the personnel list. The mixed character of the population is indicated by the fact that the iiluhlu.receive rations for a whole year, which means that they are permanently settled. That these people are from six different towns probably has to do with their legal status, and with the way they are distributed by the crown. The same phenomenon is encountered in the piierti karuJe texts found in ASSur (Freydank 1994 and 1997), and in the personnel lists from Sabi Abyad: 16 workers from Sab[lalu] and a group from Harran (T 98-136), 15 workers from ASSur (T 98-45). A text from the earliest Assyrian level records the arrival of 11 workers from ASSur "as deficit of the i i l u ~ l u "(T 93-1). An envelope of the Mannu-ki-Adad period (T 96-23, Text 9) reads: "tablet concerning the possessions (d-nu-te) of the Subaraeans from Sadikanni ($[a S]u-ub-re-e ~ ~ i a - d i - k a - n a ie-e)," undoubtedly deportees accompanied by a tablet like the one from Kar-TukultiNinurta adduced already several times (Freydank 1980). A group of badly broken grain accounts (T 98-38, T 98-44, T 98-105; Texts 3-5) mentions amounts of barley in the same range as those of T 98-1 15 as "in charge of PN from GN." The one in charge of a group of workers is sometimes recorded in the personnel lists (T 98-40). In judging the productivity of the iilu& work force it must be kept in mind that the population is mixed (only 22% male adults), and that it has to produce additional income on its own sustenance fields. A closer look at the s'iluhlu workers gives an impression of their life style and its implications for the Sabi Abyad countryside in the LBA. An adult male iiluhlu receives as yearly rations 3.6 homer of barley, or 234 kg. Recent studies suggest that 65% of a person's caloric needs were satisfied by grain, which for a male adult amounts to 230 kg. yearly. If his diet consisted of barley alone, he would go blind, get scurvy, and eventually die due to lack of vitamins A and C, but in the hypothetical case that it did it would amount to 352 kg yearly for a male adult (van der Spek 1998:248f. with literature and discussion). These figures show unequivocally that a s'ilublu could not survive on his rations alone, and that the additional sustenance fields which were postulated above on the basis of the Kar-Tukulti-Ninurta text (above 3) are a bare necessity. In fact, with the exception of Nuzi where they are the same, the MA rations are half those distributed to dependents anywhere else in the Ancient Near East (Milano 1989:224, 234f., Fales 1990:28ff.).l3 On their sustenance fields the iilulzlu produce an additional 0.13 homer of barley per average person monthly (above 3), that is 0.20 per adult male, or 2.24 homer (149 kg) yearly. Together with his rations the adult male would dispose of 383 kg of barley, just over the hypothetical 353 kg of an all grain diet. Although blindness (or bad vision) is common among populations of dependents (inflicted or through bad nourishment, cf. Freydank 1985:233, Garelli, Charpin, Durand 1982), it is rare at Sabi Abyad (T 98-136b Side A i 2', IGI.NU.TUK), and there is no reason to assume that the s'iluhlu ate only grain. In other words, the fact that a s'iluhlu produced or could produce a certain amount of grain does not necessarily imply that he consumed it himself; thus the 2.24 homer (149 kg) must be viewed as an indication of the
951.6 homer (59549.2 kg) These problems are also discussed by Kiihne, C. 1999, in relation with the personnel lists from Uarbu.
188
189
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
value of his sustenance field, not of the way it was actually used. After the circumstances the Siluhlu could produce grain or other crops, could consume his products, or barter them for others. Below we will see that dependents had further resources at their disposal, sheep, goats, oxen, pigs, and donkeys, which supplied them with additional meat, milk, wool, hair, and power. As a workforce the Siluhlu surface in the administrative record at two points, harvesting and ploughing; these are best treated in connection with the other class of dependent agricultural workers, the farmers. The officials responsible for the grain production in the pi3erti karzi'e texts from ASSur are the rub ikkcirate, the "chief farmers" (Freydank 1994). The land they supervise is crown land, and they are employees of the state, who apparently receive rations (below 5). Freydank deduces that such crown lands are to be found in conquered territories, and on state-developed farmland in the Assyrian core (Freydank 1988:83). This is exactly what Sabi Abyad is: a state developed agricultural production centre, owned and exploited by the grand vizier and king of uanigalbat himself. The "chief farmers" of the ASSur texts till standardized plots of crown land of between 100 and 300 ikii, most being 200 ikii (Freydank 1994). If this state of affairs is applied to the dunnu Sabi Abyad, its ca 2000 ikli land yearly under cultivation would require between 7 and 20 "chief farmers," 10 if each worked a plot of 200 ikli. Sabi Abyad has not yielded direct evidence on the number of its chief farmers, but decisive clues can be obtained from a text recording the distribution of sickles CT 96-16, "duplicate" T 96-17, Texts 6 and 7). From the fact that the text is made up and kept by the administration, it follows that the receivers are dependents of some sort. Another text shows that it was the chief steward (abarakku) who distributed the sickles (cf. MARV 3 67, abarakku rabii owns axe used by gardener), and makes explicit that they were meant for harvesting (T 96-3, text 8; all these texts belong to the Mannu-ki-Adad period). In the distribution text seven people receive 50 sickles each, and another eight receive smaller amounts, totaling 434 sickles. That the people receiving these sickles are the looked for state dependent "chief farmers" is proved by T 98-33 (Text 1, translated above I), in which the same Resiatri who receives 15 sickles in T 96-16:14 is explicitly called "chief farmer," and expected to settle his accounts with the administration at the end of the year. Confirmation for their status as chief farmer (rather than farmer) is supplied by a text from ASSur, in which somebody owes 50 harvesters to a "chief farmer" (GAL ik-kara-te, KAJ 91, see Postgate 1988 no. 54). The obvious assumption that the standard amount of 50 sickles is distributed to chief farmers working standard plots of 200 ikii (72 ha) is born out by the resulting total: 7 chief farmers work 1400 ikii, the other 8 (receiving in all 84 sickles, B 4 ikii per sickle) work 336 ikli, totaling 1736 ikG, quite in the range of the 1989 ikii deduced from the quantity of seed corn in T 98-1 15 (above 3): only 253 ikii short. With 100 or 300 ikli per 50 sickles the result would be either too small (868 ikli) or too large (2640 ikli). In the particular year of T 96-16/17 the bulk of the grain (SO%, 3840 homer) is produced by 7 chief farmers; the other 8 produce the remaining 20%. Clearly not every chief farmer worked 200 ikii each year, which may mean that he was involved-in other agricultural tasks, or simply was a minor chief farmer, who never cultivated 200 ikii. Moreover, in a text from the same period (Mannu-ki-Adad) as the sickles text (T 96-3, Text 8), a certain IgarSErnid receives 20 sickles, on account of which he is probably a chief farmer, one that does not occur in the sickles text. The exact number of 100 farmers
working under the chief farmers suggests yet another solution: 10 chief farmers, each standing over 10 farmers, work each 200 ikii. In the hierarchy the chief farmer would take the same position as the ten-men (rub e3arte). With the present information, the question cannot be solved. It is a reasonable assumption that the 434 harvesters handling the 434 sickles are the same s'iluhlu workers as the 422 of the personnel list and the 375 of the piSerti karzi 'e text (both Ili-pada/Tamrnitteperiod). The sickles can give us an idea of the productivity of this workforce. Each harvester reaps 4 ikii (200:50), ca 1.44 ha (rounded off to 1.5). Akkermans 1993c:216ff., discussing ethnoarchaeological experiments and modern evidence, concludes that the "harvesting of one hectare of cereals requires about 5 or 6 days' work when using a metal sickle and when working long days (15 hours)." Assuming a three week harvest season (so Akkermans, differently Wilkinson 1994:495f.: 3 haharvester assuming a harvest season of 2 months) and a mixed Slu& population (only 22% male adults) that also has to harvest its own fields, the 1.5 haharvester deduced from the text seems reasonable. Besides the s'iluhlu and the chief farmers, there is another group of agricultural workers, the farmers. Two texts of the Ili-pada/Tarnmitte period, apparently duplicates but neither of them completely preserved, contain a list of 100 men, all with their father's name added. This, and the fact that practically all names are good Assyrian, distinguishes them from the Siluhlu (never father's name, usually foreign names), and confirms their social status as free-born dependents (above 1). In the lists each man receives two different amounts of grain, the first between 0.6 and 1.8 homer, the second between 0.3 and 1.8 homer. At the end the amounts are added up as "100 [+x homer of seed corn]," and "90[+x homer of barley], rations for the mu-x-[x-x] workers." That the first amount concerns seed corn goes forth from a similar text from Dur-katlirnmu (Bonatz 1998:111 no. 105; this text does not have the second amount). If, as usual, the rations are for one month, these men would dispose of some 500 workers, that is the whole Siluhlu force (mixed Siluhlu get 0.18 homer monthly). Assuming that only the adult males plough the available 2000 ikli of land cultivated yearly, their productivity can be calculated. In a population of ca. 500 22% would be male adult, that is 110; a team of two (one for ploughing, one for sowing) would cover ca. 40 ikii. The ploughing teams would be assisted by oxen (Figs. 7 and 9). Unfortunately the p i h r t i k a r i 'e text from Sabi Abyad omits the fodder for the oxen used in agricultural work (above 3), but the figures can be supplied from one of the similar texts from ASSur; the relevant material from Dur-katlimmu is as yet unpublished. In the ASSur text 10 oxen work 200 ikii field, and receive the standard rations, that is 1 112 qii a day for 6 months, totaling 27 homers of grain (Freydank 1994:26). Compared to for instance Chagar Bazar, where the oxen receive 3 qli a day (van Driel, this volume 8 5b), these rations are low, so there must have been time for additional grazing. The dunnu with its about 2000 ikii yearly under cultivation would require 100 oxen, consuming 270 homers of grain during the half year of their employment. Thus a team of two oxen, the minimal number before a plough, would work 40 ikii (14.4 ha), exactly the same amount of land as that worked by two M.&lu. Plough teams of two oxen are attested textually in the NA period (Deller, Fadhil 1993:245 no. l), but I am not aware of any MA evidence. The seals (Fig.
190
191
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
7, 9) favour one ox, but this must be artistic convention, since one ox cannot pull a plough in a straight line.14 The area of 14.4 ha ploughed by a team of two men and two oxen compares well with NA evidence concerning the Harran region, where a team ploughs 10 to maximally (once) 15 ha (van Driel, this volume 5 5d). A problem lies in the about 120 homers of seed corn which the farmers receive, since it is about 116th of what is needed to sow 2000 ikii (663 homers of seed corn in T 98-1 15). A solution is suggested by the duration of rations for the oxen, which indicates a ploughing season lasting for six months. In that case texts recording seed corn and rations for men would form a series of six, and all seed corn would be given out to the farmers. In fact the two "duplicates" of Sabi Abyad could represent different months, but unfortunately of only one the date is preserved. The implications of a six month ploughing season will not be discussed here. That the seed corn was stored and distributed by the administration of the dunnu goes forth unequivocally from a letter sent by Mannu-ki-Adad while on duty elsewhere; the addressee is a woman, probably his wife, who takes care of business in his absence (T 96-36, Text 10). In the second part of the letter Mannu-ki-Adad orders her: "give out the seed corn of the chief farmers of the villages which they did not yet receive, and give out seed corn to Kurbinu, so that he can cultivate the field (in the district) of Sahlalu, which Sin-mudammeq himself showed him." In Dur-katlimmu the official responsible for the (agricultural) workers is the zariqu (Rollig 1988:lO n. 15, Deller 1993b:237f.). A Sabi Abyad text listing various types of dependent personnel shows that several of them served the administration of the dunnu (T 98-33, r. 8'). The 100 farmers would be accompanied by wives and children; the size of their families has been set-at 4. Contrary to the s'ilublu, where a family unit may be headed by an adult female, all farmers' families are headed by an adult male, and as the lists show, it is he who is responsible to the administration. The same is implied by the Middle Assyrian Laws (A 5 45), where an d a j u woman whose husband is taken prisoner is not allowed to head the farm by herself. Thus the population mix of the farmers differs from that of the s'iluhlu, and can be estimated as follows: male adults: 25%, female adults: 25%, children: 50%. Especially the share of female adults would be much lower than in a s'ilublu population (43.5%). One group of dependents will not be discussed here in detail, the administrative and domestic staff, including specialists and artisans of various kinds. The staff is headed by the chief steward, Tammitte, who is assisted by ten-men (rub es'arte); it consists of potters, brewers, oil-pressers, builders, leather workers, bakers (alabinnu), perfume makers, hair dressers, male and female singers, dress makers, probably a smith (tin is imported), merchants, gardeners, and shepherds (sheep, oxen, donkeys). Apparently there is a cult of ASSur nearby, since a list adduces a number of "servants of the temple of ASSur" (T 98-40 r. If), and a contract is witnessed by a "junior scribe of the god ASSur" (T 97-5, dated limu Sunu-qardii). Part of these people belongs to the s'ilublu population, the others would receive rations, and/or work sustenance fields. The highest member of
the staff, Tamrnitte, apparently disposes of a sustenance field, since he employs at least one farmer (T 98-58). By and large the domestic and administrative staff can be identified with the people living inside the walled precinct of the dunnu. One half of this area is reserved for reception and representation, a large part of the other half is occupied by the tower, so that at the most some 60 people could have lived there. Now the total population of dependents can be estimated: Eluhlu workers receiving rations and working sustenance fields: ca. 400 (422 in a list, 375 implied by the piierti kaniJe text, 434 harvesters). Most of them are agricultural workers, some are specialists. d a j u farmers working sustenance fields: ca. 400 including families. Their work is coordinated by at least 10 chief farmers, with families: 40 people. administrative and domestic staff, receiving rations and/or working sustenance fields: ca. 60 (including families). Total: ca. 900 souls. Earlier in this paragraph the amount of grain available to the dependents was calculated to be at about survival level. A text from Kar-Tukulti-Ninurta, however, shows that dependents (in this case Hurrian deportees) had other resources at their disposal (Freydank 1980; cited also above 3 in connection with the size of their sustenance fields). The text concerns the possessions of 25 families of different composition and varying wealth, and thus gives a fair picture of the means of the average family of rural dependents: Family size: between 3 and 24; total number of people: 239. Some families employ between 1 and 4 slaves, total: 13. The same phenomenon is encountered at Sabi Abyad (T 93-51 iii 17') and at Harbu (Kuhne, C. 1996:7: Elamite workers receive female prisoners of war). Arable land: between 5 and 40 ikii, total: 242 ikii Bovids: cows: 23; young bulls: 3; full grown bulls: 14 (presumably most of them castrated); total: 40 Ovids: sheep: 255; goats: 442; unspecified: 221; total: 918 Pigs: 28 Donkeys: 6 Other possessions: most families own chairs (between 1 and 4), tables (between 1 and 2), and beds (between 1 and 4); some families own wagons (total of 3), and very few bronze objects like a cauldron, an axe, or a knife. In order to get an impression of the situation at Sabi Abyad all figures must be multiplied by 3.5 (900:252); the arable land has been treated above, and will be left out of consideration:
l 4 That one ox cannot pull a plough is conventional wisdom in Assyriology, but certainly not universally true, cf. the photographs of European and Asian farmers ploughing with one animal in Netting 1993:145ff.
Living of steppe
Living off settled land Bovids Bulls People 49 91 900 0.2 Family of 4 0.4
Donkeys 22 0.1
Pigs 98 0.4
Ovids 3213 14
We will refrain from calculating the contribution of these animals to the diet and labor (castrated bulls, donkeys) of the dependents. What is clear at first sight, however, is the
192
F. A. M. WIGGERMANN
important role of the steppe in the economy of the settled community. The ovids would supply the families with meat, milk and wool, and in part insure them against the occasional bad harvests. On the basis of the calculated catchment area, the distribution of sites, and the number of dependents, the existence of 7 subcentres (including Sabi Abyad) was deduced (above 3). It is now possible to make a more definitive proposal as to the social status of the inhabitants of these subcentres. It is a reasonable assumption that the 7 major "chief farmers," the ones working fields of 200 ikfi, are the main inhabitants of the 7 subcentres; the eight minor "chief farmers" would be associated each with one of the major ones. Textual support for the idea that the chief farmers live in rural villages comes from T 97-36 (Text lo), in which Mannu-ki-Adad orders his wife to give "seed corn to the chief farmers of the villages (URU.DILI.DILI)," who are obviously under the administration of the dunnu. The farmers and other dependents such as iiluhlu and shepherds would live close by, in habitations that could be called a dunnu. In the letter T 97-17 (Text 11) Tammitte orders the chief farmer IgarSemid (encountered in text 8) to deal with the cattle of Sarniqu, apparently a farmer under his charge; he continues: "they should remove the roof beams of his farm house (dunnizu), so he can take them away; he can take away all his private possessions, and everything that belongs to his house." This man is moving out, or rather reassigned to another location, and wishes to take his belongings with him. That the permission of Mannu-ki-Adad is required shows that he and his dunnu are dependents of the central administration, and that his habitation was located in the territory of Sabi Abyad. People of Hurrian origin probably flocked together sometimes. In the letter T 937(Text 12) Tammitte writes to his two assistants, the scribes Bal(a)tu-[Sari?] and BelueriS: "The oil-pressers must dehull their sesame. If there is no sesame in your dunnu, go to the dunnu of the [Sublaraeans, and give out sesame to the oil-pressers there, so they can dehull it. They should not be idle, but dehull sesame. It is not necessary that they press oil; dehulled sesame is requested to be supplied, and they must quickly dehull it." The dunnu of the Subaraeans clearly belongs under the authority of Tammitte. Such an ethnic group must have retained parts of its identity, and possibly produced its own pottery to its own taste. Thus certain types of non-MA pottery could in fact be contemporary ethnic ware, which has a bearing on the interpretation of the survey (Lyon, this volume). For the sake of future archaeological proof or disproof we will specify the contents of a model subcentre, or "village" (ah): 1 major "chief farmer" with family; 1 minor "chief farmer" with family; 14 families of daju ("villager") farmers; 14 families of $iluhlu workers; some other dependents, such as the shepherds; total: 30 families of 4, and some other dependents: 128 people (above 3); animals: bovids: 12; ploughing oxen for their sustenance fields: 6; donkeys: 3; pigs: 12; ovids: 420; habitation for 30 families, stables for animals, especially oxen; stables for animals owned by the central administration, especially the ploughing oxen and the horses:
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
193
storage for their own grain, and for grain owned by the dunnu (below 5; 1000:7 m2 = 163 m2). It is likely that certain geographical names that occur in the texts in fact denote some of the subcentre villages. The study of these will be postponed to another occasion. 5.
YIELD AND SURPLUS
On the basis of the pizerti kGruJe text treated above (3) it could be established that in a certain year there was 1989 ikfi of land under cultivation, which yielded 4873 homer of barley, that is 2.45 homer per ikti, or 421 kgha. The seed codyield ratio is 1:7.35. In comparison with modern Near Eastern yields Sabi Abyad's 421 kgha lies in the range of bad years. Akkermans, discussing modern wheat yields in dry-farming Syria, cites averages of 689 kglha for 1949-1966, and 753 kglha for 1967-1976 (varying between 438 kglha in a bad year, and 1238 kglha in a good year); he decides that for prehistoric Sabi Abyad 600 kgha is an appropriate estimate (Akkermans, 1993c:214). For barley cultivated in the dry-farming zone of Northern Iraq Butz 1985:179 cites yields between 610 and 1240 kgha for three years between 1984 and 1957 (see also Zaccagnini 1975224f.). Compared to the results of contemporary dry-farming (Fig. 8) Sabi Abyad's seed codyield ration of 1:7.35 can be called good. In Dur-katlimmu, the only location with evidence for a series of years, this relation fluctuates between 1:1 and 1:9, averaging 1:3 or 1:4 (Kiihne 1990b:20, after Rollig 1987). With ratios of 1:9 and 1:7.2 the yields of Nemad-IStar fall in the same range as those of Sabi Abyad (Fig. 8). This place lies near Tell ar-Rimah (Freydank 1994:20 n. 19), and apparently in a well watered area (Postgate 1987a:128) within the dry-farming zone. More moderate are the results in IjiHSutu (Freydank 1994:15 n. 9, location unknown), outright bad are those of TarbaShe (TarpaSbe(na)) in the Nuzi area (Freydank 1994:23 n. 26, Miiller 1994:110). There is a wealth of information on this area for the preceding LBA I, when Nuzi fell under Mitannian rule; the yields of barley fluctuate between 1:1 and 1:8, sometimes 1:10, those of emmer fall in the same range, those of wheat are distinctly lower. The higher yields are obtained on irrigated fields (Zaccagnini 1975, 1990). Generally speaking the results of Ancient Near Eastern dry-farming are comparable to those of dry-farming in Medieval Europe (Zaccagnini 1975:217ff.). If Butz 1983:482 is right with his remark that without manure 1:7 to maximally 1:10 is the best obtainable, MA dry-farming is close to realizing its potential. While in a comparison based on yieldlarea (kglha) modern Near Eastern results clearly surpass those of LBA, the difference is less marked when the ratio seed codyield is made into the basis of the comparison. Butz 1983:483 cites evidence from Egypt, Palestine, Italy and modern Iraq, which shows a ratio 1:7.35 to lie in the range of low to normal. The difference is caused by the higher seeding rate of modern agriculture (Butz 1983:483), or, in other words, by the more intensive use of the available farmland. Manuring is probably the decisive factor. Although the relation seed cordyield may serve as a measure for the relative success of MA agriculture, it is not a realistic assessment of the relation costslprofit. According to the MA sources the production costs consist not only of seed corn, but also of labor, that is rations for workers and oxen (above 3). In Sabi Abyad the costs amount to 1777 homers of barley: seed corn 663, rations for men 844, and for oxen 270 homers. Thus the
194
195
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
cost per ikii rises to 0.89, which means that in order to pay for the additional costs the relation seed cordyield must be minimally 1:2.75 (0.30:0.89), in Sabi Abyad amply surpassed by the result. That MA state farming did not avoid the high risk zones of rainfed agriculture is shown by the fact that elsewhere in the empire, at Tarbaghe and Dur-katlimmu, results may drop to and below this critical border (Fig. 8). According to the native calculation of the costs the profit per ikii in Sabi Abyad is 2.45-0.89 = 1.56 homer, or 177%. The native calculation, however, adduces as the cost of labor only the "wages" paid directly to the workers in the form of rations. The sustenance fields, which generate the rest of the workers' "wages" do not appear on the balance. If they did, these fields (884 ikii in Sabi Abyad), which produce zero grain for their owner, would lower the yield per ikii to 1.69 homer, thereby halving the profit to 0.80 homer per ikii, or 88%. The "native" costs then serve a practical purpose, and are concerned only with the grain that needs to be minimally in store to cover the grain expenses of next year's cultivation. The comparison with modern data has shown that current agriculture is land intensive, or, in other words, that MA agriculture was land extensive. This is probably related to the low labor intensity that can be detected in the ancient sources. Basic to the Assyrian system of cereal production on crown lands is the simple fact that one family of workers can produce more than twice its own upkeep; it is their spare labor that is extracted to generate a surplus for the crown on crown lands. A good estimate on the part of its time a workers' family spends on the production 6f its own cereal needs can be obtained from the relation between the number of ikii it tills for itself, and the number of ikii it tills for the crown (above 3). The ca 400 farmers dispose of 634 ikii sustenance fields, the ca 400 filu@u would dispose of about as much, were it not that they receive rations. In fact, however, they produce the rations themselves, not on sustenance fields, but on crown land. This means that the difference between the 250 ikii they actually work and the 634 ikii they would work to generate a complete income (384 ikii) must be subtracted from the crown lands on which they earn it (1989-384 = 1605 ikii). This gives the following relations:
due to the current state of agricultural knowledge, or to pressure exerted by the Assyrian government cannot be decided. 'In any case, as long as farmland is plentiful, which it is in this period, low labor intensity poses no problem; the only limitation is the increasing distance to the centre, which will not be discussed here (Wilkinson 1994:497; in the LBA at least in part resolved by increased rural habitation). he labor and land extensive nature of MA agriculture and the availability of farmland sufficiently explain the by modem standards low kgha yields. Once this point has been cleared, and the suspicion is removed that hidden factors lower the yield figures, another question can be posed, that of the meaning of t d i t ebiiri. Above we have translated this phrase with "result of the harvest," which implies that the whole harvest is involved. The idea that this phrase denotes the complete harvest originates in the piferti kari'e texts: "wie interessant, ansichtig und sinnvoll wiiren die durchweg berechneten Durchschnittsmengen je ikii fur den oder die 'Beauftragten,' wenn sie nicht ein reales Bild von der Bodenfruchtbarkeit vermittelten, also den unverminderten Ertrag bezeichneten. Nur dieser eignete sich schlierjlich zum Vergleich mit der Saatgutmenge, wenn auch diese ... je ikii angegeben wird" (Freydank 1994:14 n. 7). The remaining problem concerns the sustenance of the producers, the farmers (Freydank 1988:83). On principle there are three solutions: either they took their sustenance from the harvest before the tdit ebiri was established, or they received rations, or they produced their own upkeep on sustenance fields. It is extremely unlikely that the farmers took their sustenance from the harvest, not only on account of what was said above about the meaning of tdit eburi, but also because it must be asked: "Sollte jemand auf den in staatlicher Regie bearbeiteten Feldern das Recht zum Zugriff auf einen festen Prozentsatz eingeraumt gewesen sein, ohne dal3 dieser soweit irgendwo greifbar wird?' (Freydank 1994:14 n. 7). If, on the other hand, they received rations like the filuhlu, these should appear among the production costs in the piferti karu e' texts, which they do not; more generally, there is no evidence that farmers received rations, not from Sabi Abyad, nor from anywhere else. This leaves only one option: the farmers produced their own upkeep on sustenance fields. It is this solution that was adopted above (see also note 4). The lands under direct control of the dunnu produce in a certain year 4873 homer (302,126 kg) of barley. After deduction of next year's production costs (seed corn, rations for filu&; oxen not included in T 98-1 15) 3366 homer (208,692 kg) remains as surplus (ribtu). Apparently a certain part of this surplus is earmarked as reserve. This goes forth from the 7393 homer (454,366 kg) "old barley" which is still in store from previous years. A similar relation between "new7' and "old" grain occurs in MARV 1 23, where the city HiSSutu stores 681 homers of "new grain" with 2612 homers of "old grain" left over from 2 (or perhaps 4) previous years (Freydank 1994:13ff.). It may also happen that the harvest is only just large enough to cover next year's production costs, and that nothing goes into storage (Freydank 1994:22ff.,concerning Tarbagbe in the Nuzi area). The Sabi Abyad text T 98-115 adds the amounts of new barley and old barley (10759 homer), and marks them as "stored" (tabki). Since next year's production costs are not yet spent and go into storage too, the term tabaku has here and in similar texts a more limited, technical meaning: "to store as reservelsurplus." At the beginning of the season the total amount of barley stored is 10759+663+844=12266homer (ca 1,000,000 liters). Inside the walled precinct of the dunnu there is no room for storage of this size, and we must assume that, as in Nuzi, a certain amount of barley was distributed over
land tilled by workers for their own needs: land tilled by workers for the crown: total:
1268 ikii (634+634) 1605 ikii 2873 ikii
The relations between the surfaces tilled indicate that a workers' family needs about 317th of its time (or labor potential) to produce its own needs (at survival level), and that the remaining 417th are extracted by the crown to create a surplus. To have workers produce their own food under the threat of starvation undoubtedly pressurizes and increases the productivity of these unwilling subjects; compared to workers that are fully rationed, however, it decreases their availability. Among legal, political, and administrative reasons availability probably explains why about half of the workforce (the filu&) is partly rationed. The fact that the workforce spends about 317th of its time to produce sufficient income shows that agriculture as practiced in this period is characterized by a low labor intensity. This means that certain labor consuming measures such as watering, weeding, and especially manuring were insufficiently applied, or altogether omitted. Whether this is
196
F. A. M. WIGGERMANN
storehouses in the countryside; on the tell itself considerable quantities of barley and wheat were found in the tower, the residence, and the building outside the walls. Two amounts are yet to be subtracted from the surplus: fodder for the oxen (270 homer), and storage losses, on which Akkermans 1993c:215 comments that "contemporary evidence from the southern Mesopotamian plains suggest at least 25%." This means that at the end of the year at most some 2500 homer remain of the "new barley." If nothing of this barley were spent, and if the yields were constant, it would take at least 3 years to amass the 7393 homer of "old barley," and certainly more, since each year further storage losses would occur. It can be shown, however, that part of the surplus was spent each year, so that, depending on the amount, it would take several more years of dependable yields to amass the "old barley;" if about half is spent each year (1250 homer), it would take some 7 years. Thus the "old barley" implies a series of at least 3, but probably more years of dependable yields, which above was taken as an indication of (additional) irrigation. The reserves were probably meant as an insurance against the unforseeable, like serious draught, enemy action, or locusts (CancikKirschbaum 1996 no. 2: l4f.). It remains to be established how the free part of the surplus was spent. The evidence is sporadic, and never sufficient to estimate the costs over a whole year: provisioning ASSur: the provinces send small quantities of barley (between 250 and 19.1 homer), honey (between nearly 2 and 0.36 homer), sesame (between 18 and 1.75 homer), and fruit (between 9 and 0.7 homer) to the ASSur temple in connection with the gini'u tax (MARV 2 21). Not every province sends all four articles, and the "province" (better: "territory in charge") of Ili-pada sends barley instead of fruit, at least in one particular year (Freydank 1992:287f.). provisioning the residence of Ili-pada: Ili-pada orders cress (100 homer, T 93-6), clothes, oil, and spice plants (T 96-1, T 97-43), but barley is not mentioned. provisioning the army: the dunnu employs horses, and at one occasion imports 15 of them (T 98-35, T 98-73). Horses are economically unproductive, and with a daily ration of 0.05 homer great consumers of grain. The 15 imported horses alone would consume yearly 275 homers of barley, and they were hardly the only horses. The dunnu has cavalry (ia pitballi, T 93-2) in its service, and produces war chariots, which, presumably fully equipped, may accompany Ili-pada (T 93-10). There is evidence for cavalry men (T 96-7) and chariot teams (T 98-12, T 98-41) being provisioned by the dunnu. The way in which such troops were used can be guessed from the police action along the Balikh described in a letter from Dur-katlimmu (Cancik-Kirschbaum 1996 no. 2). rations for personnel besides the JiZu& especially the domestic and administrative staff housed in the walled precinct of the dunnu. If all of them (estimated at 60) were fully rationed, this would amount to ca 200 homer yearly. It seems that war and policing lean heavy on the budget, but the exact amounts cannot be established. The voracity of a Neo-Assyrian army, undoubtedly larger than a MA one, can be gauged from the 3,849,000 liters of barley monthly availabIe to the king and his troops in Kar-ASgur (Fales, 1990; see also MARV 1 9, cited below 6).
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
6.
197
MAJOR AND MINOR CROPS
Besides for barley and wheat (attested archaeologically) there is some unquantifiable evidence for other crops. The relative importance of the various foodstuffs can be gauged from MARV 1 9 (Postgate 1985:148), a list of contributions to the Assyrian army active on the south-eastern border: grain (products) 351.47 homer (88%), cress 14.64 homer (3.5%), sesame 34.80 homer (8.5%). If the separate contributions of sesame at the end of the text are included (totaling 209.40 homer) the percentages are: grain (products) 57.5%, cress 2.5%, sesame 40%. Postgate 1985:147f. compares the percentages of the gina'u offerings of Arbail and Katmuhu: barley 93.6/92.6%, sesame 3.6/2.78%, fruit 2.1/3.65%, honey 0.65/0.96%. These figures show that cress, sesame, fruit and honey are much less important than grain. cress (kudimmu, St01 1983-1984):the largest quantity is 100 homer requested by Ilipada (for his residence) because "there is no cress for the meal (ana iikulte)" (T 93-6); a comparison with the 14.64 homer delivered to the Assyrian army points to a considerable number of eaters. Small quantities occur between garden products (T 98-1 17,0.7 homer, T 98-127, 0.3 homer), and Urad-kube, away on duty, requests some because "there are no spices ([ralqfitu)" (T 96-11, besides onions and lentils). sesame (Postgate 1985:147ff.): the seeds identified at Sabi Abyad are the oldest archaeobotanical evidence for the cultivation of this crop in the Ancient Near East (van Zeist 1994:546, Bottema and Cappers, this volume). It is a summer crop that needs irrigation, and is not salt tolerant. The largest quantity is 20 homer, dehulled and brought in by three different persons (T 98-1, with envelope T 98-2). Normally it would be dehulled and pressed at the dunnu (T 93-7, above 3), and delivered to Ili-pada (T 93-11, T 97-34). fruit (Postgate l987a: l28ff.): only a small quantity ('O.ll homer) of terebinth nuts (Stol 1979:l-16) is attested (T 98-117, among garden products). It may be no coincidence that the "territory in charge of M-pada" sends grain to the temple of ASSur instead of fruit (Freydank 1992:287f.). honey: T 98-92 is a contract with two merchants (one of them seals with a signet) for the delivery of 0.25 homer honey; it was presumably not produced on the dunnu. vegetables: 1 qi2 (0.84 liter) lentils (abiu; Postgate 1987b:94f., CAD s'uJu disc. section) is requested by Urad-kube for his own use (T 96-1 1, besides cress and onions); onions (Stol 1987, Postgate 1987b); ihm-ki e-li (T 96-11); rtlabiu (for antabiu "fennel"; T 98-1 17); iu-Li (chick-peas; T 98-117). spices and garden products (cf. Zaccagnini 1975:119ff.): an inventory lists small quantities (not more than 0.7 homer) of coriander (kisibirru), cress, iamuttu, firi-qu, enu-ni-h, cumin (kamzinu), black cumin (sibibiinu), ' t l a ~ ~ and u , terebinth nuts (bumitu) (T 98-1 17), another text (T 98-127) adds samidu.15 A merchant receives coriander and cumin as small wares ([ana salbirte, T 98-73). Ili-pada requests spice plants ((~~S)SIM.MES) for his perfume makers (T 96-1, T 97-34); they may be locally produced, or imported (T 98-63, gigri-qi-Liamong the possessions of a merchant from Sidon). For further evidence on MA spices see Rollig, Tsukimoto 1999, with four new texts concerning gardens and garden products in Dur-katlimrnu. In these texts e-nu-ni-a appears as naniu, possibly mint; the identification of the other spices remains uncertain; instead of SE.SES 'Bitterkorn' read S E . L ~'coriander' (kisibirru; same spelling in Sabi Abyad), see discussions ibid. 436f.
