THE ARCHAEOLOGY OF DRYLANDS The One World Archaeology (OWA) series stems from conferences organized by the World Archaeological Congress (WAC), an international non-profit making organization, which provides a forum of debate for anyone who is genuinely interested in or has a concern for the past. All editors and contributors to the OWA series waive any fees they might normally receive from a publisher. Instead all royalties from the series are received by the World Archaeological Congress Charitable Company to help the wider work of the World Archaeological Congress. The sale of OWA volumes provides the means for less advantaged colleagues to attend World Archaeological Congress conferences, thereby enabling them to contribute to the development of the academic debate surrounding the study of the past. The World Archaeological Congress would like to take this opportunity to thank all editors and contributors for helping the development of world archaeology in this way.
ONE WORLD ARCHAEOLOGY Series Editor: (Volumes 1–37): Peter J.Ucko Academic Series Editors (Volume 38 onwards): Martin Hall and Julian Thomas Executive Series Editor (Volume 38 onwards): Peter Stone
1. What is an Animal?, T.Ingold (ed.) 2. The Walking Larder: Patterns of domestication, pastoralism and predation, J.CluttonBrock 3. Domination and Resistance, D.Miller, M.J.Rowlands and C.Tilley (eds) 4. State and Society: The emergence and development of social hierarchy and political centralization, J.Gledhill, B.Bender and M.T.Larsen (eds) 5. Who Needs the Past? Indigenous values and archaeology, R.Layton (ed.) 6. The Meaning of Things: Material culture and symbolic expression, I.Hodder (ed.) 7. Animals into Art, H.Morphy (ed.) 8. Conflict in the Archaeology of Living Traditions, R.Layton (ed.) 9. Archaeological Heritage Management in the Modern World, H.F.Cleere (ed.) 10. Archaeological Approaches to Cultural Identity, S.J.Shennan (ed.) 11. Centre and Periphery: Comparative studies in archaeology, T.C.Champion (ed.) 12. The Politics of the Past, P.Gathercole and D.Lowenthal (eds) 13. Foraging and Farming: The evolution of plant exploitation, D.R.Harris and G.C.Hillman (eds) 14. What’s New? A closer look at the process of innovation, S.E. van der Leeuw and R.Torrence (eds) 15. Hunters of the Recent Past, L.B.Davis and B.O.K.Reeves (eds) 16. Signifying Animals: Human meaning in the natural world, R.G.Willis (ed.)
17. The Excluded Past: Archaeology in education, P.G.Stone and R.MacKenzie (eds) 18. From the Baltic to the Black Sea: Studies in medieval archaeology, D.Austin and L.Alcock (eds) 19. The Origins of Human Behaviour, R.A.Foley (ed.) 20. The Archaeology of Africa: Food, metals and towns, T.Shaw, P.Sinclair, B.Andah and A.Okpoko (eds) 21. Archaeology and the Information Age: A global perspective, P.Reilly and S.Rahtz (eds) 22. Tropical Archaeobotany: Applications and developments, J.G.Hather (ed.) 23. Sacred, Sites, Sacred Places, D.L. Carmichael, J.Hubert, B.Reeves and A.Schanche (eds) 24. Social Construction of the Past: Representation as power, G.C.Bond and A.Gilliam (eds) 25. The Presented Past: Heritage, museums and education, P.G.Stone and B.L.Molyneaux (eds) 26. Time, Process and Structural Transformation in Archaeology, S.E.van der Leeuw and J.McGlade (eds) 27. Archaeology and Language I: Theoretical and methodological orientations, R.Blench and M.Spriggs (eds) 28. Early Human Behaviour in the Global Context, M.Petraglia and R.Korisettar (eds) 29. Archaeology and Language II: Archaeological data and linguistic hypotheses, R.Blench and M.Spriggs (eds) 30. Archaeology and Anthropology of Landscape: Shaping your landscape, P.J.Ucko and R.Layton (eds) 31. The Prehistory of Food: Appetites for Change, C.Gosden and J.G.Hather (eds) 32. Historical Archaeology: Back from the edge, P.P.A.Funari, M.Hall and S.Jones (eds) 33. Cultural Resource Management in Contemporary Society: Perspectives on managing and presenting the past, F.P. MacManamon and A.Hatton (eds)
34. Archaeology and Language III: Artefacts, languages and texts, R.Blench and M. Spriggs (eds) 35. Archaeology and Language IV: Language change and cultural transformation, R.Blench and M.Spriggs (eds) 36. The Constructed Past: Experimental archaeology, education and the public, P.G.Stone and P.Planel (eds) 37. Time and Archaeology, T.Murray (ed.) 38. The Archaeology of Difference: Negotiating cross-cultural engagements in Oceania, R.Torrence and A.Clarke (eds) 39. The Archaeology of Drylands: Living at the margin, G.Barker and D.Gilbertson (eds)
THE ARCHAEOLOGY OF DRYLANDS Living at the margin
Edited by Graeme Barker and David Gilbertson
London and New York
First published 2000 by Routledge 11 New Fetter Lane, London EC4P 4EE Simultaneously published in the USA and Canada by Routledge 29 West 35th Street, New York, NY 10001 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge's collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” © 2000 Selection and editorial matter, Graeme Barker and David Gilbertson; individual chapters, the contributors All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data The archaeology of drylands: living at the margin/[edited by] Graeme Barker and David Gilbertson. p. cm. (One world archaeology; 39) Includes bibliographical references and index. 1. Social archaeology. 2. Landscape archaeology. 3. Human ecology. 4. Deserts—History. 5. Land settlement —History. 6. Land settlement patterns, Prehistoric—History. 7. Arid regions agriculture—Social aspects—History. 8. Climatic changes—History. I. Barker, Graeme. II. Gilbertson, D.D. III. Series. CC72.4.A735 2000 930.1–dc21 00–038257 ISBN 0-203-16573-X Master e-book ISBN
ISBN 0-203-26029-5 (Adobe e-Reader Format) ISBN 0-415-23001-2 (Print Edition)
Contents List of figures List of tables List of contributors Series editors’ foreword Preface Part I Introduction 1 Living at the margin: themes in the archaeology of drylands Graeme Barker and David Gilbertson 2 The dynamic climatology of drylands Greg Spellman Part II Southwest and Central Asia 3 The decline of desert agriculture: a view from the classical period Negev Steven A.Rosen 4 Farmers, herders and miners in the Wadi Faynan, southern Jordan: a 10,000-year landscape archaeology Graeme Barker 5 Differing strategies for water supply and farming in the Syrian Black Desert Paul Newson 6 Irrigation agriculture in Central Asia: a long-term perspective from Turkmenistan Mark Nesbitt and Sarah O’Hara Part III Sahara and Sahel 7 Conquests and land degradation in the eastern Maghreb during classical antiquity and the Middle Ages Jean-Louis Ballais 8 Success, longevity, and failure of arid-land agriculture: Romano-Libyan floodwater farming in the Tripolitanian pre-desert David Gilbertson, Chris Hunt and Gavin Gillmore 9 Twelve thousand years of human adaptation in Fezzan (Libyan Sahara) David Mattingly 10 Farming and famine: subsistence strategies in Highland Ethiopia Ann Butler and A.Catherine D’Andrea
ix xiii xv xxv xxvii
3 18
44
62
85
101
121
133 156 174
Part IV Eastern and southern Africa 11 Engaruka: farming by irrigation in Maasailand, c.AD 1400–1700 John E.G.Sutton 12 The agricultural landscape of the Nyanga area of Zimbabwe Robert Soper 13 Fifteenth-century agropastoral responses to a disequilibrial ecosystem in southeastern Botswana John Kinahan 14 Islands of intensive agriculture in African drylands: towards an explanatory framework Mats Widgren Part V North and Central America 15 Prehistoric agriculture and anthropogenic ecology of the North American Southwest Paul E.Minnis 16 The role of maguey in the Mesoamerican tierra fría: ethnographic, historic and archaeological perspectives Jeffrey R.Parsons and J.Andrew Darling Part VI Europe 17 Traditional irrigation systems in dryland Switzerland Anne Jones and Darren Crook 18 Desertification, land degradation and land abandonment in the Rhône valley, France Sander van der Leeuw Index
195 214
227
246
264
280
307
327 346
Figures 1.1 1.2 1.3 1.4 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 3.1 3.2 3.3 3.4 3.5 3.6 3.7 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 5.1
The world map of drylands A Roman-period fortified farm, northwest Libya The location of the case studies in this volume Drowning in drylands?—two vehicles sunk in a flash-flood Thermal regimes in two dryland locations: Aswan, Egypt and Jacobabad, Pakistan Mean monthly relative humidity at four locations The rainshadow effect leading to aridity The Hadley Cell circulation of the tropical northern hemisphere The structure of the trade wind atmosphere The interaction between the subtropical westerly flow and the tropical easterlies leading to the creation of Saharan depressions The monthly progression of the East African Low-Level Jet Core The tracks of Sudano-Saharan depressions over the Sahara Terraced dam system in the central Negev The wine press at Shivta (Subeita) Sketch of the Byzantine town of Shivta (Esbeita or Subeita) Map of the general settlement system of the central Negev during the Late Byzantine and Early Islamic periods View of the Byzantine town of Avdat (looking north) Elaborate raised field and dam system on Nahal Lavan The early Islamic village of Sede Boqer in the central Negev The location of Wadi Faynan within its region Looking northeast across part of the ancient field system to Khirbet Faynan The survey area of the Wadi Faynan Landscape Survey Ethnoarchaeological survey: the typical site of a winter bedouin tent in Wadi Faynan The settlement locations of the first farmers in the Wadi Faynan Part of the Wadi Faynan field system WF4, showing the early bronze age and the classical landscapes A field system on the northern side of the Wadi Faynan A field map of part of the field system WF4 The distribution of copper through sediments accumulated behind the Khirbet Faynan barrage Section through a Roman-period water conduit channel The Hauran and the Harra regions of Syria
3 4 5 7 22 24 25 26 28 30 32 34 45 46 47 48 49 50 52 63 64 65 67 70 72 73 76 77 79
86 5.2 Plan of the Roman-period settlement of Ad-Diyatheh and its water channels and field systems 5.3 The Roman-period settlement of Ad-Diyatheh and its channel walls 5.4 Plan of the main irrigated zone along the Wadi Sham by the Romanperiod settlement of al-Namara 5.5 View of the main irrigated zone along the Wadi Sham at al-Namara 5.6 Canal 3 at al-Namara, viewed from the east 5.7 The ancient reservoir at Qasr Burqu’ 5.8 Air photograph of Qasr Burqu’ and its reservoir 6.1 Turkmenistan, showing locations mentioned in Chapter 6 6.2 Bronze and iron age settlement in the Merv oasis, Turkmenistan 7.1 The eastern Maghreb, showing locations mentioned in Chapter 7 7.2 Flood deposits at Ksar Rhilane (Tunisia) 7.3 The modern Roman aqueduct crossing Wadi Bou Jbib, Carthage 7.4 Holocene terrace of Wadi Chéria-Mezeraa (Algeria) 7.5 Holocene terraces in the Wadi el Akarit (Tunisia) 8.1 Tripolitania, northwest Libya, showing the principal landforms and settlements, and the location of the UNESCO Libyan Valleys Survey 8.2 Romano—Libyan floodwater farming in the Wadi Gobbeen 8.3 Simplified distribution of early Romano—Libyan farms 8.4 A Romano—Libyan fortified farm (gasr) and its satellite buildings at Ghirza 8.5 Model of Romano—Libyan agriculture 8.6 Walls in the desert: wall systems in the Wadi Mimoun 9.1 Map showing the location of the Fezzan and the area of most detailed survey around Germa 9.2 The major climatic fluctuations of the Holocene in the Libyan Sahara 9.3 The settlement of Germa (ancient Garama), the capital of the Garamantes 9.4 Schematic cross-section of a foggara 9.5 Model of the neolithic landscape around Germa 9.6 Model of the evolved Garamantian landscape around Germa 9.7 Model of the medieval landscape around Germa 9.8 Model of hypothetical future direction of settlement and farming in Fezzan 10.1 Map of Ethiopia, showing Adi Ainawalid in Tigrai province 10.2 Residential compound near fields, Adi Ainawalid 10.3 Intercropped bread and durum wheats near Mai Kayeh, Tigrai 10.4 Harvesting grasspea by hand uprooting, Adi Ainawalid 10.5 First threshing of teff, Adi Ainawalid 10.6 Winnowing teff, Adi Ainawalid
88 89 92 93 93 95 95 103 107 122 124 126 127 128 134 134 136 144 146 147 157 159 161 163 164 167 168 169 175 176 179 181 182 183
10.7 11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 11.9 12.1 12.2 12.3 13.1 13.2 13.3 13.4 14.1 14.2 14.3 14.4 14.5 14.6 15.1 15.2 15.3 15.4 15.5 15.6 16.1 16.2 16.3 16.4 16.5 16.6 16.7 16.8 16.9
Grain storage jars, Adi Ainawalid The Rift Valley and Crater Highlands of northern Tanzania Engaruka and the Rift Valley escarpment The Engaruka escarpment from the east Engaruka south fields Grid of feeder furrows and levelled field plots The support for the ‘great northern’ canal The embanked causeway of the ‘great northern’ canal An angular cairn Sonjo: wooden house with thatch dome Location of the Nyanga area, Zimbabwe Terraced hillsides in the Nyanga lowlands Vertical aerial photograph of cultivation ridges The regional setting of Letsibogo, southeastern Botswana The distribution and linkage of Khami period sites at Letsibogo Plan and section of Letsibogo Site 125 Distribution of soil nutrient values at Letsibogo Site 125 Eastern and southern Africa, showing sites mentioned in Chapter 14 An intensively cultivated landscape at Kwermusl Preparing the field at Kwermusl Piles of manure from stalled cattle, Mama Issara An irrigation channel above Tot in Marakwet, Kenya An irrigation canal under repair above Chesoi, Marakwet The North American Southwest Prehistoric Hohokam communities and irrigation systems in the Phoenix basin of the Salt river Aerial photograph of prehistoric trincheras (checkdam) fields near Casas Grandes, Chihuahua, Mexico Gridded gardens of fields near Safford, Arizona A rock mulch field near Casas Grandes, Chihuahua, Mexico An excavated rock pile from the field shown in Figure 15.5 Middle America, showing the approximate extent of the tierra fría Field of cultivated magueys Castrating a mature maguey plant Spinning maguey fibre Pre-columbian spindle whorls used for spinning maguey fibre Use of modern iron scraper for extracting maguey fibre Examples of pre-columbian trapezoidal tabular basalt scrapers Experimental use of pre-columbian trapezoidal tabular basalt scraper A pre-columbian scraper plane
184 196 197 198 200 200 201 202 203 212 215 216 219 229 233 234 236 248 251 252 252 253 254 265 267 268 269 270 270 281 282 286 290 291 293 294 295 295
16.10 17.1 17.2 17.3 17.4 17.5 17.6 18.1 18.2 18.3 18.4 18.5 18.6 18.7
Modern iron scraper and pre-columbian obsidian scrapers The Valais canton, Switzerland, showing places mentioned in Chapter 17 Distribution of agricultural land in Vernamiège during the 1960s The Grand Bisse de Lens The irrigation sectors in Vernamiège Distribution of water during the first tour from the bisses of Vernamiège, 5th May-8th June 1964 A tessel, used by members of the consortage of the Grand Bisse de Lens The middle and lower Rhône valley, showing the progress of Roman colonization Settlement trends in the middle and lower Rhône valley, 50 BC-AD 600 The persistence of settlements in the middle and lower Rhône valley through different occupation periods GIS maps of the Haut Comtat The three levels of the investigation into modern-day urban—rural dynamics in southern France Relations between cities, individual communes and their contexts Differences in context occurring among towns of similar and/or different sizes in the Haut Comtat
296 308 309 313 317 318 320 329 331 333 335 338 339 340
Tables 2.1 2.2 2.3 2.4 2.5 2.6 6.1 6.2 7.1 8.1 10.1 10.2 13.1 13.2 16.1 17.1 17.2 17.3 18.1 18.2 18.3
The regional distribution of world drylands Rainfall regimes at selected dryland stations Estimates of the land area of arid lands Mean relative humidity at various isobaric levels in the Sahara and the Arabian peninsula Seven ITCZ zones The extent and severity of desertification Simplified chronological chart of prehistoric settlement in Turkmenistan Simplified chronological chart of settlement in the Merv oasis in the historic period Morphoclimatic evolution in the eastern Maghreb during the later Holocene Farm products of the Tripolitanian pre-desert, first to fifteenth centuries AD Crops cultivated at Adi Ainawalid, Tigrai, Ethiopia Crops no longer cultivated at Adi Ainawalid, Tigrai, Ethiopia Selected radiocarbon measurements from Letsibogo Faunal taxa from Letsibogo Site 125 The prehispanic chronology of central Mexico Approximate numbers of named irrigators using the Grand Bisse de Lens ‘aqueductis communi’ in 1457 Examples of tours, with the number of droits and sequence of irrigation hours Examples of bisse disputes Evolution of the ‘nature-culture’ debate over the last thirty years The different approaches of the historical and natural sciences to the reconstruction of the past The opposition between analytical and integrative approaches in research
19 19 20 23 29 35 112 113 122 141 178 186 231 237 283 314 316 321 341 342 342
Contributors JEAN-LOUIS BALLAIS is Professor of Physical Geography at the Université de Provence, Aix-en-Provence (France). His principal research interests focus on Holocene Mediterranean erosion (he co-directed the study of erosion and geosystems history in classical antiquity and the Middle Ages for the EU-funded Archaeomedes project described in Chapter 18), and present-day erosion, desertification and land degradation in the south of France and in the Maghreb. Relevant recent publications include: ‘Aeolian activity, desertification and the “Green Dam” in the Ziban range, Algeria’, in A.C.Millington and K.Pye (eds) Environmental Change in Drylands: Biogeographical and Geomorphological Perspectives: 177–98, Chichester, John Wiley and Sons 1994; ‘The south of France and Corsica’, in A.J.Conacher and M.Sala (eds) Land Degradation in Mediterranean Environments of the World: 29–39, Chichester, John Wiley and Sons 1998; and (with J.-C.Meffre) Le Plan de Dieu (NordVaucluse). Géoarchéologie et Histoire d’un Paysage Anthropisé, Etudes Vauclusiennes 15, 1996. (Institutional address: Institut de Geographic, ‘Université de Provence (Aix-Marseiile I), 29 Avenue Robert Schuman, 13621 Aix-en-Provence, France) GRAEME BARKER (BA, PhD University of Cambridge) taught prehistoric archaeology at the University of Sheffield (1972–84) and was then Director of the British School at Rome (1984–88), before taking up his appointment as Professor of Archaeology at the University of Leicester (UK), where he is currently Dean of the University Graduate School. His principal research interests have been in the archaeology of subsistence and agriculture, with a special focus first on archaeozoology but later in landscape archaeology. He has conducted fieldwork in Italy, Mozambique, and the former Yugoslavia, and has directed inter-disciplinary field projects in Italy, Libya and currently in Jordan. His publications include: Landscape and Society: Prehistoric Central Italy, London, Academic Press 1981; (with R.Hodges) Archaeology and Italian Society, Oxford, British Archaeological Reports 1981; Prehistoric Communities in Northern England. Sheffield, University of Sheffield 1981; Prehistoric Farming in Europe, Cambridge, Cambridge University Press 1985; (with C.S.Gamble) Beyond Domestication in Prehistoric Europe, London, Academic Press 1985; (with J.Lloyd) Roman Landscapes: Archaeological Survey in the Mediterranean Region, London, British School at Rome 1991; A Mediterranean Valley: Landscape Archaeology and Annales History in the Biferno Valley, London, Leicester University Press 1985 (two volumes); (with D.Gilbertson, B.Jones and D. Mattingly) Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume One: Synthesis, Volume Two: Gazetteer and Pottery, Paris, UNESCO 1996; (with T.Rasmussen) The Etruscans, Oxford, Blackwells 1998; The Companion Encyclopedia of Archaeology, London, Routledge 1999; and (General editor with D.Mattingly) Mediterranean Landscape Archaeology, Oxford, Oxbow 2000 (five volumes). He was elected a
Fellow of the British Academy in 1999. (Institutional address: the Graduate School, University of Leicester, Leicester LE1 7RH, UK) ANN BUTLER is an Honorary Lecturer at the Institute of Archaeology, University College London, with a BSc degree in Botany (London University), an MA in Archaeology (Manchester University) and a PhD in Archaeology (London University). Her research interests centre on legumes as a human resource in the temperate Old World, and include ancient diet and nutrition, plant domestication and crop dispersals, traditional agriculture and sustainable farming. Her fieldwork has been conducted in Europe, Southwest Asia and Highland Ethiopia. Her current research focuses on the evidence for the domestication and exploitation of legume crops. Her recent publications include: ‘Pulse agronomy: traditional systems and implications for early cultivation’, in P.C.Anderson (ed.) Préhistoire de l’Agriculture: 67–78, Paris, CNRS 1998 and ‘Traditional seed cropping systems in the temperate Old World: models for antiquity’, in C.Gosden and J.Hather (eds) The Prehistory of Food: 473–77, London, Routledge 1999. (Institutional address: Institute of Archaeology, University College London, 31–34 Gordon Square, London WC1H OPY, UK) A. CATHERINE D’ANDREA is an Associate Professor of Archaeology at Simon Fraser University, British Columbia, Canada. She completed a BSc in Anthropology at the University of Toronto, an MSc in Bioarchaeology at University College London and a PhD in Anthropology at the University of Toronto. Her research interests include palaeoethnobotany, ethnoarchaeology and early agrarian societies in Africa and the Far East. She is currently conducting ethnoarchaeological and palaeoethnobotanical research in northern Ethiopia, as well as collaborating on an excavation in northern Ghana. Her recent publications include: ‘The dispersal of domesticated plants into northeastern Japan’, in C.Gosden and J.Hather (eds) The Prehistory of Food: 163–83, London, Routledge 1999 and (with D.E.Lyons, Mitiku Haile and E.A. Butler) ‘Ethnoarchaeological approaches to the study of prehistoric agriculture in the Ethiopian Highlands’, in M.van der Veen (ed.) The Exploitation of Plant Resources in Ancient Africa: 101–22, New York, Plenum Publishing Corporation 1999. (Institutional address: Department of Archaeology, Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6) DARREN CROOK is currently a Postdoctoral Fellow in the Department of Geography, University of Liverpool, working on historical impacts of land use and climate on hydrology in a pre-alpine landscape, funded by the Leverhulme Foundation. He graduated with a BSc in Human Ecology from the University of Huddersfield. His PhD, also from the University of Huddersfield, dealt with the sustainability of the bisse mountain irrigation system in the Valais, Switzerland. Publications from this include: (with A.M. Jones) ‘Traditional irrigation and its importance to the tourist landscape of Valais, Switzerland’, Landscape Research 24 (19) 1999:49–65 and (with A.M. Jones) ‘Traditional water management in a developed world context: an example from the Valais, Switzerland’, Mountain Research and Development 19 (2) 1999:79–99. (Institutional address: Department of Geography, University of Liverpool, Roxby Building, Liverpool L69 3BX, UK) J.ANDREW DARLING completed his PhD in Anthropology at the University of Michigan in 1998 and is currently an archaeologist for the Gila River Indian
Community and Board Member for the Mexico-North Research Network, Cd. Chihuahua, Mexico. He has conducted fieldwork in Zacatecas, Mexico (1988-present), on the north coast of Peru (1989–90), in southeastern Hungary (1987) and the North American Southwest (1984–1987). His research interests include compositional studies, exchange, regional interaction and ritual in prehistoric and ethnographic complex societies, including parallel archival investigations on the development of American archaeology in Mexico during the early twentieth century. Significant publications include: ‘Anasazi mass inhumation and the execution of witches in the American Southwest’, American Anthropologist 100, 1993:1–21; (with M.Glascock) ‘Acquisition and distribution of obsidian in the north-central frontier of Mesoamerica’, in E.C.Rattray (ed.) Rutas de Intercambio en Mesoamerica, III Coloquio Pedro Posch Gimpera: 345–64, Mexico, D.F., Universidad Nacional Autonoma de Mexico 1998; ‘Trace element analysis of the Huitzila and La Lobera obsidian sources in the southern Sierra Madre Occidential, Mexico’, Journal of Radioanalytical and Nuclear Chemistry Articles 196 (2), 1995:243–52; and ‘Notes on obsidian sources of the southern Sierra Madre Occidental’, Ancient Mesoamerica 4, 1993:245–53. (Institutional address: Mexico-North Research Network, 16 de Septiembre #402, Cd. de Chihuahua, Chihuahua, Mexico CP 31020) DAVID GILBERTSON is Head of the School of Conservative Sciences at the University of Bournemouth (UK) and Distinguished Visiting Scholar in the Department of Geography and Environmental Studies at the University of Adelaide (Australia). He graduated in Environmental Sciences at the University of Lancaster and gained his PhD and DSc in Quaternary and Archaeological Geology from the University of Bristol. Previous appointments have included: a Senior Fulbright Scholar at the University of Arizona; the Directorship of the MSc Programme in Environmental Archaeology, and then Head of the Research School in Archaeology and Archaeological Science at the University of Sheffield, where he became Reader then Professor; Director of the Institute of Geography and Earth Science at the University of Wales, Aberystwyth; and Professorial Research Fellow, University College Northampton. In addition to dryland environments, past and present, his research interests include coastal geomorphology, environmental change and caves, and environmental archaeology in general. His principal publications include: (with R.D.S.Jenkinson) In the Shadow of Extinction: The Quaternary Geology and Palaeoecology of the Lake, Fissures, and Smaller Caves at Creswell Crags, Sheffield, University of Sheffield Monographs in Prehistory 1984; (with D.J.Briggs, and G.R.Coope) The Chronology and Environmental Framework of Early Man in the Upper Thames: A New Model, Oxford, British Archaeological Reports 1985; Run-Off Farming in Rural Arid Lands, Applied Geography Theme Volume 6 (1) 1986; (with G.Barker, B.Jones, and D.Mattingly) Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume One: Synthesis, Volume Two: Gazetteer and Pottery, Paris, UNESCO 1996; and (with M.Kent and J.P.Grattan) The Outer Hebrides: The Last 14,000 Years, Sheffield, Sheffield Academic Press 1996. (Institutional addresses: School of Conservation Sciences, University of Bournemouth, Bournemouth BH12 5BB, UK, and Department of Geography and Environmental Studies, University of Adelaide, South Australia 5005)
GAVIN GILLMORE is Senior Lecturer in Earth Science at University College Northampton. His research interests include the application of fossil studies to palaeoenvironmental analysis, ecotoxicology of cave environments, and microfossil and stratigraphic studies of Quaternary sedimentary basins. He has worked extensively for oil exploration companies, producing many consultancy reports on Jurassic, Cretaceous and Tertiary-Quaternary microfossil assemblages. Some current papers include: (with M.Sperrin, P.Phillips and A.Denman) ‘Radon hazards, geology, and exposure of cave users: a case study and some theoretical perspectives’, Ecotoxicology and Environmental Safety 2000; (with K.A.Smith, and S.Sinclair) ‘Palaeoenvironmental and biostratigraphical significance of Ostracoda from the Milton Formation (Quaternary), Northamptonshire, UK’, Proceedings of the Geologists Association 2000; and (with T.Kjennerud and R.Kyrkjebø) ‘The reconstruction and analysis of palaeowater depths: a new approach and test of micropalaeontological approaches in the post-rift (Cretaceous to Quaternary) interval of the Northern North Sea’, in O.J.Martinsen and T.Dreyer (eds) Sedimentary Environments Offshore Norway—Palaeozoic to Recent, Oslo, Norwegian Petroleum Society Special Publication 2000. (Institutional address: School of Environmental Science, University College Northampton, Northampton NN2 7AL, UK) CHRIS HUNT is currently a Senior Lecturer in the Division of Geographical Sciences at the University of Huddersfield. His academic education led him from a degree in Geogaphy/Geology and an MSc in Micropalaeontology (Palynology) at the University of Sheffield to a PhD at Aberystwyth (Wales) on the Pleistocene history of an area in Somerset. He has published extensively in the area of environmental archaeology in Britain, Europe, North Africa and the Middle East. Recent publications include: (with G.Barker, D.D.Gilbertson, and D.Mattingly) ‘Romano-Libyan agriculture: integrated models’, in G.Barker, D.Gilbertson, B.Jones and D.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Survey. Volume One: Synthesis: 265–90, Paris, UNESCO 1996; (with S.Campbell, J.Scourse and D.H.Keen) The Quaternary of South West England, Chichester, Chapman & Hall, Geological Conservation Review Series 14, 1998; and (with D.D.Gilbertson) ‘Context and impacts of ancient catchment management in Mediterranean countries: implications for sustainable resource use’, in D.Wheater and C.Kirby (eds) Hydrology in a Changing Environment, Volume II: 473– 83, Chichester, John Wiley and Sons 1998. (Institutional address: Division of Geographical Sciences, University of Huddersfield, Queensgate, Huddersfield, HD1 3DH, UK) ANNE JONES is currently a Senior Lecturer in the Division of Geographical Sciences at the University of Huddersfield, and Head of Division. She graduated with a BSc in Geography from Queen Mary College (University of London) and then obtained her DPhil thesis at the University of Oxford studying immigrant communities in Marseille, France. Subsequently, she has held posts at the Open University, Liverpool University and Cambridgeshire College of Arts and Technology (now Anglia Polytechnic University). Her research interests focus around the inter-relationship between demography and the allocation of scarce resources in marginal environments in alpine Europe, the Mediterranean and Africa. Her principal publications include: ‘Exploiting a marginal European environment: population control and resource management under
the Ancien Regime’, Journal of Family History 16 (4) 1991:363–79; (with C.O.Hunt) ‘Walls, wells and water supply: aspects of the cultural landscape of Gozo, Maltese Islands’, Landscape Issues 11 (1) 1994:24–9; and (with C.O.Hunt and D.S.Crook) ‘Traditional irrigation strategies and their implications for sustainable livelihoods in semi-arid areas: examples from Switzerland and the Maltese islands’, in H.Wheater and C. Kirby (eds) Hydrology in a Changing Environment, Volume II: 485–94, Chichester, John Wiley and Sons 1998. (Institutional address: Division of Geographical Sciences, University of Huddersfield, Queensgate, Huddersfield HD1 3DH, UK) JOHN KINAHAN (PhD 1989, Witwatersrand) is an independent consultant in archaeological and palaeoenvironmental studies, and has authored more than forty scientific papers in diverse fields. Recent publications include: Pastoral Nomads of the Central Namib Desert: The People History Forgot, Windhoek, New Namibia Books 1991; ‘The rise and fall of nomadic pastoralism in the central Namib desert’, in T.Shaw, P.Sinclair, B.Andah and A.Okpoko (eds) The Archaeology of Africa: Food, Metals and Towns: 372–85, Routledge, One World Archaeology 20 1993; and ‘A new archaeological perspective on nomadic pastoralist expansion in south-western Africa’, in J.E.G.Sutton (ed.) The Growth of Farming Communities in Africa from the Equator Southwards: 211–26, Nairobi, British Institute in Eastern Africa 1996. He is currently attached to the University of Uppsala in Sweden while carrying out research on the long-term environmental impacts of nomadic pastoralism in Namibia and Tanzania. (Institutional address: Quaternary Research Services, PO Box 22407, Windhoek, Namibia, and Department of Archaeology and Ancient History, St Erikstorg 5, University of Uppsala, Uppsala S75310, Sweden) SANDER VAN DER LEEUW, presently Professor in the History and Archaeology of Techniques at the Sorbonne in Paris, followed a university education in History and Archaeology at the University of Amsterdam (PhD 1976), and has taught at the Universities of Leiden and Amsterdam in the Netherlands, Reading and Cambridge in the UK, Michigan (Ann Arbor), Massachussetts (Amherst) and the Santa Fe Institute in the USA, and the Australian National University. He has undertaken fieldwork in Syria, Holland, the Philippines and France. His main research interests embrace the technology of ancient pottery making, regional and spatial archaeology, the relations between people and their environment through time as well as in the present, and the use of dynamical systems modelling for understanding social systems. Among his publications are: Studies in the Technology of Ancient Pottery, Amsterdam, University of Amsterdam, Institute for Pre- and Protohistory 1976; (with A.C.Pritchard) The Many Dimensions of Pottery, Amsterdam, University of Amsterdam, Institute for Preand Protohistory 1982; (with R.Torrence) What’s New? A Closer Look at the Process of Innovation, London, Unwin Hyman, One World Archaeology 14 1987; (with J.McGlade) Time, Process and Structured Transformation in Archaeology, London, Routledge 1997; and The ARCHAEOMEDES Project: Understanding the Natural and Anthropogenic Causes of Land Degradation and Desertification in the Mediterranean, Luxemburg, Publications Office of the European Union 1998. (Institutional address: Boit 05, Maison de l’Archéologie et de l’Etnologie, 21 Alice de l’Université, 92023 Nanterre, France)
DAVID MATTINGLY is Professor of Roman Archaeology in the School of Archaeological Studies at the University of Leicester and has conducted fieldwork in North Africa, the Near East, the Mediterranean and Britain. He took his BA and PhD at the University of Manchester, the latter a study of Roman Libya. Following the tenure of a British Academy Post-Doctoral Fellowship at the University of Oxford researching Roman-period olive oil production, he taught at the University of Michigan (Ann Arbor), before joining Leicester in 1992. He has published extensively on landscape archaeology, especially of arid zones, the Roman empire and its impact on people and environment, and olive oil production and trade in the ancient world. In addition to his current field project in Fezzan (Libya), which he discusses in his chapter (Chapter 9), he is also collaborating with Graeme Barker and David Gilbertson in the Wadi Faynan Landscape Survey, Jordan (Chapter 4). His principal publications include: (with J.A.Lloyd) Libya: Research in Archaeology, Environment, History, and Society 1969–1989, London, Society for Libyan Studies 1989; (with G.D.B.Jones) An Atlas of Roman Britain, Oxford, Blackwell 1993; Tripolitania, London, Batsford 1995; (with G.Barker, D.Gilbertson and B.Jones) Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume One: Synthesis, Volume Two: Gazetteer and Pottery, Paris, UNESCO 1996; Dialogues in Roman Imperialism. Power, Discourse and Discrepant Experience in the Roman Empire, Portsmouth, RI. 1997; and (with M.Gillings and J.van Dalen) Mediterranean Landscape Archaeology 3: Geographical Information Systems and Landscape Archaeology, Oxford, Oxbow 2000. (Institutional address: School of Archaeological Studies, University of Leicester, Leicester LE1 7RH, UK) PAUL E.MINNIS is an Associate Professor at the University of Oklahoma. His research interests include paleoethnobotany, human ecology, social evolution, human responses to food shortages, the relationships between archaeology and biodiversity, and the prehistory of the North American Southwest. For the past decade he has co-directed an archaeological project in northwestern Chihuahua, Mexico, to understand the regional setting of Casas Grandes, one of the most complex prehistoric polities in North America. His books include: Social Adaptation to Food Stress, Chicago, University of Chicago Press 1985; (with C.Redman) Perspectives on Southwestern Prehistory, Boulder CO, Westview Press 1990; (with W.Elisens) Biodiversity and Native America, Norman, University of Oklahoma Press, 2000; and Ethnobotany: A Reader, Norman, University of Oklahoma Press 2000. (Institutional address: Department of Anthropology, University of Oklahoma, Norman, OK 73019 USA) MARK NESBITT is an ethnobotanist at the Centre for Economic Botany, Royal Botanic Gardens, Kew. An undergraduate degree in agricultural botany at Reading University was followed by postgraduate training in archaeobotany at the Institute of Archaeology, University College London. Since 1985 he has been involved in a wide range of archaeological fieldwork in the Near and Middle East, including Turkey, Iraq, Bahrain and Turkmenistan. His research interests include the origins, development and sustainability of crop husbandry in arid lands, the evolution of Old World cereals, and archaeological and ethnographic evidence for wild plant foods in the temperate zones. Publications include: ‘Archaeobotanical evidence for early Dilmun diet at Saar, Bahrain’, Arabian Archaeology and Epigraphy 4, 1993:20–47; ‘Plants and people in
ancient Anatolia’, Biblical Archaeologist 58, 1995:68–81; (with D.Samuel) ‘From staple crop to extinction? The archaeology and history of the hulled wheats’, in S.Padulosi, K.Hammer and J.Heller (eds) Hulled Wheats: 41–100, Rome: IPGRI 1997. (Institutional address: Centre for Economic Botany, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AE, UK) PAUL NEWSON is a PhD student at the University of Leicester. After a first career as a graphic designer he took his BA and MA in archaeology at University College London. Funded by the Natural Environment Research Council, he is preparing his PhD thesis at Leicester on water management strategies in Roman Arabia and their implications for understanding processes of Romanization, combining field studies of dryland water-management systems of the kind discussed in his chapter (Chapter 5) with a detailed analysis of Roman-period field systems in the Wadi Faynan in southern Jordan using Geographical Information Systems. In 1999 he was Acting Assistant Director of the Council for British Research in the Levant’s British Institute at Amman for Archaeology and History. (Institutional address: School of Archaeological Studies, University of Leicester, Leicester LE1 7RH, UK) SARAH O’HARA is a Reader in Environment and Society in the School of Geography, University of Nottingham. She completed a BSc in Physical Geography and Geology at the University of Liverpool, an MSc in Geography at the University of Alberta, Canada, and a DPhil. in Geography at the University of Oxford. Her research interests include environmental reconstruction, human/environmental interactions and water resource management in the world’s drylands. Her research has been carried out in Albania, Canada, Iran, Mexico, Turkmenistan and Uzbekistan. More recently she has begun collaborating with Professor Julian Henderson on the Ancient Raqqa Industrial Project in Syria. Recent publications include: (with D.Thomas, M.D. Bateman and D.Mershahi) ‘Development, age, and environmental significance of a Late Quaternary sand ramp, central Iran’, Quaternary Research 48, 1997:155–61; (with T.Hannan) ‘Irrigation and water management in Turkmenistan: past systems, present problems, and future scenarios’, Europe-Asia Studies 51, 1999:21–41; and ‘Learning from the past: water management in Central Asia’, Water Policy (forthcoming, 2000). (Institutional address: School of Geography, University of Nottingham, Nottingham NG7 2RD, UK) JEFFREY R.PARSONS (PhD in Anthropology, University of Michigan 1966) is currently Professor of Anthropology and Curator of Latin American Archaeology at the University of Michigan, Ann Arbor, Michigan USA. In addition to ongoing fieldwork in the Valley of Mexico since 1961, he has also worked in Guatemala (Tikal, Peten, 1966), Peru (Chilca, central coast, 1969–1970 and Junin, central highlands, 1975–76), Iceland (Eyaforur, 1985) and northwest Argentina (Jujuy, 1995). His research interests include the development of pre-industrial complex society, settlement pattern studies, archaeological ethnography and (in the Andes) long-term relationships between herders and agriculturalists. Current plans include fieldwork on the pre-hispanic utilization of lacustrine resources in the Valley of Mexico, and prehispanic regional organization in the Peruvian central highlands. Significant publications include: (with C.M.Hastings and Ramiro Matos M.) ‘Rebuilding the state in highland Peru: herder-cultivator interaction during the Late Intermediate Period in
the Tarama-Chinchaycocha region’, Latin American Antiquity 8, 1997:317–41; (with G.Mastache, R.Santley, and M.C. Serra) Arqueología Mesoamericana: Homenaje a William T.Sanders, Mexico D.F., Institute Nacional de Antropología e História 1996; ‘Political implications of pre-hispanic chinampa agriculture in the Valley of Mexico’, in H.Harvey (ed.) Land and Politics in the Valley of Mexico: A Two Thousand-Year Perspective: 17–42, Albuquerque, University of New Mexico Press 1991; and (with M.H. Parsons) Maguey Utilization in Highland Central Mexico: An Archaeological Ethnography, Ann Arbor, University of Michigan Museum of Anthropology, Anthropological Paper No. 82, 1990. (Institutional address: Department of Anthropology, University of Michigan, Ann Arbor, Michigan MI 48104, USA) STEVE A.ROSEN teaches archaeology in the Department of Bible and Ancient Near East at Ben-Gurion University in Beer-Sheva, Israel. He received his Bachelor’s degree from the University of California at Berkeley in mathematics and anthropology and his graduate degrees in anthropology from the University of Chicago. Prior to his current position, he worked for eight years as a survey archaeologist for the Archaeology Survey of Israel in the Negev. His research interests include desert adaptations, the archaeology of pastoral nomadism, Levantine prehistory and lithic analysis. His major publications include: Lithics After the Stone Age, Walnut Creek, Altamira Press 1997; (with G.Avni) The ‘Oded Sites: Investigations of Two Early Islamic Pastoral Camps South of the Ramon Crater, Beer-Sheva, Ben-Gurion University Press 1997; and Archaeological Survey of Israel Map of Be’erot Oded, Beer-Sheva, Ben-Gurion University Press 1994. (Institutional address: Archaeology Division, Ben-Gurion University, PO Box 3653, Beer-Sheva, 84105 Israel) ROBERT SOPER (MA Cambridge, UK) is a Senior Research Fellow in Archaeology at the University of Zimbabwe. Between 1962 and 1985 he worked for the Nigerian Antiquities Department, the British Institute in Eastern Africa, Ibadan University, and the University of Nairobi. His principal research interests have included the later prehistory of East Africa (especially the Early Iron Age and later ceramics), the site of Oyo Ile in Nigeria, and Great Zimbabwe tradition sites in northern Zimbabwe. Significant publications include: ‘A general review of the Early Iron Age in the southern half of Africa’, Azania 6, 1972:5–37; ‘Roulette decoration on African pottery’, African Archaeological Review 3, 1985:29–51; ‘The palace at Oyo Ile, western Nigeria’, West African Journal of Archaeology 22, 1993; and (with B.E.Kipkorir and J.W.Ssennyonga) The Kerio Valley: Past, Present and Future, Nairobi, University of Nairobi Institute of African Studies 1983. He has conducted research on the Nyanga terrace complex since 1993, and a monograph on this is forthcoming. (Institutional address: History Department, University of Zimbabwe, PO Box MP 167, Mount Pleasant, Harare, Zimbabwe) GREG SPELLMAN is a lecturer in the School of Environmental Science at University College Northampton (UK). After a BA in Geography at the University of Sheffield, he took a PGDip in Applied Meteorology and Climatology at the University of Birmingham and an MA in Professional Education at the University of Leicester. His research interests include synoptic climatology, extreme hydrological events in the Iberian peninsula and the meteorology of air pollution. He is currently researching on the synoptic climatology of Spain, particularly the evaluation of various downscaling
methods in order to improve regional climate change scenarios. Recent publications include: ‘An application of artificial neural networks to the prediction of surface ozone concentrations in the United Kingdom’, Applied Geography 19, 1999:123–36; and ‘Investigating the synoptic climatology of precipitation in Mallorca, Spain’, Journal of Meteorology 23, 1998:117–30. (Institutional address: School of Environmental Science, University College Northampton, Northampton NN2 7AL, UK) JOHN E.G.SUTTON (MA Oxford, PhD East Africa, FSA) was Director of the British Institute in Eastern Africa from 1983 to 1998 and was previously Professor of Archaeology at the University of Ghana. He has been mainly concerned with later archaeology and its contribution to African history, with a special interest in field systems and agricultural technology. His first visit to Engaruka in the northern Tanzanian Rift Valley—the subject of his contribution to this volume—was in 1963, while a research scholar of the British Institute. Later, as a lecturer at the University of Dar es Salaam, he continued the investigation of the Engaruka irrigation-agricultural settlement with student assistants. That study has been extended in more recent years within a broader comparative project on African field systems and cultivation strategies. Engaruka and other prominent sites are described in detail in Archaeological Sites of East Africa: Four Studies (special volume 33 of Azania for 1998), and there is a shorter illustrated account in A Thousand Years of East Africa (Nairobi, British Institute of East Africa 1990). Another special volume of Azania (29/30, 1995), The Growth of Farming Communities in Africa from the Equator Southwards, surveyed the present state of research on the Early Iron Age and the Bantu agricultural expansion. (Address: 118 Southmoor Road, Oxford OX2 6RB, UK) MATS WIDGREN teaches at Stockholm University, where he is a Professor in Human Geography. He received his PhD in Stockholm in 1983 and has researched on the historical geography of agricultural landscapes from the Iron Age to the present in Sweden and in southern and eastern Africa. Among his publications are: ‘Is landscape history possible?’, in P.Ucko and R.Layton (eds) The Archaeology and Anthropology of Landscape: 94–103, London, Routledge 1999; ‘Fields and field systems in Scandinavia during the Middle Ages’, in G.Astill and J.Langdon (eds) Medieval Farming and Technology—The Impact of Agricultural Change in Northwest Europe: 173–92, Leiden, Brill 1997; ‘Strip fields in an iron age context: a case study from Västergötland, Sweden’, Landscape History 12, 1990:5–24; and Settlement and Farming Systems in the Early Iron Age: A Study of Fossil Agrarian Landscapes in Östergötland, Sweden, Stockholm, Almquist & Wiksell 1983. His most important works in Swedish are his contribution to the first volume of the agrarian history of Sweden Jordbrukets första femtusen år (1998) and a book on medieval field systems in Bohuslän, Sweden (1995). (Institutional address: Department of Human Geography, Stockholm University, S-106 91 Stockholm, Sweden)
Foreword One World Archaeology is dedicated to exploring new themes, theories and applications in archaeology from around the world. The series of edited volumes began with contributions that were either part of the inaugural meeting of the World Archaeological Congress in Southampton, UK in 1986 or were commissioned specifically immediately following the meeting—frequently from participants who were inspired to make their own contributions. Since then the World Archaeological Congress has held three further major international congresses: Barquisimeto, Venezuela (1990), New Delhi, India (1994), and Cape Town, South Africa (1999). It has also held a series of more specialised ‘intercongresses’ focusing on: Archaeological ethics and the treatment of the dead (Vermillion, USA, 1989), Urban origins in Africa (Mombasa, Kenya, 1993), and The destruction and restoration of cultural heritage (Brac, Croatia, 1998). In each case these meetings have attracted a wealth of original and often inspiring work from many countries. The result has been a set of richly varied volumes that are at the cutting edge of (frequently multi-disciplinary) new work, and which provide a breadth of perspective that charts the many and varied directions that contemporary archaeology is taking. As series editors we should like to thank all editors and contributors for their hard work in producing these books. We should also like to express our thanks to Peter Ucko, the inspiration behind both the World Archaeological Congress and the One World Archaeology series. Without him none of this would have happened. Martin Hall, Cape Town, South Africa Peter Stone, Newcastle, UK Julian Thomas, Manchester, UK June 2000
Preface This book stems from a symposium organized by the editors on the archaeology of drylands held at the World Archaeological Congress at Cape Town in January 1999. The Congress provided the opportunity to bring together scholars working on the archaeology of different regions of the world’s drylands, to pool experiences and in particular to investigate the extent to which we could discern common themes. Although over a third of the world’s population today lives in arid and semi-arid lands, there are many gaps in our understanding about how fragile or resilient these regions are for human settlement. To fill these gaps we need to answer questions that are likely to be of very great significance for the global community in the twenty-first century. Many dryland regions have abundant remains of ancient settlement, and people have often speculated that the actions of farmers and herders in the past must have been important in creating the degraded landscapes of the present. For decades the debate has been characterized more by speculation than informed debate and by a propensity to argue for simple processes of cause and effect in terms of climatic change or humanly induced environmental degradation. In the past fifteen years or so, however, inter-disciplinary archaeological and palaeoecological studies (especially when employed within integrated research frameworks) have demonstrated their potential to move the debate forwards by providing detailed case studies of how ancient societies actually exploited dryland landscapes, how they interacted with them, and the complex environmental and social contexts in which they variously succeeded or failed. Moreover, as we conclude in Chapter 1, the advancement of understanding about past dryland societies and environments, and of the complexity of their interactions with each other, has a particular urgency, given the way in which changing political agendas have been prone to either demonize or sentimentalize them. Nine papers were given at the Cape Town symposium, and a series of common themes about dryland settlement emerged from the discussion. Other papers were then commissioned by the editors, so that the collection as a whole would draw on the archaeology of different kinds of drylands throughout the world (the locations of which are shown in Figure 1.3), different periods of the past and different kinds of societies, but all addressing the issues we had identified at Cape Town to give a comparative perspective. Common themes, though, as we discuss in Chapter 1, do not equate with similar solutions to dryland living, or similar responses to threats and opportunities: the archaeology of drylands is eloquent testimony perhaps most of all to people’s ingenuity, as well as to their resilience. The editors would like to express their thanks to the organizers of the 1999 World Archaeological Congress for their invitation to organize the Drylands Archaeology symposium, and for every assistance from them during the conference. We would also like to thank the Society for Libyan Studies for a generous grant towards our two fares to Cape Town, augmented in the case of GB by a grant from the Staff Development Fund of
the Faculty of Arts of the University of Leicester. We are very grateful to all the contributors to the volume for their patience, especially those who contributed to the symposium, whose debates helped frame the discussion document we then circulated to them and to the authors of the papers commissioned afterwards, and who have remained committed to our idea of the integrated comparative volume, despite the timelag since the conference. Finally, we would like to dedicate this book to the memory of Barri Jones. Barri was Professor of Archaeology at the University of Manchester and introduced us both to dryland archaeology in the UNESCO Libyan Valleys Survey. He died on the eve of his retirement in the summer of 1999, leaving a legacy of a major scholarly output of books and papers, an army of professional and amateur archaeologists enthused with his passion for the subject and, for his desert companions in particular, memories of hairraising adventures in his company. He was a frenetic personality who was both enchanting and exasperating to work with—he was notorious for doing too many things at once, mostly while nominally in control of a Landrover! Amongst his many talents, though, he had an extraordinary topographical eye: he was liable to get us lost in some desert waste to visit an archaeological site once noted by a traveller 100 years ago, and thrash the vehicle in the process, but he was also by far the best person to be with to get safely back to camp again. Graeme Barker and David Gilbertson January 2000
Part I INTRODUCTION
1 Living at the margin: themes in the archaeology of drylands GRAEME BARKER AND DAVID GILBERTSON
INTRODUCTION Drylands cover 40 per cent of the land area of the Earth: their total area is about 60 million km2, of which about ten million km2 are hyper-arid deserts (Fig. 1.1). Drylands support over one fifth of the world’s population, and arid and semi-arid lands together over a third. Living conditions vary from the most affluent and profligate to the desperately poor—in some cases in close proximity. The political stability and ecological, economic and social sustainability of dryland settlement are among the most daunting challenges confronting the global community in the twenty-first century: water seems likely to be a primary flashpoint for disputes between neighbouring states, with dryland irrigation systems under strain from fast-growing populations;
Figure 1.1 The world map of drylands Source: After UNEP, 1992 and with environmental refugees from global warming predicted to be in the order of 150 million by the year 2050 under the business-as-usual scenario (Houghton, 1997), the catastrophic consequences will be particularly acute for dryland populations. Many dryland regions have archaeological remains suggesting that once upon a time
The archaeology of drylands
4
there must have been intensive phases of settlement in what are now dry and degraded environments (Fig. 1.2). People have often speculated about what must have happened to turn past glories into present-day barrenness, generally dividing in favour of climatic change or human agency as the primary culprit. Perhaps the climate shifted to greater aridity? Or was it that people sowed the seeds of their own destruction through their folly, for example by developing irrigation systems that caused salinization, or by stripping the landscape for fuelwood, or by allowing their livestock to over-graze the vegetation? In general, the debate has been characterized more by confident assertion than well-founded argument. Furthermore, as we discuss later in this chapter, contemporary ecological theory suggests that relations between dryland environments, climate and people are by no means simple (Beaumont, 1993): drylands can sometimes be remarkably resilient, for example, recovering relatively quickly from over-exploitation, and simple procedures by farmers can often protect against the latter (Mortimore, 1998; Tiffen et al., 1994). These findings are at odds with the simplistic models that have tended to dominate the archaeological literature about how climate and people may or may not have affected drylands in the past.
Figure 1.2 A Roman-period fortified farm on the desert margins of Tripolitania, northwest Libya Photograph: G.Barker Modern inter-disciplinary archaeology, especially when working in conjunction with other social and environment sciences, has the potential to move the debate forward. Archaeology deals with the entire human past, its geographical scope is regionally specific but worldwide, its scale of enquiry ranges from distributions and processes of change at the global scale and over millennia down to the actions of individuals. We can use the techniques of landscape archaeology to understand how different kinds of societies, whether recent or remote in time, exploited the different dryland environments
Living at the margin: themes in the archaeology of drylands
5
of the world. We can characterize the risks and opportunities confronting those societies, identify the solutions they reached and often the reasons for them, as well as monitoring the short- and long-term effects of those solutions. By developing a more sophisticated understanding than has hitherto characterized the debate about variability in past land use strategies and the reasons for their successes and failures, archaeology can contribute effectively to modern debates about desertification and the sustainability of dryland settlement (Beaumont, 1993). The World Archaeological Congress held at Cape Town in January 1999 provided an ideal opportunity to explore these issues from the perspectives of various scholars working on the archaeology of different regions of the world’s drylands. The symposium focused on nine contributions, discussing work in the Near East, North and sub-Saharan Africa and North America, all of which are represented in this volume. A series of common themes about
Figure 1.3 The location of the case studies in this volume; numbers refer to chapter numbers dryland settlement rapidly emerged, and the papers presented at the symposium were rewritten, and further papers commissioned, to address these common issues within a comparative perspective in this book, with case studies drawn from different kinds of dryland regions throughout the world (Fig. 1.3). Common themes, though, as we discuss below, do not equate with similar solutions to dryland living, or similar responses to risks and opportunities.