198
199
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
other: a merchant receives 405 kg "bark of the kiikanli tree" (sibpu) to buy an equal amount of tin (AN.NA BABBAR) (T 98-80).
bulls, but cows are not excluded (van Driel 1995:216, 219, 221, 224). If, as in the NeoBabylonian period, the bulls were castrated and trained during their third year, and employed in front of the plough from their fourth year onwards (van Driel 1995:225), a herd like the one published example from Dur-katlimmu (Deller, Tsukimoto 1985; cf. also Tsukimoto 1992:21ff.) would contain 32 (castrated) ploughing oxen on a total of 127 animals; the breeding bulls would have been kept separate. About the same relation (castrated) bulls : total population occurs in the herd owned by a number of deported Hurrian families (14:40; Freydank 1980:112 f.), and in a herd on a tablet from ASSur (84:439; KAJ 289). Depending on the number of cows, the 100 agricultural oxen of the dunnu imply a population of between ca 250 and 500 animals. In T 98-14 (also T 98-124) a total of 17 ox hides is delivered by 6 different people, who must have been herdsmen. If each of them was in charge of a herd of some 127 oxen, the size of the herd from Dur-katlimmu, the total number of bovids would be 762. If the herds were the size of that from AHSur (439), the total would be 2634. The number of hides delivered (17) seems too low, especially for the larger total, but it is not known what period of time is covered by the text. Besides for ploughing the oxen were used for threshing (Cancik-Kirschbaum 1996 no. 4:10), which suggests an emendation for T 9717:12: 3 G ~ D . M E Si a nam-iar(Text: K A R ) - ~"three ~ , oxen of the threshing sledge" (the text concerns farmers); and they were used in front of wagons (Cancik-Kirschbaum 1996 no. 4:9; T 96-9). Milk and meat are not mentioned in the texts from Sabi Abyad. equids: horses are used in war and policing, and economically unproductive (above 5). Mules were apparently held, since there is a $a kkdine, someone who herds, or takes care of mules (T 97-33). Donkeys are economically important, but badly attested. There was at least one ra3i emiiri, "herdsman of donkeys," who receives 23 donkeys of both sexes and various ages from Tammitte (T 98-96), rather an addition to his herd (cf. Deller, Tsukimoto 1985:325) than his whole herd. The largest number of donkeys are the 36 received by four merchants from Tammitte, undoubtedly to compose a caravan (T 9718, also T 98-76). A donkey pulling a cart (ia saparri) occurs in T 97-33, where it receives a (daily?) ration of 3 qli. pigs: pigs are not attested, only the wild boar (hunted ?), whose fat (T 93-10, T 9856) and skin (T 98-7) are used (above 2). Pork is mentioned once in an inventory of foodstuffs (T 98-3 1). In Dur-katlimmu there is archaeozoological evidence for both wild boar (sus scrofa, small quantities) and its domesticated variant (sus scrofa f. domestica): "Schweinenfleisch hat wohl regelmal3ig auf dem Speisezettel der damaligen Bewohner der Zitadelle gestanden" (Becker 1991:120, 123f.). Thus the lack of textual evidence for the keeping of pigs is coincidental,l7 rather than due to a taboo on it as in the NeoAssyrian period (Radner 1997:293; cf. van Driel 1995:233 for the Neo-Babylonian period).
7.
ANIMAL HUSBANDRY
The evidence on animal husbandry is sporadic. The basic sources for this period are the records of the census (maiartu) of oxen, donkeys and ovids, which took place each year on Hibur 20. There is a series of this type of texts from ASSur (oxen: KAJ 289; donkeys: KAJ 31 1, dated Qarratu 15) and Dur-katlimmu (Deller, Tsukimoto 1985, Tsukimoto 1992:21ff., cf. Rollig 1984:192), but none from Sabi Abyad. Another source that gives an idea about numbers are the accounts of the skins of dead animals (oxen, donkeys, ovids), of which there are two at Sabi Abyad (T 98-14, T 98-124). The skins are cooked and dyed (Rollig 1988:11, Dur-katlimmu), and then reappear in the record as issues to the leather worker, in Sabi Abyad usually in connection with the construction of chariots. There is no information on the consumption of meat or milk (products) or on wool and wool industry. A general notion of the numerical relations between oxen, donkeys, ovids, and pigs can be obtained from the possessions of Hurrian deportees listed on a tablet from KarTukulti-Ninurta (above 4): oxen 4%, donkeys l%, ovids 92%, pigs 3% (figures rounded off). Comparable are the relations suggested by Radner 1997:293f. for the Neo-Assyrian period on the basis of a debt contract: oxen 4.3%, donkeys 8.7%, ovids 87% (no pigs). Although these figures stem from the private sphere, and cannot simply be transposed on a state institute like Sabi Abyad where industrial interests (agriculture; army; trade) may play a part, they reveal the enormous importance of the steppe in the economy of the settled communities. The chasm between animals living off farmland and animals living off the steppe is so wide, that even doubling or tripling the number of oxen would not make much difference in this respect, so that we can safely assume that in every sector there was a sizable gap between oxen and donkeys on the one hand, and ovids on the other.16 The animals owned by the dependents were treated above (4, estimates based on possessions of Hurrian dependents). Here our interest concerns the animals owned or used by the dunnu. The numerical relations between the various types of animals can be estimated on the basis of the skins that surface in Sabi Abyad's leather industry (T 98-6, T 98-7, T 98-30, T 98-56): oxen: 5.6% (8.5); wild boar (iabapu): 2.4% (4); goats; 92% (141; sheep skins are not used). These figures, which do not include sheep, confirm for the dunnu what was deduced above for the private sector, the wide chasm between the number of oxen and that of sheep. bovids: on the basis of a text from ASgur the number of agricultural oxen employed by the dunnu was set at 100 (above 4). Most bovids pulling ploughs would be castrated l6
A preliminary assessment of the zoological material (1189 items taken from various locations inside the walled precinct, all dating to the Tammitte-Ili-pada period) gives a slightly different, but comparable picture (Dr. C. Cavallo, personal communication): ovicaprids 48.2%, equids (including donkeys, horses, mules, and perhaps onagers) 23.8%, sus 12% (see note 17), bos 4.1%. The remarkably high share of gazelle (7.5%) shows the importance of the hunt (ignored by the texts as administratively irrelevant), and again highlights the role of the steppe in provisioning the settled communities. Camels and dromedaries are not represented, which is revealing especially in view of Sabi Abyad's location on the NS and EW trade routes.
17 The preliminary investigation of the faunal remains revealed a similar importance of pork in the diet of LB Sabi Abyad (Dr. C. Cavallo, personal communication). The low visibility of pigs in the administrative record may be due to the fact that in the arid environment of the Near East "pigs are illsuited to large-scale centrally controlled production," so that "the optimal and perhaps only effective strategy available for taking advantage of this animal's meat-producing qualities" was "small-scale stybased management by individual ... households" (Zeder 1998:119). For occasional evidence on centralized pig keeping see Englund 1995 (Proto-literate period), van Driel, this volume: n. 85 (OB Chagar Bazar), Fincke 1993:361 (in the town Zizza in the Nuzi area).
200
20 1
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
ovids: T 98-34 is a list of ten shepherds (niqidu) in charge of someone who is probably their rub efarte, "ten-man;" the list is followed by the provision that "if they do not come back at the moment indicated to them by Tammitte, they will receive 100 blows (with a rod)." These shepherds will be away herding their sheep, to return only for the yearly census. One of the shepherds of the list is found in another document, T 98-43, in which he receives a total of 138 "assorted (ovids)" from Tammitte. Apparently not all animals in his charge were meant to be herded, since among them are 15 sheep f a u r i e' "of the stable," which means that they were being fattened by grain. Three shepherds which are not on the list of 10 receive quantities of sheep from Tamrnitte, but of only one the total is preserved: 549 ovids (T 98-91). A badly broken text of the same period is concerned with herds of small cattle (se-nu) of 260 and 300 animals respectively (T 9819), and at least one further unknown naqidu is involved. If 200 animals make a "normal" herd (cf. Radner 1997:300 for NA herds), and there are at least 14 shepherds, the dunnu disposes of at least 2800 ovids. That this is a very conservative estimate can be seen from the relation to the minimal number of oxen, 250, which implies about 20 times as many ovids, that is 5000. The skins of kids are used in the leather industry, and sometimes dyed green (dug& T 98-7, T 98-30; cf. Radner 1997:301, van Driel 1993:241); the skins of full grown goats and sheep are not used. The 10 kg hair of T 98-106 is probably goats' hair. Of historical interest is a text dated to ASSur-nirari III uklu (1 192-1187), one of the later texts of the Ili-padnammitte group (T 98-1 19). In this text the "governdrs of the land Harran" owe 164 sheep to Tammitte, and the debt is incurred "when Ili-pada came to the help of the king of Karkemish, and ... they were given them to eat." Apparently the dunnu fed a passing army of Ili-pada. A variety of sheep and goats "for the meal" (ana napteni [...I) is recorded in T 93-9; the text is seriously damaged, but the occurrence of the phrase "in the temple of Sin" suggests a relation to the cult. The "fat" (1) of sheep is attested in T 98-82. A merchant receives 3 rams "as small wares" ([ana sa]&rte, T 98-73), and a group of merchants receives besides 36 donkeys a total of 64 sheep (T 97-18) perhaps also as small wares, although it is hard to imagine a caravan being accompanied by a flock of sheep. Wool must have been important in the economy of the dunnu (for Dur-katlimmu cf. Rollig 1988:11, 14), but it is mentioned nowhere in the texts. There is evidence for a linen industry (T 97-34), but the flax may not have been cultivated at Sabi Abyad. The transport of flax is attested in letters from Dur-katlimmu (Cancik-Kirschbaum 1996, no. 7). birds: in a text recording the disbursement of small quantities of grain for various purposes it says: 0.03 homer "in front of the (female) ostriches" (pa-ni lu-ri-ma-te, T 9733), and it looks as if these birds were being kept and fattened (cf. MARV 3 3: 17 for the "fattener of birds," mufikil issurite). At least they were present in the neighborhood, since the remains of more than one egg have been excavated (for Dur-katlimmu see Becker 1991 124f., one small piece).
In T 98-77 the governor of Huziranu owes Buria (a chief steward before Tammitte) 6 doves, 4 turtle doves, and one mesukku-bird. The same birds appear in ASSur-niisir-apli's banquet (RIMA 2 292), and undoubtedly were meant to be eaten.18 8.
NATIVE VIEWS ON THE COSMIC FOUNDATIONS OF AGRICULTURE
Behind the visible reality of every day life lies a higher order, a cosmos created and steered by gods. Ultimately the success or failure of men's endeavour depends on what is beyond bis control, on divine decisions and their implementation. The standard expression of this higher order is mythology, transmitted orally (in literature) as well as visually (in art). Although the subject is complex, with a long history and manifold ramifications, two sources, a visual and an oral one, may serve to elucidate the divine world supporting the dunnu's agricultural enterprise. The first is an early Neo-Assyrian cylinder seal (Keel, Uehlinger 1990:24 Abb. 13; Fig. 9), some three centuries younger than MA Sabi Abyad, not too much for this highly conservative field of ancient thought. The subject matter divides the image into two registers, below the earth and the world of men, above heaven and gods. The earth is being ploughed by an ox (castrated bull) and a man, behind goes a second man who appears to be broadcasting seed corn from a sack. Obviously, and explicitly, the Mesopotamians understood ploughing and seeding as a sexual act in which the plough is the male, the earth the female partner (Jacobsen 1982:23, Cavigneaux 1999:258). The rhombus between the two men is a model of the female genitalia and a symbol of the goddess IStar (Wiggermann 1998:49 $ 5); it makes the sexual interpretation of ploughing explicit, and refers to the sacred marriage in the upper register. The upper register shows the higher order behind every day reality, and explains the divine roots of agriculture. The central scene is the meeting between two deities, IStar on the left, and Adad on the right. Their identities are revealed by the symbols behind them, the eight-pointed star of IStar, and the "impetuous bull calf' (bum ekdu) of Adad. The feeding of Adad's bull calf presumably symbolizes man's contribution to divine welfare, the delivery of first fruits to the temple. Adad presents IStar with three ears of corn, IStar's reaction is shown by the dove that flutters towards him, a well attested motif expressing female love (Keel 1977).l9
According to the bone material the share of birds in the diet was small (0.3%). The material from Tell Hammam shows that the rock dove (columba livia) and the turtle dove (streptopelia sp.) were common in the area. Another edible bird from Tell Hammam is the chukar partridge (alectoris chukar), at home in northern Syria, northern Iraq and Turkey, but not in southern Iraq. This edible and foreign bird is a good candidate for identification with the edible bird denoted by the foreign word mesukku, attested mainly in northern sources. The CAD bases its identification 'a bird of prey, possibly the falcon,' solely on the mesukku's iconographical role, and totally ignores its edibility; with 'Wildgans' (??) the AHw. incorporates edibility, but fails to explain why this common animal has a foreign name, and is practically limited to the north (zoological information courtesy Dr. C. Cavallo). l 9 The interpretation of the seal basically follows Keel, Uehlinger 1990:24. Von der Osten-Sacken 1999:266f. views the bird as an element of the ploughing scene, rather than as an element of the divine world above it. It is not necessary, however, to have birds in ploughing scenes (cf. the NA ploughing scene Porada 1948 no. 653), so that in principle there is no objection against its interpretation as a messenger of love. The type of bird (dove) and its position contrast with Fig. 7a (MA), where the bird does belong to the ploughing scene.
202
203
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
The rain god Adad is one of the most important gods of ancient Syria, and through time he is coupled to a variety of goddesses. Usually the name of Adad's wife is Sala, and usually it is she who holds the ear of corn, not her husband (Wiggermann 1998:51f. 5 7). The obvious relation to the fertility of the earth is made explicit by a late 2nd millennium astronomical source (MUL.APIN I ii 10): "the constellation Furrow is gala (who is) the ear of barley." On the seal the goddess does not yet hold the ear of barley, but she is going to as soon as she accepts them, that is if she consents to the act that will fertilize her, the sacred marriage. Thus both registers show different aspects of the same subject, below the sexual act of ploughing the earth, above the first steps towards the divine marriage which explains and secures the earth's fertility. With one further observation on the seal, the whole scene can be linked to textual sources. In the heart of the scene of the upper register, in the enclosed space between the two deities, are the seven dots of the Pleiades. They occur more often on seals of this period, but the prominent position they take here must have a special reason. From the following citation it will be clear that they indicate the season: "the month Ajaru, the Pleiades, the Seven Gods; the oxen are yoked, the wet grounds are opened, the ploughs are washed; the month of the warrior Ningirsu, the chief farmer (is's'akku) of Enlil" (Astrolabe B, Reiner 1981:80). The Pleiades denote the beginning of the agricultural season in MayIJune (Ajaru); the symbol of the god Ningirsu is the plough, and as such he is present on the seal as well (Wiggermann 1997:39 n. 55). Texts of the same period as the seal (NA) show that the divine counterpart of the beginning of the agricultural season, the marriage between Adad and IStar, was celebrated in the same month. The best evidence comes from ABL 1202 (Parpola 1970 no. 281 r. 24ff.), a letter from Mar-IStar to Esarhaddon: "[on the ...th] of Ajaru [the Laldy of Akkad (= IStar) will set out (in procession) (and) take up her residence [in] the akitu-[temple]," which must be combined with ABL 1197:5ff. (Parpola 1983:262): "on the 6th of Ajaru Adad will rise and take up his residence in the akitu." On the 13th of the month, during a ritual for IStar in the akitu temple, there are offerings to the Pleiades in the presence of IStar, and, among others, Adad (van Driel 1969: 89, NA ritual). According to another tradition there is a marriage between NabQ and Nanaja or TaSmetu in this month (Cohen 1993:31Iff.). While the seal effectively illustrates the current standard interpretation of fertility as due to a marriage between two deities, a myth sets out the elements of rural life in a highly unusual cosmogony. The actual manuscript of the text is Late Babylonian, but the colophon learns that it was copied from originals of Babylon and ASSur (CT 46 no. 43). The presence of the Hurrian gods Heaven and Earth (Hamurni and [HajjaSum]) indicate a date of composition in the LBA, the time of the Mitannian empire and the ascendancy of ASSur. According to a recent treatment (Jacobsen 1984, cf. Hecker 19943610f.) the myth reflects the interests of the herdsman, but, as we will see below, they are rather those of the farmer. Following a well known mythological scheme, the cosmos develops as a sequence of seven divine pairs, each producing its successor.: Harab, the soil breaking plough, and E r ~ e t u "Earth." , By ploughing they bring Tdmtu "Sea," into being, and the furrows give birth to Sumuqan. Then the first pair builds its residence, the primeval "dunnu." Dunnu here is supplied with the determinative for towns (URU), but that does not mean that it is the name of a town (so Jacobsen 1984:22, who thinks of the Old Babylonian city Dunnu in southern Babylonia), since also the generic dunnu "agricultural production centre" can be supplied with this
determinative (e. g. T 93-7, Text 12; T 97-17, text 11; T 96-20, T 96-21). It is obvious that a generic term fits the context far better than the name of a specific (and unimportant town), Znd that the meaning "agricultural production centre" or "farmstead" makes excellent sense. A cosmic dunnu founded on a mythological bagmu-dragon is shown on a kudurru-like monument from Susa (Seidl 1989 no. 40; LBA). On the earth above this cosmic dunnu goes a procession of seven lute playing gods and a goddess holding a tambourine; quietly walking beside them are seven wild animals of mountain and steppe (drawings MDP 7 [I9051 Figs. 461f.). Already Moortgat 1934:13 comments: "bei dem Ganzen denkt man unwillkiirlich an die bildliche Darstellung einer Kosmo- oder Theogonie, an die Wiedergabe eines mythischen Weltsystems." At least the unusually sedate animals contain a reference to primeval times (Wiggermann 1996:218 n.138; Fig. 10). In the myth each generation of gods is slain by its successors, and laid to rest in the dunnu. Just as the Dukug, "Holy Mound," and other elements of the primeval cosmos, the dunnu becomes a mausoleum, and part of the netherworld (Wiggermann 1992). The same can be observed concerning the primeval dunnu on the stone monument. Sumukan and his mother Er~etu,"Earth," later his sister Tdmtu, "Sea." Sumukan is an ass god, shepherd of anything, god of fourlegged animals, wild and tame, asses, goats, gazelles (Jacobsen l984:2 1, Cavigneaux l999:26 Iff.). Labar and Tdmtu, "Sea," his mother. Lahar is a god of flocks, and of domestic animals in general, occasionally associated with wool and clothing. The name literally means "Ewe;" here apparently the deity is male (Larnbert 1980-1983). Gaiu and Ida, "Cosmic River," his sister. Gaiu is a divine shepherd (An-Anum 111 86). [...I and du-a-ildag, "Pasture and Poplar," his sister. Pasture and poplar is what is found on the banks of the cosmic river, the mother of the goddess that embodies these elements. Uarnurni and NingeStina, his sister. gamurni is the Hurrian word for heaven. It is usually paired with the word for earth, e,?e (Laroche 1980:83f., Catsanicos 1996:230f., Wilhelm 1992:129, 130), which occurs in the Mesopotamian sources as Haj(j)aSu (Borger 1971:17). The Hurrian origin of these gods was discovered by W. G. Lambert (RAI 34 [I9871 unp.). NingeStina is a goddess of grapes and wine. Grapes are an important product of MA gardens (Fales 1989). [Haj(i)aSu ?] and [...I his sister. It is clear that the cosmos which is created in this myth by the successive births of its elements is one of farmers. Obvious are Earth, Soilbreaking Plough, Farmstead, the herdsmen Ass, Ewe, and Gaiu, the River with Pasture and Poplar, and the Lady of Grapes. The most serious omission is grain, but the last lines of the reverse indicate that it must have been treated in the badly damaged second half of the myth. In Jacobsen's reconstruction (Jacobsen 1984:8f.) they read "[ ...I the akitu festival of the month [Ajaru ...I I The (ploughman's) [work-solng [let him sound] shrilly [in the country]." The reason for restoring the month as Ajaru lies in a sequence of dates that runs parallel to the generations of gods. Most of the dates are broken, and their meaning is uncertain (calendaric?), so that the restoration cannot be considered full proof. That the end of the text refers to the beginning of the agricultural year, however, goes forth unequivocally from the "(ploughman's) [work-solng" ([Wit d~li[la]),the restoration and meaning (cf. CAD S.V.alda) of which can be considered certain. Since there is no ploughing without
204
F. A. M. WIGGERMANN
seed corn, the presence of grain can be safely assumed. In the north of Mesopotamia and in Anatolia Alala is a cosmogonic god (Lambert 1985:190), father of Anu "Heaven" (Hurrian Hamurni), and possibly his name is to be restored in the fifth pair. Both mythological sources, the seal and the text, stem from a northern milieu, and the concepts that inform them must have been familiar to the inhabitants of Sabi Abyad. Except for the "temple of Sin" (above 7) and the "servants of the temple of ASSur" (above 4) there is no evidence for a religious awareness, however, unless the model penises found in the tower have a cultic meaning.
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
9.
TEXTS
1 ME (ANSE?)SE-am i-na &BAN hi-bur-nj Sa mTam-me-te LU.AGRIG a-na SUKUANSE.KUR.RA.MES Ha ~DINGIR-i-pa-da 5 Ha ~ ~ U . G U R - S A G GAL X-te mRe-si-at-ri [G]AL LU.APIN.MES ma-I$-ir 10 6-Sa-dan N~G.KAS~.MES-SU i-sa-bat tup-pu-Su i-hap-pi 1
IT1 Si-ip-pu 15 UD 14' kam li-mu mdSt11-~~-~u-A-PAP
1
7393 ANSESE SUMUN 3366 ANSESE GIBIL SU.N~GIN 10759 ANSESE tab-ku
5
663 ANSESE NUMUN
300 ANSESE-am -@u-zi-ra-nu 7 1 ANSE SE-am rUruD~-nul-Sa m ~ ~ ~ - ~ I N ~ ~ ~ . ~ ~ S - n i 10 160 ANSE SE-am ~u~u-na-*~-Sur 113 ANSESE-am UNSa-a-6-a-nu 200 ANSESE-am 15 UrUDu-na-Sa-mi~-di PAP 844 ANSESE tal-pi-tu
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
6,7. [
] ~ANSE S~-aml[ [Sa S]UmfJa-be-i[a
I 1
[XM]E ANSESE SUMUN m U - ~ - [ [Sa Su] mrKi-li-lul Sa a[
1 1
] ANSESE-am Sa SU~~DINGIR-X-[ ] [Sa ?] U m ~ - [ 1 [
T 98-44 [
ANSE] 'SE-am1 [Sa SU
1 SU.SI4 ANSESE-amSa [Su 2 ME 1 Su.31ANSESE SUMUN [ PAP 6 ME 12 ANSES[E 1 ME 50 ANSESE-am SUMUN mAdad-[
T 98-105 (two pieces of one tablet, perhaps to be combined differently) ] 'ANSESE [ [
5? [ME] 34 ANSESE Sa S[U
1 ME [ANSESE Sa SU on edge probably date
I I 1
T 96-16, "duplicate" T 96-17 [50 nil-gal-lu ZABAR mu-t~-ra 50 -mUr-du 50 -mA-'pill-ti-a-du-ur 50 md~-Sur-ri-ba 50 mU-MU-SUM-na 50 mu-bal-li-fu 50 mQu-6-sa [1]4 mSil-li-ia (var.: 14) 10 -mBu-nu-uh-ni 11 -mh-6-hi 10 - m ~ - ~ d i n l - ~ U .(var. G u ~: m ~ ~ ~ - d ~ . ~ ~ ~ ) 2 ~DUMU-d~~tar 15 mMa-si-DINGIR 15 mRe-si-at-ri 7 -mBu-ra-'il-si
20 ni-g&lu ZABAR Sa Ma-nu-ki-i-U LU. 'AGRIG~
mI-ga-'a1-Se-i-mi-id a-na e-sa-di ma-@-ir A.SAi-si-id r~Lgd-li i-dan 10 h fup-pu-Su i-hap-pi ITI Sa sa-ra-ti UD 13 KAM li-mu mA-bat-tu
208
1
5
20
25
30
a-na mI-kar-Se-mi-id qi-bi-ma um-ma mTam-mi-te
a-na DUMU.SAL-rsul-sa-a-a[- ] qi-bi-[ma] um-ma m ~ a - n u - k i - i - d r ~ S ~ ~ ~ l
1
a-na ku-a-Si 'lu Sull-mu a-na ia-'a-Si Sull-mu
mSa-ar-ni-qa 5 al-tap-ra-ak-ku GUD.AB 2 Sa pi-ti mKi-te-ia mu-bur di-na-fig-Su GUD.MES-Su gab-ba lu-pa-l-j-ra 10 i-na pa-ni-ka lu-6-'Sal-zi-iz 3 GUD.MES Sa nam-Sar(Text: KAR)-te at-ta 6-su-uq 15 li-qi an-nu-te GUD.MES-Suri-ih-ta Su-ut li-il-qi gigga-Su-'ril Sa WUdu-ni-Su li-su-hu li-il-qi 20 6-[nu-t]e.M~$-Su 6-nu-ut 8-51.1 'gab-bal li-il-qi ma-am-ma la i-ka-la-Su
i-na UD 6[+x KAM Sa a-na UmT]a-i-di ni-tu-[mu-Su-ni LUGIAL iS-tu 'WTal [-i-di] 'a-nal kur~at-mu-l-j it-tu-r[a Sbm-ma LUIGAL 10 la-a i-ka-'lu-6-Sul a-di iS-tu k ~ K a t - ~ m u - ui-tu-ra-ni l i-na KUR us-ba-ni
15
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
F. A. M. WIGGERMANN
SE.'NUMUN~ Sa GAL Lo.giSA~lN.M~[S] 8a URU.DILI.DIL1 Sa la-a im-hu-ru-ni di-ni SE.NUMUN a-na mKur-ba-ni di-ni GANA.A.SASa rWulSab-la-li Ha m d ~ ~ ~ -r m ~ ~u~-5 1 - m 6-kal-li-mu[a] rSul-ni li-ru-USS6m-ma Su-ti-6.MES qa-i-ra-na-i[a]-6 i-na 'li-be Sal [mS]@-la-li i-ba-Si 'ia-nu1 Sa SA ni-ih-r[i]-ni (?) us-ba-ni l[i-q]i-e ki-i ga-6-ru-rtel a-na Su-ti-b.~ES qa-i-ra-na-ia-'el [dil-ni ma-ku-tu Sa URU.[DILI.DILI] lu dm-na-at ba-lu-ut 'ha1-ba-qi lu la 6-di-6
I
1
5
a-na mBa-la-t[u- ] mEN-KA[M] qi-bi- 'ma1 um-ma mrTarn-mi1-[tel-ma
-
LU.$~-l-j-tu SE.~.GIS-SU-~U.MES li-gI-rlu-$ul 'Stirn1-ma SE.3.'GIS~ i-na Wdu-ni-k~-~nul 10 la-AS-Su a-na Wdu-ni %a1[Su-ub]-re-e 'a-all-ka SE.~.GI$ a-na ~u.?a-'l-jl-te 15 di-na li-gI-rlul-$u
210 S E . ~ . Gli-rHI1-lu-$u I~ ~.MES la i-$a-hu-tu SE.~.GISl_lal-'?u1 20 a-na e-ma-di [-(x)] i-qa-bi-[a] ar-rl_liE1 li-H[I-lu-su]
E A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
21 1
REFERENCES Akkermans, P. M. M. G. 1993a with J. Limpens and R. H. Spoor, On the Frontier of Assyria. Excavations at Tell Sabi Abyad, Akkadica 84-85, 1-52. 1993b with B. Wittmann, Khirbet esh-Shenef. Ein spatneolithische Siedlung im Balikhtal, Nordsyrien, MDOG 125, 143-166. 1993c Villages in the Steppe: Later Neolithic Settlement and Subsistence in the Balikh Valley, Northern Syria Tell Sabi Abyad. The Late Neolithic Settlement 1996 (forthcoming) with F. A. M. Wiggermann, Middle Assyrian Tell Sabi Abyad, Syria. Excavation and Texts 1988-1999. Arnaud, D. Textes syriens de l'age du Bronze Ricent 199 1 Aynard, M.-J. and Durand, J.-M. Documents d7Epoque Mtdio-Assyrienne, Assur 311, 1-54. 1980 Barnett, R. D. More Balawat Gates: A preliminary report, 19-22 in M. A. Beek et al. eds., 1973 Symbolae Biblicae et Mesopotamicae Franciso Mario Theodoro de Liagre Bohl dedicatae Baal, K. Khirbet esh-Shenef. A Late Bronze Age Settlement in the Balikh Valley, 1990 Northern Syria, Akkadica 67, 10-32. Becker, C . Erste Ergebnisse zu den Tierknochen aus Tall Seb Hamad - Die Funde aus 199 1 Raum A des Gebaudes P, 117-132 in Kiihne ed. 1991. Biagov, L. N. Zur Interpretation der Termini E und frdunnulURUdunnuin den Urkunden 1976 der Mittelassyrischen Periode, AcAnHung 22, 333-335. Boerma, J. A. K. Soils and Environment of Tell Hammam et-Turkman, 1-9 in M. N. van 1988 Loon ed., Hammam et-Turkman I Bonatz, D. et al. eds., Rivers and Steppes. Catalogue to the Museum of Deir ez-Zor 1998 Borger, R. 1971 Gott Marduk und Gott-Konig h l g i als Propheten: Zwei Prophetische Texte, BiOr 28, 3-25. Butz, K. 1980-83 Landwirtschaft, RLA 6, 470-486. 1985 with P. Schroder, Zu Getreideertragen in Mesopotamien und dem Mittelmeergebiet, BUM 16, 165-209. Cancik-Kirschbaum, E. C. Die mittelassyrischen Briefe aus Tall St?b Hamad 1993 Catsanicos, J. 1996 L'apport de la bilingue de Hattusa ? lailexicologie hourrite, Amurru 1, 197296. Cavigneaux, A. 1999 A Scholar's Library in Meturan?, 251-173 in T. Abusch, K. van der Toorn eds., Mesopotamian Magic: Textual, Historical, and Interpretative Perspectives
212
F. A. M. WIGGERMANN
Cohen, M. E. The Cultic Calendars of the Ancient Near East 1993 Deller, K. 1985 with A. Tsukimoto, Ein mittelassyrisches Protokoll iiber eine Rinder- und Eselmusterung, BUM 16, 3 17-326. 1993a with A. Fadhil, Neue Nimrud-Urkunden des 8. Jahrhunderts v. Chr., BUM 24, 243-269. 1993b with A. R. Millard, Die Bestallungsurkunde des Nergal-apil-kumuja von Kalhu, BUM 24, 217-242. Driel, G. van The Cult of As's'ur 1969 with R. Jas, A Second Middle Assyrian LB Text on Gold, JEOL 31, 33-39. 1989 Cattle in the Neo-Babylonian Period, BSA 8, 215-240. 1995 Englund, R. K. Late Uruk pigs and other herded animals, 121-133 in U. Finkbeiner et al. 1955 eds., Beitrage zur Kulturgeschichte Vorderasiens (Fs. R. M. Boehmer) FalL?i3 1990a
A Middle Assyrian Text concerning Vineyards and Fruit Groves, SAAB 3, 53-59. Grain Reserves, Daily Rations, and the Size of the Assyrian Army, SAAB 4, 23-34. The Rural Landscape of the Neo-Assyrian Empire, SAAB 4, 81-141.
1990b Fincke, J. Die Orts- und Gewassernamen der Nuzi-Texte (RGTC 10) 1993 Noch einmal zum mittelassyrischen s'ilu& AoF 21, 339-351. 1994 Finkelstein, J. J. Cuneiform Texts from Tell Billa, JCS 7, 111-147. 1953 Freydank, H. Anmerkungen zu mittelassyrischen Texten 2, OLZ 80, 229-234. 1985 Zur Lage der deportierten Hurriter in Assyrien, AoF 7, 89-117. 1980 Zu den Grundeigentumsverhiiltnissen in mittelassyrischen Zeit, 79-88 in B. 1988 Brentjes ed., Das Grundeigentum in Mesopotamien Beitrage zur mittelassyrischen Chronologie und Geschichte 1991 d'%2/ Das Archiv ASSur 18764, AoF 19, 276-321. Drei Tafeln aus der Verwaltung des mittelassyrischen Kronlandes, AoF 21, 13-30. Noch einmal zum Vorgang pis'erti karii'e, 129-143 in B. Pongratz-Leisten et 1997 al. eds., Ana s'ad2 Labnani lii allik (Fs. W. Rollig) Garelli, P. with D. Charpin and J.-M. Durand, R61e des prisonniers et des dbport6s B 1982 l'tpoque m6dio-assyrienne, 41-46 in H. Klengel ed., Gesellschaft und Kultur im alten Vorderasien Giiterbock, H. G. The Cuneiform Tablets, 86-9 1 in C. W. McEwan et al., Soundings at Tell 1958 Fakhariyah, OIP 79. Hecker, K. Der Charab-Mythos, 610-611 in 0 . Kaiser ed., Texte aus der Umwelt des 1994 Alten Testaments 11114 Heimpel, W. 1976-1980 Jagd, RLA 5, 234-236.
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
Helzer, M. 1979 Hrouda, B. 196 1 Hunt, R. C. 1987
213
Dimtu-gt, pyrgos. An Essay about the non etymological Sense of these Terms, JNSL 7, 31-35. Tell Fekheriye: Die Keramik, ZA 54, 201-239. The Role of Bureaucracy in the Provisioning of Cities: A Framework for Analysis of the Ancient Near East, 161-192 in McG. Gibson, R. D. Biggs eds., The Organization of Power. Aspects of Bureaucracy in the Ancient Near East
Jacobsen, Th. The Harab Myth, SANE 213 1984 Jankowska. N. B. Communal Self-government and the King of the State of Arrapha, JESHO 1969 12, 233-282. The Role of the Extended Family in the Economic Life of the Kingdom of 1986 Arraphe, Oikumene 5, 33-42. Jas, R. Two Middle-Assyrian Lists of Personal Names from Sabi Abyad, Akkadica 1990 67, 33-39. Keel, 0 Vogel als Boten 1977 with C. Uehlinger, Altorientalische Miniaturkunst 1990 King, L. W. Babylonian Boundary-Stones and Memorial-Tablets 1912 Kiihne, C. Ein mittelassyrisches Verwaltungsarchiv und andere Keilschrifttexte, 2031995 222 in W. Orthmann et al. eds., Ausgrabungen in Tell Chut3-a in NordostSyrien I Aspects of the Middle Assyrian Harbu Archive, SAAB 1012, 3-7. 1996 Gersterationen im mittelassyrischen Uarbu, BBVO 18, 179-188. 1999 Kiihne, H. 1990a Ein Bewasserungssystem des ersten Jahrtausends v. Chr. am unteren Habur, 193-215 in B. Geyer ed., Techniques et pratiques hydro agricoles traditionelles en domaine irrigue' 1990b The Effects of Irrigation Agriculture: Bronze and Iron Age Habitation along the Khabur, Eastern Syria, 15-30 in S. Bottema et al. eds., Man's role in the Shaping of the Eastern Mediterranean Landscape ed., Die rezente Umwelt von Tall S$J Hamad und Daten zur 1991 Umweltrekonstruktionen der assyrischen Stadt Dur-Katlimmu Dur-Katlimmu and the Middle Assyrian Empire, in 0 . Rouault, M. Wafler 1994 eds., La djkzire' et l'euphrate syriens de la proto-histoire d la fin du second millinaire av. J. -C. Lambert, W. G. Near Eastern Seals in the Gulbenkian Museum of Oriental Art, University of 1979 Durham, Iraq 41, 1-45. 1980-83 L&x, RLA 6,431. The Pair Lahmu-Lahamu in Cosmogeny, OrNS 54, 189-202. 1985 Laroche, E. Glossaire de la langue hourrite 1980
214 Lewis, N. 1988
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
The Balikh Valley and its People, 683-694 in M. N. van Loon ed., Hammam et-Turkman II
Liverani, M. Economy of Ugaritic Royal Farms, 127-168 in C. Zaccagnini ed., 1989 Production and Consumption Matthews, D. M. Principles of Composition in Near Eastern Glyptic 1990 McEwan, C. W. et al. Soundings at Tell Fakhariyah, OIP 79 1958 Milano, L. Food and Diet in Pre-Classical Syria, 201-271 in C. Zaccagnini ed., 1989 Production and Consumption Miller, R. Elephants, Ivory and Charcoal: An Ecological Perspective, BASOR 264 , 291986 43. Moortgat, A. Bildwerk und Volkstum Vorderasiens zur Hethiterzeit 1934 Assyrische Glyptik des 13. Jahrhunderts, ZA 47, 50-88. 1942 Muller, G. G. W. \ Studien zur Siedlungsgeographie und Bevolkerung des Mittleren 1994 Osttigrisgebietes, HSAO 7 Nashef, Kh. Die Orts- und Gewassernamen der mittelbabylonischen und mittelassyrischen 1982 Zeit, RGTC 5 Netting, R. McC. Smallholders, Householders. Farm Families and the Ecology of Intensive, 1993 Sustainable Agriculture Nissen, H. Aus dem Geschiiftsleben assyrischer Kaufleute im 14. Jhdt. v. Chr., 111-120 1967 in D.O. Edzard ed., Heidelberger Studien zum Alten Orient (Fs. Falkenstein) Osten-Sacken, E. von der Vogel beim Pflugen, BBVO 18, 265-278. 1999 Parpola, S. Letters from Assyrian Scholars to the Kings Esarhaddon and Assurbanipal, 1970 Part Z Letters from Assyrian Scholars to the Kings Esarhaddon and Assurbanipal, 1983 Part II Porada, E. Corpus of Ancient Near Eastern Seals in North American Collections I: The 1948 Collection of the Pierpont Morgan Library Postgate, J. N. Land Tenure in the Middle Assyrian Period: A Reconstruction, BSOAS 34, 197 1 496-520. Ilku and Land Tenure in the Middle Assyrian Kingdom - A Second Attempt, 1982 304-313 in M. A. Dandamayev et al., eds. Societies and Languages of the Ancient Near East (Fs. Diakonoff) 1985 The 'Oil-Plant' in Assyria, BSA 2, 145-152. 1987a Notes on Fruit in the Cuneiform Sources, BSA 3, 115-144. 1987b Some Vegetables in the Assyrian Sources, BSA 3, 93-100. I
215
Postgate, J. N. The archive of ~rad-Seriiaand his family. A Middle Assyrian household in 1988 government service (CMA 1) The Ownership and Exploitation of Land in Assyria in the First Millenium 1989 B.C., 141-152 in M. Lebau et al., eds., Reflets des deux fleuves (Fs. A. Finet) Powell, M. A. Late Babylonian Surface Mensuration. A Contribution to the History of 1984 Babylonian Agriculture and Arithmetic, AfO 31, 32-66. Radner, K. Die neuassyrischen Privatrechtsurkunden als Quelle fur Mensch und Umwelt 1997 (SAAS 6) Reiner, E. Eniima Anu Enlil Tablets 50-51 (BPO 2) 198 1 Rollig, W. Preliminary Remarks on the Middle-Assyrian Archive from Tell Schech1984 HamadIDur-Kattlimmu, AAAS 34, 189-194. Zur Landwirtschaft am unteren Habur im 2. Jt. v. Chr., RAI 34, Istanbul, 1987 unpub. Das Archiv von Dur-Katlimmu, unpub. 1988 with A. Tsukimoto, Mittelassyrische Texte zum Anbau von Gewiirzpflanzen, 1999 427-439 in B. Bock et al., eds., Munuscula Mesopotamica (Fs. J. Renger). Roth, M. T. Law Collections from Mesopotamia and Asia Minor 1995 Sassmannhausen, L. Bauern in der Kassitenzeit, BBVO 18, 155-160 1999 Schirmer, W. Landschaftsgeschichte um Tall Bica am Syrischen Euphrat, MDOG 119, 571987 117. Seidl, U. 1957-1971 Gottersymbole, RLA 3, 483-490. 1989 Die babylonischen Kudurru-reliefs Simonet, G. Irrigation de PiCmont et Cconomie agricole Z i Assur, RA 71, 157-168. 1977 Spek, R. J. van der Cuneiform Documents on Parthian History: The Rabimesu Archive. 1998 Materials for the Study of the Standard of Living, 205-258 in J. Wiesenhofer ed., Das Partherreich und seine Zeugnisse Sterks, M. De Koninklijke Jacht, unpub. MA thesis 1996 Stiehler-Alegria Delgado, G. 1996 Die Kassitische Glyptik Stol, M. On Trees, Mountains, and Millstones in the Ancient Near East 1979 1983-84 Cress and its Mustard, JEOL 28, 24-32. Garlic, Onion, Leek, BSA 3, 57-80. 1987 Tsukimoto, A. 1992 Aus einer japanischen Privatsammlung: Drei Verwaltungstexte und ein Brief aus mittelassyrischer Zeit, WdO 22, 21-38. Villard, P. 1987 Un conflit dYautoritCsB propos des eaux du Balih, MAR1 5, 591-596.
216
F. A. M. WIGGERMANN
Wiggermann, F. A. M. 1992 Mythological Foundations of Nature, 279-304 in D. Meijer ed., Natural Phenomena. Their Meaning, Depiction and Description in the Ancient Near East The Middle Assyrian Texts from Tell Sabi Abyad, RAI Berlin, unpub. 1994 Scenes from the Shadow Side, 207-230 in M. E. Vogelzang and H. L. J. 1996 Vanstiphout eds., Mesopotamian Poetic Language: Sumerian and Akkadian (CM 6) Transtigridian Snake Gods, 33-55 in I. L. Finkel, M. J. Geller eds., Sumerian 1997 Gods and their Representations (CM 7) Nackte Gottin, RLA 9, 46-53. 1998 Wilhelm, G. Hurritische Lexikographie und Grammatik: die hurritisch-hethitische 1992 Bilingue aus Bogazkoy, OrNS 61, 122-141. Wilkinson, T. J. The Structure and Dynamics of Dry-Farming States in Upper Mesopqtamia, 1994 Current Anthropology 3515, 483-520. Sabi Abyad: The Geoarchaeology of a Complex Landscape, 1-24 in 1996 Akkermans 1996. 1998a Water and Human Settlement in the Balikh Valley, Syria: Investigations from 1992-1995, Journal of Field Archaeology 25, 63-87. 1998b Settlement and Irrigation in the Balikh Valley, Syria, from the 3rd 'to the 1st Millenium BC. A Preliminary View, Subartu 4 (1998) 151-170. Wirth, E. Syrien. Eine Geographische Landeskunde 197 1 Zaccagnini, C. The Yield of Fields at Nuzi, OA 14, 181-225. 1975 The Rural Landscape of the Land of Arraphe 1979 Again on the Yield of Fields at Nuzi, BSA 5, 201-217. 1990 Zeder, M. A. Pigs and Emergent Complexity in the Ancient Near East, 109-122 in S. M. 1998 Nelson ed., Ancestors for the Pigs: Pigs in Prehistory, MASCA Research Papers in Science and Archaeology 15 Zeist, W. van Some Notes on Second Millenium B.C. Plant Cultivation in the Syrian Jazira, 1994 541-553 in H. Gasche et al. eds., Cinquante-deux r$lexions sur le ProcheOrient ancien (Fs. L. DeMeyer).
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
tkt-Qd.H~.--f~;c~+J COPY OF TEXT 1 = T 98-33
* COPY OF TEXT 2 = T 98-1 15
218
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
F. A. M. WIGGERMANN
edge
COPY OF TEXT 3 = T 98-38
COPY OF TEXT 4 = T 98-44
TWO
pace, d o n e
hzbfd. which
do notjoin.
(
COPY OF TEXT 6 = T 96-16 COPY OF TEXT 5 = T 98-105
I
c+ rn
-r)
n w.
Lwer edge comp[ctcly abraded
CD
COPY OF TEXT 7 = T 96-17
220
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
COPY OF TEXT 8 = T 96-3
m v g r irnqpr of
inside of envelop f r a p e n t
COPY OF TEXT 9 = T 96-23
222
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
Fig. 1. WESTERN PART OF THE MIDDLE ASSYRIAN EMPIRE. The map shows the western part of the Middle Assyrian Empire at ca. 1200 B.C. West of the Euphrates lies the Hittite Empire, Carkemish is the seat of a Hittite viceroy.
o last
cign
o f [. 9
COPY OF TEXT 11 = T 97-17
COPY OF TEXT 12 = T 93-7
Fig. 2. TELL SABI ABYAD WITH FARMSTEAD OF ILI-PADA, GRAND VIZIER OF ASSYRIA. The Assyrians tried to resettle depopulated areas by founding state farms. The fortified farmstead (dunnu) of the grand vizier, Ili-pada, is presently being excavated at Tell Sabi Abyad. Inside the walled precinct, we find the following functional units (buildings rendered schematically): 1) owner's residence, 2) tower with stores, treasury, and jail, 3) residence of the chief steward (abarakku), 4) reception court with 5) a small office of the latter, 6) servant quarters, with office of the lower administration near a second entry. Outside the walls there are more buildings (7), but not on the whole tell (8, prehistory only). The dots indicate findspots of tablets.
F. A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
/ '/
Fig. 4. NATIVE VIEW ON THE RURAL LANDSCAPE OF THE BALIKH VALLEY. Middle and Neo-Assyrian kings hunted lions, wild bulls, and elephants in the area. Ashumasirpal I1 (883-859) showed his lion hunt on a bronze band of a gate at Balawat (drawing after Bamett 1973).
Rain-fed c u l t i v a t i o n i n modern S y r i a :
vA
I n n e r zone with r e l a t i v e l y c e r t a i n y i e l d s p a r t i a l ;;I/; sporadic
dominant
I / ,
Outer zone with h i g h y i e l d u n c e r t a i n t y
irrigated
.
.. ., ,,....., ..,.,.....,.: ,,,. ... .
:.: ... ;. :.,.. ,;. :,,::. .:;: . :
..t.
,:
' /
dominant /
,
sporadic
V, L'
partially irrigated
st eppe
:2
Fig. 3. TELL SABI ABYAD IN ITS AGRICULTURAL SETTING. Tell Abyad (2, ancient Dunni-ASSur ?), Harran (5), Urfa (7), and Carkemish ( 6 ) are well within the rainfed zone, Tell Chuera (3, ancient Harbu) and Tell Sabi Abyad (1) are located on the margin, Tell Bi'a (ancient Tuttul) is outside. Source: TAVO A X 4.
F. A. M. WIGGERMANN
Fig. 5. CATCHMENT AREA OF SABI ABYAD. From textual evidence the catchment area of the farmstead at Tell Sabi Abyad (BS 189) can be calculated (36 km2), and plotted on a map as a circle (radius of ca. 3.5 km), or a square (sides of 6 km). The deduced area gives the farmstead plausible natural boundaries in the Balikh and the non-arable areas (indicated by lines) in the north and southeast, and includes Khirbet esh-Shenef (BS 170) and Hammam et-Turkman (BS 175), where Middle Assyrian presence has been established. The asterixes indicate the locations of possible further subcentres (see Fig. 6). Tell Jittal (BS 211, Dunni-Dagal?) and BS 161 are probably independent of Sabi Abyad.
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
Fig. 6. SABI ABYAD AND THE LOCATION OF ITS SUBCENTRES. The location of Khirbet esh-Shenef (BS 170) suggests subcentres with a radius of ca. 1.4 km, distributed on a circle around the centre. Hammam et-Turkman (BS 175) can function in this circle, although it is at the limit, BS 161 clearly falls outside the series of deduced subcentres (indicated by an asterix).
E A. M. WIGGERMANN
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
Sources: Sabi Abyad: T 98-115 (Text 2); HiSSutu: MARV 2 23; NEmad-IStar: MARV 3 4; Tarbagbe: MARV 3 10 (see Freydank 1994).
1 locality Sabi Abyad
I
total:
I area in
I seed corn I yield in
ikGs
in homers homers
homerlika
1989
[0.30]
2.45
1900
I
~ ~ m a & ~ S t1a r30b
.
-- -
total:
Dur-katlimmu Fig. 7. NATIVE VIEWS ON AGRICULTURAL WORK. Three Middle Assyrian seals (a, b: Moortgat 1942:81 Figs. 66, 67; c: Lambert 1979 no. 63) show a ploughman at work. On the better preserved example the animal is clearly male (a; cf. Fig. 9), and must have been castrated. That only one ox pulls the plough is artistic convention rather than reality. Here, as well as on Fig. 9, the plough is not the Babylonian seeder plough encountered on Kassite seals (d: Matthews 1990 no. 156, also 155, cf. Stiehler-Alegria Delgado 1996: 143, Seidl 1957-1971 3 5d), and in Assyria in the NA period (Radner, this volume). On c a second figure is sowing broadcast (cf. Fig. 9). The type of plough apparently does not influence the seeding rate (0.30 homerlikci in Assyria, as well as in Kassite Babylonia, cf. King 1912:xiii).
1 700
I
4873
1 1512.50 1 [0.30] [0.30]
2 300 total: 600
1 yield:
1 817.50
1
I seed codyield I
kg/ha 42 1
I 2.70 2.17
810 650 1460
1 yield:
1 : 7.35
I 465 372
I 1:9 1 : 7.2
I
the relation between seed corn and yield varies between 1 : 1 and 1 : 9, averaging 1 : 3 or 1 : 4 (Kiihne 1990b:20, after Rollig 1987). The seeding rate is 0.30 (Freydank 1994:29).
Fig. 8. CULTIVATED SURFACES AND YLELDS OF MIDDLE ASSYRIAN CROWN LANDS. Sources: Sabi Abyad: T 98-1 15 (Text 2); HiZSutu: MARV 2 23; NEmad-IStar: MARV 3 4; Tarbaghe: MARV 3 10 (see Freydank 1994).
F. A. M. WIGGERMANN
Fig. 9. A NATIVE VIEW ON THE COSMIC FOUNDATIONS OF AGRICULTURE. The lower register shows the real world, a man and an ox ploughing the earth, a second man sowing broadcast. The upper register shows the mythological foundation of agricultural fertility, the marriage between the rain god Adad and Mar.
AGRICULTURE IN THE NORTHERN BALIKH VALLEY
Fig. 10. POSSIBLE DEPICTION OF A COSMIC DUNNU. The monument shows the cosmos in three registers, below, founded on a mythological dragon, a depiction of the netherworld, in the middle a scene with unusually sedate wild animals, probably a reference to primeval times, above the symbols of the great gods. The stone, although uninscribed, has the form of a kudurru (boundary stone), most of which are concerned with royal grants of agricultural land. The only myth that fits the agricultural context of the stone is a farmers cosmogony, in which the primeval gods Plough and Earth inhabit a cosmic dunnu (farmstead). They produce animals, divine herdsmen, and ultimately agriculture. Source: Seidl 1989 no. 40, LBA, from Susa.
.
PIHANS LXXXVIII, 2000
How did the Neo-Assyrian King Perceive his Land and its Resources? Karen Radner (Helsinki) Slightly modifying the topic originally suggested to me by the organisers of this symposium, I will not discuss "how the Assyrian King perceived his land and its resources," but attempt to investigate "how the Neo-Assyrian king perceived his land and its resources" in the present paper. Focusing on the Neo-Assyrian period is, I think, justified by the sources which in that age are more diverse than in the preceding periods. It should be noted, however, that the Neo-Assyrian kings' views on land can be expected to reflect their predecessors' as the attitude towards land tends to be determined by tradition to a large extent. I am aware of the limited possibilities at our disposal to find answers to the question "how the Neo-Assyrian king perceived his land and its resources" as no king, and indeed nobody else, from that period left a treatise elaborating on his personal views on the matter. However, we are in the fortunate position to command a variety of different sources which at least allow us to attempt an approximation of the king's attitude towards his land and its resources. The royal inscriptions are probably our most informative source. But as they illustrate the way in which the king chose to portray himself and his actions they do not necessarily reflect the way he actually thought and acted.1 The same is true for our second source, the royal grants and decrees with which privileges were bestowed upon individuals and temples. A more unbiased view may be found in our third source, the correspondence between the king and his officials. However, as authors and addressees were familiar with the overall situation, the letters do not offer much general information, but primarily deal with specific details. Another important source is the visual record left, most prominently the reliefs found in the Neo-Assyrian palaces. On these reliefs close attention is paid to the depiction of landscape. By combining the information gathered from these sources, we may try to approximate the king's views on land and its resources. For the purpose of this paper, I will focus on two kings, Sargon I1 (721-705 BC) and Sennacherib (704-681 BC), as texts covering all the aforementioned categories as well as reliefs have survived from the reigns of these kings.2 Both Sargon and Sennacherib had new residence cities built, the first ~ur-Sarrukinand the latter Nineveh; in their inscriptions they explain why they chose to do this. These passages contain the most personal accounts on how these Assyrian kings saw their land and its potential. While Sargon relishes the thought that none of his predecessors, 350 in number, had realised the great promise shown by the
On the difficulties of using royal inscriptions as a source for a characterisation of a king see recently E. Frahm, Einleitung in die Sanherib-Znschriften. AfO Beiheft 26, Vienna 1997, 19. But note that no letters from the reign of Sennacherib have survived.
234
K. RADNER
city of Maganuba which he had chosen as a location for his residence ~ur-Sarrukin,3 Sennacherib takes pride in being the one to raise the ancient city of Nineveh to the eminent position it had so long deserved.4
As a first step to approximate the Assyrian kings' attitude towards land a reaction to a recent description of this attitude as that of a dominant male penetrating a passive female is called for. In an attempt to interpret the landscape imagery of the Assyrian reliefs, modelled after interpretative work on sixteenth century Italian and seventeenth century Dutch landscape paintings, M. I. Marcus, claiming that "it seems worthwhile exploring the radical idea that the Assyrian imperial "landscapes" were likewise expressions of (male) sexual anxieties and ideologies - expressions of (royal) manhood in the face of Nature, seen as female and Other,"5 tries to connect Assyrian territorial conquest, characterised as "penetration followed by possession," with "male (hetero)sexual anxieties."6 She comes to the conclusion that "all of this imagery implies a double desire by the king and state: to dominate the land as a man might dominate a woman."7 While it is clear that the Assyrian king indeed wished to master the land the sources hardly lend themselves to the equation with a sexual relationship between a dominant male and a passive female. As sexual connotations are entirely missing both in the written and in the visual accounts on the Assyrian conquest^,^ Marcus' approach may be less radical than rather unduly influenced by pre-conceived notions about the nature of Assyrian imperialism as well as her theoretical f r a m e ~ o r kEspecially .~ the idea that the Assyrians
THE NEO-ASSYRIAN KING AND HIS LAND
equated the land with the female and the female body cannot be supported at al1.10 Note also that the Assyrians called their country ASSur, after their male god.11
Recently J. Reade and I. J. Finkel12 have suggested to interpret a sequence of symbols depicted repeatedly on glazed-brick reliefs and bronze appliquks of Sargon's buildings in ~ur-&irrukinand a similar sequence found on Esarhaddon's so-called Black Stone13 and on some of his prisms as hieroglyphic writings of these kings' names and titles.14 In Sargon's sequence, the combination of a fig-tree and a seeder-plough is supposed to stand for "the country of Assyria" (fig. 1) while in Esarhaddon's sequence the combination of a seeder-plough and a date-palm stands for "Assyria and Babylonia" (fig. 2). The symbols chosen to represent the countries stress the importance of agriculture and horticulture. In our study of the king's attitude towards his land, we shall highlight the king's role as farmer and as gardener, a role which is both propagated in those sources promoting the king's self-image but also emphasised in less biased texts.
When a bad omen threatening the king's well-being made it necessary that the country was pro forma ruled by a substitute king (Zar puhi),l5 the king adopted the title of a "farmer"l6 and, bearing this title, carried on his functions. The title cannot be explained 10
Inschrift auf den Tonzylindern:44-47, see A. Fuchs, Die Inschriften Sargons II. aus Khorsabad, Gottingen 1994, 38f and 293; Inschrift auf den Stierkolossen:43-46, see Fuchs 1.c. 67 and 304f; Kleine Prunkinschrift des Saales XIV:29f; see Fuchs 1.c. 78 and 309f. D. D. Luckenbill, The Annals of Sennacherib. OIP 2, Chicago 1924, 80: H 3:17f and 94f: A 1:6370. M. I. Marcus, "Geography as Visual Ideology: Landscape, Knowledge, and Power in Neo-Assyrian Art," Neo-Assyrian Geography. QGS 5, ed. M. Liverani, Rome 1995,200. Marcus 1.c. 202. Marcus 1.c. 202. Quite on the contrary, it is noteworthy that depiction of sexual violence of Assyrian soldiers against women and children are altogether missing although abuse certainly took place. This is even more remarkable as the Assyrian stonemasons were perfectly able to portray extreme physical violence in all its horror, cf. the detailed depiction of skinning, impaling and decapitation. The reluctance to make sexual violence and sexuality the subject of art and literature is certainly a topic worthy of further study. Note that the scene on an Assurbanipal relief quoted by Marcus (1995202 with n. 66) as showing a rape scene (cf. also J. Reade's subtitle for the illustration "rape of Arab woman" in S. Parpola and K. Watanabe, Neo-Assyrian Treaties and Loyalty Oaths. SAA 2, Helsinki 1988, 47 fig. 13) shows two Assyrian soldiers brutally manhandling a (fully clothed) captive in a tent; although this may well have led to a rape the scene can hardly be described as depicting one. Cf. also the review of A. Fuchs, AfO 44/45 (1997198) 408.
235
Marcus 1.c. 200, using, e.g., "the European imperialist tradition of naming colonial territories after women" as a comparison. 1 1 See W. G. Lambert, "The God ASSur," Iraq 45 (1983) 82-86 on the nature of the god. I. Finkel and J. Reade, "Assyrian Hieroglyphs," ZA 86 (1996) 244-268; see already J. Reade, "The Khorsabad Glazed Bricks and Their Symbolism," Khorsabad, le palais de Sargon II, roi d'Assyrie, ed. A. Caubet, Paris 1995, 235 with 248 fig. 12 and 250 fig. 14 and compare the additional remarks of J. A. Scurlock, "Assyrian Hieroglyphs Enhanced," NABU 1997192. l 3 The Black Stone, also known as Lord Aberdeen's Stone, currently kept in the British Museum (BM 91027), has received a lot of attention during the last few years. Aside from adorning the cover of L. Kataja and R. Whiting, Grants, Decrees and Gifts of the Neo-Assyrian Period. SAA 12, Helsinki 1995, the stele and its decoration has been studied by B. N. Porter, "Conquest or Kudurru's? A Note on Peaceful Strategies of Assyrian Government," The Tablet and the Scroll. Fs. W. W. Hallo, ed. M . Cohen et al., Bethesda 1993, 194-197 and by P. A. Miglus, "<
236
K. RADNER
THE NEO-ASSYRIAN KING AND HIS LAND
by Rollentausch between the king and his substitute alone as calling the king "beggar" or "fool" would have served that purpose just the same, if not better.17 It seems more likely to me that the title of a "farmer" was chosen because it reflected one of the king's supreme duties and primary concerns, the cultivation of land.18 Rain was always good news for the king, as witnessed by a letter written by Issarduri, the governor of Arrapha, to Sargon: "It has rained a lot. The harvest will be good. The king, my lord, can be glad."lg Another letter20 which is unfortunately very broken is concerned with more alarming news: the river running past Samaria in Palestine had dried up and the area was facing a period of severe drought. Rain which was so essential but just as impossible to control was therefore considered a divine blessing of the highest level which is best illustrated by Sargon's inscription on the threshold of the temple of the storm-god Adad in ~ur-Sarrukin:~l
The Assyrian kings tried to reduce the risk in the uncertainty zone and to raise the yield of the arable land in general by supplementing rainfed irrigation with artificial irrigation by means of wells and canals. The importance of this task is stressed in the foundation inscription of the New Year festival house in Assur in which Sennacherib calls himself the one "who causes canals to be dug, who opens wells, who causes irrigation-ditches to murmur, who establishes plenty and abundance in the wide regions of Assyria, who puts irrigation water inside of Assyria."26 Sennacherib bore this title with full right as, other such enterprises aside, he had five major irrigation projects completed for providing the Nineveh area with water in order to guarantee maximum profitability." Whereas Sennacherib's irrigation projects for the Nineveh area were necessitated by the need to provide an over-sized metropolis28 with the means to produce enough food to function properly without being overly dependent on imports from the provinces, the situation in the rest of the empire was rather different. Although a number of fairly big cities existed, Assyria was essentially a rural, not an urban country. As various surveys in northern Mesopotamia have shown the country was covered by a multitude of settlements in the Neo-Assyrian period, most of them rather sma11.29 The same situation is found in the textual record, most explicitly in the texts of the so-called Harran Census which are possibly to be dated to the reign of Sargon.30 To quote F. M. Fales, "No urban sites appear directly in the texts: the listed real estate pertains to the rural world of the village and the hamlet (kapru), which may at best be located (ina) qanni, 'near' a larger town or a city, but on the other hand may be deep ina madbar, 'in the steppe, ' its sole reference being a particular province."31 Wherever possible, steps to convert steppe lands into arable land were taken. A good illustration of this is provided by a letter written by NabQ-Sumu-iddina, the governor of
"0 Adad, canal inspector of heaven and earth who illuminates the sanctuaries, for
Sargon, king of the world, king of Assyria, governor of Babylon, king of Sumer and Akkad, the builder of your sanctuary, bring rain from the sky and flood from the springs, amass corn and oil in its (i.e., the city of Dur-Samkin's) surroundings, let your subjects graze in the meadows in plenty and abundance, strengthen the foundation of his (i.e., Sargon's) throne (and) let his reign last for long!" It is usually thought that rainfalls of a minimum of 200 mm per annum are needed for rainfed agriculture. A glimpse at a modern map shows that although most of the Assyrian heartland with the cities Nineveh, Kalhu and Dur-Sarrukin is situated within this area, a large part of the empire, including the city of Assur and most of the Jezireh, is outside it. Most of this land is steppe land. In the inscriptions, composed in the Standard Babylonian dialect, the synonyms nama and madbaru are used to denote steppe land whereas in the texts written in Neo-Assyrian such as letters, legal and administrative documents only the term mudaburu22is employed. Steppe land could always be used for grazing the herds23 but cultivated only when the rainfalls were sufficient. For this reason the region has been dubbed the risk or uncertainty zone24 by modern researchers.25 17 The title has nothing to do with the original profession of the substitute king as he could be anybody whose life could be dispensed with (he was subsequently killed). For a study of the Neo-Assyrian vocabulary on land cultivation as found in the legal documents see F. M. Fales, "The Rural Landscape of the Neo-Assyrian Empire: A Survey," SAAB 4 (1990) 81-142. l 9 ABL 157 r. 7-11: A AN.MESma-a'-da (8) a-dun-niS i-ta-lak (9) BURU14 de-e-qe (10) SA-bi Sa LUGAL be-li-ia (1 1) lu-u DUG.GA. 20 CT 53 458 = SAA 1 255. 21 A. Fuchs, Die Inschrifen Sargons 11. aus Khorsabad, Gottingen 1994,282 and 370. 2 2 Neo-Assyrian for madbaru. 23 ABL 547 = SAA 1 82 illustrates the tensions between the semi-nomads grazing their herds in the Jezirah and the sedentary population of Assur whose relations with each other were especially fragile at times of hunger. 24 See, e.g., H. Kiihne, Die rezente Umwelt von Tall S2h Hamad und Daten zur Umweltrekonstruktion der assyrischen Stadt Dur-Katlimmu. BATSH 1, Berlin 1991, 28. 25 The evaluation of the large archive found in 1998 in the so-called Red House in Dur-Katlimrnu (mod. Tall S E :Hamad) ~ on the river Habur promises for the first time the possibility to gain insights into the economics of a major Neo-Assyrian city situated in the risk zone. Although the city of Assur is today situated outside the 200 mrn per annum border, the texts from this city offer little information on
237
artificial irrigation; an exception is the reference to Sennacherib's construction of a canal with the name "The one that purifies the New Year festival" feeding the gardens of the New Year festival house, see VAT 9656 = SAA 12 86:19. Note that in the building inscription of the New Year festival house, KAH 2 122, Sennacherib says that he had two canals dug in order to irrigate the gardens, see D. D. Luckenbill, The Annals of Sennacherib. OIP 2, Chicago 1924, 137: I 2:33-35. 26 KAH 2 122:ll-14, see Luckenbill 1.c. 135 (I 2) and cf. M. Liverani, "Critique of Variants and the Titulary of Sennacherib," Assyrian Royal Inscriptions: New Horizons in Literary, Ideological, and Historical Analysis. Orientis Antiqui Collectio XVII, ed. F. M. Fales, Rome 1981, 248f. 27 See in detail J. E. Reade, "Studies in Assyrian Geography. Part I: Sennacherib and the Waters of Nineveh," RA 72 (1978) 47-72, 157-180 and cf. E. Frahm, Einleitung in die Sanherib-Inschriften. AfO Beiheft 26, Vienna 1997, 13. 28 Nineveh was much larger than any other Assyrian city; thanks to 0 . PedersCn, Archives and Libraries in the Ancient Near East, 1500-300 B.C., Bethesda 1998, the size of Nineveh (p. 159: Plan 74) can now easily compared with that of various other Neo-Assyrian cities: Assur (p. 133: Plan 62), Kalhu (p. 144: Plan 66), ~ur-$armkin(p. 156: Plan 72), Imgur-Illil (p. 166: Plan 77), Sibaniba (p. 168: Plan 79), Dur-Katlimmu (p. 170, Plan SO), Guzana (p. 173: Plan 82), Til-Barsip (p. 176: Plan 84), Burmarina (p. 178: Plan 86) and Huzirina (p. 179: Plan 87). 29 T. J. Wilkinson, "Late-Assyrian Settlement Geography in Upper Mesopotamia," Neo-Assyrian Geography. QGS 5, ed. M. Liverani, Rome 1995, 139-159. 30 The texts of the Harran Census, originally published by C. H. W. Johns and later edited by F. M. Fales, were recently re-edited as SAA 11 200-219. F. M. Fales and J. N. Postgate, Imperial Administrative Records, Part 11. Provincial and Military Administration. SAA 11, Helsinki 1995, xxxi.