The archaeology of drylands
6
THEMES IN DRYLAND ARCHAEOLOGY The term ‘drylands’ obviously fixes attention upon low precipitation. Common knowledge emphasizes that the climatic significance of this shortage depends upon the other aspects of the atmospheric environment—the radiation budget, thermal regime, wind regime, the sources and pathways of moisture, including fog, as well as the many other components of the biosphere and lithosphere that play significant parts in the hydrological cycle. The meteorology and climatology that produce drylands are not simple (Spellman, Chapter 2): understanding them requires an appreciation of variability of precipitation and drought in both space and time. Rainfall in many drylands is typically characterized as consisting of erratic, short, localized downpours of high intensity. Low average precipitation totals are associated with notable variability. Fierce and localized downpours creating sudden and dangerous floods are the primary resource base that many indigenous peoples have had to utilize to maintain themselves, their crops and their animals for millennia (as well as environmental hazards for archaeologists working in arid lands; see Fig. 1.4!). However, it is the prospect of prolonged and severe drought that dominates thinking about drylands. Instrumental, historical and palaeoenvironmental records show that episodes of severe drought lasting decades or more in length have not been uncommon in many drylands over the last few thousand years (e.g. Bureau of Meteorology, n.d; Fritts, 1991; Lamb et al., 1995; Nicholson, 1994). Ingenuity, flexibility and enterprise have been required from individuals, communities and organizations to negotiate their survival in the face of such uncertainty and risk. It is important to remember that, apart from such fluctuations at the scale of seasons and decades, modern drylands have also been part of tremendous fluctuations in climate operating at the global scale in the remote past, from major shifts in the world’s oceanographic and atmospheric systems. The period from approximately 18,000 to 10,000 years ago, the last phase of the Pleistocene (the ‘Ice Age’), saw the last major ice advances of glaciers and icesheets in the high latitudes. Regions of the world that now enjoy a temperate climate, such as Europe and North America, were cold and arid. Regions such as the Sahara were hyper-arid: long-term reductions of rainfall reached well beyond the present desert margin, as far south as present-day Nigeria, with much of interior North Africa having to be abandoned by human populations. However, between 10,000 and 8,000 years ago, during
Living at the margin: themes in the archaeology of drylands
7
Figure 1.4 Drowning in drylands?—two vehicles of the UNESCO Libyan Valleys Survey sunk in a flash-flood Photograph: G.Barker the opening millennia of the Holocene or Postglacial, the environment across the Sahara was notably wetter than occurs today, with the development of lakes and woodland in favoured locations such as those that are now desert oases and savannah-like habitats on the surrounding plateaux (Barker et al., 1996; and see Mattingly, Chapter 9). The people who colonized these places were able to live by plant and shellfish gathering, fishing and hunting not just steppe animals like gazelle but wetland species such as turtle, hippo and crocodile (Cremaschi and di Lernia, 1998; Wendorf and Schild, 1980). In the Near East, wetter environments at this time were the context for the development of mixed farming systems of the kind found in the Wadi Faynan in southern Jordan (Barker, Chapter 4), and the first farming in Turkmenistan may also have developed in wetter conditions than today (Nesbitt and O’Hara, Chapter 6). Desiccation started to develop about 6,000 years ago in North Africa and the Near East, with people responding differently. In the Sahara, people shifted to cattle and sheep/goat herding (Barich, 1987, 1998; Wendorf et al., 1984, 1989), whereas the farmers of Wadi Faynan started to experiment with methods of trapping and storing water, which were the beginnings of recognizable systems of dryland farming there (Chapter 4). It was also about this time, in the fourth millennium BC, that prehistoric farmers in Turkmenistan started to build canals to divert floodwaters and, via small feeder channels (aryks), bring it to their fields (Chapter 6). So far as we can tell, it was not until notable aridification developed in the Sahara around 4,500 years ago, as seasonal streams replaced perennial streams, salt pans replaced fertile lake floors, and the modern regime of flash-floods and droughts developed, that similar systems of dry farming were developed by farmers living in the oases (Chapter 9) and on the desert margins such as the Tripolitanian Pre-desert (Barker et al., 1996). Archaeological evidence for the development of irrigation-based farming is a recurrent theme throughout this book because, whilst far from all drylands are warm or hot for all
The archaeology of drylands
8
or part of the year, the provision of an adequate and reliable supply of water in warm and sunny regions has been a goal for innumerable communities over time. Rainwater harvesting and floodwater farming are described for North America, Africa and Asia in a variety of chapters. Four major themes emerge from these case studies, the first three of which are closely inter-related. The first is the resilience of many farmers in antiquity to cope with harsh and riskprone arid environments and their climatic vicissitudes over long periods of time. This longevity of occupation points to an inherent robustness of many of these past communities, their attitudes and ways of life. The second is the repeated evidence for the similarity of the ‘building blocks’ or ‘tactics’ employed by most ancient farmers in drylands—building walls to trap soil and divert or stem water flow, building channels (including underground in the case of the Turkmenistan qanats [Chapter 6] and Libyan foggaras [Chapter 9]) to divert water, and so on. The third is the remarkable variability in the overall systems that were put together from such building blocks and the way they invariably reflect detailed local knowledge of topography, weather patterns, and so on: ancient farmers knew from observation exactly how and where the water would flow after a storm, and so knew how best to manage that flow to suit their purposes. The fourth critical finding from our survey, though, is that the diversity we can observe in the archaeology of dryland farming systems is in no sense just a straightforward matter of commonsense observation by ancient farmers of what was the ‘best fit’ to particular environmental or economic circumstances. We can see today how dryland communities attempt to manage themselves and their habitats within the context of a whole nexus of attitudes, beliefs, as well as economic, social, geographical, educational, agricultural and technological processes; and whilst many of these details will elude archaeologists, given the nature of our evidence, the case studies illustrate how people took choices, and not always the right ones, within a complex mix of factors, including perceptions of risk, the need or desire for economic advantage, and the institutional and regional context in which they were operating. As Widgren comments (Chapter 14), models of agrarian development too often assume even developments of farming systems in response to particular environmental, social or economic pressures, but the archaeological record emphasizes above all the unevenness of development in both space and time. We do not see the kind of evolutionary development of land use systems so often assumed for the past, for example from simple to complex or from extensive to intensive—they were ‘formed and changed within specific, place-bound, social, historical, and ecological contexts’ (Widgren, p. 262). One key influence on the character and scale of an irrigation system was, not surprisingly, related to the extent to which the agricultural product was to serve only the local community. Examples of such systems, ‘islands’ of relatively intense landscape development, are described for societies varying widely in time, place and social complexity, for example in the North American Southwest in prehistoric times (Minnis, Chapter 15), the Achaemenid or Parthian periods in Turkmenistan (Chapter 6), the Libyan oases (Mattingly, Chapter 9) and the Tricastin region of southern France (van der Leeuw, Chapter 18) in the Roman empire, and at various localities in sub-Saharan Africa in recent centuries (Sutton, Chapter 11; Soper, Chapter 12; Widgren, Chapter 14). Widgren illustrates how both hierarchies and the absence of hierarchies can be associated
Living at the margin: themes in the archaeology of drylands
9
with labour-intensive agriculture and how both market orientation and subsistence farming can be connected with labour-intensive farming. These landscapes and irrigation systems stand in contrast not only to the examples described in Israel (Rosen, Chapter 3), Jordan (Barker, Chapter 4), Syria (Newson, Chapter 5) and northwest Libya (Gilbertson et al., Chapter 8) of archaeological landscapes that were once very productive, if peripheral, parts of the imperial economies, characterized by the large-scale import and export of goods, products and information to the core areas of the Mediterranean, but also to the somewhat similar relationship of highland Mexico to the rest of the Inca state (Parsons and Darling, Chapter 16). ‘Patchiness’ in distribution has also been identified in the studies by Jones and Crook (Chapter 17) of the Swiss bisses—canal systems that have tapped and redistributed water within the surprisingly dry environment of the Canton of Valais for over a millennium but that remain a relatively unknown but vital component of the economy of one the world’s richest and technologically sophisticated countries. Clearly, archaeological investigation of ‘marginal’ landscapes has to engage with the need for complex inter- and intra-regional articulations of explanations of cause and effect. All archaeologists also have to recognize the commonplace dictum that ‘the past is a foreign country’: things were thought and done differently there. Then, as now, it seems that many individuals, organizations and polities have behaved in relation to their situation and their environment in manners that do not appear rational to the modern external observer or to those with the wisdom of hindsight. Then, as now, people made poor decisions, foolish decisions, self-interested decisions, carried out actions that they or their neighbours had cause to regret; or, more generally, they misunderstood their land and situation. One striking example of long-term devastation caused by the economic needs of the Roman empire was the pollution of the Wadi Faynan in Jordan by copper and lead mining (Barker, Chapter 4), but it is important that we do not dismiss such actions as the exclusive domain of market-driven economies (like the profligacy of Turkmenistan irrigation farmers once they lost their sense of ownership of the system in the Soviet period: Chapter 6): as Minnis (Chapter 15) comments in the case of North America, indigenous subsistence foragers and farmers, equally, have not always been environmentally ‘correct’, ‘sound’ or ‘neutral’. A good example of the importance of perception affecting decision making by dryland farmers responding to adverse conditions occurred during the great drought that affected the wheat-arid far north frontier of South Australia between 1881 and 1884, the climatic effects of which were documented in remarkable detail by a sophisticated network of instrumental records maintained by, amongst others, telegraph operators (Bureau of Meteorology, n.d.). The well-known account of this episode by Meinig (1962) based upon parliamentary records and newspapers reveals that the human and economic impacts of the first two years’ drought were devastating. Many wheat farmers were sustained by their belief that the ‘rain followed the plough’: more ploughing and tilling, they thought, would release further soil moisture into the atmosphere, and so break the drought. Others farmers, desperate to maintain overall production totals and to pay their mortgages to the Government for their newly acquired lands (pastureland hitherto), ploughed and planted ever larger areas of bush, believing that minimal returns from vast areas would compensate for the poor yields per acre. Both strategies, of course, made
The archaeology of drylands
10
things far worse—as the drought grew worse and human distress grew ever more profound, ever larger areas of land were being ploughed and farmed. It is salutary for archaeologists, who rarely have access to such sophisticated documents or precise chronologies, to reflect on the complete reversal in present thinking of the nature of human—environment interactions that underpinned these farmers’ behaviour just over a century ago. The case studies demonstrate repeatedly that drought, however pernicious and sustained, is not necessarily the sole cause of ‘abandonment phases’ identified in the archaeological record of drylands. Chapter 10 by Butler and D’Andrea, for example, shows that episodes of famine and distress (like, indeed, those of success and prosperity) cannot be explained in drylands by consideration of one factor, even such vital factors as drought or flood. Discussing the Northern Highlands of Ethiopia, a region almost synonymous with drought and famine for most readers, they emphasize the potency for understanding famine in the area of the following, sometimes archaeologically invisible, factors: human smallpox; cattle rinderpest; plagues of predators such as locusts, ants and army worms; conflict; and social and political circumstances. In fact, drought alone seldom causes famine. The sustainability of many dryland communities, pastoral (Kinahan, Chapter 13) as well as agricultural, is underpinned especially by the flexibility of traditional practices: their capacity to avoid, to mitigate and to create buffers against risk and adversity; more generally their ability to organize themselves effectively in drought-prone habitats; and, in the last resort, their willingness to relocate (Mortimore, 1998:122). We can see from the archaeological record that systems without such flexibility did not have the necessary resilience for longterm survival, as in the case of the productive but short-lived systems of cash-crop farming that Romanized Libyans developed in the Tripolitanian Pre-desert to supply the local military and coastal urban markets (Gilbertson et al., Chapter 8), or the intensively irrigated fields built to feed the large industrial workforce of Roman miners in the Wadi Faynan in Jordan (Barker, Chapter 4), or the massive centralist-administered irrigation systems of Soviet Turkmenistan (Nesbitt and O’Hara, Chapter 6).
ARCHAEOLOGY AND DESERTIFICATION During the last few decades, many of the world’s drylands—the hotter drylands especially but not exclusively—have been viewed as threatened by ‘desertification’. The term was coined by Aubreville (1949) in a report on the vegetation of Africa, and its meaning has developed through time. Thomas and Middleton (1994:9–10) defined it as ‘land degradation in arid, semiarid, and dry sub-humid areas resulting mainly from adverse human impact’, and though some authors have also used the term to refer to land degradation caused by a sustained aridification of climate, most prefer to use the term to refer to the effects of human actions, though climate and people can clearly work in tandem to produce deterioration in dryland environments, as may be the case in the context of global warming today (Barrow, 1995; Millington and Pye, 1994; Spellman, Chapter 2). The key ideas focus upon significant and long-term degradation producing a loss of potential in biological, soil and water resources. Manifestations of humanly
Living at the margin: themes in the archaeology of drylands
11
caused degradation may include decreased vegetation cover, timber loss, salinization, reduced water supplies, lower crop yields, outbreaks of disease, accelerated erosion of soils and dust storms, and induced regional climatic change, all encapsulated in the popular metaphor of the conversion of pastoral or arable lands within drylands into desolate and sterile desert. The Green agendas of recent decades have rightly and repeatedly focused on dryland ecosystems and the sometimes appalling consequences of human impacts upon them (Fantechi and Margaris, 1986), often induced or certainly exacerbated by top-down programmes of economic development (IFAD, 1992). Beaumont (1993:474) concluded his book with a bleak prediction of the inevitability of this process for the world’s poorest nations: ‘in certain cases land degradation may be a sacrifice which has to be paid in order that local populations can survive future drought or famine’. According to Tolba and el-Kholy (1992:134), the current rate of desertification is about 60,000 km2 per year, amounting to 0.11 per cent per year of the total area of dryland. On this calculation, desertification today threatens no less than 70 per cent of the world’s drylands, which represents over 25 per cent of the world’s land surface. Grainger (1990) and Spooner (1989) argued that such desertification can be recognized in Australia, North America and South Africa in the first half of the twentieth century, and that it was likely to have been a factor in antiquity, too. There are, however, conflicting views about the extent of desertification today. Thomas and Middleton (1994), for example, questioned whether desertification in recent decades is actually a global problem of such vast dimensions, as opposed to local manifestations of local problems of smaller significance. There are also examples of spirals of land degradation being reversed by indigenous technological adaptations working in combination with population growth and market opportunities. The Machakos district of Kenya was considered an environmental disaster in the 1930s because of massive soil erosion and famine, but by 1990 terrace construction had protected arable land, farmed and protected trees provided sufficient fuel-wood, and agricultural production per person and per hectare had increased, sustaining a population five times larger than that of the 1930s (Tiffen et al., 1994). Deforestation and massive erosion on the Yatenga plateau in Burkino Faso were exacerbated by mechanization programmes funded by ill-judged development aid, but the reinstatement of traditional systems of terrace building stopped erosion and doubled crop yields (Lean, 1994). Mortimore (1998:149–56) described an examples of c. 150 years of sustainable intensification by smallholders in Kano, Nigeria, in the context of population growth and monetization. Archaeological remains in many drylands have been grist to the mill of the desertification debate. For example, Hughes with Thirgood (1982:60, 74) wrote that in the more arid regions, forests that formerly moderated the climate and equalized the water supply were stripped away, permitting the desert to advance. The image of ruined cities in North Africa, from which olive oil and timber were exported in ancient times but which were buried beneath desert sand, epitomizes the environmental factor in the decline of civilization…. Roman dams and canals stand in dry wadis today as witness to the fact that the destruction of the vegetation and consequent desiccation have changed the
The archaeology of drylands
12
environment. Terms such as excessive land use, population pressure, loss of biological diversity and vegetation cover, mis-use of water, accelerated soil erosion and ever-larger human needs all recur in archaeological explanations (Gilbertson et al., Chapter 8), often in association with reference to times of perceived political unrest, military invasion, conflict and drought. In his review of the historical likelihood of humanly induced desertification, Spooner (1989) cautioned against the simplistic tendency to assume that desertification, famine, drought and poverty will inevitably be found together, and several case studies in this book support those who argue for the complexity of desertification processes today. For example, according to Barker et al. (1996) and Gilbertson et al. (Chapter 8, this volume), the vast Romano-Libyan and Islamic settlements and farms of the Libyan Pre-desert at the edge of the northern Sahara seem to have neither produced nor experienced the kind of self-induced environmental degradation described by Hughes with Thirgood (1982). Indeed, the increase in human population numbers, farming intensity and land management probably promoted a greener, more diverse and infinitely richer and productive environment than has occurred since the great aridification in climate that afflicted the region some 4,500 years ago. There are, in fact, good reasons to suspect that Romano-Libyan farmers did have significant and deleterious impacts upon their arid environment, but there are few reasons to believe that catastrophic long-term climatic change, short-term catastrophic drought, or anthropogenically induced environmental degradation played a central role in the progressive abandonment of these settlements—a process that is still underway after nearly 1,500 years. Rosen (Chapter 3) also suggests that substantial cultural changes in the ancient Negev desert are not best explained by climatic catastrophe, invasion or the inabilities of people to manage their desert environment: rather, periods of cultural florescence were related to increased economic and social input from, or integration with, the Mediterranean ‘core area’, with desert pastoralism strongly geared to active markets in the settled zone, and likely to be sorely afflicted by the latter’s collapse. Yet in the adjacent deserts of southern Jordan, the Wadi Faynan Landscape Survey (with many of the same members as the UNESCO Libyan Valleys Survey, and using similar methodologies) has found convincing evidence for dramatic humanly-induced land degradation in the wake of agricultural and industrial intensification in the context of Roman imperialism. In the Saharan Fezzan, Garamantian development of foggara irrigation systems may have been a key factor leading to the decline of their civilization as a result of over-extraction from a non-renewable groundwater source (Mattingly, Chapter 9). Ballais (Chapter 7) argues that increases in soil erosion in the eastern Maghreb in classical antiquity reflected specific combinations of climatic change and human activities, and affected parts of the landscape in different ways. In the Roman imperial centuries, increased intensity of rains, or the annual amount of precipitation, badly affected regions already made vulnerable by vegetation degradation or unwise cultivation systems, whereas the irrigated zones and terraced mountains were more resilient. In the middle and lower Rhône valley in southern France, episodes of climatic change are out of step with archaeologically visible episodes of human impact; in the Tricastin region here, accelerated erosion can be tied clearly with
Living at the margin: themes in the archaeology of drylands
13
the lack of maintenance of Roman drainage systems (van der Leeuw, Chapter 18). The role of pastoralism in desertification, like dryland farming, is much debated. It has often been asserted to be particularly pernicious, the prime cause of a legacy of apparently exhausted, depleted and deserted drylands today. The archaeological literature on drylands is replete with references to the likelihood of ‘over-grazing’, ‘excessive grazing’, and so on, implying that there is today, and that there was in times past, some knowable and realizable intensity of pastoral activity in drylands—a ‘carrying capacity’ which, if exceeded, must have had dire consequences for the pastoralists and their habitats. However, the core of this view—that such a carrying capacity exists—is now being challenged by ecologists who consider that ‘boom-or-bust’ models of animal population numbers may be more appropriate (Thomas and Middleton, 1994). There may never have been sufficient time for any medium- or long-term balance to be struck between livestock numbers and arid environments because arid lands are too variable in their production of forage—this variability itself a consequence of precipitation, which is variable in space and time (Holling, 1973; Noy-Meir, 1978; Olsvig-Whittaker et al., 1993; Scoones, 1995; Walker et al., 1981). Typically, drought is likely to have reduced animal numbers drastically before irreparable damage was done to pastures (Noy-Meir, 1978). Indeed, grazing drylands pastures may have had some overall beneficial impacts (Warren and Khogali, 1992), whilst Olsvig-Whittaker et al. (1993) argue that many dryland pastures may in some sense be ‘adapted’ to grazing stress and that pastoral disturbance could be regarded as a natural component of many arid environments. In brief, not all dryland environments are as fragile as some popular literature suggests (Thomas and Allison, 1993). Parallel arguments about the likely complexity of past relations between pastoralism and ecological change are put forward on the basis of archaeological evidence by Gilbertson (1996) and Kinahan (Chapter 13). In the case of the Maghreb, Ballais (Chapter 7) also concludes that periods of conquest, often assumed in this region to be periods of environmental devastation wrought by nomadic pastoralists, were probably in fact characterized by less arable land, a progressive development of ‘natural’ vegetation and pastures, and so less soil erosion. Holling’s (1973) ecological view that arid lands are ‘non-equilibrium but persistent’ may have utility for many archaeologists working in drylands, not least because it serves as a disincentive to extrapolate from the local diagnosis of an ancient cause and effect to the inference of causality at regional or global scales. The possibility of non-linear relationships within and between environmental processes and human activities must also be considered. Relatively minor changes in the human or biophysical environment can, in principle, set in train self-sustaining sequences of events and processes that can cause the environment to transform from one state to another, with cause and effect entangled. In drylands today, relationships between individuals, communities, institutions and the landscapes with which they interact are clearly neither simple nor linear in form (Ellis, 1995; Phillips, 1993; Chapter 10). The principal argument of the case studies in this volume is that the same was certainly the case in the past, even though the nature of archaeological and palaeoecological evidence is such that it may sometimes be difficult or impossible to identify the key players, critical species or ideas, the nature of underlying trends or pre-disposing factors, the agencies of stability and the triggers of
The archaeology of drylands
14
change.
CONCLUSION The archaeology of drylands is one of the richest bodies of archaeological data and frequently the most visible for archaeological surveyors, though it is also often amongst the most vulnerable to destruction by development programmes far from the eye of national archaeological services. Moreover, the deflated landscapes of most drylands frequently pose daunting challenges of archaeological analysis, given the paucity of deep stratigraphies and organic remains, though in some cases both the latter are present in rich abundance; palaeoecologists often face similar challenges, requiring similar persistence and ingenuity in response. However, the practical (and often logistical) difficulties of dryland archaeology should not dissuade us from attempting to understand its significance: whilst the details of structure and agency in past dryland settlement will often be problematical to determine, a better understanding of the complexity of people’s interactions with dryland environments must surely underpin the desertification debate. In many parts of the world, too, investigating how past societies lived in drylands is critical for understanding not just how and why they lived as they did, and desertification theory, but also, in the case of ancient water-management systems, the extent to which the latter could or should be rebuilt to the advantage of local communities and their ecosystems today (e.g. Barker et al., 1996). Similar arguments apply to pastoral development programmes (Kinahan, Chapter 13). We need a sophisticated understanding of the environmental and social contexts of ancient dryland farmers and herders, detailed knowledge of modern dryland ecologies, and sympathetic awareness of issues such as the ownership, empowerment and organization of local technologies by indigenous peoples (Cullis and Pacey, 1992; Reij, 1991; and Gilbertson et al. [Chapter 8], Minnis [Chapter 15] and Jones and Crook [Chapter 17]). Finally, as Steve Rosen also points out in Chapter 3 (pp. 57–8), the advancement of understanding of past dryland societies and landscapes through the combination of good archaeological science and social archaeology is critical most of all to combat the politicization of much past theorizing on these matters. Relations between the desert and the sown underpin many origin myths of ethnicity, and the ‘biography’ of arid lands has frequently been rewritten to changing political agendas. The role of desert pastoralists in ancient Palestine as told to us through the Old and New Testaments is an obvious case in point, where it is repeatedly represented as a moral force for good in the history of the Israelites: the preferment of the shepherd Abel over his brother Cain, the farmer; the substitution of the ram for Abraham’s son Isaac; the commandment that lists the ox and the ass before the wife; the parable of the sheep and goats, and the lost sheep; and the role of Christ himself as the lamb of God, the Lord our Shepherd of Psalm 23. A pastoral ideology with numerous parallels to the Biblical stories underpinned the origin myth of the Incas and their sense of their right to rule subject peoples (Brotherston, 1989). Yet in both the Near East and North Africa, simplistic notions of Islamic pastoralist invaders as the prime causes of environmental and cultural decline have stemmed primarily from political agendas (Rosen, Chapter 3; Ballais, Chapter 7), and one of the planks of modern
Living at the margin: themes in the archaeology of drylands
15
Zionism has often been the contrast between the ‘greening of the desert’ of the kibbutz movement (never mind its long-term implications for the River Jordan!) and the depiction of recent bedouin pastoralisrn as inefficient and environmentally destructive. The dryland farming systems of Native American peoples have been variously portrayed as environmentally destructive or in sympathy with the landscape, according to changing colonial and post-colonial perspectives (Minnis, Chapter 15). At the root of the 1990s massacres in Rwanda was the ‘Tutsis’ and Hutus’ belief that they are derived respectively from Nilotic cattle-herders and Bantu farmers—a note left with a group of European tourists killed in Bwindi National Park in Uganda by Hutu guerillas in 1999 read in broken French ‘here is the fate of all the Anglo-Saxons who betray us to the Nilotics against the Bantu cultivators’ (Hannan, 1999)—but in fact there is very little to distinguish the two groups, and the Nilotic/Bantu dichotomy is almost certainly a mistaken creation of nineteenth- and twentieth-century scholarship (Hall, 1987, 1996). As dryland peoples face the uncertainties of the twenty-first century, understanding the richness, diversity and, above all, the complexity of the archaeology of their antecedents has never been more urgent.
REFERENCES Aubreville, A. (1949) Climats, Foràts et Désertification de l’Afrique Tropicale . Paris, Societé d’Editions Géographiques Maritimes et Coloniales. Barich, B.E. (1987) Archaeology and Environment in the Libyan Sahara: The Excavations in the Tadrart Acacus 1978–1983 . Oxford, British Archaeological Reports, International Series 368. Barich, B.E. (1998) People, Water and Grain: The Beginnings of Domestication in the Sahara and the Nile Valley . Rome, ‘L’Erma’ di Bretschneider. Barker, G., Gilbertson, D., Jones, B. and Mattingly, D. (1996) Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume One: Synthesis. Volume Two: Gazetteer and Pottery . Paris, UNESCO Publishing. Barrow, C.J. (1995) Developing the Environment: Problems and Management . London, Longman. Beaumont, P. (1993) Drylands: Environmental Management and Development . London, Routledge. Brotherston, G. (1989) Andean pastoralisrn and Inca ideology. In J.Clutton Brock (ed.) The Walking Larder . 240–55. London, Unwin Hyman. Bureau of Meteorology, Commonwealth of Australia (n.d.) Results of Rainfall Observations Made in South Australia and Northern Territory 1839–1950 . Melbourne, Bureau of Meteorology, Commonwealth of Australia. Cremaschi, M. and di Lernia, S. (1998) (eds) Wadi Teshuinat: Palaeoenvironment and Prehistory in South-western Fezzan (Libyan Sahara) . Florence, Insegna del Giglio. Cullis, A. and Pacey, A. (1992) A Development Dialogue: Rainwater Harvesting in Turkana . London, Intermediate Technology Publications. Ellis, J. (1995) Climatic variability and complex ecosystem dynamics: implications for pastoral development. In I.Scoones (ed.) Living with Uncertainty: New Directions in
The archaeology of drylands
16
Pastoral Development in Africa : 37–46. London, International Institute for the Environment and Development. Fantechi, R. and Margaris, N.S. (1986) (eds) Desertification in Europe. Proceedings of the Information Symposium in the EEC Programme on Climatology Held in Mytilene, Greece, 15–18 April 1984 . Dordrecht, D.Reidel Publishing Company. Fritts, H.C. (1991) Reconstructing Large-scale Climatic Patterns from Tree-Ring Data . Tucson, University of Arizona Press. Gilbertson, D.D. (1996) Explanations: environment as agency. In G.Barker, D. Gilbertson, B.Jones and D.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume One: Synthesis : 291–317. Paris, UNESCO Publishing. Grainger, A. (1990) The Threatening Desert . London, Earthscan. Hall, M. (1987) The Changing Past: Farmers, Kings and Traders in Southern Africa . Cape Town, David Philip. Hall, M. (1996) Archaeology Africa . Cape Town, David Philip. Hannan, L. (1999) Tourist massacre in Bwindi National Forest, Uganda. The Independent , 3 March 1999. Holling, C.S. (1973) Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4:1–23. Houghton, J. (1997) Global Warming: The Complete Briefing . Cambridge, Cambridge University Press, second edition. Hughes, J.D. with Thirgood, J.V. (1982) Deforestation, erosion and forest management in ancient Greece and Rome. Journal of Forest History 26:60–75. IFAD (International Fund for Agricultural Development) (1992) Soil and Water Conservation in Sub-Saharan Africa: Towards Sustainable Production by the Rural Poor . Amsterdam, Free University of Amsterdam, Centre for Development Cooperation Services. Lamb, H.F., Gasse, F., Benkaddour, A., el-Hamouti, N., van der Kaars, S., Perkins, W.T., Pearce, N.J. and Roberts, C.N. (1995) Relations between century-scale Holocene arid events in tropical and temperate zones. Nature 373:134–7. Lean, G. (1994) How stones can hold back the Sahara. The Independent on Sunday , 16 October 1994:16. Meinig, D.W. (1962) On The Margins of the Good Earth: The South Australian Wheat Frontier 1869–1884 . Adelaide, Rigby Limited. Millington, A.C. and Pye, K. (1994) (eds) Environmental Change in Drylands: Beogeographical and Geomorphological Perspectives . Chichester, John Wiley and Sons. Mortimore, M. (1998) Roots in the African Dust: Sustaining the Drylands . Cambridge, Cambridge University Press. Nicholson, S.E. (1994) Rainfall fluctuations in Africa and their relationship to past conditions over the continent. The Holocene 4(2):121–31. Noy-Meir, I. (1978) Grazing and production in seasonal pastures: analysis of a simple model. Journal of Applied Ecology 15:809–35. Olsvig-Whittaker, L.S., Hosten, P.E., Marcus, I. and Schochat, E. (1993) Influence of grazing on sand field vegetation in the Negev desert . Journal of Arid Environments
Living at the margin: themes in the archaeology of drylands
17
24:81–93. Phillips, J.D. (1993) Biophysical feedbacks and the risks of desertification. Annals of the Association of American Geographers 83(4):630–40. Reij, C. (1991) Indigenous Soil and Water Conservation in Africa . London, International Institute for the Environment and Development, Gatekeeper Series no. SA27. Scoones, I. (1995) (ed.) Living With Uncertainty: New Directions in Pastoral Development in Africa . London, International Institute for the Environment and Development. Spooner, B. (1989) Desertification: the historical significance. In R.Huss-Ashmore and S.H.Katz (eds) African Food Systems in Crisis. Part One: Micro-Perspectives : 111– 62. New York, Gordon and Breach. Sutton, J.E.G. (1977) The African Aqualithic. Antiquity 51:25–34. Thomas, D.S.G. and Allison, R.J. (1993) (eds) Landscape Sensitivity . Chichester, John Wiley and Sons. Thomas, D.S.G. and Middleton, N.J. (1994) Desertification: Exploding the Myth . Chichester, John Wiley and Sons. Tiffen, M., Mortimore, M. and Gichuki, F. (1994) More People, Less Erosion: Environmental Recovery in Kenya . Chichester, John Wiley and Sons. Tolba, M.K. and el-Kholy, O.A. (1992) (eds) The World Environment 1972–1992 . London, Chapman Hall. Walker, B.H., Ludwig, B., Holling, C.S. and Peterman, R.M. (1981) Stability of semiarid savannah grazing systems. Journal of Ecology 69:473–98. Warren, A. and Khogali, M. (1992) Assessment of Desertification and Drought in the Sudano-Sahelian Region 1985–1991 . London, International Institute for Environment and Development. Wendorf, F. and Schild, R. (1980) Prehistory of the Eastern Sahara . New York, Academic Press. Wendorf, F., Schild, R. and Close, A.E. (1984) (eds) Cattle-Keepers of the Eastern Sahara: The Neolithic of Bir Kiseiba . Dallas, Southern Methodist University. Wendorf, F., Schild, R. and Close, A.E. (1989) (eds) The Prehistory of the Wadi Kubbaniya . Dallas, Southern Methodist University, two volumes.
2 The dynamic climatology of drylands GREG SPELLMAN
DEFINING DRYLANDS Surprisingly, given that the critical and unifying variable for dryland environments is a shortage of water on a seasonal or longer-term basis, there has been a long-standing difficulty in determining their geographical extent (Beaumont, 1989; Wallen, 1967), though it is generally estimated that hyper-arid, arid and semi-arid lands in total cover a third of the Earth’s land surface (UNEP, 1992; see Fig. 1.1). The absence of significant moisture is manifest in the characteristics of the soils, vegetation and topography. Consequently, Oliver (1973) and Nir (1974) have suggested ways of identifying arid lands by a variety of non-climatic criteria. Straightforward classical approaches create regionalizations using isopleths of climatic elements with respect to associations with vegetation and agricultural conditions, such as the 250 mm rainfall limit as the arid boundary (Oliver, 1981). In contrast, indexing methods delimit regions with differing levels of aridity by the application of objective standard formulas. The best-known classical method is that of Koppen (1931), who defined dryland regions in terms of an annual precipitation and temperature index. Assuming a mean annual temperature of 18°C, his formula gives a maximum precipitation of 640 mm for semi-aridity with summer rainfall, and 360 mm with winter rainfall, and the calculation that drylands occupy about 26 per cent of the total Earth surface, with the desert region covering 12 per cent and semi-desert and steppes the other 14 per cent. The system was criticized by Mather (1974) for failing to consider water supply and having no physical meaning or indication of the atmospheric processes involved. Water-balance models were developed independently by Penman (1948) and Thornthwaite (1948). Penman’s model is more sophisticated and considers turbulent transfer and energy balance approaches. Thornthwaite’s model considers the energy balance alone, using P, the mean annual precipitation (mm), and a calculation of Pe, the mean annual potential evaporation (mm), in the calculation of a moisture index (Im), resulting in: Im=100[P/Pe−1]. In this system arid regions have an index value of under −66.7, whereas a semi-arid region is defined where Im lies between −33.3 and −66.7. The method was criticized by Wallen (1967) for tending to over-estimate water supplies. Other water-balance methods are reviewed by Jones (1997). The definitive map of the spatial distribution of dryland areas produced by UNEP (1992; see Fig. 1.1) divides the globe on degrees of bioclimatic aridity using the values of the ratio P/PET, that is P=the mean annual precipitation (mm) and PET=the mean annual potential evapotranspiration (mm), as calculated by Penman’s formula. Three categories are relevant here: hyperaridity, where the P/ETP ratio is less than 0.05; aridity, from 0.05
The dynamic climatology of drylands
19
to 0.20; and semi-aridity, from 0.20 to 0.45. Some classifications of drylands also include the dry sub-humid regions (0.45
Table 2.1 Regional distribution of world drylands (103km2). (After Le Houerou, 1996) Zone Africa Asia AustralasiaEurope North South Total % America America Hyper-arid 67202773 0 0 31 257 9781 7.5 Arid 50356257 3030 110 815 4451569212.1 Semi-arid 51386934 3090 1052 4194 26452305317.7 Dry sub2687 352 513 1835 2315 207012947 9.9 humid Table 2.2 Rainfall regimes at selected dryland stations. (Data from Pearce and Smith, 1984) Place Altitude (m) Location Annual Total (mm) J F M A M JJ A S ON D Fiya, Chad 225, 18°0′N, 18 0 0 0 0 0 00 18 0 0 0 0 19°10′E Khartoum, 390, 15°37′N, Sudan
32°33′E
157
0 0 0 0 2.5 753 71 185 0 0
The archaeology of drylands
20
Kashgar, 1309, 39°24″N, China 76°07′E
86
15 3 13 5 8 510 8 3 3 5 8
Amman,
777, 31°57′N,
278
69 74 31 155 00 0 0 5 33 46
Jordan
35°57′E 43
3 0 0 0 5 58 8 8 3 3 0
Lima, Peru 120, 12°05′S, 77°03′W
Table 2.3 Estimates of the land area of arid lands using the climate classifications of Koppen (1931), Thornthwaite (1948), Meigs (1953) and UNESCO (1977) Koppen Thornthwaite Meigs UNESCO 12.0 15.3 20.5 19.5 Hyper-arid Semi-arid 14.3 15.2 15.8 13.3 Total 26.3 30.5 36.3 32.8 brought by the Inter Tropical Convergence Zone (ITCZ), so regions with summer rainfall lie on equatorward-margins of drylands, whereas regions with winter rainfall, under the influence of mid-latitude disturbances, are on the poleward sides. Dryland types based on temperature classifications are: tropical deserts exhibiting little change in monthly temperatures (e.g. Somalia); subtropical deserts experiencing considerable temperature changes (e.g. the Thar and Australian deserts); temperate drylands with cold winters (e.g. drylands in Iran, Syria and Mongolia); and cold highland areas (e.g. Tibet). Yet another scheme is that of Thomas (1989): hot arid lands (coldest month temperature 20–30°C); drylands with mild winters (coldest months 10–20°C); drylands with cool winters (coldest months 0–10°C); and drylands with cold winters (coldest month less than 0°C). Despite these varieties of definition, however, the overall purpose of dryland classification is the same: to identify their global significance; to examine the processes that operate to create them; and to assess whether any major changes are occurring. Agnew and Anderson (1992) remark that there is a grave danger that arid lands are treated by water resource managers as homogeneous entities with similar environments and similar problems, when this is clearly not the case.
CLIMATIC CHARACTERISTICS Precipitation By definition, all dryland areas receive low annual precipitation, and in most dryland areas, as rainfall amounts diminish, there is a corresponding increase in variability and unreliability (Le Houerou, 1996). Mean values, therefore, do not adequately describe the true nature of the precipitation regime, because annual totals will show significant year-
The dynamic climatology of drylands
21
to-year departures from long-term norms. Nir (1974), for instance, mentions a rain event frequency of once every eight years at certain sites in the Sahara and once every eighteen years in Peru. The interquartile or 10–90 percentile ranges are more useful indicators (Beaumont, 1989). The variability of rainfall in arid areas is greater than that of temperate regimes, because of the character of the measure used: the coefficient of variation (COV), calculated by the division of the standard deviation by the mean. Areas with low rainfall will inevitably record the highest variation, even though the magnitude of that variability away from the mean is smaller. The variability in absolute terms may not be much greater than that of temperate regions (Cooke and Warren, 1973), but for areas with low rainfall, even small variations are extremely significant (Agnew and Anderson, 1992). When rainfall events do occur in dryland areas, it is often when rainbearing frontal systems or tropical cyclones penetrate the region. Incursions are therefore more frequent at the margins of dryland areas. The usual mechanism in poleward areas is the southward movement of cold ‘upper lows’—areas of cold air in the upper atmosphere that have been cut off from the prevailing westerly circulation under conditions of low zonal flow in the mid-latitude index cycle (Barry and Chorley, 1998). On the equatorward margins, the remnants of tropical cyclones and the seasonal advance of the ITCZ are important, giving very localized, short-lived and often high-intensity rainfall events (Fig. 1.3). Examples in Algeria listed by Barry and Chorley (1998) include 8.7 mm in three minutes (El Golea), 38.5 mm in 25 minutes (Beni Abes) and 46 mm in 63 minutes—though such catastrophic events are not a universal characteristic of drylands (Gordon and Lockwood, 1970). What is certainly typical is the highly localized nature of rainfall (Beaumont et al., 1988; Sharon, 1972, 1981). Synoptic climatological methods have long demonstrated their validity for the analysis of regional rainfall variability (Barry and Perry, 1973; Sweeney and O’Hare, 1992), and to model regional scenarios of climate change (Wilby and Wigley, 1997). A weather-type indexing method originally developed in an investigation of rainfall variability in Egypt by El Dessouky and Jenkinson (1975) has been adapted for investigating the role of atmospheric circulation pattern on rainfall in dry areas of Spain: surface index values can be correlated with rainfall amounts and statistically significant associations have been identified (Spellman, 2000). There are therefore distinct opportunities for the analysis of historic rainfall events and drought. Drought The World Meteorological Organization (1975) defines drought as “a deficit of rainfall in respect to the long term mean, affecting a large area for one or several seasons or years, that drastically reduces primary production in natural ecosystems and rain-fed agriculture”. Drought is commonly defined in terms of its impacts rather than its causes: hence the terms ‘agricultural drought’ and ‘hydrological drought’ have been proposed (Smith, 1992). Drought and aridity are not the same thing: aridity refers to a negative ratio between mean annual rainfall and mean annual potential evapotranspiration. The degree of aridity is inversely related to the magnitude of this ratio, but drought is more or less related to aridity because arid regions experience frequent droughts. Drought impacts are worst in dryland areas because the low mean annual variability of rainfall is
The archaeology of drylands
22
associated with high variability, and drought duration is greater—the drought in the Sahel began in the late 1960s and continues, rainfall still not reaching the 1931–1960 mean (Hulme and Kelly, 1993; Morel, 1992; Nicholson, 1993; Nicholson et al., 1988). Temperature It is far harder to generalize about the thermal regimes of dryland areas. Annual temperature ranges are greatly affected by altitude and the distance from the sea, but in general, high summer temperatures as a consequence of high radiation loads are common to all regions (e.g. Fig. 2.1). In the Sahara, maximum average daily temperatures of more than 45°C are recorded in the interior, and July temperatures of 37.5°C are recorded elsewhere except in the highlands.
Figure 2.1 Thermal regimes in two dryland locations: Aswan, Egypt (112m 24º 02′ N, 32° 53′ E), and Jacobabad, Pakistan (57m 28° 17′ N, 68° 29′ E) Relative humidity Atmospheric moisture content is typically very low above dryland areas (Table 2.4): dryness is a response partly to the lack of local evapotranspiration and partly to the lack of horizontal moisture advection. Coastal drylands have high humidities—60 per cent in parts of Western Australia, for example, compared with under 20 per cent 150 km inland. Contrasting humidity regimes are shown in Figure 2.2: Walvis Bay, Namibia, in coastal drylands, has high humidity; In Saleh, in the interior of Algeria, has similarly low rainfall but a much drier atmosphere; Damascus, Syria, and Timbuktu, Mali, respectively north and south of the subtropical anticyclones, have annual regimes that mirror each other, depending on the timing of the rainy season.
The dynamic climatology of drylands
23
Wind Persistent strong winds are common in drylands, often a consequence of extensive flat areas with little vegetation cover to disturb air movement in the boundary layer. Wind is a major agent of erosion, lifting significant quantities of dust from the dry soil; hazy atmosphere with low visibilities is commonplace. Wind has an important indirect climatic role, because the amount of dust influences the surface energy balance, aiding the process of desertification (Le Houerou et al., 1993).
ATMOSPHERIC PROCESSES CAUSING ARIDITY Condensation of moisture in clouds occurs when moist air is cooled to the point whereby saturation is reached. This occurs through ascent, mixing, radiation-cooling, or contactcooling with a colder underlying surface. The clouds that form must then grow to a sufficient depth in order that drops of water can grow to a size to overcome air resistance and fall to the ground. The
Table 2.4 Mean relative humidity at various isobaric levels for radiosonde stations in the Sahara and the Arabian peninsula. (After Lockwood, 1974) Isobaric level January July 850 33 16 Fort Trinquet (25°14′N, 11°35′W 360m)
Aulef-el-Arab (27°04′N, 1°06′E 275m)
700 500 850
22 15 28
22 31 14
Habbaniya (33°22′N, 43°34′E, 45m)
700 500 850
22 15 53
17 21 15
700 500
33 26
18 7
The archaeology of drylands
24
Figure 2.2 Mean monthly relative humidity at four locations, with their mean annual rainfall (mm) shown in parentheses atmospheric processes that promote aridity are thus those that (1) result in a complete lack of atmospheric moisture, or (2) inhibit cooling of the air through the prevention of convection or the creation of inversions, or (3) reduce humidity by warming the atmosphere. Climatological processes that cause aridity operate at both global and regional scales. Thompson (1975) outlines four main processes that help to explain the distribution of arid lands. The first and most important (Hills, 1966) is atmospheric subsidence on the poleward side of the subtropical anticyclones. Aridity results as descending air is slowly compressed and subsequently adiabatically warmed, leading to a dry, stable atmosphere (Fig. 2.1). Subsidence within the anticyclones does not extend right to the surface, since normally the warm, dry, subsiding air is insulated from the surface by a shallow layer of relatively cool air. The properties of this boundary layer can be completely different to that of the sinking air, and are usually maintained by a source outside that of the main anticyclone. If the air forming this surface layer originated over the sea, it may be moist and contain layer cloud, which can result in light rain or drizzle. Sinking air will also prevent significant depth of thermal convection, despite high radiation receipt and subsequent strong surface heating under clear skies. Dryland areas are centred beneath the subtropical anticyclones in both hemispheres. Wind direction is also important: air flowing over the interior of a continent has a reduced opportunity to absorb moisture at its base, so strong stability and low humidities will develop in the lower levels. In the northern hemisphere, dry northeasterly winds (the returning flow of the Hadley Cell circulation) contribute to much of the aridity of
The dynamic climatology of drylands
25
Southwest Asia and the Middle East. The third factor is topography: natural obstacles across the path of prevailing winds can cause aridity on their leeward side. Thus as moist air is forced to rise over a mountain range (Fig. 2.3), air cools at the dry adiabatic lapse rate (A to B) until saturation is reached and then at the moist adiabatic lapse rate (B to C) until the cloud top; in the lee of the mountain range, the descending air warms at the dry adiabatic lapse rate (C to D) and will be warmer than the ascending air at each corresponding altitude. The air stream will arrive at the other side as a dry, desiccating wind as, for example, occurs on the leeside of the Sierra Nevada in North America or the Andes in South America. The fourth process is cold ocean currents. Onshore winds that pass over cool, equatorward-flowing, ocean currents close to the shore will be rapidly cooled in the lower layers (up to 500 m). This induces atmospheric stability, which then reduces the potential for rainfall production by promoting thin extensive sheets of stratiform cloud cover and persistent coastal mists and fogs. At higher altitudes the air will be warm, thus creating a strong inversion that further prevents convection. Examples where this effect is important include the Atacama and Namib deserts, under the influence respectively of the Peru and Benguela currents.
Figure 2.3 The rainshadow effect leading to aridity Source: After Agnew and Anderson, 1992
SURFACE ATMOSPHERIC CIRCULATION The low latitudes are dominated by the meridional circulation of the Hadley Cells, a thermally driven rising limb of air in the equatorial zone, a poleward-moving flow in the
The archaeology of drylands
26
upper atmosphere, a sinking limb in the region of the subtropics, and a returning trade wind flow at the surface that converges with corresponding winds from the opposite hemisphere at the ITCZ (Fig. 2.4). On the poleward side of the upper atmosphere, above the return branch of the cell, is the subtropical jet, a relatively narrow band of highvelocity westerly winds encircling the Earth. The Hadley Cells exhibit marked seasonal variation in intensity, geographical extent and latitudinal position. Subtropical anticyclones Between about 20° to 40° mean surface pressure patterns are dominated by a discontinuous belt of subtropical high (STH) pressure areas, broadly elliptical in shape and oriented in an east-west direction. On average, they dominate the ocean basins in these latitudes. The geographical positions of the centres of the subtropical highs fluctuate. In winter in the southern hemisphere, they intensify and spread over the adjacent continental areas, and an almost closed belt of high pressure can be formed. In summer, thermally produced low-pressure centres over land masses (Australia, southern Africa) disrupt the pattern. In the northern hemisphere, higher central pressure is exhibited in summer, which is a time when the STHs also show their greatest extent. The high-pressure centres display a regular movement. During the winter season, STH centres exhibit equatorwards movement, which is reversed in summer. A change of only one third of a degree of latitude (about 35 km) in the position of the Atlantic high (almost unobservable) causes a one degree
Figure 2.4 The Hadley Cell circulation of the tropical northern hemisphere Source: After Musk, 1988 change in latitude in the position of the ITCZ, with an immense effect on rainfall in the Sahel (Oliver, 1981). For reasons that are unclear (Hastenrath, 1985), the STH centres also migrate longitudinally—in winter in the northern hemisphere all subtropical highs are centred over the eastern regions of their respective ocean basins, whereas during summer they migrate to the west.
The dynamic climatology of drylands
27
Changes in the intensity of STHs (as measured by sea level pressure) are clearly displayed in the southern hemisphere. The South Pacific anticyclone tends to be strongest in the southern hemisphere spring. Jones (1991) has shown that the centre of the South Pacific anticyclone has declined in strength over the period 1951–1985, yet the northern flanks have strengthened. In the northern hemisphere, temporal variations in the STH intensity have also been identified: the North Atlantic anticyclone, for instance, showed a significant increase in surface pressure between 1946 and 1987 (Inoue and Bigg, 1995). Subtropical highs are generally asymmetrical in structure, with highest pressures in the east at the Earth’s surface and maximum pressure to the west at altitude. Consequently, the circulation around the centre is not parallel to the Earth’s surface but slopes gently towards the west, with subsidence dominating the eastern half and rising air currents more frequent in the west. Thus western air masses are more unstable and humid, whereas in the east conditions are generally cloud-free or, if clouds are present, they display limited vertical development, like thin stratocumulus. Origins of the subtropical anticyclones Classical dynamic explanations of the origins of the subtropical anticyclones attribute their existence to the ‘piling up’, or convergence, of the poleward-flowing upper air ‘antitrades’ at about 20°, inducing a downward movement of air and high pressure at the surface. Alternatively, the main cause may be the movement of polar air (McIlveen, 1992). As a result of changes in the Coriolis Force with latitude, anticyclonic cells near the polar front have a tendency to move equatorwards, while low-pressure centres generally migrate towards the poles (Rossby, 1947). These travelling cold anticyclones frequently rejuvenate the subtropical highs. Polar outbreaks would therefore prefer the eastern parts of the ocean basins, where cold ocean currents prevail and where friction along the continental coasts gives a strong meridional influence on these movements. Pulses in the intensity of subtropical highs on a daily time scale might be explained by this idea (McGregor and Nieuwolt, 1998). In addition, the interaction between cold polar outbreaks and surface ocean currents maintains the observed higher pressure over the eastern oceans at low levels and the stronger development of STHs over the southern oceans. Thermal explanations have also been proposed, involving cooling in the upper air and cooling at the Earth’s surface. Upper air cooling will occur when air in the upper poleward-moving branches of the Hadley Cells loses heat by long-wave radiation to space. The air thus becomes progressively denser, subsides, and leads to high pressure at all levels. Cooling at the Earth’s surface is seen as a response to cold ocean currents and the cool continental land areas in winter. These features may correlate with the cellular pattern of the subtropical highs and their extension over continents. An explanation of the existence of subtropical highs can also be found in the consideration of both thermal and dynamic mechanisms. According to McIlveen (1992), the Coriolis effect may impose dynamic constraints on the flow of air in anticyclonic systems: once upper level convergence occurs, the Coriolis Force prohibits the outflow of air at surface, leading to atmospheric mass build-up in the anticyclonic centres, with the warming effect of air subsidence reducing the vertical pressure gradient so that air pressure falls more slowly with increasing altitude. Isobaric surfaces therefore tend to
The archaeology of drylands
28
‘dome’, rising to greater heights than in the surrounding air, which produces deep and warm anticyclonic systems. Trade winds Between the subtropical highs and the ITCZ, the low-level circulation of the atmosphere is dominated by the persistent easterly winds known as the ‘trade winds’. These have a distinct three-layer structure, the heights of which increase towards the equator (Fig. 2.5). The height and intensity of the inversion layer will have a strong influence on precipitation mechanisms. These features are generally dependent on latitude, or distance from the STH centre, yet lower, more intense, inversions (and subsequent weaker convection) are associated with the east side of ocean basins: on the coast of West Africa, for example, the intensity can be between 5–8°C, markedly reducing rainfall potential.
Figure 2.5 The structure of the trade wind atmosphere The Inter-tropical Convergence Zone At the equator flank of dryland areas, in regions classified as semi-arid, rainfall is governed by the seasonal fluctuations in the position of the Inter-tropical Convergence Zone (ITCZ). This is commonly perceived as a belt of low pressure encircling the globe where the two Hadley circulations meet, but Waliser and Gautier (1993) have identified seven separate ITCZ zones (Table 2.5) differentiated by structure and behaviour. In the northern hemisphere, dry conditions are associated with hot continental tropical air, which moves in behind the ITCZ as it migrates southwards during the winter. In Africa, at its most southerly extent, in January or February, the ITCZ lies at about 8° north of the equator in the west but about 15–20° south of it in the east. This is the dry season of the north. The ITCZ moves north during the northern summer, but the extent of the progression (up to about 20°N) shows considerable year-to-year variability, with latitudinal departures of up to 6° (800 km) for some regions and up to 2° for the global
The dynamic climatology of drylands
29
average. These departures, which can last from 3 to 18 months, may produce lengthened drought periods.
UPPER AIR CIRCULATION Explanations of the spatial variability of precipitation conditions can be afforded by reference to upper atmospheric flow, particularly the position and intensity of the subtropical jet stream. During the winter dry season, when anticyclonic conditions prevail over North Africa, the jet stream becomes convergent towards the equator and produces a downward shift of air to feed high pressure at the surface. In the summer months, upper air divergence results from the convergence of southwesterly and northeasterly winds at the surface. This tends to draw in moist air and increase the likelihood of precipitation. Low rainfall totals in dryland areas are explained by weak easterly jet streams associated with weaker circulation in the middle and upper troposphere.
Table 2.5 Seven ITCZ zones (after Waliser and Gautier, 1993) Zone Longitude limits (°) Africa 10–40E Indian 60–100E West Pacific 100–150E Central Pacific 160E–160W East Pacific 100–140W South America 45–75W Atlantic 10–40W Subtropical Westerly Jet Classical models commonly portray the upper air poleward-moving section of the Hadley Cell as a meridional flow (the ‘anti-trades’), yet in reality this will be strongly redirected by the Coriolis Force as soon as it moves away from the equator, resulting in a narrow band of high-velocity westerly winds known as the Subtropical Westerly Jet, which is found on average at around 30° from the equator in both hemispheres. Palmen and Newton (1969) describe the SWJ as a persistent long-wave pattern encircling the globe with wave troughs at 20°W, 150°W and 90°E and wave crests at 70°W, 40°E and 150°E. Maximum wind speeds of up to 100 m per second are found in the vicinity of the wave crests. The mean position of the SWJ and the year-to-year variability of the crests and troughs influence precipitation patterns in the low latitude regions. Flow reaches its maximum intensity in the winter months when the pole-to-equator thermal gradient is greatest: the core moves towards the poles as the Hadley circulation strengthens. At 200 mb, the SWJ will lie over the poleward flanks of the STHs. If individual high-pressure cells contract away from one another, as meanders develop in the jet between them, the troughs can extend southwards to interact with low-level (850 mb) underlying tropical easterlies (Fig.
The archaeology of drylands
30
2.6). In the central parts of the Sahara, rainfall occurs under the variable northward penetration of the West African monsoon trough, which allows tongues of moist southwesterly air to travel comparatively far north, producing short-lived low-pressure centres. Low-pressure centres then move north along the meander trough, though they are often ‘rained out’ when they reach the central Sahara. Related to the SWJ but far less common is the southward movement of Mediterranean cold fronts: Barry and Chorley (1998) noted such an event in December 1976 in southern Mauretania, which yielded 40 mm of precipitation. Tropical Easterly Jet The Tropical Easterly Jet extends from Southeast Asia (80°E) to North Africa (50°E), at approximately latitude 15°N. It spirals out clockwise from the subtropical high pressure centres and flows in the northern summer months (June to September), because it is related to the seasonal heating cycle (Hastenrath, 1985). The strongest intensity is at about 15 km altitude, where maximum speeds are in the region of 40 m per second; there is a second, weaker, easterly flow at about 5 km. The TEJ owes its existence to the strong surface heating in summer over the land masses in Africa and Asia, where very intense
Figure 2.6 The interaction between the subtropical westerly flow and the tropical easterlies leading to the creation of Saharan depressions, which move eastwards along the trough axis Source: After Barry and Chorley, 1998
The dynamic climatology of drylands
31
heat lows promote the ascent of air to the upper atmosphere. Mass convergence in the upper high-pressure system is so intense that pressure surfaces bulge upwards, creating an atmospheric thickness difference between the subtropics and the equatorial and mid latitudes, resulting in a reversal in the normal equator-to-subtropics temperature gradient, with warmer temperatures recorded in the subtropical upper atmosphere (McGregor and Nieuwolt, 1998). Convergence in the jet over Africa induces subsidence over the Sahel and may be responsible for preventing the advancement of the West African monsoon rains. Generally, rainfall is greatest north of the jet entrance in the southern Asian region and south of the jet exit in the West African region. West African Mid-Tropospheric Jet The West African Mid-Tropospheric Jet is associated with rainfall patterns in African drylands and arises in response to mid-tropospheric temperature gradients between the warm Sahara desert area and the cool waters of the Gulf of Guinea off West Africa. It has its core at 15°N at about 4,500 m (600 mb) and is located in a region south of the central area of anticyclonic outflow. Maximum intensity occurs in the northern winter, when hemispheric temperature gradients are steepest: flow can reach 10 m per second. Rainfall occurs on the equator side of the jet above the Saharan heat low. East African Low-Level Jet The East African Low-Level (850 mb) Jet has a wandering parabolic course over the western Indian Ocean (Fig. 2.7). It exists in all months, but its
The archaeology of drylands
32
Figure 2.7 The monthly progression of the East African LowLevel Jet Core Source: After McGregor and Niewolt, 1998 greatest development is related to the onset of the African-Asian monsoon circulation. In winter, it is confined to the southern hemisphere, but it is an integral part of the northern summer monsoon circulation in the African-Indian area. Maximum coolness, moisture and cloudiness coincide with the jet core and its eastern regions over the coast of eastern Africa, where maximum ascent of air occurs (Kamara, 1986). Minimum cloudiness occurs above the jet core and to the west in the direction of the footslopes of the East African plateau, where the air is descending, creating the warmest, driest and most stable air (Findlater, 1972). The core of this jet occurs at about 1,500m, where velocities can
The dynamic climatology of drylands
33
reach 25–50 m per second. Branches of the jet can penetrate inland over eastern Africa through topographic breaks in the East African plateau. It has been related to rainfall occurrences in the northern parts of Ethiopia and the tracks of so-called Sudano-Sahelian depressions (Fig. 2.8).