238
K. RADNER
THE NEO-ASSYRIAN KING AND HIS LAND
Lahiru in northern Babylonia, to S a r g ~ n"Concerning :~~ what the king, my lord, wrote to me: '[Surlvey in detail the surroundings [of] the fort in regard to cultivating the ste[ppe]!' I have surveyed it: it is very resourceful." That the cultivation of barren land was one of the king's primary concerns is made abundantly clear in Sargon's prism inscription:33
reconstruct foreign landscapes in their reliefs; they also brought them into Assyria by laying out parks in their residence cities. The first Assyrian king to bring back trees from the conquered lands to Assyria was Tiglath-pileser I (1114-1067); in any case, he is the first to mention this in his inscriptions and he prides himself in the fact that "I took cedar, box-tree (and) Kanish oak from the lands over which I had gained dominion, such trees which none among the previous kings, my forefathers, had ever planted, and I planted (them) in the orchards of my land; I took rare orcbard fruit which is not found in my land (and) filled the orchards of Assyria."38 When Ashurnasirpal I1 (883-859 BC) had exhaustive building work performed in Kalhu in order to make the town his new residence city, he also had numerous kinds of trees which he previously had encountered on his campaigns planted in the new pleasure gardens in the Tigris meadows, giving a detailed list of the 41 varieties of trees in one of his inscriptions, the famous Banquet Stela.39 The gardens were irrigated with the help of a newly constructed canal dug from the Upper Zab named "Canal of plenty" (patti hegalli). Sargon had extensive gardens built around ~ u r - h m k i nof which his inscriptions, a number of letters from the royal correspondence and scenes on his reliefs in the palace of Dur-&mukin (fig. 3)40 bear witness. Sargon states in his inscriptions that the model for the park in Dur-Sanukin was the Amanus mountain range: "Around it (i.e. the city) I constructed a park, an exact replica of the Amanus mountains, in which all the aromatic plants of the Hatti land and the fruit trees of every mountain are planted."41 The letters illustrate how several thousands of plants were brought in, especially from the region of the Habur and the Middle Euphrates, but also from Assyria's northern border.42 An enterprise of such dimensions necessarily needed a great deal of organisation and planning in advance. As a letter from the governor of Kalhu to the king shows, plans were used to construct the gardens.43 The most famous among the royal Assyrian horticulturists is, however, Sargon's son and successor Sennacherib who had lavish gardens designed in Nineveh. In his
"The well versed king, who constantly considers plans of good things and who directs his attention to the settlement of desolate steppes, to the cultivation of fallow land and to the plantation of fruit groves, contemplated causing steep rocks, from which never before green had sprouted, to produce yield. He had in mind to let furrows arise in waste barren land which had not known the plough under the previous kings, to let the work song resound, to open a spring as a karattu in an area without well and have (everything) irrigated in abundance from top to bottom (with) water, like with the masses of the flood (of a river in spring)." The cultivation of barren land was either achieved with the work power of the local population or of deportees who were settled in the area at the same time. Sargon reports in his inscriptions that he had people from the country of Kurnmuhi in Anatolia (classical Commagene) brought to Bit-Iakin in Babylonia where they were settled and where they cultivated the barren land.34 This method is also well documented in the letters from the royal correspondence. Hence we read in a letter written by Bel-liqbi, the governor of Supat in Syria (possibly mod. Homs), to S a r g ~ n"The : ~ ~ town of HCsa, a roadstation of mine, lacks people; the postmaster and the commander of the recruits are there alone and cannot attend to it. Now, let me get together thirty families and place them there. There are men of the prefect Naba-u~allaliving in HCsa, a cohort of craftsmen; let him move them out, settle them in the town of Argite, and give them fields and gardens." Note that it was clearly seen as a priority to have the population of a conquered country return to their fields as quickly as possible. Nabil-hamatu'a, the deputy governor of Mazamua, reported to Sargon on the orders he gave to the people from the newly conquered country of Allabria:36 "[I said]: 'Do your work, each in [his house and] field, and be glad; you are now subjects of the king.' They are peaceful and do their work. I have brought them out from six forts, saying: 'Go! Each of you should build (a house) in the field and stay there!"'
The Assyrian kings' keen interest in plants and trees is illustrated by the detailed depiction of plants in their reliefs.37SOnaturalistic is the representation of the flora that it is possible to identify geographic regions with its help. But the kings did not only 32
ABL 685:12-15: Sa LUGAL be-li' iS-pur-an-ni (13) [ma]-aT A Sid-di ar-ki ina bat-bat-ti (14) [ i a ] (15) [a-mu]ra-ta-mar ma-a '-da a-dun-ni?. URU.HAL.$Ua-na a-ra-si Sa KUR.~U-[da-bir] 33 Inschrift auf den Tonzylindern:34-37, see A. Fuchs, Die Inschriften Sargons II. aus Khorsabad, Gottingen 1994, 37 and 292. 34 Annalen:378-381, see Fuchs 1.c. 169f and 335; GroBe Prunkinschrift:137-139, see Fuchs 1.c. 229f and 351f. 35 ABL 414 = SAA 1 177. 36 ABL 208 = SAA 5 210. 37 See E. Bleibtreu, Die Flora der neuassyrischen Reliefs. WZKMIS 1, Vienna 1980.
239
3 8 A. K. Grayson, Assyrian Rulers of the Early First Millenium BC I (1114-859 BC). RIMA 2, Toronto 1991, 27: A.0.87.1 vii 17-27. 39 Grayson 1.c. 290: RIMA 2 A.0.101.30: 36b-52. 40 See the relief sequence of room 7, especially the slabs 12-13, see P. E. Botta and E. Flandin, Monument de Ninive 11, Paris 1849, pl. 114 and P. Albenda, The Palace of Sargon King of Assyria. fiditions Recherche sur les Civilisations. Synthkse 22, Paris 1986, pl. 89-90; cf. also S. Parpola, The Correspondence of Sargon II, Part I. Lettersffom Assyria and the West. SAA 1, Helsinki 1987, 64 fig. 23 and Bleibtreu 1.c. 102 Abb. 34 (side-inverted!) and 110 Abb. 40. 41 Inschrift auf den Stierkolossen:4lf, see Fuchs 1.c. 66f and 304. Kleine Prunkinschrift des Saales XIV:28f, see Fuchs 1.c. 78 and 309. 4 2 ABL 938 = SAA 1 222 (tree saplings from Sadikanni), ABL 813 = SAA 1 226 (tree saplings from Nemed-Issar, Suhu and LaqC), ABL 814 = SAA 1 227 (tree saplings from Nemed-Issar), ABL 510 = SAA 5 27 (tree saplings from the northern border, possibly from Tidu), ABL 544 = SAA 5 105 (tree saplings from Tamnuna), CT 53 836 = SAA 5 268 (tree saplings of unknown provenance), CT 53 36 = SAA 5 281 (tree saplings from Suru in the Tur Abdin region), see also S. Parpola, "The Construction of ~ur-Sarrukinin the Assyrian Royal Correspondence," Khorsabad, le palais de Sargon II, roi dJAssyrie, ed. A. Caubet, Paris 1995, 58f. 43 NL 16 = SAA 1 110; it is not entirely clear whether this letter refers to the park around Dur&mukin or to another garden project, see Parpola 1.c. 74 n. 70.
240
K. RADNER
inscriptions he gives extremely detailed descriptions of them.44 Whereas Sargon attempted to bring the lush landscape of the northern Syrian mountain regions to Assyria, Sennacherib tried to recreate not only this paradise-like environment, but also the southern Babylonian marsh landscape in N i n e ~ e h . ~As 5 S. Dalley has shown his gardens, which are depicted in an Assurbanipal relief (fig. 4),46 were supported by an innovative irrigation system using the so-called Archimedes' screw. The story of the Hanging Gardens of the fabled Queen Semiramis of Babylon, praised as one of the wonders of the world in classical antiquity, may well be merely a distorted recollection of Sennacherib's pseudo-Babylonian park.47 But Sennacherib had other gardens built as well, such as the one surrounding the New Year festival house erected in the steppe outside the city of Assur. In a royal decree dedicating personnel to this temple Sennacherib describes the garden: "I encircled it (i.e. the New Year festival house) with [trees] of the orchard, all kinds of fruit trees and aromatic plants, as with a garland."4* The Assyrian kings used their parks as recreational areas. K. Delle14~was able to show that the Neo-Assyrian term for the bowery of vines depicted in the famous scene on an Assurbanipal relief, dubbed "Assurbanipal in der Gartenlaube" .by modem scholars (fig. 5),50 was qersu. The combination of textual and visual sources illustrates how the king and his closest confidants retreated to this bowery to find relaxation and peace.51 If access to the king's bowery was strictly denied to unwanted visitors, a letter to Sargon shows how officials would speculate to find the king accessible in the surroundings of his parks.52 But the gardens were not only meant for the enjoyment of the king. As Sennacherib states in his inscriptions he gave part of the garden land, subdivided in plots of a size of two pinu each, to the people of Nineveh so that they could have orchards of their own.53 It may seem that the image of the Assyrian kings as patrons of horticulture contrasts sharply with their frequently used war strategy to destroy the enemies' orchards. Both the accounts in the inscriptions as well as the depiction on the reliefs bear witness to the devastation of fruit groves by ruthlessly cutting down trees.S4 But the use of tree
(I
THE NEO-ASSYRIAN KING AND HIS LAND
24 1
destruction in war was the other side of the Assyrian kings' love for horticulture. As they deeply appreciated trees and their gifts to mankind they perfectly understood how to use their destruction as an instrument of terror. As S. Cole was able to show, orchards were not destroyed at once but the precious trees which take decades to reach maturity were gradually cut down one by one in order to persuade the inhabitants of a besieged city to surrender.
We may conclude that the Assyrian kings, being familiar with regions where agriculture and horticulture were a matter of course, were aware of the possibilities to create a similar environment in less fortunate areas and were eager to optimise the land at their disposal. Big cultivation projects in the steppe and large scale landscape gardening were the direct results of the kings' role as supreme fanner and gardener. But whereas the kings portrayed themselves as the masters of earth in their inscriptions, stressing how they conquered deserts, mountains and oceans, cultivated barren lands and connected distant parts of the world by constructing cities, canals, bridges and roads, they were at the same time well aware of the fact that it was impossible to fully dominate nature with manpower and advanced engineering. With these means alone, plagues such as epidemicss5 and locust swarm@ and natural disasters such as flooding,57 storms58 and earthquakes59 were impossible to control. The king60 tried to prevent such catastrophes preferably before they even came into existence. Information on them was collected with the help of a diversity of scholars who interpreted astronomical and terrestrial omens gained from watching the skies, analysing sheep livers and generally keeping the eyes open for any kind of unusual event. Once the impending danger was detected, the scholars endeavoured to prevent the catastrophe's
-
44
For attestations see E. Frahm, Einleitung in die Sanherib-Inschriften. AfO Beiheft 26, Vienna 1997,269 sub 32) and cf. 277f. 45 J. A. Brinkman, "Reflections on the Geography of Babylonia (1000-600 B.C.)," Neo-Assyrian Geography. QGS 5, ed. M. Liverani, Rome 1995, 29. 46 BM 124939, see S. Dalley, "Nineveh, Babylon and the Hanging Gardens: Cuneiform and Classical Sources Reconciled," Iraq 56 (1994) 51 fig. 1. 47 Dalley 1.c. 45-58 and see also S. Dalley, "The Hanging Gardens of Babylon at Nineveh," Assyrien im Wandel der Zeiten. HSAO 6 (= CRRA 39), ed. H. Hauptmann and H. Waetzoldt, Heidelberg 1997, 19-24. 48 VAT 9656 = SAA 12 86:20, cf. D. D. Luckenbill, The Annals of Sennacherib. OIP 2, Chicago 1924, 137: KAH 2 122 = I 2:34f. 49 K. Deller, "Assurbanipal in der Gartenlaube," BaM 18 (1987) 229-238. BM 124920, see Deller 1.c. pl. 13 for a reproduction of the relief and 1.c. 229 n. 3 and n. 7 for the most important earlier literature. Deller 1.c. 238. 5 2 ABL 843 = SAA 1 160:4-9: "[I] stood [alongside] the king's road, [in fronlt of the gardens, but the king did not pay attention to me, speaking as he was with Rasappaiu." 53 D. D. Luckenbill, The Annals of Sennacherib. OIP 2, Chicago 1924, 97: A 1:88 and 101: B 1:58. 54 See S. W. Cole, "The Destruction of Orchards in Assyrian Warfare," Assyria 1995, ed. S. Parpola and R. M. Whiting, Helsinki 1997, 29-40 for a survey of the textual and visual sources; cf. also B.
Oded, "Cutting Down Orchards in Assyrian Royal Inscriptions - The Historiographic Aspect," JAC 12 (1997) 93-98. 55 Such as the one reported in NL 18 = SAA 1 171 to Sargon. 56 Locust swarms are reported in several letters of Sargon's correspondence: ABL 1015 = SAA 1 103, CTN 2 240 = SAA 1 104, ABL 910 = SAA 1 221, cf. also NL 103. Sennacherib compares his enemies who constantly engaged him in battle with the onset of locust swarms in spring, see D. D. Luckenbill, The Annals of Sennacherib. OIP 2, Chicago 1924,43: H 2 v 56. 57 A report by the crown prince Sennacherib to his father Sargon on flooding in central Assyria is found in the letter ABL 713 = SAA 1 36. Sennacherib quotes the heavy storms and the dangers caused by the swollen mountain rivers as the reason for the termination of his seventh campaign in 693 against Elam, see Luckenbill LC. 41: H 2 v 7-11. See also F. M. Fales, "Rivers in Neo-Assyrian Geography," Neo-Assyrian Geography. QGS 5, ed. M. Liverani, Rome 1995, 205f. 5 8 storm ~ which destroyed a camp is reported to Sargon in the letter CT 53 197 = SAA 5 249. The Assyrian kings compared the destructive force of their army frequently with that of a storm, for references in Sargon's and Sennacherib's inscriptions see A. Fuchs, Die Inschrifen Sargons II. aus Khorsabad, Gijttingen 1994, 150 and 330: Annalen:296, and Luckenbill 1.c. 28: H 2 ii 15. 59 For a report on an earthquake at ~ur-Sarrukinsee ABL 191 = SAA 1 125. The subject has been recently studied by A. Fadhil, "Erdbeben im Alten Orient," BaM 24 (1993) 271-278. 60 Whereas all known letters and reports from scholars to the king date into the reigns of Esarhaddon and Assurbanipal, it is clear that also their predecessors used the service of scholars, see, e.g., the letter ABL 1216 = SAA 10 109 for a reference to the activity of scholars under Sennacherib.
0
242
K. RADNER
THE NEO-ASSYRIAN KING AND HIS LAND
realisation with the help of the appropriate rituals61 While modern man may be reluctant to put faith in these measures they certainly succeeded in comforting the Assyrian king and his people. The general awareness of possible catastrophes alone will have contributed its share to restrict the damages once they took place.
Although privately owned land existed and could be freely bought and sold, the majority of the land was owned by the state, especially as newly conquered and newly cultivated land automatically belonged to the king. However, the king chose to cede the possessory rights of large quantities of land to officials, temples and soldiers. By distributing state-owned land to have it worked without giving up the absolute titles to it the central administration was greatly relieved and at the same time the emergence of independent great landowners who would endanger the absolute power of the king was prevented. State-owned land given as maintainance land, called ma'uttu,67 to the holder of an office such as a provincial governor was the primary means to sustain the administration. Maintainance lands are also attested for such institutions as the royal tombs in Assur.68 While a fixed share of these estates' yield had to be handed over to the state to support the central administration the remaining share was to sustain the holder's office. This is especially clear from a letter to Sarg0n,~9written by a dismayed provincial official who could not see how he should be able to provide the state with the assigned quota of 1000 homers of corn and still maintain his office. He claimed that unlike his colleagues in adjoining provinces who were able to meet their quota and still could feed both humans and horses as well as use the corn as seed he was not able to do so. In addition to the taxes the proprietors of state-owned land had to complete state service, called ilku.70 By using its crops as temple offerings and as provisions for the temple staff, land was the most important means to support the temples. It was the king's responsibility to see to it that the temples had sufficient land at their disposal. He did so by donating land to the temples,71 but also by bestowing land onto people who were supposed to use part of their fields' yields as provisions for the temples. This practice is best attested in a grant by Sargon who, when some land in the town of bakers, which Adad-nirari 111 had earlier exempted from taxes and given to certain families in order to provide the A63ur temple with offerings, became useless due to the expansion of the city of Maganuba into Sargon's new residence ~ur-Sarrukin,exchanged this land against fields in the town of clergymen in the district of N i n e ~ e h . ~ ~
As a conclusion to this paper we shall briefly touch upon the role of the Assyrian king as the supreme landlord by summarising how he both administered and held together his land by distributing it among officials, temples and soldiers without ceding the absolute titles to it. The character of land tenure in the Neo-Assyrian period was the direct result of the developments in the Middle-Assyrian period.62 Whatever the exact nature of the evolution of land tenure, by the first millennium landed property could not only be owned by the king (i.e., the state), but also by private individuals, individually or jointly. There is no direct evidence for land owned by cities and villages. However, as the organisation of concerted use of agricultural land must have been one of the major responsibilities of the municipal government, the community's importance in respect to landed property should not be underestimated63 and it seems that the consent of the municipal government was needed for the transfer of ownership within the community's jurisdiction.64 Unless the king granted tax exemption,65 the tenure of land was linked to the duty to pay taxes, specifically the 3ib3u tax on corn and the nusikz.2 tax on straw.66 6
See most recently S. M. Maul, Zukunftsbewiiltigung: eine Untersuchung altorientalischen Denkens anhand der babylonisch-assyrischen Loserituale (Namburbi). BaF 18, Mainz 1994, especially chapters 11. and 111. 62 J. PeEirkovB, "On Land Tenure in Assyria," Fs. L. Matous' 11, ed. B. HruSka and G. Komorbczy, Budapest 1978, 187-200 summarises the different opinions of various scholars, notably Diakonoff, Garelli, Jankowska and Postgate, on Assyrian land tenure and its evolution. See further F. M. Fales, "A Survey of Neo-Assyrian Land Sales," Land Tenure and Social Transformation in the Middle East, ed. T. Khalidi, Beirut 1984, 1-13 and "The Neo-Assyrian Period," Circulation of Goods in Non-Palatial Context in the Ancient Near East. Incunabula Graeca 82, ed. A. Archi, Rome 1984, 207-220, J. N. Postgate, "The Economic Structure of the Assyrian Empire," Power and Propaganda. A Symposium on Ancient Empires. Mesopotamia 7, ed. M. T. Larsen, Copenhagen 1979, 193-221, "ilku and Land Tenure in the Middle Assyrian Kingdom - A Second Attempt," Societies and Languages of the Ancient Near East. Fs. I. M. Diakonoff, ed. M. A. Dandamayev et al., Warminster 1982, 304-313 and "The Ownership and Exploitation of Land in Assyria in the 1st Millennium B. C.," Reflets des Deux Fleuves. Fs. A. Finet. Akkadica Supplementum 6, ed. M. Lebeau and Ph. Talon, Louvain 1989, 141-152 (= "Grundeigentum und Nutzung von Land in Assyrien," Grundeigentum in Mesopotamien. Jahrbuch fiir Wirtschaftsgeschichte 1987/S, ed. B. Brentjes, Berlin 1988, 89-110). 63 Cf. J. N. Postgate, Akkadica Supplementum 6, 144. 64 This can be shown best in the case of property within the city of Assur: the sale documents not only have to be sealed by the seller, but also by the city officials, see E. Klengel-Brandt and K. Radner, "Die Stadtbeamten von Assur und ihre Siegel," Assyria 1995, ed. S. Parpola and R. M. Whiting, Helsinki 1997, 137-143. According to CTN 2 44, a text from Kalhu, a building plot is bought and the mayor (hazannu) of the city of [...I seals the text alongside with the seller, see 1.c. 138. 65 Postgate LC. l49f. On Sibs'u and nusah? see J. N. Postgate, Taxation and Conscription in the Assyrian Empire. Studia Pohl SM 3, Rome 1974, 174-18 and P. Garelli, "Le systkme fiscal de l'empire assyrien," Points de vue sur lafiscalit6 antique, ed. H . van Effenterre, Paris 1979, I If.
67
For ma'uttu Sa s'arri "maintainance land of the king" see CTN 3 14:4, CTN 3 16:4 and NL 52:5, for ma'uttu s'a ekalli "maintainance land of the palace" see CTN 3 87 r. 11. It seems that in contrast to privately owned land, maintainance lands were always described as the land of a certain official, without giving his proper name, see Postgate, Akkadica Supplementum 6, 146f. The problem of distinguishing maintainance lands from privately held estates is not restricted to officials, but also arises in the case of the king and his family. 68 TIM 11 33:s': ma-%-u-te $a fi LUGAL-n[i]. 69 CT 53 79 = SAA 5 225. 7O On ilku see Postgate, Studia Pohl SM 3, 63-79, also in Mesopotamia 7, 203-205, in Fs. I. M. Diakonoff, 304-313 and Garelli 1.c. 8-11; on the link between ilku and land tenure see also J. N. Postgate, Fifty Neo-Assyrian Legal Documents, Warminster 1976,24f and cf. A. K. Grayson, "Assyrian Civilization," The Assyrian and Babylonian Empires and other States of the Near East, from the Eighth to the Sixth Centuries B.C. The Cambridge Ancient History IW2, ed. J. Boardman et al., Cambridge 1991 (second edition), 213f. 7 1 The letter ABL 480 = SAA 1 106 concerns land given by Sargon to the Nabfi temple in Dur$mukin. In general cf. Postgate, Akkadica Supplementum 6, 145f. 72 NARGD 32 = SAA 12 19:23'-33'.
244
K. RADNER
THE NEO-ASSYRIAN KING AND HIS LAND
A reference to "bow land" in a letter to Sargon shows that soldiers received shares of state land as "fiefs" which at least sometimes could be exempted from taxes.73 This system was adopted by the Chaldean empire74that passed it on to the Achaemenid empire.75 Due to the scarcity of references we do not know whether these fiefs could be inherited by the original recipient's heirs or whether the land was returned to the state after the recipient's death. The latter, however, would seem to be more likely. Fig. 1. Sargon's hieroglyphs: a fig tree and a seeder-plough, standing for "the country of Assyria".
Sources of illustrations
Fig. 1 Fig. 2 Fig. 3
Fig. 4 Fig. 5
I. Finkel and J. Reade, ZA 86 (1996) 249 (fig-tree) and 250 (seeder-plough). I. Finkel and J. Reade, ZA 86 (1996) 260 (seeder-plough and palm-tree). Author's drawing based on E. Botta's original drawings of slabs 12-13 of room in Sargon's palace in ~ur-&mukin (Chorsabad), published in P. Albenda, The Palace of Sargon King of Assyria. ~ditionsRecherche sur les Civilisations. Synth6se 22, Paris 1986, pl. 89-90. BM 124.939, drawing reproduced from S. Dalley, Iraq 56 (1994) 51 fig. 1. BM 124.920
Fig. 2. Esarhaddon's hieroglyphs: a seeder-plough and a palm-tree, standing for "Assyria and Babylonia".
73 74
Postgate 1.c. 148 and in Studia Pohl SM 3, 223 on ABL 201 = SAA 5 16:6: A.SA GIS.BAN-Su. See M. Jursa, Der Tempelzehnt in Babylonien vom siebten bis zum dritten Jahrhundert v. Chr. AOAT 254, Miinster 1998, 14-18 and M. Jursa, "Bogenland schon unter Nebuchadnezar II.," NABU 19981124 on the attestations for bit qaSti in the Chaldean period. 75 Thanks to the evidence presented by Jursa it is unnecessary to hypothesise that the system was transmitted from the Assyrians to the Achaemenids via the Medes, as has been suggested earlier due to the lack of attestations from the Chaldean period (most recently M. Stolper, "Milit;irkolonisten," RLA 813-4 (1994) 206a and M. Dandamayev, "Assyrian Traditions during Achaemenid Times," Assyria 1995, ed. S. Parpola and R. M. Whiting, Helsinki 1997, 43). ,
, after reliefs from ~ u r - h m k i n . Fig. 3. Sargon's park in ~ u r - k r r u k i ndrawing
246
K. RADNER
PIHANS LXXXVIII, 2000
*
Land Tenure in Northern Mesopotamia: Old Sources and the Modem Environment R. M. Jas (Amsterdam) 1.
INTRODUCTION
That the different agricultural regimes in North and South Mesopotamia have caused different systems of land tenure can still be seen in modern Iraq. D. Warriner, in Land and Poverty in the Middle East* has remarked on this in the following words: "There are two different varieties of tenure in the two zones, determined by the different conditions of agriculture. In the north, the forms of tenure are similar to those of Syria, with a class of small proprietors taking some, but not all, of the land. In the south large owners or sheiks own virtually all of the land, letting it to share-tenants, through a series of intermediary lessees, minor sheiks, sirkals (formerly sub-sheiks), or town notables." As the title of Warriner's book already indicates, the role peasant poverty plays in perpetuating these particular forms of land tenure is great, also not exactly a new phenomenon, as we will see when we look at the old texts. In fact, no other topic lends itself better to illustrate the ever threatening "easy slide into debt" and the frequently following total dependency of Mesopotamian smallholders than that of land tenure. For a better understanding of a few characteristic traits of the written documentation on land tenure in Northern Mesopotamia it is perhaps useful to start with a description of the current precipitation regimes and agro-ecological zones in the areas from which the textual material used in this paper comes from, and to give a definition of the term rainfed agriculture. In our attempt to bring out the precise nature of the written documentation, we will stress the fact that only unusual situations were written down and suggest that the majority of 'regular' land leases probably were oral, like many still are in the Near East today. Those texts that were written down a11 have a clear creditor-debtor relationship as a background. Moreover, it is not the wealthy landowner who leases out his land, but the farmer borrowing money from him, who places his own land as a pledge in return. The creditor thus gains the usufruct of the land of his debtor, serving as interest, and only after the term of the loan has passed the debtor is allowed to pay back the capital and take back the pledged land. To conclude, we will describe the puru system, a poorly documented risk-evading strategy we nevertheless believe to be typical for rainfed agriculture.
Fig. 5. Assurbanipal's vine bower in Nineveh, scene of an Assurbanipal relief from Nineveh (BM 124.920).
Warriner 1948: 104.
248 2.
LAND TENURE IN NORTHERN MESOPOTAMIA
R. M. JAS
THE MODERN ENVIRONMENT
Iraq consists of four major natural regions: the mountains, the foothills or terrain region, the alluvial plain and the desek2 A simple division of Northern Mesopotamia, which has the benefit that originally it was made for practical purposes, and therefore can be expected to correspond more to the conception of the environment found in the old sources than the strictly scientific one used by A. Abdulsalam for his TAVO map,3 is the division made for the Naval Intelligence Division in 1944.4 Northern Iraq is divided into the following three regions: 1) The Jazira of Iraq (Middle-Assyrian and Neo-Assyrian sources). 2) Foothills and Plains of Southern Assyria: the regions south-east of the Little Zab (Nuzi, Middle Assyrian and Neo-Assyrian sources). 3) The Foothills and Plains of Central and Northern Assyria. Central Assyria consists of the land between the Tigris and the two Zabs and Northern Assyria of the land between the Tigris, the Great Zab, and the Jabal Bakhair and Aqra Dagh (Middle Assyrian and Neo-Assyrian sources).
Within the North various zones can be distinguished. The definition of zone by the Oxford English Dictionary as "a definite region or area of the earth, or of any place or space, distinguished from adjacent regions by some special quality or condition (indicated by a defining word or phrase)" makes it possible to use the word on different levels. To avoid confusion, therefore, we will have to decide first which zones will have our main interest. A good example of a large-scale zone is the one owing its existence to the paucity of rainfall, the so-called low-rainfall zone of West Asia and North Africa, or WANA in F A 0 terminology. This zone includes twenty countries stretching from Morocco to Afghanistans and obviously can be of very little help to us here. Even the several major climates6 that can be distinguished within this zone still cover areas that are too large for us, but here the contours at least begin to emerge. Iraq, with its sub-tropical, continental, arid climate? with wetter parts in the north and an arid region in the south,8 is clearly different from the western parts of Syria, where a more Mediterranean climate prevails. The circumstances in the parts of Turkey with cold winters are remarkably similar to those in Northern Iran and Afghanistan.9 It is not, however, until we take other factors into account as well, such as soil and relief, that more specific zones start delineating themselves. For these particular zones
249
F A 0 and ICARDA have suggested the term agro-ecological zones. Their main characteristics can be deduced from the following statement by the FAO: "The Agroecological Zone Study takes into account soil production capabilities, length of growing season, as defined by moisture availability and cold hazards, identifies suitability classes for cultivation of the major crops under rainfed conditions."lO The F A 0 definition of rainfed agriculture is as follows: "Under rainfed agriculture, moisture available to the growing plant is the host significant single factor limiting yields."ll This seems to be exactly the type of zone that we are looking for. First of all, this type of zone can be neatly, but not necessarily rigidly, defined. In addition, a type of zone that takes not only the various natural regions and precipitation regimes into account but land use as well has excellent possibilities for making the distinction between positive land-use and negative areas, taking a fluctuating uncertainty area into account as we11.12 It seems only obvious to assume that we find this most important distinction either directly expressed or, at least, implied in much of the ancient written documentation. Whether we call them rainfall based agricultural stability zones13 or speak of positive land use versus negative areas and an uncertainty zone,14 or distinguish, as TAVO does, between an inner zone of rainfed agriculture with relatively certain yields, and an outer zone with high yield uncertainty, each with three gradations of its own,l5 or use the term low-potential rain-fed areas16 is in the end not very important. What is important is that these zones fluctuate, which could at times cause major upheavals. Because much more research on this topic has been done in Syria than in Iraq, we will start with that country. Based on the intensity and reliability of rainfall, Syria has been divided in five agro-climatological zones in F A 0 1982:85 (plus table 52 on p. 86): Apricultural Zone No. 1 (high rainfall) : This zone is subdivided into two regions: Region (a) has rainfall of over 600 mm (humid region) where a number of crops can be grown under rainfed conditions; and Region (b) with rainfall of over 300 mm (350 to 600 mm) where two good wheat crops can be obtained each three seasons, and summer crops or pulses can be grown in rotation with winter cereals. Zone No. 2 (medium rainfall) : Rainfall in this zone ranges from 250 to Agricultural 300 mm in 66 percent of the seasons, and this zone produces a good crop of barley in two out of three seasons. Wheat and food legumes may be planted also in this zone.
2
See Boesch 1939; Lebon 1953; Davies 1957; Wirth 1957; Zuidervliet 1981:ll; FA0 1982:41; Ockerman and Samano 1985; Beaumont et al. 1988. 3 TAVO A VII 1 Vorderer Orient. Naturraumliche Gliederung (Middle East. Natural Regions), 1988 (with a separate sheet). Geographical Handbook, 1944. Oram and de Haan 1995:5. 'A broad climatic division, defined in terms of monthly temperatures, seasonality of rainfall, and temperature regime', see FA0 1983:229. Buringh 1960:42; Thalen 1979:17. Ockerman and Samano 1985:192. Oram and de Haan 1995:xv.
FA01980:15. FA0 1980:14, see also Beaumont et al. 1995:140. See Davies 1957. Tutwiler 1995:18-9. l4 Davies 1957. 15 Hkle 1985: *inner zone with relatively certain yields (mainly wheat and barley, also legumes) and *outer zone with high yield uncertainty (virtually only barley and wheat). Both zones know the three gradations: *dominant, *partial, *sporadic (favorable locations). ti Adams and Holt 1985:66. lo
250
25 1
R. M. JAS
LAND TENURE IN NORTHERN MESOPOTAMIA
Agricultural Zone No. 3 (low rainfall) : Rainfall in this zone remains around 250 mm in 50 percent of the seasons. This zone is suitable for barley growing, and some legumes are also included in rotation with small grains.
In his article, Mahdi concentrates mainly "on the farming system that has evolved in the drier marginal transitional zone of unreliable rainfall, with an annual average ranging between 200 and 350 mm"21 and concludes that "apart from the well-watered valleys in the Kurdistan mountains, current rainfed land-use practices do not appear to vary widely between the drier and less dry areas."22
These are marginal lands with rainfall of 200 to : 250 mm. Rainfall does not drop below 200 mm in 50 percent of the seasons. Lands in this zone are utilized for growing barley or grazing. This is the desert region of the country which Agricultural Zone No. 5 (desert) : has very low rainfall and supports no crops at all. The only use of this region is for grazing. This division, with several refinements, is used in Cocks et al. 1988:5f. He speaks of the six agricultural stability zones: Zone No. 1 : 1 (b) The main crops are wheat, chickpeas, lentils, fruit, vegetables and nonirrigated summer crops. Zone No. 2 : Barley, wheat, food legumes, and summer crops are grown, and livestock are of increasing importance. Zone No. 3 : Barley and livestock are the main products, wheat is also grown, while food legumes are of only slight importance. Zone No. 4 : Barley and livestock are again the main products, the former chiefly grown to feed the latter. Zone No. 5 : Not considered useful for arable agriculture. It is however widely used for grazing. This is the division T. J. Wilkinson uses for his description of the agricultural potential of Northern Syria and Iraq.17 It is not certain, however, that these zones are valid for Northern Iraq as well. The article by K. A. Mahdil8 in Jones, Mathys and Rijks (1993) on Farming Systems, Market Conditions and the Production of Barley in Northern Iraq seems based on a rather different set of data, more into line with the figures from Harnmam al-Alil, 30 krn south of Mosul, the most recent and reliable set of data available from this area.19 Mahdi starts out by stating that "in the rainfed zone of Northern Iraq, extensive cultivation has continued to be widely practiced from the foothills and upland plain belt to the north and north east, extending into the marginal steppe beyond that belt and into the wide expanse of the Jazira plain west of the Tigris up to the Syrian Wilkinson 1997:72. Mahdi 1993:216-229. l 9 Tamimi and Young 1980; Fakhry and Sultan 1980. 20 Mahdi 1993:216.
With these figures in mind we still need to take Wirth's chapter On the sense and countersense of a "drought limit for rainfed agriculture23 into account as well. This chapter is far from outdated by the mass of recent data, and the clarity with which Wirth defines the problem remains admirable. The most useful supplements to E. Wirth's study are the maps and the accompanying book that M. Alex prepared for the Tubinger Atlas des Vorderen Orients.24 According to Wirth setting the drought limit for rainfed agriculture in the Near East is first of all difficult because the various agricultural systems do not neatly fall into the two categories rainfed agriculture and irrigation agriculture. A second problem is that the drought limit for rainfed agriculture cannot be localized on a map as either a line or a strip. Mean annual rainfall and the drought factor25 are less important for the drought limit than relief and the soil-moisture depending on it. The drought limit for rainfed agriculture is the border between areas that are taken up exclusively by fields with rainfed agriculture, and areas where rainfed agriculture shows lacunae.26 All attempts to set the drought limit for rainfed agriculture at 250mm are seriously hampered by the fact that the annual variability of rainfall in the arid zones of the Near East is very high. Cultural-geographical factors play a considerable role as well. Wirth points out that in the past the actual distribution of rainfed agriculture never coincided with the possible distribution. Above all things, the drought limit for rainfed agriculture is essentially the limit of rentability.