PREVAILING WINDS AND MOUNTAIN BARRIERS Where topographical factors are added to those caused by the general circulation, aridity is greatly increased and it is in these areas that the most severe deserts are found. Moisture available for precipitation is trapped in a shallow layer beneath the subtropical inversion. The depth of this moist layer varies, but if a mountain barrier projects throughout this moist layer, it interrupts the surface flow and the surface moist layer will not penetrate behind the mountain range. Even if the range does not completely block the moist layer, the reduction in moisture advection to the lee of the range can still be substantial, of the order of 60–70 per cent (Lockwood, 1974). Dryness can further be enhanced by subsidence of air near the inversion down the lee slopes. Such mountainenclosed inland basins can be extremely arid; Death Valley in California is a prime example.
OCEAN TEMPERATURES Ocean temperature has a considerable influence on climate, particularly in coastal regions: cool ocean currents moving towards the equator stabilize the atmosphere and reduce atmospheric instability. When cold water along the equator is well developed, the air above will be too cold even to take part in the ascending motion of the Hadley Cell circulation. Along the coast of Peru, the surface moist layer is less than 800 m deep and normally only drizzle will fall from a deck of stratus. The coasts of South America and Southwest Africa are sheltered respectively by the Andes and Namib escarpment from the dynamically stable easterly trades, allowing shallow tongues of cold air to roll in from the west. These are capped by strong inversions at c.600–1,500 m, which reinforce the trade wind inversions, precluding the development of intense convective cells, except where orographic ascent occurs. Precipitation from fog may also result from ocean currents. When rain does fall, it is on those rare occasions when large-scale pressure changes prevent sea breezes and fog.
DESERTIFICATION Desertification has been defined in various ways, recently (McGregor and Nieuwolt, 1998) as the process by which dryland conditions are brought into
The archaeology of drylands
34
Figure 2.8 The tracks of Sudano-Saharan depressions over the Sahara Source: After Barry and Chorley, 1998 areas where such conditions did not previously exist. According to Le Houerou (1996), if hyper-arid zones are excluded (as not susceptible to further desertification), 38 per cent of drylands can be described as desertified, which is 16 per cent of the overall land area (Le Houerou, 1996; Table 2.6). However, it remains debatable how extensive desertification is or how fast it is proceeding (Thomas and Middleton, 1994). Two major factors are involved in the desertification process (though their relative magnitudes are unknown): human activities and drought as a consequence of climatic variability. Observing that desertification occurred in the Sahel during the 1950s and early 1960s in spite of the fact that rainfall was well above the long-term average, Le Houerou (1996:146) concluded that ‘desertification may therefore result from land abuse alone’. Most meteorological models for dryland expansion or the occurrence of episodes of drought point ultimately to local changes to the surface energy balance or to large-scale shifts in atmospheric circulation. ‘Human impact’ theories generally focus on areas with sparse vegetation, which will commonly have surface air temperatures that are lower than their surroundings due to the increased amounts of surface reflectivity of solar radiation (Otterman, 1974). Charney (1975) suggested a mechanism (‘biogeophysical feedback’) whereby over-grazing of desert margins can increase surface albedo, decreasing the total energy absorbed at surface and reducing thermal convection, thereby enhancing stability and reducing rainfall potential; this in turn provides a positive feedback because lower moisture availability leads to even less surface vegetation amounts. Since much dryland rainfall comes from re-evaporated rainfall, not from advected moisture from elsewhere,
The dynamic climatology of drylands
35
declining soil moisture may intensify drought conditions (Hulme, 1989; Laval, 1986), though the theory is disputed (Courel et al., 1984; Idso, 1977; Williams and Balling, 1994). A large artificial body of water, such as Lake Nasser in Egypt, does not increase rainfall in the Nubian desert, despite its low albedo, which extends over an area of 5,000 km2 and which contrasts with the very high reflectivity of the surrounding desert (Le Houerou, 1996).
Table 2.6 The extent and severity of desertification (after Le Houerou, 1996) Region Light* Moderate* Strong* Severe* Total** area % area % area % area % area % Africa 1,180 9 1,272 10 707 5.0 35 0.2 12,860 56.0 Asia 1,567 9 1,701 10 430 3.0 5 0.1 16,718 42.0 Australasia 836 13 24 4 11 0.2 4 0.1 6,633 32.0 North America 134 2 588 8 73 0.1 0 0.0 7,324 50.0 South America 418 8 311 6 62 0.2 0 0.0 5,160 29.0 Total 4,273 8 4,703 9 1,301 2.5 75 0.1 51,691 39.7 *Area desertified in 103km2; % area desertified as % of total drylands (where drylands = arid+semi arid+dry subhumid) **Total drylands; %=percentage of desertified areas in the non hyper-arid drylands Bryson and Murray (1977) suggested that surface desiccation would lead to large amounts of soil particles being entrained and then lifted aloft by the wind, increasing the atmospheric albedo, cooling the air aloft and causing it to subside, warm adiabatically and form an inversion, hence preventing convection and cloud development. They suggested that this was illustrated by the Rajputana Desert on the borders of India and Pakistan, where extreme dustiness stifles rain processes even though atmospheric humidity is as high as in humid tropical forests. Large-scale climatic explanations for very dry episodes have recently focused on teleconnections. Of importance are the possible impacts of anomalous patterns in environmental variables, particularly sea surface temperature anomalies (SSTAs), which influence the flux of moisture and sensible heat at the ocean-atmosphere interface at locations geographically remote from the region under investigation. One well-known example is the association between dry episodes in the subtropics in the mid-twentieth century and El Niño (ENSO) events. Attempts have been made to link SSTAs in the tropical Atlantic to rainfall in the Sahel. Owen and Ward (1989) have linked recurring SSTA patterns to notably wet and dry conditions in sub-Saharan Africa. Another example of teleconnections is suggested by Gray (1990), who identified a positive association between rainfall patterns in West Africa and the frequency of intense hurricanes reaching the Atlantic coast of the United States. During the period of drought in the western Sahel (1979–1987), there was a mean annual incidence of only fifteen hurricane days in the Atlantic basin compared with thirty per year in the wetter phase (1947 to 1969). ENSO events in the Pacific have been seen to influence some drought events: for
The archaeology of drylands
36
instance, the strong 1982–1983 ENSO showed good correlations with drought in Australia (Nicholls, 1987), Indonesia (Malingreau, 1987) and western South America (Serra, 1987). In other areas, the relationship was dubious or had very low statistical significance, for example northeast Brazil (Gasques and Magalhes, 1987) and southern Africa (Nicholson et al., 1988). No relation exists between the present twenty-five years of drought in the Sahel and ENSO (Glantz, 1987), although there is a clear link between the drought and positive SSTAs in the Gulf of Guinea, which are in turn related to the Benguela current. There seems to be a South Atlantic Oscillation (SAO), comparable to ENSO, with many similarities between the Humboldt and Benguela currents, their upwellings and the generation of coastal deserts. In the Mediterranean basin, the history of drought does not seem to be related to ENSO events: ENSO events occur at regular intervals of about 6.4 years, yet Mediterranean droughts are totally acyclical and unpredictable, especially in North Africa and the Near East (Le Houerou, 1996). A considerable amount of work has been carried out on this subject (e.g. Folland et al., 1986; Kane, 1999; Kiladis and Diaz, 1989). Trenberth (1993) describes El Niño as having ‘different flavours’. Consequently, finer classifications have been attempted: Kane (1999), for instance, has identified ‘unambiguous’ ENSO events in which the Tahiti-Darwin sea level pressure minima occur in the middle of the calendar year. It is these events that have more impact on drought conditions elsewhere. Some General Circulation Models (GCMs) have predicted a slight increase in rainfall variability, others a decrease; some indicate an increase in winter rain and a decrease in summer precipitation, others forecast the opposite (Williams and Balling, 1994). Commonly the resolution for rainfall predictions is very coarse. Le Houerou (1996) concludes that, in view of the fact that there have been no significant observed trends in rainfall in any dryland area, no change must be assumed for the not too distant future. In contrast to the predictions about rainfall, however, GCMs agree (at a 50 per cent confidence level) that the twenty-first century is likely to be characterized by an increase in temperature of 2–3°C in the subtropics and 1–2°C in the tropics. Statistical analysis of temperature and mean annual evapotraspiration (PET) shows that each degree of temperature corresponds to 72 mm of PET a year, using the Penman equation (Le Houerou, 1996). A temperature rise of 1–3°C would therefore correspond to a PET increase of 72–232 mm a year—a significant increase in climatic aridity. Furthermore, effects on the movement of the ITCZ and patterns in the westerlies will have an impact on the regime in semi-arid regions at the desert margin. This increase in aridity can only enhance the current expansion of the dryland areas.
CONCLUSION This climatologically-based analysis has emphasized that drylands should not be seen as homogeneous entities with broadly similar environments. Many different types of drylands can be recognized. The reasons for the shortage of precipitation are many and complex. The physical processes inherent in the maintenance of drylands involve synergies and subtleties at a variety of time and spatial scales. Similarly, it is clear that difficult problems remain to be resolved before the magnitude and significance of human-
The dynamic climatology of drylands
37
environment—climate interactions in drylands today can be fully elucidated, let alone those in the distant past.
REFERENCES Agnew, C. and Anderson, E. (1992) Water Resources in the Arid Realm . London, Routledge. Barry, R.G. and Chorley, R.J. (1998) Atmosphere, Weather and Climate . London, Routledge. Barry, R.G. and Perry, R.J. (1973) Synoptic Climatology: Methods and Applications . London, Methuen. Beaumont, P. (1989) Drylands: Environmental Management and Development . London, Routledge. Beaumont, P., Blake, G.H. and Wagstaff, J.M. (1988) The Middle East: A Geographical Study . London, Fulton. Bryson, R.A. and Murray, T.J. (1977) Climates of Hunger: Mankind and the World’s Changing Weather . Madison, University of Wisconsin Press. Bullock, P. and Le Houreou, P. (1996) Land degradation and desertification. In Climate Change 1995: Impacts, Adaptations and Mitigations of Climate Change. Scientific and Technical Analysis : 171–90. Cambridge, Cambridge University Press, Intergovernmental Panel of Climate Change. Charney, J. (1975) Dynamics of deserts and drought in the Sahel. Quarterly Journal of The Royal Meteorological Society 101:193–202. Cooke, R.U. and Warren, A. (1973) Geomorphology in Deserts . London, Batsford. Courel, M.F., Kandel, R.S. and Rasool, S.I. (1984) Surface albedo and the Sahel drought. Nature 307:528–31. El Dessouky, T.M. and Jenkinson, A.F. (1975) An Objective Daily Catalogue of Surface Pressure, Flow, and Vorticity Indices for Egypt and its Use in Monthly Rainfall Forecasting . Bracknell, Meteorological Office, Synoptic Climatology Branch Memorandum 46. Findlater, J. (1972) Aerial explorations of the low level cross equatorial current over eastern Africa . Quarterly Journal of the Royal Meteorological Society 98:274–89. Folland, C.K.Palmer, T.N. and Parker, D.E. (1986) Sahel rainfall and worldwide sea surface temperatures. Nature 320:602–7. Gasques, J.G. and Magalhes, A.R. (1987) Climatic anomalies and their impact in Brazil during the 1982–83 ENSO event. In M.Glantz, R.Katz and M.Krenz (eds) The Societal Impacts Associated with the 1982–83 Worldwide Climate Anomalies : 30–6. Boulder CO, National Center for Atmospheric Research. Glantz, M. (1987) Impacts of the 1982–83 climate anomalies in the West African Sahel. In M.Glantz, R.Katz and M.Krenz (eds) The Societal Impacts Associated with the 1982–83 Worldwide Climate Anomalies : 62–4. Boulder CO, National Center for Atmospheric Research. Gordon, A.H. and Lockwood, J.G. (1970) Maximum one day falls of precipitation in Tehran. Weather 25:2–8.
The archaeology of drylands
38
Gray, W.M. (1990), Strong association between West African rainfall and US landfall of intense hurricanes. Science 249:1251–6. Hastenrath, S. (1985) Climate and Circulation of the Tropics . Dordrecht, D.Reidel. Hills, E.S. (1966) Arid Lands . London, Methuen. Hulme, M. (1989) Is environmental degradation causing drought in the Sahel? An assessment from recent empirical research. Geography 74:38–46. Hulme, M. and Kelly, M. (1993) Exploring links between desertification and climate change. Environment 35:4–11, 39–45. Idso, S.B. (1977) A note on some recently proposed mechanisms of the genesis of deserts. Quarterly Journal of the Royal Meteorological Society 103:369–70. Inoue, M. and Bigg, G.R. (1995) Trends in wind and sea level pressure in the tropical Pacific Ocean for the period 1950–1979. International Journal of Climatology 15: 35– 52. Jones, J.A.A. (1997) Global Hydrology: Processes, Resources, and Environmental Management . Harlow, Longman. Jones, P.D. (1991) Southern hemisphere sea level pressure data: an analysis and reconstruction back to 1951 and 1911. International Journal of Climatology 11: 585– 608. Kamara, S.I. (1986) The origins and types of rainfall in West Africa. Weather 41: 48–56. Kane, R.P. (1999) Rainfall extremes in some selected parts of central and South America. ENSO and other relationships re-examined. International Journal of Climatology 19:423–55. Kiladis, G.N. and Diaz, H.F. (1989) Global climatic anomalies associated with extremes of the Southern Oscillation. Journal of Climate 2:1069–90. Kodama, Y. (1992) Large scale common features of subtropical precipitation zones (the Baiu Front, The South Pacific Convergence Zone, the South Atlantic Convergence zone). Part 1—Characteristics of the Subtropical frontal zones . Journal of the Meteorological Society of Japan 70:813–35. Kodama, Y. (1993) Large scale common features of subtropical precipitation zones (the Baiu Front The South Pacific Convergence Zone, the South Atlantic Convergence zone). Part II—Conditions for generating the subtropical convergence zones. Journal of the Meteorological Society of Japan 71:581–610. Koppen, W. (1931) Die Klimate der Erde . Berlin. Lamb, H.H. (1982) Climate History and the Modern World . London, Routledge. Laval, K. (1986) General circulation model experiments with surface albedo change. Climatic Change 9:91–102. Le Houerou, H.N. (1977) Biological recovery vs desertization. Economic Geography 53:413–20. Le Houerou, H.N. (1996) Climate change, drought, and desertification. Journal of Arid Environments 34:133–85. Le Houerou, H.N., Popov, G.F. and See, L. (1993) Agrobioclimatic Classification of Africa . Rome, Food and Agriculture Organization, Agrometeorology Series Working Paper No. 6. Lockwood, J.G. (1974) World Climatology: An Environmental Approach . London, Edward Arnold.
The dynamic climatology of drylands
39
Malingreau, V.P. (1987) The 1982–83 drought in Indonesia. Assessment and monitoring. In M.Glantz, R.Katz and M.Krenz (eds) The Societal Impacts Associated with the 1982–83 Worldwide Climate Anomalies : 11–18. Boulder CO, National Center for Atmospheric Research. Mather, J.R. (1974) Climatology: Fundamentals and Applications . New York, McGrawHill. New York. McGinnies, W.G. (1988), Climatic and biological conditions of arid lands: a comparison. In E.E.Whitehead, C.F.Hutchinson, B.N.Timmerman and R.G.Vardy (eds) Arid Lands Today and Tomorrow : 61–8. Boulder CO, Westview Press. McGregor, G.R. and Nieuwolt, S. (1998) Tropical Climatology . Chichester, John Wiley and Sons, second edition. McIlveen, R. (1992) Fundamentals of Weather and Climate . London, Chapman Hall. Meigs, P. (1953) World distribution of arid and semiarid homoclimates. UNESCO Arid Zone Program 1:203–10. Middleton, N.J. (1991) Desertification . Oxford, Oxford University Press. Morel, R. (1992) Atlas Agroclimatique de Pays de la Zone de CILSS . Niamey, AGRHYMET. Musk, L.F. (1988) Weather Systems . Cambridge, Cambridge University Press. Nicholls, N. (1987) The El Nine/Southern Oscillation phenomenon. In M.Glantz, R.Katz and M.Krenz (eds) The Societal Impacts Associated with the 1982–83 Worldwide Climate Anomalies : 2–10. Boulder CO, National Center for Atmospheric Research . Nicholson, S.E. (1993) An overview of African rainfall fluctuations of the last decades. Journal of Climate 6:1463–6. Nicholson, S.E., Jeeyong, K. and Hoopingarner, J. (1988) Atlas of African Rainfall and its Annual Variability . Tallahassee, Florida State University. Nir, D. (1974) The Semi-Arid World . London, Longman. Oliver, J.E. (1973) Climate and Man’s Environment . Chichester, John Wiley and Sons. Oliver, J.E. (1981) Climatology. Selected Applications . London, Edward Arnold. Otterman, J. (1974) Baring high albedo soils by over-grazing. Science 86:531–3. Owen, J.A. and Ward, M.N. (1989) Forecasting Sahel rainfall. Weather 44:57–64. Palmen, E. and Newton, C.W. (1969) Atmospheric Circulation Systems . New York, Academic Press. Pearce, E.A. and Smith, C.G. (1984) World Weather Guide . London, Hutchinson. Penman, H. (1948) Natural evaporation from open water, bare soil, and grass. Proceedings of the Royal Society A193:120–45. Rossby, C.G. (1947) On the general circulation of the atmosphere in the middle latitudes. Bulletin of the American Meteorological Society 28:255–80. Serra, R.B. (1987) Impact of the 1982–83 ENSO on the southeastern Pacific fisheries, with emphasis on Chilean fisheries. In M.Glantz, R.Katz and M.Krenz (eds) The Societal Impacts Associated with the 1982–83 Worldwide Climate Anomalies : 24–9. Boulder CO, National Center for Atmospheric Research. Sharon, D. (1972) The spottiness of rainfall in a desert area. Journal of Hydrology 17: 161–75. Sharon, D. (1981) The distribution in space of local rainfall in the Namib desert. Journal of Climatology 1:69–75.
The archaeology of drylands
40
Smith, K. (1992) Environmental Hazards: Assessing Risk and Reducing Disaster . London, Routledge. Soliman, K.H. (1953) Rainfall over Egypt. Quarterly Journal of the Royal Meteorology 79:389–401. Spellman, G. (2000) An objective weather type method for the Iberian peninsula. Weather (in press) Sweeney, J.C. and O’Hare, G.P. (1992) Geographical variations in the precipitation yields and circulation types in Britain and Ireland. Transactions of the Institute of British Geographers 17:448–63. Thomas, D.G. (1989) (ed.) Arid Zone Geomorphology . London, Bellhaven Press. Thomas, D.G. and Middleton, N.J. (1994). Desertification: Exploding the Myth . Chichester, John Wiley and Sons. Thompson, R.D. (1975) The Climatology of the Arid World . Reading, University of Reading, Department of Geography Paper No. 35. Thornthwaite, C.W. (1948) An approach towards a rational classification of climate. Geographical Review 38:55–94. Trenberth, K.E. (1993) The different flavours of El Niño. Proceedings of the 18th Annual Climate Diagnostics Workshop : 50–3. Boulder CO, National Center for Atmospheric Research. UNEP (1992) World Atlas of Desertification . Nairobi, UNEP, and London, Edward Arnold. UNESCO (1977) Map of the World Distribution of Arid Regions: Man and Biosphere . Paris, Technical Note 7. Waliser, D.E. and Gautier, C. (1993) A satellite-derived climatology of the ITCZ. Journal of Climate 6 :2162–74. Wallen, C.C. (1967) Aridity definitions and their applicability. Geografiska Annaler 49a: 367–84. Wilby, R.L. and Wigley, T.M.L. (1997) Downscaling general circulation model output: a review of methods and limitations. Progress in Physical Geography 21: 530–48. Williams, M.A.J. and Balling, R.C. (1994) Interactions of Desertification and Climate . Geneva, World Meteorological Organization. World Meteorological Organization (1975) Drought in Agriculture; Technical Note No. 138 . Geneva, World Meteorological Organization. Yair, A. and Berkowicz, S.M. (1989) Climatic and non-climatic controls of aridity: the case of the northern Negev of Israel. Catena Supplement 14, Arid and Semi Arid Environments .
Part II SOUTHWEST AND CENTRAL ASIA
3 The decline of desert agriculture: a view from the classical period Negev STEVEN A.ROSEN
INTRODUCTION The presence of sophisticated, large-scale, agricultural systems dating to classical times in the arid regions of the central Negev, southern Jordan and Sinai has long served both to illustrate the ingenuity of the ancient peoples of the region and as an inspiration to modern peoples as to the potential of wise exploitation of the desert. Archaeological survey has demonstrated that agriculture was practised throughout the Irano-Turanian desert steppe zone in areas that today receive as little as 75 mm average annual rainfall (compare Evenari, et al., 1982:32, fig. 13 to Kedar, 1967). Virtually every wadi worthy of the name shows terrace systems for the damming of flash-floods and their exploitation for farming (Fig. 3.1). The amazing efficacy of these systems has been repeatedly demonstrated. Both texts (Bruins, 1986:87; Kraemer, 1958: Document 82; Mayerson, 1960:224–69) and experimental archaeology (Evenari et al., 1982:191–219) have indicated that yields from the desert zone using run-off water catchment systems could, in fact, approximate those of the Mediterranean zone (Bruins, 1986:87; Evenari et al., 1982:191–219). Excavations and surveys have revealed the existence of large and numerous wine presses (Mazor, 1981; Rubin, 1996:54; Shershefski, 1991:198–200; Fig. 3.2), suggesting industriallevel production of grapes and wine. The reconstruction and operation of some of these systems over several decades have demonstrated that in some ways they constitute an agricultural regime more resistant to drought than their counterparts in the better-watered areas farther north. Finally, in the central Negev, there were six towns, which, together with their village and homestead hinterland, comprised an urban system proper with a population of over 20,000 people, whose subsistence was based on this agricultural regime (Broshi, 1979; Elliot, 1982:103–14; Shershefski, 1991:200–14; Fig. 3.4). In the light of the impressive nature of these systems, their decline is all the more marked. By the tenth or eleventh centuries AD, the entire settlement system of the central Negev had been abandoned. All previously occupied
The decline of desert agriculture: a view from the classical period Negev
45
Figure 3.1 Terraced dam system in the upper reaches of Nahal Nizzana, in the central Negev Note: The terrace dams are marked by the lines of vegetation across the wadi bed; the system of dams extends for several km along this stretch of the wadi Photograph: S.Rosen sites, including towns, villages, farmsteads and even nomadic encampments, had been deserted, and there is no evidence for any alternative settlements, either permanent or nomadic (for example: Avni, 1996; Nachlieli, 1992; Rosen, 1987a; Rosen and Avni, 1993). The desert had reverted to desert. The stark contrast between the rich archaeological remains and the contemporary desolation has struck every traveller through the region (Fig. 3.3), and there has been no shortage of attempts to explain this apparent ‘triumph of the desert’. Two general factors have been suggested as primary causes for Negev desertification: (1) the Moslem or Arab conquests, and the ensuing destruction of Byzantine civilization (for example: Lowdermilk, 1945:136; Negev, 1988:15; Palmer, 1872:243; Reifenberg, 1955:98; Sharon, 1969); and (2) climatic deterioration, rendering habitation impossible due to shifting sands, increased erosion and reduced water for agriculture (for example: Huntington, 1911; Issar, 1995; Issar and Govrin, 1991). Additional subfactors have included the negative effects of over-grazing by the flocks and herds of bedouin (Reifenberg, 1955:98), the destructive effects of earthquakes (Fabian, 1994) and increased marauding by nomads (Sharon, 1976; cf. also Lowdermilk, 1945:129). Critical examination of these factors in the light of recent intensive archaeological research carried out in the Negev indicates that each of these explanations is fundamentally flawed as a prime mover in the desertification
The archaeology of drylands
46
Figure 3.2 The wine press at Shivta (Subeita) Note This is a relatively small press, located in one of the central squares of the town. The actual pressing floor is the square area in the background, while the collecting and settling vat is in the foreground. An intermediate settling or filtering area is poorly preserved, located to the left of the vat Photograph: S.Rosen of the Negev, although each plays a role within a larger perspective. The key issue, rarely discussed in reviews of the decline of classical civilization in the Negev, is that periods of cultural florescence can usually be tied to increased economic and social input from, or integration with, the Mediterranean core area. The collapse of the economic core will inevitably result in the collapse of its dependants, unless alternative economic paths are available.
ARCHAEOLOGICAL OVERVIEW The central Negev in the sixth century AD, the Byzantine period in local terms, was the well-integrated frontier province of Palestina Tertia of the Late Roman empire (Mayerson, 1994; Rubin, 1997; Shershefski, 1991; also see Isaac, 1992). Although the lucrative trade route of the Nabatean period had long since been eclipsed by alternative trade systems (Crone, 1987; Negev, 1988), the province functioned both as a strategic southern buffer zone protecting the Levantine heartland (Gihon, 1980; Mayerson, 1986, 1990; also Isaac 1992 for a differing view) and as a gateway to both the holy pilgrimage destinations of the Sinai and to the mineral-rich desert regions farther east and south (Mayerson, 1982, 1983).
The decline of desert agriculture: a view from the classical period Negev
47
Figure 3.3 Palmer’s pen-and-ink sketch of the Byzantine town of Shivta (Esbeita or Subeita) in the central Negev, showing the rich archaeological remains amidst the desert environment Source: After Palmer, 1872:314 Archaeologically, the region is marked by two complementary settlement systems (Avni, 1996; Elliott, 1982; Haiman, 1995a; Mayerson, 1989; Negev, 1988; Rubin, 1990; Rosen, 1987b; Rosen and Avni, 1993; Shersehfski, 1991; Fig. 3.4). First, in the north and in the higher mountains, both better watered than areas farther south, large towns such as Avdat (Fig. 3.5), supported by intensive run-off agricultural systems (Fig. 3.6), evolved out of the Limes Palestina and the preceding Nabatean caravanserai over the course of several centuries. By the sixth century AD, the six towns of Elusa (modern Haluza), Ruheiba (Rehovot), Subeita (Shivta), Nessana (Nizzana), Oboda (Avdat) and Mampsis (Mamshit or Kurnub) represent the expansion of Byzantine society and economy deep into the desert. The design and construction of these towns are dominated by an architecture whose roots are undeniably in the Mediterranean zone, with little adjustment for local conditions (Shershefski, 1991:228), excepting the use of local raw materials (Negev, 1980). Christianity is the only religion represented at these sites in this preIslamic period, and classic basilica-style churches are present, in the plural, at each town. The wealth of the towns is especially evident in these churches, which showed such features as wall facings and furniture of marble imported from Anatolia, elaborate mosaics, and vaults of large wooden beams imported from the Mediterranean zone (Negev, 1974).
The archaeology of drylands
48
Figure 3.4 Map of the general settlement system of the central Negev during the Late Byzantine and Early Islamic periods Key: /// urban zone with agricultural support; = village and farmstead agricultural hinterland; \\\ pastoral nomadic region lacking evidence for agricultural exploitation; | | | agro-pastoral region showing combination of pastoral sites with agricultural exploitation. The major cities were Elusa (modern Haluza), Ruheiba (Rehovot), Subeita (Shivta), Nessana (Nizzana), Oboda (Avdat), and Mampsis (Mamshit or Kurnub). For detailed discussion, see Rosen and Avni (1993)
The decline of desert agriculture: a view from the classical period Negev
49
Figure 3.5 View of the Byzantine town of Avdat (looking north) Note: The left edge of the cliff shows the remains of churches, and a late Nabatean temple and a Byzantine fortress are located to the right of this. The domestic quarter is located on the slopes, and to the right (foreground) Photograph: S.Rosen Although defensively postured, defence does not seem to have been a primary consideration in the settlements. Aside from the isolated nature of many of the villages and farmsteads, only Mampsis shows a circumference wall, although Avdat shows a fortification wall on one side of the settlement. Neither is especially massive. Both Avdat and Nessana show internal forts, indicating military presence. Subeita presents a limited number of access gates to the town, but these gates are in fact breaks in the continuum of attached structures and not the gates of a city wall (Shershefski, 1991:184–8). The agricultural systems surrounding these towns, both those in direct association with the towns and those that were part of the village-farmstead hinterland, are perhaps the most impressive evidence of the wealth and long-term stability of the Byzantine regime (Bruins, 1986; Evenari et al., 1982; Kedar, 1967; Mayerson, 1960). Vast areas of both wadi floodplain and upper alluvial terraces show elaborate systems of terraced dams, drainage channels, sluice gates and support walls. Hill slopes are covered with tuleiliot el anab—rows of stone mounds and stone lines—whose function was presumably connected to either ground clearance for run-off enhancement or some other form of agricultural activity (Evenari et al., 1982:127–47). Calculations based on aerial photography, pedestrian survey and farm reconstruction demonstrate that the average ratio of drainage catchment to farmed area was
The archaeology of drylands
50
Figure 3.6 Elaborate raised field and dam system on Nahal La van, just south of Shivta (Subeita) Note: Notice the wadi bed to the right of the fields; water flow was drained onto the raised fields several km upstream Photograph: S.Rosen approximately 21:1, so that with run-off estimated at 15 per cent of actual rainfall, an average annual rainfall of 100 mm could be transformed to an effective annual rainfall for the farmed fields of more than 400 mm (Evenari et al., 1982:95–119). Not only is this more than sufficient for growing barley and wheat (the basic cereal staples of the period), but it sufficed for growing grapes and olives as well. The presence of olive and wine presses at each town, sometimes at an industrial scale, demonstrates clearly the practice of arboriculture and viticulture; dates, figs and even pomegranates were also grown (Mayerson, 1960; Mazor, 1981; Rubin, 1996). Rubin (1996) characterizes this system as the adoption of the Mediterranean agricultural system into the Negev. The second system, which is less well documented than Palestina Tertia, is that of the pastoral hinterland, located in the deserts beyond the village-farming hinterland (Avni, 1996; Haiman, 1995a; Rosen, 1987b, 1994; Rosen and Avni, 1993). Aside from the significantly lower rainfall associated with these southern areas, the region is marked by the general scarcity of agricultural remains and the presence of the larger-scale pastoral encampments. The remains of pastoral encampments are found throughout the desert and steppe zones, but the larger aggregate camps are located only south of the agricultural areas. These camps are, obviously, smaller than the Byzantine towns and villages, but they also differ in their basic architecture and organization. In essence, the structures revealed at such encampments are to be interpreted as ephemeral tent bases or, in some cases, as hut foundations that carried brush or tent superstructures. The settlements align along secondary and tertiary drainages, in patterns dictated by topography. Analyses of
The decline of desert agriculture: a view from the classical period Negev
51
material culture also support the interpretation of these settlements as basically pastoralnomadic (Rosen and Avni, 1997:62–81), and textual references (Mayerson, 1989) accord well with this. A key point in analyses of these pastoral systems is their essential dependence on the settled system to the north for both their subsistence and their material culture. The markets of the settled zone were a sine qua non for pastoral existence in the desert (cf. Khazanov, 1984). Relations between the desert and the sown, while perhaps occasionally tense, must have been essentially stable for the nomadic system to have thrived. In summary, when the agricultural exploitation of the desert was at its peak in the classical period, the region had been well integrated into the Roman-Byzantine (and later Ummayad) empire (Rubin, 1996, 1997). That integration in essence established the economic and social stability that enabled the desert to bloom.
THE ISLAMIC CONQUESTS AS CAUSE FOR DESERTIFICATION The battle of Gaza, in AD 633–4, marks the beginning of the political end of the Byzantine empire in the Negev. Although the events leading up to that battle, and the causes behind the Byzantine collapse, have been much discussed, and are beyond the scope of this paper, in terms of desertification, several important points require attention. Archaeologically, there is no evidence for the destruction or violent conquest of any of the Negev towns (per contra Negev, 1988:15). In fact, the processes of urban decline seem to have been initiated well before the Islamic period. Mampsis (Negev, 1988:15) does not appear to show an occupation in the seventh century at all. Avdat shows evidence for a major earthquake at the beginning of the seventh century, after which the city seems to have been abandoned for two centuries, and eventually reoccupied during the Islamic period (Fabian, 1994). Significantly, an earlier fourth- or fifth-century earthquake resulted in repairs and ‘retro-fitting’ of various structures against further earthquake damage. Nessana shows continued occupation at least into the late seventh century and probably well into the eighth, both in the archaeology (Shershefski, 1991:5, 50) and in the archives recovered from the site (Kraemer, 1958:213), with little obvious disruption, although a clear decline can be traced. At Subeita, the presence of a mosque wedged into an open space next to a church (Baly, 1935; Segal, 1983; Shershefski, 1991:74) indicates both clear continuity of occupation well into the eighth century and its contemporaneity with at least one church on the site, indicating the peaceful coexistence of the two religions during the Ummayad period. Ruheiba (Shershefski, 1991:95) also seems to show continued occupation into the early Islamic period. Recent excavations at Elusa have not revealed any evidence for Islamic occupation, but nor is there any evidence for destruction (although it must be admitted that the areas excavated are still quite limited). The excavator (Goldfus, pers. comm.) suggests an abandonment prior to the Islamic period. The city of Beer-Sheba (Figueras, 1979), in the northern Negev, seems to show archaeological decline as well, although again with no evidence for either abandonment or destruction. In this context, it is important to recognize that the first decades of the seventh century were catastrophic for the Byzantine empire, as a result of its long wars with the
The archaeology of drylands
52
Sassanids. Although it is unlikely that the Persian armies that devastated the Levant actually came as far south as the central Negev, the havoc wreaked on the Mediterranean heartland could not but have been felt on the periphery as well. On the other hand, in spite of the decline marked in the cities, the Ummayad and early Abassid periods seem to show a rural florescence. The central village and satellite farms at Sede Boqer (Nevo, 1985, 1991; Fig. 3.7) are the best example of this phenomenon. Another example is the farmstead at Nahal Mitan (Haiman, 1995b). Avni (1994) has indicated the presence of at least thirteen mosques in the Negev highlands in this period, some of
Figure 3.7 The early Islamic village of Sede Boqer in the central Negev Note: The site is surrounded by numerous agricultural terraced field systems, not pictured Photograph: S.Rosen which are clearly associated with farming settlements and others with pastoral encampments. Finally, a series of large homesteads that were colonized during the Ummayad and early Abassid periods has been excavated recently around the outskirts of Beer-Sheba (for example: Bar-Ziv and Katz, 1993; Gilead et al., 1993; Katz, 1993; Katz and May, 1996; Nachshoni et al., 1993; Negev, 1993). Evidence from the nomadic periphery also shows continuity, with little evidence for destruction or invasion. Pastoral settlements dating to the eighth and perhaps ninth centuries AD have been excavated in the southern central Negev (Rosen and Avni, 1997). Some of these, in the higher areas, seem to show the adoption of floodwater farming into the pastoral subsistence system (Rosen and Avni, 1993). The continued import and use of
The decline of desert agriculture: a view from the classical period Negev
53
typologically Byzantine ceramics (and other elements of material culture) from the settled regions into the pastoral sites demonstrate underlying economic continuities. There was no break in relations between the nomads and the farmers in the transition to the Ummayad administration. Importantly, there is no incursion of nomadic settlement types into the agricultural zones in this period. Although erosion is a dominant feature in the desert landscapes today, it cannot be linked to the over-grazing that is often tied to such pastoral incursions, since there is no evidence of such incursions. In short, the Islamic conquests—a problematic concept in itself for the Negev—did not bring any desertification. Whilst the late Byzantine period saw an urban decline in the Negev, the early Islamic period seems by contrast to have seen a rural renaissance.
CLIMATIC DETERIORATION AS CAUSE FOR DESERTIFICATION Establishing climatic change as a prime factor in cultural transformation requires three distinct steps. First, one must establish the reality of the climatic change itself. Second, the suggested climatic change must be correlated chronologically with the cultural transformation. Third, a reasonable scenario or mechanism for causality must be established, beyond the mere fact of correlation: it is not enough to establish a climatic change, indicate a contemporaneity with a cultural change, and then claim a causal link. There are several lines of evidence suggesting a change in climate some time following the classical period settlements. The most obvious of these is the deposition of extensive terraces sometime during the classical period (Bruins, 1986:189; Goldberg, 1994), followed by their erosion and wadi downcutting (Ben-David, 1997; Bruins, 1986:189; Reifenberg, 1955). It is clear that there has been landscape degradation, but it is not clear either when this degradation occurred or whether it was the result of climatic changes or of other factors such as microtectonics or human intervention. Certainly, accelerated erosion can be expected if terrace systems are not maintained (cf. Butzer, 1974), and some of the gullying that can be seen in the Negev today is the undoubted result of breached dam systems and not climatic change. One possible indication of a climatic component is the existence of post-classical downcutting in areas where agriculture, or its abandonment, can be discounted as a factor. The pastoral encampment of Nahal Oded (Fig. 3.4), south of the Ramon Crater, shows two post-classical wadi channels, one a modern one and the other an earlier and somewhat higher one that cuts several eighth-century structures located on higher alluvial terraces. In the absence of any agriculture in the area, Ben-David (1997) suggests that these downcutting events reflect episodes of extreme aridity, both of which post-date the Ummayad-period occupation of the site. The infiltration and movement of sand dunes, blocking drainages and burying settlements, have also been suggested as reflective of climatic deterioration. Issar (1995) claims that the burial of the Byzantine towns of Elusa and Ruheiba, beginning c.AD 800, is the result of an increased supply of Nile sands on the Levantine littoral, to be correlated with increased monsoon rains in East Africa. Dead Sea water levels, as established from salt cave evolution and analysis of core sediments, have also been used to reconstruct climatic sequences (Frumkin et al., 1991,
The archaeology of drylands
54
1994; Geyh, 1994; Issar, 1995; Neev and Emery, 1995:62). Summarized briefly, higher Dead Sea levels are evident during the first two centuries AD (the early Roman period), indicating greater humidity. This period was followed by a warmer, more arid, period in the middle of the first millennium BC that was not ameliorated until the beginning of the second millennium AD. Analyses of oxygen isotope ratios from cave speleothems and marine molluscs (Gat and Magaritz, 1980; Geyh, 1994) show high 18O ratio peaks of c.2300 BP and c.1500 BP, indicating cooler temperatures (and presumably higher humidity), with cooler (and moister) periods between and following. These analyses accord well with the studies of the Dead Sea water level. Of further interest is the apparently significantly warmer (and drier) period prior to 2300 BP, so that, although it was not especially cool or moist on any absolute scale, the c.2300 BP episode is a relatively significant amelioration. Later episodes do not approach this first in the scale of change. Given the above data, from different sets of evidence, it is hard to argue that climate remained stable during the first millennium AD (per contra Rubin, 1989). The next issues are whether the climatic fluctuations outlined above do indeed correspond with, and can explain, the cultural and physical desertification of the Negev. The weakest link in the argument is that of dating, since shifts of a few hundred years, quite within the range of radiocarbon errors given problems of fractionation, intrusion and so on, significantly affect historical interpretation (Gat and Magaritz, 1980). However, given current dating of the climatic events, it is hard to reconcile them with the desertification of the Negev. Thus, the Nabatean and early Roman periods, in the final centuries BC and first two centuries AD, when agriculture was incipient at best (Bruins, 1986:189; Mayerson, 1963), seem to have been at a climatic optimum, whereas the cultural peak in the succeeding Byzantine period seems to have been climatically dry. The Byzantine collapse and rise of the early Islamic empire seem to have been either stable climatically or marked by only minor fluctuations. Although sand dunes did indeed bury those cities built in the dune areas, Goldfus (pers. comm.) suggests that Elusa was in fact abandoned relatively early, prior to the eighth-century dune invasions claimed by Issar (1995). Notably, Avdat, Subeita and Mampsis were not affected by dunes at all. It is important to stress here that the gradual abandonment of the Byzantine cities is not equivalent to either the abandonment of the Negev or desertification, for, as indicated earlier, there is a significant Early Islamic agricultural presence in the Negev at least until the ninth or tenth centuries AD. The final abandonment of the central Negev, probably in the tenth or perhaps even eleventh centuries AD, may in fact even be associated with the beginning of climatic amelioration. In short, climatic change does not adequately explain the decline of classical civilization in the desert or the reversion of the desert to desert.
THE RISE OF THE DESERT To understand the rise of the desert, we must understand first its domestication. The essence of the classical period ‘Green Revolution’ in the Negev was the transplantation of a Mediterranean-zone agricultural complex into the arid zone. This complex, in the
The decline of desert agriculture: a view from the classical period Negev
55
Mediterranean zone, consists of cereal (wheat and barley) farming, the cultivation of fruit crops, including grapes, olives, figs and dates, and animal husbandry based especially on sheep, goat and cattle, with significantly less emphasis on pig. Landscape management, in the form of hill slope terracing and various forms of irrigation, is integral to the complex as well (Grigg, 1974:123–8, 132–4; Stager, 1985; compare also with Braudel 1972:56, 59, 423). Despite claims concerning the inappropriateness and instability of Mediterranean-zone farming systems in the New World and other non-Mediterranean environments (see Butzer, 1996), the expansion of the Mediterranean zone into the desert, in terms of culture, society and subsistence, in fact proved a remarkably stable phenomenon, enduring for at least half a millennium. The stability of this system is even more marked given the political perturbations that occurred during this period, perturbations that included the rise and decline of urban centres, the rise of Christianity, the collapse of Byzantine administration and the rise of Islam. Two points are particularly relevant for comprehending the success of the Mediterranean system. The first is the integration of the desert economy, both in terms of trade and subsistence, into the larger state. Even beyond the fact of active imperial subsidy, that the desert settlement system was well embedded in the classical world is reflected in virtually all aspects of material culture, economy and society. The second is that the Mediterranean economy itself should be seen as a flexible strategy, fluctuating between emphasis on cash crops and subsistence staples, depending on the historical and economic contexts. Within the Mediterranean zone, during periods of social collapse, the complex shifts towards subsistence mode, whereas during times of economic prosperity, cash crops play a larger role (Stager, 1985). In the desert zone, the subsistence mode may be insufficient by itself, especially given large urban populations that were at least partially supported by trade. Even without reference to climatic change, the desert environment exerts pressures on settlement systems not felt in better watered areas. Thus, regardless of the effectiveness of run-off irrigation systems, agriculture in the Negev must have required significantly more labour input than further north, for example in the construction and constant maintenance of terrace dam systems. Subsistence is more difficult in a desert, and therefore the raison d’être of permanent settlement in the Negev has always been its integration with some other core region. The decline of the core-region economy resulted in a reversion towards the subsistence end of the Mediterranean complex spectrum, one that may not have been sustainable in the desert at the high population levels there of village and urban society. The Mediterranean complex continued well into the early Islamic period. In this context, it is important to understand that the early Islamic horizon in the Negev, in spite of its rural character, still shows a high degree of social and economic integration with the Mediterranean core area. This is most obvious in the material culture continuities between the core and the periphery. However, it is especially impressive in the ideological integration, such that Negev rock inscriptions from this period follow very standardized Islamic formulae (Sharon, 1990), burials are typically Moslem (Rosen and Avni, 1997:13) and mosques follow standard definitions (Avni, 1994). It was in the later Abbassid period, following the political and economic shift of the Caliphate to Baghdad, that the Levant itself declined. With that decline, the means for the integration of the desert and the sown were no longer available, and the entire desert system, both settled
The archaeology of drylands
56
and nomadic, was abandoned. Glantz (1994) defines desertification as the creation of unproductive desertlike landscape in a place where none had existed in the recent past. In the sense that the central Negev reverted from being a productive and integrated component of a Mediterranean state system to its original desert state, the processes reviewed here are indeed those of desertification.
FINAL NOTE The history of research on the rise of Near Eastern deserts is one inextricably tied to the political and ideological struggles of the region. For example, the nineteenth-century British Orientalist Edward Palmer (1872:241–3) viewed the decline of civilization and the rise of the desert as the result of invasion and indigence on the part of the local inhabitants. Ellsworth Huntingdon’s (1911) environmental determinism, in which he claimed that settlement and the rise and decline of civilization were dictated by the carrying capacity of a region, in turn determined by climate and environment, was in antithesis to attitudes like Palmer’s, it was adopted as state policy by the British Foreign Office in its administration of Palestine, and used as a rationale for limiting Jewish immigration. In response, Zionist ideologues such as Ben-Gurion and Ben-Zvi (1979 [1918]) claimed that the decline of Palestine and the rise of the desert were the result of negligent administration, discounting the role of climatic change (Troen, 1989). Indeed, Ben-Gurion (1961) idealized the rebirth of the desert. As a part of the scientific background to the Zionist vision of the blooming desert, the role of the black goat as a factor in the reduction of vegetation, and in the consequent rise in erosion, has often been stressed (Orni and Efrat, 1980:470; Reifenberg, 1955:98; see also Kohler-Rollefson, 1992, for a claim for destructive over-grazing in the Neolithic), thus legitimising expropriation of bedouin grazing lands. In response, some scholars have denied traditional pastoral nomadism and grazing as a significant factor in landscape degradation (Thomas and Middleton, 1994:13, 67–73). Desertification is the result of a complex chain of causality. On its most simple level, it is clear that land degradation is the result of physical processes. However, these physical processes are often set in motion by human activities (Glantz, 1994), such that the issues are social and historical (Blaikie and Brookfield, 1987). The historical causalities are also complex: it is accepted knowledge that over-grazing by pastoralists causes erosion (Orni and Efrat, 1974:470; Reifenberg, 1955:98), but Danin (1983:17) notes that ‘during the few years that several Negev and Sinai areas were closed to bedouin and their domestic animals, no substantial changes in the list of species and plant communities could be discerned’. As noted above, land degradation as a consequence of over-grazing may be only the latest stage in desertification. In the circum-Mediterranean region, the steppe zones inhabited by bedouin in the nineteenth and early twentieth centuries were almost all exploited for agriculture during the classical period, and subsequently abandoned, to be exploited by pastoralists only later. The simplistic notions of Islamic invasion or climatic catastrophe as prime causes in the decline of the Negev in fact mask political agendas. It is the historical complexities, in all their richness and texture, that need to be
The decline of desert agriculture: a view from the classical period Negev
57
addressed before we can critically understand desertification as a social phenomenon.
ACKNOWLEDGEMENTS I am grateful to Graeme Barker and David Gilbertson for the opportunity to participate in the symposium on the Archaeology of Drylands at the Fourth World Archaeological Congress and to the other participants for their stimulating and eye-opening papers. Haim Goldfus was good enough to read an early draft of this paper and make valuable comments. Arlene Miller Rosen shared her knowledge of climate and climate-change freely and happily. The photographs were developed from slides by Alter Fogel, and the map was drafted by Patrice Kaminsky, both of Ben-Gurion University.
REFERENCES Avni, G. (1994) Early mosques in the Negev highlands—new archaeological evidence on Islamic penetration of southern Palestine. Bulletin of the American Schools of Oriental Research 294:83–100. Avni, G. (1996) Nomads, Farmers and Town-Dwellers . Jerusalem, Israel Antiquities Authority. Baly, C. (1935) Sbaita. Palestine Exploration Fund Quarterly Statement 62:171–81. Bar-Ziv, H. and Katz, O. (1993) Beer Sheva—Nahal Ashan. Archaeological News 100: 100–01. (In Hebrew.) Ben-David, R. (1997) The geology and geomorphology of the Nahal ‘Oded site area. In S.A.Rosen and G.Avni, The ‘Oded Sites: Investigations of Two Early Islamic Pastoral Camps South of the Ramon Crater . 109–18. Beersheva, Ben-Gurion University Press, Beersheva XI, Studies by the Department of Bible and Ancient Near East. Ben-Gurion, D. (1961) Introduction. In Y.Morris, Masters of the Desert : 11–20. New York, Putnams. Ben-Gurion, D. and Ben-Zvi, Y. (1979 [1918]) Eretz Israel in Past and Present . Jerusalem, Yad Ben Zvi Press. (Translated from Yiddish to Hebrew by D. Niv.) Blaikie, P. and Brookfield, H. (1987) Defining and debating the problem. In P.Blaikie and H.Brookfield (eds) Land Degradation and Society : 1–7. London, Methuen. Braudel, F. (1972) The Mediterranean and the Mediterranean World in the Age of Philip II . New York, Harper. Broshi, M. (1979) The population of western Palestine in the Roman-Byzantine period. Bulletin of the American Schools of Oriental Research 236:1–10. Bruins, H.J. (1986) Desert Environment and Agriculture in the Central Negev and Kadesh Barnea during Historical Times . Nijkirk, Netherlands, Midbar Foundation. Butzer, K.W. (1974) Accelerated soil erosion: a problem of man-land relationships. In I.R.Manners and M.W.Mikesell (eds) Perspectives on Environment . 57–77. Washington D.C., Association of American Geographers. Butzer, K.W. (1996) Ecology in the long view: settlement histories, agrosystemic strategies, and ecological performance. Journal of Field Archaeology 23:141–150.
The archaeology of drylands
58
Crone, P. (1987) Meccan Trade and the Rise of Islam . Princeton, Princeton University Press. Danin, A. (1983) Desert Vegetation of Israel and Sinai . Jerusalem, Cana. Elliott, Jack D. Jr. (1982) The Elusa Oikoumene: A Geographical Analysis of Ancient Desert Ecosystem Based on Archaeological, Evironmental, Ethnographic and Historic Data . Mississippi State, Mississippi State University, Cobb Institute of Archaeology Occasional Papers 82–01. Evenari, M., Shanan, L. and Tadmor, N. (1982) The Negev: The Challenge of a Desert . Cambridge, Mass., Harvard University Press. Fabian, P. (1994) New evidence for earthquakes and their influence on the urban development of Avdat. Paper presented at the 20th Archaeological Congress in Jerusalem, Israel. Figueras, P. (1979) The Roman-Byzantine period. In Y.Grades and E.Stern (eds) Beersheva : 39–52. Jerusalem, Keter. (In Hebrew.) Frumkin, A., Magaritz, M., Carmi, I. and Zak, I. (1991) The Holocene climatic record of the salt caves of Mount Sedom, Israel. The Holocene 1:191–200. Frumkin, A., Carmi, I., Zak, I. and Magaritz, M. (1994) Middle Holocene environmental change determined from the salt caves of Mount Sedom, Israel. In O. Bar-Yosef and R.Kra (eds) Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean : 315–22. Tucson, University of Arizona, Radiocarbon Press. Gat, J.R. and Magaritz, M. (1980) Climatic variations in the eastern Mediterranean sea area. Naturwissenschaften 67:80–7. Geyh, M.A. (1994) The paleohydrology of the eastern Mediterranean. In O.Bar-Yosef and R.Kra (eds) Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean: 131–45. Tuscon, University of Arizona, Radiocarbon Press. Gihon, M. (1980) Research on the Limes Palaestina: a stocktaking. In W.S.Hanson and L.J.F.Keppie (eds) Roman Frontier Studies 1919 :843–64. Oxford, British Archaeological Reports, International Series 71. Gilead, I., Rosen, S.A. and Fabian, P. (1993) Horvat Beter (Be’er Matar) 1990–1991. Archaeological News 99:88–89. (In Hebrew.) Glantz, M.H. (1994) Drought, desertification and food production. In M.H.Glantz (ed.) Drought Follows the Plough : 7–32. Cambridge, Cambridge University Press. Goldberg, P. (1994) Interpreting late Quaternary continental sequences in Israel. In O.Bar-Yosef and R.Kra (eds) Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean : 89–102. Tucson, University of Arizona, Radiocarbon Press. Grigg, D.B. (1974) The Agricultural Systems of the World: An Evolutionary Approach . Cambridge, Cambridge University Press. Haiman, M. (1995a) Agriculture and nomad-state relations in the Negev desert in the Byzantine and early Islamic periods. Bulletin of the American Schools of Oriental Research 297:29–54. Haiman, M. (1995b) An early Islamic period farm at Nahal Mitnan. Atiqot 26:1–13. Huntington, E. (1911) Palestine and Its Transformation . Boston, Houghton & Mifflin. Isaac, B. (1992) The Limits of Empire . Oxford, Clarendon Press. Issar, A. (1995) Climatic change and history of the Middle East. American Scientist 83:350–5.