Mahdi 1993:217. Mahdi 1993:218. 23 Von Sinn und Widersinn einer "Trockengrenzedes Regenfeldbaus," Wirth 1962:12-23. 24 TAVO A IV 1 Vorderer Orient. Mittlere Jahrestemperaturen (Middle East. Mean Annual Air Temperature) 1983. TAVO A IV 2 Vorderer Orient. Mittlere Januartemperaturen (Middle East. Mean Air Temperature, January) 1983. TAVO A IV 3 Vorderer Orient. Mittlere Jahresniederschlage und Variabilitat (Middle East. Mean Annual Rainfall and Variability), Wiesbaden 1984. TAVO A IV 5 Vorderer Orient. NiederschlagsverlaJlichkeit (Middle East. Rainfall Reliability), Wiesbaden 1985. TAVO A IV 6 Vorderer Orient. Jahresgangtypen des Niederschlags (Middle East. Seasonal Rainfall Patterns) 1985, with Alex (1985). 25 The ratio between precipitation and evaporation. 26 Wirth 1971:146 again raises this point, this time in relation to ancient farming: 'Allerdings diirfen wir zu keiner Zeit eine flachenhafte Besiedlung des Landes mit liickenlosem Anbau annehmen. Schon damals trat die Insel- und Kernstruktur der syrischen Kulturlandschaft klar zutage.' 21 22
252
R. M. JAS
LAND TENURE IN NORTHERN MESOPOTAMIA
Finally it is important to distinguish between areas where only arable farming takes place, and areas where farmers have settled as well.27
The most frequent amount (five times31) is 10 shekels of silver per emiiru, with twice32 a term of six years (= 1 213 shekels per year) and once33 eight years (1 114 shekels per year). This standard rate34 occurs in both the pledge and the 'lease' group. The lowest rate is 112 shekels per emiiru per year (term: six years).35 Postgate has compared these rates with the prices of fields.36 A few uncertainties remain, but the list drawn up by C. Zaccagnini37 shows that three times38 there is question of a price of 3 shekels per emiiru, but that prices as high as 15 shekels are attested as well.
3.
PLEDGE OR LEASE? - THE NATURE OF THE EVIDENCE
Neo-Assyrian lease documents without a debt as the prime motive for the transaction are unknown to me. In all texts28 real estate is given as a pledge against a loan. We are not looking at absentee landlords who lease out their land to tenants, but, instead, at smallholders borrowing money from wealthy people, in return for which they deposit their land in pledge. Whatever produce the pledged lands yield serves as interest on the loan. Only after the stipulated term of the loan (usually six or eight years) has elapsed, the debtor is allowed to pay back the capital, thus, theoretically at least, regaining his property. In case he could not pay the loan-and we may assume this frequently was the case-the pledged object automatically became property of the creditor, who until then only enjoyed the usufruct of it. The distinction made by K. Radner between Verpfandung von Feldem (XI. 2. C. b. 2. 1) and Verpachtung von Feldem (XI. 2. C. b. 2. 2) is based on the formal criterion whether the phrase ana Saparti Zakanu 'to place as a pledge' occurs in the text or not. Texts lacking this phrase are seen not as pledges but as leases. Looking at the contents of these texts it turns out there are very few differences between these two groups, consisting of thirty-six texts in all. The most important resemblance between the two groups is the presence of a clause in every text laying down the right of the debtor to regain his property by paying off the loan once the term of the loan has passed. A second indication that texts in which the ana Saparti Sakiinu clause does not occur are still to be interpreted as pledges instead of leases is the occurrence in both groups of the (an)duriiru29 clause. ADD 629 (SAA 6 226), strictly speaking a lease document, states: "In the case of a remission of debts S[ilim-ASSur = the creditor] will see his money." ADD 73-74 (SAA 6 259-260), from the pledge group, has a similar clause. A clause like this, exclusively protecting the interests of the creditor, seems to be more in place in texts that have a debt as their background than in lease documents. J. N. Postgate has called the entire group "pledge leases,"30 and connected them with unpaid debts. He has stressed the fact that in all cases we are dealing with pledges instead of leases. This is important for a correct appraisal of the 'rates' found in these texts.
27 Compare Davies 1957:127-8: 'Figure 3 also indicates the southern limit of rural settlement in unirrigated areasie., dependent upon rainfall agriculture. This is vital to an appraisal of land use patterns, since virtually all cultivation is carried on by permanently settled farmers; and in Iraq there are practically no settled rural people who are not cultivators.' Because of the improved means of transportation in this century, this is the least relevant of the Wirth rules for a better understanding of ancient farming. 28 Studied most recently by Radner 1997:384-89. 29 An (an)durGru is a debt remission proclaimed by the government. In Nuzi, sometimes the Hurrian word kirenzi is used instead, see Fincke 1995:36. 30 Postgate 1989:150 n.103.
253
Of very special interest are two unsealed texts from Kalhu39 and one from Nineveh40 which appear to be yearly contracts between owner and cultivator, but due to the excessively curt phrasing not all details are ~ l e a r . The ~ l gap between notes like these and oral agreements appears not to be very large. We also have to take the possibility into account that 'regular' leases were, as a rule, destroyed after the harvest. For the former possibility one could compare the modern Near-Eastern practice as described, for t ~is~ at the threshing floors43, of course, that sharecropping example, by P. B e a ~ m o n 'It contracts are fulfilled with the division of the bagged grain.' The rates in these three documents are 2 shekels of silver for 7 sutu of field, 3 shekels of silver for 3 emiiru of fallow land44 and 2 shekels for one emiiru of fallow land 31 ADD 58 = SAA 6 81, ADD 73-74 = SAA 6 259-260, ADD 83-84 = SAA 6 223-224, ADD 622 = NALK 105 and ADD 1154 = SAA 6 268. 32 ADD 622 = NALK 105 and ADD 1154 = SAA 6 268. 33 ADD 83/84 = SAA 6 223-224. 3 4 MADD 8 comes close with 1 315 shekels. 35 ADD 623 = NALK 128. 36 Postgate 1989:151. 37 Zaccagnini 1979:25. The extreme amount of 160 shekels of silver for one emiiru (ARU 436+ = ADD 384+ = NALK 319) can easily be corrected by restoring copper instead of silver. This restoration is more likely because the text dates from the early 8th century (see Postgate 1973, 29 n. 6). 38 ARU 427 = ADD 381 = SAA 6 126, ARU 431 = ADD 410 = NALK 291, ARU 397 = ADD 374 = SAA 6 161. 39 ND 3457 (Iraq 15 [I9531 145) and ND 7095 (CTN 3 No. 45). 40 IM 76890 = TIM 11 4. See Postgate's commentary in CTN 3, 102. 42 Beaumont et al. 1988:176. 43 Assyrian adru (Postgate 1976:38 3 3.3.1) and Nuzi magrattu (Richter 1995:69-79). Against a temporal meaning of the clause 'he will pay at the threshing floor' (i.e. after the harvest) speak those occurrences in which a geographical name follows, and the fact that threshing probably took several months and would therefore be a rather inexact term for paying a loan. According to Barazi 196357, threshing lasted from June until October. 44 See Isom and Worker 1979:202: 'Many of the semi-arid regions of the world can only produce a crop successfully and economically through an alternating crop and fallow system.' According to these authors 'the objectives of an alternate crop-fallow system are: (1) to conserve moisture and nutrients and make them more available to the crop, (2) to control weeds and minimize losses from weeds, insects and diseases, (3) to expedite seedbed preparation for the crop which follows the fallow.' It is not so much the question how useful fallowing is under rainfed circumstances, but the question whether it was practiced, and according to what schedule (biennial, triennial) that should concern us here. Modern Near Eastern
254
R. M. JAS
LAND TENURE IN NORTHERN MESOPOTAMIA
(rates that according to Postgate need to be halved before they can be compared with the long-term contracts that reckoned with fallow for half of the years45).In ND 3457 (see below) and TIM 11 4 the leaving of the lessee (el6 'to riseY46)and freedom of straw- and grain-taxes are stipulated. TIM 11 4 has a penalty clause: in case he does not pay the costs of the field the lessee has to deliver the yield of a field of one emaru (the surface of the field in question).
The s'ibs'u and nusahi taxes (respectively 114 and 1/10 of the yield in ADD 623 [NALK 1281) occur in both the pledge and in the 'lease' group and the importance these agricultural taxes have in this context has been repeatedly stressed by Postgate. In the pledge group two cases of freedom of these taxes are attested48 against one obligation,49 while in the 'lease' group we see four cases of freedom50 and two obligations.51 On the basis of these facts it is impossible to conclude what was normal under which circumdances.52 Zaccagnini 1979, has compared the Neo-Assyrian texts with the tidennutu contracts from Nuzi (p. 25f.) while treating the pledge and the 'lease' group as one in his chart on ~p. 26. The first thing that needs to be stressed when talking about land tenure in Nuzi is the exceptionally high customary rate of interest of 50 percent.53The other possibility to obtain credit in Nuzi was to give a piece of real estate in tidennutu in return for a sum of money or for something else like grain, cattle, tools, textiles or a slave. After a fixed term (the longer the better for the creditor) the capital had to be returned, the yield of the pledged field serving as interest only, exactly as in the Neo-Assyrian pledges. That the value of the borrowed good was much less than the yield of the land placed as a pledge quickly becomes obvious from the table in Jordan 1990.
ND 3457 (Iraq 15 [I9531 145)
(1) 3 ANSE ka-rab-be (2) Sa m S h l - m u - ~(3) ~m d ~ ~e-ra(!)-AS ~ (4) - e-si-id ~ G~~ ga-lab (5) e-li la SE Si-ib-Se (6) la nu-sa-be (7) 3 GIN KU.BABBAR gi-mir (8) A . S A - S ~i-ti~ ~ IGI - ~m ~ D a - g i l - (1 ~ 1) ~ IGI ~~~~ Si (9) ITU.GAN (Nov./Dec.) lim-mu m A S - S u r - ~(10) msam-si-i (12) IGI mdp~-zu '3 homers of fallow land, belonging to Sulmu- el, SamaS-Sm-uSur will plow, harvest and stubble47 -he will leave. No straw- or grain-taxes. He took 3 shekels of silver, the costs of his field. Date; witnesses.' SamaS-Sm-uSur was an elderly and wealthy bird-catcher in Kalbu, who hardly can have been expected to perform the work on the field himself, although, formally, he is without a doubt the lessee. That eventually also these texts could have had a debt as their background not only follows from the use of the verb naSQ(as in: ina puhi nag0 'to borrow'), but also from the following text: ND 2342 = IM 64003 = FNALD 28
(I) NA~.KISIBmSi-qi-15 (2) 10 G~N.MESK~J.BABBAR(3) Sa mEz-bu [three seal impressions] (4) ina IGI-Sdina 1TU.APIN Sa ur-ki BURU14 (5) K~.BABBARina SAG.DU-SG SUM-an (6) Sfim-ma la id-din-ni (7) ina 4-ut-ti-SG GAL-bi (8) 1 ANSE 3-BANka-rap-@(9) e-ra-6S e-si-di (10) 6-ga-la-pa e-li 'Seal of Sinqi-IHtar. 10 shekels of silver from Ezbu are at his disposal. In the 8th month (Oct./Nov.), after the harvest, he will pay back the silver in its original amount. If he does not pay it, it will increase by one fourth. One homer and three sutu's of fallow land he will plow, harvest and stubble-he will leave. Date and witnesses.'
practices point to a tradition of biennial fallow (see for example Morrison 1939:31-32). Studies of ancient Greek agriculture are characterized by curious reversals of opinion on this issue. See Scheidel in Garnsey 1998:196. The most recent studies apparently again agree on biennial fallow in classical Greece. 45 With TCL 9 66: 8-9 as the exception: 7 Sandu 7 m2ruSu '7 years: 7 years under cultivation.' 46 The suggestion to translate el6 as 'leave' in this context was first made by Parpola 1975:295. 47 In Nuzi qaqqada ;a eqli gullubu, literally 'shaving the head of the field,' is used, see Grosz 1988:92.
-
~
~
48
ADD 83-84 = SAA 6 223-224: a total of six emaru of field, freedom of taxes; ADD 70 = SAA 6 252: x emiru of field, freedom of taxes, no grain- or straw-taxes. 49 ADD 81-82 = NALK 352: he will pay grain- and straw-taxes according to the custom of the city. 50 BT 136; ADD 621 = NALK 126; ADD 625 = SAA 6 287; ADD 629 = SAA 6 226. ADD 622 = NALK 105; ADD 623 = NALK 128. 5 2 The question why more straw than grain tax could be asked. A possible explanation, perhaps too far-fetched, is that grain grown under rainfed conditions has shorter stalks and thus produces less straw (Wirth 1962:29: 'Meistens haben allerdings die Regenfelder kiirzere Halmen, also weniger Stroh.'). In combination with the large demand for straw (tiles and fodder) this could have caused the higher percentage of the straw tax. More important, it seems, is the conclusion that twice as many exemptions as obligations occur in both the pledge and the 'lease' group. These taxes (Postgate 1974:174-199) are more than just straw and grain taxes. ADD 370 (SAA 6 191) is a sale of a GILSAR,a garden, which is said to be exempt of 'grain taxes' and of the city ilku. It was possible to pay these agricultural taxes in silver. The amount of 4 minas of silver in Dalley and Postgate 1984: no. 70 is in no proportion to the plot of land involved. The grain tax in silver, in this instance, had to be paid to the rab ekalli, the 'palace manager.' Other officials that are attested as collectors of straw and grain taxes are the governor, or his replacement (ND 2648 = NL 74; CTN 2 196), and a guard (qurbute) in ABL 1012 = SAA 5 82. Not all land was subject to agricultural taxes. In TIM 11 1: 37 a piece of land that is exempt from straw and grain taxes changes hands (A.SAza-ku-ti la s'ib9i la nuhibe). The seller is a weaver, and the buyer a high military official. The price is one and a half (Carchemish) mina of silver for 20 emaru, threshing floor and house included (= 4,5 shekel per emclru, which is not exceptionally high). This relatively low price could be taken as an indication that the straw and grain taxes were not considered insurmountable. Also the unique bow field (A.SAGIS.BAN)of an Itu'ean (SAA 5 16) is free of straw- and grain taxes. In several Neo-Assyrian grants exemptions are found as well (instead of Ebiu, tibnu 'straw' is used). Not exempt, on the other hand, are the ma'uttu fields (CAD rnu'untu; from mdnu 'to provide [with food, etc.].' See the discussion by Radner, this volume, par. VI). Three debt notes from Fort Shalmaneser (Dalley and Postgate 1984: no. 14 [9 homers of barley, grain tax from the ma'uttu land of the king]; no. 15 [5 bales of straw, straw tax of the ma'uttu land]; no. 16 [15 bales of straw, straw tax of the ma'uttu land of the king]) deal with straw and grain taxes 'of the ma'uttu field (of the king).' In all three texts the granary officer has the administrative control. 53 Owen 1969: passim.
256
R. M. JAS
LAND TENURE IN NORTHERN MESOPOTAMIA
The Middle-Assyrian sources show a similar picture. No lease documents have survived, only loans with real estate as a pledge. These texts have been dealt with thoroughly by C. Sap0retti.5~The description of the real estate can only be described as vague, and it is very hard to arrive at conclusions. What is clear, however, is that this genre of transaction again appears to be the first phase of permanent expropriation of the pledged land by the creditor. What is striking in the Middle-Assyrian group is the short term of most of the loans, and the small surface of the pledged fields To conclude with a question: Do we, with the possible exception of the two Kalhu texts and TIM 11 4, indeed have no 'real' leases from Northern Mesopotamia? It seems unlikely the benefits of pledging plots of land by a debtor outweighed the risks involved in such transactions. On the other hand, the advantages for the creditor are obvious: he assures himself of a steady income of agricultural products, and at best ends up with owning the land as well as enjoying the usufruct of it. Can we also conclude from this 'reversed' situation that the demand for both labor and choice plots of land in stable zones, or plots with possibilities of additional irrigation was much greater than the supply?55 What we see happening here very much resembles the situation of Roman peasants as described by P. Garnsey: "It is hard for subsistence farmers to survive and to remain independent. Their margins are narrow. There is an easy slide into debt; small proprietors are readily converted into tenants of rich money-lending landowners."56 Warriner's observations on the situation of Syrian peasants in this century paint in the same direction: "Thus the custom of share-cropping reflects the instability of graingrowing in arid climates with uncertain rainfall and great variations in yields. It is peasant indebtedness which supports the large landowners. Two bad harvests in succession will mean that the peasants become so much indebted that they will actually reduce the area sown for lack of seed, and will therefore be obliged to sell to a money-lender or city merchant."57 To illustrate what could be considered a bad harvest, and what the effects of one or consecutive crop failures could be, the figures given by W.-D. Hiitteroth and D. Tully are helpful comparative material.58 Hutteroth59 has given figures for yields (Reinernte) in Turkey, dating from 1928-1961 (before the enormous increase in the use of artificial fertilizers). He distinguishes the following seven categories of harvests: 1) catastrophic crop failure 2) crop failure 3) bad harvest 4) moderate harvest 54 55
less than 400 kgha 400-500 kgha 55 1-650 kgha 65 1-750 kgha
2 years out of 34 3 years " 4years " 6years "
Saporetti 1978-1979, and Saporetti 1981. Compare, on the situation in modern Iraq, the following remark by Davies 1957:128 '... the farmer has to be certain of his minimum precipitation before establishing a farm on the Assyrian plains.' 56 Garnsey 1998:96. 57 Warriner 1948:85. 58 But see the reservations of van Driel, this volume, par. 6. 59 Hiitteroth 1982:126.
5) average harvest 6) good harvest 7) very good harvest
75 1-850 kgha 85 1-950 kgha 95 1- 1050 kgha
8 years 7 years 4 years
" " "
For nutricial needs, we can quote D. Tully's estimate of cereal consumption in the rainfed areas of the Near East at approximately 200kg/person/year, providing over half of the calories and protein consumed (the other main source of protein being milk products) .60 The figures Hutteroth gives for loss of cattle in dry years are staggering: 82 percent of the large cattle and 96 percent of the small cattle in Keskin did not survive the successive dry years 1873-1874. These high percentages were caused by a combination of two factors: the meadows withered much earlier than normal, meaning less grazing possibilities. Secondly, because of the crop failure, there was no straw to feed the animals throughout the winter. That this indeed was a disastrous combination is proven by the fact that in Turkey in 1873-1874 250.000 people starved to death. 4.
RISK EVADING BEHAVIOR
It is well known that a high level of uncertainty is the cause of risk evading behavior of peasants all over the world.61Apart from the practice of fallowing, we have one other indication of behavior of this kind in Northern Mesopotamia: the puru system. The information the Middle Assyrian Laws contain about land tenure is so important because few other sources inform us about the ways in which the agricultural hinterland of the city of ASSur was exploited in Middle-Assyrian times.62 We have seen that pledging of land against loans occurred frequently in this period, which can be taken, with J. Renger, as proof that private ownership of land did exist on a certain scale in Middle-Assyrian times.63 Additional proof for this is the group of land sales, all coming from archive M 9,64 dealing with land on 'the far bank' of the SiSSar.'65 J. N. Postgate66 has pointed out this specific background of these sales and also has made clear that these texts are not tuppu dannutu's (= definitive sales contracts that could only be drawn up after several formalities were complied with67), but provisional texts. Postgate suspects --
6O Tully 1990:4. We are well informed about food rations in Nuzi. See Zaccagnini 1984:251 with reference to the personnel of Silwa-teSSup: 'The food allowances allotted to these people are differentiated according to sex and age: they range between a maximum of 3 sati (ca 20 liters) and a minimum of 4 qa (ca 3.5 liters) of barley per month.' 6 l See Ellis 1993, 82-104, Chapter 5: The risk-averse peasant. For ancient Greek agriculture, see Garnsey 1988, and Gallant 1991. 6 2 Postgate 1989:143 stresses the fact that the Middle-Assyrian laws 'apply particularly to the city of Assur and its environs.' 63 Renger 1995:297; for Middle-Assyrian attestations see CAD Sapartu. 64 PedersCn 1985:89f. 65 Whether this is Wadi Tharthar or a canal in the neighborhood of ASSur is not really important: imgation is possible, which makes the location desirable. 66 Postgate 1982:308. 67 It is interesting to compare these with the restrictions of first millennium land sales from ASSur. See Klengel-Brandt and Radner 1997.
258
259
R. M. JAS
LAND TENURE IN NORTHERN MESOPOTAMIA
the reason for this was to enable the sellers to profit from the high prices of land, although the opposite, to let the buyers profit from an exceptionally low price, would be more in line with what we have seen thus far. J. Renger has suggested that these 'sales' only transfer the usufruct of the field in question and not the ownership.@ He substantiates this theory with evidence from the Middle-Assyrian laws: the existence of the so-called p i r u (= lot) system.69 This system, in my opinion, is characteristic of rainfed agriculture, and survives into the Neo-Assyrian period.70 In Professor Renger's interpretation, the arable land surrounding the village was under collective control of the community, and was divided into two parts. Half of the land was under the plow; the other half remained fallow. The cropped land was divided yearly among the villagers by casti-This is a form, not so much of communal ownership but of communal m a n ~ g e r n e n t .Strict ~ ~ rules regarding supplementary irrigation, crop rotation, and the timing of agricultural activities during the day lie at he basis of this system, which can best be described as some sort of Flurzwang.
due to its poor state of preservation the importance of HSS 15 60 (SMN 3391, from D 6) has been underestimated. According to Kendall (1975) 93 this text mentions only four aziiibu's, but the structure of the text rather suggests it is a list of men with aiiiibu's in their charge. Even though it is much broken, the text still is the most important a&bu document we have. A detailed study of it will appear in my forthcoming The Administration-of Food and Cattle in Nuzi. While it is a clear ai3ibu's do sometimes perform agricultural work, and can be claimed for service by the state, the evidence is not specific enough to allow a translation 'tenant farmer.'
EXCURSUS
aEabu 'tenant farmer'?
In Nuzi studies one comes across the translation 'Pachter' for aiiabu, which we will have to discuss here briefly. The latest treatment of the word is Dosch 199395-87 (with previous literature), to which we should add Mayer 1976 who has suggested that the foreign word JeJtuhlu is the equivalent of Akkadian a3abu. s'es'tublu is a hapax. In HSS 13 208:15 (Mayer Cat 18) mAkkul-enni ie-es'-tu-ub-lu is listed among male and female palace employees receiving textiles. The other attestations quoted by the CAD S.V.have nothing to with this word. There is no evidence that SES.MES 'brothers; colleagues' (for which see CAD ahu A mng 2b where HSS 15 27 already was quoted) should be interpreted as an abbreviation of s'es'tuklu. The Hittite origin of the word suggested by Mayer is unlikely (personal communication R. H. Beal), so we will not follow Mayer's suggestion, quoted by the CAD, that s'es'tuhlu is perhaps to be regarded as a foreign word for as's's'abu. CAD aJ's'iibu has the Nuzi references under usage c) 'alien(?) resident (of low status).' Fincke 1993:424 translates Pachter, as does G . Wilhelm in his translations of A ~ 154 S and A ~ 260 S (without commentary). M. P. Maidman 1995:942 translates 'dweller' and in his discussion of the various social classes in Arrapba suggests they were the inferior, dependent class living on other people's land. Maidman suspects their main tasks were agricultural. Most of the references are quoted by the CAD, to which a few are added by Dosch. The plural ai-ia-be-e-na (Conteneau 1931:38, no. 6:1, see Fadhil 1983:33) can also be read in CT 51 10:8 New information is provided by Muller 1998 no. 20 (BM 26225:915) in which two men receive emmer from Zike, and promise to do his laundry. Neither the herald nor the mayor can claim them because they are now Zike's a%ibuys. Probably 68 69 70
Renger 1995306. Described by Renger 1995:306,without mentioning the Akkadian word. Postgate 1989:144; Radner 1997:257. Postgate 1989:145.
260
LAND TENURE IN NORTHERN MESOPOTAMIA
R. M. JAS
Fincke, J. 1993
REFERENCES
Adams, M. E. and J. M. Holt The Use of Land and Water in Modem Agriculture: An Assessment, 63-83 in 1985 Beaumont and McLachlan 1985. Alex, M. Klimadaten ausgewahlter Stationen des Vorderen Orients (Beihefte zum 1985 Tubinger Atlas des Vorderen Orients, A 14), Wiesbaden. Al-Barazi, N. K. The Geography of Agriculture in Irrigated Areas of the Middle Euphrates 1963 Valley, Vol. 2, Baghdad. Beaumont, P. et al. The Middle East: A Geographical Study, 2nd ed., London. 1988 Boesch, H. El-'Iraq, Economic Geography 15, 325-61. 1939 Buringh, H. Soils and Soil Conditions in Iraq, Baghdad. 1960 Conteneau, G. Textes et Monuments , I: Tablettes de Kerkouk du MusCe du Louvre, RA 28, 193 1 27-39. Dalley S. and J. N. Postgate The Tablets from Fort Shalmaneser, London. 1984 Davies, D. H. Observations on Land Use in Iraq, Economic Geography 33, 122-134. 1957 Dosch, G. Zur Struktur der Gesellschaft des Konigreichs Arrapbe (HSAO 5), 1993 Heidelberg. Ellis, F. Peasant economics, 2nd ed., Cambridge. 1993 Fakhry, A. K. El and A. M. Sultan Studies on Crop Rotations and Tillage Practices on Wheat Production under Rainfed Conditions in Northern Iraq, 59-65 in F A 0 1980. Rainfed Agriculture in the Near East and North Africa. Proceedings of the FA0 seminar on rainfed agriculture in the Near East and North Africa held in Amman, Jordan, 5-10 May 1979. FA0 Near East Regional Office, and Soil Resources, Management and Conservation Service, Land and Water Development Division. Food and Agriculture Organization of the United Nations, Rome. World Soil Resources Report: Regional Study on Rainfed Agriculture and Agro-climatic Inventory of Eleven Countries in the Near East Region (Algeria, Cyprus, Iraq, Jordan, Lebanon, Libyan Arab Jamahiriya, the Sudan, Syrian Arab Republic, Tunisia, Turkey and Yemen). Near East Regional Officenand and Water Development Division, Food and Agriculture Organization of the United Nations, Rome. Soils Bulletin 52. Guidelines: land evaluation for rainfed agriculture. Soil Resources Management and Conservation Service, Land and Water Development Division, Food and Agriculture Organization of the United Nations, Rome.
Re'pertoire Ge'ographique des Textes Cune'iformes Band 10: Die Orts- und Gewassernamen der Nuzi-Texte, Wiesbaden. Einige Joins von Nuzi-Texten des British Museum, SCCNH 7, 23-36.
26 1
1995 Gallant, T. W. Risk and Survival in Ancient Greece. Reconstructing the Rural Domestic 199 1 Economy, Cambridge. Garnsey, P. Famine and Food Supply in the Graeco-Roman World: Responses to Risk 1988 and Crisis, Cambridge. Cities, Peasants and Food in Classical Antiquity, ed. by Walter Scheidel, 1998 Cambridge. Gelb, I. J. et al. Nuzi Personal Names (OIP 57), Chicago. 1943 Geographical Handbook Iraq and the Persian GulJ; Geographical Handbook Series B.R. 524, Naval 1944 Intelligence Division. Grosz, K. Archive of the Wullu Family, Copenhagen. 1988 Hall, A. E. et al. (eds.) Agriculture in Semi-Arid Environments (Ecological Studies 34), Berlin. 1979 Hkle, J. TAVO A X 4 Mesopotamien. Landnutzung (Mesopotamia. Land Utilization), 1985 Wiesbaden. Hutteroth, W.-D. Turkei (Wissenschaftliche Liinderkunden Band 21), Darmstadt. 1982 Isom, W. H. and G. F. Worker Crop Management in Semi-Arid Environments, 200-23 in A. E. Hall, G. H. 1979 Cannell and H. W. Lawton, eds., Agriculture in Semi-Arid Environments (Ecological Studies 34), Berlin. Jones, M. et al. (eds.) The Agrometeorology of Rainfed Barley-based Farming Systems. 1993 Proceedings of an International Symposium Tunis, 6-10 March,1989, ICARDA Aleppo. Jordan, G. D. Usury, Slavery, and Land tenure: The Nuzi tidenniitu Transaction, ZA 80, 1990 76-92. Kendall, T. Wagare and Military Matters in the Nuzi Tablets, Ph.D dissertation, Brandeis 1975 University. Klengel-Brandt, E. and K. Radner 1997 Die Stadtbeambten von Assur und ihre Siegel, 137-143 in S. Parpola and R. M. Whiting eds., Assyria 1995, Helsinki. Lebon, J. H. G. 1953 Population Distribution and the Agricultural Regions of 'Iraq, Geographical Review 43, 223-228. Mahdi, K. A. 1993 Farming Systems, Market Conditions and the Production of Barley in Northern Iraq, 216-29 in Jones, Mathys, Rijks 1993.
262
R. M. JAS
LAND TENURE IN NORTHERN MESOPOTAMIA
Maidman, M. P. Nuzi: Portrait of an Ancient Mesopotamian Provincial Town, 931-947 in: J. 1995 M. Sasson ed., Civilizations of the Ancient Near East, Volume 2, New York. Mayer, M. Beitrage zum ljurro-Akkadischen Lexikon I, UF 8, 209-210. 1976 Morrison, J. A. Alishar: A Unit of Land Occupance in the Kanak Su Basin of Central 1939 Anatolia, Chicago. Muller, G. G. W. Londoner Nuzi Texten, SANTAG 4, Wiesbaden. 1998 Ockerman, H. W. and S. G. Samano The Agricultural Development of Iraq, 189-207 in Beaumont and 1985 McLachlan 1985. Oram, P. A. and C. de Haan Technologies for Rainfed Agriculture in Mediterranean Climates. A Review of 1995 World Bank Experiences (World Bank Technical Paper 300), Washington, D.C. Owen, D. The Loan Documents from Nuzu, Ph.D. dissertation, Brandeis University. 1969 Parpola, S. review of Postgate 1974, ZA 65, 293-296. 1975 PedersCn, 0 . Archives and Libraries in the City of Assur, Part I, Uppsala. 1985 Postgate, - J. N. The Governor's Palace Archive (CTN 11), London. 1973 Taxation and Conscription in the Assyrian Empire, Rome. 1974 Fifty Neo-Assyrian Legal Documents, Warminster. 1976 Ilku and Land Tenure in the Middle Assyrian Kingdom - A Second Attempt, 1982 304-13 in Societies and Languages of the Ancient Near East (Fs. Diakonoff), Warminster. The Ownership and Exploitation of Land in Assyria in the 1st Millenium 1989 B.C., 141-152 in Reflets des deux fleuves (Fs. Finet), Louvain. Radner, K. Die neuassyrischen Privatrechtsurkunden als Quelle fur Mensch und Umwelt, 1997 Helsinki. Renger, J. Institutional, Communal, and Individual Ownership or Possesion of Arable 1995 Land in Ancient Mesopotamia from the end of the fourth to the end of the first millenium, Chicago Kent Law Review 71, 269-319. Richter, T. Die Tenne in den Texten aus Nuzi, SCCNH 7, 69-79. 1995 Saporetti, C. 1978-79 I1 prestito nei documenti privati dell'Assiria del XrV e XI11 Secolo, Parte I: L'analisi dei testi, Mesopotamia 13-14, 5-90. I1 prestito nei documenti privati dell'Assiria del XIV e XI11 Secolo, P a t e 11: I 1981 testi in trascrizione, Mesopotamia 16, 5-41. Tamimi, S. A. and J. R. W. Young Preliminary Soil Moisture Studies at Hammam al-Alil, 239-47 in F A 0 1980. 1980
Thalen, D. C. P. Ecology and Utilization of Desert Shrub Rangelands in Iraq, Ph.D. 1979 dissertation Rijksuniversiteit Groningen. Tully, D. (ed.) Labor and Rainfed Agriculture in West Asia and North Africa, Dordrecht. 1990 Tutwiler, R. N. The Great Chickpea Challenge: Introducing Winter Sowing in the 1995 Mediterranean Region (ICARDA Social Science Papers 4), Aleppo. Warriner, D. Land and Poverty in the Middle East, London. 1948 Wilkinson, T. J. Environmental Fluctuations, Agricultural Production and Collapse: A View 1997 from Bronze Age Upper Mesopotamia, 67-106 in H. Nuzhet Dalfes et al. eds., Third Millennium BC Climate Change and Old World Collapse, Berlin. Wirth, E. Das lhdliche Haus im Irak, Verh. 30. Dt. Geographentag Hamburg 1955, 1957 416-22. Agrargeograpie des Irak (Hamburger Geographische Studien 13) Hamburg. 1962 Syrien. Eine geographische Landeskunde (Wissenschaftliche Lhderkunden 197 1 Bd. 415) Darmstadt. Zaccagnini, C. The Price of the Fields at Nuzi, JESHO 22, 1-31. 1979 review of G. Wilhelm, Das Archiv des Silwa-teg5up. Heft 2, OLZ 79, 249-53. 1984 Zuidervliet, A. C. Irak (Koninklijk Instituut voor de Tropen, Landendocumentatie 1981 no. 2) 198 1 Zutphen.
PIHANS LXXXVIII, 2000
The Mesopotamian North: Land Use, An Attempt* G. van Driel (Leiden)
1.
PRELIMINARY REMARKS
a. Introduction The question can only be tackled with some success if it is realised that the answer depends on the evidence available and that that available evidence is extremely limited as to time and place. Only if these limitations are accepted some attempt at a generalisation can be made. Northern Mesopotamian society has never been the butt of generalisations of the type more common for the South, particularly not for the Third Millennium. The reason is that the evidence from the South suggests utter domination by the institutions, especially the temples, where everybody is supposed to find an embedded existence. That there is room for a certain amount of skepticism has been argued earlier, especially half a year ago at the second MOS symposium. In the Mesopotamian North the presence of a pastoralist population is well-documented, if not in general, at least for particular periods. Already before the Mari period this tribal population had penetrated the South and contributed to the downfall of the Ur 111monarchy. In the longer run the shift of the balance of power would result in the disappearance of some of the best documented Southern divine households. The penetration of the South was not a sudden event, it had, to all likelihood, an earlier history and it would be repeated later in history, by the Arameans, with different, but equally far-reachng results. The North plays its role in this. In the North we do not have evidence for strong institutions other than those based on the personalities of outstanding monarchs. Temples are not documented in the way they are in the South and there is nothing to suggest that Northern divine households had a strong grip on the economy. The role of the palaces, on the other hand, cannot be underestimated,l though many had a very limited sphere of influence. Though the nature of the institutions differs, the structure through which they organise (their) land use is comparable. Though in studying land use archaeology plays the dominant role, the impact of the remnants of ancient bureaucracy should not be underrated. They underline the importance of organization in producing the surplus on which the palatial system is built. When constructing a model of the ancient Mesopotamian economy this phenomenon *Not all the implications of the contributions to this collection of papers have been digested completely. Thanks to the editor the final version of B. Lafont's article was available just in time to remove at least some of the most notable overlaps. I could also read the final versions of the contributions by F. A. M. Wiggermann and T. J. Wilkinson. Not in agriculture, not in animal husbandry, not in crafts production, but especially not in the "circulation of goods."