The decline of desert agriculture: a view from the classical period Negev
59
Issar, A. and Govrin, Y. (1991) Climatic changes and the desertification of the Negev at the end of the Byzantine period. Katedra 61:67–81 (In Hebrew.) Katz, O. (1993) Beer Sheva—Nahal Beka’ 1. Archaeological News 99:87–8. (In Hebrew.) Katz, O. and May, V. (1996) Beer Sheva, Ramot B. Archaeological News 106:162–4. (In Hebrew.) Kedar, Y. (1967) Ancient Agriculture in the Negev Highlands . Jerusalem, Bialik Institute. Khazanov, A.M. (1984) Nomads and the Outside World . Cambridge, Cambridge University Press. Kohler-Rollefson, I. (1992) A model for the development of nomadic pastoralism on the Jordanian steppe. In O.Bar-Yosef and A.M.Khazanov (eds) Pastoralism in the Levant: Archaeological Materials in Anthropological Perspectives : 11–18. Madison, Prehistory Press. Kraemer, C.J. (1958) Non-Literary Papyri. Excavations at Nessana Volume III . Princeton, Princeton University Press. Lowdermilk, W.C. (1945) Palestine, Land of Promise . London, Gollancz. Mayerson, P. (1960) The Ancient Agricultural Regime of Nessana and the Central Negev. Excavations at Nessana Volume I . London, Colt Institute. Mayerson, P. (1963) The desert of southern Palestine according to Byzantine sources. Proceedings of the American Philosophical Society 107:160–72. Mayerson, P. (1982) The pilgrim routes to Mount Sinai and the Armenians. Israel Exploration Journal 32:44–57. Mayerson, P. (1983) The city of Elusa in the literary sources of the fourth-sixth centuries. Israel Exploration Journal 33:247–53. Mayerson, P. (1986) The Saracens and the Limes. Bulletin of the American Schools of Oriental Research 262:35–47. Mayerson, P. (1989) Saracens and Romans: micro-macro relationships. Bulletin of the American Schools of Oriental Research 274:71–9. Mayerson, P. (1990) Toward a comparative study of a frontier. Israel Exploration Journal 40:267–79. Mayerson, P. (1994) Monks, Martyrs, Soldiers and Saracens. Papers on the Near East in Late Antiquity ( 1962–1993 ). Jerusalem, Israel Exploration Society. Mazor, G. (1981) The wine presses of the Negev. Qadmoniot : 53–54:51–60. (In Hebrew.) Nachlieli, D. (1992) The Negev Highlands and the Arava During the Early Arab Period . Tel Aviv University, unpublished MA thesis. Nachshoni, P., Ustinov, Y. and Bar-Ziv, H. (1993) Beer Sheva—Nahal Kovshim. Archaeological News 99:84–5. (In Hebrew.) Neev, D. and Emery, K.O. (1995) The Destruction of Sodom, Gomorrah, and Jericho . Oxford, Oxford University Press. Negev, A. (1974) The churches in the central Negev: an archaeological survey. Revue Biblique 81:400–22. Negev, A. (1980) House and city planning in the ancient Negev and Provincia Arabia. In G.Golany (ed.) Housing in Arid Lands: Design and Planning : 3–32. New York, John
The archaeology of drylands
60
Wiley and Sons. Negev, A. (1988) The Nabatean Cities in the Negev . Jerusalem, Ariel 62–63. Negev, N. (1993) Beer Sheva—Kiryat HaUniversita (Mizrach). Archaeological News 99:85–6. (In Hebrew.) Nevo, YD. (1985) Sede Boqer and the Central Negev in the 7–8th Centuries AD . Jerusalem, Israel Publication Services. Nevo, YD. (1991) Pagans and Herders . Jerusalem, Israel Publication Services. Orni, E. and Efrat, E. (1980) Geography of Israel . Jerusalem, Israel University Press. Palmer, E.H. (1872). The Desert of the Exodus . New York, Harper and Bros. Reifenberg, A. (1955) The Struggle Between the Desert and the Sown . Jersualem, The Jewish Agency. Rosen, S.A. (1987a) Demographic trends in the Negev highlands: preliminary results from the Emergency Survey. Bulletin of the American Schools of Oriental Research 266:45–58. Rosen, S.A. (1987b) Byzantine nomadism in the Negev: results from the Emergency Survey. Journal of Field Archaeology 14:29–42. Rosen, S.A. (1994) The nomadic periphery: archaeology of pastoralists in the south central Negev during late antiquity. Aram 6:295–309. Rosen, S.A. and Avni, G. (1993) The edge of empire: the archaeology of pastoral nomads in the southern Negev highlands in late antiquity. Biblical Archaeologist 56: 189–99. Rosen, S.A. and Avni, G. (1997) The ‘Oded Sites: Investigations of Two Early Islamic Pastoral Camps South of the Ramon Crater . Beersheva, Ben-Gurion University Press, Beersheva XI, Studies by the Department of Bible and Ancient Near East. Rubin, R. (1989) The debate over climatic changes in the Negev, fourth-seventh centuries CE. Palestine Exploration Quarterly 121:71–8. Rubin, R. (1990) The Negev as Settled Land . Jerusalem, Hebrew University Press. Rubin, R. (1996) Urbanization, settlement and agriculture in the Negev desert—the impact of the Roman-Byzantine empire on the frontier. Zeitschrift des Deutschen Palstini-Vereins . 112:49–60. Rubin, R. (1997) The Romanization of the Negev, Israel: geographical and cultural changes in the desert frontier in late antiquity. Journal of Historical Geography 23: 267–83. Segal, A. (1983) The Byzantine City of Shivta (Esbeita), Negev Desert, Israel Oxford, British Archaeological Reports, International Series 179. Sharon, M. (1969) The history of Palestine from the Arab conquest until the Crusades (633–1099). In M.Avi-Yonah (ed.) A History of the Holy Land : 185–220. London, Weidenfeld & Nicolson. Sharon, M. (1976) Processes of destruction and nomadization in Eretz Israel under Islamic rule (633–1517 CE). In M.Sharon (ed.) Notes and Studies on the History of the Holy Land Under Islamic Rule : 7–32. Jerusalem, Yad Ben Zvi Press. Sharon, M. (1990) Arabic rock inscriptions from the Negev. In M.Sharon and M.Halloun (eds) Supplement to the Map of Har Nafha (196) : 9–35. Jerusalem, Israel Antiquities Authority. Shershefski, J. (1991) Byzantine Urban Settlements in the Negev Desert . Beersheva, Ben-Gurion University Press, Beersheva V, Studies by the Department of Bible and
The decline of desert agriculture: a view from the classical period Negev
61
Ancient Near East. Stager, L.E. (1985) The first fruits of civilization. In J.N.Tubb (ed.) Palestine in the Bronze and Iron Ages: Papers in Honor of Olga Tufnell : 172–88. London, University of London, Institute of Archaeology. Thomas, D.S.G. and Middleton, N.J. (1994) Desertification: Exploding the Myth . New York, John Wiley and Sons. Troen, I. (1989) Calculating the economic absorptive capacity of Palestine: a study of the political uses of scientific research. Contemporary Jewry 10:19–38.
4 Farmers, herders and miners in the Wadi Faynan, southern Jordan: a 10,000-year landscape archaeology GRAEME BARKER
INTRODUCTION The Wadi Faynan Landscape Survey is a study of the landscape evolution of Wadi Faynan in southern Jordan from prehistoric times to the present day, as a contribution to the issue that is the central theme of this volume: the importance of providing a long-term archaeological perspective on how people have lived in arid lands. How did past societies in marginal environments learn to cope with risk? What solutions did they develop and how successful were they? Why did they take the choices they took? To what extent did their actions affect their landscape, and for good or ill? The rationale of the project has been to bring together an inter-disciplinary team of archaeologists, geographers and environmental biologists in the investigation of the landscape history of the chosen study area within a single integrated research framework. (The Acknowledgements at the end of the chapter detail the numerous colleagues working in the project whose researches are summarized in these pages.) The Wadi Faynan is situated about 40 km from Petra, the world-famous capital of the Nabatean kingdom that flourished in the last few centuries BC before the Roman conquest of the region (Fig. 4.1). The catchment of the wadi forms a transect about five km wide running for some 15 km westwards from the rim of the Jordanian plateau c.1,500 m above sea level to the floor of the Wadi Arabah rift valley at about sea level. The main wadi today is a bleak landscape, arid and largely denuded of vegetation (Figs. 4.2, 4.5), though where they cut through the plateau escarpment the channels of the three main feeder tributaries (the Dana, Ghuwayr and Shayqar) are in places well watered and comparatively well vegetated from ground springs. The Wadi Faynan today (part of the Dana Nature Reserve of Jordan’s Royal Society for the Conservation of Nature) is used mainly by nomadic bedouin herders, but is also well known for its abundant archaeological remains. The principal archaeological monuments of the Wadi Faynan, long known to early travellers, are the Khirbet Faynan (Fig. 4.2), a major settlement of
Farmers, herders and miners in the Wadi Faynan, southern Jordan
Figure 4.1 The location of Wadi Faynan within its region
63
The archaeology of drylands
64
Figure 4.2 Looking northeast across part of the ancient field system to Khirbet Faynan (the prominent hill in the right middle ground), thought to be the ancient settlement of Phaino mentioned by classical writers Photograph: G.Barker Nabatean, Roman and late Roman (Byzantine) date located at the head of the Wadi Faynan near the confluence of the three main tributaries, and, nearby, an aqueduct, reservoir and water mill of Roman/Byzantine date. To the west of this complex is a substantial (c.5 km long) field system of rubble walls, its surface pottery indicating primary use contemporary with that of the Khirbet Faynan settlement (Fig. 4.3). Before our project began in 1996, reconnaissance surveys had also located a variety of prehistoric sites both in the main wadi and in its tributaries, some of which are being excavated by other teams. Wadi Faynan and its environs are also characterized by rich mineral deposits, and from the work especially of the Bochum Mining Museum, the history of copper exploitation here is comparatively well documented (Hauptmann, 1989, 1992; Hauptmann and Weisgerber, 1987; Hauptmann et al., 1992). Although Faynan copper was used by neolithic and chalcolithic societies, the first intensive exploitation seems to date to the Early Bronze Age, c.3500–1900 BC. There was a second significant episode in the first part of the first millennium BC, the Edomite Iron Age. Copper was then extracted on a major scale in Nabatean and especially Roman and Byzantine times; it is generally agreed that Khirbet Faynan must be the settlement of Phaino mentioned by classical writers as the place to which prisoners such as Christians from Palestine and Egypt were transported in the third and early fourth centuries AD to work the copper mines it controlled.
Farmers, herders and miners in the Wadi Faynan, southern Jordan
65
Figure 4.3 The survey area of the Wadi Faynan Landscape Survey, showing the ancient field systems and the archaeological sites recorded up to 1999 Note: The topography shown is from a photogrammetric map, the boundaries of which do not extend as far as the boundaries of the survey area The Wadi Faynan seemed, therefore, a particularly attractive location for investigating the ‘archaeological history’ of interactions between a desertic landscape and its human inhabitants, given the rich archaeological record that appeared to be prima facie evidence for episodes of intensive settlement and sedentary farming in the past that were very different from settlement and land use today.
METHODOLOGIES The project began in 1996 and the fieldwork ends in 2000; ongoing results have been reported in annual papers in Levant (Barker et al., 1997, 1998, 1999, 2000). Geomorphological mapping and palaeoecological analysis of exposures and cored sediments are establishing an environmental sequence for the past 200,000 years, with a particular focus on the past 10,000 years. The resulting event sequence is being dated variously by radiocarbon and Optical Spin Luminescence (OSL) dates and by stratigraphic association with dated archaeological sites. The changes that are being observed in the palaeoenvironmental sequence can be linked with increasing confidence variously to natural processes of change, such as climatic shifts and phases of tectonic activity, and to cultural processes of change, such as arable, pastoral and industrial activities. Geochemical analysis of sediments using EDMA (Energy Dispersive X-Ray Microanalysis) is also being used to monitor the bioaccumulation of heavy metals as a
The archaeology of drylands
66
result of metalliferous pollution, providing an invaluable independent indicator of the changing scales, technologies and environmental impacts of mining and smelting activities to compare with the Bochum team’s studies of the mining and smelting sites. In the first three seasons of the project the programme of archaeological fieldwork concentrated on the detailed exploration of the relict field system (termed WF4 in the survey catalogue) that covers the lower terraces on the southern side of the present-day wadi channel (Figs. 4.2 and 4.3). This involved the verification on the ground of walls shown on an earlier photogrammetric survey, the systematic collection and counting of artefacts in each individually numbered field of the system (c. 900 fields in total) and the detailed recording of constructional details of wall types and of other structures within or adjacent to the fields. A series of smaller satellite field systems on the northern side of the main channel has also been recorded in the same way. These studies, combined with test excavations and identification of the prolific lithic and pottery collections, have confirmed that the main phase of construction and use of the field systems is broadly ‘classical’ (Nabatean/Roman/Byzantine—a period of some 1,000 years), but have also established that the evidence is at the same time a complex palimpsest of reuse and adaptation, of agricultural activities and systems of land management spanning the past 6,000 years or so. The focus of the fieldwork has now shifted to frame these data within the broader landscape context of the study area defined for the archaeological investigation shown in Figure 4.2, which measures just over 30 km2. Extensive field walking and recording of the block of terrain around the ancient field systems in 1999 were facilitated by using hand-held Garmin 12 GPS units within a grid based on UTM (Universal Transverse Mercator) coordinates. This survey located over 1,000 ‘sites’, varying from lithic scatters to settlement structures and enclosures dating to all archaeological periods from prehistoric times to the recent past. The investigation of a representative sample of these sites, to attempt to refine our understanding of their chronological and functional patterning, formed the primary focus of the archaeological fieldwork in the final (2000) season. The archaeological survey is drawing on the results of another component of the project—a programme of ethnoarchaeological research (Fig. 4.4). This involves elucidating how farmers (fellaheen) and pastoralists (bedouin) exploit the landscape of the study area and adjacent zones of the Wadi Arabah and plateau today, and how they have done so in the recent past. Within the
Farmers, herders and miners in the Wadi Faynan, southern Jordan
67
Figure 4.4 Ethnoarchaeological survey: the typical site of a winter bedouin tent (beit sha’ar) in Wadi Faynan Note: Gullies to direct rainwater away from the cleared men’s and women’s sections are clearly visible. The men’s hearth is under the tent poles to the left of the photographic scale. The women’s hearth is in the far left-hand corner of the cleared area where there are fire-blackened stones. To the rear, there is a thick, dark accumulation of animal dung where the goats were housed at night. Scale: 1 m Photograph: C.Palmer study area, planning recently abandoned bedouin structures and analysing their floor sediments, combined with interviewing the families who used the structures, is establishing archaeological signatures of seasonality and different age and gender groups, to inform our interpretations of the settlement archaeology. The final major component of the project’s methodology is the development of a Geographical Information System integrating all the above data. This is attempting to refine further our understanding of changing relationships between arable, pastoral and industrial activities; between the ‘economic’, ‘social’ and ‘ritual’ landscapes that are being defined; and between all these cultural activities and the development of the natural landscape, which is the core issue of the project’s research agenda.
THE NEOLITHIC: EARLY FARMING c.9500–4000 BC The transition from the Pleistocene to the Holocene or Postglacial, the modern climatic era, occurred at approximately 9500 BC. Our geomorphological studies have established
The archaeology of drylands
68
that, as elsewhere in the region, the Pleistocene late glacial environment was cold and dry. We then have strong and widespread evidence from sediment sequences and the fauna and flora within them that the early Holocene was characterized by a significantly wetter environment than today, which lasted until about 6,000 years ago. The Near East has some of the earliest evidence in the world for agriculture, which can be recognized very soon after the beginning of the Holocene—the culmination of trends in settlement, subsistence and social change that can be observed amongst the Natufian peoples of the late Pleistocene, following the peak of glacial conditions c.20,000 years ago (Sherratt, 1997). After about 15,000 years ago there was a sudden dramatic warming, and the Natufians were able to develop semi-permanent settlements by lakes and springs in the Jordan valley (the ‘Levantine corridor’). Excavations show that Natufian settlements such as Jericho and Abu Hureyra were sustained by a combination of fishing, fowling, hunting (especially gazelle), collecting forest foods in the valley and gathering wild cereals and other grasses on the steppelands above (Hillman, 1996). With the return to cold and dry conditions termed the Younger Dryas (11000–9500 BC), the steppelands returned to being a resource-poor environment and lake levels shrank. Natufians responded in various ways (Bar-Yosef and Belfer-Cohen, 1992): some moved, others diversified their subsistence, but in the Levantine corridor in particular the signs are that people began to concentrate even more on spring-side locations and on collecting cereals, perhaps engaging in activities that can be regarded as incipient horticulture. So far we have only lithic evidence (flint implements found on the surface) for the presence of Natufian hunter-fisher-gatherers in the Wadi Faynan, but, significantly, most of it has been found in the upper tributaries near the springs (Fig. 4.5). With the beginning of the Holocene c.9500 BC there was a sudden return to warmer conditions, coinciding with the first generally accepted evidence in the Levantine corridor that the main settlements (Pre-Pottery Neolithic A or PPNA) were being sustained at least in part by the cultivation of plants, though wild seeds and fruits continued to be staple foods augmented by fishing, fowling and hunting a variety of game (Bar-Yosef and Kislev, 1989; Byrd, 1992). Sedentary mixed farming, in which goat herding was combined with cereal cultivation, then developed throughout the Near East about 1,000 years later (the Pre-Pottery Neolithic B or PPNB, c.8500–6500 BC), coinciding with major changes in architecture (the appearance of substantial square or rectangular dwellings rather than the circular or oval rubble shelters of Natufian and PPNA sites). PPNA and PPNB sites were invariably located by springs, presumably because the latter provided naturally irrigated land for cereal fields (Bar-Yosef, 1995). This transformation is exemplified in the development of the well-known PPNA and PPNB settlement of Beidha (Byrd, 1994; Kirkbride, 1966), on the plateau near Petra, but is also clear from current investigations of PPNA and PPNB settlements in the Wadi Faynan. A PPNA site of simple rubble shelters and pits has evidence for mixed hunting, gathering and plant cultivation (Finlayson and Mithen, 1998). The inhabitants of a substantial PPNB settlement of well-built stone houses were mixed farmers, growing wheat, barley and legumes, and herding domestic sheep and goats (Simmons and alNajjar, 1996). The two sites are only 100 m apart, in the Wadi Ghuwayr at the junction between the mountains and the main wadi, by the spring where Natufians also camped (Fig. 4.5, upper photograph). We have also found traces of similar settlements in the
Farmers, herders and miners in the Wadi Faynan, southern Jordan
69
upper Wadi Dana, by the main spring there. Although these first agricultural communities clearly preferred the well-watered upper tributaries for their primary settlements, other lithic scatters indicate that they also used the main wadi, presumably for hunting and herding. By the sixth and fifth millennia BC, the zone of principal arable settlement had expanded out into the main wadi. Excavations a few years ago revealed a late neolithic/early chalcolithic settlement of simple rectangular drystone houses at Tell Wadi Faynan, 1 km west of Khirbet Faynan (al Najjar et al., 1990; Fig. 4.5, lower photograph). Our geomorphological investigations show that, when these people settled at Tell Wadi Faynan, the climate was significantly wetter than either before or afterwards: there was a more or less perennial stream by the site—the archaeological sediments contain, for example, frustules of the diatom Navicula, a freshwater organism, and the pottery and mortar contain fragments of reeds and grass as well as straw. The likelihood is that the primary farming zone was able to expand from the tributaries to the main wadi floor at this time, because the climatic amelioration allowed farmers to exploit the seasonal floodwaters of the main wadi for their crops with methods akin to those used by the earlier neolithic farmers in the upper tributaries. Presumably they sowed their crops on either side of what was then the Faynan stream, after spring floods had soaked the soils on either side of its course.
The archaeology of drylands
70
Figure 4.5 The settlement locations of the first farmers in the Wadi Faynan Note: (above) Looking east from near Khirbet Faynan up the Wadi Ghuwayr: the PPNB settlement was on the low terrace immediately above, and the PPNA settlement on the low terrace immediately to the right of the wadi channel where it issues from the hills at the centre of the picture (the spring is behind the PPNB settlement); (below) looking west from Khirbet Faynan towards the Wadi Arabah: the late neolithic settlement of Tell Wadi Faynan is the prominent cliff at the channel edge in the distance on the right-hand side of the photograph; when it was occupied there was a perennial stream flowing down the wadi Photographs: G.Barker
Farmers, herders and miners in the Wadi Faynan, southern Jordan
71
THE BRONZE AGE, c.4000–1200 BC: THE BEGINNINGS OF METALLURGY AND DRY FARMING Aridification began to develop in the fifth millennium BC, leading to the development of a relatively steppic landscape by the fourth and third millennia BC, the period of the Early Bronze Age. This was a period of immense social change in the Near East, characterized by the development of metallurgy, long-distance trade networks and, in some regions, complex polities with quasi-urban settlements (Finkelstein, 1995; Gophna, 1995). Nuggets of surface copper were collected by neolithic (and probably earlier) people in Wadi Faynan, presumably for ornamental purposes, and Faynan copper was exploited by chalcolithic people and traded with the surrounding region. However, the first clear evidence for the systematic mining of the copper ores and their processing at Faynan settlements is in the Early Bronze Age (Adams and Genz, 1995; Hauptmann, 1989; Wright et al. 1998). This was the context for the emergence of local elites who controlled copper production and the exchange to other regions of smelted copper ores and finished artefacts. The research by the Bochum Mining Museum suggests that, at first, ores visible at the surface were mined by open-cast methods and then smelted in simple crucibles in the settlements, but as demand increased, deeper ores were mined by galleries and then smelted in smelting ovens located on the windward side of ridges near the settlements. Our geochemical analyses of sediments at Tell Wadi Faynan indicate that these smelting activities caused small-scale localized pollution. The primary settlement zone shifted during this period into the main wadi and expanded throughout it. The survey has revealed a series of discrete zones of bronze age settlements associated with field systems, both on the southern side of the wadi within the area demarcated by the later classical field system and in the small tributary wadis on the northern side. One zone of the classical field system encompasses the most substantial of these settlements, where excavations by Dr Karen Wright have revealed evidence for irregularly built drystone structures together with enclosures, middens, pits and storage bins, and evidence of working smelted copper into ingots and finished artefacts (Wright et al. 1998). In the area of this settlement, we have been able to recognize a series of boulder walls, within and underlying the later field system, that appear to be vestiges of bronze age structures and field boundaries, some of the latter terraced (Fig. 4.6). They are associated with circular or oval cisterns 30–50 cm deep, fed by short feeder walls. The northern settlement zones include sequences of roughly built terrace walls and check dams built across the shallow floors of tributary wadis, with pottery in associated sediment sequences indicating a bronze age date (Fig. 4.7). In the hills to the north, the 1999 survey found a series of small sites with crudely built one- and two-roomed structures with circular enclosures, presumed to be pastoral encampments (they have analogies with the pastoral encampments of the Negev: Chapter 3), and a few larger, more complex
The archaeology of drylands
72
Figure 4.6 Part of the Wadi Faynan field system WF4, showing (above) the early bronze age and (below) the classical landscape in unit WF4.13 settlements near ridge-top spreads of bronze age slag and furnace waste, presumed to have had—at least in part—an industrial function. The indications are, therefore, that early bronze age settlement in Wadi Faynan was characterized by three different archaeological complexes linked to three overlapping but diverse economic orientations: agricultural, pastoral and metallurgical. Whereas neolithic farmers in the Wadi Faynan were able to exploit well-watered locations, bronze age
Farmers, herders and miners in the Wadi Faynan, southern Jordan
73
farmers were having to develop
Figure 4.7 A field system on the northern side of the Wadi Faynan Note: The outer diversion walls and many of the field walls in the centre very probably date to the Iron Age, although potsherds in sediment exposures indicate that some of the simple check dams at points A and B are bronze age. strategies for coping with the more arid environments evidenced for the fourth and third millennia BC, such as building walls to collect and trap seasonal floodwaters in storage cisterns and in terraced fields laid out along the direction of water flow. If correctly dated (and the dating is still tenuous), this will be the earliest evidence for floodwater farming yet found in the Near East, making this another indicator of the social and economic transformations that characterized this phase of settlement in the region. A degree of pastoral specialization may have been another way in which bronze age societies were able to respond to aridity, whilst also being, like metallurgy, an indicator of complex economic structures of production and exchange. What is also interesting is that we have evidence for strong soil erosion through the second and first millennia BC and palynological indicators that this reflects the impact of human activities on the landscape such as clearance of fuelwood for smelting, and intensification in systems of cultivation and herding, rather than climatic change. What happened in the Wadi Faynan during the second millennium BC (the Later Bronze Age) remains unclear, as throughout the entire region. The end of the millennium was marked by the widespread disintegration of urban polities throughout the East Mediterranean and Levant. There is some evidence for climatic deterioration: this is the
The archaeology of drylands
74
period of the Thera, or Santorini, eruption in the Aegean Sea—and Egyptian scribes make references to raiding by ‘Sea Peoples’, so there has been a lively debate about whether ‘external’ environmental or cultural factors such as these caused economic collapse or whether (more likely) they exacerbated internal processes of social change that were already in train. In the Levant it has commonly been argued that societies turned to nomadic pastoralism at this time (Finkelstein, 1995; LaBianca, 1990), though convincing archaeological signatures for such pastoralism are unclear (Bar-Yosef and Khazanov, 1992). The debate is further obscured by Biblical archaeologists and historians looking for the nomadic peoples of Old Testament origin myths. In the ethnohistorical record, furthermore, specialized pastoralism is a highly complex economic system that invariably operates not in isolation but in close relationship with adjacent agricultural and urban systems, as it may have operated in articulation with arable and metallurgical activities at Faynan in the Early Bronze Age. It is, therefore, difficult to separate absence of settlement evidence from evidence of settlement absence in the Wadi Faynan at this time, but it does seem likely that smaller-scale systems of mixed cultivation and herding characterized life in the wadi during the second millennium BC.
THE IRON AGE c.1200–300 BC: TRANSFORMATIONS IN SETTLEMENT, FARMING AND MINING The early first millennium BC saw the emergence of iron age states west of the Jordan and tribal kingdoms in the Jordan valley and steppeland to its east: Ammon in northern Jordan, Moab in central Jordan and Edom in southern Jordan (LaBianca and Younker, 1995). The State of Edom, in which Wadi Faynan was situated, was the least densely settled of the three. Although historians have accepted Josephus’ statement that these kingdoms were destroyed by the Neo-Babylonians in the sixth century BC, archaeological surveys in fact indicate continuity in rural settlement, for example around Madama in Ammon (LaBianca, 1990). In the Wadi Faynan, the iron age landscape was quite different from that of the Bronze Age, dominated by a single substantial settlement, site WF424 in our survey record, built immediately below its successor, Khirbet Faynan, at the strategic centre of the Faynan region, at the point where the three major tributary wadis come together to form the main channel of the Faynan. We have also found zones of iron age settlement along the southern margins of the field system and on the northern side of the wadi. The evidence suggests, therefore, that the Edomite settlement system consisted of a few large and discrete habitation units, probably organized hierarchically, with WF424 the dominant site. Recent work in the neighbouring Wadi Fidan indicates that there may have been other more ephemeral settlement forms as well (Levy et al., 1999). By this period, deep ores were being mined extensively in the hills and then smelted at settlements such as WF424, where we found thick deposits of slags, charcoal-rich unlike the bronze age slags, suggesting experimentation with new technologies to deal with the far larger quantities of ore being processed for the Edomite economy. Geochemical analyses confirm the increasing scale of smelting pollution.
Farmers, herders and miners in the Wadi Faynan, southern Jordan
75
WF424 was associated with a field system of boulder-built walls, often set orthostatically, and also with substantial boundary walls built upstream of these fields along the junction between the hill and the wadi floor. These boundary walls collected water from the surrounding slopes and channelled it to exit sluices above the terraced fields, so that maximum water flow could be directed down the central part of the field system. Similar boundary walls enclosed iron age field systems on the northern side of the Wadi Faynan, and in part at least they had a water-diversion function (Fig. 4.7). We cannot be sure of the dating of these boundary walls, but the fact that we have only found them enclosing field systems with significant iron age material, and the constructional similarities between the walls and the fields they enclose, suggest an iron age date. If this dating is correct, it implies that, whereas bronze age farmers built simple terrace walls at right angles across wadi beds to check floodwater flow and to try to spread it laterally over surrounding fields, together with small catchments to collect water in cisterns, iron age farmers in Faynan had learned to construct substantial and rather sophisticated walls to divert the flow of floodwaters sometimes hundreds of metres from their natural line, so that far greater quantities of water could be collected and sent down a field system than was possible with bronze age technology.
NABATEAN SETTLEMENT, c.300–63 BC The Nabatean state, with its capital Petra, developed at the time that Rome’s power was expanding across the eastern Mediterranean in the last three centuries BC, and flourished until Palestine was annexed by Rome in 63 BC. The Nabatean settlement system in the wadi, like that of the Iron Age, was dominated by one central settlement: the community at Khirbet Faynan. This settlement presumably controlled copper production, which, on the evidence of both mining archaeology and sediment geochemistry, increased dramatically in scale in the later centuries of the first millennium BC. The landscape now consisted of a series of adjacent settlement units, organized in some kind of hierarchical relationship with respect to Khirbet Faynan: a series of large farmsteads of broadly Nabatean date on the southern slopes overlooked the classical field system; excavations by Wright et al. (1998) found small buildings of Nabatean date in the area of major bronze age settlement within the classical field system (Fig. 4.6); and our survey has identified a variety of structures with similar pottery elsewhere within the field system. The technology of floodwater farming was further refined by Nabatean farmers. The particular focus of their wall-building activities was a series of small tributary wadis that run parallel to the main wadi along its southern side (Fig. 4.8). Water was dammed as it issued from the adjacent hills, diverted westwards by boulder walls along the contour of the slope, and then channelled through simple sluices (gaps) and spillways (stepped structures) onto terraced fields below. Nabatean technology on the southern slopes may also have included channels formed by parallel walls that fed water directly into the fields on either side through sluice gaps.
The archaeology of drylands
76
THE ROMAN IMPERIAL LANDSCAPE, 63 BC-c.AD 600 With the Roman annexation of Palestine, Faynan became one of the principal suppliers of copper and lead for its eastern empire, with its extraction probably state-organized. A garrison was located at Khirbet Faynan, the Phaino of the classical sources, and the surrounding hills were honeycombed with deep mine shafts. There are references to Christian slaves being sent to work the mines as punishment (according to some sources they were crippled to prevent their escape by blinding, having a hand removed, or having their Achilles tendon severed), though most of the mining and smelting was probably done by a workforce paid well for skilled and dangerous work. The quantities of ore being smelted by the Phaino labour gangs have left kilometre-long spreads of tap slag on the ridges above Khirbet Faynan. The EDMA geochemical studies of the heavy metals found in a 2,500-year long sequence of sediments behind a barrage at Khirbet Faynan indicate extraordinary levels of air pollution in Roman and Byzantine times, with levels several times lethal in terms of modern pollution criteria (Fig. 4.9).
Figure 4.8 A field map of part of the field system WF4 after ground verification by the survey teams Note: The photograph of Khirbet Faynan shown as Figure 4.2 is looking across the fields mapped as Units 4, 3 and 2 in this plan. The upper (southern) part of this field system was probably laid out by Nabatean farmers, who diverted water as it flowed out of tributary wadis onto the upper terraces, for example diverting water at point F from its channel F-G along the wall F-H, and then through sluices onto the terraced fields below. The entire system was managed as an integral unit by Roman farmers, who also built the mill complex
Farmers, herders and miners in the Wadi Faynan, southern Jordan
77
The satellite farms were abandoned, leaving Khirbet Faynan as the single dominant settlement. Our studies of field layout and construction, and of the surface material, indicate that the entire agricultural landscape was now managed as a more or less integral unit or estate. Systems of long parallel walls were built to divert water from the main wadi into adjacent fields on low-elevation terrain, and from the southern tributary wadis (Fig. 4.10). Further down the tributary wadis, similar channels were built at c.45 degrees to water flow to collect any water that had bypassed the higher diversion walls or drained back into the wadis from the higher terraced fields, to force it once more onto adjacent cultivable land. The effectiveness of the system is in part explained by the uniformly low levels of water infiltration we have found at sample sites from the upper slopes to the lowest fields, but organizational factors were also important. The field evidence supports the hypothesis of cooperation between areas of the field system fed by the parallel channels: rather than farmers with land upslope having exclusive access to the floodwaters of particular wadis at the expense of other farmers with land
Figure 4.9 The distribution of copper (in parts per million—ppm) through the sediments that accumulated behind the Khirbet Faynan barrage Note: The sequence extends from c.2,500 years ago (far right) to the present day (far left) further down the direction of flow, the internal linkages between the system imply that water resources were shared down the length of the field system. The construction of major parallel channels to carry floodwater through the system demonstrates the same engineering skills in moving water relatively long distances over gentle gradients as are displayed by the Roman engineers who designed the rock-cut feeder channel that brought water several kilometres from the Wadi Ghuwayr spring to the aqueduct feeding the reservoir and ore-crushing mill near Khirbet Faynan.
The archaeology of drylands
78
This imperial landscape was highly organized, with large-scale industrial processing sustained by an integrated agricultural and hydraulic system. But it was also increasingly barren, fast eroding and grossly polluted. Charcoal samples from the smelting sites studied by the German team show that, whereas Nabatean miners cut local firewood for their smelting activities, by the Roman period timber was having to be brought down from the plateau because local supplies had been exhausted (Engel, 1993; Hauptmann, 1992). Our pollen evidence indicates the same process of humanly induced desertification: by the time of Christ the landscape consisted of very degraded steppeland, and this degradation then accelerated significantly through the first millennium AD. By the end of the Roman period, the steppic component of the pollen diagrams collapses, evidence of olive cultivation disappears, and signs of cereal cultivation drastically reduce—the flora at this time was analogous to the modern pollen rain in the Dead Sea. We have also found widespread evidence that Roman farmers were trying to combat the effects of wadi-downcutting in
Farmers, herders and miners in the Wadi Faynan, southern Jordan
79
Figure 4.10 Section through a Roman-period water conduit channel, visible on the surface as two parallel walls between the ranging pole in the foreground and the right-hand tree in the distance Note: The section shows that the parallel walls on the surface overlie buried walls of an ancient conduit filled with water-lain sediments that contained Roman pottery Photograph: G.Barker their alterations to the floodwater farming systems, though these were ultimately ineffective (the main wadi now flows at least 5 m below the parallel channel systems that diverted the Faynan floodwaters into the ancient fields).
The archaeology of drylands
80
THE POST-BYZANTINE LANDSCAPE The desertic environment has persisted to the present day, though there is evidence for an episode of even greater aridity in the period c.AD 1600–1850. The nature of the Islamic and later settlement systems following the eventual abandonment of Khirbet Faynan remains unclear, but it is at least evident from our survey work so far that it was not characterized by a renewal of substantial settlement within the field system zone, and there are indications of ephemeral settlements on the surrounding slopes akin to the sites of recently abandoned Bedouin encampments. Small-scale increases in smelting pollutants in levels above the post-Byzantine collapse (such as the peak at c.90 cm depth in Figure 4.9) indicate a revival of industrial activity at some time probably in the early second millennium AD (radiocarbon dates are awaited), and the range of pollutants suggest the reworking of Roman and Byzantine slag deposits rather than renewed mining on any scale. The likelihood is that the degraded landscapes of the post-Byzantine period have for the greater part supported only systems of land use like those of the bedouin in the region today.
CONCLUSION As described above, we are beginning to detect oscillations in environmental change, settlement forms and agricultural and industrial activity over the past 10,000 years. If we can understand how they do and do not inter-relate, we should be able to write a landscape history in the Braudelian sense of complex interactions between short-term processes, medium-term processes and the longue durée that can provide a significant archaeological contribution to the desertification debate. In terms of environmental change, after the wetter phase of the early Holocene we can discern a principal trend of progressive aridification and degradation, culminating in extremely degraded landscapes by the mid-first millennium AD. However, it is clear that the trend was not constant in its progression and that it contains oscillations. In terms of land use, from the Late Neolithic onwards a number of increasingly sophisticated systems of water control can be discerned, but again it is clear that there is no simple progression in land use from simple to complex, but rather oscillations between the two. Another complex non-linear sequence is emerging regarding the impact of people on landscape. The expansion of farming down the wadi in the Later Neolithic appears to have had no significant impact on the landscape, but it is possible that the erosion we can detect during and after the Early Bronze Age, whilst probably mainly a response to aridification, partly reflects poor land management techniques. Given the evidence of the geochemistry for the beginning of smelting pollution at this time, wood cutting for metallurgical processing may also have been a factor. However that may be, it is clear that the demands of Nabatean and, in particular, Roman/Byzantine mining, in parallel with the intensive agricultural practices developed to feed the workforce, had an ultimately devastating impact on the landscape. Whether or not climatic change was also a factor, it is clear that
Farmers, herders and miners in the Wadi Faynan, southern Jordan
81
large-scale stripping of the landscape of vegetation made it extremely vulnerable to erosional forces. The geochemical evidence also demonstrates that the effects of Roman and Byzantine mining and smelting are still felt today—the milk, urine and faeces of the bedouins’ goats today have significant levels of heavy metals from grazing the polluted ground, and cereal growth is also badly affected around the smelting sites. Does the extraordinary density of Roman and Byzantine potsherds carpeting the field system indicate large-scale manuring in an attempt to deal with falling cereal yields? Certainly there is a strong possibility (currently being tested by skeletal analysis) that the health of the Roman and Byzantine population in particular was directly affected from inhalation, skin contamination and bioaccumulation of polluted animal and plant foods. Whilst the collapse of intensive farming and mining in Late Roman/Byzantine times no doubt in part reflects changing economic relations between Faynan and the wider world, it seems inescapable that the activities of these farmers and miners had a profound impact on their landscape (which has still not recovered) and probably directly on their own well-being. The project has succeeded so far in establishing the principal components of the environmental and settlement sequences in the Wadi Faynan. What we can tell so far of their inter-relationships indicates something of the potential complexity of the interplay between long-term, medium-term and short-term processes that is likely to emerge as the project develops. However, the richness of the data also suggests that, especially through Geographical Information Systems (GIS), we should be able to integrate our findings on how the landscape has changed, and the role of people in this, to investigate also how different populations in the past perceived their changing landscape and their place within it. It should be possible, for example, to model the spatial characteristics of air pollution at different distances from the major smelting settlements and its different effects on surrounding populations. There is much still to be learned about the extent to which farmers using the field system operated independently or collaborated, and, in the latter case, whether they did so by cooperation or coercion. There is also the question of how the sacred and the secular related to one another in different periods. There is intriguing evidence, for example, that whereas bronze age people kept their cemeteries and fields apart, classical farmers deliberately constructed water diversion systems so that they incorporated pre-existing burial cairns at key nodal points. Reconstructing landscape histories needs natural scientists to analyse changing forms of landscape, and archaeologists to analyse changing settlement morphologies and systems. We hope that the effective partnership of disciplines in our project will allow us ultimately not just to describe these two landscape histories for Faynan but also to understand their interactions, looking at the perceptions and choices that underpinned human actions in this landscape and shaped the latter with ultimately devastating consequences.
ACKNOWLEDGEMENTS This chapter represents the work of the Wadi Faynan Landscape Survey team, especially: R.Adams (University of Bristol: prehistoric pottery analysis), O.Creighton (University of Exeter: archaeological survey), P.Daly (University of Oxford: archaeological survey and
The archaeology of drylands
82
GIS analysis); D.D.Gilbertson (University of Bournemouth: geomorphology); J.P.Grattan (University of Wales Aberystwyth: geomorphology, geochemistry); C.O.Hunt (University of Huddersfield: palynology); D.J.Mattingly (University of Leicester: archaeological survey); S.J.McLaren (University of Leicester: geomorphology), H.Mohammed (University of Benghazi, Libya: palynology); P.Newson (University of Leicester: archaeological survey and GIS analysis), C.Palmer (University of Leicester: ethnoarchaeology), H.Parton (British Institute at Amman for Archaeology and History: Roman and post-Roman pottery analysis); F.B.Pyatt (Nottingham Trent University: environmental biology, geochemistry), T.E.G.Reynolds (Cambridgeshire County Council: lithic analysis); H.Smith (University of Bournemouth: ethnoarchaeology, environmental archaeology); R.Tomber (Museum of London: Roman and post-Roman pottery analysis); and A.Truscott (University of Wales Aberystwyth: OSL dating)— together with the archaeological field team that has done most of the foot work! Grateful thanks are also due especially to the Arts and Humanities Research Board, the Council for British Research in the Levant, the Natural Environment Research Council, the Society of Antiquaries of London, and the Universities of Leicester, Huddersfield and Aberystwyth for funding the project, and to the Jordanian Department of Antiquities, CBRL’s British Institute at Amman for Archaeology and History and the Royal Society for the Conservation of Nature for other essential support.
REFERENCES Adams, R. and Genz, H. (1995) Excavations at Wadi Fidan 4: a chalcolithic village complex in the copper ore district of Feinan, southern Jordan. Palestine Exploration Quarterly 127:8–20. Barker, G., Gilbertson, D., Jones, B. and Mattingly, D. (1996) Farming the Desert. The UNESCO Libyan Valleys Archaeological Survey. Volume One: Synthesis . Paris, UNESCO; London, Society for Libyan Studies; Tripoli, Department of Antiquities. Barker, G., Creighton, O.H., Gilbertson, D.D., Hunt, C.O., Mattingly, D.J., McLaren, S.J. and Thomas, D.C. (1997) The Wadi Faynan Project, southern Jordan: a preliminary report on geomorphology and landscape archaeology. Levant 29: 19–40. Barker, G., Adams, R., Creighton, O.H., Gilbertson, D.D., Grattan, J.P., Hunt, C.O., Mattingly, D.J., McLaren, S.J., Mohammed, H.A., Newson, P., Reynolds, T.E.G. and Thomas, D.C. (1998) Environment and land use in the Wadi Faynan, southern Jordan: the second season of geoarchaeology and landscape archaeology (1997). Levant 30:5– 26. Barker, G., Adams, R., Creighton, O.H., Crook, D., Gilbertson, D.D., Grattan, J.P., Hunt, C.O., Mattingly, D.J., McLaren, S.J., Mohammed, H.A., Newson, P., Palmer, C., Pyatt, F.B., Reynolds, T.E.G. and Tomber, R. (1999) Environment and land use in the Wadi Faynan, southern Jordan: the third season of geoarchaeology and landscape archaeology (1997). Levant 31:255–92. Barker, G., Adams, R., Creighton, O.H., Daly, P., Gilbertson, D.D., Grattan, J.P., Hunt, C.O., Mattingly, D.J., McLaren, S.J., Newson, P., Palmer, C., Pyatt, F.B., Reynolds, T.E.G., Smith, H., Tomber, R. and Truscott, A.J. (2000) Archaeology and
Farmers, herders and miners in the Wadi Faynan, southern Jordan
83
desertification in the Wadi Faynan: the fourth (1999) season of the Wadi Faynan Landscape Survey. Levant 32:27–52. Bar-Yosef, O. (1995) Earliest food producers—pre-pottery neolithic (8000–5500 BC). In T.Levy (ed.) The Archaeology of Society in the Holy Land : 190–204. London, Leicester University Press. Bar-Yosef, O. and Belfer-Cohen, A. (1992) From foraging to farming in the Mediterranean Levant. In A.B.Gebauer and T.D.Price (eds) Transitions to Agriculture in Prehistory : 21–48. Madison, Prehistory Press, Monographs in World Archaeology 4. Bar-Yosef, O. and Khazanov, A. (1992) (eds) Pastoralism in the Levant: Archaeological Materials in Anthropological Perspectives . Madison, Prehistory Press, Monographs in World Archaeology 10. Bar-Yosef, O. and Kislev, M.E. (1989) The pre-pottery neolithic B period in eastern Jordan. Paléorient 15 (2):150–6. Byrd, B. (1992) The dispersal of food production across the Levant. In A.B.Gebauer and T.D.Price (eds) Transitions to Agriculture in Prehistory : 49–61. Madison, Prehistory Press, Monographs in World Archaeology 4. Byrd, D. (1994) Public and private, domestic and corporate: the emergence of the southwest Asian village. American Antiquity 59 (4):639–66. Engel, T. (1993) Charcoal remains from an iron age copper smelting slag heap at Feinan, Wadi Arabah (Jordan). Vegetation History and Archaeobotany 2:205–11. Finkelstein, I. (1995) Living on the Fringe: The Archaeology and History of the Negev, Sinai and Neighbouring Regions in the Bronze and Iron Ages . Sheffield, Sheffield University, Monographs in Mediterranean Archaeology 6. Finlayson, B. and Mithen, S. (1998) The Dana-Faynan (South Jordan) Epipalaeolithic Project: report on reconnaissance survey, 14–22 April 1996. Levant 30: 27–32. Gophna, R. (1995) Early bronze age Canaan: some spatial and demographic observations. In T.Levy (ed.) The Archaeology of Society in the Holy Land : 269–80. London, Leicester University Press. Hauptmann, A. (1989) The earliest periods of copper metallurgy in Feinan, Jordan. In A.Hauptmann, E.Pernicka and G.A.Wagner (eds) Archaemetallurgie det Alten Weltt/Old World Archaeometallurgy : 119–36. Bochum, Deutsches Bergbau-Museum, Der Anschnitt, Beiheft 7. Hauptmann, A. (1992) Feinan/Wadi Feinan. American Journal of Archaeology 96:510– 12. Hauptmann, A. and Weisgerber, G. (1987) Archaeometallurgical and miningarchaeological investigations in the area of Fainan, Wadi ‘Arabah (Jordan). ADAJ 31:419–37. Hauptmann, A., Begemann, F., Heitkemper, E., Pernicka, E. and Schmitt-Strecker, S. (1992) Early copper produced at Feinan, Wadi Araba, Jordan: the composition of ores and copper. Archaeomaterials 6:1–33. Hillman, G.C. (1996) Late Pleistocene changes in wild plant-foods available to huntergatherers of the northern Fertile Crescent: possible preludes to cultivation. In D.R. Harris (ed.) The Origins and Spread of Agriculture and Pastoralism in Eurasia : 159– 203. London, UCL Press.
The archaeology of drylands
84
Kirkbride, D. (1966) Five seasons at the pre-pottery neolithic village of Beidha in Jordan. Palestine Exploration Quarterly 98:8–72. LaBianca, Ø.S. (1990) Hesban 1: Sedentarization and Nomadization . Berrien Springs (MI), Andrews University Press. LaBianca, Ø.S. and Younker, R.W. (1995) The kingdoms of Ammon, Moab and Edom: the archaeology of society in late bronze age/iron age Transjordan (ca.1400–500 BCE). In T.Levy (ed.) The Archaeology of Society in the Holy Land : 399–415. London, Leicester University Press. Levy, T., Adams, R. and Shafiq, R. (1999) The Jebel Hamrat Fidan Project: excavations at the Wadi Fidan 40 cemetery, Jordan (1997). Levant 31:293–308. al-Najjar, M., Abu Dayyeh, A. es-S., Suleiman, E., Weisgerber, G. and Hauptmann, A. (1990) Tell Wadi Feinan: a new pottery neolithic tell in southern Jordan. Annual of the Department of Antiquities of Jordan 34:27–56. Sherratt, A.G. (1997) Climatic cycles and behavioural revolutions: the emergence of modern humans and the beginning of farming. Antiquity 71:271–87. Simmons, A.H. and al-Najjar, M. (1996) Test excavations at Ghwair I, a neolithic settlement in the Wadi Feinan. ACOR Newsletter 8.2:7–8. Wright, K., Najjar, M., Last, J., Moloney, N., Flender, M., Gower, J., Jackson, N., Kennedy, A. and Shafiq, R. (1998) The Wadi Faynan Fourth and Third Millennia Project, 1997: report on the first season of test excavations at Wadi Faynan 100. Levant 30:33–60.
5 Differing strategies for water supply and farming in the Syrian Black Desert PAUL NEWSON
INTRODUCTION Increasing evidence has been gathered through the twentieth century of largescale settlement across the high plateau of the Jebel al-Arab in Syria, part of the so-called Hauran, during the Roman period, between the first and the third centuries AD (Tate, 1997:55; Fig. 5.1). This region is within the 200 mm rainfall isohyet, which is the accepted limit for dry farming without irrigation. Especially intriguing, however, is evidence for apparently permanent settlements of the same antiquity on the desert plateau beyond in the region long called the al-Harra (‘Burnt Land’ in Arabic) by the bedouin and the Black Desert by European travellers (Braemer et al., 1996b:1). The water management features associated with some of these sites provide the focus of this chapter. A number of reasons to explain the development of permanent settlement in the Harra in the Roman period has been suggested, ranging from significant changes in climate (that is, increased rainfall compared with today) to the sedentarization of previously nomadic tribes. Certainly the new pattern of settlement did not endure: by the Late Byzantine and Early Islamic periods (the seventh to ninth centuries AD), most of the settlements had dwindled drastically in size or had been abandoned. The region where these ancient remains are to be encountered is within the Syrian desert (the Badiyat al-Sham)—a huge area of arid plateau (Hamada) forming the northern part of the Arabian steppe bounded by more fertile regions to the west (the Mediterranean littoral), the north (the Taurus foothills), and the east (the River Euphrates), and by the Nafudh desert to the south. The average annual rainfall within this region declines with latitude, from c.150 mm in the north to less than 50 mm in the south. The topography of the region is dominated by an extensive area of lava flows and basalt rocks up to 100 km wide and reaching 250 km in length towards the southeast, making it difficult to traverse (Helms, 1981:19). The lava emanated from a number of fissures in at least six successive flows over a relatively short period of time. These flows solidified into layers of hard basalt,
The archaeology of drylands
86
Figure 5.1 The Hauran and the Harra regions of Syria, and other regions and sites discussed in Chapter 5 each on average 30 m thick. The resulting flattish plateau is broken by a number of fissure cones and dissected by a series of wadis generally flowing in an easterly direction. The wadis radiate out from the high relief of the Jebel al-Arab, cutting across the lava flows, and have long formed the main lines of communication across this difficult area.
Differing strategies for water supply and farming
87
In a quite waterless region they also provide the main access points to water, so naturally form the main foci for settlement. The Black Desert has experienced limited permanent settlement in certain periods, linked to the utilization of water-harvesting techniques that have been characterized by varying degrees of sophistication. Some of the earliest systems have been dated to the third millennium BC, notably at Jawa on the southeastern edge of the Jabel al-Arab (Helms, 1981) and on a smaller scale at Khirbet el-Umbashi to the northeast (Braemer et al., 1996a). However, this chapter concentrates on the methods of water management that current evidence suggests were used at dryland settlements in the Harra during the Roman period, taking three sites as case studies where I have conducted fieldwork. The first of these sites, ad-Diyatheh, is located on the western edge of the Harra near the steep descent from the Jebel Al-Arab. The second site, al-Namara, is located in the middle of the northern part of the Harra at the confluence of the Wadi Sham and a tributary wadi. The third site, Qasr Burqu’, is on the eastern edge of the Harra near its junction with the Ruhba the large fertile alluvial plain that, in season, ‘has provided highly-prized grazing for nomads from time immemorial’ (Braemer et al., 1996b:1).
FARMING THE BLACK DESERT: THREE CASE STUDIES Ad-Diyatheh This site comprises a number of connected elements based around the well-preserved remains of a small Roman fort (Fig. 5.2). Ad-Diyatheh lies at the junction between the settled Hauran in the west and the Hamada steppelands to its east. This demarcation line also echoes the important 100 mm precipitation isohyet, which follows the relief edge of the Jebel al-Arab. The site is also situated along the edge of the Wadi Sham, which is one of a number of wadis whose floodwaters have etched deep-sided valleys into the Jebel as they flow eastwards into the Harra. In relation to this wadi, the site of the fort was carefully chosen, being a small flat plateau above the north bank of the wadi, with commanding views across the Harra. Clustered around the fort and along the edge of the north bank of the wadi are the remains of c.100 stone-built structures, all of a similar construction and with a homogeneous layout. Many of these houses are in a very good state of preservation; some have underground rooms with extensions excavated into the wadi bank. It has been convincingly suggested that this collection of structures forms a village of contemporaneous construction (Villeneuve, 1986). Apart from a walled meeting or function area, no public building of any other sort has been identified and there seems to be no indication of a stone-lined birket or reservoir (Sadler, 1993). The latter is contrary to normal expectations, as ancient settlements of the Jebel region invariably have such reservoirs. However, in the wadi bed south of the village and fort, and continuing up onto the right bank, can be seen the entrances to some
The archaeology of drylands
88
Figure 5.2 Plan of the Roman-period settlement of Ad-Diyatheh and its water channels and field systems Key: Stippling marks areas of ancient fields Source: Adapted from Sadler, 1993 stone-lined wells, which are still utilized, tapping into the underground water flow of the Wadi Sham. In addition, there are the extensive remains of stone-cleared fields and walls (Fig. 5.3). The construction of the village houses on two floors, with the lower floor in some buildings containing evidence for stone troughs, suggests an economy based on cattle rearing, as in the areas of the limestone massif in northern Syria (Tate, 1992; Villeneuve, 1986). As the French survey of the village was coming to a close, it was realized that the agricultural operation was on a very large scale. Initially it had seemed like a simple diversion of wadi floodwater onto the Harra plain below the fort and village, but further investigation revealed the remains of a complex floodwater farming system, together with other significant features such as watermills built onto leats extending out onto the Harra plain on the northern side of the wadi course. Sadler (1993) outlined the main features of these field systems in the following terms (Fig. 5.2). At a point 300 m downstream of the village occurs the first major diversion, across the wadi bed, and there are at least two other such diversions situated a further 2 and 3 km downstream (Sadler, 1993:428). The first is by far the best preserved, constituting a long, low barrage of medium-sized stones crossing the wadi bed at an oblique angle to the flow. The barrage leads directly into a long, straight canal constructed on a shallow gradient, which would have allowed the momentum of waterflow to overcome the height difference between the Harra plain and the wadi bed several
Differing strategies for water supply and farming
89
Figure 5.3 The Roman-period settlement of Ad-Diyatheh and its channel walls in the Wadi Sham, viewed from the west Photograph: P.Newson metres below. Although the networks have been eroded somewhat by the action of floodwaters and neglect, the course of the main canal can be identified as running more or less in a parallel direction to the main wadi. As it approaches the plain, a succession of tributary canals leads off it to feed water into the irrigated zones. This process is repeated with the other diversion barrages and their associated primary, secondary and tertiary canals further downstream. Two secondary canals in the first network feed their own selfcontained network of smaller-sized tertiary channelways or canals. These smaller tertiary canals (up to several hundred metres in length) lead off from the secondary canals to feed a number of rock-cleared fields scattered at intervals across the gently sloping plateau. The second of these secondary canals is the longest (around 3 km) and serves a larger but more dispersed field system. The canal heads in a northeasterly direction almost as far as the next wadi coming down from the Jebel, the Wadi Gharaz. Running from this canal in an easterly direction, and at regular intervals along its course, is a succession of parallel tertiary canals up to 800 m in length feeding a large number of rock-cleared spaces that constitute the fields of this sub-network. It has been calculated that the total area that could be irrigated by this first barrage and its associated system of primary, secondary and tertiary canals amounts to around 1,200 ha (Sadler, 1993:431). At the end of each sub-network, there is evidence for the collection of excess water into a small canal, which appears to carry this along to the following network. No water was wasted or allowed to erode the canals or fields by ponding. Around the slopes to the north of the village, along the edges of a shallow valley leading off from the Jebel, evidence for another water-capture strategy has been recorded.
The archaeology of drylands
90
This valley is at too high an altitude to be irrigated by the canals and received too little floodwater to allow cereals to be cultivated. Nonetheless, there are remains of stonecleared fields and the vestiges of long, low parallel walls, some leading down the slopes at regular intervals and others at lower levels forming low terraces (Sadler, 1993:433). This complex appears to represent a different strategy of water management, concerned with collecting surface run-off water from higher up in the valley and leading it in a controlled way down to a terrace system where water and water-borne sediment could be captured and controlled. Similar systems dating to the Roman period have been documented in comparable dryland environments such as the Negev desert in Israel (Evenari et al., 1982) and the pre-desert plateau of Tripolitania in northwest Libya (Barker et al., 1996; and see also this volume, Chapter 8). The fact that much of the central part of this system at Ad-Diyatheh has been destroyed by erosion since it was abandoned testifies to the extent to which this concentration of run-off was successful when it was constructed and maintained (Sadler, 1993:434). The watermills are another striking illustration of the overall success of this scheme of floodwater farming in the Black Desert. These were presumably utilized for grinding the winter cereals that were grown here on a large scale in Roman times and are still cultivated by the local bedouin when the conditions allow. The remains of eight such mills have been located so far. All display a similar construction and are usually located along short mill-race canal sections leading off from the secondary canals. The position of the first two mills immediately below the village settlement at ad-Diyatheh, and the similarities in construction between these and the village houses and other mills that can be found situated amongst the canal networks of the Harra plateau, have led Sadler (1993:435) to suggest that the field system and the village must be contemporary. From a cursory assessment of the pottery from the remains of the fort and village, Villeneuve (1986:713) concluded that the main phase of occupation lasted from the late third century AD to the fifth and perhaps even into the seventh centuries, but this date range is by no means certain. Al-Namara The site of al-Namara (Fig. 5.4) lies some 60 km east of ad-Diyatheh, fully on the Harra plateau, at the confluence of the Wadi Sham and a small tributary called the Wadi Saad. The name al-Namara refers to the large basin etched into the plateau by the confluence of these two wadis; at the basin’s centre is an ‘island’ of resistant rock, the remains of a volcanic (basalt) plug. Water flowing down from the Jebel in the spring floods pools naturally at points in the wadi beds around this island and lasts well into the summer. Therefore, the combination of ample water in an otherwise arid region, and the strategic vantage point of the island, has long attracted local pastoralists as well as the attention of regimes trying to control them. On the flat top of the island or ‘citadel’ are the few remains of reused structures of the Roman army and a later Arab occupation, whilst in the surrounding basin and beyond are the extensive remains of water catchment systems and encampments (Figs. 5.5 and 5.6). Little archaeological investigation has taken place at the site, apart from an initial assessment and topographical plan of the citadel and its immediate environs within a 2×1 km area, completed while work was being done to
Differing strategies for water supply and farming
91
construct a modern reservoir on the wadi course at the eastern end of the site (Braemer et al., 1996b). The Wadi Sham has eroded the Harra plain to an average depth of 15 m and an average width of 200 m. At some points, as a result, the wadi floor has widened to form small alluvial plains 500–1,000 m wide. Al-Namara is situated in an S-bend of the wadi, where the erosive action of the floodwaters has created a small plain measuring some 800 m long by 400 m wide. The island intrudes into this plain, rising to around 10 m in height. The wadi has to curve its course round the island to the south, as a result of which the south bank of the curve has gentle slopes, whereas the north bank forms a cliff some 2–8 m high. Downstream, the banks lower to around 1 m in height, and the distance between them widens into a plain up to 1 km wide. The northern terrace of the plain has substantial evidence for simple irrigation systems up to 1 km long, whereas the southern side has more limited evidence (Braemer et al., 1996b:4).
The archaeology of drylands
92
Figure 5.4 Plan of the main irrigated zone along the Wadi Sham by the Roman-period settlement of al-Namara Source: After Braemer et al., 1996b
Differing strategies for water supply and farming
93
Figure 5.5 View of the main irrigated zone along the Wadi Sham at al-Namara, looking north from the ‘citadel’ Photograph: P.Newson
Figure 5.6 Canal 3 at al-Namara, viewed from the east Photograph: P.Newson
The archaeology of drylands
94
The remains of around five diversion barrages have been located, and these are associated with canals leading from them. The four barrages on the Wadi Sham are of a similar construction, consisting of a line of large othostatic boulders positioned on an outcrop of hard basalt, placed at an oblique angle to the direction of flow of wadi floodwaters. The boulders were bound together with smaller stones placed around them. The low dams thus formed obstructed the wadi course only partially, in order to capture water in a controlled manner without being liable to being destroyed by the force of the flood. (The same thing can be observed at ad-Diyatheh.) Three of the barrages in fact only partially cross the full width of the wadi course (Braemer et al., 1996b:9). The canals that lead off from these barrages seem to have served two purposes. The two canals positioned upstream from the citadel are fairly short in length (around 800 m long) and seem to have been built to capture water from the wadi and bring it to the basin south of the island where the water naturally pools. Downstream of the island, two barrages capture water from the wadi, and their associated canals lead it to areas where fields have been cleared of stones and laid out for irrigation. The first canal (Canal 3 on Figure 5.4) lies on the northern terrace of the wadi and is around 2.5 km long. The length of the other canal (Canal 4) cannot be measured accurately because it has been partially obscured by the new reservoir. The area irrigated by the first canal lies immediately adjacent to the wadi, forming a small terrace up to 100 m in width and roughly 2 km long. This canal follows a similar pattern in its construction to all the other canals, with a channel way 1–2 m wide and a wall 0.5–1 m high of medium-sized boulders edging the downslope of the channel, which follows the contour of the terrace bank (Fig. 5.6). The wall was made waterproof with the packing of smaller stones and earth in the gaps between the larger stones. The floodwaters flowed behind this low wall and were let into the fields below at certain points by the means of simple spillways. At the end of the main canal is a small secondary stretch, which stops the remaining water from entering back into the wadi, redirecting it back into the irrigated area. There are two further canals along the course of the Wadi Saad; these are short in length (250 m and 800 m respectively) and seem to have been constructed to capture water to irrigate small, stone-cleared fields immediately adjacent to this wadi. Qasr Burqu’ The site of Qasr Burqu’ lies some 100 km east of ad-Diyatheh and some 70 km southeast of al-Namara. At this location stone structures have been built in a natural shallow basin where water from surface and subsurface run-off from the surrounding slopes naturally ponds (Betts et al., 1990:6; Fig. 5.7). The most significant structure is a tall, rectangular, stone tower rising to a height of c.5 m (Fig. 5.8); the study of its arches and floor levels suggested that it may have reached a height of 13 m (Betts et al., 1991:16). The tower is surounded by a series of rooms forming a rough square, which
Differing strategies for water supply and farming
95
Figure 5.7 The ancient reservoir at Qasr Burqu’ Source: After Kennedy and Riley, 1990
Figure 5.8 Air photograph of Qasr Burqu’ and its reservoir, looking southeast Photograph: B.Bewley
The archaeology of drylands
96
appears to have been constructed at a later date. An analysis of the architectural details of the tower and surrounding structures, along with the evidence from two burial inscriptions written in Greek, led Svend Helms to suggest the site to be a monastic foundation dating to the Late Antique and Early Islamic periods (Betts et al., 1991:16). Potsherds found within the structure for the most part date to the Late Antique and Early Islamic periods, and initially it was thought that the tower was probably constructed at this time as the residence for a recluse (Gaube, 1974). However, earlier Roman pottery has also been found in the area (Betts and Helms 1989:8; Betts et al., 1991:22), and there is a strong possibility that the Roman army established themselves at the site earlier, including perhaps constructing the tower (Kennedy and Riley, 1990). Later, this tower could have served as the focus for a monastery that grew up around it, and changed its function to that of a watch tower and possible refuge from hostile nomads. Alternatively, the tower and enclosure buildings may have had a military purpose for the Arab government in the sixth-century AD Ghassanid period (Betts et al., 1991:17). Whatever the function of these buildings, it was vital for the occupants to secure an adequate all-year-round supply of water. Natural ponding of flood-and rainwater occurs at this point due to a bed of more resistant basalt rock crossing the wadi valley. Presumably the quantity of water collected behind this natural barrier was not sufficient to sustain an adequate population throughout the year, so the volume of ponded water was enlarged and secured by the building of a dam downstream of the buildings, on top of the low ridge that was responsible for the natural pooling of floodwater within the wadi bed (Fig. 5.7). This dam, the lower courses of which still survive, was composed of large-sized, roughly dressed, basalt boulders, laid in a series of stretcher courses, capped by one of headers. Two such walls were built about 10 m apart, following the top of the ridge, the space between probably being filled with earth and rubble. Along the side facing the water is evidence for plastering, which, together with the form of wall construction, indicates that the dam is probably of a similar date to the tower and its surrounding enclosure (Betts et al., 1991:12). The edges of the reservoir are lined with low, roughly-coursed, stone walls, from which two short stone staircases lead down to the reservoir, one on each side of it. No evidence has been found for a sluice gate of any type within the wall of the dam, which suggests that the water was not used to irrigate a network of fields in the manner of the impressive dam and field network further to the north at Qasr el-Gherbi near Palmyra (Schlumberger, 1986). However, one of the ‘rooms’ of the enclosure (room 11) has been tentatively identified as a windmill (Betts et al., 1991:17), an interpretation that, if correct, clearly implies that cereals were being grown in the vicinity of the settlement. Around the dammed lake and its associated buildings are huge numbers of encampments, burial cairns and corral remains, along with scatters of artefacts from all periods, in a manner similar to that at al-Namara. Given the location of the settlement and its very large facility for water storage, the likelihood is that the main function of Qasr Burqu’ was as an oasis settlement serving a series of north—south and east—west desert routes (Betts et al., 1990:6). No evidence for field systems has been found by the recent survey. However, in the nearby more fertile and stone-free Ruhba there are ‘fields of barley, planted for animal fodder and watered by flood irrigation’ (Lancaster, 1981; Helms, 1989) beside the fixed
Differing strategies for water supply and farming
97
bedouin encampments of Feytha, arRisha al-Fawq and ar-Risha al-Taht. It is likely that any permanent population living at Qasr Burqu’, whether Roman or Arab, was served by equivalent field systems to those situated in the Ruhba.
DISCUSSION On first consideration, all three sites exhibit an attempt to impose Roman control on the peripheral regions adjacent to the settled provinces of Arabia and Syria. The exercising of control would have been specifically aimed at the local populations of pastoralists, the traffic of caravans and the activities of occasional groups of bandits (Isaac, 1990). This took the form of strategically-placed military structures at all three sites, where water could be collected and which would have attracted the travellers and pastoralists of the region, whose movements could thus be more easily monitored. At present it is almost impossible to say at which sites settlement of local people preceded or post-dated the building of the Roman structures. What is certain is that at all three sites substantial settlements did occur. At ad-Diyatheh, this took the form of a village with substantial permanent stone buildings. At al-Namara and Qasr Burqu’, the settlements mostly consist of stone-cleared corrals for temporary structures such as tents, but this does not necessarily imply that the settlements were not of a long-term nature: they could have been permanent on a year-to-year basis, with families periodically changing the location of their campsites, or they could have been occupied for substantial periods of the year by different family groups of nomadic pastoralists. All the sites exhibit measures for the control and use of limited supplies of water, for both drinking and agricultural purposes, but the methods utilized were on different levels and scales. At ad-Diyatheh, varied and sophisticated techniques were employed for floodwater farming. The systems for floodwater farming at al-Namara are simpler and smaller. At Qasr Burqu’, even though floodwater farming systems have not yet been identified, cereal cultivation is implied by the probable presence of a mill, and the site certainly displays an impressive scale of planning for the storage of much larger quantities of water than at the other sites. These substantial differences in floodwater farming techniques and strategies for water storage imply that the inhabitants of these sites, whether Roman, indigenous bedouin, or both, were dealing with different issues in building the settlements and their associated structures. Local environmental considerations were undoubtedly one important factor affecting strategies for water management. At ad-Diyatheh, for example, there would have been a greater quantity of water moving at a higher speed of flow than at the other two sites, requiring substantial and carefully designed structures to allow some of this water to be brought under control. The topography at ad-Diyatheh, at the juncture of the Jebel slope and the gently descending alluvial fan below it, allowed also for the construction of a large extensive network of irrigation channels and a large dispersed spread of fields. The speed of waterflow would have decreased at al-Namara, 60 km downstream of ad-Diyatheh, and diverting the waterflow would have been especially impeded by the fact that at this point the wadi bed lies up to 15 m below much of the surrounding plateau, so irrigation is really effective only along areas of adjacent
The archaeology of drylands
98
floodplain. On the other hand, Qasr Burqu’ lies in a natural dip within the surrounding plateau, with the majority of water collecting here through groundwater seepage (Lancaster and Lancaster, 1999:132). A second area of differences, though one harder to identify with the archaeological evidence available, relates to the social and economic conditions of the inhabitants of the three settlements. At ad-Diyatheh, the field system can be directly related to the substantial permanent village at the site, whose inhabitants relied to a great extent upon the cereals they grew, judging by the number of mills. In addition, construction of the dwellings echoes in form structures found on the Jebel and other regions of Syria, in the provision of accommodation for livestock below the family rooms (Villeneuve, 1986). The reliance on crops for subsistence would have been much less for the pastoral populations postulated for al-Namara and Qasr Burqu’, and so the need for large farming systems proportionally less, though the surrounding Harra could have supported only low-density populations. At both al-Namara and Qasr Burqu’, it may well have been only the residents of the substantial permanent structures, be they soldiers or monks, who attempted to grow crops on any scale. The pastoral populations at these settlements may have grown cereals on any scale only when favourable climatic conditions allowed, so would not have been inclined to spend valuable time in investing in the substantial infrastructure required beforehand, using the locations simply as convenient watering places (Macdonald, 1993). Caution is needed in advancing such an hypothesis, however, for the modern bedouin of the region in fact try to irrigate areas when settled in a particular location over a period of time (Lancaster and Lancaster, 1999:132–66). It is probably reasonable to assume, though, that the occupants of the substantial structures were generally more reliant on provisions provided from the immediate area, given that pastoralists would have had the capacity to move to new areas when they had exhausted supplies at any particular location. It may simply be the case that people with different traditions and cultures were responsible at various periods of time for constructing the different systems of water control and floodwater farming at all three sites. The people at ad-Diyatheh may have come from more settled areas on the Jebel, though the poor construction of the houses at ad-Diyatheh compared with the more refined architecture of the adjacent Jebel villages suggests that ad-Diyatheh was built by ‘sedentarized nomads’ (Villeneuve, 1986:710). The small field system at al-Namara could have been constructed either by a small military garrison or by semi-sedentary pastoralists. The water storage system at Qasr Burqu’ may well be a later enhancement of a much older catchment system. Whatever the origins of these systems, though, the water management techniques employed in the Roman period at both ad-Diyatheh and al-Namara reveal an agricultural culture not extensively practised in the region since the Bronze Age, a fact strongly suggesting that these new approaches to agriculture came as a direct consequence of political and social changes brought about by the imposition of Roman hegemony. The extent to which these new agricultural regimes were imposed, created and operated by the Roman army, or reflect a variety of complex responses by indigenous populations to the opportunities of Romanization, still remains unclear. However, as the preceding discussion has shown, the likelihood is that the archaeology of the Syrian Black Desert
Differing strategies for water supply and farming
99
reflects both the external imposition of, and indigenous adaptions to, the forces of Roman imperialism.
ACKNOWLEDGEMENT I would like to thank Bob Bewley for permission to reproduce, as Figure 5.8, his air photograph of Qasr Burqu’ and its reservoir from his Aerial Archaeology in Jordan Project.
REFERENCES Barker, G., Gilbertson, D., Jones, B. and Mattingly, D. (1996) Farming the Desert. The UNESCO Libyan Valleys Archaeological Survey. Volume One: Synthesis. Paris, UNESCO; London, Society for Libyan Studies; Tripoli, Department of Antiquities. Betts, A. (1993) The Burqu’/Ruweishid Project: preliminary report on the 1991 field Season. Levant 25:1–11. Betts, A.V.G. and Helms, S.W. (1989) A water harvesting and storage system at Ibn alGhazzi in eastern Jordan. Levant 21:3–11. Betts, A., Helms, S., Lancaster, W., Jones, E., Lupton, L., Martin, L. and Matsaert, F. (1990) The Burqu’/Ruweishid Project: preliminary report on the 1988 field season. Levant 22:1–20. Betts, A., Helms, S., Lancaster, W. and Lancaster, F. (1991) The Burqu’/Ruweishid Project: preliminary report on the 1989 field Season. Levant 23:7–28. Braemer, F., Echallier, J.-C. and Taraqji, A. (1996a) Khirbet el Umbashi (Syrie): Rapport préliminaire sur les campagnes 1993 et 1994. Syria 73:117–29. Braemer, F., Echallier, J.-C., Hatoum, H. and Macdonald, M.C.A. (1996b) Archaeological and Epigraphic Rescue Survey at Al-Nemara. Report on the First Season Sept-Oct 1996. Damascus, Department of Antiquities and Museums of Syria, unpublished report. Evenari, M., Shanan, L. and Tadmor, N. (1971) The Negev: The Challenge of a Desert. Cambridge, Mass., Harvard University Press. Evenari, M., Shanan, L. and Tadmor, N. (1982) The Negev: The Challenge of a Desert. Cambridge, Mass., Harvard University Press, second edition. Gaube, H. (1974) An examination of the ruins of Qasr Burqu’. Annual of the Department of Antiquities of Jordan 19:93–100. Helms, S.W. (1981) Jawa: Lost city of the Black Desert. London, Methuen. Helms, S.W. (1989) Jawa at the beginning of the Middle Bronze Age. Levant 21: 141–68. Isaac, B. (1990) The Limits of Empire: The Roman Army in the East. Oxford, Oxford University Press. Kennedy, D. and Riley, D. (1990) Rome’s Desert Frontier From the Air. London, Batsford. Lancaster, W. (1981) The Ruwala Bedouin Today. Cambridge, Cambridge University Press.
The archaeology of drylands
100
Lancaster, W. and Lancaster, F. (1999) People, Land and Water in the Arab Middle East: Environments and Landscapes in the Bilâd ash-Shâm. Studies in Environmental Anthropology, Volume 2. Amsterdam, Harwood Academic Publishers. Macdonald, M.C.A. (1993) Nomads and the Hawran in the Late Hellenistic and Roman periods: a reassessment of the epigraphic evidence. Syria 70:303–413. Sadler, S. (1993) Le terroir agricole de Diyateh: l’irrigation comme condition d’existance de ce terroir. In B.Geyer (ed.) Techniques et Pratiques Hydro-Agricoles Traditionnelles en Domaine Irrigué, Approche Pluridisciplinaire des Modes de Culture avant la Motorisation en Syrie (Actes du Colloque de Damas, 27 juin-1 juillet 1987): 421–51. Paris, P.Geuthner, Bibliothèque Archéologique et Historique 136. Schlumberger, D. (1986) Qasr el-Heir el-Gharbi. Paris, P.Geuthner, Bibliothèque Archéologique et Historique 120. Tate, G. (1992) Les Campagnes de la Syrie du Nord di IIe au VIIe Siècle: Un Exemple d’Expansion Demographique et Economique a la Fin de l’Antiquité. Paris, P.Geuthner, Bibliothèque Archéologique et Historique 133. Tate, G. (1997) The Syrian countryside during the Roman era. In S.E.Alcock (ed.) The Early Roman Empire in the East 55–70. Oxford, Oxbow Books. Villeneuve, F. (1986) Ad-Diyatheh: village et castellum romains et byzantins a l’est du Jebel Druze (Syrie). In P.W.M.Freeman and D.L.Kennedy (eds) Defence of the Roman and Byzantine East. Volume II: 697–715. Oxford, British Archaeological Reports, International Series 297.
6 Irrigation agriculture in Central Asia: a longterm perspective from Turkmenistan MARK NESBITT AND SARAH O’HARA
INTRODUCTION Agriculture in the newly independent republics of the former Soviet Central Asia is almost entirely dependent on irrigation. Consequently, access to water is essential and it has long played an important role in the social, environmental, economic and political situation of the region. Today, as in the past, agriculture represents the single most important economic activity throughout the region, and currently over 40 per cent of the population is employed in the commercial agricultural sector, with the vast majority of Central Asians either partially or wholly dependent on subsistence agriculture. The agricultural sector throughout Central Asia, however, is under threat because of the rapid deterioration in the water distribution and irrigation since the collapse of the Soviet Union (O’Hara, in press). Central Asia boasts a long history of irrigated agriculture, but the exploitation of the region’s water resources and the expansion of the irrigation network peaked during the latter part of the Soviet era. During this period huge water diversion and irrigation projects were constructed to satisfy Moscow’s continual demands for cotton. In order to maximize agricultural output, water was taken from areas of surplus to those of deficit, often involving transfers over considerable distances and in some case from other republics. Today, however, this huge, highly integrated, network serves five independent states, each following its own agenda for reform. The implications for the region’s water resources are immense and it is becoming increasingly difficult to reach a consensus on how the water distribution and irrigation system should be managed and maintained (Bedford, 1996; O’Hara, in press). Further complicating the matter is the fact that Central Asia’s irrigation zones are plagued by secondary salinization and high water tables (O’Hara, 1997; Smith, 1992), and it is evident that these large-scale Soviet-built systems are environmentally unsustainable. The situation is not likely to improve and indeed could be exacerbated by changing land and agricultural policies, coupled with an increased demand for water as population rises. Should the system fail, the consequences would be enormous and could ultimately undermine regional security. The question of sustainable irrigation is therefore urgent. Given that Central Asia not only has a long history of irrigated agriculture but has witnessed the rise and fall of a number of major empires over the last few thousand years, it may well be that lessons can be learned from the past. An assessment of former irrigation and water management practices may highlight whether sustainable irrigation is a feasible option, and if so how it might be
The archaeology of drylands
102
achieved. Here we review the history of settlement, agriculture and irrigation over some 8,000 years in southern Turkmenistan (Fig. 6.1). A large body of archaeological evidence is available for this region, much of it resulting from the establishment of the South Turkmenistan Multi-Disciplinary Archaeological Expedition (YuTAKE) in 1946. Many of its publications were not widely distributed, even within the former Soviet Union, but we have been able to draw on a wide range of useful syntheses published in western journals. A more recent phase of fieldwork involving a number of international research teams has resulted in a series of renewed excavations at several important sites including Jeitun, Anau, Gonur Depe and Merv. Although many of these projects are ongoing, important papers pertaining to the area have emerged (Harris et al., 1993, 1996; Herrmann, 1997; Herrmann et al., 1998, Hiebert, 1994), providing valuable information on changes in environment and society over this period. Historical sources are more problematic. Although literate civilizations have existed in the region since the Achaemenid period, there is no systematic body of texts comparable to the clay tablets of Mesopotamia. For the medieval period, we are largely dependent on short descriptions in accounts by Arab or Chinese travellers or Arab historians. Some Sasanian records have survived through their use by the Arab historians. Prior to this, we are again dependent on brief travellers’ accounts and histories compiled far away to the west, in classical Greece and Rome. Our understanding of the political dynamics underlying the increasingly welldocumented settlement archaeology is therefore currently less sophisticated than in the Near East proper.
ENVIRONMENT Turkmenistan covers an area of 480,000 km2, 90 per cent of which is covered by the virtually uninhabited Kara Kum Desert (Babaev, 1996). Most of Turkmenistan comprises lowlands, with mountains being confined to the southern and western parts of the country. It lies within the temperate desert zone (Babaev, 1994) and has a marked continental climate (Orlovsky, 1994). Precipitation mainly falls as snow or rain in winter, with almost none in the agriculturally-active summer months of June through to September. Average annual precipitation varies from 90 mm in Dashouz to nearly 400 mm in the southwest highlands of the Kopet Dagh, but in much of the country it
Irrigation agriculture in Central Asia
103
Figure 6.1 Turkmenistan, showing locations mentioned in Chapter 6 is less than 200 mm per annum and therefore insufficient for dryland agriculture. Average temperatures are high, varying from 12 to 18°C. The coldest months are December to February, with temperatures frequently falling below 0°C, and the hottest months June to August, when temperatures often exceed 45°C. Potential evaporation rates vary accordingly, from 1–2 mm per day in winter to 10–15 mm per day in summer. Total annual potential evaporation rates are of the order of 2,500–3,000 mm, which are far higher than the precipitation rates. The hydrological network is weakly developed and all major sources of water rise outside the country’s borders (Fig. 6.1). The headwaters of the Amu Darya, the largest river in Central Asia, are in the Pamirs, and the river flows through a number of countries before discharging into the Aral Sea. It displays two periods of peak discharge, one during the spring, associated with snow melt, the other later in the summer when ice melt increases its flow. The other main rivers all rise in the mountains to the south, the Atrek flowing into the Caspian Sea and the Murgab and Tejen draining into the Kara Kum Desert. Although small when compared with the Amu Darya, they are an important source of water and have long been used by people occupying the region. Fed by winter rains and snowmelt, they have only one period of peak discharge during the spring. In addition to these rivers, there is a number of smaller intermittent rivers and springs, most of which cease to flow during the summer. Today, as in the past, human settlement in Turkmenistan is concentrated in two zones: the piedmont at the foot of the steep slopes of the Kopet Dagh mountains, and the desert oases.
The archaeology of drylands
104
THE NEOLITHIC: EARLY FARMING IN THE PIEDMONT The beginning of agriculture in Turkmenistan is best documented by the important excavations at Jeitun (Djeitun) located some 25 km north of Ashgabat in the piedmont between the Kopet Dagh mountains and the Kara Kum Desert. First discovered in the early 1950s, Jeitun has been the subject of a number of detailed excavations that have produced one of the bestknown archaeological sequences in Central Asia. Today, fourteen such Jeitun culture sites have been identified across southwestern Turkmenistan. Jeitun comprises some thirty excavated small, rectangular, mudbrick houses located on the distal reach of the Kara Su (Black Water), a small ephemeral stream that rises in the Kopet Dagh and discharges into the Kara Kum Desert. The settlement covers less than 0.7 ha, and estimates of the cultivated land surrounding it in the neolithic period are between 15 and 33 ha. First excavated by Masson in the 1950s and 1960s, the site was dated to c.6000 BC on the basis of ceramic assemblages (Masson and Sarianidi, 1972). This date was later confirmed by a series of eleven radiocarbon dates from the BritishRussian—Turkmen excavations of 1989–94, which indicates that the site was occupied between c.6300 and 5600 cal BC (Harris et al., 1993, 1996). Recovery of animal bones and charred plant remains from these new excavations has allowed a reassessment of the site’s subsistence base. The results confirm earlier evidence for a primarily agricultural system of subsistence based on cereals and domestic sheep and goat, augmented by hunting, primarily of gazelle. The cereals are dominated by einkorn wheat (Triticum monococcum), with small amounts of emmer (T.dicoccum) and naked and hulled forms of barley (Hordeum sativum). Other artefacts from the site point to the importance of cereal cultivation for the inhabitants of Jeitun, with sickle blades accounting for 37 per cent of all tools found in Masson’s early investigation of the site (Masson and Sarianidi, 1972). In addition to cereal cultivation, the inhabitants of Jeitun herded goats and sheep; faunal analysis shows that, although raised primarily for meat, these animals could also have been an important source of milk, wool, hair and skins. The dominance of domesticated plants and animals from the very bottom of the Jeitun sequence, together with the absence of wild progenitors of wheat and sheep in Central Asia, supports the view that agriculture and its attendant domesticated species did not evolve independently in the region, but rather reached it from the Fertile Crescent of southwest Asia, via the Zagros mountains of Iran (Harris and Gosden, 1996). Jeitun is often cited as one of the oldest known sites of irrigation in the world (Dukhovny, 1995; Harris et al., 1993; Lisitsina, 1984). There is, however, some difference in opinion as to how crops were irrigated at this time. Lisitsina (1981), for example, assumed that cultivation at Jeitun was entirely dependent on run-off from the Kopet Dag, with Lewis (1966) suggesting that Jeitun’s location on the distal reaches of the Kara Su was due to the fact that neolithic farmers were better able to control and manipulate flows in this part of the river system. However, Kohl (1981) argued that Jeitun was in fact located on the distal reaches of the Tejen Delta, which at this time discharged further into the Kara Kum Desert than today. The presence of many seeds of the weeds club-rush (Scirpus maritimus) and goat-face grass (Aegilops tauschii), in
Irrigation agriculture in Central Asia
105
association with the charred cereal remains, led Harris et al. (1993) to conclude that cereals were being grown in areas with high water table and high salinity (‘takyrs’), rather than on stream sides irrigated by less saline floodwaters. Takyrs are highly impermeable, almost flat, clay surfaces that retain water and are of considerable importance to communities living in the desert today. All these different scenarios would largely draw on naturally irrigated land, with only relatively small-scale channels or embanking required. No definite evidence of such features has been discovered, though this may in part reflect subsequent processes of deposition or erosion. The belief that early agriculturists at Jeitun irrigated their fields, however, is based largely on the assumption that the climate during the Neolithic was similar to today. There is some evidence to suggest that this assumption may be unfounded, for the base of the dunes overlying fluvial deposits at Jeitun has yielded an OSL date of c.4,500–5,000 years BP, so it is possible that arid conditions similar to today developed somewhat later than previously thought (just as wetter environments characterized the early Holocene in the Levant: Chapter 4). Further support for this hypothesis is provided by a recent analysis of plant remains from the site, suggesting that cultivation may have been possible without irrigation (M.Charles, pers. comm.). Despite this uncertainty, though, it is evident that there is a long history of agriculture in this region and that by the fifth millennium BC agricultural settlements were spread along the piedmont from Kyzal Arvat in the west to Tejen in the east.
ENEOLITHIC TO IRON AGE: PIEDMONT SETTLEMENT AND EXPANSION TO THE OASES The establishment of agricultural communities such as Jeitun in the Neolithic was followed by several millennia of continuing settlement, largely in the piedmont of the Kopet Dagh, during the Eneolithic or Chalcolithic (c.4800–3000 BC), Early Bronze Age (c.3000–2500 BC) and Middle Bronze Age (c.2500–1900 BC). Sites grew significantly in size—eneolithic settlements such as Altyn Depe, Anau and Namazga cover up to 25 ha— and there were two key changes in settlement pattern: the expansion into the Geoksyur oasis in the Eneolithic and the emergence of state-level urbanism in the piedmont zone in the terminal Early Bronze Age. The Geoksyur oasis (Fig. 6.1) is situated on the Tejen River delta and, unlike other oases in the region, is contiguous with the foothills of the Kopet Dagh. Nine prehistoric sites comprising large, widely scattered, mudbrick houses, have been found in the oasis, dating from the earliest Eneolithic (Kohl, 1984), but the oasis appears to have been abandoned by the end of the Eneolithic. Earlier eneolithic settlements are at the end of branches of the river delta, suggesting that a modified form of floodwater irrigation was practised. Later in the Eneolithic, well-developed artificial irrigation systems are documented for the first time in Turkmenistan (Namazga III period, c.3500–3000 BC). Aerial photographs and excavations have shown that land around the site of Geoksur I was irrigated by three parallel canals, each up to 3 km long and 5 m wide, possibly irrigating an area of about 50 ha by means of small aryks (irrigation canals) branching off and leading to fields (Lisitsina, 1969). The water flow into secondary canals was
The archaeology of drylands
106
controlled by inlet structures where they joined the main canals (Lisitsina, 1981). In the piedmont proper, the last part of the Early Bronze Age witnessed a transformation of settlements with the appearance of specialized production areas, fortification walls around settlements, increased status differentiation in burials, and evidence of much interaction between settlements throughout the Kopet Dag region, all consistent with a state-level society (Hiebert, 1994). These trends continued into the Middle Bronze Age, and by its terminal phase (2200–1900 BC) the foothills contained a number of very large sites such as Namazga (50 ha) and Altyn Depe (25 ha). This period of expansion came to an end in the Late Bronze Age. The settlements at Anau and Namazga, for example, were considerably smaller, now covering only a few hectares. Relatively little is known about agriculture in the piedmont zone in the Eneolithic and Bronze Age. Irrigation canals have not been located in the piedmont, but this may reflect deposition and erosion in this geomorphologically-active zone. The presence of bread wheat and six-row hulled barley in lencolithic samples from Anau dated to c.4500–3000 BC has been cited as possible evidence for irrigation (Miller, 1999), but both cereals were grown in many regions of the Old World without irrigation (Maier, 1996). Paralleling the decline of settlement on the northern piedmont was the spread of irrigation to the lower reaches of the Murgab river at the end of the Middle Bronze Age, although this occurred while some sites such as Altyn Depe were still very large. A number of factors has been cited for this shift in agricultural settlement. Masson (1957), for example, suggested that a rise in population stretched resources to the extent that people were forced to migrate, whilst some authors have highlighted the potential impacts of climate change. Lewis (1966) argued that there is no evidence of a major shift in climate during this period, but, as mentioned above, evidence is emerging for a shift to drier conditions c.5,000–4,500 years ago, coinciding with the rise of agriculture in the Murgab oasis. It is possible, therefore, that conditions became sufficiently dry to precipitate change. The bronze age settlements of the Merv oasis covered an area of 100 km north—south by 50 km east—west, which is almost five times larger than the later medieval and classical oasis to the south. Hiebert’s recent re-analysis of the ceramic chronology and survey data suggests that the colonization of the oasis was rapid (Hiebert, 1994). The sites cluster in ‘micro-oases’, forming linear patterns that presumably followed old river branches (Fig. 6.2). Settlements are characterized by large fortified building complexes with intervening fields, which, as Hiebert points out, typify Central Asian oasis architecture of the time. Initially, settlements were located on the northern margins of the oasis, with the system expanding southwards some 400 years later (Hiebert’s Gonur Period 3). Initial settlement was at the northern fringe of the oasis because large-scale canals were not used. Instead, fields were irrigated by ditches carrying water from the smaller streams into which the Murgab river split near the edge of the delta. As Bader et al. (1996) comments, settlers from the Kopet Dagh would already have been familiar with the technology of using streams of the piedmont. Archaeobotanical analysis indicates that, over time, greater numbers of plants and animals were domesticated. By the Bronze Age, the variety of crops grown had increased significantly compared with in the Neolithic: samples from the middle bronze age site of Gonur Depe in the Merv oasis, for instance, are dominated by hulled and naked barley,
Irrigation agriculture in Central Asia
107
with free-threshing wheat, lentils, peas, chickpeas and grape also present (Miller, 1993;
Figure 6.2 Bronze and iron age settlement in the Merv oasis (adapted from Hiebert, 1994)
The archaeology of drylands
108
Moore et al., 1994). These finds are consistent with those from the neighbouring Geoksyur oasis (Lisitsina, 1969; Lisitsina and Prishchepenko, 1976), and are typical of bronze age settlements throughout the Near East. Late bronze age samples from Tahirbaj Tepe in the Merv oasis were also dominated by hulled barley but add broomcorn millet (Panicum miliaceum) to the repertoire of crops (Nesbitt, 1994). Iron age settlements in the Kopet Dagh foothills are widely distributed and often continue on the same sites as bronze age settlements, but are smaller and marked by less material complexity (Kohl, 1984). In the Merv oasis, iron age sites are concentrated in four ‘micro-oases’. The northernmost two, Takhirbai and Togolok, contain bronze age settlements, while the southernmost two, Yaz depe and Aravali, represent new occupation, thus forming part of the pattern of southward movement of settlements that continues until the Achaemenid period (Bader et al., 1996; see below). This shift in settlements is most plausibly explained by increased extraction of water upstream by settlements using more sophisticated canal systems, collecting water near the head of the delta. However, early sites in the upper part of the oasis may have been masked by alluvial deposition, accounting in part for this pattern.
ACHAEMENID TO MEDIEVAL: URBAN SOCIETIES The Achaemenid period (530–330 BC) marked two important transitions for the Merv oasis: it was the first of several periods when Merv came under the control of an empire based to the south; and for the first time a series of urban centres emerged in the oasis. From this time onwards, Merv was also militarily important as a frontier city at the northeastern part of firstly the Achaemenid and later the Seleucid (330–140 BC), Parthian (140 BC–AD 220) and Sasanian (AD 220–651) empires. Surveys of the magnificent ruins of Merv’s urban centre show a steady increase in its size. The earliest city, Erk Kala, had walls enclosing an area of 20 ha. It later became the citadel of the adjoining Seleucid city of Gyaur Kala (400 ha) (Fig. 6.2), which continued to be occupied for a period of over a millennium, even after the construction of the nearby city of Sultan Kala in the eighth century AD. Survey work in rural areas in the north of the oasis confirms this basic pattern of expansion, with increasing residential areas from Achaemenid to Seljuk times (Bader et al, 1993/94; Gubaev et al., 1998). At its greatest extent, the oasis covered c.700 km2. The area cultivated appears to have fluctuated, with a decrease in rural settlement in the Hellenistic period, and a marked increase in cultivation and, probably, the first construction of a large central dam and canal network in the Parthian period. Although written sources state that Merv was destroyed by Mongol invasions in AD 1221–2, there is archaeological evidence for a substantial post-Seljuk occupation, and in the early fifteenth century a new, much smaller, city was built by the Timurids. Notwithstanding this, the oasis declined in importance, and the Timurid city was abandoned by the nineteenth century. Overall, therefore, changes in settlement pattern suggest three key phases in the occupation of the Merv oasis: the initial colonization by dispersed but numerous bronze age settlements c.2200 BC; urban development in the Achaemenid period c.600 BC; and the gradual abandonment of intensive settlement in
Irrigation agriculture in Central Asia
109
most of the oasis in the centuries after the Mongol invasions of AD 1221–2. The large-scale sampling of contexts carried out at the city of Merv by the International Merv Project (Boardman, 1997; Nesbitt, 1994) has provided some the best archaeobotanical evidence for the Sasanian period. It indicates that, during the Late Sasanian period (the fifth to seventh centuries AD), cereals consisted, as before, of abundant hulled barley and free-threshing wheat and rarer broomcorn millet, lentils (chickpea seems to disappear after the Bronze Age), very abundant cotton seed, and a wide range of fruits and vegetables including cucumber/melon, grape, almond, peach and nuts. Two changes are apparent in comparison with the Bronze Age: first, an increase in crop diversity, particularly in the fruits; and second, and more importantly, the addition of cotton, which is a source of both textile and oil and, like millet, a crop that expands the growing season through the summer, after the wheat and barley harvest. Overall, the range of crops seems similar to that mentioned in Islamic times: in the tenth century Merv’s famously soft cotton textiles were exported as far as Africa and Spain, and there are thirteenth-century references to Merv’s fine grapes and other fruits (Serjeant, 1972). Of the range of crops grown in the Sasanian period, barley and cotton are relatively tolerant of soil salinity (though not, of course, of heavily salinized soils), bread wheat is moderately tolerant, melon and grape are moderately sensitive, and almond and peach are sensitive (Maas, 1987). That crops sensitive to salinity are present throughout the late Sasanian sequence from Merv, and that the full suite of crops is present in broadly similar quantities throughout the sequence, strongly suggest that irrigation agriculture was sustained through this period without occurrence of catastrophic salinization. This is consistent with evidence for unbroken intensive settlement in the oasis from the Parthian to Seljuk periods. Although salinization has often been viewed as an inherent and imminent threat in ancient irrigation systems in the Near East, particularly in Mesopotamia (Jacobsen and Adams, 1958), there is increasing evidence that soil management practices that avert salinization (and which are ethnographically documented in Mesopotamia) were applied effectively in the past (Powell, 1985). The success of Merv and other settlements in the region depended to a large extent on how water resources were managed. There were two major technological innovations during the urban period. In the foothill zone the qanat was introduced in the first millennium BC. Like the foggaras of the Sahara (Chapter 9), this system allows groundwater to be tapped by underground tunnels cut into the foothills, and is most widely used in the highland area of Iran. Qanats are difficult to date directly, but associations with sites suggest that they became widespread in the highlands of Iran and neighbouring areas at this time. In the Merv oasis, there is indirect evidence from survey work of large, state-sponsored, irrigation works in the Parthian and Sasanian periods (Bader et al., 1993/94, 1996; Gubaev et al., 1998), like the contemporary transformations occurring further south in Susiana (Wenke, 1975–76) and Mesopotamia (Adams, 1981). Major changes in irrigation technology in the Merv oasis are, therefore, later (if the dating is correct) than the first urbanization at Erk Kala and Gyaur Kala but coincide with the increase in the population of the oasis in the Parthian and Sasanian periods. In contrast with the bronze age canals, these later irrigation systems are difficult to investigate because they have been largely destroyed by twentieth-century agriculture, but the medieval irrigation system of the tenthtwelfth centuries—during which time the
The archaeology of drylands
110
oasis nourished—may give a good parallel. The Arab historians and geographers such as Muqaddasi, Al-Biruni and Yakut provide valuable accounts of water distribution and irrigation systems (see Bartold, 1914, 1928 and Le Strange, 1905 for translations and discussions of their works). It is evident from these that the administration of scarce water resources was central to the way in which the social and political hierarchy of settlements operated: water was viewed as a ‘gift from God’ that could not be owned or controlled by an individual. The city of Merv had access to only one source of water—the Murgab river, which rises in the Afghan mountains and drains northward into the Kara Kum Desert. The river’s annual discharge is about 1.2 km3, which is approximately 5 per cent of the total amount of water available for use in Turkmenistan at present (O’Hara, 1997). The oasis was renowned for its productivity, not only producing enough food to feed its large population but exporting produce to adjacent areas (Herrmann and Petersen, 1997). The region’s agricultural success was in part due to the land and water management strategies of the time. Land, for example, was divided into small plots that were intensively cultivated, receiving water on a regular basis. The amount of land cultivated in any given year depended on water availability. Muqaddasi, writing in the tenth century AD, described how a depth gauge situated at the Razik Dam to the south of the city was used to determine whether there would be a surplus or deficit of water that year. If the level reached the 60th point, water would be plentiful that year and the order would be given to increase the amount of land cultivated, whereas in years of low water availability the area was reduced and only the best lands were cultivated. The dam was extremely important and was, in effect, the only water storage facility for the city. Its maintenance was assured by 400 divers employed around the clock, each diver having to deliver a specified amount of wood and mud to the dam each day (Bartold, 1914). Yakut, who resided in the city at its zenith in the early thirteenth century, provides further details. He described how water gauges were installed at the head of every canal throughout the city. The whole system was headed by the mirab bashi (chief water master), and hourly reports on the water level in the main canal were passed to his office, so that he could decide which off-takes were to be opened or closed. The system was managed by elected senior officials and maintained by over 12,000 workers, paid by the water users, who were also expected to take part in major construction schemes and in the annual maintenance programme.
EXPANDING THE IRRIGATION SYSTEM: THE TSARIST AND SOVIET PERIODS When Central Asia finally came under Tsarist control in the late 1880s, the new administration attempted to introduce reforms in the irrigation sector. These failed, however, and the authorities declared that irrigation would be run ‘by custom’. Notwithstanding this, a number of subtle changes was made: most important, irrigation officials became part of the Tsarist civil service and as such were no longer controlled by water users. This act severed the link between water users and providers, so effectively undermining the traditional system of water management. State salaries for officials were
Irrigation agriculture in Central Asia
111
low and there was no longer any incentive to control the system. The situation was exacerbated by the imposition of irrigation officials unaccustomed to the traditional method of management, resulting in increased problems within the system, which became subject to corruption and abuse by the wealthy and more powerful water users. More significant than Tsarist interventions in water management, however, were the changes in agricultural policies. The authorities in Moscow, keen to end their reliance on America for cotton (particularly following the American Civil War when supplies almost ceased), recognized that Central Asia had the potential to become a major cotton growing region; in fact, the main factor behind initiatives to increase the amount of land irrigated was cotton production (Lipovsky, 1995). The subtle, but nonetheless important, changes in water management, coupled with increased demand and use of water, appear to have caused widespread land degradation. In the Merv oasis, for example, the irrigation network was expanded by some 33,000 ha (Zaharchenko, 1990), but poor management of the system caused local water tables to rise, resulting in salinization and widespread surface ponding that not only degraded the soils but also led to outbreaks of malaria (Pierce, 1960). The Bolshevik Revolution and the subsequent emergence of the Soviet Union heralded a period of radical change in the way water was managed in Central Asia. In 1923 the Soviet administration decreed that water management was to be taken ‘out of the hands of traditional elders and councils with whom it resided’ (Black et al., 1991) and, like land, was to become a common resource for the benefit of all. Various government bodies were established to be responsible for the development of a regional water management strategy that would allow centrally-determined production targets to be met. With cotton production the priority for Moscow, huge sums of money were invested in the region in the development of massive, highly integrated systems of water distribution and irrigation (Micklin, 1991). Land was irrigated no longer by a single local source, as in the past, but by water often piped over considerable distances: the Kara Kum Canal, for example, considered to be one of the engineering feats of the Soviet era, now transfers in excess of 12.9 km3 of water from the Amu Darya along its 1,400 km length every year, irrigating an area of c.1 million ha (Hannan and O’Hara, 1998). There has been much criticism of the management and maintenance of Soviet irrigation systems and the inefficiency of water use (e.g. Micklin, 1991). Losses occurred throughout the system, with problems of seepage and evaporation from the many thousands of kilometres of unlined irrigation canals creating huge problems with waterlogging and soil salinization. Within a few years of the Kara Kum Canal being constructed, for instance, the water table in the Merv region had risen over 20 m (Kornilov and Timoshinka, 1975) and vast tracts of land had become salinized (O’Hara, 1997). Water use at the field level also rose, as field size increased to accommodate increasingly bigger agricultural machinery, not only increasing the amount of time that it took to water fields, but also causing the traditional practice of night-time watering to be replaced by daily, and often continuous twenty-four-hour, irrigation. Yet despite an emphasis on the need to modernize the agricultural sector, furrow irrigation continued to dominate, with large and poorly levelled fields creating huge problems for irrigators. Moreover, unlike in the past, access to water was not a problem, with diversion schemes bringing what to many seemed an infinite supply of free water; people who had long
The archaeology of drylands
112
viewed water as a scarce commodity forgot its worth and wasted it. Further exacerbating the situation was the fact that government agencies rather than individual water users were responsible for, amongst other things, maintaining the irrigation infrastructure, dredging canals and ensuring that the drainage system was clean. At the farm level, maintenance became the responsibility of a few collective workers. In all cases, the bulk of the work was done using heavy equipment. Consequently, water users had little if anything to do with the management or maintenance of the water distribution and irrigation system. Despite Soviet successes in expanding the irrigation network and increasing agricultural output, the systems they built were (and still are) inflexible and highly inefficient. By the 1980s, agricultural land in Turkmenistan was being abandoned at a rate of over 50,000 ha per annum (Zaharchenko, 1994): clear testimony to the fact that this huge irrigation system is not sustainable.
CONCLUSION In Tables 6.1 and 6.2 we summarize the major trends in settlement and agriculture in southern Turkmenistan. It is evident that there is a strong correlation between the degree of urbanization and population size (themselves correlates of centralized political control) and the sophistication of irrigation technology. The range of crops likewise increases through time. Although
Table 6.1 Simplified chronological chart of prehistoric settlement in Turkmenistan Archaeological Settlement Irrigation systems Crops period and date (cal BC/AD) Neolithic Small farming villages on Crops cultivated in Main crop 6300–4800 BC piedmont of Kopet Dagh. areas of high water- einkorn; also table; possibly also emmer, hulled Key site: Jeitun. simple diversion of and naked streams. barley Hulled barley, Larger, complex Simple irrigation Eneolithic free-threshing (Chalcolithic) settlements, to 15 ha with assumed for 4800–3000 BC shrines and fortifications in piedmont and early wheat Middle Eneolithic (4000– occupation of 3500 BC); spread of Geoksyur; large settlements to Geoksyur irrigation canals oasis in Middle Eneolithic identified in but abandonment of oasis byGeoksyur oasis in late Eneolithic E.B.A. Key sites: Altyn(Namazga phase III). depe, Anau, Geoksyur, Namazga (phases I–III). Early Bronze Sites to 25 ha, restricted to Irrigation assumed Hulled barley,
Irrigation agriculture in Central Asia
113
piedmont zone. Key sites: for large settlements free-threshing Altyn-depe, Namazga (IV). in piedmont but no wheat, grape direct evidence. Main crop: Sites to 50 ha; monumental Smaller-scale architecture. Abandonment irrigation at northern hulled barley; of piedmont sites at end of fringe of Merv oasis. also freethreshing M.B.A. Major fortified wheat, lentil, settlements appear in Merv pea, chickpea, oasis in terminal period grape (2200–1900 BC). Key sites: Altyn-depe, Namazga (V), Gonur depe More dispersed, smaller Sophisticated canal Late Bronze sites (to 2 ha) in piedmont. irrigation in Merv Age Period of Bactrian-Margianaoasis, using water (L.B.A.) 1900–1500 BC Archaeological Complex; from main channels abundant large sites in Merv of rivers. oasis. Key sites: Namazga (VI), Gonur depe. Early Iron Age Abundant settlements (to 15 Introduction of qanat Broomcorn 1500–550 BC ha) in piedmont and oases. (kiariz) irrigation to millet Key sites: Tahirbaj, Yaz- piedmont. In Merv depe (Merv oasis). oasis settlement continues to shift to south. Age (E.B.A.) 3000–2500 BC Middle Bronze Age (M.B.A.) 2500–1900 BC
Table 6.2 Simplified chronological chart of settlement in the Merv oasis in the historic period Historical period Settlement Irrigation systems Crops and date (cal BC/AD) Achaemenid Founding of Achaemenid 550–330 BC city at Erk Kala c.500 BC; dispersed settlement centred on large buildings throughout the oasis and continuing to Seljuk period. Construction of the city Marked reduction in Seleucid rural settlement. of Antiochia (Gyaur (Hellenistic) Kala), incorporating Erk 330–140 BC Kala as citadel. Parthian New settlements in north Expansion of Main crops:
The archaeology of drylands
114
cotton, hulled 140 BC–AD 220 of the oasis; fortifications cultivated area; barley, freepossible on perimeter such as threshing wheat; construction of Gobekli. also lentil, central Murghab dam and extensive grape, almond canal network. Main crops: Sasanian Peak of settlement in the Cultivated area stable. cotton, hulled AD 220–651 Merv oasis; much barley, freecontinuity with Parthian threshing wheat; period. Possible also lentil, pea, construction of Wall of melon, grape, Antiochus (usually dated almond, peach, to the Hellenistic period), broomcorn which marks northern millet limit of most post-bronze age settlement. Umayyad Continuity in settlement Continuity in area Abbasid 8th–9th after Arab conquest of cultivated until centuries AD 651. Merv is capital Mongol conquest Samanid Seljuk of Seljuk empire in 11th results in destruction and Post Seljuk and 12th centuries. Sultan of dam system. 11th–13th Kala established 9th Abundant textual centuries Mongol century; fortified 12th evidence for conquest 1221 century. function of irrigation Devastating conquest of system in 10th–13th city by Mongols, but centuries. archaeological evidence suggests some postconquest occupation. Historical period and date (cal BC/AD) Timurid 14th–15th centuries Safavid 1502–1736
Settlement
Irrigation systems
Crops
New city built 1409 at Central dam rebuilt Abdullah Khan Kala but in Timurid period; destroyed in war of decline continues. 1727. Irrigation system Turkmen 18th–19th Dispersed settlement centuries with semi-independent functioning but small-scale landlords. cultivation. Introduction of large- American Russian conquest 1890 Establishment of cotton scale irrigation Soviet Union 1919– modern settlement at Mary; planned villages systems for cotton; species 1991 Republic of Karakum canal Turkmenistan 1991– and communal farms
Irrigation agriculture in Central Asia
present
throughout oasis.
115
completed 1967.
we would hesitate to identify simple cause and effect, it would appear that increased population was linked, through a complex sequence of interactions, with the expansion of irrigated agriculture and increased centralization of authority. The increase in population not only required an increase in the amount of land irrigated but also provided the workforce necessary for this expansion to take place. Irrigation flourished during periods of political stability, often when a single polity ruled over the area, and declined in periods of invasion or unstable internal political conditions. The decline of Merv can clearly be traced to the Mongol destruction of AD 1221–2. The Mongols took advantage of the fact that Merv, like most other settlements throughout Central Asia, was reliant on a single water source. In their rapid conquest of the region, the Mongols frequently forced communities to capitulate by disrupting water supplies and damaging irrigation structures, and all they needed to do at Merv was to destroy the main dam that controlled water in the oasis. Whilst the city was in part rebuilt, the irrigation systems were never fully reconstructed until the region once again came under the influence of another empire—that of the Soviets. Significantly, the widespread environmental degradation that plagues Soviet-built irrigation systems in the region does not appear to have been a major problem in the past, suggesting that sustainable irrigation in Turkmenistan is not only feasible but has been the norm. Traditional irrigation systems were generally localized and often dependent on a single water supply that was not only limited but also liable to fluctuate considerably from year to year. Water management required considerable skill, hence the mirab bashi, responsible for highly important and often contentious decisions on water allocation and distribution, was one of the most senior officials in central government—indeed, the success of many political officials often hinged on their skill at managing local water resources. Yet whilst water was managed centrally, all water users were responsible for the upkeep of the system, with those gaining more being expected to contribute more. The fact that individuals could benefit as a result of their efforts gave all users a vested interest in ensuring that the irrigation network was maintained and that water was used efficiently. The Soviet system effectively broke this link, with the system managed centrally but from afar. Together with the collectivization of land, the imposition of central planning meant that benefits were no longer linked to duty; water users had no say in how the system was managed, nor were they responsible for its maintenance. The establishment of myriad agencies to oversee different parts of the network resulted in unnecessary bureaucracy and waste. In sum, traditional irrigation and water distribution systems tended to be small, highly productive, well managed, extremely efficient and sustainable over the long term. In contrast, Soviet-built systems are huge, inefficient, inflexible, poorly managed and, for the most part, unsustainable. The decline in the water distribution and irrigation network since the break-up of the Soviet Union is thus unsurprising. What remains to be seen, however, is how this decline will be managed, and what can be done to ensure the future sustainability of Turkmenistan’s (and indeed Central Asia’s) water distribution and irrigation network. The Central Asian Republics have inherited a Soviet-built system and must learn to work
The archaeology of drylands
116
with the system and resources that are available. While we cannot revert to the past, Central Asia’s water managers would do well to look to the past for some of the answers to their current and future problems.
REFERENCES Adams, R.M. (1981) Heartland of Cities. Surveys of Ancient Settlement and Land Use on the Central Floodplain of the Euphrates. Chicago, University of Chicago Press. Babaev, A.G. (1994) Landscapes of Turkmenistan. In V.Fet and K.I.Atamuradov (eds) Biogeography and Ecology of Turkmenistan: 5–22. Dordrecht, Kluwer. Babaev, A.G. (1996) Problems of Aridland Development. Moscow, Moscow University Press. Bader, A., Gaibov, V. and Koshelenko, G. (1993/94) The northern periphery of the Merv oasis—from the Achaemenid period to the Mongol conquest. Silk Road Art and Archaeology 3:51–70. Bader, A.N., Gaibov, V.A. and Koselenko, G.A. (1995) Walls of Margiana. In A. Invernizzi (ed.) In the Land of the Gryphons: Papers on Central Asian Archaeology in Antiquity: 39–50. Florence, Le Lettere. Bader, A., Gaibov, V., Gubaev, A. and Koshelenko, G. (1996) The oasis of Merv: the dynamics of its settling and irrigation. Ancient Civilizations from Scythia to Siberia 3:49–60. Bartold, V.V. (1914) K istorii Orosheniya Turkestana. St. Petersburg, Sel’skago Vestnika. (In Russian.) Bartold, V.V. (1928) Turkestan down to the Mongol Invasion. London, Luzac, second edition. Bedford, D.P. (1996) International water management in the Aral Sea basin. Water International 21:63–9. Black, C., Dupree, L., Endicott-West, E., Naby, E., Matuszewski, D.C. and Waldron, A.N. (1991) The Modernization of Inner Asia. Armonk, NY, M.E.Sharpe. Boardman, S. (1997) Plant use in the Merv oasis. Iran 35:29–31. Dukhovny, V. (1995) Civilisation and Water Resources Management in Central Asia. Tashkent, World Bank. Gubaev, A., Koshelenko, G. and Tosi, M. (1998) (eds) The Archaeological Map of the Murghab Delta: Preliminary Reports 1990–95. Rome, ISIAO. Hannan, T. and O’Hara, S.L. (1998) Managing Turkmenistan’s Kara Kum canal: problems and prospects. Post-Soviet Geography and Economics 39:225–35. Harris, D.R. and Gosden, C. (1996) The beginnings of agriculture in western Central Asia. In D.R.Harris (ed.) The Origins and Spread of Agriculture and Pastoralism in Eurasia: 370–89. London, UCL Press. Harris, D.R., Masson, V.M., Berezin, Y.E., Charles, M.P., Gosden, C., Hillman, G.C., Kasparov, A.K., Korobkova, G.F., Kurbansakhatov, K., Legge, A.J. and Limbrey, S. (1993) Investigating early agriculture in Central Asia: new research at Jeitun, Turkmenistan. Antiquity 67:324–38. Harris, D.R., Gosden, C. and Charles, M.P. (1996) Jeitun: recent excavations at an early
Irrigation agriculture in Central Asia
117
Neolithic site in southern Turkmenistan. Proceedings of the Prehistoric Society 62:423– 42. Herrmann, G. (1997) Early and medieval Merv: a tale of three cities. Proceedings of the British Academy 94:1–43. Herrmann, G. and Petersen, A. (1997) The Ancient Cities of Merv, Turkmenistan. London, University College London, Institute of Archaeology, International Merv Project. Herrmann, G., Kurbansakhatov, K. and Simpson, S.J. (1998) The International Merv Project. Preliminary report on the sixth season (1997). Iran 36:53–75. Hiebert, F.T. (1994) Origins of the Bronze Age Oasis Civilization in Central Asia. Cambridge, MA, Harvard University, Peabody Museum of Archaeology and Ethnology, American School of Prehistoric Research Bulletin 42. Jacobsen, T. and Adams, R.M. (1958) Salt and silt in ancient Mesopotamian agriculture. Science 128:1251–8. Kohl, PL. (1984) Central Asia: Palaeolithic Beginnings to the Iron Age. Paris, Editions Recherche sur les Civilisations, Synthèse 14. Kornilov, B.A. and Timoshinka, V.A. (1975) The impact of the Kara Kum canal on the environment. Soviet Geography 15:308–14. Le Strange, G. (1905) The Lands of the Eastern Caliphate. Cambridge, Cambridge University Press. Lewis, R.A. (1966) Early irrigation in West Turkestan. Annals of the Association of American Geographers 56:467–91. Lipovsky, I. (1995) The central Asian cotton epic. Central Asian Survey 14:529–42. Lisitsina, G.N. (1969) The earliest irrigation in Turkmenia. Antiquity 43:279–88. Lisitsina, G.N. (1981) The history of irrigation agriculture in southern Turkmenia. In PL. Kohl (ed.) The Bronze Age Civilization of Central Asia: 350–8. Armonk, NY, M.E.Sharpe. Lisitsina, G.N. and Prishchepenko, L.V. (1976) The significance of paleoethnobotanical remains for the reconstruction of early farming in the arid regions of the USSR. Folia Quaternaria 47:83–8. Maas, E.V. (1987) Salt tolerance of plants. In B.R.Christie (ed.) CRC Handbook of Plant Science in Agriculture. Volume II: 57–71. Boca Raton, FL, CRC Press. Maier, U. (1996) Morphological studies of free-threshing wheat ears from a Neolithic site in southwest Germany and the history of the naked wheats. Vegetation History and Archaeobotany 5:39–55. Masson, V.M. (1957) Jeitun and Kara-tepe. Sovetskaya Arkheologiya 1:143–60. (In Russian.) Masson, V.M. and Sarianidi, V.I. (1972) Central Asia: Turkmenia before the Achaemenids. London, Thames & Hudson. Micklin, P.P. (1991) The Water Management Crisis in Soviet Central Asia. Pittsburgh, University of Pittsburgh Center for European and Russian Studies, Carl Beck Papers in Russian and East European Studies. Miller, N.F. (1993) Preliminary archaeobotanical results from the 1989 excavation at the central Asian site of Gonur Depe , Turkmenistan. International Association for the Study of the Cultures of Central Asia. Information Bulletin 19:149–63.