266
267
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
must be incorporated. The surplus produced is not (only) a matter of umpteen small farmers each producing more than they need, either through taxation, through the workings of a system of land holding which turns them into tenants or through their indebtedness, or even through the requirements of "the market," but it is the result of institutional organization. "Collapse" is also a result of eliminating or reducing the importance of this system of production, it is not only a consequence of, for instance, a worsening climate. Of great importance is the fact that the geographical circumstances and the ethnic situation differ in North and South. Neither can be dissociated from the two principal ways of life which both are intimately associated with the natural circumstances. The way in which land is exploited is the central question to be addressed. In the Ancient Mesopotamian societies of the past we are dealing with, only access to the products of agrarian ways of life could guarantee a stable society, even if this stability finds its most natural expression in an equilibrium of what we would regard as opposites, settled and non-settled. In reality all transitions are fluid, and they can be so on account of the geographical circumstances and the not very intensive use of the land.
If we want to understand what happens in the North, we must reconcile the natural circumstances, the archaeological map, with its implications for land use, and the ethnic situation with the historical facts as we know them now. That knowledge is both limited and seductive. Can we attribute a decisive role in the later Third Millennium to the short episode of the Akkad kings? Are the twelve years of tribal Mari really representative? Are the Assyrian kings colonisers or depleters of the countryside? Three related questions of prime importance need to be addressed. The first is that of dimorphism, the degree to which the various human groups from sedentaries to nonsedentaries share in the exploitation of the land.3 The second is the degree of institutionalisation, the nature of the grip of the centre on its "hinterland." The third is the structure of arable agriculture and, provided it is visible, of animal husbandry, in short the circumstances under which production takes place. That (written) information on animal husbandry is nearly always disappointing does not diminish its importance. For the question of the effectiveness of the production there is not much evidence, but what is available causes considerable problems.
b. The nature of the evidence In the area immediately South and West of the main chain of the Turkish and Iranian mountains the predominant dry farming is supplemented by irrigation wherever the occasion arises. Moving towards the South the inhabitants increasingly have to cope with drier circumstances, which, however, provide a natural niche for animal husbandry as practised by both the settled and the non-settled population. Essential is that, as stressed, transitions are fluid. The key to understanding Northern developments, at least from the Uruk period onwards is understanding how the centre, or the centres, dominated agrarian production. This implies the possibility that there are no centres in a given situation. In the agrarian world of Ancient Mesopotamia political dominance is impossible without control of agrarian production. Agrarian production has two faces, the one arable, the other based on animal husbandry. Who strives after hegemony cannot fail to dominate both, at least to a degree. The system of farming practised leaves much room to the non-sedentary pastoralist, both as to actual space and as to time. He penetrates deeply into settled land, aided by the fact that in a given year the arable farmer uses only half his land, and that in a limited part of the year only, in which there are alternatives for the herdsman elsewhere. This is a world where it is difficult to separate sedentaries, half-sedentaries, half-nomads and nomads effectively.2 Wirth's scheme is informative on the where, the when and the how intensive, but it lacks one specific type of information, because this factor has lost its importance in the recent past. This is ethnicity, the importance of which is emphasised twice, in a different manner, in the period which concerns us: Mari and the Amorites, Assyria and the Arameans. That, at any rate in the Second, and to all probability in the later Third Millennium, if not earlier, the Hurrians are of equal importance, we may surmise, but are unable to fathom. In the area of study there are at least three centres of ethnic gravity, all three of which develop recognisably in the course of time. E. Wirth, Syrien, Darmstadt 1971, 257, fig. 32. Though lacking the aspect of ethnicity, which is, in this context, of recently greatly reduced importance, this figure is basic for the understanding of the Northern situation.
All evidence has archaeological roots. Survey and excavation contribute, or should contribute, in a complementary manner to our understanding. Northern archaeological survey is fragmentary, hampered not only by the present national borders, but also by the lack of that unity of vision -even if it has grown in the course of the work- which characterises the monumental torso of the work of Adams and Nissen in the Alluvium. In the North the limited excavation which has taken place has made it only too clear that though many of the major settlements look very similar, that implies no uniformity. These settlements, marked as a kind of irregular spiders on Van Liere's map and inserted into a typology of his making,4 all have their own individual development and history.5 That was perhaps to be expected from the outset in view of the spacing, and from the little that is historically known. Tell Halaf and Fekheriya are, for example, very close to one another. A regional binary development is something that occurs more often, Mozan and Amouda(/Shermola)6 could be mentioned and it is also historically attested: Andarig and Allabad, Karana and Qaffara,Kurda and Kasapa.7 These towns are close to one another, their importance alternates. Other combinations would seem to be equally possible. Essentially not all can be important at the same moment. The administrative archives partially illustrate why. The commonest type of document in the archaeologists' MB in the North deals with the allocation of bread, beer and the like. The fact that this naptan s'arri, "the king's meal," is consumed in a given place influences its status directly. It induces the presence of a public related to the importance of the s'arru concerned, it creates a "market" for all sorts of goods with all that that implies. The presence of a ruler We are not dealing with the purely non-sedentary steppe areas. AAS 5 (1954-5) after 146 and 133f. The "Kranzhiigel," a creation of Freiherr von Oppenheim, Van Liere's type IIIa, have acquired special notoriety. But what is the role of the other Tell Amouda across the border? Cf. especially F. Joannbs, "L'organization de l'espace en Irak du nord," Cahiers du Centre G. Glotz 3 (1992) 5-9 and J.-M. Durand, LAP0 17,239-240.
268
G. VAN DRIEL
contributes notably to the number of "inhabitants" of a given place. If this ruler is not present much of the built up area of that place has a reduced population, though the palace will continue to function as a local economic, i.e..agrarian, centre. Especially the presence of the ruler decides whether the palace concerned plays a role in the "circulation of goods." There is a skeleton staff, perhaps a demoted queen, but these towns are in an alternating way empty. A binary development is not general, of course. It can mean alternating importance on a long term basis, perhaps like Halaf and Fekheriya, or it can be strictly period specific as in Mozan and Amouda. The individual nature of the development of the various core settlements is to be related to the carrying capacity of their "hinterland." If this hinterland is artificially enlarged, sudden expansion can take place, followed by the inevitable contraction as the supra regional function disappears. The enumeration Kalhu-Khorsabad-Nineveh can serve as a reminder. Nineveh was perhaps a not too big town, with a fairly important sanctuary, Kalbu a small provincial town built against a prehistoric tell and Khorsabad a small village of which we happen to know the name, before temporary expansion based on a royal whim took place. This is not specifically Neo-Assyrian. Sam%- dad's Tell Leilan is probably in the same company, though he uses something older. We have in this context, with little doubt, Tell Taya with its visible growth rings, somewhere before or at the beginning of the Akkad period, and we may perhaps start to compare Middle Uruk Brak. But these are the extreme examples, almost inevitably results of very specific circumstances. More important is the long-term development of regional settlement as a whole and its correlation with land use. Here we sorely miss elementary sets of maps, and the information on the size of settlements, of the type found in Adams' Heartland.8 We must acknowledge, as is the case in the South, that archaeological survey and reasonable levels of textual evidence also in the North do often not deal with the same, promising, area. The Nuzi case is the most striking, especially in view of the existence at least two or three detailed text-based geographical studies.9 Archaeological information on land use in the Arrapba area is minimal and, admittedly, the text-based maps are mainly valid for limited periods only. Even if we do not have an archaeological map which encompasses the whole rainfed and marginal zones of Northern Mesopotamia, we have at least the "Wilkinson-mode1,"lO which combines the observed size of settlements, the available arable, the derived population and the calculated labour requirements in such a manner that theoretical
Perhaps the situation will change for the Syrian part of the area under discussion by the Lyonnet survey, which will probably provide a kind of outline based on the bigger tells and hopefully contribute consistent information on their size, but such a map would require a filling in of details 2 la Tell el Hawa, which would take much time. A. Fadhil, Studien zur Topographie und Prosopographie der Provinzstadte des Konigreichs ~ r i a ~ bBaF e , 6, Mainz 1983; G. G. W. Miiller. Studien zur Siedlungsgeographie und Bevolkerung des mittleren Osttigrisgebietes, HSAO 7, Heidelberg 1994; C. Zaccagnini, The rural landscape of the Land of Arraphe, QGS 1, Rome 1979. T. J. Wilkinson, "The Structure and Dynamics of Dry-Farming States in Upper Mesopotamia," CA 3515 (1994) 483-520, plus R. C. Hunt, "On Dry Fanning in Upper Mesopotamia," CA 3612 (1995) 289290.
THE MESOPOTAMIAN NOH'I'II. I,ANI) 1I\I:
limitations can be discussed.ll The model provides up to a degree an explanation, from an economic ~ o i nof t view, of the extremely fragmented political situation we perceive in the Mari archive, even if what we perceive in Mari is no more than a late stage in a development which had started much earlier. In the rainfed area settlements have an (arable based) economic zone of maximally 10 to 15 km across, and if they need to expand this is only possible through subjection and exploitation of neighbours. Agriculture takes place in a high risk-low yield environment, but with ample land, which must be studied, ideally, per individual settlement system. The problem is defining what a settlement system is. A sequence capital - district centre - rural centre village/farm would be appropriate for Mari itself, but in the smaller kingdoms levels two and three would merge. Though levels three and four will differ little in size, they differ in function, at level three considerable administrative responsibility with the accompanying risk is located. Whether the possession of defences is significant, is dubious. The Mari correspondence indicates the ease with which small settlements are turned into strong points, and a certain degree of fortification is to be expected as soon as there is talk of dimtu's, which are, probably, little more than farms. We are not dealing with a market oriented system, but with one which is intended to produce an agricultural surplus in order to maintain a palatial system. The way in which agriculture is organised and delegated is important. The main tool is the plough team which is managed directly or through temporary transfer to someone with an interest in the system. Small settlements with a managerial role12 can be more important than small towns. A four tier system seems to be acceptable for Third Millennium Ebla,l3 Leilan 11 could represent something similar,14 but what in the Middle to Eastern Babur triangle where kingdoms are thickly sown in the period of the Mari archives? A system can be, geographically spoken, double headed, as we have seen.l5
Crucial is the area which can be harvested by one person. This point requires, perhaps, special attention, both regarding what can be done by one worker and the practical length of the harvesting season. If the first is adjusted upwards and the second downwards the net result may maintain the balance. Under Mesopotamian institutional (irrigation) circumstances a harvester can manage, or is supposed to manage, much more than the 1 hectare in 20 days the model operates with. In Umma there is an occasional norm of half an ikk per day, which means that a hectare takes a little more than 5.5 man days with a copper or bronze sickle. This is of course for land planted in rows with very wide, say 50 to 75 cm, intervals between furrows. But this is not irrigation specific, it is an element of an institutional agriculture which is based on the use of the seeder plough. On the other hand the inhabitants of the villages I stayed in in Syria were extremely keen to finish harvesting within a fortnight, preferably less, ten days. Wiggermann's Sabi Abyad. A. de Maigret, "I1 fattore idrologico di Ebla," OA 20 (1981) 1-36, esp. 28, cf. J. W. Meyer, AoF 23 (1996), esp. p. 145-6. l 4 P. Wattenmaker, "The 1987 Tell Leilan regional survey: preliminary report," 9-18 in N. F. Miller ed., Economy and Settlement, MASCA Papers, Supp. Vol. 7, 1990. l 5 Leilan and Moh. Diyab, where the second is over half the size (50 ha) of the first (90 ha) at a good hours walk, perhaps also belongs here, but the smaller of the two continued after the end of the bigger town.
lag
270
27 1
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
There are moreover ethnic aspects.16 A complication is that a settlement system in Northern Mesopotamia does not exist without the influence exerted by the non-sedentary element. The leadership of this tribal world exerts a constant political and military influence and the tribesmen participate in arable agriculture,just as the settled population needs access to pasture. The practical results of this situation can work in a far from uniform way. The period of Sam~i- dad's kingdom can be characterised as a period of wars along the borders, but these borders encompassed more or less the area from the Euphrates in the West to beyond the Tigris in the East, where the situation was unstable. Internally the situation appears to be well in hand. In that respect the period of Zimri-Lim differs completely: his restores the "old families to their thrones" and the result is utter instability. But perhaps the Mari episode is not a good yardstick to measure the settlement system in the North with: Mari represents the high water of the political and military influence of the primarily non-sedentary tribes. A different insight in the urbanised world of Syria would probably appear if we knew more of the period of the iakkanakku's or their predecessors, or of that of the Ebla documents, where, at least for the present, the non-sedentaries do not play a dominating role. Mari illuminates the way in which an elite with a non-sedentary background reacts to a pre-existing urbanised world. These are renamers not founders.17 Not for nothing does the Old Babylonian period end in a massive decline of urbanisation, both in the North and the South. If this is true the rich Mari documentation illustrates a world in decline. After the decline we are in a changed world. No longer do we find a town based settlement system in much of the area we are dealing with, but one of smaller settlements. Written sources could be interpreted as corroborating this. The social basis on which political structures are based has changed, though the plough team system continues to provide a surplus, but directions have changed.
suggest that the way in which institutional agriculture was organised in Mari will not have differed greatly from that more to the North. Mari makes a major contribution in that respect. The archives of the Assyrian monarchy are in general very incomplete and have come to us in a haphazard manner, but they contain some of the most interesting and detailed information concerning Northern villages we posses.21 The importance of the fragments of the Assyrian Harran-survey can hardly be overrated, and the Assyrian sale contracts are often usefully detailed. The Assyrian royal inscriptions pose a special challenge: they convey a message, which means that they are far from truthful. Though they are stereotype, there remains an underlying message with a bearing on the way the Ancient Mesopotamian exploited this area. This is the world of the "big towns with the small towns of their surroundings" and the open space to be assumed between such clusters.
There is not much published written evidence on land use in the North in the Third Millennium, the somewhat later written material from Mari colours all. The Ebla archive deals with agricultural land only in a very marginal way and indirect information on a (possibly) non-sedentary population is restricted in meaning.18 We will be presented with the results of a special study of agriculture in Mari.19 Mari agriculture is, however, the agriculture of the river valleys of the Euphrates and the Habur, and not that of the true rain-fed zone. Therefore we will have to go to Nuzi, and may perhaps hope that in the end Leilan will provide information. The most important aspect of the Mari-archives is that they inform us on the non-sedentary world of the years of Zimri-Lim with an otherwise unprecedented intensity.20 But important is also that there is just sufficient evidence to In J. R. Kupper, ARMT 28 79:35-6 Zimri-Lim is quoted as having said: "(The town of) Zalluhan is not of the Idamara~ians,it is Simalite," that is : "of my own (tribal) group." Ethnic terminology with regional implications is also used near Jebel Sinjar: N u a a and Jamutbal. Jahdun-Lim is a possible exception. l 8 A. Archi, AAS 40 (1990) 50-55; L. Milano, "Ebla: gestion des terres et gestion des ressources alimentaires," Amurru 1 (1996) 135-171. On the MAR.TU in Ebla see A. Archi, OrNS 54 (1985) 7-13. Cf. the contribution of B. Lafont. 20 Documents from about the reign of the Ur 111 rulers, but also from that of their immediate successors, that is so to say roughly pre-Hammurabi Old Babylonian, are encompassed by the misleading term "Sakkanakku-texts."This type of material is known from Mari, cf. H. Limet ARMT 19,
2.
METROLOGICAL ASPECTS
a. Basics The main difficulty which hampers all attempts at correlating the contents of the texts, archaeological observations and the modern "facts of life," is the near impossibility of combining the Northern systems of surface and solid volume metrology with our systems. Northern metrology is badly known, in fact is based on what can be classified as "reasonable suppositions." Understanding surface and solid volume metrology, in practise the relation between land and seeding rate, is vital for our understanding of agricultural practises and for the comparison of models like those proposed by T.J. Wilkinson and R. C. Hunt with textual evidence. Down to the Old Babylonian period the ikti was the main surface measure for arable in the North (GasurINuzi and Mari).22We, light-heartedly, assume that this was the same ikii as the one used in the South, and thus acquire with little trouble a means for comparing Northern practises with those of Southern irrigation agriculture. We use there the bureaucratic norms from the best known period, Ur III: 18 ikti constitute 1 bur(3), which in Southern thinking has a kind of standard relation to plough teams employed, seeding rate, and it would seem, yield expected. Let us say 6.5 bur per team, a royal kor, 300 sila(3) of seed per bur, and a (bureaucratic) yield of 30 kor per bur, with all sorts of practical variations, especially as to the last amount mentioned. M.A. Powell23 has explained the workings of the metrological system, but for practical application choices have to be made concerning the specific value of the cubit, which determines the size of the ikii, and also for the sila, the basic solid volume measure. Powell proposes (p. 462-3) 50 cm for the cubit and one litre for the sila. An ikti is thus J.-M. Durand MAR1 4 (1985) 147-72, and Terqa, cf. 0. Rouault, 105-109 in 6. Tunca and D. Deheselle eds., Tablettes et images auxpays de Sumer et d'Akkad (Fs. H . Limet) Likge 1996. 21 Cf. F. M. Fales, "I1 villaggio assiro Bit Abu-Ila'a," Dialoghi di Archeologia NS 3 (1981) 66-84, Republished as: F. M. Fales, "The Assyrian Village of Bit Abu-ila'a," 169-200 in C. Zaccagnini ed., Production and Consumption in the Ancient Near East, Budapest 1989. 22 Note that the essential piferti kariiJe documents of the Middle Assyrian period still use the ikC and not the emaru, in a period later than the Nuzi documents, cf. the contribution of F. A. M. Wiggermann. 23 "MaJe und Gewichte," RLA 7 (1987-90) 480.
272
G. VAN DRIEL
(almost) 3600 m2 (and a bur 64800 m2, 1/50 ha less than the 6.5 ha. I propose to use, which makes a difference of 2 ha per square kilometre). A cubit of about 50 cm seems acceptable, but the equation one sila is one litre causes more difficulties (Powell p. 5034). F. Thureau Dangin24 proposed 0.84 litre, which is accepted by many Assyriologists. A kor of 300 sila is under these circumstances 252 litres and not 300, a fact which really influences outcomes.25 I prefer to accept this more "pessimistic" amount, but the choice is admittedly arbitrary. A further source of difficulties is that the texts measure cereals as solid volumes and that modern evidence comes to us in kgs. Conversion causes problems, not just because humidity influences weight or different types of cereals have different volume/weight ratios. All cereals are "barley" (and "barley" is "barley"). Among those confronted by the problem P. Halstead works with 1 litre is 0.74 kg, 1 kg is 1.4 litres, J. Renger prefers 1 kg equals 1.66 litres. Let us cut our losses, be arbitrary, and equate 1 kg with 1.5 litres.26 One kor of 300 sila equals 252 litres, identical with 168 kgs. Those who prefer other equations can easily recalculate the outcomes, but we must accept that there is much arbitrariness in this all.27 b. Seeding rates
The best calculable seed ratio's in Northern texts using the ikfi are those collected by C. Zaccagnini, UF 11, 1979, 854f, where he arrives for Akkad period Gasurmuzi, with one exception, at 60 sila per ikfi, or "1080 sila per bur", just over 166 sila per ha (93 kgha). This is considerable compared with Southern standards (300 sila per bur, 252 litred168 kgs. per 6.5 ha, 25.85 kgha) and therefore leads Zaccagnini to the suggestion (p. 856) that the Northern ikfi is not based on the 50 cm-cubit but on the "big cubit" of 75 cm and therefore measures almost 8000 m2 instead of (nearly) 3600 m2.28 60 sila per ikfi of this size means 75 sila per ha, 63 litres or 42 kgha, which, though more than in the South, is not only much less, but actually means very light sowing compared to modern practises. Even if we equate sila and litre the sowing remains with 50 kg per ha close to the lower margin of what seems to be feasible n0wadays.~9Does this really suggest light sowing or are we dealing, in the first case, that of the ordinary cubit, with the well known Southern "seed plus fodder" assignment? Such assignments do occur, cf. what will be said about Rimah, but 100 kgha is still possible today. The argument for this early "big24
RA 18 (1921) 134, but note that the assumption not only concerns the size of the sila, but also the fact that it represents a constant factor as to time and place. 25 I understand Powell's insistence on the consistent use of cubits of 50 cm and sila's of a litre as a purely practical device, not as something based on strict metrological theory. Important is that by accepting fixed dimensions for cubit and sila (and stepping over a number of other practical difficulties) we assume that we have a means for directly comparing North and South. Those who prefer the kor of 300 litres make the same assumption and must add 48 litres for every (Southern, royal) kor involved in a calculation here. 26 That this must be "properly" winnowed stands to reason, otherwise no comparison is possible. 27 It should be emphasised that whenever there is talk of the "bur" in a Northern context, that does not mean that this surface measure is really used, its occurrence here always indicates a recalculation intended to make possible comparison with a (theoretical) Ur I11 "Southern standard." 28 This is the surface measure in the South in the Kassite period. 29 Between 120 and 50 kglha.
THE MESOPOTAMIAN NORTH: LAND USE
273
cubit ikfi" is perhaps not so strong as it seems to be and fairly dense sowing could be a fact. Mari poses a special problem as to the Northern iki2, but the obligation implied in not equating the Northern and the Southern ikfi is that we would have to fiddle either the number of sar's per ikfi or their size, neither of which is attractive.30 In Nuzi and later Assyria the imirulemiiru is used for surface and solid volume measurements, indicating the direct relation between land measured and seed used. Powell (p. 485) has discussed the theoretical basis. F. A. M. Wiggermann in his contribution points out that the standard Middle Assyrian seeding rate must be put at 0.0.3, or one ~imdu,per ikfi.3l Or, recalculated, 1.4.0 kor per bur, considerably more than the "Southern standard." But it also complicates matters. A jimdu of seed sounds suspiciously similar to what is the basis for the standard surface measure in Kassite Babylonia, but there the ikfi has become bigger.32 So either this could caution us in assuming that the situation in the North was perceived as less favourable, therefore entailed the use of more seed: North and South use the same standard amount, -or could we add: "in irrigation circumstances"? But that is only the case if the ikfi's are equal. The opposite could be the fact. Zaccagnini, UF 11, p. 85 1, has gathered the evidence for a Nuzi surface imeru of , one exception giving 120x60. The G ~ Rcannot be shorter than 100x80 "feet", G ~ R with three cubits. If we remain faithful to the 50 cm cubit, we arrive at the 18000 m2 which , ~ which ~ is what Powell expects for theoretical had been calculated by H. L e ~ y and reasons: it equals 5 ikfi in the earlier system (p. 487). This 1.8 ha takes one imeru of seed, which, for Zaccagnini, for sound Nuzi reasons contains 80 sila: 44.4 sila per ha, 37.3 litres, 24.8 kgha. This is a very low outcome, though close to what is (our) "standard" in the South (25.85 kglha). Zaccagnini also considers a 40 cm cubit, with a surface imeru of 11520 m2, which results in about 41 kgha, which is also to be considered a low seeding rate. A 1.1 ha surface is Zaccagnini's preferred equation for the surface imeru. This depends on acceptance of a 40 cm cubit. Whatever we do, seeding rates in Nuzi seem to be low in the imeru system, sometimes even lighter than in the South. We could increase the amount of seed by working with a volume imeru of 10 situ of 10 sila each. This situ is used by Zaccagnini in the table on p. 184 of Oriens Antiquus 14 30
Even if we could, perhaps, brush aside ARM 2 61 as the complaints of a (high-ranking) tenant who overacts his protestations, the same high rent of 20 Mari kor per ikli occurs elsewhere, notably in ARMT 24 4(+)5, rev. 6' and 9', a most interesting text, which is difficult to understand in its present state. Recalculating, this would mean 8 Southern kor per ikli, 144/bur, which is incomprehensible. This remains a real problem. Middle Assyrian texts retain the ikli, though earlier Nuzi already has the irneru. It should be noted that from the texts quoted in Wiggermann's fig. 8 in this collection, the only published evidence on seeding rates is from Tarbagbe where the yields are on the one hand dangerously low and on the other twice obtained by sowing 35 sila and not 30 as on the other two occasions. There is therefore some reason for caution, all the more so as the support for the seeding rate of 30 sila per ikli derives from (unpublished material from) Dur-Katlimmu, where grain will be produced almost inevitably in irrigation circumstances. There too the yields are low for Southern ideas. But cf. also H. Freydank, AoF 21 (1994) 29 on VAT 18097, probably from Kar-Tukulti-Ninurta. 32 This could pose a problem, but possibly also contribute solutions. The Kassite ikli is measured by "the big cubit" of 75 cm and is 118 of the old bur instead of 1/18. The backgrounds are explained by Powell, RLA 7 (1987-90) 481-82. 33 RA 35 (1938) 33-35.
275
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
(1975) which contains the evidence he could gather on seeding rates calculable from four Nuzi texts. Accepting this, and using the 1.8 ha imeru we arrive for barley at seeding rates of 44,7, 37.33 or 27.5 kgfha, which remains low. If we used the 1.1 ha imeru we would arrive at 73.25, 56.36 or 45 kglha. This looks better, but the basis for accepting it is arbitrary. Looking at (later) texts from Assyria we may observe that arable is not measured in units lower than one sutu. But the number of sila's per sutu is variable, not the number of sutu's per emaru. The specific sutu used is known locally and needs not be specified for those who know what they are dealing with. We must compare NALK 126 (ADD 621) and 127 (ADD 414), both from the Kakkullanu file, the texts on which Fales' article on Bet-Abi-ila'i is based. The most important text, NALK 127, does not specify which emaru is used locally, but from NALK 126 it seems to be clear that it is the one of 9 sila, and not that of 10 sila: the emiiru of 18000 m2 becomes one of 16200 m2 and if we were dealing with an 8 sila sutu we would arrive at 14400 m2. Practical difficulties are considerable. Why such differences? We do not know and can only speculate. The social background of those involved in NALK 126 and 172 suggests at some point in time some kind of division of land. In such a case the adaptation of nominal amounts measured on principle in the 100 sila emaru to the practical situation of a land division scheme with insufficient land could suggest the use of a smaller emaru. Another possibility could be adaptation to the quality of the land. With a seed emaru which stays at 100 sila the actual standard land unit assigned is reduced, which means that relatively spoken the amount of seed used increases.34 Practical Assyrian seeding rates are not known beyond the fact that the standard surface measure, the emaru, on principle should receive an emaru of seed. This volume emiiru will, however, like the surface emiiru, not always have contained the same number of sila's. The question is briefly discussed by He equates an Assyrian surface emaru with 5 (Southern) ikti and a volume emiiru with 100 sila, which would mean that 1 ikti receives 20 sila, against a "standard" Southern ikti 16.66 sila. When discussing the scant Nuzi evidence collected by C. Zaccagnini it was already concluded that there are arguments for a somewhat heavier sowing pattern in the North compared to that in the South. The equation of the Assyrian emiiru with 5 ikti rests, however strictly spoken, on a theoretical basis, not on a practical, actual calculation.
studied by Milan036 ad ARET I1 5 1, but remains problematic, not because speculation is wrong, but because its reasonings are difficult to follow. As to solid volume measures Milano's considerations depend on the reasonable supposition that it was not the intention to let valuable agricultural staff starve. The (implied) outcome is interesting: 1 (Ebla) an-zam, = 116 (Ebla) sila which is merely another word for 1 ni.sagSu, of which 10 constitute 1 barizu, of which there are in turn 2 in a gibar. We have therefore 1 gubar of 120 anzam, (one of) which is supposed to be the Ebla equivalent of the Southern sila. In a really suspicious way the Ebla gubar is thus identical with the Mari kor of the Zimri-Lim period, which we will encounter later on. (Such an Ebla gubar and a Mari kor at 0.84 litre per (120) sila equal 100,8 litres, and at 1.5 litre per kg, 76.2 kg of barley.) Regarding the surface measures we accept as a working hypothesis the supposition that the surface measure ghna.keSda, gha.kbSda, gAna.ki and ki in texts dealing with land all represent (for our purposes) the same,37we call this the "ki." There are speculations on its size by Pomponio, OLP 14, 1981, 5f and Milano, ASJ 9, cited. Pomponio regards his proposals as supposition^,'^ and he adapts the text in such a way that 1.5 nisaggu of seed is used for one "ki." Milano rightly protests. For comparative reasons Pomponio puts the "ki" on account of the amount of seed used at 1/10 of an ikii. This implies 15 sila per ikti, a value found in Fara, thus, as Milano objects, in the Southern irrigation area, whereas Ebla is situated in a basically dry-farming area. The result of Milano's own involved reasoning is finally based on Zaccagnini's conclusions regarding Nuzi. TM 76.G.188 deals with seed and land in (generally) the same places, but explicitly not in such a way that the amounts of land and seed are directly connected, which would have been a very easy to do. The seed mentioned could rather represent a partial restocking. The constructed seeding rates are for that reason "remarkably lower" (than the author concerned expects). TM 7562143 (ASJ 9, 1987, 184) is of much more direct importance. Structure: x cereals (as) seed (for) y "ki" (at) z (anzam each). The text obviously suggests what could be a seeding rate in the shape of a number of anzam's at the end of a passage, but the calculated relation between the surface unit mentioned and the amount of the cereal sown is in several cases lower than the number mentioned. Rates for various types of cereals vary, as should be expected. For barley it is given as 9 (Southern) sila per "ki," but actually provided is 7.5, which, in itself, makes the interpretation of ARET I1 51 as resulting in 7.8 instead of the prescribed 9 acceptable. But we are no nearer to a determination of the size of the "ki" as we cannot compare 7.5, 7.8 or 9 per "ki" to for instance 270 sila per bur (Fara) or 300 per bur ("Ur 111") and put it at somewhere between 1/30 or 1/33 of a bur: Milano rightly stresses the ecological differences. He can only go to Nuzi and manipulate Zaccagnini, in which way he arrives at 1 "ki" equals 116 of a (Southern) ikti. This is arbitrary: "average 4.3 anzam (= Southern sila) per "ki" equals 62 litresfha" (that is 41.33 kglha, low for rainfed areas). Direct translation in Southern terms: 25.8 sila per ikti, 464,6 sila = 309.6 kg per bur, results in 47.6 kgfha. Decisive is what the author regards as the appropriate amount of seed.
274
c. Ebla
It remains very difficult to work with the Ebla material. But there is enough evidence to suggest that agriculture in Ebla worked along the same general lines as that of other "palatial economies" in the North. The question of surfaces and seeding rates has been
3 4 In Nuzi similarly the 100x80 G ~ Rcould represent a reduction from an imeru which should have been 100x100 G ~ R in , which 100 square G ~ R receive one sila of seed. A volume imiru of 100 sila is known in Nuzi. 35 M. A. Powell, RLA 7 (1987-90) 488. Note that Powell duly expresses his reservations in view of the fact that the surface emaru is occasionally defined in s i t u ' s of unequal size. This is all very uncertain.
36
L. Milano, "Barley for rations and barley for sowing (ARET I1 51 and related matters)," ASJ 9 (1987) 177-201. 37 Cf. B. R. Foster, BiOr 40 (1983) 300ff.
277
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
The really important fact remains that the Ebla administration has fixed ideas about seeding rates, clearly adapted not only to type of crop sown but also to (local?) circumstances. Only an example of a measured field with stated surface can solve our surface problems. Institutional agriculture in Ebla works with a fixed relation between area and plough team: a team is responsible for a standard 200 "ki," see, for instance, Milano, A m u m I, 138-9, but the fact is immediately visible in many (of the not too many) texts dealing with land. We do not know whether this is an "iikaru," which on principle strains practical possibilities, or an attempt at setting a realistic task. If we follow Pomponio 200 "ki" is equal to 20 ikfi, 1.0.2 (bur) and with Milano we have 33 113 ikfi, 1.2.3 (bur) for the same 200 "ki," all of which seems to be unacceptably low for an institutional team, even if the furrow distance is drastically reduced and the fact is considered that Ebla teams seem to be of "donkey-type," thus less strong than the oxen preferably, but far from exclusively, used in the South.38
supported by an agricultural population of ostensibly the same ethnic background. There we are dealing with the results of a Landnahme, not simply with a military aristocracy which constitutes an upper layer of society. In Nuzi, and no doubt in the whole Mitannwanigalbat dominated world, Alalab included, we can assume a way of land holding which is directly related to the military structure on which society rests. But it would be wrong to suppose that this structure, once introduced, is peculiar to a specific ethnicity. It is an answer to technical possibilities, which, if successful, finds expression in a social structure of which land holding is inevitably the most important aspect in an agrarian society. We know little about a formal aristocracy in Ancient Mesopotamia. In the South we can sometimes recognise the leading institutional families from the fact that certain functions in certain periods tend to gravitate towards them. In the North we rather recognise, though dimly, a basically aristocratic land-holding elite. Assuming such an elite makes, at any rate, the personnel policy of the Assyrian kings easier to understand: they must create a countervailing power against this aristocracy, which, occasionally, can aspire to the title of king, ?arm, even in Assyria. That there was, at least in certain periods, a certain, for us very vague, idea about the existence of a kind of aristocracy which as of right was entitled to leading functions in society occasionally transpires, for instance in the case of the muikznu Silli-Sin who becomes king of ESnunna without being a m a d i i r ~or~in~ certain prejudices against ISbi-Irra of Isin, whose origins can, not for nothing, be traced back to the Mari area.
276
What we understand about the Northern seemand ratio determines in practise our interpretation of the metrology involved, not the other way round. All interpretation, in the end, depends on what is regarded as a "reasonable amount of seed." But seeding rates often seem to be low compared to modern usage. 3.
THE ETHNIC
FACTOR^^
a. General aspects Rowton has coined for the world we are dealing with the term dimorphic, stressing the fact that we are involved with an environment exploited in a more or less easy equilibrium by a sedentary and a non-sedentary population, which often has a different ethnic background, not coinciding exactly with the division sedentary-non-sedentary. The concept is typically one based on the observations of an outsider: the insider will react much more to the "them and us" of daily contact and there will have been an acute awareness of ethnicity. The ethnic factor in Northern Mesopotamia can be studied in a limited way only. That is a direct consequence of our sources, which are especially restricted as far as the North and East of our area is concerned. Nobody will deny the importance of the Hurrians, but little more than lip service to this factor can be paid. We can note that the Assyrian military machine in the first half of the first millennium is not able to hold the mountainous areas to the immediate East and North of the heartland, that the Turukku in the Mari period are a permanent menace to their neighbours and that at the same time Guti soldiers try their fortune in the whole area. But this is not sufficient for a balanced view of the importance of the role of the "people from the mountains," who after all are an overwhelmingly important factor in the third quarter of the second millennium. Factors of military technique will be very important in this. But that is not all. In Nuzi we can note the importance of a military, land-holding class of Hurrian extraction, but that class is
b. The Old-Babylonian world: Mari under Zimri-Lim or the stresses of a dimorphic society
In Zimri-Lim we have someone who is well aware of his own tribal backgrounds. The area he tries to dominate is the area of irrigation agriculture in the valleys of the Euphrates and the Habur, which is also that of the seasonal migrations of the pastoralists who are in part his supporters and in part his rivals. He is one of them and the other "great kings" of his time are also aware of this common background. On reflection we can perhaps question whether we are dealing with a local, Akkadian or Akkadianised population and tribal Amorites, or rather Hana as they seem to prefer to be called, of various quarrelling denominations: we actually see little of the Akkadians. The muikZnu's who seem to constitute the normal agricultural population are not specifically identified as Akkadians, as far as I can see. Yet it would seem to be realistic to assume that there is an element of the population which descends from that of the period of the ?akkanakkuYs,or rather the established rulers of the Ur 111period and their successors, which after all is not so long ago, somewhat more than a generation in Zimri-Lirn's time, which perhaps could be called "Akkadian." That the Amorites or Qana are acutely aware of their tribal identity does not mean that they are automatically pastoralists. Both their confederations4l are engaged in both 40
We will not consider the solid volume measures. 39 The importance of this subject for land holding in general in Mesopotamia can only be outlined. The matter deserves a fuller treatment. Most important is that (mobile) ethnic diversity makes domination of the economy by (static) institutions less likely.
The term has been studied by G. Dossin, "Le rnadiirurn dans les <
27 8
279
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
arable agriculture and in animal husbandry, and B la Wirth, it will be difficult to determine whether a person is to be classified as sedentary or non-sedentary. The available documentation could suggests that the Benjaminites are primarily pastoralists in the West. That includes the area near the mouth of the Balib where the power of that Simalite, Zimri-Lim, becomes very restricted. But their princes, kings, iarru's, are eager to possess (defensible) villages and land in the more easterly area dominated by the king of Mari. The Simalites cultivate land in the Mari kingdom: an unspecified number of the tribesmen is sedentary. The royal administration of Mari deals with the exploitation of the river valleys of Euphrates and Habur and the necessity to retain access for the nawii to the Habur triangle involves Zirnri-Lim constantly in Northern affairs, where his hold is precarious. Politics and land use are inseparable here, but it is not only sedentary land use which determines policies. Rowton's dimorphism has a clear social aspect. Sedentary and non-sedentary are closest together at the opposite social ends of society. The leading families of the sedentaries and the non-sedentaries intermarry and at the other end of the scale economic circumstances compel people to become sedentary instead of non-sedentary and viceversa. This last aspect will not be discussed, but it is evident that in the Mari area people belonging of the same family can participate in both ways of life, again B la Wirth. R. M. Whiting has provided for E6nunna a classical example of intermarriage between a local sedentary ruling family and a non-sedentary family at the end of the Ur In period.42It may be a sign of the times, but in Mari the daughters of the king are used in a policy which tries to bind vassals to a family which pretends to be a step higher on the ladder. Such daughters are not wanted, it would seem, for families which are a cut higher. Whether accepting, or soliciting, a daughter of another king (as first wife) is, in this period a sign if not of submission, at least one of subordination, requires further elaboration. Female members of the Zimri-Lim family occasionally marry the Benjaminite princes, who can, as a rule, be found on the Balib. The policy can be observed both before and after the war against the Benjaminites. Doubts on the status of the daughters of Zimri-Lim has been expressed by B. Lafont, who stresses the miserable size of their dowries.43 Marrying one of them at any rate means the presence of eyes and ears, and
probably also has consequences for property holding at home, but it does hardly strengthen the position of the son-in-law.
"confederation" in these circumstances means nothing more than that both groups have an identity of their own. The movements of the Benjaminites as pastoralists are lead by (two) merhu's who are not identical with their princes, such transpires at any rate from A.2741, translated with the relevant note as LAP0 16 433, a letter of Atamar-El, presumably to Sumu-dabi, an obvious rival of Zimri-Lim early in his reign. Also with the Simalites two (?) merhu's are in charge of the nawa, a word which seems to indicate primarily the whole conglomeration of men and animals moving through the landscape, perhaps including the land where it happens to stay. But in MAR1 6 (1990) 51, n. 64, Meptum, a faithful adherent of Zimri-Lim, is called "the merhu of Suly" by Durand. nawa is not a phenomenon which encompasses all pastoralists and herds, it seems to be group specific. But the merhu does not seem to be specifically in charge of a nawci, he could be, it would seem, a regional functionary. ARMT 2 90, republished as ARMT 2611 220, a Kibri-Dagan letter, differentiates between the nawa of the Benjaminites and that of the Bana. In ARMT 2611 114 the Benjaminite king of the Rabeans, DadiUadun, has a merhu of his own. In the RHM period the functions of kipifu and merhu could be combined, it seems. The matter has been studied in the case of g a b d u m a - ~ a ~ aand n Tuttul by W. Mayer, UF 19 (1987) esp. 154f. The merhu remains one of the least understood non-sedentary institutions, co-existing with primarily tribal and primarily regional "kings." 42 R. M. Whiting, Old Babylonian Letters from Tell Asmar, AS 22, Chicago 1987, 26ff. 43 CRRA 33, 119-120.
c. From Old Babylonian to Neo-Assyrian
The later Old-Babylonian period and its aftermath are thinly documented in a manner relevant for our subject in TerqdSirqu44 and in the texts from Alalab IV.The important question of the disappearance of the Amorites and the beginning of the Aramean penetration is practically undocumented. In Terqa the published documentation is dominated by documents dealing with private property, real estate, which constitutes primarily a preoccupation of a sedentary population. But the titles of the local kings, which stress the relation to Dagan on the one hand and the "king of (the) Hana" on the other, could underline the non-sedentary, still Amorite, connection. Dimorphism seems to continue, possibly at least into the Mitanni period. Even more interesting is perhaps the situation in Alalab in the period it was a Mitanni vassal. When (re-)editing what could be called a group of census-texts M. Dietrich and 0. Loretz45 have cautiously included the baniahbe among the "social groups," among whom the hups'u and the mariannu are immediately recognisable and the s'uzubu, also indicated by the term ehelena, can more or less be identified. The authors deem the meaning of baniabbe sufficiently explained by what they regard as a synonym, ekii, "poor,"46 but they also accept a derivation from hanG, "banaisch," obviously related to the in Mari frequent ethnic name Uana. Of some interest is that these hania&e are sometimes subsumed in the bupiu's$7 who are also called ERIN/DUMU.MES nu-me-(e) and who constitute obviously the ordinary agricultural population. We also find the haniahbe in the "house-lists."48 We are dealing with a social group with an ethnic background which is in the process of being absorbed into the ordinary (agricultural) population. We are watching the disappearance of the Amorites. Absorption of other elements with the same ethnic background into what will be the Arameans will not be documented. Ethnicity and social status are obviously related in other respects in Alalab IV and in the wider Mitanni dominated world, but it would probably be incorrect to suggests that a mariannu is automatically someone with a Mitanni (or Banigalbat) background. Being a mariannu is possessing a status which can be conferred locally in a (vassal) kingdom.49 But also the status of hanigalbat~itu~~ can be conferred, even by a private person, though 44
G. Buccellati, "The rural landscape of the ancient Zor, the Terqa evidence," 155-169 in B. Geyer ed., Techniques et pratiques hydro-agricoles traditionelles en domaine irrigue', BAH 136, Paris 1990; A. H. Podany, "A Middle Babylonian Date for the Uana Kingdom," JCS 43-45 (1991-93) 53-62 and also D. Charpin's note NABU 1994123. 45 M. Dietrich and 0 . Loretz, "Die soziale Structur von Alalab und Ugarit, 11," WdO 5 (1969-70) 5793. 46 Wd05(1969-70)91. 47 W d 0 5 (1969-70) 83. 48 Cf. M. Dietrich and 0 . Loretz, ZA 60 (1970) 88-123, esp. 115. 49 Ugarit: PRU 3 p. 140 (RS 16.132). Brak: N. J. J. Illingworth, Iraq 50 (1988) 99 no. 23. The son in question is made a free subject of the king, to avoid the term citizen. For Alal* cf. AT 13.
280
G. VAN DRIEL
in the presence of the Mitanni/IZanigalbat king. That does, however, not mean that the ethnic element of the question is meaningle~s.~~ The written documentation from Northern Mesopotamia in the second half of the Second Millennium is not rich, but it does not lack traces of the importance of the phenomenon under discussion: ethnic diversity is a constant factor in the North, it influences economic and political structures and it is a counterweight against centralisation. d. The Neo-Assyrian period: the response to ethnicity
For the Neo-Assyrian period the subject of dimorphism is difficult to treat. That is the immediate result of the documentation available. The Assyrian royal inscriptions are primarily concerned with the deeds of a particular king, they only tell about the enemy of the moment, not about the broad outlines of the situation. General policies are rarely indicated. After the ninth century the Euphrates area was securely in Assyrian hands: no campaign equals no information as far as royal inscriptions are concerned. Only in the earlier, eleventh century, inscriptions of Tiglathpileser I and in the Broken Obelisk, assigned to ASSur-bel-kala, do we find the internal small-scale military operations which are a clear sign of the breakdown of dimorphism caused by the increasing Aramaic pressure. The reasons for this pressure have been discussed often. Worsening of the climate has been adduced as the main reas0n.5~Increased insecurity inevitably means reduced dependability of arable production. The result is hunger, even in the core area of Assyria.53 The Assyrian reaction is that of an arable based society. If Mari represents a dimorphic society, the term is far less applicable to Assyria, whatever the first passage of the Assyrian Kinglist may suggest. The nomadic aspect of SamSi- dad's ancestry, which is incorporated artificially into the list, is not ideologically exploited by the later Assyrian kings, even though they try to dominate a large area where a pastoralist way of life is natural. It is not a concept to which they adapt easily. The core of the Assyrian kingdom is the arable land East of the Tigris where all their later capitals are. We cannot see the Assyrians as migratory pastoralists, they are farmers living in an area of dry farming, no doubt, where feasible, with additional irrigation,54 as is well attested in neighbouring Nuzi in the fourteenth century. A sedentary way of life is natural to them. There is no better illustration than the reaction of their kings at the moment their country starts to recover from the Aramean invasions, which had cost them all their land West of the Tigris. ASSur-dan I1 tells us that he brings back the terrified Assyrian population, which had fled as a result of hunger, to their towns and houses, and, this is the returning refrain we also find with Adad-nirari 11and Tukulti-Ninurta 11: ekalliiti ina Jiddi miitija arsip epinniiti ina Siddi matija arkus, "I built palaces in the
Whether we are dealing with Arians or not is immaterial for the present discussion. The groups concerned primarily carry meaning for those involved. 52 J. Neuman and S. Parpola, "Climatic Change and the eleventh-tenth-century eclipse of Assyria and Babylonia," JNES 46 (1987) 161-182 53 For this cf. A. K. Grayson, TCS 5 (1975) 189, Assyrian Chronicle Fragment 4. 54 R. Jas reminds me of the importance of the Middle Assyrian Laws in this respect.
THE MESOPOTAMIAN NORTH: LAND USE
28 1
(various) districts of my land and constructed ploughs,"55 after which follows a phrase dealing with the gathering of huge amounts of cereals and the increase in the number of horse teams, this last no doubt a military and not an agricultural measure. Returning to what for the Assyrian is normal circumstances means palace building, i.e. the restorations of regular administration and the reintroduction of a plough-based system of arable agriculture. We can assume that this is royal policy. A3Sur-dan talks about resettling Assyrians, not about others. Assyrian policy in general can be classified as one of incorporation with attempts at weakening spots of persistent resistance by selective deportation. The non-sedentary population plays no special role in this, but it has been noticed that Assyrian proceedings against Arabic enemies shows a particular harshness.56 Awareness of the ethnic aspect cannot be denied the Assyrians. There is no sense in trying to eliminate the practical importance of the factor in the Ancient Near East. The ethnic soldier is not unknown in Ancient Mesopotamia. The Aramaic Itueans and Gurreans of the Assyrians are perhaps the best known, but they also kept, it would seem, the Samaritan charioteers as a special unit.57In later Babylonia we find all sorts of ethnic groupings settled, it would seem, as compact groups already at the beginning of the Persian period: Misiraja, BaniSaja, Ludu.58 But this differs from dimorphism: the Assyrians incorporate by settling foreign ethnic elements, in this Chaldeans and Persian follow them. The best illustration of the natural inclination of the Assyrians to solve problems on the basis of organised arable agriculture, even in areas which are on the face of things less than suitable from an ecological perspective, is the attempt to create villages, or perhaps farmsteads is a better translation for the term URU.SE, in the steppe East of the Habur, South of Jebel Sinjar. The Palil-ere; steles9 has been discussed often, the details need no repetition. It is indicative of a mentality: real exploitation implies sedentary arable agriculture. That security of cornrnunicationsplayed its role can be accepted. Assyrian evidence on the seasonal movement of pastoral tribes in what is now Syria is strikingly absent. It is unlikely that the Assyrians would be able to stop such movements. The explanation must lie in the documentation. But also the local rulers of the Euphrates area in the ninth century are in their inscriptions primarily settled rulers.60 There is evidence concerning herding, as is to be expected, in the Harran Census documents, but it is all too obvious that for the scribes who wrote down the register, the herdsman had to have a fixed tax address. ADB 8 (SAA 11 206 I 10') is a good example: "170 sheep, quote (iJkaru)6l total of (township) garbat-ila'i-Be1 in the steppe 55
Cf. A. K. Grayson, RIMA 2, Assyrian Rulers of the Early First Millennium BC I(1114-8-79), 1345: 60-65; 154: 120-121 and 178: 132-3. 56 J. Reade has observed that only in the wars against the Arabs are Assyrian soldiers shown actually manhandling women. Cf. the pertinent remarks by K. Radner in her contribution. 57 S. Dalley, "Foreign chariotry and cavalry in the armies of Tiglath-Pileser 111 and Sargon 11," Iraq 42 (1985) 31-48. 58 Cf. Camb 85,22 IX Camb kB; Dar 351,25 V Dar 13. 59 S. Dalley, Iraq 30 (1968) 139-153. 60 A. Cavigneaux, B. Kh. Ismail, "Die Statthalter von Subu und Mari im 8. Jh.v.Chr.," BaM 21 (1990) 321-456. For instance the beginning of inscription no. 2 is very informative. Whether "quote" is an acceptable translation is open to doubt. When arable holdings or vineyards are listed no taxes are mentioned. iikaru is not strictly a tax-term, it denotes an obligation in a general
283
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
(madbaru) of the (town) of Balilp". This is probably the obligation of the shepherd (SIPA UDU.MES) mentioned with his family of five in the damaged preceding passage. The difficulties the situation caused for the scribes are illustrated by ADB 3 (SAA 11 203 I11 6) in a passage dealing with a small estate belonging to the cook Sin-ere8, which Adad-remanni the chief cook of &man, presumably his superior, caused to be registered: not included among those actually seen by those performing the census is 1 SIPA UDU.MES p a - ~ u"1 , shepherd, dispersed."62 No name is given.63 This is herding based on fixed settlement^.^^
temple for non-sedentaries, that of ASSur in Assur perhaps a second. The governor of Terqa can limit access to the sanctuary to a certain degree or characteristically refer back the decision to the king: dominating the institution does not give them much power over those who want to be admitted, and who pose a danger. Though the king of Assyria is traditionally the SANGA of ASSur, bleakly translated as "vice-regent," but, if the Southern parallels of the title are taken seriously, in principle the person who dominates the economic aspects of the cult of the god, thereby gaining considerable social and economic influence, that does not turn the ASSur temple into an economically important institution. The use of this type of title is rare in the North, only the case of the rulers of Sadikanni springs to mind. Not a sign that "temples" are politically or economically that important. Finding faithful adherents amongst all groups of society is not translated into economic power for the temples of the North. Before the Neo-Assyrian period documented palaces were in the main, economically spoken, small and local, though the claims for Ebla may retain some credibility after verification. As agriculture is generally the basis, the number of plough teams employed should be taken as the yardstick of economic importance. Evidence for textile production is also pertinent. But not even the palace at Mari seems to have been economically a strong institution, as Lafont agrees. Before the Neo-Assyrians Northern political entities are regional,65 they operate in an unstable hierarchical system where mutual gifts express relative position. Resort to force is the ultimate decider. Whether that in itself is a sufficient reason to assume that the structure of Northern society was completely different from that in the South must be considered as doubtful. The important difference between North and South are the big divine households of the South which do not occur in the North. But they are the human tools of those dominating the palace sphere. In the South their existence contributes to the possibilities of political centralisation. Important is also that the role of the non-sedentaries is more pronounced in the North. We do not possess the administration of an arable estate managed directly by a palace.66 But we can demonstrate that where traces occur exploitation takes place through a plough team based structure. What is also visible in the North as a more or less permanent feature is that the palace transfers to individuals, members of the royal family, persons of rank, elements of its holdings, often for lifetime, after which all, or most, of what is granted is re~uperated.6~ Such practises will not have been unknown in the South, but they are less well documented. What is granted serves as remuneration, but also reduces the burden of administrating possessions: the beneficiaries must supply the palace with some of the yield of the estate. The workings of this system are particularly well visible in Mari.68
282
4.
THE DEGREE OF INSTITUTIONALISATION: THE GRIP OF THE CENTRE ON ITS HINTERLAND
Temples and palaces: hot all While the earlier economy of the Mesopotamian alluvium can still be seen by many as an economy totally dominated by the institutions, especially, in the Third Millennium, the temple, such an approach is impossible for the North. This is the direct result of the natural circumstances in combination with the ethnic situation which exclude complete institutional domination of both arable agriculture and animal husbandry, certainly in the longer term. A static, localised, institution cannot dominate a largely non-sedentary population which drifts in and out of both arable agriculture and animal husbandry. That does not mean that such an institution cannot find faithful adherents among such a population. The temple of Dagan in Terqa is a prime example of the importance of a sense, which can mean that the amount mentioned is to be delivered, or for instance that there exists an obligation to care for a specified number of animals. 62 I follow the SAA translation. 63 Specialised cow herding is mentioned by ADB 1, SAA 11 201 I1 16. In ADB 9+ (SAA 11 209) the second column, which contains a list consisting of names followed by numbers of sheep is sadly damaged. That there are owners of considerable numbers of sheep is indicated by ADB lo+, SAA 11 210 I11 l', which mentions 7621 sheep actually seen and 450 not inspected. The herding of cows presupposes the presence of sufficient, that is ample, water. 64 Though I have not the slightest doubt that the Sabi Abyad population was engaged in herding, I find it difficult to agree with Wiggennann that the importance of that activity can be reconstructed from VAT 18087+, Freydank, AoF 7 (1980) 89-1 17, esp. 104ff. That text deals with a very specific group of families, of whom some have no possessions at all. As the group stays in the capital of the day, they cannot be regarded as representative. Nor would I regard a procedure which derives the total number of cattle from the number of plough oxen as sound. At any rate in the South the herding of cattle and the keeping of plough oxen are separate activities, plough teams having (at least occasionally) additional cows (two cows to four oxen) to keep numbers up to strength. Plough oxen are the responsibility of the plough teams, not of the herdsmen. The head of the team is responsible, he must hand in the hide and tendons of a fallen animal. The for our ideas incomprehensibly high number of (uncastrated) bulls in herds of cattle and rams and billies in flocks of sheep and goats are explained by the fact that these animals are required for offerings, they are not identical with the (castrated) oxen of the plough teams. The fact that high numbers of (intact) male animals occur in texts does not mean that they were kept (during the whole year) with the other animals, the documents are only administrative lists, as the total number of the animals often indicates. Hides of animals used for offerings will also turn up in administrative documents. In general animals received from herdsmen are rarely slaughtered straight away, as a rule they are fattened on barley first.
65
I want to avoid the word "state." The most important "political entities" are monarchies with all that that implies. The monarch can not rule on his own and that influences all aspects of government and especially agriculture. Cf., however, F. A. M. Wiggermann's contribution to this volume which describes the workings of a segment of a palace-type estate. 67 But not always as the development of Alalab into a vassal kingdom demonstrates. 68 The question has been studied in the unpublished master's thesis of F. van Koppen.
284
285
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
The importance of the "palace" as focus of production in general and agriculture in particular in the thinking of M. Liverani and C. Zaccagnini69 has been underlined in the contribution of I. MArquez Rowe to this symposium.70 But the palace, in this way of thinking, is more an exploiter than an organiser. The palace cannot survive if it does not actively stimulate production. The Alalah texts with which MArquez Rowe deals, coming from a palace administration, inevitably illustrate the subordination of the villages to the, or their, palace. The economy of the Balih Valley dunnu studied by Wiggermann illustrates particularly well the status of a "village with its farms" assigned to a high official -whatever his title, this "king of Hanigalbat" is not an independent ruler- and thus belonging to an estate which for every practical purpose can be regarded as a segment of the palace. It can be reclaimed if the occasion arises. The other possibility is that such an estate comes into the hands of the high land holding aristocracy, which we can assume to be an important factor in the North. It will be difficult to draw a line between junior lines of the royal families and other landed magnates. That is demonstrated by the Nuzi material. We can assume, along Mari lines, that royal daughters married exactly this type of person. Aristocratic estates are divided through inheritance and disintegrate through indebtedness, there is no reason why North Mesopotamian estates should not be subjected to these forces. A special palace related type of village would be that where the villagers, from the possessor of a chariot to the bowman, with all sorts of other specific obligations in between, for instance that of serving as a horseman or of being subject to transport duties, have been assigned fields which enable them to fulfil their specific obligations. This is not explicitly documented, but is certainly behind much of what is visible in Nuzi and in Alalah IV. There the inhabitants of villages are socially stratified in several layers, of some of which the military backgrounds are clear by their association with, for instance, "bows" or "chariots." Such villages, or the part of them with the military groups, come into being as the result of a formal arrangement between the palace and the (ancestors) of the people concerned. For the moment such formal arrangements are not documented, but that is probably a question of time.7l The military often will have had a tribal background, but in these villages we are not dealing with the spontaneous settlement of tribal elements.72That villages, or rather their fields could be burdened by special obligation is clear from Sargon's Maganuba d0cument.~3
The important question is whether all villages belonged to palace-headed hierarchical structures, thus were either (completely) closed estate villages with a plough team based structure or military settlements, or whether there were villages of a different type, that of the village owned by a group of owners, more or less identical with the inhabitants, possibly with the involvement of additional outsiders, for instance town dwellers who were drawn into its affairs through the workings of debt laws. These villages need not be seen as "communities" in an ideological sense. In the first case the villages produced for the organizations to which they belonged, which took all but the costs which included what is produced on sustenance fields by the subject population. In the second the government is mainly involved through taxation and conscription. The existence of nonestate villages is -for me- inevitable in an area with a population which is to a degree nonsedentary where there is ample land. That there is ample land, especially of indifferent quality, is unproblematic. I have no difficulty in accepting the possibility that villages are incorporated in tribal structures, but traces will be difficult to find in Mesopotamian type written documentation. Such villages can (partially) escape from taxation and conscription, but they can be plundered, in which case they occur perhaps in royal inscriptions. In contrast to the South, where institutional archives dominate, the North has archival material which demonstrates the existence of villages inhabited by people with rights to the land they till. As the archives come from cities like Nineveh texts will attest in general the alienation of possessions by inhabitants of the villages. Whether this is symptomatic for the period concerned remains the question, but there is always a possible link with the agricultural debt problem resulting from the lack of reasonable credit facilities. There are villages with one owner and there are villages with many owners.
69 The relevant matter is best exposed in C. Zaccagnini, "Asiatic Mode of Production and Ancient Near East, notes towards a discussion," 1-126 in C. Zaccagnini ed., Production and Consumption, Budapest 1989. 70 Regrettably enough, this will not be published. M. Jursa found a document specifying the contents of a Neo-Babylonian rent farm very soon after the existence of such a document had been predicted. The formal, palace related, arrangements behind the, ostensibly private, Old Assyrian trade are becoming clear: K. Hecker and, and J. Eidem have published the relevant documents. The formal arrangements behind military land divisions will also be better documented than the anecdotal evidence in, for instance, royal letters. 72 That we are dealing in these cases with a population that came into the system from the outside is suggested by the explicit military obligations: the ordinary subject would have had to serve on another basis. 7 3 For the Maganuba document, ADD 809, see now SAA 12 19.
A discussion of the "Wilkinson-model"74 from the point of view of an Assyriologist can only be attempted with better practical insight in the (written) Northern economy. The situation differs considerably from that in the South, where the Hunt model can be compared with (institutional) agricultural practise: the bureaucratic evidence from the North which must be the basis for all attempts at quantification is much less secure. The Wilkinson model suggests ceilings and makes "stress" visible. It does not consider the possible effects, in the North, which an economic sector has which operates in a manner which differs from the operation of an ordinary, primarily subsistence oriented farmer. That sector is not market oriented, but it is geared to producing a surplus which supports a local "institution." The palace works as a hierarchy of smaller local palaces which in the end need not be more than a one plough team oriented "farm." Institutional agriculture can in this way reduce some of the practical negative results of good land being situated at too great a distance from a primary or secondary settlement. If it works like a Southern plough team it extensifies agriculture, uses a considerable surface with a limited amount of seed, manand ox-power. Its structures will be ephemere and in the long run add to the scatter of settlement debris marking the intensive (agrarian) use zones.75 The institutional, palace,
See note 10. I have no difficulty in accepting the fertilising of the environment of settlements with their detritus, it is practically possible with donkeys and the big Mari-type plough teams would have had the man
74 75
286
G. VAN DRIEL
sector has the means to situate its man and animal power where suitable land can be found. It works according to sets of norms, which, when known, help understanding. That such norms are not understood in the same manner by all is clear, not even within one institution. In Mari norms are not those of Ur I11 LagaS, they are adapted to local circumstances and habits. 5.
THE STRUCTURE OF ARABLE AGRICULTURE IN THE NORTH
We can expect a structure which is not too dissimilar from what can be assumed for the South, but reconstructing that structure from textual material is considerably more difficult because the material is much poorer. The texts deal with the institutions and less intensively with the holders of military "fiefs." Subsistence farming and the tribal world do not appear in the administrative texts, but that is not unexpected. a. The Old-Babylonian world This must deal with generalities. These are best visible in Mari, ecologically a very specific area, not of necessity representative for the whole North. The particulars of Mari agriculture are treated here in detail by B. Lafont, but for the general outline we may draw on what has already been published. The importance of the Mari evidence is that it illustrates the use of plough teams in the exploitation of the palatial domain, that parts of that domain are hived of to "persons of rank," who exploit those parts for their own purposes, but who are subject to certain levies and certainly are not the owners of the land concerned, as it is taken back on death or disgrace. Even more important is that Mari indicates that not all arable is exploited by the institutions or given out as sustenance lots: there is land in the hands of the ordinary mus'kenu's and members of various tribes also exploit land. In the South, both in earlier and in later times, the importance of institutional land use can be measured by the number of plough teams employed. For a moment disregarding animal husbandry, the number of ploughs determines the economic size of an institution. On principle the plough teams are directly related to the settlement system, as they are family based, even if that is not always directly visible. In the earlier periods (Akkad, Ur 111, perhaps Old Babylonian) the strength of team depends on the circumstances, some teams are much stronger than others and their tasks are obviously related to their strength. In later periods (Neo-Babylonian) the strength of the team is, at any rate in an institutional context, on principle standardised, but it often fails to reach the norm. It is often possible to calculate a standard task. What is the situation in the North? The first difficulty, already discussed, we encounter is that of the system of surface and solid volume measures used, the second is that we know relatively little about the practicalities of Northern agriculture, which anyway cannot be approached as something uniform. What is done in one way in the irrigation areas of the Euphrates, the Habur and on a more modest scale the Balikh, may (but not always must) have been done differently in the dry farming areas. Perhaps a detailed study of arable farming in Nuzi could help: are there notable differences in the tilling of irrigated and non-irrigated fields? Only in
THE MESOPOTAMIAN NORTH: LAND USE
Nuzi is there a documentation which differentiates between such fields, as irrigation or non-irrigation are not self-evident in the region. Elsewhere texts automatically deal with the obvious way of planting as far as the user is concerned and contain no specification. First there is the common, almost -but not completely so- deafening silence about fallow. In later Assyrian leases we find the contrast between karapbu and mZris'~,7~ which we can with some confidence understand as respectively fallow and sown/actually used, as the number of years indicated in leases for both states and the amounts of land are equal. This leaves, however, the possibility that several years of continuous use and fallow followed one another. Important is also the question of the use of the seeder plough. To date, I can explicitly document the use of the seeder plough in the North only for the river valley agriculture of Mari: as M. Birot states in his edition of ARM 27 3, a letter send by IlSuna~ir,one of the governors of Qattunan on the Habur, land sown with sesame, a summer product, and therefore certainly irrigated, contrasts A.SA s'irbam and A.SA sipbum, respectively "sown in furrows" and "scattered (by hand)." That the seeder plough is known in the North is clear, but that does not imply general use all over the whole area. One could suppose that with the expected extensive land use the seeder plough inevitably will prove its efficiency, just as in the (institutional) South it makes possible an efficient type of extensive agriculture. The correspondence, especially that of Zimri-Lim's governors on the Habur, also indicates that the land belonging to the palace, and also that exploited directly by persons, generally high officials, who are provided by the palace with land and wherewithal, is tilled in a plough team system. But the system does not give the impression of being very stable as both the personnel available and the land to be tilled vary. As we have seen, there are norms, but they require "interpretation." The central administration seems inclined to impose high(er) work loads, sometimes perhaps against local opinion. Explicit is ARM 2611 76, Enlil-epuS to Asqudum, in MiSlan in the Euphrates valley. It had been arranged that a plough(-team) would plant 70 ikli, the king orders 100 ikli. If we take the ikli to be the same surface-unit as in the South, 3600 m2, that would be a task of 36 ha instead of 25.2 ha. For comparative purposes it could be stressed that in the preceding Ur 111period in LagaS, in the South, a plough was assigned a practical workload of some 6 or 7 bur, 108 to 126 ikfi, around 40 ha. In the Mari period in the same LagaS area of the South, in the contemporary kingdom of Larsa, a work load of 10 bur Ijkr plough is attested, almost 65 ha.77 If the same extensive, but effective, manner of tilling as in the South was applied the royal 100 ikli imposed by the central administration in Mari, does not seem to be excessive. All this is based on an assumed identity of the Northern and the Southern tkli and on the uncertain assumption that plough teams are of comparable strength. The Mari correspondence suggests that the Mari palaces, especially those on the uabur where not very large agricultural enterprises, and that the land and ploughs assigned to officials (and members of the royal family) did not considerably increase the surfaces tilled. 1000 ikLi sown is mentioned as a major achievement, which characteristically causes harvesting problems. This means that a local palace on the uabur 76
power. A problem might have been that for it to have been effective the material would have had to be ploughed under rapidly, which might have been difficult.
287
Cf. CAD K 206-7. Note that AT 218, from level IV, contrasts ba-ra-up-9eand qa-ra-up-he,for fields belonging to the same official, which suggests the possibility of biennial fallow in the West. 77 D. Charpin, BiOr 36 (1979) 198-9, Archives W.
288
would require, inclusive of fallow, some 720 ha of land - but is this really all palace land? In ARM 13 39 we find a warning to the king that he should not restitute to the Benjaminites lands which have been taken from them after a revolt, the result would be that the seeding efforts of 19 plough teams of various origins and owners would be lost. This seems about the maximum number of teams mentioned in one go in Mari known to me. The personnel strength of a team is best assessed by ARM 29 1 (LAPO 17 754): ASqudum had assigned 12 men to a team, governor IlSu-nasir wants 15, who then proposes that 10 is enough? It is possible that the strength of a team varies with the season or is influenced by the way irrigation is practised: that both gravity irrigation and drawing of water occurred seems certain, cf. ARM 2611 171 on the nature of land in MiSlan in the Euphrates valley, presumably North of Mari. Drawing water is always a labour and time consuming affair. In general we will have to suppose that the central administration, -the king- has a much more optimistic idea about what is feasible than the practical men on the spot, who, of course, had a vested interest in underestimating the mean,s available to them. As Durand has stressed, the real problem may have been the lack of sufficiently trained and capable "farmers," engar's, who could lead a team. In ARM 1 44 (LAPO 17 753) SamSiAdad complains that he has constructed many ploughs, or perhaps, rather, has formed many plough teams, but that capable engar's are wanting. The continuation of the letter makes clear that it is not the lack of men in general, but lack of skill which constitutes the drawback.78 We can conclude that a plough team in the Mari irrigation area tills about 36 ha with some 12 men. That is not as much land as an Ur I11 team in the South and it requires more men. The number of oxen is not easily established, unknown in this particular case. We find teams of 8 or 12 oxen, which might indicate that they could field two or three actual plough teams of four oxen and four men e a ~ h . ~This 9 would reduce the amount of land per (real) team to 18 or even 12 ha.80 For the size of the parcels available as sustenance plots two texts are of some importance. In Qattunan under governor Zimri-Addu we find that an ordinary soldier receives a sustenance field of 5 ikii, 1.8 ha: ARM 27 107. That actual fields were often much smaller transpires from ARM 2212 328, a list of fields bought on various occasions and in various locations by a certain Warad-Sin. The text could be taken as an indication 78
THE MESOPOTAMIAN NORTH: LAND USE
G. VAN DRIEL
In this connection reference should be made again to F. van Koppen's unpublished master's thesis, as he argues that there are reasons for the assumption that in Mari institutional plough teams were managed by entrepreneurs. He should publish these findings himself, especially because some solution for the problems with the exceptionally high yields in Mari could possibly be found here. I must thank him for allowing me to refer in this way to his findings. 79 ARM 24 13 mentions one team of 7, two of 8 and one of 12. Comparatively an institutional team of four oxen is normal, especially in the late period. What the exact benefit of four against twice two is, I do not know, but in the private MuraSu archive the actual change over from teams of two to teams of four can be observed. Less than four men per team caused obvious problems in the Neo-Babylonian period. I could imagine that preparatory ploughing was done by two teams of two and the actual sowing, for which there is a narrower time slot, with the complete team. This supposition is certainly at variance with what can be seen in Umma, where preparatory ploughing definitely requires a team of three men. Compare this with Wiggennann's interpretation of the Sabi Abyad evidence: a team with two oxen tills 40 ikfi or 14,4 ha.