The archaeology of drylands
118
Miller, N.F. (1999) Agricultural development in western Central Asia in the Chalcolithic and Bronze Ages. Vegetation History and Archaeobotany 8:13–19. Moore, K.M., Miller, N.F., Hiebert, F.T. and Meadow, R.H. (1994) Agriculture and herding in the early oasis settlements of the Oxus Civilization. Antiquity 68: 418–27. Nesbitt, M. (1994) Archaeobotanical research in the Merv Oasis. Iran 32:71–3. O’Hara, S.L. (1997) Irrigation and land degradation: implications for agriculture in Turkmenistan, central Asia. Journal of Arid Environments 37:165–79. O’Hara, S.L. (in press) Central Asia’s water resources: contemporary and future management issues. International Journal of Water Resources Development. Orlovsky, N.S (1994) Climate of Turkmenistan. In F.Fet and K.I.Atamuradov (eds) Biogeography and Ecology of Turkmenistan: 23–48. Dordrecht, Kluwer Academic Press, Monographicae Biologicae 72. Pierce, R.A. (1960) Russian Central Asia 1867–1917: A Study in Colonial Rule. Berkeley, University of California Press. Powell, M.A. (1985) Salt, seed and yields in Sumerian agriculture. A critique of the theory of progressive salinization. Zeitschrift für Assyriologie und Vorderasiatische Archäologie 75:7–38. Serjeant, R.B. (1972) Islamic Textiles. Material for a History up to the Mongol Conquest. Beirut, Librairie du Liban. Smith, D.R. (1992) Salinization in Uzbekistan. Post-Soviet Geography 33:21–33. Wenke, R.J. (1975–76) Imperial investments and agricultural developments in Parthian and Sasanian Khuzestan: 150 BC to AD 640. Mesopotamia 10–11:31–221. Zaharchenko, B.T. (1990) Voda v Turmenskoy Zhizni (Water in Turkmen Life). Ashgabat. (In Russian.) Zaharchenko, B.T. (1994) A Brief History of the Construction of the Niyazov Kara Kum Canal. Ashgabad. (In Russian.)
Part III SAHARA AND SAHEL
7 Conquests and land degradation in the eastern Maghreb during classical antiquity and the Middle Ages JEAN-LOUIS BALLAIS
INTRODUCTION Ibn Khaldoun’s description of the eleventh-century Arab invaders of the Maghreb (Ibn Khaldoun, 1968), in which he likened them to plagues of locusts, is well known. Since then, many historians have tended to believe that the Arab invaders were responsible for the land degradation that is so visible today in many parts of north Africa. More than twenty years ago, after French decolonization, controversy was particularly strong (Berque, 1970, 1972; Cahen, 1968; Idris, 1968a, 1968b; Poncet, 1967, 1968), though interest has since decreased. Since this period, however, advances in geoarcheological research allow a reassessment of the role of Arab nomads in land degradation in north Africa, the basis for which was more ideological than factual. The purpose of this chapter is to discuss the geoarchaeological record of the eastern Maghreb (Fig. 7.1) and to compare it with the historical record of conquests, invasions and occupations, in order to assess the respective roles of climate and people in shaping this region’s landscape in classical antiquity and the Middle Ages.
THE ARRIVAL OF THE PHOENICIANS The Phoenicians settled on the eastern coasts of present-day Tunisia between the eleventh and ninth centuries BC (Decret, 1977). This period coincides with the beginning of the Late Holocene, a phase characterized throughout the region by a trend to aridification following the Middle Holocene climatic optimum (Ballais, 1991a). This aridification was the reason for the reappearance of aeolian deflation on the great ‘chotts’ or ‘sebkhas’ (salt flats) of southern Tunisia, as well as the incision of the lower prehistoric holocene terrace, especially in the present-day semi-arid subzone (Ballais and Benazzouz, 1994; Table 7.1). The phase is contemporaneous with the erosional crisis at the transition between the Bronze and Iron Ages in the northern Mediterranean
The archaeology of drylands
122
Figure 7.1 The eastern Maghreb, showing locations mentioned in Chapter 7 Key: 1. Ksar Rhilane; 2. chott Rharsa; 3. Ksar Rheriss; 4. Wadi KébirMiliane; 5. Henchir Rayada; 6. Wadi Chéria-Mezeraa; 7. Wadi el Akarit; 8. Haïdra. Stippling denotes major areas of sand dunes, horizontal dashed lines denote salt flats (‘sebkhas’ or ‘chotts’)
Table 7.1 Morphoclimatic evolution in the eastern Maghreb during the later Holocene; dates in radiocarbon years BP Isotopic chronology Morphogenesis Shifting sands incision 610 +/− 110 Terrace aggradation Incision 1350 +/− 70 Terrace aggradation 1470 +/− 190 Terrace aggradation 1730 +/− 185 Pedogenesis 2380 +/− 155 Flood deposits 2420 +/− 70 Gyttja deposits 2590 +/− 90 Pedogenesis Deflation
Conquests and land degradation in the eastern Maghreb
3680 +/− 160 4220 +/− 50
123
Incision Terrace aggradation Terrace aggradation
(Jorda et al., 1993). The characteristics of this sediment morphogenesis in the eastern Maghreb, as well as those of the associated pollen (Brun, 1989), indicate the occurrence of strong floods due to intense rainfalls, the latter probably highly concentrated and episodic. Thus, the climate was probably characterized by greater seasonality than before and a hydric balance less favourable than today. As far as we can tell, Phoenician colonization in the eastern Maghreb—for long limited to a few coastal sites—does not appear to have been responsible for this morphogenic crisis.
THE ROMAN CONQUEST AND COLONIZATION From the Punic Wars up to the Vandal conquest, during almost six centuries, three morphogenic phases can be discerned in the region: the end of an aggradation period; an incision period; and a second alluvial aggradation period (Table 7.1). The end of an aggradation period Between c.2400 and 2200 BP, aggradation became widespread once more. The evidence includes: washed sand deposits with Helicella molluscs in dunes on the eastern margin of the Grand Erg Oriental sand sea, at Ksar Rhilane in the Saharan subzone (Fig. 7.1: site 1; Fig. 7.2); gyttja accumulation in the Rharsa chott in the arid subzone (Ballais, 1992; Fig. 7.1: site 2); and fine alluvium along the watercourses flowing down from the Nemencha Mountains in the semi-arid subzone (Ballais and Benazzouz, 1994). Today at Ksar Rhilane sands are blown by the wind, and rillwash and sheetwash never occur. In the Rharsa chott, the principal deposit is sodium chloride. In the Nemencha Mountains, the fine alluvium has been organized in continuous beds by slow streams in a large channel. The mean annual accumulation rate was 1.4–2.2 mm, which is very close to the rate calculated for the low prehistoric Holocene terrace (Ballais, 1991b). The few pollen grains taken from those deposits show that a woodland could have colonized the slopes. All these characteristics are compatible with a more positive hydric balance in this phase than either during the previous climatic phase or today, especially in the arid and Saharan subzones. In comparison with the climatic optimum period (Ballais, 1991a), the degradation of climate in the semi-arid subzone is shown by a probable increase in summer evaporation, though winters remained sufficiently cool to permit the growth of Cedrus on the summits of Nemencha 1800 m above sea level. During this period, Roman penetration of the interior seems to have been very limited: no evidence for Roman agricultural works has yet been observed.
The archaeology of drylands
124
An incision period No deposit has been identified for the period from c.2200 BP to c.1650 BP (the second and third centuries AD), but there is widespread evidence for
Figure 7.2 Flood deposits at Ksar Rhilane (Tunisia): bedded silts and sands (dated to 2380 +/− 155 BP) with ripplemarks in the upper part, with the present-day dunes behind Photograph: J.-L.Ballais watercourse downcutting. This period was thus characterized by streams having the capability to incise their channels and to transport their alluvial sediments up to the base levels. It is possible precisely to measure neither the scale of the incision nor its annual rate, owing to the lack of chronological markers. Nevertheless, it should be remembered that the mean annual rate of downcutting of the low prehistoric Holocene terrace was 1.2 mm between 4000/3500 BP and 1700/1600 BP (Ballais, 1991b). Thus, during this period, slopes furnished very few colluvial sediments, indicating either that they were well protected by rather dense vegetation or that the agricultural techniques did not produce soil erosion. The beginning of this period was marked by the almost continuous expansion of Roman colonization, both westwards and southwards (Février, 1989). Even though they cannot be dated precisely, numerous dams were built at about this time on the watercourses of the Nemencha and Aurès mountains (Ballais, 1976; Birebent, 1964; Leveau, 1974–75) and in southern Tunisia (Ballais, 1990; Trousset, 1974). Sometimes their construction can be shown to be associated with the construction of the Roman frontier works (limes) at the boundary of the arid and Saharan subzones
Conquests and land degradation in the eastern Maghreb
125
(Baradez, 1949; Trousset, 1974). Presumably it was the combination of favourable conditions of climate, vegetation, and soils, and well-organized agricultural activities, that was responsible for such limited soil erosion through the second and third centuries AD. An aggradation period Evidence for an aggradation terrace dating to historic times is widespread along most of the watercourses, from the north to the south and from the humid subzone to the upper Saharan subzone. The exception is the lower Saharan subzone: so far, the historic aggradation terrace, unlike the late prehistoric terrace, has not been identified in the far south of Tunisia. As with the late prehistoric terrace, the later feature developed for the most part in those catchments formed in loose, erodible rocks. The sediments are generally fine in texture, beige in colour in the south and greyer in the north. This terrace also has an extensive surface area, particularly in the north, and forms a sediment unit that varies in thickness from 1 m (along small watercourses) to 5–6 m (along major rivers). On the largest rivers, detailed studies of sedimentation patterns show variations with latitude. Thus in the Wadi Leben terrace at Ksar Rheriss, towards 35°N in the arid subzone (Ballais, 1991b; Fig. 7.1: site 3), desiccation cracks appear in thin beds formed by washed clay and have been filled by sand during later flooding. These patterns, which are characteristic of intermittently-flowing wadis, disappear in the more humid subzones, with the exception of the Kébir-Miliane wadi (Fig. 7.1: site 4), which is today a perennial stream. Conversely, further to the north, the terraced sediments of many watercourses contain dark, hydromorphic, silt-like facies, rich in organic matter or in manganese oxide. In eighteen different cases, it is possible to correlate the presence of these facies with the perennial nature of the watercourse or, inversely, their absence with the intermittent nature of the watercourse. In three other cases, this correlation is not apparent. The sediments nearly always contain fragments of Roman or even earlier pottery (more than twenty recorded examples have been noted) and often charcoal, hearths or other artefacts. They cover the base of Roman bridges or aqueduct piles (Fig. 7.3) or fill in low-volume dams built during the second and third centuries AD. In another case, along the Wadi es Sgniffa, a whole ancient settlement is covered by alluvial sediments. There are still very few isotopic dates for the very low main historic Holocene terrace, but two examples of dated sediments are in the Wadi Chéria-Mezeraa (Fig. 7.1: site 6), which was radiocarbon-dated using samples of land molluscs to 1370 +/− 70 years BP (Fig. 7.4) and in the Wadi el Akarit (Fig. 7.1: site 7), where the terrace is dated to 1470 +/− 190 years BP (Ballais, 1995; Fig. 7.5). The rate at which the low terrace sediments of the historic period were accumulating became considerable: at the eighteen locations examined, the average reached 7.4 mm per year, which is five times the rate of accumulation of the lower (prehistoric) Holocene terrace, suggesting that the geosystems in which the historic terrace formed differed significantly from those when the earlier terrace accumulated. It is of interest to note that this historic aggradation started only when the area in question, with the exception of the Sahara, was occupied by sedentary populations as a result of a lengthy process of political and economic
The archaeology of drylands
126
Figure 7.3 The modern aqueduct crossing Wadi Bou Jbib, Carthage, built on the piles of the Hadrianic aqueduct Note: The right-hand pile is partly covered by grey silts belonging to the historic-period Holocene terrace Photograph: J.-L.Ballais evolution. Generally speaking, the northeast of Tunisia, near Carthage, was cultivated from perhaps the fifth century BC onwards and the cereal plains in the northwest of Tunisia and some parts of eastern Algeria from the third and second centuries BC. The steppes of Algeria, Tunisia and Libya were cultivated from the first and second centuries AD and the borders of the Sahara during the second century AD, at the time of the construction of the Severan limes (Trousset, 1986). If the observed variation in alluvial deposits was linked solely to the political and economic development of this vast area, we might expect to find the imbalance of the geosystems resulting from this development occurring at the same time in all places, with the threshold producing the change from incision to aggradation taking place at the same time in the north and south after eight centuries of sedentary settlement in the former region and after a few decades of such settlement in the latter. However, this coincidence, though not impossible, seems highly unlikely, particularly if the tremendous differences in mean annual rainfall between the north and the south are taken into account. In addition, it has now been shown that comparable terrace deposits were deposited throughout the Mediterranean, particularly towards the end of classical antiquity and in the early Middle Ages (Ballais and Crambes, 1992; Brückner, 1986; Vita-Finzi, 1969). The extensive occurrence of this feature can best be explained by an
Conquests and land degradation in the eastern Maghreb
127
Figure 7.4 Holocene terrace of Wadi Chéria-Mezeraa (Algeria) Note: The stratigraphy visible to the left of the standing figure is, from bottom to top: grey silts; bedded pebbles (Early Holocene); beige-grey silts (dated to 1370 +/− 70 BP); and a Roman drainage ditch filled with pebbles Photograph: J.-L.Ballais underlying climatic shift across the region. In view of the characteristics of the alluvial deposits, it seems likely that they originated from erosion of soils, particularly those that developed during the Holocene Climatic Optimum and at the beginning of the Roman period. However, land use systems in classical antiquity probably exacerbated these erosional trends. Presumably the spread of cultivation and ploughing destroyed a large part of the vegetation on the watersheds, reducing the cohesion of soils and superficial formations. It then required only a small change in rainfall characteristics, perhaps in the annual total, or at least in intensity and periodicity, to produce considerable soil erosion and the start, if not the return, of water in the stream channels and increased discharge, though this increase was not enough to carry the large sediment load from the slopes to the base levels. The absence of the very low historic terrace in the lower Saharan subzone is probably due to the lack of agriculture on the watersheds. For these very recent periods it is difficult to compare the climatic situation with that of Sahara. However, the fact that Acacia and Tamarix could be found in central Serir Tibesti at around 1700 BP and 1400 BP indicates that, at least in tropical Sahara, the mean annual rainfall was more than the present day rainfall of 5 mm (Pachur, 1974).
The archaeology of drylands
128
Figure 7.5 Holocene terraces in the Wadi el Akarit (Tunisia). A very low terrace of post-Islamic date (610 +/− 110 BP) is fitted into a gypseous terrace of late prehistoric date Photograph: J.-L.Ballais
THE VANDAL CONQUEST AND OCCUPATION The Vandals are not one of the beloved peoples of eastern Maghreb history. Too often it seems that their brief passage in the fifth and sixth centuries AD has no longer left traces (Courtois, 1955). In fact, despite new studies (Modéran, 1988), the Vandal period remains badly known for two main reasons. The first one was the lack of interest of French archeologists of the colonial period in the post-Roman civilizations (Février, 1989). The second one is a consequence of the first one: the destruction of the Vandal sites established on top of Roman towns. As a result, neither the limits of Vandal territory (in the present day Constantinois, for example) nor their modes of soil occupation and land use are well known. There is insufficient evidence on which to base any detailed discussion of climate, soils, vegetation and people’s possible effects on the landscape at this time. We can note, however, that alluvium continued to accumulate as the very low historic terrace (Table 7.1).
BYZANTINE CONQUEST AND COLONIZATION The Byzantines have been less neglected by French archeologists and historians of the colonial period because they presented themselves as the legitimate heirs of the Roman
Conquests and land degradation in the eastern Maghreb
129
emperors (Ducellier, 1988). They occupied the eastern Maghreb from the sixth to the seventh centuries AD. As for the Vandal phase, there is great uncertainty regarding modes of rural settlement and land use. Most information is available for the wars against Berbers (Modéran, 1989). In particular, most of the numerous small forts still visible today were attributed to Byzantine colonization, but it now appears that some are later than the Arab conquest (Mahjoubi, 1978). According to isotopic datings, the end of the aggradation of the very low historic alluvial terrace coincides more or less with the Byzantine period. This can be confirmed at Haïdra (Ammaedara; Fig. 7.1: site 8), where the foundations of a Byzantine bridge constructed at the same time as the sixth-century fort (Baratte and Duval, 1974) were dug into the alluvial deposit (Ballais, 1991c).
ARAB CONQUEST AND COLONIZATION Seventy years were necessary for successive waves of Arab forces to conquer the eastern Maghreb during the second part of the seventh century AD (Marçais, 1946), presaging the high Islamic period, which was a time of general economic prosperity (Vanacker, 1973). The last wave of Arab invaders, the well-known Hilalian, arrived in the middle of the eleventh century. They are described as nomadic shepherds coming from Upper Egypt, the ‘plague of locusts’, in Ibn Khaldoun’s memorable phrase, who ‘pushed their flocks into the middle of the fields, devastated the gardens, stripped and ill-treated the country persons, plundered the hamlets’ (Marçais, 1946). In theory, the consequences of such devastation would have been so disastrous that they would have provoked catastrophic and long-lived decline in the Tunisian economy (Al-Idrisi, 1983; Ibn Khaldoun, 1968; Marçais, 1946; Vanacker, 1973). Following the aggradation of the very low historic terrace, the general trend for watercourses in the study area was vertical incision (Table 7.1), with two to three interruptions. The main interruption can be seen in the very low post-Islamic terrace, which, as far as we know today, is little represented, presumably because of its small size. With only one exception (Wadi Kébir-Miliane), this terrace covers very small areas, in particular in convex meander lobes, and its height above the major bed rarely exceeds 2 m. Occasionally it appears as a rocky terrace that was breached in the previous buildup; elsewhere, the facies can sometimes be compared to that of the previous terrace, though it is sometimes considerably coarser, at least at the base. The age of the terrace is still rather uncertain, because appropriate means of dating are not available, but at Henchir Rayada (Fig. 7.1: site 5) it contains Islamic pottery from the tenth/eleventh centuries and in Wadi el Akarit (Fig. 7.1: site 7) it was dated by radiocarbon using collagen to 610 +/− 110 years BP (Fig. 7.5). This terrace is thus much younger than the period of the presumed Hilalian invasions. As for the previous terrace, the widespread presence of a terrace of the same age can be seen throughout the Mediterranean basin (Ballais and Crambes, 1992; Vita-Finzi, 1969). Moreover, in contrast to the final years of classical antiquity (the third to the fifth centuries AD), population was probably low at the beginning of the Hafside period in Tunisia during the twelfth century AD. This period marked the end of the medieval climatic optimum and the beginning of the Little Ice Age in Europe (Lamb, 1977; Le
The archaeology of drylands
130
Roy-Ladurie, 1967). In Morocco, the existence of a Little Ice Age is controversial (El Bouch and Ballais, 1997; Lamb et al., 1989; Stockton et al., 1985). Once again it seems most likely that such a widespread terrace resulted from climatic causes, but further studies will be required to clarify this point. This conclusion emphasizes the extreme ideological character of the theories regarding the impact of nomadic shepherds on the landscape at the time of the Arab invasions. Even if those shepherds did in fact cut down trees to any extent, their main effect would have been to substitute pastures for cultivated fields. The consequences would have been as follows: an increase in the rate of vegetation cover over the soil; a consequent decrease in the direct splash impact of rain drops on the soil and thus of pluvial erosion and sheet erosion; and an increase in the ‘roughness’ of the terrain and thus a diminution in wind erosion. Within such a model it is necessary to moderate the intensity of such processes according to the climatic subzones and different grazing intensities, but it seems realistic to suppose that, in general, such a move from arable to pastoral land use is likely to produce less, rather than more, soil erosion.
CONCLUSION In the eastern Maghreb during classical antiquity and the Middle Ages, fluctuations in geosystems and, in particular, increases in soil erosion can be seen to have reflected specific combinations of climatic change and human activities. A climatic fluctuation that increases the intensity of rains, or the annual amount of precipitation, affects slopes only if they have been made vulnerable by vegetation degradation or by cultivation systems that have not been designed to counteract erosion. In other words, phases of massive agricultural colonization and phases of extension of the cultivated surface are very favourable to such erosion. This was the case in the study area, as in many parts of the Mediterranean, during the Roman period. On the other hand, periods of conquest generally seem to have been characterized by a contraction of the cultivated surface and a progressive development of ‘natural’ vegetation or of pastures that limited soil erosion. This may have been the situation in the case of the nomadic Hilalian shepherds of the Arab conquest. However, there were exceptions to these trends, in particular in the irrigated zones and in the terraced mountains.
REFERENCES Ballais, J.-L. (1976) Morphogenèse holocene dans la region de Chéria (NementchasAlgérie). Actes du Symposium sur les Versants en Pays Méditerranéens: 127–30. Aixen-Provence, CEGERM 5. Ballais, J.-L. (1990) Terrasses de culture et jessours du Maghreb oriental. Méditerranée 3.4:51–3. Ballais, J.-L. (1991a) Evolution holocene de la Tunisie présaharienne et saharienne. Méditerranée 4:31–8. Ballais, J.-L. (1991b) Vitesse d’accumulation et d’entaille des terrasses alluviales
Conquests and land degradation in the eastern Maghreb
131
holocènes et historiques au Maghreb oriental. Physio-Géo 22–23:89–94. Ballais, J.-L. (1991c) Les terrasses historiques de Tunisie. Zeitschrift fur Geomorphologie Suppl. Bd 83:221–6. Ballais, J.-L. (1992) Le climat au Maghreb oriental: apports de la géomorphologie et de la géochimie. Les Nouvelles de l’Archéologie 50:27–31. Ballais, J.-L. (1995) Alluvial Holocene terraces in eastern Maghreb: climate and anthropogenic controls. In J.Lewin, M.G.Macklin and J.C.Woodward (eds) Mediterranean Quaternary River Environments: 183–94. Rotterdam, Balkema. Ballais, J.-L. and Benazzouz, M.T. (1994) Données nouvelles sur la morphogenèse et les paléoenvironnements tardiglaciaires et holocènes dans la vallée de l’oued ChériaMezeraa (Nemencha, Algérie orientale). Méditerranée 3.4:59–71. Ballais, J.-L. and Crambes, A. (1992) Morphogenèse holocene, géosystèmes et anthropisation sur la montagne Sainte-Victoire. Méditerranée 1.2:29–41. Baradez, J. (1949) Vue Aérienne de l’Organisation Romaine dans le Sud Algérien. Fossatum Africae. Paris, Arts et Métiers Graphiques. Baratte, F. and Duval, F. (1974) Les Ruines d’Ammaedara-Haïdra. Tunis, Société Tunisienne de Diffusion. Berque, J. (1970) Les Hilaliens repentis, ou l’Algérie rurale au XVIe s., d’après un manuscrit jurisprudentiel. Annales Economic Société Civilisation 5:1325–53. Berque, J. (1972) Du nouveau sur les Banû Hilâl? Studia Islamica 36:99–113. Birebent, J. (1964) Aquae Romanae: Recherches d’Hydraulique Romaine dans l’Est Algérien. Alger, Baconnier frères. Brückner, H. (1986) Man’s impact on the evolution of the physical environment in the Mediterranean region in historical times. GeoJournal 13 (1):7–17. Brun, A. (1989) Microflores et paléovégétations en Afrique du Nord depuis 30 000 ans. Bulletin de la Société Géologique de France 8(1):25–33. Cahen, C. (1968) Quelques mots sur les Hilaliens et le nomadisme. Journal of Economic and Social History of the Orient 11 (1):130–2. Courtois, C. (1955) Les Vandales et l’Afrique. Paris, Arts et Métiers Graphiques. Decret, F. (1977) Carthage ou l’Empire de la Mer. Paris, Editions du Seuil. Diehl, C. (1896) L’Afrique Byzantine: Histoire de la Domination Byzantine en Afrique (533–709). Paris, Imprimerie Nationale. Dore, J.N. and van der Veen, M. (1986) ULVS XV: radiocarbon dates from the Libyan Valleys Survey. Libyan Studies 17:65–8. Ducellier, A. (1988) Les Byzantins. Histoire et Culture. Paris, Editions du Seuil. El Bouch, A. and Ballais, J.-L. (1997) Travertinisation, détritisme et anthropisation a Fès (Maroc). Würzburger Geographische Arbeiten 92:213–24. Fèvrier, P.-A. (1989) Approches du Maghreb Romain. Aix-en-Provence, Edisud. Ibn Khaldoun, A. (1968) Muqqadima. Beirut, Commission Libanaise pour la Traduction des Chefs d’Oeuvre, translated by V.Monteil. Idris, H.R. (1968a) L’invasion hilalienne et ses consequences. Cahiers de Civilisation Médiévale 3:353–71. Idris, H.R. (1968b) De la réalité de la catastrophe hilalienne. Annales Economic Société Civilisation 23:390–6. Al-Idrisi, A. (1983) Le Maghrib au 6e siècle de l’Hégire. Paris, Publisud, translated by
The archaeology of drylands
132
Hadj. Sadok. Jorda, M., Parron, C., Provansal, M. and Roux, M. (1993) Erosion et détritisme holocene en Basse Provence calcaire. L’impact de l’anthropisation. Travaux du Centre Camille Jullian 14:225–33. Lamb, H.F., Eicher, U. and Switsur, V.R. (1989) An 18,000-year record of vegetation, lake-level and climatic change from Tigalmamine, Middle Atlas, Morocco. Journal of Biogeography 16:65–74. Lamb, H.H. (1977) Climate: Present, Past and Future. London, Methuen. Le Roy-Ladurie, E. (1967) Histoire du Climat depuis l’An Mil Paris, Flammarion. Leveau, P. (1974–75) Une vallée agricole des Némenchas dans l’Antiquité romaine: l’oued Hallail entre Djeurf et Aïn Mdila. Bulletin Archéologique du Comité des Travaux Historiques et Scientifiques 10–11b:103–21. Mahjoubi, A. (1978) Recherches d’Histoire et d’Archéologie à Henchir El-Faouar (Tunisie). Tunis, Publications de l’Université de Tunis. Marçais, G. (1946) La Berbérie Musulmane et l’Orient au Moyen Age. Paris, Aubier. Modéran, Y. (1988) Les premiers raids des tribus sahariennes en Afrique et la Johannide de Corippus. Histoire et Archéologie de l’Afrique du Nord 2:479–90. Modéran, Y. (1989) La découverte des Maures. Réflexions sur la ‘reconquête’ byzantine de l’Afrique en 533. Cahiers de Tunisie 43:155–6, 211–38. Pachur, H.J. (1974) Geomorphologische Untersuchungen im Raum des Serir Tibesti (Zentrasahara). Berliner Geographische Abhandlungen 17:6–58. Poncet, J. (1967) Le mythe de la ‘catastrophe’ hilalienne. Annales: Economie, Société, Civilisation 23:660–2. Poncet, J. (1968) Encore à propos des Hilâliens: la mise au point de R.Idris. Annales: Economie, Société, Civilisation 23:660–2. Stockton, C.W., et al. (1985) Long-Term Reconstruction of Drought in Morocco. Tucson, University of Arizona Press. Trousset, P. (1974) Recherches sur le Limes Tripolitanus du Chott el Djerid à la Frontière Tuniso-Libyenne. Paris, CNRS. Trousset, P. (1986) Limes et frontière climatique. Histoire et Archéologie de l’Afrique du Nord: 55–84. Paris, CTHS. Vanacker, C. (1973) Geographic économique de l’Afrique du Nord selon les auteurs arabes du IXe au milieu du XIIe siècle. Annales: Economie, Société, Civilisation 3: 659–80. Vita-Finzi, C. (1969) The Mediterranean Valleys. Geological Changes in Historical Times. Cambridge, Cambridge University Press.
8 Success, longevity and failure of arid-land agriculture: Romano—Libyan floodwater farming in the Tripolitanian pre-desert DAVID GILBERTSON, CHRIS HUNT AND GAVIN GILLMORE
INTRODUCTION This chapter examines the successes and failures of Romano—Libyan and later floodwater farmers in the Tripolitanian pre-desert in Libya (Fig. 8.1). This vast region of rocky plateaux and incised wadis lies between the higher, better watered, Gebel Nafusa and the Mediterranean coastlands to the north, and the desert of the Hamada al-Hamra to the south. Critical to floodwater farming in this region were complex networks of walls that were used to manage occasional storm-water to sustain agricultural settlement, with many impacts on soils, geomorphology and biogeography (Fig. 8.2). The vast scale of the ancient settlement stands in stark contrast to the depopulated modern landscape. As long ago as 1857, the explorer Heinrich Barth recorded that the landscape displayed a ‘sea-like level of desolation’ (1857:125). Today, the region remains empty and inhospitable, except for a few pastoralists with mixed herds of sheep and goats. The pastoralists exploit bore water, rare springs and small wells. In the hotter and drier parts of the year, herds may be taken north to the better watered and cooler Gebel. The modern towns of Beni Ulid and Mizda are the only significant settlements in the region. The area around Beni Ulid is a dense mixture of modern development and remains of ancient buildings, evidencing substantial occupation from Romano—Libyan times to as recently as only 400 years ago (Gilbertson and Hunt, 1988). The Wadi Merdum through Beni Ulid is also one of the last, if not the last, wadis in this region where active floodwater farming continues. Date palms, figs, plums and ancient olive trees can still be found growing along 4–5 km of the modern wadi floor (Goodchild and Ward-Perkins, 1949), together with eucalyptus, whilst the sheltered floors of small side-wadis sometimes yield a crop of barley.
The archaeology of drylands
134
Figure 8.1 Tripolitania, northwest Libya, showing the principal landforms and settlements, and the location of the UNESCO Libyan Valleys Survey Key: The 200 mm, 100 mm and 20 mm isohyets are shown as dashed lines; the contours are in metres
Figure 8.2 Romano-Libyan floodwater farming in the Wadi Gobbeen in the Tripolitanian pre-desert Note: A wadi-edge diversion wall is visible in the right foreground, and a series of crosswadi walls down the wadi in the distance, with the fortified farm (gasr) on the horizon on the right Photograph: G.Barker
Success, longevity and failure of arid-land agriculture
135
ROMANO-LIBYAN AND LATER SETTLEMENT The character of ancient settlement and land use in the Tripolitanian pre-desert was summarized in the two-volume report of the UNESCO Libyan Valleys Survey (ULVS) (Barker et al., 1996a, 1996b). The survey was a combined enterprise between the Department of Antiquities in Tripoli and a group of archaeologists and geographers from the Universities of Leicester, Huddersfield, Manchester and Sheffield in England. The main archaeological evidence is presented in a Gazetteer, based upon 2,437 site records (Barker et al., 1996b). Many individual sites are themselves complex: for example, a single entry deals with the complex networks of hundreds of substantial wadi walls that were mapped over 10 km of the Wadi Umm el-Kharab (Barker et al., 1996a). The scale and significance of the past occupation of the predesert are clear from a cursory examination of the distribution of large open (that is, unenclosed or undefended) farms and farmsteads, some built in the Opus Africanum style, attributed to the first to the third centuries AD, and the imposing, enclosed and possibly fortified barn-like gsur (Fig. 8.3). The
The archaeology of drylands
136
Figure 8.3 Simplified distribution of Opus Africanum and other early Romano-Libyan farms and farmsteads (above) and fortified farms (gsur) (below) in the UNESCO Libyan Valleys Survey study area Source: After Barker et al., 1996:162, 165
Success, longevity and failure of arid-land agriculture
137
latter are mainly attributed to the third to fourth centuries AD, though many examples also date to the Islamic period, and at some gsur, significant activity continued until the sixteenth century AD (Gilbertson and Hunt, 1988). These substantial settlement remains appear to have been built by local, neo-Punicspeaking, Macae tribespeople, here referred to as Romano-Libyan (Mattingly, 1989). Initially transhumant pastoralists, they took advantage of the extension of Roman influence into Tripolitania rapidly to develop a robust, long-lasting, mixed farming economy and a substantial increase in population, in a desert environment broadly similar to that of today. The ancient farming was more than self-sufficient, producing surpluses of olive oil, perhaps other tree crops, grapes and cereals. Stock keeping no doubt continued through the Romano—Libyan period, and it is transhumant pastoralism that characterizes the human geography of the modern landscape. A substantial trade took place between the pre-desert, the Mediterranean coast and beyond: significant quantities of olive oil and perhaps cereal crops were sent north to the coastal cities, whence some were exported to the wider Roman world. Products produced by better-watered regions were imported into the pre-desert, including even ‘exotic’ foodstuffs such as deep-water sea fish. The density of Romano—Libyan and later settlement in the pre-desert varied significantly through both space and time (Flower and Mattingly, 1995; Mattingly with Flower, 1996). A dramatic transformation of pre-desert settlement, with a rise in population to about 20,000 in perhaps 2,000 farms, occurred in the study region during the first century AD. The longevity of this occupation is rarely securely known. The estimated 1,000 gsur built from the third century AD were perhaps fortified farms. These must have provided massive and secure storage of crops and food as buffers against sequences of adverse drought years. By the sixth to seventh centuries AD the entire network had probably become consolidated into major ‘agricultural estates’ controlled by powerful, rich, local elites or warlords (Mattingly, 1996). These ‘estates’ are manifest in the changing density of both walls and gsur, with no obvious linkages to topographic or hydrological features (Gilbertson and Chisholm, 1996; Mattingly, 1996). Curiously, this intensification and reordering of settlement in the pre-desert were associated with a decreasing import of goods from the Mediterranean countries and decreasing quantities of olive oil sent to the coast (Mattingly, 1996). GIS-based analyses suggest both a general ‘thinning’ of settlement and a slight northward and westward shift, from the early extensive phase through this late Romano—Libyan period, followed by a notable shift northwards and a trend towards clustering of settlement during the Islamic period (Mattingly with Flower, 1996). Eventually, the only remaining large settlement with significant modern floodwater farming was Beni Ulid on the northern edge of the predesert. The information that is now available to explore these ideas is vastly superior to that compiled before the ULVS survey. It is nevertheless limited in scope and reliability and is often incapable of sustaining sophisticated theoretical enquiry. For example, the logistical reality of access by off-road vehicle in the difficult terrain and generally arduous circumstances forced a concentration of surveys upon the more accessible wadis, at the expense of other wadis and the vast plateaux between them. This survey pattern is likely to have under-represented features such as ancient farms in basins or on ancient
The archaeology of drylands
138
route-ways on the plateaux, as found, for example, near the Wadi Umm el-Kharab (Barker et al., 1991), or detected through Landsat images and field survey (Dorsett et al., 1984). Detailed surveys on foot were carried out only in the Wadi Umm el-Kharab and in small sectors of the Wadis el-Amud, Gobbeen, Mansur and Mimoun (Barker et al., 1996a), probably locating only 70–80 per cent of the walls initially present (Barker with Gilbertson, 1996). Some wadi-floor walls were totally buried by sediment and detected only in gully exposures or as lines of bushes (Gilbertson et al., 1994). Elsewhere, our understanding is mostly based upon field sketches and simple field maps (Gilbertson et al., 1984) as a consequence of the lack of air photographs and appropriate base maps for field workers in this remote and politically sensitive area.
COMPETING EXPLANATIONS Reasons for the initiation, growth, stability and eventual decline of settlement and farming in this pre-desert region remain remarkably unclear. Some possible explanations are set out below. First we consider the possible explanations—singly or in combination—for the abandonment of Romano-Libyan settlements and farms in the Tripolitanian pre-desert. Human processes • Political, social and economic changes at the coast and in the wider Roman empire brought about the loss of the market for pre-desert produce. • Political, social and economic changes at the coast and in the wider Roman empire brought about the loss of the ‘Roman’ technocrats who made the system work. • The arduous life did not give sufficient rewards to the farmers. • It was too unrewarding and unexciting for young people—the opportunities and good life in the coastal cities were too attractive. • Insecurity—raids and menace from the desert tribes to the south could no longer be managed. • The first and second Arab invasions prompted the abandonment of the farms. • The quelling of nomadism by the development of farms/settlements and the provision of better water supplies in cisterns caused unacceptable, albeit localized, environmental degradation through over-grazing, loss of pastures, excessive soil loss and so on. • The demands of the imperial economy and imperial attitudes undermined the ideology, attitudes, self-sufficiency and ultimately the vigour, agricultural success and subsistence basis of the indigenous population at the floodwater farms. • The settlers were pioneers, not developers, and were followed by parasitic professionals who failed to support the development process: the professionals so drained the economic basis of the region that the economy failed. • The region was drained by the activities of the equivalent of itinerant ‘carpetbaggers’ who drained this region of its vitality and wealth. • The question is based upon a misinterpretation: the people were never fully
Success, longevity and failure of arid-land agriculture
139
sedentary, and failed to return rather than left—it was a threat to, or loss of, their mobility that was critical. Natural processes • A disease epidemic removed the capacity of people, crops or livestock to continue in this demanding desert environment. • The ‘long-term’ climate became too arid. • There was one or more relatively short but pernicious droughts that lasted too long for people unable to import food or water in adequate quantities to sustain themselves and their plants and animals through such adverse times. • The inherent instability of the biophysical systems in dryland environments led to the growth of mutually reinforcing links between any of many possible cause-andeffect relationships, which resulted in the non-reversible growth and persistence of an originally minor human or environmental disturbance, and subsequent desertification or non-sustainable intensification of grazing. ‘Induced environmental change’ • There was a local version of the ‘Charney Effect’: there was an increased exposure of the soil and rock at ground surface as a result of more intensive and widespread livestock and arable farming in the pre-desert, which eventually brought about a downward spiral resulting in progressive desiccation. • Intensive and widespread livestock and arable farming raised such large quantities of dust into the atmosphere that a regional climatic change was induced, resulting in greater aridity. • Accelerated soil erosion made arable and pastoral farming too difficult on the plateaux. • Goats and sheep ‘ravaged’ the pastures. • Excessive trapping of water in soil produced soil salinization in plateau-basins and on wadi floors. • Excessive removal of vegetation led to the salinization of soil: the reduction of evapotranspiration caused a greater deposition of salts as a result of the induced (periodic) soil water-logging. • Excessive cropping caused the loss of soil fertility and unexpected crop failure. • The supply of timber for fuel and other domestic purposes was effectively exhausted. • The inherent instability of the biophysical systems in dryland environments led to the growth of mutually reinforcing links between any of many possible cause-andeffect relationships, which resulted in the non-reversible growth and persistence of an originally-minor environmental disturbance, and subsequent desertification or non-sustainable intensification of grazing. Now we turn to the possible explanations for the expansion of these same settlements.
The archaeology of drylands
140
Human processes • There was a demand from the coastal cities and wider Roman economy for olive oil, grapes, wheat, barley, dates, figs, pistachios and animal products, which was met by indigenous people and/or settlers. • The southern ‘frontier’ of the Roman empire was secured by a ‘defence in depth’ made up of farming communities and some military installations. • Driven by people who may have had (variously) particular types of ideology or belief, misunderstanding, adventure, anticipation of quick or long-term profit, a need to escape from the confines of contemporaneous life, optimism, a pioneer culture, military imperatives, a need for new lands, and so on, the ‘frontier’ moved south bringing with it farmers and settlers. • The arrival of Roman ‘know-how’ and ‘can-do’. Environmental processes • The climate was ‘wetter’ (caused by cloud cover in greater quantity, differently distributed, more reliable and/or more frequent), so agriculture prospered and settlement extended deeper into the pre-desert. • The climate worsened, and the settlers or indigenous people caught between the desert and hostile neighbours were obliged to develop intensive agriculture in the wadis. • ‘Pioneer’ or ‘eccentric’ people had experimented with small-scale cultivation (perhaps experimental in outlook) using ideas from indigenous people or elsewhere and started a ‘fashion’ that was thought worthy or otherwise good for personal development. • The farms were started and maintained as a tax avoidance or tax mitigation measure. • The effect of the water-harvesting and the planting of tree and other crops as part of floodwater farming was to so change the nature of the relationships between climate, soil and vegetation that the pre-desert became transformed by many other occupants who created, through their type of land use, a biologically-productive, as well as more wooded and economically-productive, environment. • Relatively minor, small-scale developments associated with water-harvesting and plant production produced a series of biophysical feedbacks between the various components of the pre-desert environmental system. These proceeded to reinforce each other, eventually leading to the transformation of the pre-desert from one stable state, characterized by relatively low biological productivity, to a different stable state characterized by a much higher level of biological activity (and soil developments, precipitation and so on), which appeared to encourage the extension and/or the intensification of farming developments. Several explanations appear to be contradictory, some are counter-intuitive, whilst others probably exceed the strength of the present evidence. Several possible explanations cannot be investigated with the methods currently available. Indeed, many explanations would be difficult to explore even in arid lands that have been the subject of sustained,
Success, longevity and failure of arid-land agriculture
141
intensive and extensive systematic study—which is far from the case here. Through the filter of nearly 2,000 years, it is difficult to separate cause and effect and ‘triggers’ from pre-disposing or maintaining factors, as well as to disentangle feedback effects and synergies. Key events will rarely have occurred in isolation. Numerous combinations of processes will have operated at different times, leading to a variety of outcomes. The most critical information and key ideas are set out below, together with new evidence produced since the publication of the UNESCO survey in 1996.
AGRICULTURE The products of Romano—Libyan agriculture in the pre-desert are summarized in Table 8.1 (Gilbertson et al., 1994; van der Veen et al., 1996). They are essentially similar to the products of modern intensive mixed farming in many Mediterranean lands, including the Libyan coastal plain and its hinterlands. One noteworthy archaeozoological pattern was the increasing proportion of wild animals consumed with increasing distance south into the more arid parts of the pre-desert. Overall, for most people, meat was likely to have been a luxury item. Of at least equal importance were the wool, hair, milk, labour and manure that domestic animals produced. There were many similarities to, but also some differences from, the more ‘normal’ agricultural economy of the Mediterranean lands to the north. For example, the crops produced are similar to those of rain-fed agriculture to the north, but the details of the agricultural practices used must have been notably different, since precipitation in the pre-desert is both minimal and unreliable—less than 25–100 mm a year, with substantial variability in both time and space. In common with ancient arid-land farming in many other deserts, agriculture and settlement in this arid land were dependent upon
Table 8.1 Farm products of the Tripolitanian pre-desert, first to fifteenth centuries AD Farm products Centuries AD PLANT CROPS 1–5 10–16 Hordeum vulgare (hulled six-rowed barley) + + Triticum (wheat) + + Pisum sativum (field pea) + + Lens culinaris (lentil) + + Other pulses + + Ficus carica (fig) + + Vitis vinifera (grape) + + Phoenix dactylifera (date) + + Olea europea (olive) + + Prunus amygdalus (almond) + + Pistacia atlantica (wild pistacia) + ANIMAL PRODUCTS Sheep/goat + +
The archaeology of drylands
142
Gazelle + + Bovid + + Pig + Canid + + Camel + + Hare/rabbit + + Equid + + Antelope + + The present archaeobotanical evidence suggests that there were no fundamental differences between the agricultural economies of the Romano—Libyan, Late Antique and Islamic periods. In general, hunted as opposed to herded animals became increasingly more important further south into the desert. harvesting rainwater, with particular emphasis upon managing overland flow and controlling floodwater in the wadi floors—practices originally termed ‘floodwater farming’ by Bryan (1929) in his studies in the American Southwest. Unfortunately, the palaeoeconomic analyses that underpin these ideas derive from a minute subset of sites in the study region. Nevertheless, the ULVS survey design did attempt to ensure that material was available from the primary types of agricultural buildings located during the project. The studies of seeds and animal bones that underpin present understanding derive from excavations at middens or in buildings at only four open farms, three of which were associated with olive presses, and six gsur, one of which was associated with an olive press.
ENVIRONMENT Knowledge of the detailed environmental history of the pre-desert from the midHolocene to the period of instrumental records is vital to understanding its human history. Unfortunately, such understanding is often rudimentary. Palaeoenvironment Only four palaeoenvironmental studies have been reported previously. Three of these suggested the cultivation of olives in the Romano—Libyan period. A study of the sedimentary fill of a karstic plateau-basin north of Beni Ulid indicated the presence of a wetter climate during the early Holocene, with shallow semi-permanent lakes surrounded by a grassy steppe, perhaps with some scrub or trees, in what are nowadays dry basins (Gilbertson et al., 1994; Gilbertson et al., 1994). Aridification took place from 4,000 to 5,000 years ago, creating an environment essentially similar to the modern arid steppe. A study of cave deposits near Beni Ulid indicated the essential similarity of RomanoLibyan and modern conditions (Gale et al., 1993). A third study analyzed pollen from sediments infiltrated into a Romano—Libyan cross-wadi wall in the Wadi Mansur (Hunt et al., 1986), suggesting a degraded steppe flora very similar to that of the modern Wadi
Success, longevity and failure of arid-land agriculture
143
Mansur, and the cultivation of cereals. The fourth study was a multi-disciplinary assessment of sediments infilling the conduit that fed water to gasr Mm10 in the Wadi Mimoun (Hunt et al., 1987). These deposits probably date from the abandonment of the gasr in the late Romano-Libyan period. A landscape of steppe and scrub was suggested, more biodiverse and perhaps wetter than occurs today. Cereals were cultivated. The high frequency of charcoal recovered suggests burning nearby. Interestingly, pods of Medicago sp. were also excavated from these deposits. This plant is native to the pre-desert and is of considerable interest. Its seeds were also recovered from Romano—Libyan deposits further south at Ghirza (van der Veen et al., 1996; Fig. 8.4). During the nineteenth century in South Australia, a system of medic/cereal rotation was developed by dryland farmers to improve nitrogen levels in their soils—a system still used today (Chatterton and Chatterton, 1984). Medic-enriched grasses are sown and allowed to flower and produce seeds in the first season; cereals are grown in the second; seeds from the first medic pasture then germinate to create another pasture in the third season. The adoption of this ‘Australian’ system in parts of the Tripolitanian and Cyrenaican Gebel in the 1970s and 1980s increased cereal yields by as much as 50 per cent, and allowed stocking rates to rise dramatically. Chatterton and Chatterton (1984) argued that, if Romano-Libyan farmers had left land fallow for two or three years between cereal crops, the resulting substantial medic pasture would have improved soil fertility and grazing. Such a scenario is probable, because in many areas rainfall would not have been sufficient every year to justify planting a cereal crop. Over time, the ancient farmers may well have noticed the benefit of this type of crop rotation. A new palaeoecological study (Hunt, unpublished data) is reported in outline here. This is a palynological study of modern and Romano-Libyan coprolites from the middens and room fills of the farmstead Lm4 at Wadi el-Amud in the south of the pre-desert (Gilbertson and Hunt, 1990). The modern samples reflect an extremely degraded environment, with low local
The archaeology of drylands
144
Figure 8.4 A Romano—Libyan fortified farm (gasr) and its satellite buildings at Ghirza Kite photograph: G.D.B.Jones pollen productivities and the local flora dominated by drought-resistant species. In contrast, the samples from contexts dating to the Romano—Libyan and Arab periods contain pollen of grasses and a diverse steppic flora, with abundant pollen of cereals and olives reflecting crop plants. Critically, also, the taphonomic patterns suggest that animals were fed on monoculture crops—grasses, cereal waste and chenopods. Evidence from
Success, longevity and failure of arid-land agriculture
145
Lm4 had previously suggested the stalling of animals at this site (Gilbertson and Hunt, 1990). This is a very different type of husbandry pattern than occurs today in the predesert, where goats forage widely and indiscriminately. Palaeoclimatology Palaeoclimatic evidence from Morocco and the Nile basin suggests severe, prolonged, Late Holocene drought events. The Late Holocene was notably drier than much of the Early and Middle Holocene (Hassan, 1981, 1996, 1997, 1998; Lamb et al., 1994, 1995). Significant arid phases were identified for 4600–4000 BP, 3800–3600 BP, 2500 BP, 2000 BP, 1500 BP and approximately 700 BP (radiocarbon years). The flood record of the River Nile is especially interesting for the last 1,500 years, indicating low to very low flows from AD 760 to 1070, with especially low flows between AD 930 and 1070 and between AD 1180 and 1350 (Hassan, 1981). Parallel evidence has not been found in the Tripolitanian pre-desert, mainly because deposits suitable for investigation are rare and these phases may not be resolved at the studied sites. The interior of Libya is heterogeneous and environmentally complex, and climate changes occurring elsewhere in North Africa may not necessarily have manifested themselves there in quite the same way. Gilbertson and Hunt (1996) and Nicholson (1989, 1994) describe the regional climatology. The quantity, variability and reliability of precipitation are not well known. In general, annual precipitation averages below 100 mm north of Beni Ulid to less than 25 mm in the south. Thus, Wadi Umm elKharab and Wadi el-Amud can be anticipated to have an unreliable and variable precipitation regime, averaging about 30 mm a year. Nowadays, drought may occur in many consecutive years. A year’s rain may fall in just one or two very intense and localized rainstorms, with adjacent areas remaining completely dry. In other arid regions of north Africa, it is known that ‘desert farming’ was not sustained by harvesting rainwater or floodwater. Rather, it was supported by a reliable underground water supply: perhaps a spring, as at Lemasba, Algeria (Shaw, 1982), or oases, as in the Fezzan (Mattingly, this volume Chapter 9; van der Veen, 1992). Springfed oases supporting ancient agriculture are known at Gheriat el-Gharbia in the study region. Geomorphology Understanding of the regional geomorphology is summarized in Anketell et al., 1995; Gale et al. (1993), Gilbertson et al. (1993), Gilbertson and Hunt (1988, 1996a) and Hunt et al. (1986). Plateau-basins near Beni Ulid contain a well-developed palaeosol indicative of former wetter conditions, presumably dating to the Early Holocene. Screes and alluvial fans may well have developed on several occasions during the Holocene. Overall, in the period during and after extensive farming, it is evident that there were several episodes of slope erosion, fluvial aggradation, incision and aeolian reworking. Anthropogenic deposits—large middens, layers of ash or dung—occur, notably at the olive farm at Wadi el-Amud (Gilbertson and Hunt, 1990).
The archaeology of drylands
146
Erosion, deposition and floodwater farming on wadi floors Of particular interest are changes to the intensity of run-off and patterns of erosion and deposition resulting from floodwater-based agriculture on wadi floors. During the passage of a storm pulse, the roots of modern olives and date palms bind the floodplain sediments, whilst intervening gullies can be scoured over 1 m deep, leaving the trees on eroded pinnacles. Elsewhere, modern barley is successfully grown on the level surface of recently deposited flood-loams. Cross-wadi walls promote fine-grained sedimentation and resultant increases in soil moisture, seed catch and shrub growth on their upstream side. Downstream from walls, waterfall-effects during flood promote gullies. Later, long after the walls are over-topped, subsequent subsurface flow may promote temporary springs, sapping and piping. These observations led to the development of a spatial model to explain agricultural practice on wadi floors (Fig. 8.5). The model also predicts where browse and shelter would have been available for stock—and thus an immediate source of the manure necessary to sustain intensive cereal cultivation (Barker et al., 1996a; Chatterton and Chatterton, 1984).
Figure 8.5 Model of Romano—Libyan agriculture Note: Zone A is a zone of deposition in quiet water upstream of a wadi wall, used for cereal cultivation; Zone B is the zone of turbulence downstream from a wadi wall, where tree crops were grown Source: After Gilbertson et al., 1984 The alluvial and biological materials on the wadi floors are mobile and frequently reworked by wind, rain, storm and, occasionally, by burrowing or grazing animals. Subsurface processes are less securely known. Field and laboratory evidence indicates that near-surface water is sometimes saline and it is not unreasonable to question whether soil salinization may have been locally important in the past, especially given the
Success, longevity and failure of arid-land agriculture
147
deliberate introduction of large quantities of water on to wadi floors. At present, there is no evidence for large-scale salinization of wadi floors in Romano—Libyan or more recent times (Gilbertson, 1996; Gilbertson et al., 1993).
WALLS AND WALL NETWORKS IN RELATION TO RUN-OFF AND FLOODWATER FARMING The spread of desert walls The immense numbers of walls are one of the most important signs of the ancient farming in the pre-desert (Fig. 8.6). The vital role of walls in facilitating ancient farming in drylands by trapping water and sediment has been recognized by numerous
Figure 8.6 Walls in the desert: wall systems in the Wadi Mimoun near Gasr Lebr (Mm10) Source: After Hunt et al., 1987
The archaeology of drylands
148
archaeologists and earth scientists (see Gilbertson, 1986; Pacey with Cullis, 1986). Brushwood fences were used, in addition to stone walls, to divert floodwaters at times of rain-storm in the American Southwest (Bryan, 1929; Nabhan, 1986a, 1986b; Nabhan and Sheridan, 1977; and see Minnis, Chapter 15). However, fences for water control have not been detected in the pre-desert, although thorn scrub is still in widespread use to corral animals. The antiquity of walls Inevitably, the age relationships of walls are often unknown. They are inferred from nearby archaeological features, whose antiquity was typically determined from associated ceramics—a necessary first assumption, but one that may often be incorrect. It is also clear that walls are likely to have been frequently rebuilt, reused, repaired, repositioned or reformed. In some areas, most walls were perhaps associated with the second major phase of Romano-Libyan settlement characterized by the construction of gsur. Elsewhere, many walls relate to the open farms of the earlier Romano—Libyan settlement phase. Some may even be older and associated with the modest numbers of later prehistoric settlements; other walls may post-date the Romano—Libyan period, as demonstrated by Hunt et al. (1986) in the case of the Wadi Mansur. Design principles In nearly every case studied, the position of a plateau or wadi-side wall was apparently intended to maximize the quantity—and perhaps the rate—at which run-off was delivered to the wadi floor. Often, water was led from the plateau or hill slopes into cisterns, many with sediment traps, or into caves at the wadi edge. On the wadi floor, the primary objective was apparently to capture floodwater, causing it to sink into the long-term storage provided by the wadi-floor alluvium. Occasionally, water was conducted into cisterns adjacent to ancient settlements. Many cisterns remain in use today, or at least they still function. Erosion appears to have been understood and managed by the Romano-Libyan farmers. Numerous wadi-floor walls contain ‘drop structures’: reinforced gaps through a boulder-built wall often leading onto a stone-reinforced area immediately downstream. They appear to be devices to avoid walls being overwhelmed and breached during flood peaks: the reinforced surfaces downstream prevent scour and gully erosion. These features still appear to be operating effectively, with few displaying evidence of damage. A substantial literature describes the role of wall-managed floodwater for contemporary dryland management and development—notably improving subsistence farming, or as a means of soil reclamation (for example Evenari et al., 1971, 1982; Pacey with Cullis, 1986; Reij, 1991; van der Wal and Zaal, 1990 and references therein). Even though the wall systems of the Libyan pre-desert were originally constructed two millennia ago, the robustness of the technology is evident, since they continue to harvest and channel run-off and storm water, with marked ecological and biogeographic consequences (Gilbertson et al., 1994).
Success, longevity and failure of arid-land agriculture
149
Walls and risk management Floodwater farming in Romano-Libyan times had to cover large areas and, to some extent, to be opportunistic: it had to cope with the patchiness and unreliability of desert rainfall. Substantial permanent investment of time and human energy in the construction and maintenance of walls would have been necessary. People had to sow seed on wetted soil and may have travelled to wadis that had received run-off from localized rainfall events. Major confluences or positions down-wadi must have been more reliable places to grow cereals, especially in times of general drought. A balance had to be struck in such locations between the opportunity to use the more frequent and larger run-off events and the risks posed by sequences of floods, which would have eroded seed sown after earlier floods.
WALL FUNCTIONS Six hypotheses have been proposed that, singly or in combination, might explain the function of the wall systems observed in the pre-desert (Gilbertson et al., 1984). Walls may: • capture, store and redistribute surface water for human and animal consumption and irrigation; • control fluvial erosion, sediment entrainment, transport and deposition; • control the movements of animals, either acting as pens and enclosures for domesticated herds, or by excluding animals from cultivated areas, or by controlling wild animals during hunts; • delineate areas of different land use; • represent the by-products of stone clearance to ease cultivation; • define parcels of land owned or controlled by different individuals or groups. These functions are not necessarily exclusive. The same wall may have had one or more uses when first constructed and later may have acquired or lost other roles. Walls whose primary purpose was clearly to delineate ownership have not been found in the pre-desert. Detailed surveys of wall distributions in the Wadi Umm al-Kharab indicate that cross-wadi walls were grouped and associated with different communities at various points along the length of the wadi (Gilbertson and Chisholm, 1996). It is not known how widespread this practice was. Overall, most walls appear related to hydrological/geomorphic factors, whilst an absence of walls may reflect their deep burial (Gilbertson et al., 1984). It is quite possible that the delineation of land ownership, tenure or management—if it was marked on the ground—operated within the hydrological constraints of the wall systems. The archaeological consequence is that it is very difficult to distinguish factors such as past community ownership or social groupings from the present information on wall networks.
The archaeology of drylands
150
ENVIRONMENTAL INFLUENCES Within the long periods in which the wall systems were in use, many droughts and other environmental vicissitudes must have occurred. Indeed, the evidence described previously from the Nile basin and Morocco suggests that droughts of 10–50 years’ duration probably occurred several times during the last two millennia. Large parts of the pre-desert are likely to have been abandoned at times of prolonged drought, once buffer stocks and imported feed were exhausted. Nevertheless, no direct evidence for temporary abandonment yet exists—at present both the environmental and archaeological data are too coarse to distinguish such events. Similarly, there are no clear indications that the development and major shifts in the form and distribution of Romano—Libyan settlement or its subsequent decline were associated with some form of climatic change or fluctuation. Nevertheless, the existing palaeoecological and palaeoclimatic information from the study area suggests that the climate during much of the period of Romano— Libyan settlement was not dissimilar to that which prevails today, though vegetation was generally less degraded. At present, the essential robustness and the long-term duration of floodwater farming in the pre-desert, as well as the available palaeoecological evidence and modern ecological theory, provide no support for many possible explanations of region-wide changes in settlement or movement out of the Libyan pre-desert (Gilbertson, 1996; see above). In brief, the widely argued litany of anthropogenic agencies of desertification does not seem to have played a central role in transforming the widely farmed and settled Romano-Libyan pre-desert into the modern arid wilderness. The possible significance of disease and synergistic or feedback effects, though, remains completely unknown.
HUMAN AGENCIES As a result of the analysis described previously, broad-scale interpretations of the ancient settlement and farming in the pre-desert must focus upon human agencies of change: the outcomes of developments in the economic, military, political, psychological and social worlds (Barker et al., 1996a; Mattingly, 1996). In brief, the prime factor encouraging the Macae tribal pastoralists to become sedentary floodwater farmers appears to have been regional stabilization resulting from the expansion of Roman influence into Tripolitania. Effective incorporation into the wider imperial economy produced different patterns of land use, greater stability, access and trade with the vast new market, and a major increase in population. There are no grounds for believing that a widespread military colonization by soldier-farmers (limitanei), or a frontier army, ever played a significant role, as was once suspected (Goodchild and Ward-Perkins, 1949). Neither is there any indication that the expertise, ideas or technology of floodwater farming were introduced from outside. Probably, the desert dwellers developed these approaches indigenously. As demonstrated in many chapters in this volume, the essence of this technology was repeatedly invented in antiquity in very different places.
Success, longevity and failure of arid-land agriculture
151
It remains unproven whether the replacement of open, undefended farms by gsur should be related to a greater sense of insecurity, or whether the rich and powerful of the time adopted these imposing enclosed structures to follow fashion, as a display of prestige, or because the shade, size and airiness of such buildings were well adapted to the rigours of desert life (Fig. 8.4). The progressive abandonment of settlement and farms in the southern part of the pre-desert region is perhaps best attributed to the wider political and economic changes throughout the Mediterranean at the end of the Roman era, with the development of smaller, more regionalized, group identities (Mattingly, 1996). There are no grounds for suspecting that the arrival of the Arab armies in the AD 640s and later brought about major changes in pre-desert settlement or farming. Floodwater farming was to continue at smaller scales for another thousand years; indeed it continues to be practised today in the region.
THE DECLINE OF FLOODWATER FARMING It is clear that dramatic explanations of the abandonment of the floodwater farming systems as the result of climatic, economic or political change are not congruent with the history of the pre-desert as presently understood. It is also clear that ecological degradation of the landscape, for example at the Lm4 farm, post-dated the end of Romano—Libyan floodwater farming. Floodwater farming seems to have come to an end gradually, on a piecemeal basis in some areas, though there may have been rapid early retreats from the southernmost outposts such as Wadi el-Amud as these became uneconomic with the collapse of long-distance trade networks in late Romano—Libyan times. It is clear that partial use of systems such as at Mm10 and Lm4 continued after formal use of the Romano—Libyan buildings ended. People continued to grow cereal crops and keep stalled animals, though they often no longer lived in the Romano— Libyan buildings. At this stage the landscape still had a distinctly steppic aspect. The population of the pre-desert was never very large. For maximum efficiency, labour-intensive maintenance of the wall networks is essential. One might envisage that, as Roman influence waned and the political landscape became unstable, intensive investment in farming complexes became a risky strategy. People began to readopt ‘bedouin’ ways of life, which are flexible and in many ways less arduous than living in fixed settlements in this region. As people abandoned buildings for tents, a transhumant lifestyle became possible and people started to move to where rain had fallen most intensively each year to grow their crops. Because of increased mobility, it was no longer important that walls were rigorously maintained. Systems would be abandoned as they became inefficient. The end of animal stalling, as seen at Lm4, would have placed additional stress on the landscape because grazing removed the steppe vegetation and led to the modern pre-desert ecology. Rapid alluviation events in the medieval or early postmedieval periods may perhaps have been linked with landscape degradation of this type (Barker with Gilbertson, 1996; Gilbertson and Hunt, 1996a).
The archaeology of drylands
152
CONCLUSION The systems of floodwater farming developed by Romano-Libyan farmers in the Tripolitanian pre-desert seem to have been ‘sustainable’ in the true sense of the word. They have persisted in some localities for two millennia, surviving the fall of empires, major economic catastrophes, climate fluctuations and changes, and other adversities. Much of the resilience of the floodwater farming systems is clearly the result of the exploitation of detailed (if informal) understanding of run-off and fluvial processes and local geomorphological conditions by the local population from the Macae tribespeople onwards, together with their engineering skills and their capacity to take advantage of patchy and unreliable storms. Their farming systems seem to have been well adjusted to local conditions. Details are, however, still sparse. The hypothesis of a geomorpbiologically-adjusted polyculture, with tree crops in erosive areas and grain crops under-planted with medic pasture in depositional areas, is plausible but unproven. The possibility that Romano—Libyan farmers stalled their stock is significant, because animals kept this way would be less able to de-vegetate and thus degrade the landscape. The end of floodwater farming seems to have been piecemeal and gradual, and not linked to most of the cited ‘push-factors’ such as the fall of Rome, which were relatively rapid. It may be that the bedouin lifestyle simply became more attractive to the small population of the Tripolitanian pre-desert.
REFERENCES Anketell, M.J., Ghellali, S.M., Gilbertson, D.D. and Hunt, C.O. (1995) Quaternary floodplain and wadi floor infill deposits in northeastern Libya and their implications for landscape development. In J.Lewin, M.Macklin and J.M.Woodward (eds) Quaternary Mediterranean River Environments: 231–44. Amsterdam, A.A. Balkema. Barker, G., with Gilbertson, D.D. (1996) Farming the desert: retrospect and prospect. In G.Barker, D.D.Gilbertson, G.D.B.Jones and D.J.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis: 343–63. Paris, UNESCO Publishing. Barker, G. and Gilbertson, D.D., with Hunt, C.O. and Mattingly, D.J. (1996) Romano— Libyan agriculture: integrated models. In G.Barker, D.D.Gilbertson, G.D.B.Jones and D.J.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis: 265–90. Paris, UNESCO Publishing. Barker G., Gilbertson D.D., Jones, G.D.B. and Welsby, D.A. (1991) The UNESCO Libyan Valleys Survey XXIII: the 1989 season. Libyan Studies 22:31–60. Barker, G., Gilbertson, D.D., Jones, G.D.B. and Mattingly, D.J. (1996a) Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis. Paris, UNESCO Publishing. Barker, G., Gilbertson, D.D., Jones, G.D.B. and Mattingly, D.J. (1996b) Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 2. Gazetteer and
Success, longevity and failure of arid-land agriculture
153
Pottery. Paris, UNESCO Publishing. Barth, H. (1857) Travels and Discoveries in North and Central Africa. London, Longman. Bryan, R.K. (1929) Floodwater farming. Geographical Review 19 (3):444–56. Chatterton, B.A. and Chatterton, L. (1984) Medicago—its possible role in Romano— Libyan dry farming and its positive role in modern dry farming. Libyan Studies 15:157–60. Dorsett, J.E., Gilbertson, D.D., Hunt, C.O. and Barker, G. (1984) The UNESCO Libyan Valleys Survey IX: image analysis of Landsat data and its application to environmental and archaeological Surveys . Libyan Studies 15:71–80. Evenari, M., Shanan, L. and Tadmor, N. (1971) The Negev: The Challenge of a Desert. Cambridge, Mass., Harvard University Press. Evenari, M., Shanan, L. and Tadmor, N. (1982) The Negev: The Challenge of a Desert. Cambridge, Mass., Harvard University Press, second edition. Flower, C.P. and Mattingly, D.J. (1995) ULVS XXVII: mapping and spatial analysis of the Libyan Valleys Data using GIS. Libyan Studies 26:49–78. Gale, S.J., Gilbertson, D.D., Hoare, P.G., Hunt, C.O., Jenkinson, R.D.S., Lamble, A.P., O’Toole, C., van der Veen, M. and Yates, G. (1993) Late Holocene environmental change in the Libyan pre-desert. Journal of Arid Environments 24:1–15. Gilbertson, D.D. (1986) (ed.) Run-off Farming in Rural Arid Lands. Theme Volume 6 (1 and 2) of Applied Geography. Gilbertson, D.D. (1996) Explanations: environment as agency. In G.Barker, D.D. Gilbertson , G.D.B.Jones and D.J.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis: 291–318. Paris, UNESCO Publishing. Gilbertson, D.D. and Chisholm, N.T. (1996) Manipulating the desert environment: ancient walls, floodwater farming and territoriality in the Tripolitanian pre-desert of Libya. Libyan Studies 27:17–52. Gilbertson, D.D. and Hunt, C.O. (1988) The UNESCO Libyan Valleys Survey XIX: the Cenozoic geomorphology of the Wadi Merdum, Beni Ulid, in the Libyan pre-desert. Libyan Studies 19:95–121. Gilbertson, D.D. and Hunt, C.O. (1990) The UNESCO Libyan Valleys Survey XXI: geomorphological studies of the Romano-Libyan Farm, its floodwater control structures and weathered building stone at site Lm4, at the confluence of the Wadi el Amud and the Wadi Umm el Bagul in the Libyan pre-desert. Libyan Studies 21: 25–42. Gilbertson, D.D. and Hunt, C.O. (1996a) Quaternary geomorphology and palaeoecology. In G.Barker, D.D.Gilbertson, G.D.B.Jones and D.J.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis: 49–82. Paris, UNESCO Publishing. Gilbertson, D.D. and Hunt, C.O. (1996b) Romano—Libyan agriculture: walls and floodwater farming. In G.Barker, D.D.Gilbertson, G.D.B.Jones and D.J. Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis: 191–216. Paris, UNESCO Publishing. Gilbertson, D.D., Hayes, P.P., Hunt, C.O. and Barker, G. (1984) The UNESCO Libyan Valleys Survey VII: a classification and functional analysis of ancient irrigation and
The archaeology of drylands
154
wall systems in the Libyan pre-desert. Libyan Studies 15:45–70. Gilbertson, D.D., Hunt, C.O. and Fieller, N.R.J. (1993) ULVS XXVI: sedimento-logical and palynological studies of Holocene environmental changes from a plateau basin infill sequence at Grerat D’nar Salem, near Beni Ulid, in the Tripolitanian pre-desert. Libyan Studies 24:1–19. Gilbertson, D.D., Hunt, C.O., Fieller, N.R.J. and Barker, G. (1994) The environmental consequences and context of ancient floodwater farming in the Tripolitanian predesert. In A.C.Millington and K.E.Pye (eds) Environmental Change and Geomorphic Processes in Arid Lands: 229–51. Chichester, John Wiley and Sons. Goodchild R.G. and Ward-Perkins J.B. (1949) The Limes Tripolitanus in the light of recent discoveries. Journal of Roman Studies 39:81–95. Hassan, F.A. (1981) Historical floods and their implications for climatic change. Science 212:1142–5. Hassan, F.A. (1996) Abrupt Holocene climatic events in Africa. In G.Peti and R. Soper (eds) Aspects of African Archaeology: 83–9. Harare, University of Zimbabwe Publications. Hassan, F.A. (1997) Holocene palaeoclimates of Africa. African Archaeological Review 14 (4):213–30. Hassan, F.A. (1998) The archaeology of North Africa at Kiekrz 1997. African Archaeology Review 15 (1):85–93. Hunt, C.O., Mattingly, D.J., Gilbertson, D.D., Barker, G., Dore, J.N., Burns, J.R., Fleming, A.M. and van der Veen, M. (1986) The UNESCO Libyan Valleys Survey XIII: interdisciplinary approaches to ancient farming in the Wadi Mansur, Tripolitania. Libyan Studies 17:7–47. Hunt, C.O., Gilbertson, D.D., van de Veen, M., Jenkinson, R.D.S., Yates, G. and Buckland, P.C. (1987) The UNESCO Libyan Valleys Survey XVII: the palaeoecology and agriculture of the abandonment Phase at Gasr Mm10, Wadi Mimoun in the Tripolitanian pre-desert. Libyan Studies 18:1–14. Lamb, H.F., Duigan, C.A., Gee, J.H.R., Keits, K., Lister, G., Maxted, R.W., Merzouk, A., Niessen, F., Tahri, M., Whittington, R.J. and Zeroual, A. (1994) Lacustrine sedimentation in a high altitude semi-arid environment: the palaeo-limnological record of Lake Isli, High Atlas, Morocco. In A.C.Millington and K.E.Pye (eds) Environmental Change and Geomorphic Processes in Arid Lands: 229–51. Chichester, John Wiley and Sons. Lamb, H.H., Gasse, F., Benkaddour, A., El Hamouti, N., van der Kaar, S., Perkins, W.T., Pearce, N.J. and Roberts, C.N. (1995) Relations between century-scale Holocene arid intervals in tropical and temperate zones. Nature 373:134–7. Mattingly, D.J. (1989) Farmers and frontiers: exploiting and defending the countryside of Roman Tripolitania. Libyan Studies 20:135–53. Mattingly, D.J. (1996) Explanation: people as agency. In G.Barker, D.D.Gilbertson, G.D.B.Jones and D.J.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis: 319–42. Paris, UNESCO Publishing. Mattingly, D.J. with Flower, C. (1996) Romano—Libyan settlement: site distribution and trends. In G.Barker, D.D.Gilbertson, G.D.B.Jones and D.J.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis:
Success, longevity and failure of arid-land agriculture
155
159–90. Paris, UNESCO Publishing. Nabhan, G.P. (1986a) Papago Indian desert agriculture and water control in the Sonoran Desert, 1697–1934. Applied Geography 66 (1):42–3. Nabhan, G.P. (1986b) ‘Ak-ciñ ‘arroyo-mouth’ and the environmental setting of Papago Indian fields in the Sonoran Desert. Applied Geography 6 (1):61–75. Nabhan, G.P. and Sheridan, T.E. (1977) Living fencerows of the Rio San Miguel, Sonora, Mexico; traditional technology of floodplain management. Human Ecology 5:97–111. Nicholson, S.E. (1989) Long term changes in African rainfall. Weather 44:46–56. Nicholson, S.E. (1994) Rainfall fluctuations in Africa and their relationship to past conditions over the continent. The Holocene 4(2):121–31. Pacey, A. with Cullis, A. (1986) Rainwater Harvesting: The Collection of Rainfall and Run-off in Rural Areas. London, Intermediate Technology Productions. Reij, C. (1991) Indigenous Soil and Water Conservation in Africa. London, International Institute for the Environment and Development, Gatekeeper Series no. SA27. Shaw, B.D. (1982) Lamasba: an ancient irrigation community. Antiauités Africaines 18: 61–103. van der Veen, M. (1992) Garamantian agriculture: the plant remains from Zinchecra, Fezzan. Libyan Studies 23:7–39. van der Veen, M., Grant, A. and Barker, G. (1996) Romano—Libyan agriculture: crops and animals. In G.Barker, D.D.Gilbertson, G.D.B.Jones and D.J.Mattingly, Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis: 227–64. Paris, UNESCO Publishing, van der Wal, A. and Zaal, F. (1990) Bibliography on Indigenous Soil and Water Conservation with Special Reference to Africa. Amsterdam, Vreije Universiteit, Center for Development Cooperation Services.
9 Twelve thousand years of human adaptation in Fezzan (Libyan Sahara) DAVID MATTINGLY
INTRODUCTION With a few notable exceptions (Bousquet, 1996; Nesson et al., 1973; Trousset, 1986), the archaeology and long-term history of the Saharan oases remain poorly documented. In many cases, pioneer studies have not been followed up in recent decades (Ball and Beadnell, 1903; Fakhry, 1974; RSGI, 1937; Scarin, 1934, 1937). Yet, there is undoubtedly much to learn from the manner in which desert people have exploited resources and mastered the limitations of their environment. A better understanding of human adaptation to the desert environment has clear relevance for modern concerns about the sustainability of oasis farming. To illustrate this theme, this chapter will focus on the Fezzan Project, which I direct and which has completed four seasons of work (1997–2000). The project is investigating the archaeology of a region of the Libyan Sahara c.1,000 km south of Tripoli (Fig. 9.1) and follows on from earlier British work carried out by the late Charles Daniels. His exploration and excavations from 1958 to 1977 accumulated a vast dossier of information on one of the most important Saharan peoples of classical antiquity—the Garamantes (Daniels, 1969, 1970, 1971, 1989; cf. also Pace et al., 1951). The full publication of his work is being undertaken in parallel with the renewed work (Edwards et al., 1999). The Garamantes were the dominant power in the Libyan Sahara from c.500 BC to c.AD 500, and at the height of their influence they controlled a vast desert territory of c.250,000 km2, at times threatening both the Romanized cities of the Mediterranean coast and the sub-Saharan populations of Chad and Niger: Liverani (1999), for example, describes Garamantian forts on the routes south of Ghat, itself 300 km southwest of the Garamantian capital. They were several times defeated by Roman armies sent against them, but their territory was never annexed to the Roman empire and for much of the Roman period they seem to have thrived on a combination of oasis agriculture and trade (Mattingly, 1995:33–7, 68–77 on relations with Rome).
Twelve thousand years of human adaptation in Fezzan
157
Figure 9.1 Map showing the location of the Fezzan and the area of most detailed survey around Germa The renewed fieldwork has aimed to amplify this picture by setting the Garamantes in a longer-term framework of human lifeways in the region—broadly focusing on the Holocene, but with backward glances at the very different Pleistocene environment (Mattingly et al., 1997, 1998a, 1998b, 1999a, 1999b). At the heart of the project is a concern with human interaction with the environment and a study of how this has varied over an extended period of time and changing conditions. The Fezzan Project, then, has relevance to wider debates than parochial Libyan ones, though the evolution of Libyan culture and society in the desert is of major importance in its own right (Bates, 1914; Brett and Fentress, 1996; Camps, 1980). The transition to farming, the emergence of social complexity and the formation of a distinctive Saharan culture were all achieved in a region undergoing massive climatic degradation and desiccation.
DESERT LANDSCAPES The Libyan Desert is an area almost the size of India (Bagnall, 1935: map 1), but with a tiny population—Libya itself has under five million people, most concentrated in the coastal cities. The desert is an extraordinarily difficult environment to live well in, and, as in most desert regions, water is a critically scarce resource in large parts of the country. Most of Fezzan has negligible annual rainfall and today depends entirely on subterranean fossil water sources for sustaining its human population, its livestock and its areas of cultivation. In this respect, the region has very different characteristics to the Libyan predesert zone between Fezzan and the Mediterranean coast, studied in an earlier project
The archaeology of drylands
158
(Barker et al., 1996; Gilbertson et al., Chapter 8), where careful harvesting of the limited rainfall proved the key to past exploitation. The story of farming the desert in Fezzan is one of people evolving strategies to utilize more effectively the huge groundwater resources at the limited number of locations where they are relatively accessible at or close to the surface. Over the last few millennia, the groundwater levels appear to have fallen significantly, leading to small lakes drying up, spring-lines ceasing, irrigation systems being abandoned, wells being deepened, and so on. As described below, the strategies evolved to tap into the groundwater were at certain times highly sophisticated and in all probability highly organized within the society. The peak population level (before the modern era) seems to have been reached in the Garamantian period, when the material prosperity of the region also reached its apogee. The landscapes of Fezzan are very variable—a mixture of great sand seas, gravel and boulder-strewn wastelands, and hyper-arid rock plateaux (Fig. 9.1). The project focuses on a long depression aligned east-west called the Wadi el-Agial (also known as al-Hayat), though it is not in fact a true dry river (the normal meaning of wadi). The el-Agial depression contains a chain of small oases over a length of about 150 km, drawing on a series of aquifers. The traditional pattern of cultivation involves scattered palm groves, with intensively irrigated plots of wheat, barley and sorghum. Duveyrier (1864:147–216, 439) and Lyon (1821:270–78) both give good accounts of Fezzanese plants and cultivation systems in the nineteenth century. There are no perennial springs here today, though there are hints in the landscape that at some point in the past there was an active spring-line along the south side of the valley. Since the invention of diesel and electrical pumps, extraction of water from the aquifers has accelerated greatly and has caused water levels in many wells to fall by up to 100 m in the last century. As we shall see, there are important lessons for the present to be learnt from the past history of human activity and over-exploitation of this resource. A vast sand sea (erg/edeyen) rises on the northern side of the oases. Although the scale of the dunes is forbidding and the crossing of them can be perilous (Denham and Clapperton, 1826:177–85), water was once more abundant within the sands and even today there is a number of small relict lakes, which sustain small stands of date palms. In the neolithic and classical periods there were undoubtedly more of these lakes, facilitating travel across, and life within, the sands. The south side of the Wadi el-Agial is dominated by a sheer, cliff-like escarpment, behind which extends a great sandstone plateau (hamada), turned black by desert varnish and dissected by deep gorge-like wadis running off to the south and southeast. Some of the strata in this formation are fine-grained silicified sand- and mud-stones, which were exploited extensively as a source for stone tools in the palaeolithic period and, to a lesser extent, in the mesolithic and neolithic periods. There are no perennial water sources on the plateau itself, the main aquifer lies deep beneath it, but this is the one part of the Fezzan to receive rain with any regularity; at certain times of year, pools of water can be found in the wadi beds, along with some rough grazing. Engraved rock art, dating to both the later prehistoric and historic periods, is abundant in many of the hamada wadis, representing seasonal exploitation of this forbidding landscape by hunters and mobile pastoralists (Lutz and Lutz, 1995). Even at the peak periods of oasis cultivation, it is clear that pastoral groups operated alongside cultivators in exploiting the potential of the desert
Twelve thousand years of human adaptation in Fezzan
159
landscapes (Nicolaisen and Nicolaisen, 1997; UNESCO Nomades, 1962). The consensus scientific view is that the Saharan region in the middle palaeolithic period (say 250,000–40,000 BP) was very much wetter than today, with an abundance of vegetation and wildlife flourishing around vast inland lakes (Petit-Maire, 1982; PetitMaire et al., 1980; Ziegert, 1995). Thereafter, in the late Pleistocene and early Holocene, conditions became subject to a series of dramatic swings from wetter to drier conditions and back again (Fig. 9.2). The neolithic period (c.6000–1000 BC) was generally one of worsening conditions (Lutz and Lutz, 1995; Petit-Maire, 1988; Shaw, 1976), and by about 3000 BC it is likely that the Saharan climate was much as today (Cremaschi, 1998). However, subterranean water sources, based on the huge
Figure 9.2 The major climatic fluctuations of the Holocene in the Libyan Sahara Note: The lower part of the diagram shows possible phases of Saharan rock art Source: After Lutz and Lutz, 1995 Continental Intercalate aquifer system (Edmunds and Wright, 1979; Zaluski and Sadek, 1980), may have been more abundant and more readily accessible at that time (more springs, small lakes and shallow aquifers), with wild fauna more diverse as a result. Neolithic rock engravings show that, at the start of this period, the Sahara supported a large and rich wild fauna, including species like the crocodile, which require permanent water, but that this was crucially changed with increasing aridity, leading to extinction of many species north of the Sahara and to major changes in the lifestyle of the surviving human groups (Barker, 1989; Encyclopédie, 1997; Le Quellec, 1987; Lutz and Lutz, 1995; Mori, 1969, 1988). In the rock art we see evidence of the domestication of animals and an increasing emphasis on fertility and ritual—perhaps reflecting the social stress caused by environmental change (Encyclopédie, 1997:2791–96, 2800–2; Lutz and Lutz, 1995:145–65, 169–75). The same trends towards domestication of animals are evident also in the few well-excavated neolithic rock shelters (Barich, 1987; Cremaschi and di Lernia, 1998).
The archaeology of drylands
160
AIMS OF THE FEZZAN PROJECT The project addresses a number of key questions: • the transition to farming in the Saharan region and in particular the origins of agriculture; • the diffusion or invention of farming/hydraulic technology and the spread of different cultivated plants; • the response of human populations to the climatic and environmental changes; • the origins of urbanization in the Sahara and its evolution over time; • the construction of identity through material culture; • inter-regional contact across the Sahara (trade); • processes of desiccation (and desertification?) in the northern Sahara; • the recognition of palaeo-hydrological features in the landscape (spring-lines, lakes, marsh); • the dating of changes in the hydrology; • the identification and dating of evidence of climatic and environmental change. The project geomorphologists have been studying the regional hydrology, using field data and remote sensing techniques to map gypsum formations, which are indicative of ancient springs and small dried-up lakes, sampling palaeo-lake sediments, crosssectioning spring mounds and assessing dune morphology (Mattingly et al., 1998b:117– 22, 1999b:129–31). A series of possible prehistoric lake sediments has been identified at various points in the landscape, and dating of these is a priority of continuing work, to establish whether they are of Pleistocene or Holocene date. Laboratory techniques being used in support of the programme of palaeoenvironmental reconstruction include stable isotope analysis, particle size distributions, and mineral magnetic analysis of the putative lake deposits to characterize them, combined with uranium-thorium and optically stimulated luminescence techniques of dating. The uranium-thorium dating is being used on both Melanoides tubercolota shells, found in association with dark organic lake-edge deposits, and on gypsum crystals from a line of defunct springs at the foot of the escarpment. With these methodologies we are hoping to be able to track through welldated contexts the shrinkage and disappearance of the lakes with the onset of desiccation, perhaps accompanied by the drying up of the palaeo-spring-line at the foot of the escarpment. The primary archaeological component of the project is the excavation of a site within the major ancient urban centre of the region, Old Germa (or Garama as it was known in antiquity) (Fig. 9.3). This is a still-standing medieval caravan town, controlling one of the larger and more fertile oases of the el-Agial and situated on a trans-Saharan trade route. There is a complex stratigraphy of a sequence of earlier cities superimposed one on another to a depth of 4–5 m. Some earlier clearance excavation (Ayoub, 1967a) has revealed a group of Garamantian buildings at the core of the site. Unlike most of the later structures, these have stone walls and reflect the power and wealth of the site in its heyday in the period between the first and fourth centuries AD. The origins of the
Twelve thousand years of human adaptation in Fezzan
161
settlement go back until at least the fifth century BC, again with a sequence of mud-brick buildings.
Figure 9.3 The settlement of Germa (ancient Garama), the capital of the Garamantes Photograph: D.Mattingly The current excavations here are designed to refine knowledge of this long urban sequence, producing a series of time-slices illustrating the entire history of this remarkable site. The material culture revealed demonstrates clear change over time— some phases are unmistakably more impoverished than others. For instance, much of the medieval and early modern periods is characterized by relatively low numbers of imported goods, despite the existence of trans-Saharan trade at this time. In the Garamantian period, by contrast, an abundance of wine and olive oil amphorae, ceramic finewares and glass ware was imported from the Roman world (Fontana, 1995). Systematic sieving of deposits is recovering a sample of the plant fragments and animal bones present in each phase. The good preservation of many plant fragments in desiccated form is significant, because it means that a reasonably full range of cultivated and weed species is represented in the samples. This gives useful information about the local environment: the cultivated plants all require irrigation, and the weeds often reflect the arid and salty background conditions. We are also identifying a series of significant botanical horizons, including a ‘maize horizon’, representing the coming of New World crops, and a ‘sorghum horizon’, representing northward transference from the Sahel or Sudan, probably within the Garamantian period. This study builds on earlier work by van der Veen (1992) and is extending our knowledge of changing patterns of plant cultivation back from the present to c.900 BC (for the broader North African context, see van der Veen, 1995, 1999). Analysis of the faunal remains is also indicating change over time,
The archaeology of drylands
162
with the bones providing important information about not only what stock was raised but also the age at which animals were slaughtered and the butchery techniques used. In order to provide a wider context for the picture of life in the town, the excavation is complemented by fieldwalking and by more extensive survey in the Germa region. This aims to build upon Daniels’ earlier survey, which was successful in locating a large number of cemeteries of Garamantian date, particularly in the form of cairn graves along the foot of the hamada escarpment. He was less successful in locating Garamantian settlements, though he was aware of a few village-like sites in the oasis and a number of hillforts along the escarpment. Systematic fieldwalking has now revealed that the Garamantian settlement pattern was far denser than previously suspected, with numerous satellite villages all around Old Germa (Fig. 9.1). The fieldwalking essentially logs the density of archaeological material (humanly-made or imported goods—lithics, ceramics, ostrich shell beads and so on) and isolates significant concentrations of such material as ‘sites’. Topographic survey of a selection of these sites has added structural detail and confirms that we are dealing with settlements and not simply rubbish disposal. The survey complements the evidence of a series of excavations by Daniels on additional Garamantian sites, which we are also preparing for publication; Zinchecra, an early Garamantian hillfort and cemetery (Daniels, 1968), Saniat Ben Howedi, a rich Roman period cemetery (Ayoub, 1968; Daniels, 1989), and Saniat Gebril, an oasis village. Our fieldwork has discovered sites of many different phases of activity, not simply the Garamantian phase. On the hamada to the south of Germa, we have recorded a series of important lithic scatters of palaeolithic date, comprising tools such as 100,000-year old handaxes together with chunky waste flakes and chippings at the locations where tools were produced. In our 1999 season, a series of neolithic occupation sites was also identified close to the edge of the sand sea to the north of Germa. These sites, yielding extremely finely worked lithics, early pottery, grindstones, ostrich eggshell fragments and beads, were probably occupied in the last few millennia BC when climatic conditions were rapidly deteriorating. Their inhabitants seem to have exploited a shallow and now vanished lake site. A gazetteer of ancient sites throughout the el-Agial is being compiled, combining both the Daniels’ material and the new work. Transcription of a series of air photographs taken in the 1950s and 1960s is revealing a wealth of information now destroyed by modern development. This work complements a programme of remote sensing using modern satellite imagery; comparison of the satellite imagery and the air photographs has revealed the extent to which deep-bore artesian wells have expanded the area under cultivation in the last twenty years, but at the cost of dramatically lowering the regional water table. The extension of the cultivated area, and the growth of modern villages that has accompanied it, have particularly affected the preservation of one of the most important and enigmatic classes of monument—the foggaras. These are underground irrigation canals, similar to the Persian qanat or the Arabian falaj or aflaj, which tapped into an aquifer below the foot of the escarpment and led flowing water out into the oasis proper (Bousquet, 1996; Goblot, 1979; Klitzsch and Baird, 1969; Lô, 1953, 1954; Mattingly et al., 1998a: 190–2; 1998b:137–42; 1999b:139–42; Nesson et al., 1973). They are readily identifiable at the surface, where traces survive, from the regularly spaced vertical shafts
Twelve thousand years of human adaptation in Fezzan
163
that were dug to facilitate construction and maintenance of the channels, though they must have added hugely to the labour involved. The shafts can be up to 20 m deep, gradually diminishing in depth until the channel emerges at the surface (Fig. 9.4). The available dating evidence indicates that the foggara system was introduced to Fezzan during the Garamantian period, with their use probably extending into the early Islamic period. It is clear that these structures were a key to ancient irrigation in the region, though evidently they have been dry for many centuries now. There are many hundreds of these structures visible on the air photographs, most being at least several km in length. The labour involved in their construction and maintenance was on a significant scale (Mattingly et al., 1999b:140–1).
HISTORICAL RECONSTRUCTION What does all this new evidence add up to? It is clear that there has been dramatic change in environment, climate and human activity over time. What
Figure 9.4 Schematic cross-section of a foggara, tapping into water-bearing strata below the escarpment and leading flowing water along a tunnel to the oasis zone in the valley floor follows is a very simplified and provisional analysis, with suitable disclaimers attached. The reconstruction proposed at this stage is essentially a series of models, designed for further testing and elaboration. The clear trend running through, though, is one of an overall decrease in water availability over time. Climatic change and the onset of desertification have reduced rainfall to negligible levels and caused old surface water sources, such as lakes and springs, largely to dry up. During the Upper Pleistocene, the region is known to have been very different from the desert environment it has become. The hamada plateau is assumed to have been wellvegetated savanna, with abundant rainfall supporting a large range of animals and huntergatherer human groups—whose tool assemblages occur in profusion across its surface. It is generally agreed that last phase of the Pleistocene—the period c.40,000–10,000 BP— was one of high aridity in North Africa, reflected in the Fezzan in a dearth of evidence for
The archaeology of drylands
164
the characteristic lithic assemblages of this phase. The reappearance of substantial human populations in the Early Holocene, after 10,000 BP, can be related to a new period of increased rainfall (Fig. 9.5). The landscape was well vegetated in this phase, supporting a wide range of wild animals, which were initially exploited through hunting, especially on the plateau and wadis of the hamada. However, in successive phases of further climatic change—whether major oscillations as indicated on Figure 9.2, or a more step-like progression towards acute aridification—human settlement became increasingly focused on locations where water was to be found at shallow depth. Thus many sites, presumably seasonal camp-sites, have been located in the el-Agial depression and around small lakes on the edge of the sand sea. Because the mesolithic and neolithic phases were far from a uniform period climatically, it is necessary to undertake more work on the phasing of sites of these periods, through further analysis of tool types, rock art phases and the
Figure 9.5 Model of the neolithic landscape around Germa, with settlement and activity (stock raising and later cultivation) based around perennial water sources (lakes and springs) in the valley and the edge of the sand sea
Twelve thousand years of human adaptation in Fezzan
165
evolution of pottery use. In this respect, the research by Cremaschi and di Lernia (1998) marks an important advance. For the moment, our model of later prehistoric settlement (Fig. 9.5) includes sites from both wetter and drier phases of the Early Holocene. The huge climatic fluctuations of this period form a backdrop to the transition to farming here. The domestication of animals can be traced both in the rock art (which can be dated only in relative terms at present) and from some of the excavated rock shelters. The exploitation of plant resources is most clearly signalled by the abundant grindstones at the neolithic campsites by the lakes and water sources. What is particularly interesting about this transition in Libya is that it seems to arise as a response to adversity rather than to opportunity: people turned to stock raising and cultivation here during the fifth and fourth millennia BC when a dramatic change in the availability of water made a huntergatherer existence increasingly more precarious (Barker, 1989, 1996). The Fezzan Project will hopefully make an important contribution to these debates. It is likely, though has not yet been demonstrated, that neolithic farmers grew their crops in small patches of soil naturally irrigated by higher groundwater levels, in contrast with the floodwater farming systems developed on the northern margins of the Sahara fringe by Romano-Libyan farmers (Barker et al., 1996; see also Chapter 8). With pastoralism and small-scale cultivation established, there is then little evidence for significant change in subsistence through the third and second millennia BC. The period of the Garamantes, however (between 900 BC and AD 500), marked a dramatic development in farming technologies and systems, associated with transformations in cultural complexity. These transformations included: • the rise of a major polity and civilization in the Sahara (Daniels, 1970; Ruprechtsberger, 1997); • the development of urbanism (Daniels, 1971:262–5); • the evolution of a hierarchical and probably slave-using society (Daniels, 1970:27– 35); • the adoption of a written script for the Libyan language (Daniels, 1975); • the further development of agriculture to encompass a range of Mediterranean and desert crops that require intensive irrigation (cereals, grapes, olives, dates) (Daniels, 1989:56–58); • the introduction of the horse, the camel and wheeled transport to the Sahara (Camps, 1989); • the creation of trade and political relations that extended north to the Mediterranean, east to Egypt and south to sub-Saharan Africa (Bovill, 1968:1–44; Fontana, 1995); and • a massive demographic expansion to a level that was probably not equalled again until the last forty years—Daniels (1989:49) estimated that there were at least 120,000 Garamantian burials in the el-Agial alone. The Garamantes represent in part a continuation of the local neolithic tradition, as is clear from lithic and ceramic finds at their early settlements. But they probably comprised a great confederation of tribes, and there are indications that some elements may have migrated from oases further east, nearer Egypt, bringing with them knowledge of improved technology for oasis cultivation—notably the foggara. There are clear parallels,
The archaeology of drylands
166
for instance, between the Libyan tribesmen on Egyptian reliefs and in rock art of southern Libya and Algeria (Lutz and Lutz, 1995:140–1; Ruprechtsberger, 1997:66–9). Most of the early Garamantian settlements currently known are situated along the edge of the escarpment, many in defensible positions, such as the classic hillfort site of Zinchecra. Botanical remains from sites like Zinchecra dating to the first half of the first millennium BC demonstrate that irrigated cultivation had begun by that early date (Daniels, 1989:56– 8; van der Veen, 1992). The main phase of occupation at Zinchecra ended around 500 BC (van der Veen, 1992:12–13), at which point it appears that an urban site originated at Germa. Over time, Garama emerged as the Garamantian capital and in the Roman period was adorned with substantial public buildings and temples utilizing stone on a scale and with a quality of dressing not previously witnessed. Since there is no evidence to suggest a Roman occupation of Fezzan, these must be the result of contact, diplomacy and trade between the Roman empire and the Garamantian kingdom. Garamantian culture, nowhere better illustrated than in its extraordinary funerary architecture, was extremely eclectic—though the variety of tomb types in contemporary use may also reflect the maintenance of discrete tribal identities within the structure of the polity (Ayoub, 1967b, 1968; Daniels, 1971:265–8; el-Rashedy, 1988; Ruprechtsberger, 1997:51–65). The evolved settlement pattern (Fig. 9.6) reflects the increasing localization of farming activity in the oases along the base of the depression. In addition to the large urban centre at Garama, there were regularly-spaced village settlements all along the valley, to match the extensive evidence of cemeteries along the foot of the escarpment (tens of thousands of graves have been recorded, as noted above). Hundreds of foggaras facilitated the largescale and extensive cultivation of the valley-floor oasis area. A crucial question we are still seeking an answer to is why these systems were abandoned; perhaps it was because of falling water levels in the aquifer. The settlement density, the number and scale of the cemeteries, and the foggara systems, all combine to highlight the Garamantian period as one of peak population and oasis cultivation. Garamantian civilization was thus the result of raised population levels in the northern Sahara following the development of advanced irrigation systems. The concentration of tens of thousands of people in the largest of these oases allowed them to dominate a large expanse of the Sahara—raiding and trading in equal measure to all points of the compass. Classical sources speak of the Garamantes hunting the troglodytae and ‘Ethiopians’, which gives a strong hint of slave raiding against neighbouring peoples (Herodotus 4.183; cf. Tacitus Hist. 4.50). Quite apart from the possibility of selling-on such captives north across the Sahara, the intensive irrigated cultivation and the dangerous task of foggara construction could have absorbed large numbers of slaves. The evidence for the existence of trans-Saharan trade at this date is partial at best, but the large quantities of Roman trade goods found at Garamantian sites and in their burials indicate that something of value must have been passing the other way. Apart from slaves, it is possible that the Garamantes also traded in salt, gold, semi-precious stones and natron (the latter used in glass making) (Bovill, 1968). The funerary evidence indicates the emergence of a social hierarchy, with a prominent elite order enjoying significantly greater wealth than the majority of the population, who were still buried in relatively simple cairn graves.
Twelve thousand years of human adaptation in Fezzan
167
In the early Islamic period some at least of the Garamantian villages appear to have continued and may have been embellished with castle-like structures (gsur. see Ruprechtsberger, 1997:77–81, for examples) built of mud brick.
Figure 9.6 Model of the evolved Garamantian landscape around Germa, with its extensive irrigation systems, urban centre and satellite villages, and numerous cemeteries Over time, however, the number of villages seems to have declined markedly, perhaps linked to a shift from foggara to well irrigation (Fig. 9.7). The problem with irrigation based on wells is that water must be mechanically raised by bucket before being fed into irrigation canals, with the result that in general each well can irrigate only a limited area of fields around it. The late medieval and early modern pattern is thus of small clumps of palms and cultivated fields, clustered around many scattered wells, in contrast with the evidently more extensive areas that appear to have been cultivated whilst the foggaras were operating.
The archaeology of drylands
168
Garama was displaced as the regional capital by sites further east and south (Murzuk, Traghen, Zuila), but its substantial walls and kasbah guaranteed it a role in the politics and warfare of the period. Nonetheless, when the earliest
Figure 9.7 Model of the medieval landscape around Germa, showing shrinkage of the cultivated area and demographic decline after the failure of the foggara systems and the refocusing of agriculture around wells in the valley centre European travellers penetrated into the Sahara in the late eighteenth and early nineteenth centuries, they found the Wadi el-Agial a desperately impoverished region, with many of its villages underpopulated and crumbling, and the bulk of its agricultural production taken as taxes and rents by absentee sheiks and Turkish officials (Barth, 1857:143–9; Bruce-Lockhart and Wright 1999; Denham and Clapperton 1826:169–77). Only in the last forty years have modern artesian wells reversed the trend of decline and revived the population and agricultural productivity. However, this has been at a cost
Twelve thousand years of human adaptation in Fezzan
169
to the aquifer levels, which have already fallen significantly. In the long term (Fig. 9.8), it is possible that agriculture will be forced to contract around a limited group of agricultural settlements with very deep bore artesian wells serving clusters of individual irrigated crop
Figure 9.8 Model of hypothetical future direction of settlement and farming in Fezzan, with the concentration of population around a series of agricultural settlements, irrigating large circular fields with very deep artesian wells circles, each of c.300 m diameter. This system developed elsewhere in Libya for exploiting fossil water supplies deep below the Sahara (cf. Allan, 1979). The Fezzan Project cannot offer solutions to the problem of where water is to come from next, but it has graphically illustrated the human consequences of past changes in water availability in the desert. Whilst we may take pride in human ingenuity in finding ways to live in the desert, we may also reflect on the environmental costs that such ‘mastery’ brings in its wake. Garamantian development of the foggara irrigation systems may, in the long term, have been a key factor leading to the decline of their civilization as a result of overextraction from a non-renewable groundwater source.
The archaeology of drylands
170
ACKNOWLEDGEMENTS The Fezzan Project is sponsored by the Society for Libyan Studies, the Arts and Humanities Research Board, the British Academy and the University of Leicester. The final publication of the earlier work by Daniels is supported by a major grant from the Leverhulme Trust. This chapter was written during the tenure of a Research Readership award from the British Academy. The project involves the work of many individuals, who are thanked as a group here, but whose contributions are clearly acknowledged in the multi-authored interim reports. Thanks are also due to D.Miles-Williams for Figure 9.2 and L.Farr for Figures 9.3–9.7.
REFERENCES Allan, J.A. (1979) Managing agricultural resources in Libya: recent experience. Libyan Studies 10:17–28. Ayoub, M.S. (1967a) Excavations in Germa between 1962 and 1966. Tripoli, Ministry of Education. Ayoub, M.S. (1967b) The Royal cemetery at Germa. A preliminary report. Libya Antiqua 3–4:213–19. Ayoub, M.S. (1968) The Cemetery ofSaniat Ben Howedy. Tripoli, Ministry of Education. Bagnall, R.A. (1935) Libyan Sands. Travel in a Dead World. London, Hodder & Stoughton. Ball, J. and Beadnell, H.J.L. (1903) Banana Oasis: Its Topography and Geology. Cairo, National Printing Department Barich, B.E. (1987) (ed.) Archaeology and Environment in the Libyan Sahara. The Excavations in the Tadrart Acacus 1978–1983. Oxford, British Archaeological Reports, International Series 368. Barker, G.W. (1989) From classification to interpretation: Libyan prehistory 1969–1989. Libyan Studies 20:31–43. Barker, G. (1996) Prehistoric settlement. In G.Barker, D.Gilbertson, B.Jones and D.Mattingly, Farming the Desert. The UNESCO Libyan Valleys Archaeological Survey, Volume 1 Synthesis: 83–109. Paris, UNESCO; London, Society for Libyan Studies. Barker, G., Gilbertson, D., Jones, B. and Mattingly, D. (1996) Farming the Desert. The UNESCO Libyan Valleys Archaeological Survey: Volume 1 Synthesis.; Volume 2 Gazetteer and Pottery. Paris, UNESCO; London, Society for Libyan Studies. Barth, H. (1857) Travels and Discoveries in North and Central Africa. London, Longman, Brown & Green (reprint Longman 1965). Bates, O. (1914) The Eastern Libyans. London, Frank Cass (reprint 1970). Bousquet, B. (1996) Tell-Douch et sa Region, Geographic d’une Limite de Milieu à une Frontière d’Empire. Cairo, Institut Française d’Archéologie Orientale. Bovill, E.W. (1968) The Golden Trade of the Moors. Oxford, Oxford University Press
Twelve thousand years of human adaptation in Fezzan
171
(second edition). Brett, M. and Fentress, E. (1996) The Berbers. Oxford, Blackwell. Bruce-Lockhart, J. and Wright, J. (1999) Difficult and Dangerous Roads. Hugh Clapperton’s Travels in Sahara and Fezzan (1822–1825). London, Sickle Moon Press. Camps, G. (1980) Berbères. Aux Marges de l’Histoire. Toulouse, Hespérides. Camps, G. (1989) Les chars sahariens. Images d’une societé aristocratique. Antiquités Africaines 25:11–40. Churcher, C.S. and Mills, A.J. (1997) Reports from the Survey of Dakhleh Oasis, Western Desert of Egypt, 1977–1997. Oxford, Oxford Archaeological Monographs. Cremaschi, M. (1998) Late quaternary geological evidence for environmental change in south-western Fezzan (Libyan Sahara). In M.Cremaschi and S.di Lernia (eds) Wadi Teshuinat: Palaeoenvironment and Prehistory in South-western Fezzan (Libyan Sahara): 13–47. Florence, Insegna del Giglio. Cremaschi, M. and di Lernia, S. (1998) (eds) Wadi Teshuinat. Palaeoenvironment and Prehistory in South-western Fezzan (Libyan Sahara). Florence, Insegna del Giglio. Daniels, C.M. (1968) Garamantian excavations: Zinchecra 1965–1967. Libya Antiqua 5:113–94. Daniels, C.M. (1969) The Garamantes. In W.H.Kanes (ed.) Geology, Archaeology and Prehistory of the Southwestern Fezzan, Libya: 31–52. Castelfranco-Veneto, Petroleum Exploration Society of Libya. Daniels, C.M. (1970) The Garamantes of Southern Libya. London, Oleander. Daniels, C.M. (1971) The Garamantes of Fezzan. In F.F.Gadallah (ed.) Libya in History. Proceedings of a Conference Held at the Faculty of Arts, University of Libya 1968:261–85. Benghazi, University of Libya. Daniels, C.M. (1975) An ancient people of the Libyan Sahara. In J.Bynon and T. Bynon (eds) Hamito-Semitica: 249–65. The Hague, Mouton. Daniels, C.M. (1989) Excavation and fieldwork amongst the Garamantes. Libyan Studies 20:45–61. Denham, D. and Clapperton, H. (1826) Narration of Travels and Discoveries in Northern and Central Africa in the Years 1822–1824. London, John Murray (reprinted 1965 as Missions to the Niger III, Haklyt Society, second series CXIX, edited by E.W. Bovill). Duveyrier, H. (1864) Les Touaregs du Nord. Paris, Challamel Ainé. Edmunds, W.M. and Wright, E.P. (1979) Groundwater recharge and palaeoclimate in the Sirte and Kufra basins. Journal of Hydrology 119:71–87. Edwards, D., Hawthorne, J., Dore, J. and Mattingly, D.J. (1999. The Garamantes of Fezzan revisited: publishing the C.M.Daniels’ archive. Libyan Studies 30:109–27. el-Rashedy, F. (1988) Les pratiques funéraires des Garamantes et leurs relations avec celles d’autres peuples d’Afrique du Nord. In Libya Antiqua, Histoire Générale de l’Afrique. Etudes et Documents III: 85–114. Paris, UNESCO. Encyclopédie (1997) Encyclopédie Berbère fasc. xviii, s.v Fezzan: 2777–817. Aix-enProvence, Edisud. Fakhry, A. (1974) The Oases of Egypt. II, Bahariyah and Farfara Oases. Cairo, American University of Cairo Press. Fontana, S. (1995) I manufatti romani nei corredi funerari del Fezzan. Testimonianza dei commerci e della cultura dei Garamanti (I–III sec.d. C.). In Productions et
The archaeology of drylands
172
Exportations Africaines. Actualités Archéologiques: 405–20. Paris, VI Colloque International sur l’Histoire et l’Archéologie de l’Afrique du Nord. Goblot, H. (1979) Les Qanats: une Technique d’Acquisition de l’Eau. Paris and New York, Mouton éditeur, Industrie et Artisanat 9. Kanes, W.H. (1969) (ed.) Geology, Archaeology and Prehistory of the Southwestern Fezzan, Libya. Castelfranco-Veneto, Petroleum Exploration Society of Libya. Klitzsch, E. and Baird, D.W. (1969) Stratigraphy and palaeohydrology of the Germa (Jarma) area, southwest Libya. In W.H.Kanes (ed.) Geology, Archaeology and Prehistory of the Southwestern Fezzan, Libya: 67–80. Castelfranco-Veneto, Petroleum Exploration Society of Libya. Le Quellec, J.-L. (1987) L’Art Rupestre du Fezzan Septentrional (Libye): Widyan Zreda et Tarut (Wadi esh-Shati). Oxford, British Archaeological Reports, International Series 365. Liverani, M. (1999) Ultime scoperte nella terra dei Garamanti. Archeo. Attualità del Passato 15.8:30–9. Lô, Capitaine (1953) Les foggaras du Tidikelt. Travaux de l’Institut de Recherches Sahariennes 10:139–79. Lô, Capitaine (1954) Les foggaras du Tidikelt. Travaux de l’Institut de Recherches Sahariennes 11, 49–77. Lutz, R. and Lutz, G. (1995) The Secret of the Sahara. Innsbruck, Golf Verlag. Mattingly, D.J. (1995) Tripolitania. London, Batsford. Mattingly, D.J., al-Mashai, M., Balcombe, P., Chapman, S., Coddington, H., Davison, J., Kenyon, D., Wilson, A.I. and Witcher, R. (1997) The Fezzan Project 1997: methodologies and results of the first season. Libyan Studies 28:11–25. Mattingly, D.J., al-Mashai, M., Balcombe, P., Chapman, S., Coddington, H., Davison, J., Kenyon, D., Wilson, A.I. and Witcher, R. (1998a) The Fezzan Project I: research goals, methodologies and results of the 1997 season. Libya Antiqua ns 3: 175–99. Mattingly, D.J., al-Mashai, M., Aburgheba, H., Balcombe, P., Eastaugh, E., Gillings, M., Leone, A., McLaren, S., Owen, P., Pelling, R., Reynolds, T., Stirling, L., Thomas, D., Watson, D., Wilson, A.I. and White, K. (1998b) The Fezzan Project 1998: preliminary report on the second season of work. Libyan Studies 29:115–44. Mattingly, D.J., al-Mashai, M., Aburgheba, H., Balcombe, P., Eastaugh, E., Gillings, M., Leone, A., McLaren, S., Owen, P., Pelling, R., Reynolds, T., Stirling, L., Thomas, D., Watson, D., Wilson, A.I. and White, K. (1999a) The Fezzan Project II: preliminary report on the 1998 season. Libya Antiqua ns 4:63–93. Mattingly, D.J., al-Mashai, M., Balcombe, P., Drake, N., Knight, S., McLaren, S., Pelling, R. Reynolds, T. Thomas, D., Wilson, A. and White, K. (1999b) The Fezzan Project 1999: preliminary report on the third season of work Libyan Studies 30:129– 45. Mori, F. (1969) Prehistoric cultures in Tadrart Acacus, Libyan Sahara. In W.H.Kanes (ed.) Geology, Archaeology and Prehistory of the Southwestern Fezzan, Libya: 21–30. Castelfranco-Veneto, Petroleum Exploration Society of Libya. Mori, F. (1988) L’art rupestre préhistorique dans le Sahara libyen comme aboutissement d’un long processsus bioculturel. In Libya Antiqua, Histoire Générale de l’Afrique. Etudes et Documents III: 157–63. Paris, UNESCO.