289
that privately owned land was scattered, even over more than one community. Plots of one ikii or less are not uncommon.81 b. Other archival material from the Mari erag2 None of the other archives throws light on the question of dimorphism, though this does not mean that it is not a matter which deeply affects the situation in the area of the Habur triangle and the area North and South of Jebel Sinjar. This is an area where agriculture is practised on an institutional basis, but which is also essential for the survival of the herds of all manners of pastoralists. Mari is in itself sufficient as a source in this respect. For Subat-~nlil,Sehna, Tell Leilan, for the present we only have SamSi- dad's remark in the letter found at Mari on the lack of experienced plough men, other published material from the site itself is not informative on the structure of agriculture. But it is enough to indicate that institutional, that is palatial, agriculture probably had the same plough team-based structure as found in Mari. This is important as we have crossed the dry farming boundary. We can look for more information in other more or less contemporary text groups from roughly the same area. The texts from Chagar B a ~ a rhave , ~ ~been republished by Ph. Talon. They obviously tell something about the functioning of a (smaller) local palace and its economic base in the period of SamSi- dad. The texts on the assignation of barley as fodder for various teams of oxen and donkeys indicate that we are dealing with an official organization, a "palace." Four texts are of considerable importance: Talon 30,53,60 and 57, all from the eponymous year of Adad-bani, with the dates 10 111 (twice), 1 IV and 1 VI. The last three texts mention an Adad-bani as the person responsible for the animals, -who is, incidentally, not likely to be the same person as the eponym-, in Talon 30 it is Jakua. Adad-bani is in charge of 50 GU4 e-ri-5, "plough oxen," Jakua has 32 plus two one-year olds, and additionally some sheep and goats, a combination which is rare in texts, as cattle on the one hand and sheep and goats on the other are always and everywhere managed
8l
Lafont has grasped the nettle of the Mari evidence regarding yields, especially in his Table in 3 5. There is no reason to discuss the matter here: his considerations coincide with what was suggested during the meeting of the symposium. The problems remain, the exceptionally good results are embarrassing, just as low local results are caused by a mishap. 82 No more than mention needs to be made of the few texts from Tell el Hawa, J. A. Black, Iraq 51 (1989) 41-5, cf, for a for our subject important elimination J. Eidem NABU 1993178; Tell Taya, J. N. Postgate Iraq 35 (1973) 173-5 and Qalat el Hadi, see J.-M. Durand NABU 1987137. Here also Hamam et-Turkman, for a published text see W. H. van Soldt, 275-76 in Th. P. J. van den Hout and J. de Roos eds., Studio Historiae Ardens (Fs. Houwink ten Cate) Leiden 1995. For texts reported from Kazane Hiiyiik in the Urfa area, NABU 1996190. More numerous are the tablets from Tell Bi'a, Tuttul, M. Krebernik MDOG 122 (1990) 67-87 and MDOG 123 (1991) 57ff. from the Sam~i- dad period, but, for the moment, they do not contribute to our subject. All these texts constitute an impressive sign of basically uniform administrative habits obtaining in Sam~i- dad's kingdom. Much less clear, as yet, if we discount the Anatolian aspect, is the equally impressive network of commercial contacts we can surmise. 83 On the identification of this place cf. Wu Yuhong, NABU 1994167, but this is probably not the final word, cf. M. Falkner AfO 18 (1957-58) 19 and 35, supported RGTC 3 139.
290
29 1
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
separately.84 We are dealing with an organization which could field at least 20,5 teams of four oxen, which compares well with published evidence from Mari. All oxen receive 3 sila of barley per day, which presupposes additional grazing. The real interest lies in the dates. In MAR1 4 (1985) Charpin has synchronised the "Ekallatum-calendar" used in SamSi- dad's kingdom and thus by our texts, and the Mari calendar. He has noted that harvest time in Mari falls in months I and 11of the local calendar, answering to months VIII and I X of the Ekallatum calendar. That means that at least month VI of that calendar, for the complete month of which the oxen received fodder, is much too late for sowing. The sowing period is normally the one in which the oxen employed in ploughing receive barley as fodder, but here the period is much longer, even if it does not encompass the whole year. The possibility could be considered that these animals were also used for transport duties. The same text group contains texts dealing with fodder for horse and ox teams specifically used for transport. Land transport will have been unavoidable in the Uabur triangle. Some donkey teams could possibly have been employed in agriculture, but that is not indicated explicitly in the texts. It should be noted that, as elsewhere, cattle were fed barley on a regular basis in preparation for slaughter.85 Both land transport and fattening require a share in the surplus produced. The tablets from Tell Rimah, Qatfara, also touch only marginally on matters of agricultural organization. But there are some texts which refer to an underlying routine. In OBRT 303, an undated letter order the "governor," Siipilu, orders AhamarSi to hand out the SE.BA SE.NUMUN z.2 SA.GAL.@A (of day x?), a combination which clearly indicates the provision of the rations, the seed and the fodder required by (a) plough team(s) in the sowing period.86 Another letter, OBRT 280, speaks of the @e-pi-nu-um Sa h a - l a - a s PN, the plough teams (collective) of the district of PN.
Important is OBRT 322, a damaged tabulated list registering plots of land, amounts of cereals (barley, rarely burru, "wheat") and the person concerned.87 The text is divided into (at least) four sections, each representing a geographical unit, the first of which is Qatlara, with 604 ikii, in Southern terms 33.1.4 bur or 217.5 ha of land. The editor has established that if the first column represents a certain number of ikfi's, the second lists ten times that number in sila's of cereals. He prefers an interpretation that makes the amount of cereals seed (and not yield to be delivered). In Southern terms, accepting for the moment that an ikii is an ikii, that would mean that 1 bur of land is sown with 180 sila, which is strikingly light in view of the 300 sila, one kor per bur, which would seem to be common in the South. As a (share in the) yield destined for the owner, the institution, this would seem to be excluded, unless a nominal recognition is meant. The text shares many of the personal names occurring with other texts from the archive and we are probably dealing with land assigned to personnel belonging to the palace. The strikingly light seeding rate requires some explanation. Light sowing could be regarded as a sign of speculative planting, in the hope that there will be enough rain, but the text seems to deal with sustenance land where this type of speculation would seem to be out of place. In OBRT 317 we find again the 5 ikii (1.8 ha) plot which occurred in Mari as the plot of an ordinary soldier, with other, smaller plots of 2 and 3 ikii. In Mari we can assume that such land could be irrigated, otherwise it will not have been usable. This type of small plot does not warrant the use of an ox-drawn plough, at any rate not in isolation. OBRT 308 indicates that there were also (standard) plots totalling 10 ikii, and the assignations could take place in different locations. South of Jebel Sinjar we will be in a marginal area, though on long term average it will obtain some 300 mm rainfall per annum. It therefore stands to reason that there too some sort of irrigation is practised. Letter order OBRT 297 deals with a plot of 1 ikii to be given to a person, with the water required for cultivation. Similar is OBRT 296, with insistence that the water must be given "this very night," or otherwise tomorrow. The land provided is taken from that of the governor. The interpretation of OBRT 295 is somewhat difficult "give water to the township of x, that they may cultivate Sa-pi-il-ti A.SA-Su-nu, (alsotat any rate?) the lower part of their field."88 We can continue in this vein, but it is obvious that we are dealing with dispersed plots of land managed by the institution and handed out to various individuals or groups, who have rights to certain facilities. OBRT 299 is, for instance, a complaint by a woman concerning, probably, the sowing of her cereals. In OBRT 294 20 ikii in a certain place is assigned to someone with the stipulation that "they must work it with the majiiru." The majaru is a plough used in the South for preparatory ploughing, before the seeder plough comes into play. "They" are probably the personnel of the institution. The probably related OBRT 156 and 157 suggest that harvesters were allotted centrally, the texts are concerned with fields assigned to the queen. In general the Rimah texts deal with land allocated to personnel related to an institution, to all likelihood "the" palace, even if a number of relevant texts have been found in a temple. SemSara, SuSarra, has sadly, not received the separate treatment promised. It was heavily involved in the affairs of the Turukku. Their presence is of some importance for the question of dimorphism, whether they are a really foreign element which not so much
84 That Jakua was (also?) a fattener could be indicated by Talon 76: 1, cf. 77: 9, which incidentally provides a reference to an engar located in a place called Zammaranum. The in the Sam~i- dad kingdom secondary centre from which the archive derives, controls, as is to be expected, plough teams located in small sites in the neighbourhood. That Adad-bani had a supervisory function concerning cattle in general is indicated by the colophon of Talon 35 and 69. 85 Talon 23 mentions a herd consisting of full-grown bulls (capable of breeding, G u ~ . A B , only eight, each receiving 10 sila (one sutu) of barley per day), and 35 three year olds (4 sila), 40 two-year olds (3 sila), and 131 one year olds (1 sila) being fed barley in month V by an unmentioned mus'akilu. This indicates that we are not dealing with a real breeding herd but with animals being prepared for slaughter, especially as none of the animals seems to be female. The term G U ~ . A Bin this case stresses that the animals are not castrated, not that they are intended for breeding. Cf. also Talon 44. Talon 8 represents the feeding of barley to a breeding herd of donkeys. The ANSE rakibi of line 3 therefore are possibly not primarily "riding donkeys," but are to serve the donkey mares, i.e. they are the riders, not the ridden. The importance of over-land transport is underlined by the mentioning of pack animals receiving fodder: Talon 27,21. Six oxen of the majaltu, "wagon," of Jasmab-Adad are mentioned in Talon 65 and 74. Transport using oxen probably also forms the background of Talon 45. The 210 pigs being fed barley recorded in Talon 54 constitute perhaps the biggest herd of these animals mentioned by a cuneiform text, and this text does not stand alone. Another small indication that the Chagar Bazar organization employed plough teams can be found in Talon 10 where people classified as engar's are mentioned in connection with lard. The phrasing suggests that it is being returned. Such lard is probably not to be regarded as part of the rations, but it is issued in connection with the ploughing, as parallel texts from the South suggest. 86 i.e. as in the South what is called "seed" may, in the North, include "fodder."
87 The text is discussed by R. M. Whiting, 193ff. in S. Eichler et al. eds., Tall al-Hamidiyah 2, Freiburg and Gottingen 1990. The sowing of the summer product sesame mentioned in text no. 280 also implies irrigation.
292
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
acts as a co-user of the land as as a real conqueror or settler would be of some importance for the role of people on the eastern flank of Mesopotamia. The same is probably the case with the Guti, the Lullu and other inhabitants of the Eastern mountains. They penetrate into Mesopotamia and establish themselves, they were obviously ethnically recognisable. That there is a palatial organization in $emSara is self-evident, some of its agricultural base is documented in J. Eidem's Archive 2.89
latter fact is important.93 In VS 7 204 we have a document in which a king grants land, over 18 ha, to a "~ervant,"9~again a sign of persisting administrative habits. Terqa is a small town, with a surface of no more than 10 ha as estimated by its present excavator, but in this not very strong political entity the manner of land-holding seems to continue in much the same way as it was in the Mari period. Alalab VII, though not Mesopotamian, is of considerable interest for further developments. It should be stressed that in the Level-VII period we are primarily dealing not with a fully-fledged local kingdom, but with the estate of a younger branch of the ruling family of Jambad, which has a local basis founded on land holding. Complete settlements are transferred and bartered.95 That probably means that what is exchanged and bartered are rights over land and inhabitants, who thereby cease to be the direct dependants of the central administration. This type of society is ripe for beheading, it can be taken over with little trouble by a new aristocracy based on new military technology. There are signs that this technology is developing.
Outside the sphere of interest of Sam~i- dad's kingdom we have in the North only the Old-Babylonian documents excavated in Alalah which was subject to Aleppo in the period, and therefore come from the same general ambience.90
c. From the Old-Babylonianperiod to the period of Neo-Assyrian domination The Old Babylonian material from the North primarily documents the extent of
SamSi- dad's kingdom and its administration by a network of local palaces, which each had a local agricultural basis.91 The fundamentally dimorphic structure of the area of dryfarming is, however, best illustrated by the Mari archival material from the time of ZimriLim, where also the frequently weak basis, seen from a quantitative viewpoint, of the local "palaces" transpires. It is not possible with the available evidence to provide an idea about the size of sedentary and non-sedentary herds which used the land together with the institutional herds and those of other arable farmers. Institutional land requirements are limited, especially because man power is limited. There is ample space for grazing. For the later Old-Babylonian period and its aftermath we are restricted to the private texts from TerqaJSirqu92 which are supplemented up to a certain degree by, in this case, the (later) material from Alalab W. There is just enough to suggest that along the Habur and near TerqaISirqu a system of land holding which combines (small) palaces and private owners remains en vigeur, and that an irrigation system continues to function. The
*9 J. Eidem, The Shemshara Archives 2, The Administrative Texts. Copenhagen 1992. But cf. also and especially text 137 from Archive 1, which no longer, like the text from Chagar Bazar quoted by J. Laessoe, The Shemshara Tablets, Copenhagen 1959 p. 68-70, contains a reference to irrigation machines: the majaltu is a wagon. 90 Cf. the remarks by D. Charpin, MAR1 8 (1997) 793 (ad p. 44 of H. Klengel's Syria 3000-300, Berlin 1992.) That the Assyrian Kinglist arrogates, so to say, Sam~i- dad and family does not implicate that he was an Assyrian king, Durand and Charpin are unquestionably right in this, and that the family was distinctly regarded as non-Assyrian by its contemporaries is well-documented. In the Puzur-Sin inscription, which is one of the most explicit utterances of ethnicity originating from ancient Mesopotamia the dislike is virulent. But it remains a fact that the town of Assur, though politically not very important, or its main god, though hardly ever mentioned, or the ramifications of its commercial organization, somehow exerts a supra-regional influence in that dating by limu's occurs in places where there are no signs of Assyrian political domination: Qattara, Sehna, and perhaps most important of all, the much later text from the Habur area published by H. M. Kiimmel and C. Wilcke, ZA 79 (1989) 191f, which, incidentally, again shows a functioning irrigation system in the Habur area. A shadow of a continuation of a political Mari-tradition is visible in the texts from Tell Bderi from the time of Tiglatpileser I, cf. S. M. Maul, "Die Inschriften von Tall Bderi," Berlin 1992. 9 2 See note 44.
293
Nuzi and the Middle Assyrian material receive special attention in this collection and should be skipped here. Without suggesting that we are dealing with something completely new, the importance of the period in general is that it documents a type of land holding that has military-aristocratic overtones. The tragedy of the Nuzi material is that for the exceptionally well-documented 14th century there exists no archaeological equivalent which could act as a check on the textbased map of the kingdom of Arrapba. There are only a few known fixed points, but, even there, what is the seize of the settlements? Having an archaeological map not only means a better insight in the relative size of settlements, but also that the distances between the actual settlements, and thereby the amount of land available, becomes visible. Though it is probably unlikely that an archaeological survey would be able to find all dimtu's in the always restricted time available, such a survey is essential for obtaining an idea about the actual space available. In Nuzi there is ample evidence for herding,96 but it is herding controlled by the settled population, while the herdsmen lead a transhumance-like existence. There is no evidence about a non-sedentary element. We find many examples of ethnic instability, especially of people classified as bapiru, originating in adjacent countries settling under the protection of local magnates. But these people cannot be classed as a non-sedentary 93
The year name of BRM 4 52 indicates that the upkeep of canals continued. Several of the texts are sealed with the royal seal presumably as a guaranty for the validity of the document, even though the king is no party. Land bordering on that of "the palace" is mentioned in VS 7 204, RBC 779 (A. H. Podany, G. M. Beckman and G. Colbow, JCS 43-5 [1991-931 390; RA 41 (1947) 42f; TCL 1 237, 238. In RA 34 (1937) 184f the palace of Saggaratum is mentioned. (Land in Qattunan, on the Habur, occurs in the text published by J. Nougayrol in Syria 37 [I9601 205f) Continuing local royal involvement with canal building in the area further down the Euphrates is documented by the inscription found in Babylon published by F. H. Weissbach, Babylonische Miscellen no. 4, which incidentally also shows that such canals could be used by shipping. 94 Confirmation of the grant by an oath which invokes the gods and the person of the king himself, presumably by the king, is strange, but it also occurs in TCL 1 237. 95 E.g. B. Kienast, "Die altbabylonischen Kaufurkunden aus Alala," WdO 11 (1980) 35-63. 96 M. A. Morrison, "Evidence for Herdsmen and Animal Husbandry in the Nuzi Documents," SCCNH 1 (1981) 261.
294
G. VAN DRIEL
element. They include Assyrians and Akkadians, i.e. Babylonians and there are skilled craftsmen among them. The dominating Hurrian element of the population also contains a Kassite component. Land-holding in Nuzi is a much discussed subject, especially since the sale of land is generally masked by pseudo-adoption and mortgage. The question is complicated by ilku obligations which clearly derive from the holding of land. Whether this type of obligation was general or existed only in certain parts of the country is difficult to assess, as there seem to be local scribal practises, which always create the possibility that clauses which describe circumstances which obtain generally, are left out. We know that in the later Third Millennium the area had an Akkadian population, at least that it was administered in Akkadian and had Akkadian personal names. We can assume a subsequent "Landnahme" in which settlers were assigned land with attending military obligations.97 The result is on the one hand a palace, the royal administration, which consists of a hierarchy of larger and smaller local palaces on the one hand, and on the other of a rigorously stratified population, consisting of holders of E Sa riikib narkabti, B Sa nakkuSSi, E Sa d i k ilki and E Sa aSSabe98. Status and obligations are obviously related, and the obligations have military implications as far as we do understand them. The system was under economic -and military- pressure in the documented period and it will have existed earlier than the first of the documented five (Tehip-Tilla) generations. The passing of time in such a system will mean a mixed process of concentration and fragmentation which blurs outlines. Here, for comparative purposes only, some basic facts. Irrigation was practised where possible, the facts about the canals have been collected recently by G. G. W. Miiller.99 The use of the four-ox team, occasionally of double or half-strength is demonstrated by HSS 16 427, A ~ 437, S therefore from a household with palatial connections. The seeding density practised and questions relating to the yield, studied by C. Zaccagnini, who duly noticed the various weaknesses of the evidence,loOhave already been discussed.
97
Subjection to the Mitanni overlord may have formalised a military type of land holding, the king of Arrapba will have been obliged to provide a contingent of fixed size for fixed amounts of time, if the system operated in the manner such systems tend to work. 98 This is best illustrated by Gadd 63, RA 23 (1926) 158. 99 G. G. W. Miiller, HSAO 7, 204f, cf. 228f K. Grosz, JNES 42 (1983) 307 has pronounced herself against irrigation in Nuzi, but I follow C. Zaccagnini, BSA 5 (1990) without hesitation. R. Jas pointed out to me two clear texts, both from the Nuzi palace: W. Mayer, Nuzi Studien I, AOAT 205, nos. 101 and 151. loo C. Zaccagnini, "The Yield of Fields at Nuzi," OA 14 (1975) 181-225, cf. BSA 5 (1990) 201-217. The calculated yields in the tables on p. 187 and 190 vary considerably, and attempts to arrive at a "notional" yield by using evidence from the pseudo-adoption texts, and the (equally pseudo-)pledges involving land do not result in a clear outcome. G. G. W. Miiller has put all available evidence into one table (with additions) illustrating local variations, yield being dependent on on local circumstances: HSAO 7, Table 1, 229-232, but he rightly points out that documents from one locality often were written down on the same occasion, which councils caution (p. 232). Only the fines imposed for withholding land and yields provide a firm relationship: for 1 imeru landlyear 10 imeru of cereals is imposed. This is "notional," but definitely not for general use.
THE MESOPOTAMIAN NORTH: LAND USE
295
d. AssyrialOl: the structure of land use102 Direct exploitation by the palace organization is not really documented, but it is implied by the programmatic palace building and plough constructing referred to above. Land use in Assyria is primarily documented by contracts and grants, and by the remnants of registers of ownership or land holding, of which the Harran census is the best example, whatever its background. As it happens, the Harran region is also an area, beyond the upper reaches of the Balib, where we may expect to find traces of pastoralist activities. Contracts dealing with land document the importance of, especially military, servants of the crown and of a group of high officials. Assyria is not the monolithic state dominated exclusively by the person of the king which the royal inscriptions suggest, it is a land where the kings have to be careful about their power base, must create and maintain a group of landowners on which they can depend. A text like ADD 675 (SAA 11 221) suggests the ease in which estates are shuffled around in court circles. Estates are spread over the whole area dominated by the king. It is a society which can be beheaded easily: its big cities are a creation of a predatory system and equally easy its land owners can be replaced by a new group or loose their reason of existence by the disintegration of the political system. There is little sense in a long discussion of the matter. The documentation, as far as from Nineveh deriving from the palace files and kept there for very specific purposes, documents land holding by those dependent on the king. These persons are larger and smaller absentee landowners, who sometimes obtain the means to acquire their scattered estates from their royal master. Whether this is really representative of large scale Assyrian land holding is not easily answered. An answer to the question whether this land constituted the donation of the high official, or whether the high official held his rank because he was a land owner, greatly depends on the views held regarding the basic structure of Assyrian society. If one reads the texts belonging to the "@rran census," which is easily done in the SAA 11 re-edition,l03 the text also provides some insight in actual land use. There are many orchards and vineyards and even plantations of "timber," It is not a barren landscape with barley monoculture. But the information on the arable is more interesting for us. The land belongs to estates of larger and smaller owners, or perhaps rather "landholders." These owners employ -possess, is probably a better word as in this period it is clear that at least part of the rural population was tied to the soil and sold with it- on specified tracts of land a specific farmer and his family.104 In these texts the amount of land is mentioned with specification of what is actually being tilled.lo5 In a number of cases the amount under the plough is half the land For the Middle Assyrian period cf. the contribution by F. A. M. Wiggermann. l o 2 In general, from a fairly prolific literature: J. N. Postgate, "Some remarks on conditions in the
Assyrian countryside," JESHO 17 (1974) 225-243; "Grundeigentum und Nutzung von Land in Assyrien im 1. Jt. v. u. Z.", 89-110 in B. Brentjes ed., Grundeigentum in Mesopotamien (Jahrbuch fur Wirtschaftsgeschichte 1987lSonderband) Berlin 1988; F. M. Fales, "The Neo-Assyrian Period," 207-220 in A. Archi ed., Circulation of Goods in non-palatial context in the ancient Near-East, Rome 1984. l o 3 Mention should be made of F. M. Fales, Censimenti e catasti di epoca Neo-Assira, Rome 1973. l o 4 On these families cf. the most stimulating article by M. T. Roth, "Age at Marriage and the Household: a study of Neo-Babylonian and Neo-Assyrian Forms," CSSH 29 (1987) 715-747. lo5 Infrequently a number of oxen is mentioned, e.g. ADB 3 (SAA 11 203) I1 6 one, Rev. I 8' two, I11 7' six, but not generally. This should not be regarded as proof that the plough was not in general use. It
296
THE MESOPOTAMIAN NORTH: LAND USE
G. VAN DRIEL
available, a possible suggestion of biennial fallow.lO6 But what is actually under the plough is often less.107 Very suggestive in this respect is ADB 1, SAA 11 201, where the summary (m 34-5) indicates that out of 1250 emaru's of arable only 280 are sown. It is not only that fallow is practised, but also that the means to cultivate are not available. The amounts both available and sown vary in size per plough. The unavoidable conclusion seems to be that a plough team in the Harran area tills much less than a late Third Millennium Southern team: 10 to maximally (once!)l5 ha, against 45 ha. This is becoming something of a pattern: Mari, Sabi Abyad, Harran. It may be deceptive, however. The amounts could suggest a very different arable practice. The 45 ha in the South represents a very wide spacing of the furrows, with 75 to at least 50 cm between each pair. What really counts is the actual distance covered in ploughing. If furrows are less widely spaced in the North108 the actual difference is reduced.109 There is little sense in summing up all similar evidence in the complete set of texts. Much land, that is land having known owners, is unused.110 ADB 5 (SAA 11 219), which does not belong to the Harran census proper, probably represents an attempt to settle deportees on under-used land belonging to various estates. Little can be said on the actual practise of Assyrian agriculture. The use of the plough is universal, but whether this was the seeder plough is not strictly clear.111 Otherwise we can only refer to what has been said in the section on seeding rates. After the fall of the Assyrian monarchy our (written) documentation becomes almost non-existent. The big Assyrian capitals will have lost their raison d'e^tre, which did not automatically imply disappearance, but rather meant reversion to former localised importance. The exact division of the spoils between the Medians and the Chaldeans is not known, but at least the Euphrates area came temporarily into Chaldean hands,ll2 and we find some evidence of participation of Southern institutions in the herding of sheep in both the Euphrates and the Tigris area, no doubt a direct result of the prevailing pax chaldaeica or achaemenidica.l13 Somewhat later in the Persian period, the Elephantine archive indicates the existence of large noble estates in the general area.
could mean one of several things, possibly that the farmer himself owned the oxen mentioned, or perhaps more likely, that a standard complement (possibly 4 ) is available but only deviations are registered. The fact that farmers possessed land and animals of their own is mentioned occasionally e.g. in ADB 2 (SAA 11 202) . lo6 ADB 3 (SAA 11 203) 12: 7 emaru out of 15, I1 4: 15 out of 30, IV 13: 10 out of 20. lo7 ADB 3 (SAA 11 203) IV 4: 10 out of 24, Rev. 14': 10 out of 40, 111 11': (a subtotal) 10 out of 140. lo8 Cf. preceding notes. During the meeting S. Bottema pointed out that the distance between the furrows cannot be reduced beyond a certain limit when oxen are used for ploughing: the weight of the ox would cause the seed to be trodden in too deeply, more than 5 cm being decidedly undesirable. O9 This might also explain the heavier sowing which seems to be suggested occasionally. Though the texts mention the presence of water (ADB 2, SAA 11 202 I 7 , a gubbu), there is for the Harran area no explicit reference to irrigation. Unmistakably a seeder plough is the one depicted on the glazed brick panel in Khorsabad: V . Place, Ninive et Z'Assyrie 111,Paris 1867, P1. 31. Cf. also K. Radner in this volume. The best new evidence are the dates on the Assyrian-style NB documents from Seib Hamad, J. N. Postgate, SAAB 7 (1993) 109-124. Cf. BSA 7 (1993) 225-6.
6.
297
CONCLUSION
In general the question to be answered is whether higher seeding rates and different, i.e. in general tending to be lower, yields per surface unit in comparison to the South, are compensated in the North by tilling larger surfaces or result in a less developed system as to population, urbanisation or political muscle. There is sufficient evidence to propose that in the North institutional agriculture was based on a plough team organization, but the actual strength of these teams is not too clear. There seems to emerge a pattern indicating that the amount of land per team was less, perhaps only a third, of what was the Ur III norm in the South. Seeding rates could in practise be slightly higher than in the South, but they were apparently much lighter than nowadays. The evidence is, however, minimal. The actual resilience of the underlying system may have been great, with a reduced superstructure in comparison to the South. As to yields Jas (this volume) quotes the Hiitteroth table of exceptionally good to particularly bad harvests, based on evidence gathered by Christiansen-Weniger for the years 1928-1961 in Turkey. This table is not concerned with evidence on seeding practise and obviously deals with "large scale" generalisations, but the evidence is, to say the least, suggestive when compared to known results in antiquity. A "good harvest" of 1000 kgha, 1500 litres, equals 1785 sildha or 11600 per bur, that is 38.3.2 kor from dry farming, if we forget about seeding rates, compares well with the bureaucratically expected 30 (or exceptionally 40) kor per bur in irrigated circumstances. A moderate harvest, 700 kgha means 27 kor and the "catastrophic" 400 kglha is still a good 15 kor per bur, not reached by many a field in many a year in irrigated Southern Mesopotamia. I find it very difficult to compare these "modem" results with Ancient Mesopotamian outcomes. It is even more difficult to incorporate this type of evidence in model building. On p. 503 of CA 3515 Wilkinson works with a "base yield" of 300 kgha (less 50 kglha for seed.) and the "average yields for Syrian areas" on p. 497 should also be mentioned. Is this really pre-fertiliser (or -pre-what)? All discussion on ancient yields suffers from the fact that the Northern evidence relates always to one particular moment, not to a range of years. Neither comparison with modern results nor the evidence from the ancient texts really inform us about "average" productivity. That makes it also illusory to attempt to figure out what the net advantages were of the use of the institutional plough team over and above small scale subsistence farming. Yet the plough team was a persistent phenomenon. Points for further discussion could be the fundamental difference between North and South suggested by the next to invisible role of "temples" in the North.l14 The North , especially in the upper Habur area, will have looked in the later Assyrian period very different from what it was in the Mari period. In the North we seem to find, that is, perhaps as yet, more a suggestion than an established fact, at first a land administered by a hierarchy of palaces, either directly dominated by a centralising monarch, if he could grasp the opportunity, or indirectly manipulated through vassals of various degrees of (economic) independence, or otherwise, in a disintegrated state in the hands of a mob of The fact that "temples" are not important in questions regarding the exploitation of land is stressed for Mari, the main source of our knowledge, by J.-M. Durand, LAP0 17, 524-525.
298
299
G. VAN DRIEL
THE MESOPOTAMIAN NORTH: LAND USE
small local potentates. From such a small king to a local aristocrat is only a step. Later on the open country with few natural barriers favours a military organization based on the chariot, which requires in turn a certain amount of income, which, in its turn, in the main will have been derived from land.115 Towns had, when this military structure had come into being, presumably lost some of their raison d'etre: a bigger share of the surplus may have remained in the country side, but those who benefited had to keep in contact with the (fewer) big capitals. In the South the much larger scale of economic and political systems at an early point in time meant that the centralising monarchs needed an effective means of political domination at a local level, which left not too much power in the hands of a local representative and which yet provided an outlet for local, so to say, patriotic sentiment. Nominal concentration of potential in the hands of the local god could be regarded as an ideal solution. The economic role of the Southern temples thereby became a corollary of political concentration of supra-local economic potential in the hands of rulers striving after supremacy. Such a system could be regarded as more stable than the system we observe in Mari, with its small bi-polar states and rapid turn-over of kings, without economically important divine estates. But also in the South the degree of urbanisation declined after the Old Babylonian period.
Before that we are dealing with the herding of sheep and goats. We do not hear about non-sedentaries using the horse-drawn chariot. Did this feature give the upper layer of sedentary society an advantage during the third quarter of the second millennium, perhaps partially lost by the gradual penetration of the camel? The basically military structure of much of Northern land holding after the Mari period seems to be an essential feature of the system. But the permanent presence of a non-sedentary population at least in the Euphrates area must be regarded as a certainty.
The available documentation does not suggest that land was in short supply, neither in the South nor in the North. Together with the natural circumstances in general, this suggests that there were ample opportunities for a non-sedentary element and for the breeding of sheep by sedentaries. The importance of the production of wool is insufficiently documented in the written material from the North.l16 In its dimorphic guise the sedentary-non-sedentary continuum is well documented only in Mari, in a more antagonistic shape we find it probably in Assyria, especially in the period of Aramaic penetration. But only when Arabic tribes came within the arc of vision of the Assyrians, the economic basis of nonsedentary life, as seen from sedentary Mesopotamia, started to change.117 Cf. the mobilisation blanks from the North-West Palace at Nirnrud published by B. Parker, Iraq 23 (1961) ND 2431 etc., which presumably were addressed to a specific person on their envelope, and which are chariot based, comparable, probably, to the mediaeval "lance," based on a mounted warrior, also a unit of assessment. The strength of such a unit could be suggested by AfO Beih. 6 no. 48. l 6 Not discussed here is the importance for a sheep-breeding based society of the shift from a linenbased textile production to a wool-based one. That such a shift is of great importance for Northern Mesopotamian land use should be self-evident. The general theory is presented by J. McCorriston, "The Fiber Revolution. Textile Extensification, Alienation, and Social Stratification in Ancient Mesopotamia," CA 3814 (1997) 517-549. A tantalising preview of Northern archaeozoological material from what is suggested to be the relevant period is presented by M. Zeder, "Environment, Economy and Subsistence on the Threshold of Urban Emergence in Northern Mesopotamia," 55-67 in M. Fortin and 0. Aurenche eds., Espace naturel, espace habite' (Canadian Society for Mesopotamian Studies QuCbec, Bulletin 33) Qut5bec and Lyon 1998, 55-67, cf. the contribution by J. McCorriston, 43-53. The question of a (Northern) "second urbanisation" (especially Leilan), after an Uruk-period dead-end, is, perhaps, to be approached with some caution. Their present excavators regard both Mari and Chueira as (in Southern terminology) ED I creations on an urban scale, and this period is also in the South a high-water mark of big towns, cf. R. McC. Adams, Heartland of Cities, Chicago 1981, 93. Perhaps it is better to regard the North as an area of shifting big towns, dependent on a specific political constellation. l 7 This transpires from the material published by S. W. Cole, OIP 114, Chicago 1996.
Little enough can be said, when things come to the crunch, about the evidence concerning the actual practise of Northern agriculture as derived from texts. There is, as far as we can see for the moment only one certain reference to the seeder plough, but use of that type of plough must have been universal.l18The amount of land per plough seems to vary, but clearly tends to be less than in the South. That there are a number of indications that sowing was heavier than in the South is not remarkable, but the evidence is not very consistent, for occasionally the sowing rate is strikingly low. A heavier rate might have been expected on the real sustenance fields, with, as always, light seeding indicating a certain amount of speculation. The question of the intensity of the ploughing in relation to the area sown cannot be answered, information on weeding and similar activities, which could explain the high number of men in plough teams, is not available, as far as I can see. This ends on a suitably negative note. Others must provide the positive new evidence. Final version, Leiderdorp October 20, 1999
118 We must await confirmation for the two GIS.APIN Sa GILGIGIR of ARM 19 458:2 before postulating the early use of a charrue. (H. Limet proposes an unidentified part of a chariot, M. Lambert translates "2 charrues & chariot") For the moment a wheeled plough must be regarded as an anachronism.