Twelve thousand years of human adaptation in Fezzan
173
Muzzolini, A. (1991) Proposals for updating the rock-drawing sequence of the Acacus (Libya). Libyan Studies 22:7–30. Nesson, C., Rouvillois-Brigol, M. and Vallet, J. (1973) Oasis du Sahara Algérien. Paris, Institut Géographique National. Nicolaisen, J. and Nicolaisen, I. (1997) The Pastoral Touareg: Ecology, Culture and Society . London, Thames & Hudson, two volumes. Pace, P., Sergi, S. and Caputo, G. (1951) Scavi Sahariani. Monumenti Antichi 41: 150– 549. Petit-Maire, N. (1982) (ed.) Le Shati. Lac Pleistocene au Fezzan. Paris, CNRS. Petit-Maire, N. (1988) Climatic change and man in the Sahara. In J.Bower and D. Lubell (eds) Prehistoric Cultures and Environments in the Late Quaternary of Africa: 19–42. Oxford, British Archaeological Reports, International Series 405. Petit-Maire, N., Delibrias, G. and Gaven, C. (1980) Pleistocene lakes in the Shati area, Fezzan (27°30′N). In M.Sarnthein, E.Seibold and P.Rognon (eds) Sahara and the Surrounding Seas: 289–93. Rotterdam, Balkema, Palaeoecology of Africa 12. RSGI (1937)=Real Società Geographica Italiana, Il Sahara Italiano. Fezzan e Oasi di Gat. Rome, Società Italiana Arti Grafiche. Ruprechtsberger, E.M. (1997) Die Garamanten, Geschichte und kultur eines Libyschen Volkes in der Sahara. Mainz, Verlag P. von Zabern. Scarin, E. (1934) Le Oasi del Fezzan. Bologna, Zanichelli, two volumes. Scarin, E. (1937) Le Oasi Cirenaiche del 29° Parallelo. Florence, Sansoni. Shaw, B.D. (1976) Climate, environment and prehistory in the Sahara. World Archaeology 8.2:133–49. Trousset, P. (1986) Les oasis présahariennes dans l’antiquité: partage de l’eau et division du temps. Antiquités Africaines 22:161–91. UNESCO Nomades (1962) = Recherches sur la Zone Aride XIX, Nomades et Nomadisme au Sahara. Paris, UNESCO Publications. van der Veen, M. (1992) Garamantian agriculture: the plant remains from Zinchecra, Fezzan. Libyan Studies 23:7–39. van der Veen, M. (1995) Ancient agriculture in Libya: a review of the evidence Acta Palaeobotanica 35 (1):85–98. van der Veen, M. (1999) (ed.) The Exploitation of Plant Resources in Ancient Africa. New York, Plenum. Zaluski, M. and Sadek, K.E. (1980) Hydrogeology of mesozoic aquifers in the western part of Wadi al Ajal . Symposium on the Geology of Libya 2:635–42. Ziegert, H. (1995) Das neue Bild des Umenschen. UniHH Forschung. Beitrage aus der Universität Hamburg 30:9–15.
10 Farming and famine: subsistence strategies in Highland Ethiopia ANN BUTLER AND A. CATHERINE D’ANDREA
INTRODUCTION The Highlands of Ethiopia have an environment that is governed by the high altitude, and within relatively low longitudes they have a temperate climate. This supports a wide range of crops, which include both indigenous African domesticates and the cool-season grain crops developed in, and introduced in antiquity from, southwest Asia. Cultivation is rainfed, and the technology is largely ox-plough. The region therefore presents an ideal situation for the study of traditional dryland agriculture and provides an opportunity to understand some of the rationale that underlies these farming practices. The results of a new ethnoarchaeological study in the Ethiopian Highlands are presented here, integrated with some published accounts of traditional agriculture in the region.
FIELDWORK Ethnobotanical studies were carried out between 1996 and 1998 in the Ethiopian Highlands (Tigrai province) about 2,000 m above sea level, in the mid-altitude region. This is the agro-climatic zone known as dry woina dega (Bekele-Tesemma, 1993:6). The average temperature range is between 5 and 40°C (Gebremedin and Haile, 1997). Fieldwork was concentrated on the northern edge of the Giba plateau, in the Enderta administrative region (woreda) and village group (tabia) of Mahabere Genet, about 15 km northwest of the provincial capital Mekelle. Adi Ainawalid, a village (kushet) of 180 households, was selected for a detailed study (Fig. 10.1). Supplementary records were made both at further kushets within the same tabia, and also at others within the woreda of Entalo-Wajeret near Adi Gudem, about 30 km south of Mekelle. Farming practices were observed and farmers were interviewed between May and June and during the main harvest time between November and December (Butler, in press; Butler et al., 1999; D’Andrea et al., 1997, 1999).
Farming and famine: subsistence strategies in Highland Ethiopia
175
Figure 10.1 Map of Ethiopia, showing Adi Ainawalid in Tigrai province During the political upheavals in Ethiopia between 1973 and the early 1990s, there was large-scale compulsory resettlement. Individually owned and managed farmland was taken and incorporated into large co-operatives. By 1996 these had been dispersed: land holdings had only recently been reallocated to individual farmers and traditional farming practices had been resumed. However, several families at Adi Ainawalid were spared the turmoil: throughout the conflict they occupied their original family homes, and they now retain some land farmed by their grandparents.
ENVIRONMENT, AGRARIAN SYSTEMS AND CROPS IN THE NORTHERN HIGHLANDS Subsistence is largely vegetarian and depends on the household production of grain crops, supplemented by a few resources, such as salt and oil, bought from the regional market in Mekelle. Land holdings are based on units (tsumdi) representing the area of land that can be ploughed by one ox-team in a day, which is estimated at a quarter of a hectare (Adebo, 1993:48; Konde, 1993: 18). Individual holdings are very small, ranging from one to eight tsumdi, commonly consisting of at least two plots, usually one adjacent
The archaeology of drylands
176
to the dwelling and the other(s) up to an hour’s walk away towards the perimeter of the village. This is similar to the situation described for other areas in Ethiopia (Tsegaye, 1997). It was noted that some families in the kushets investigated are newly settled and have missed an allocation of land. They have had either to rent land from a household lacking the means to work it or to find an alternative to farming as a means of livelihood. Plot boundaries may be defined by stones, shallow drainage channels or spiny brushwood. The houses are round or rectilinear, built of local stone, and with thatched, earthen or wooden roofs. A separate kitchen building is common as well as an enclosed area for animals within the surrounding stone-walled compound (D’Andrea et al., 1997; Fig. 10.2). Livestock is also kept inside houses, especially donkeys, horses and calves, to protect them from predation by hyenas. Gardens are common in larger mature compounds but are rarely found with small houses. The soils are largely derived from limestones weathered to vertisols and cambisols, which are clays and sandy clays, and in this region they are typically stony, thin and eroded (Hunting Technical Services Ltd, 1973–4; Mitiku Haile, pers. comm.). The natural vegetation in the region is described as Acacia savannah (Bekele-Tesemma, 1993:6), but today there are few trees. The action of heavy rains and trampling by livestock on the treeless and uncultivated soils tend to cause surface crusting; this restricts penetration by
Figure 10.2 Residential compound near fields, west end of Adi Ainawalid, facing southwest, November 1997 Photograph: C.D’Andrea water and promotes run-off (Butzer, 1981). Attempts are made to catch and retain rainwater in clay-lined artificial ponds, which are used mainly for watering livestock and for washing. The main anti-erosion strategy in the region is the use of soil-retentive
Farming and famine: subsistence strategies in Highland Ethiopia
177
terracing: low walls are constructed on the surrounding slopes using stones dug out from local outcrops by the farmers during the slacker farming periods or in food-for-work programmes. To further reduce erosion and conserve rainfall, a tree-planting scheme has recently been undertaken on the slopes around Adi Ainawalid, using native species of genera such as Acacia Mill, and Erythrina L. In the settlement area, small stands of Eucalyptus species have been introduced mainly for shade and fuel. There are occasional compounds with small, mainly leguminous, trees planted for shade and supplementary livestock feed. Until recently, water for household use has had to be carried from the river, up to two hours’ walk away, often twice daily, but at Adi Ainawalid three wells are now available for use. Irrigation is rare and confined to plots near the river, which are rented for the cash-cropping of introduced crops such as tomatoes, potatoes and maize. Livestock In Tigrai, oxen play a central role in the household economics (Bauer, 1975; McCann, 1995:48–56). Their availability is essential to cultivation, although donkeys and, more rarely, camels and mules also supply labour. A 2-year-old ox takes a year’s training and can give up to five years’ work (Spiess, 1994). During the study period, of those farmers questioned at Adi Ainawalid only about one third owned an ox, which is a similar finding to other surveys in the region (e.g. FAO, 1986). Animals are commonly loaned to make up a ploughing team and for threshing, when up to eight or more oxen may be used to trample the yield from a single harvest. Also a man with a team will plough land for others for a payment of half the harvested crop. Occasionally donkeys, mules or mixed teams, may be used for ploughing. The number of animals is restricted by a shortage of feed. This is most scarce just prior to the heaviest ploughing season: thus the oxen tend to be undernourished and least fit when their labour is most in demand (Konde, 1993:70). Cattle, small ruminants, equids and the few camels graze field edges and stubble. Feed crops are not cultivated, nor is land set aside for hay. To conserve the pasture and reduce erosion from over-grazing, the availability of the communal village grazing lands is carefully restricted to certain periods and to particular animals, mainly oxen and cattle. By-products of cultivation, such as weeds and crop-processing residues, and food-preparation residues, are the most important feed resources, the choicest of which are fed to the oxen. Tree and shrub vegetation provides useful supplementary fodder. Dung is the most valuable animal by-product. It is used primarily for fuel, and also for fertilizers and as the raw material to make various household features and effects, such as house floors, storage-jar lids and pot stands. Skins are an important source of income and are used, for example, to store honey and grain and to make baby carriers and as mats. Plant resources The wide range of grain crops that is cultivated includes species of indigenous African domesticates, members of the assemblage of southwest Asian founder crops and some introductions from the New World (Table 10.1). Farmers possess much detailed local knowledge of the habit and environmental requirements of the different varieties of each
The archaeology of drylands
178
cultigen, which is a skill widely recognized (Tsegaye, 1997; Worede and Mekbib, 1993). This rural knowledge has been a very important factor in the successful reinstatement of traditional farming systems following nearly twenty years of agricultural changes. Each year the choice of crop and the selection of the particular variety or mixture of species are based primarily on an estimate of which is most likely to succeed best in a particular field under the environmental conditions anticipated during the following season. A range of traits in each taxon is desirable, such as both early and late maturing varieties (Gebremedin and Haile, 1997). Also, crop types are selected for variables such as the colour of resulting food products, and the baking quality and storage quality of grains are more important than the grain yield and the size (Haile, 1995; Webb and von Braun, 1994). To extend the range of crop types available to individual farmers, locally developed populations of grains are exchanged and further varieties or species may be purchased from the regional market. These measures help to perpetuate recognized landraces (Worede and Mekbib, 1993). To intensify the yield produced from the small land-holdings, and to spread risk, mixtures of different species are inter-cropped. At Adi Ainawalid, a wheat and barley mixture (hanfetse) and mixtures of wheat species are common (Fig. 10.3). In other areas, mixtures such as pea and faba bean (ater-abie), or sorghum with chickpea, are sown. When possible, the varieties of the crops are chosen for their synchroneity of development: for example, a hanfetse of shahan wheat and burguda barley can be sown, harvested and processed as a single crop. The grains may then be treated as a single resource and prepared for food as one, or the constituent grain types may be separated in the home. When a single species is planted, several varieties may be mixed. The proportion of different grains in the mixture at harvest differs from that sown, so the mixtures have to be reassembled each sowing season. Following periods of drought, crop failures can result in a reduction in the number of crop species and varieties harvested, as well as the total yield. In the few house gardens, chilli (berbere), garlic (ta’eda shigurtee), onions (shigurtee), basil (seseg) and other spice plants and herbs may be cultivated. These provide important nutritional components and, when available, are always added to the staple carbohydrate foods. As an example of the exotic drought-tolerant New World species that have been introduced, prickly pear, Opuntia ficus-indica Mill. (beles), is commonly planted as hedging, the leaves also being a valuable source of fodder and the fruits a human food. Wild plant resources are collected mainly for medicinal and other non-food uses. For example, grasses such as Hyperrhenia hirta (L.) Stapf. (sa’ri awald),
Table 10.1 Crops cultivated at Adi Ainawalid, Tigrai, Ethiopia Crop species Common name and varieties Sorghum bicolor (L.) Moench sorghum (5 varieties) Eragrostis tef (Zucc.) Trotter teff (red, white) Triticum turgidum conv. durum durum (black with hexaploid (Desf.) MacKey characters) bread wheat (shahan, Canada Triticum aestivum subsp. vulgare wheat) (Vill.) MacKey
Local name mashella taff tselimoi sindai
Farming and famine: subsistence strategies in Highland Ethiopia
179
Hordeum vulgare L. Triticum L. sp./Hordeum L. sp. Eleusine coracana (L.) Gaertn.
barley (burguda, sa’saa) segem wheat/barley intercrop mixture hanfetse finger millet (black, red, dagousha white) Lens culinaris ssp.culinaris Medikus lentil bersheem Lathy rus sativus L. grasspea gwayya, sebero Cicer arietinum L. chickpea shimbra Zea mays L. maize (arigo, beraho) efoon Trigonella foenum-graecum L. fenugreek abaka Linum usitatissimum L. linseed indata
Figure 10.3 Intercropped bread and durum wheats near Mai Kayeh, Tigrai, November 1997 Photograph: C.D’Andrea H.rufa (Nees) Stapf. and Eleusine floccifera (Forssk.) Spreng. (rigaha), are gathered to weave into baskets, and the labiate Otostegia integnfolia Benth. (chi’indogwee) has insecticidal properties, the juices being smeared onto livestock to prevent damage to their hides. Cultivation systems In Ethiopia as a whole, the environmental factor that has the greatest control over the farming schedule is said to be rainfall (McCann, 1995:28–31), which characteristically is bimodal. The small spring rains (belg) and the main summer rains (kremt) support two cropping seasons. However, throughout the year there can be unpredictable rains,
The archaeology of drylands
180
sometimes with heavy hailstorms damaging to crops. Cyclical episodes of low rainfall occur, associated with the El Niño Southerly Oscillation (ENSO), with frequencies varying between three and fifteen years (Bekele, 1997; Wolde-Georgis, 1997; and see Chapter 2). There is also evidence that for several decades the basic annual rainfall has been decreasing. Although a belg season still occurs in the central Highlands in the provinces of Wello and Shewa (Rahmato, 1991:54), and even in parts of the northern Highlands (Emergencies Unit for Ethiopia, 1997), over the past thirty years the total annual rainfall in much of the northern province of Tigrai has been drastically reduced, and the spring rains have virtually ceased (Emergencies Unit for Ethiopia, 1994). This situation has largely restricted the cropping periods to a single season (meher), associated with the kremt (Adebo, 1993:75; Konde, 1993:79). Today, despite an annual rainfall of between 450 and 900 mm (Butzer, 1981), Tigrai is known as one of the most droughtprone regions of Ethiopia (Webb and von Braun, 1994). Thus in southeastern Tigrai, the land is tilled for a single growing season. The production from the small plots is optimized by measures such as inter-cropping, double cropping, the ploughing-up of headland and land rental, so that at any one time the maximum area is cultivated. Soil preparation begins in late winter or early spring with ploughing. The seed bed for cereals usually receives several ploughings, but the plots for pulses may be ploughed only once. Stones and tree stumps are retained in the soil to reduce erosion from wind and water. Grain crops are usually broadcast sown between May and July—a period associated with the start of the big rains, which are concentrated between June and September. Cereal crops attract the priority of farming input, but even for this crop category it appears minimal. Cereal fields may be manured with dung from grazing ruminants; although government supplies of chemical fertilizers are sometimes available, their application is usually precluded by their expense. Pulses are very rarely fertilized. Weeding also is rare, but is more common for cereals. Cereal plots may be ploughed again to aerate the soil and facilitate drainage and reduce the weeds. The main harvesting season falls between October and December. Crop plants are commonly uprooted individually by hand (Fig. 10.4), or they may be either uprooted or cut by sickle. Weeds may also be uprooted and harvested separately for feed. Unpalatable or spiny weeds remain standing in the fields. Cultivation of the different crops is commonly staggered across the growing season to increase the breadth of the harvest season, thereby preventing an excessive concentration of labour and resources at one period. Finger millet, sorghum, maize and lentil are usually sown earlier than wheats, barleys and grasspea; chickpea is often the latest crop (D’Andrea et al., 1997). Grain separation takes place on threshing floors of compacted soil, constructed at the edges of fields or within the settlement area, as described in detail by D’Andrea and others (1997, 1999). The harvested crop is carried to the edge of the floor, where it is piled to dry or threshed immediately. A threshing team of up to eight oxen is driven around the floor over the crop, and the crop fractions are sorted with forks and brushes (Fig. 10.5). Winnowing is a complex set of operations involving the use of several implements, and it results in a cleaned pile of grain and the crop residues, which are normally amalgamated for feed (Fig. 10.6). The threshing floor is swept clean ready for the next crop, and the separated crop fractions are carried to the house. Attempts are made to maximize the yield of crop varieties that are prone to shatter, by
Farming and famine: subsistence strategies in Highland Ethiopia
181
harvesting them either when they are still slightly under-ripe, or else early in the day prior to the highest temperatures. Crop losses from the predation of birds and rodents are minimized by child scarers or scare-crows stationed in the fields. Crop rotation is used to break cycles of plant parasites, such as Striga asiatica (L.) Kunze. (selemi) on cereals, particularly sorghum, and Orobanche minor Sm. (m’andat tali) on legumes. Barley and
Figure 10.4 Harvesting grasspea by hand uprooting, Adi Ainawalid, November 1996 Photograph: A.Butler tselimoi wheat are said to be the cereals most resistant to Striga. Crop rotation usually consists of three to four years of cereals planted to every one of pulses. Pulses, especially chickpeas, and flax are planted to reinvigorate plots depleted of nutrients. Because of the shortage of land, fallowing is unusual. A fallowed plot will often signal a shortage of oxen or human labour (Adebo, 1993:83; Konde, 1993:80).
The archaeology of drylands
182
Individual families normally provide the labour to work their own land, with assistance when required from the extended family or from near neighbours. Sometimes labour will be hired. Ploughing is undertaken by men and older boys. In cases where no male family member is available to work the land, it may be rented out in return for half the yield at harvest. It was reported that, although not unknown, it was very rare for a woman to plough. The whole family may be engaged in weeding and harvesting. Men
Figure 10.5 First threshing of teff, Adi Ainawalid, November 1997 Photograph: C.D’Andrea
Farming and famine: subsistence strategies in Highland Ethiopia
183
Figure 10.6 Winnowing teff, Adi Ainawalid, November, 1996 Photograph: A.Butler and children perform the threshing and winnowing stages of crop processing. Women are concerned with small-scale winnowing and fine grain cleaning within the household, and the preparation of food. Importantly, they also play a significant role in discussions on the annual farming schedule and on crop and seed selection. Grain storage Storage is overseen by the women (Tsegaye, 1997). Cleaned grain is stored inside the houses in clay or bamboo vessels about 1 m tall and sealed with dung (Fig. 10.7). Fumigants or insecticides are not added, but it is believed that the dung acts as an insect repellant. Small crop yields may be kept in skin bags or sacks. Unthreshed crops are stacked in the house compounds, as are the threshing residues for animal feed. Cats are kept to deter rodents.
The archaeology of drylands
184
Figure 10.7 Grain storage jars, Adi Ainawalid, November, 1996 Photograph: A.Butler The long-term storage of grains in clay-lined underground pits is said to be a practice in highland regions where the soil is dry. The pits are concealed to minimize the loss of grain through plunder. Sorghum is known to have survived such emergency storage for at least five years (McCann, 1995:67–8; Rahmato, 1991:31; Worede and Mekbib, 1993). However, owing to its secret nature, this storage system was not investigated during this study. Farming and fuel In the Ethiopian Highlands, the paucity of trees is believed to be long-standing, and is increasing (Ståhl, 1993). Historical descriptions give varying accounts of the vegetation
Farming and famine: subsistence strategies in Highland Ethiopia
185
of Tigrai; and the nineteenth-century illustrations of Salt (1814) show moderate tree cover rather than dense woodland near Mekelle. In the early 1900s it was estimated that there was about 40 per cent tree cover in the country as a whole, but by the late 1980s this had fallen to 5.6 per cent (FAO, 1986). In Tigrai, the demands for fuelwood and for timber for manufacturing appear to have reduced the number of mature trees mainly to the wooded conservation areas around churches or to single specimens used as community assembly points. This loss has been accelerating, due partly to continuing household demands, and also to the need to supplement income outside farming: the sale of firewood has been a traditional supplementary source of income until the recent past. Nowadays, in order to protect the remaining trees, gathering timber for fuel is licensed (Derege Asefa, pers. comm.). Wood continues to be the raw material for house supports and farming implements such as plough beams, yokes and winnowing forks. Up to 55 per cent of the fuel resources are provided by alternatives to timber (World Bank, 1984), and are mainly farming by-products. Dung is perhaps the most valued. It accumulates in residential compounds where livestock are penned overnight, and is collected by children from the grazing areas; it is then spread on walls to dry and be stacked. The culms of sorghum and other vegetable material are also important fuel resources. Thus fuel is sparingly used, and dried grasses are the usual kindling. For each cooking episode, small fires are lighted individually within the stoves. The latter are usually permanent fixtures of clay and stone constructed inside the kitchen building, but small portable stoves are also used (D’Andrea et al., 1999).
CROP DIVERSITY Many Ethiopian crops are noted for an impressive diversity of form under environmental conditions that were described by Vavilov (1935:347) as relatively uniform within the high altitude. This morphological diversity incorporates traits adapted to various stress conditions, and it seems to be maintained by both environmental and human agencies. Recent studies of Ethiopian wheats have shown that crop varieties with deeply pigmented black and purple grains appear to be adapted to high altitudes (Tesemma, 1991). Interestingly, at Adi Ainawalid a type of black wheat (tselimoi) is grown, which has been identified as a hybrid form of durum with hexaploid characteristics (Gordon Hillman, pers. comm.). Temperature and drought stress are important factors that affect variables such as plant height, the protein content of grains and the timing of heading (Annicchiarico et al., 1995). Many varieties have been developed that are pigmented, and also of low height and early maturation; these desirable traits confer resistance to falling in wet or windy weather (lodging) (Bejiga et al., 1996; Belay et al., 1995). The selection of similar traits is seen in other crops such as barley (Demissie and Bjornstad, 1996) and pea (Govorov, 1930). Many of these crop types tend to be low-yielding (Tesemma, 1991; Tsegaye, 1997) and are officially regarded as having low industrial quality, but because of their performance under potentially stressful environmental conditions they continue to be selected by farmers (Belay et al., 1995). However, while a general diversity of crop varieties appears to be continually maintained by farmers’ selection, the cultivation of some species of food plants is
The archaeology of drylands
186
becoming restricted, due to the changing pattern of climate and land shortages. This is affecting the long-term local availability of some crops. Following a succession of bad years, there has been a long-term reduction in the range of grains planted in the study area. Table 10.2 lists the crops that are no longer found at Adi Ainawalid, or indeed in the tabia of Mahabere Genet as a whole. However, they are familiar to many farmers of Adi Ainawalid, and they have been grown locally in the past. These crops are still grown in areas of higher rainfall, particularly further south in Tigrai, and occasionally they can be found in the regional market in Mekelle. As opposed to the black wheat (tselimoi) mentioned above, ‘classic’ durum wheat, recently still said to be one of the major cereals in Ethiopia (Engels and Hawkes, 1991), is an uncommon crop in the study area; one farmer in the kushet of Adi Akel immediately adjacent to Adi Ainawalid had twice obtained some durum grain, but this had produced only sterile plants. All the rare grains are valued as traditional resources with special properties: durum makes a heavy, solid loaf for sustaining field lunches at harvest-time; emmer wheat is made into a nutritious and easily digested gruel for invalids and babies; peas and faba beans, although expensive, are still regularly bought in for meals on
Table 10.2 Crops no longer cultivated at Adi Ainawalid, Tigrai, Ethiopia Crop species Common name Local name Triticum turgidum spp. diccocum (Schrank) Emmer wheat ares Thell. T.turgidum conv. durum (Desf.) MacKey Durum wheat kinkinai Pisum sativum ssp. abyssinicum A.Br. Ethiopian pea dekoko P.sativum ssp. sativum var arvense L. Field pea ater Vicia faba L. Faba bean abie Pisum sativum L./Vicia faba L. Pea/faba bean ater/abie intercrop Holy Days throughout the year. The Ethiopian pea is particularly sought after, and the addition of even a few seeds to a festival legume dish is held to enrich the celebrations.
CULTIVATION UNDER DROUGHT CONDITIONS The annual precipitation in Tigrai is well within the levels generally considered to be enough to support dryland agriculture (Butzer, 1981). At the same time, however, this rainfall is acknowledged to be insufficient (Emergencies Unit for Ethiopia, 1994, 1997). The discrepancy can be explained by changes in pattern: the expected rains of the spring and summer monsoon seasons have diminished, but the total annual rainfall may be augmented to near-normal levels by unpredictable showers, often of heavy hailstorms destructive to crops. The opportunistic use of watered soil is a feature of Ethiopian highland agriculture. The study periods between 1996 and 1997 fell within an episode of low precipitation,
Farming and famine: subsistence strategies in Highland Ethiopia
187
and it was possible to observe the effects of water-stress on farming. Rainfall in the region of Mekelle was generally reduced, scattered in its distribution, and of spatially differing amounts. The reservoir at Adi Ainwalid was dry in both years, although the one shared by nearby kushets appeared to have retained some water. In 1997 some farmers in adjacent settlements appeared to be producing yields of most grain crops, yet the farmers of Adi Ainawalid, although they had cultivated a restricted range of the most droughttolerant crop species, experienced severe crop failure, with very reduced harvests. Following a very dry summer, there were outbreaks of heavy rain in November that caused lodging and ruined many of the surviving small harvests. At the adjacent kushet of Adi Akel, where the soils appear to be deeper and more water-retentive, harvests seemed to be less affected, and following the unseasonal rains some farmers ploughed for extra end-of-season crops of barley and chickpeas, which would have been ready for harvesting at the end of the following spring. In other highland regions, when the summer rains are especially heavy and flooding occurs, double-cropping is common. Short-season species such as grasspea can be sown and harvested on the semi-waterlogged fields prior to the main growing season on the drained land (Abate Tedla, pers. comm.). Rainfall above 200 mm is not officially classified as a drought (World Bank, 1984), yet in Tigrai a chronic food shortage prevails. Between 1988 and 1992 three-quarters of the families produced insufficient food (Holt and Lawrence, 1993:26–31), and in 1997 the regional food production was deficient by about 20 per cent (Gebremedin and Haile, 1997). This is in contrast with previous times: documentation from the sixteenth century, for example, describes southern Tigrai as a land of great abundance of production, with great yields of cereals and pulses (Alvares, 1520:31). The current shortfall is thought to be less a reflection of the demands of an increasing population than the result of the dual effects of the changing pattern of climate and the political circumstances (Pankhurst, 1992:318; Zewde, 1991:195–6). On the local level, in 1996 crop production throughout Tigrai was officially reported by the aid agencies as being poor (Ahrens and Spiess, 1997), and Enderta was singled out as being in particular need of food aid (Emergencies Unit for Ethiopia, 1997). At Adi Ainawalid, some supplies of grain had been distributed, although towards the end of 1997 food aid had been discontinued. These food shortages prompted surveys in the Highlands that have highlighted a number of farming problems, with priorities that vary slightly at different seasons and in different areas: lack of rain, small land holdings, insufficient oxen, lack of human labour, shortage of fuel, high market prices, shortage of livestock feed, livestock disease and unclean drinking water (Adebo, 1993:70–89; Konde, 1993:76–86). Most of these problems are familiar to the farmers of Adi Ainawalid.
CONCLUDING REMARKS The northern Highlands of Ethiopia are a region commonly regarded as being associated with both drought and famine. It is now believed that drought alone seldom causes famine: more often, a combination of factors is involved. These include: epidemics of human and veterinary diseases such as smallpox and cattle rinderpest; plagues on crops of insect predators such as locusts, ants and army worms; and social and political
The archaeology of drylands
188
circumstances (McCann, 1995:89–91). The recent situations of conflict appear to underlie many of the nutritional difficulties seen in the region today. Sustainable agriculture in Ethiopia is characterized both by maintained traditional practices and flexibility. Farming as described prior to the revolution of the early 1970s (e.g. Westphal, 1975; Simoons, 1960) was abruptly and severely altered following resettlement and ‘villagization’ until the early 1990s (Rahmato, 1985; Rock, 1994). Families were split up, relocated in regions of unfamilar ecology, and expected to cultivate alien (to them) crop assemblages on collective farms. Now reinstated, traditional farming is demonstrating, in the selected crops and technology, an essential stability that has been successfully supported by long-term rural knowledge. The farming strategies devised and developed in the Ethiopian Highlands, as witnessed during this study, include mechanisms for survival during periods of climatic stress. Crop germplasm is carefully conserved to allow the best selection of crop types to suit the agricultural situation. Within a wide repertoire, agrarian systems are flexible and encorporate strategies to maximize production under whatever conditions pertain. Mechanisms have been developed to minimize erosion, water is reserved at run-off, soilwater is exploited, grazing is controlled and alternative sources of fuel have been found. It appears that periods of low rainfall and crop failure can be endured for several years by the careful storage of grain during good years. Social networks of exchange and the sharing of resources, such as oxen, are essential mechanisms to bridge periods of shortage. These strategies appear to be able to promote survival for at least two to three years of high aridity. When droughts last longer than a few years, or when epidemics of disease or pests or bouts of conflict or political unrest are superimposed upon drought periods, then it appears that the social and farming mechanisms described above may be insufficient to avert severe food shortage. A better understanding of climatic perturbations could speed the implementation of measures to lessen future threats of famine (Wolde-Georgis, 1997).
ACKNOWLEDGEMENTS The fieldwork was funded by the Social Science and Humanities Research Council of Canada (Grant No. 410–96–1520) and supported with the valuable assistance of Dr Mitiku Haile, Dean of Mekelle University College (MUC). We are grateful to the Committee for Research and Conservation of Cultural Heritage (CRCCH), Addis Ababa, and the Tigrai Bureau of Culture, Tourism and Information, Mekelle, for granting permission to undertake the study. Field assistance was provided by Shewiaye Belay, Zelealem Tesfay, Derege Asefa and Alemtsehay Tsegay of MUC. Figure 10.1 was drawn by Shannon Wood of Simon Fraser University (SFU). Useful comments on the text by Diane Lyons (SFU) are acknowledged. Our deepest thanks go to the kind, generous and patient farmers of Adi Ainawalid and neighbouring settlements in south-central Tigrai.
Farming and famine: subsistence strategies in Highland Ethiopia
189
REFERENCES Adebo, S. (1993) Report of Diagnostic Survey of Debre Tabia in Enderta Wereda. Addis Ababa, FARMAfrica. Ahrens, J. and Spiess, H. (1997) Field Trip to Amhara and Tigray Regions: 5/15/–6/1/96. Situation Report. Emergencies Unit for Ethiopia (UNDP-EUE). (http://www.sas.upenn.edu/African_Studies/EUE/amhra696.htm) Alvares, F. (1520) The Prester John of the Indies. Volume I. Cambridge, Cambridge University Press, translation by C.F.Buckingham and G.W.B.Huntingford (eds) 1961 edition, Hakluyut Society, vols 114–15. Annicchiarico, P., Pecetti, L. and Damania, A.B. (1995) Relationships between phenotype variation and climatic factors at collecting sites in durum wheat landraces. Hereditas 122:163–7. Bauer, D.F. (1975) For want of an ox…: land, capital and social stratification in Tigre. In H.G.Marcus (ed.) Proceedings of the First United States Conference on Ethiopian Studies: 235–48. East Lansing MI, Michigan State University, African Studies Center. Bejiga, G., Tsegaye, S., Tullu, A. and Erskine, W. (1996) Quantitative evaluation of Ethiopian landraces of lentil (Lens culinaris). Genetic Resources and Crop Evolution 43: 293–301. Bekele, F. (1997) Ethiopian use of ENSO information in its seasonal forecast. Internet Journal for African Studies 2. Bekele-Tesemma, A. (1993) Useful Trees and Shrubs for Ethiopia. Nairobi, Regional Soil Conservation Unit, Swedish International Development Authority. Belay, G., Tesemma, T., Bechere, E. and Mitiku, D. (1995) Natural and human selection for purple-grain tetraploid wheats in the Ethiopian highlands. Genetic Resources and Crop Evolution 42:387–91. Butler, E.A. (in press) Sustainable agriculture in a harsh environment: an Ethiopian perspective. In F.Hassan (ed.) Drought, Food and Culture: Food Security in Africa’s Later Prehistory. New York, Plenum Publishing Corporation. Butler, E.A., Tesfay, Z., D’Andrea, A.C. and Lyons, D.E. (1999) The ethnobotany of Lathyrus sativus L. in Highland Ethiopia. In M. van der Veen (ed.) The Exploitation of Plant Resources in Ancient Africa: 123–36. New York, Plenum Publishing Corporation. Butzer, K.W. (1981) Rise and fall of Axum: a geoarchaeological approach. American Antiquity, 46:471–95. D’Andrea, A.C., Haile, M., Butler, E.A. and Lyons, D.E. (1997) Ethnoarchaeological research in the Ethiopian highlands. Nyame Akuma 47:19–26. D’Andrea, A.C., Lyons, D., Haile, M. and Butler, A. (1999) Ethnoarchaeological approaches to the study of prehistoric agriculture in the Ethiopian highlands. In M. van der Veen (ed.) The Exploitation of Plant Resources in Ancient Africa: 101–22. New York, Plenum Publishing Corporation. Demissie, A. and Bjornstad, A. (1996) Phenotypic diversity of Ethiopian barleys in relation to geographical regions, altitudinal range and agro-ecological zones: as an aid
The archaeology of drylands
190
to germplasm collection and comservation strategy. Hereditas 124:17–29. Emergencies Unit for Ethiopia (1994) Situation Report for Region 1 (Tigray). Emergencies Unit for Ethiopia (UN-EUE). (http://www.sas.upenn.edu/African_Studies/EUE/tigray0494.html) Emergencies Unit for Ethiopia (1997) Field Trip to Amhara and Tigray National Regional States. Emergencies Unit for Ethiopia (UN-EUE) Development Programme. (http://www.sas.upenn.edu/African_Studies/EUE/north0296.html) Engels, J.M.M. and Hawkes, J.G. (1991) The Ethiopian gene centre and its genetic diversity. In J.M.M.Engels, J.G.Hawkes and M.Worede (eds) Plant Genetic Resources of Ethiopia: 23–41. Cambridge, Cambridge University Press. FAO (1986) Ethiopia: Economic Analysis of Land Use. Technical Report 8. Rome, FAO. Gebremedin, B. and Haile, M. (1997) Food Security and Dryland Agriculture: the Case of Tigray. Utvikingsfundet (the Development Fund), (http://www.ufondet.no/engelsk/tema/konf/1-3.html) Govorov, L.I. (1930) The peas of Abyssinia. A contribution to the problem of the origin of cultivated peas. Essay II. Bulletin of Applied Botany, Genetics and Plant Breeding (Leningrad) 24:399–431. Haile, M. (1995) Indigenous knowledge and agricultural practices in Central Tigray. Unpublished paper presented at Rural Development Workshop, Mekelle, Tigray. Holt, J. and Lawrence, M. (1993) Making Ends Meet: A Survey of the Food Economy of the Ethiopian North-East Highlands. London, Save the Children UK. Hunting Technical Services Ltd. (1973–4) Tigray Rural Development Studies: Map of Landforms in Mekelle District: Gradients, Soil Depth and Soil Types (1–6). London, Ministry of Overseas Development. Konde, A. (1993) Report of Diagnostic Survey of Debre Medhanit Tabia in Dedebama Derga-Agen Wereda. Addis Ababa, FARMAfrica. McCann, J.C. (1995) People of the Plow. Madison, Wisconsin, University of Wisconsin Press. Pankhurst, R.A. (1992) A Social History of Ethiopia. Trenton, New Jersey, The Red Sea Press. Rahmato, D. (1985) Agrarian Reform in Ethiopia. Trenton, New Jersey, The Red Sea Press. Rahmato, D. (1991) Famine and Social Strategies. Uppsala, Scandinavian Institute of African Studies. Rock, M.J. (1994) Famine and Food Insecurity in Ethiopia: A Critical Assessment of the Notion of “Coping Strategies”. University of Leeds, unpublished PhD thesis. Salt, H. (1814) A Voyage to Abyssinia and Travels in the Interior of that Country. London, F.C. and J.Rivington. Simoons, F.J. (1960) Northwest Ethiopia. Madison, University of Wisconsin Press. Spiess, H. (1994) Report on Drought Animals under Drought Conditions. Emergencies Unit for Ethiopia, http://www.sas.upenn.edu/African_StudiesEUE/drought 0794.html Ståhl, M. (1993) Foreward. In A.Bekele-Tesemma (ed.) Useful Trees and Shrubs for Ethiopia: vii. Nairobi, Regional Soil Conservation Unit, Swedish International Development Authority. Tesemma, T. (1991) Improvement of indigenous durum wheat landraces in Ethiopia. In
Farming and famine: subsistence strategies in Highland Ethiopia
191
J.M.M.Engels, J.G.Hawkes and M.Worede (eds) Plant Genetic Resources of Ethiopia: 288–95. Cambridge, Cambridge University Press. Tsegaye, B. (1997) The significance of biodiversity for sustaining agricultural production and role of women in the traditional sector. Agriculture, Ecosystems and Environment 62:215–27. Vavilov, N.I. (1935) The phyto-geographical basis of plant breeding. Theoretical Bases for Plant Breeding, Moscow, 1. Reprinted and translated in D.Löve (1992) Origin and Geography of Cultivated Plants: 316–66. Cambridge, Cambridge University Press. Webb, P. and von Braun, J. (1994) Famine and Food Security in Ethiopia. Chichester, John Wiley and Sons. Westphal, E. (1975) Agricultural Systems in Ethiopia. Wageningen, Centre for Agricultural Publishing and Documentation (PUDOC). Wolde-Georgis, T. (1997) El Niño and drought early warning in Ethiopia. Internet Journal for African Studies 2. Worede, M. and Hailu Mekbib, H. (1993) Linking genetic resource conservation to farmers in Ethiopia. In W.de Boef, Kojo Amanor and K.Wellard (eds) Cultivating Knowledge: 78–84. London, Intermediate Technology Publications. World Bank (1984) Ethiopia: Recent Economic Developments and Future Prospects. Washington, D.C., World Bank. Zewde, B. (1991) A History of Modem Ethiopia 1855–1974. London, James Curry.
Part IV EASTERN AND SOUTHERN AFRICA
11 Engaruka: farming by irrigation in Maasailand c.AD 1400–1700 JOHN E.G.SUTTON
INTRODUCTION: THE RIFT VALLEY AND CRATER HIGHLANDS OF NORTHERN TANZANIA The equatorial highlands of East Africa are bisected by the north-south trough of the Rift Valley. They contain marked variations in altitude, precipitation and vegetation, as well as in their exploitation in recent centuries by hunters, herders and cultivators. The contrasts are especially sudden and striking at Engaruka, situated at the foot of the eastfacing Rift wall at three degrees south (Fig. 11.1). At an altitude of 1000 m (which is low for this interior region) and with unreliable and variable rainfall, estimated at not more than 400 mm in an average year, it is a relatively hot, dry and dusty place with high evapotranspiration. Despite the attraction of a permanent supply of clear water in the Engaruka river, no cultivators would ever have contemplated settling here by relying on the rain alone for their crops. Immediately behind Engaruka the escarpment rises to 2,000 m above sea level, and towering above that are the Crater Highlands, with the wide dome of Lolmalasin reaching to some 3,500 m. These highlands catch two or three times the rainfall of the Rift floor; their vegetation ranges from montane forests to open grasslands. The latter have in recent times supported wild herbivores and pastoral communities with cattle, sheep and goats. Since the seventeenth or eighteenth centuries AD, these pastoralists have been Maasai. However, the history of pastoralism here stretches back some 3,000 years, during which time successive groups, of which the Maasai are the most recent, have replaced or assimilated those who preceded them (Sutton, 1993). Although these cool highlands have not attracted agricultural settlement, the run-off, which descends the escarpment in deeply cut gorges, has been essential for that of Engaruka at its foot. These gorges of different sizes spectacularly incise the escarpment face at Engaruka along a stretch of 9 km (Figs. 11.2 and 11.3). Nowadays only one of them carries water permanently, and this is the Engaruka river itself (no. 2 on Figure 11.2). This is a fast but shallow stream, usually 3–4 m wide
The archaeology of drylands
196
Figure 11.1 The Rift Valley and Crater Highlands of northern Tanzania, showing Engaruka and related sites
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
197
Figure 11.2 Engaruka and the Rift Valley escarpment, showing the main river (2) and seasonal streams (1, 4 and 5), the area of ancient fields as surveyed (stippled), the artery canals as traced (broken lines) and villages (black circles)
The archaeology of drylands
198
Figure 11.3 The Engaruka escarpment from the east, with the gorge of Engaruka river (central) Photograph: J.E.G.Sutton as it descends the rocky scree; it is easily fordable, except during spates following storms on and behind the escarpment. With its speed compensating for its small dimensions, its discharge into the plain is considerable. The other streams (nos. 1, 4 and 5) flow seasonally or, in the case of certain escarpment gullies and clefts (notably no. 3), open very occasionally after exceptional rain. Of the seasonal streams, the Makuyuni (no. 4) on the northern side is the most reliable, in good years flowing for six to nine months and very occasionally lasting throughout. That at the south end, Olemelepo (no. 1), may carry nearly as much water overall but is extremely temperamental, liable to open in spate and then to fail equally suddenly. The effects of extreme spates occurring at intervals over many millennia (of the Pleistocene presumably, as well as the Holocene) are clear from the sizes of the outwash fans. These consist of soil mixed with water-worn lava boulders of all sizes, which have accumulated immediately below the points where the gorges, of the seasonal as well as the main river, open onto the escarpment foot. As a result, the streams enter the plain at a commanding level on the crests of these fans, the land falling away not only towards the Rift floor but also on either side of the stream beds. This situation continues, in the case of the main river, for a distance of nearly 2 km downstream of the gorge. Not surprisingly, this source of permanent water, with its rapid descent and advantageous level, has been exploited for irrigated agriculture, but at two separate periods. The second of these persists: the present community of Engaruka continues to expand and enlarge its cultivation area. This dates from the 1890s when a few farmers, from different parts of what was then German East Africa, became established a short
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
199
distance down the river (that is, largely below the area of the archaeological field remains of the earlier period), and began cultivating the soft soil with the help of furrows taken off both banks. This community has been reinvigorated by new settlers on occasions, in the 1920s and 1940s, during the British mandateship of Tanganyika, and again in the 1970s, within the Tanzanian ‘villagization’ (ujamaa) movement. The latter involved the incorporation of numbers of Maasai who previously herded in the surrounding plain, so that the character of what used to be called the ‘Swahili’ village of Engaruka has altered. In a series of good years, some of these farmers cultivate in the Makuyuni basin (on the north side) too, by relying on a combination of rainfall and water furrowed from that stream (no. 4) while its flow lasts. Others farm in the Olemelepo basin (to the south) by using a long cross-valley furrow (following close to the line of an ancient one) taken off the Engaruka gorge. Before the 1890s, however, Engaruka was, according to available reports, deserted, except for some pastoral Maasai whose cattle and goats grazed and browsed the sparse pasture of the Rift floor within reach of the river, that being the only permanent water in the district. Information about previous inhabitants gleaned from local Maasai early in the twentieth century is vague and is probably not genuine tradition so much as guesses offered in response to direct questions about the ‘ruins’ (for discussion see: Fosbrooke, 1938; Sassoon, 1966:80–81; Sutton, 1978:67–68). This negative reaction indicates that the place was deserted before the nineteenth century at latest. The recent and existing cultivating community does not appear to be descended in any way from the earlier irrigation farmers who lived here between approximately the fifteenth and the seventeenth centuries. The evidence for that settlement and its fields is exclusively archaeological.
THE ANCIENT FIELDS AND IRRIGATION SYSTEM These earlier fields and irrigation works—which cover some 2,000 ha at the base of the escarpment, around the foothills and into the plain (Fig. 11.2)—are distinguished from the modern ones by their use of stone for dividing and levelling the plots for irrigation (Figs. 11.4 and 11.5) by means of revetments and mild terracing, and for lining and embanking the artery canals (Figs. 11.6 and 11.7) and feeder furrows. Equally distinctive among these ancient fields are two other types of stone features. The first of these consists of numerous square or angular cairns standing up to 2 m high, with rubble cores retained by drystone casings of larger boulders (Fig. 11.8). They are interpreted as stone clearance devices necessitated by the thinness of soil and abundance of surface stone, these conditions being doubtless exacerbated by intensive cultivation with irrigation over a considerable period. Secondly, there are round enclosures up to 10 m across, consisting of thick stone walls
The archaeology of drylands
200
Figure 11.4 Engaruka south fields: stone field divisions and feeder furrow Photograph: J.E.G.Sutton
Figure 11.5 Grid of feeder furrows and levelled field plots below the intermediate north gorge (no. 3 on Figure 11.2) Photograph: J.E.G.Sutton
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
201
Figure 11.6 The support for the ‘great northern’ canal running along the escarpment foot Photograph: J.E.G.Sutton
The archaeology of drylands
202
Figure 11.7 The embanked causeway of the ‘great northern’ canal Photograph: J.E.G.Sutton of similar construction, faced both outside and in, with a narrow entrance gap. Most probably these were built not to contain houses but for cattle. The latter would have been valued for providing manure for the fields as well as milk and meat, and would have needed to be stall-fed, owing to the intense cultivation all around and the lack of pasture in the vicinity for much of the year. This suggestion of stall feeding and manuring at Engaruka is deduced from the evidence for cattle keeping obtained when excavating rubbish deposits in the villages
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
203
(Thorp, 1986)—these villages being situated on the escarpment scree, above the level attainable by channelled water—considered alongside examples of certain recent and existing compact and integrated agricultural
Figure 11.8 An angular cairn (later colonized by termites) in the fields on the south side of Engaruka river, its rubble core revealed by breakage in the faced casing (on right) Photograph: J.E.G.Sutton systems in Africa (Sutton, 1986, 1989). Of relevance also is the archaeological example of the Nyanga terraced fields and connected stone-walled farmsteads with sunken stock-
The archaeology of drylands
204
pens in Zimbabwe (Sutton, 1988; and Soper, this volume, Chapter 12). Similarly, information about the crops that were cultivated on the Engaruka fields is in large part deduced from ethnographic examples of communities currently cultivating, in some cases with irrigation, at comparable altitudes with medium to low rainfall in this interior region of East Africa. Of particular value in this exercise are the Sonjo villages and irrigated basins 100 km or so to the north, close to the Tanzania—Kenya border. Here, sorghum—an ancient African grain, and the principal crop of the savanna regions across the continent throughout the Iron Age—maintains (despite the progressive popularity of maize in the twentieth century) its dominant position, with varieties selected and developed locally both for withstanding droughts and for tolerating waterlogging and irrigation (Adams et al., 1994). Confirmation that sorghum was grown at Engaruka is attested from excavations in the villages, where charred seeds from the hearths and granaries have been recovered. Other crops suggested by the examples of Sonjo and drier parts of the Rift Valley generally would be finger-millet (eleusine) and varieties of pulses. The latter, in rotation with grains, can provide valuable nutrition both for the soil and for the farming community. Finger-millet, while less tolerant than sorghum of heavy irrigation (and perhaps less productive in grain harvested per hectare), has the advantage of ripening on a low rainfall and also of long storage qualities. Probably it would have been sown to take advantage of the rains in the main and, if the crop were successful, stored in the roofs or homestead granaries as reserve against a famine year. It would also have been valued for beer, as would surplus sorghum too, doubtless improved by honey obtained in the forests above Engaruka. Much, but not all, of Engaruka’s ancient field area lies closer to the escarpment than that now settled and cultivated, so that it survives, most unusually, as an expanse of fossilized fields and irrigation devices. Despite the effects of subsequent erosion in places, with gullies damaging and destroying features on cutting through the loose and stony soil, the upper part of the field system is preserved in a remarkably pristine state. Visibility depends on the season and the amount of grass and greenery on the trees and bushes. Moreover, there has been a noticeable increase of thornbush over the last forty years, so that certain photographs—notably Figure 11.5, taken in 1971—cannot be repeated. This vegetational change apparently relates to a reduction of grazing by various wild herbivores and also by Maasai cattle, as well as to a cessation of burning, as the agricultural population of the new Engaruka villages has increased. The lower part of the old field area has been subject to the opposite experience—that of redeposition of soil eroded from the upper part—so that the stone features there tend to be obscured. But the typical field divisions can be seen in gully sides, and the irrigation grid pattern is very clear from air photographs taken in the 1960s before the recent expanse of bush. It is on this relatively level soil with less surface stones that the present inhabitants have chosen to cultivate, avoiding, not surprisingly, the stone-strewn terrain, much of it bereft of soil, closer to the escarpment and on the outwash fan. Stonework was used for dividing and terracing the fields, for lining the canals and feeder furrows, for the stone clearance devices and stock enclosures in the fields, and also for terracing and revetting the numerous homestead platforms in seven large villages that are situated immediately above the top canals (Fig. 11.2). The sheer density of these
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
205
remains over such a wide extent, combined with complete abandonment of this system at least two, more probably three, centuries ago, makes old Engaruka unique as an archaeological field system, and one that can be mapped and studied on the ground. (For more detailed description, see Sutton, 1998; for discussion of particular features, see Sutton, 1978 and 1986; the villages and excavations undertaken in them are further described by Sassoon, 1966 and 1967, and by Robertshaw, 1986.) This intense use of stone, which in older ethno-historical literature of eastern Africa (such as Murdock, 1959) was labelled ‘megalithic’, is, as explained, partly attributable to the ubiquity of the surface gravel and boulders that needed to be moved if one was to put the land to any use. The obvious solution was to utilize these stones in the field divisions and terraces and in the furrow and canal sides, with any remaining excess being piled in the enclosure walls and especially the cairns, which were built as neatly and vertically as possible in order to minimize the waste of cultivable ground. As pressure on resources of both soil and water increased in this isolated situation—one circumscribed by the limits to which irrigated water could be led by gravity—it appears that this commitment to stonework (which is explicable in the first place in functional and environmental terms) developed into a cultural attachment, if not a hallmark, of the old Engaruka community. However, before becoming unduly enthralled by the stone ‘ruins’ of Engaruka and the accidents of survival—‘the tyranny of the monuments’ in Ian Farrington’s phrase—it is encumbent to consider Engaruka in its regional ethnographic and historical context. There is, in fact, nothing very unusual about irrigated agriculture with ‘indigenous roots’ in the precolonial past along the western wall of the Rift in northern Tanzania and Kenya: there are examples from four degrees south, through the celebrated instances of Sonjo (Adams et al., 1994), Baringo (Anderson, 1989) and Marakwet (Hennings, 1951; Soper, 1983; Watson et al., 1998) to two degrees north; or again to the east of the Rift in the highlands of northeastern Tanzania and the Kenya border, notably Pare, Taita, Kilimanjaro and Mount Meru. (For a survey of these remains, see Sutton, 1973, 1984, 1989; Widgren and Sutton, 1999; and see also Widgren, this volume, Chapter 14.) Among these communities numerous varieties of field systems are found, prepared for different crops and combinations and depending in greater or lesser measure on artificial irrigation of the ‘hill furrow’ sort (Adams, 1989), that is by constructing small gravity-fed canals off springs or mountain streams. But, since these present and recent fields are obliged for obvious reasons (the water sources and the basic hydraulics of gravity-fed furrows) to use much the same irrigable land as the older ones, the latter are not recognizable as such on the landscape. Or, rather, where there is a strong suggestion of continuity of settlement and of cultivation dependent on irrigation over a long period, one may, as a historian, have to be content with regarding the ancient and the existing fields as all one. At best, therefore, in these favoured areas of concentrated agricultural settlement, one rarely gains anything more than an impression of the cultivation and irrigation system that operated in the past, and one cannot discern the ancient fields as physical units. Engaruka—for the dual reasons of its being a deserted site and a conspicuous one because of its stonework—is different therefore, and extremely valuable as a research resource, being (with a few minor related sites in the district) a rare example of an archaeological field and irrigation system that can be studied directly on the ground. It
The archaeology of drylands
206
differs from many of the existing East African examples, moreover, in the degree of its dependence on irrigation. Few, if any, of those cited are in quite so arid terrain and some, like those on the southern and eastern slopes of Kilimanjaro and other mountains, receive a high rainfall, adequate to support forest if cultivation is restricted. In a number of these cases, therefore, a fair amount of cultivation is possible without any artificial irrigation, and in places the latter option may be barely activated in average years, necessary though it may be for survival through droughts and bad runs. But the more important point is that irrigation—or the ability to turn to it—has become an essential element in these agricultural systems, because of the success of the latter over time and the size to which the communities have grown. This has necessitated more production per hectare than is afforded by the rain alone (at least in years of low rainfall), and therefore the extension of planting assisted by irrigation devices into the dry season, and the adoption of special crops, varieties or combinations to suit the complex regime. In this way, devices that may at first have been considered optional, or supplementary in difficult years, would in time have become permanent and essential complements to developing agricultural systems and the communities dependent on them. A further factor in the nineteenth century, at particular favoured locations adjoining dry plains such as Taveta, South Pare and Baringo, was the supplying of trading caravans. This required the production of more than a normal surplus, or at least facilities for growing a fast second crop to restock the granaries. Although trade-routes and transport methods have changed in the twentieth century, new opportunities have arisen for production for local and more distant markets. On the slopes of Kilimanjaro and the hills of North Pare, for instance, the production on smallholdings of coffee for export, interplanting it with subsistence food crops, has encouraged farmers to maintain the irrigation systems to ensure watering around the year. And when coffee prices are low, a channelled water supply is still valued for domestic needs in a highly populated rural area, with stalled cattle kept by these means in some locations. Besides its excessively stony terrain, it is doubtless Engaruka’s extreme situation—an impossible one, in fact, in the eyes of most cultivators—that explains the exquisite layout and detail of its fields and irrigation system over so wide an area. Moreover, as argued below, this situation is demonstrably harsher now than it was when Engaruka was first inhabited, which further explains why most of the ancient field area has remained untouched since it was abandoned. The first settlers doubtless began on a small scale with rudimentary irrigation works, probably on the easier terrain some way downstream of the gorge. In time, however, as the community increased in numbers on the success of a cultivation system evolved to handle the peculiarities of the location, it would have been obliged to expand its cultivation area into the more broken and stony ground closer to the escarpment, and eventually as high onto the scree as could be reached by water channelled from the gorges of the Engaruka river and the seasonal streams. The top canals were accordingly led from the highest practicable points in the gorges, that is at the vertices of the outwashes, and carried along the rocky escarpment base at the maximum level attainable while permitting a gravity-induced flow. The actual take-off works on the sides of the stream beds have, of course, not survived the spates since the time of their abandonment, but one must imagine improvised structures of stone and trash, as in existing irrigation systems, requiring annual maintenance if not complete
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
207
rebuilding. The positions of these take-offs can be estimated fairly accurately by tracing the visible upper stretches of the artery canals back to source. The stone linings and embanking of these upper artery canals are preserved, quite spectacularly in some stretches, along the base of the escarpment scree (Fig. 11.6), around the small hills on the edge of the plain, and through the whole area of ancient field remains (Fig. 11.2). Every effort was made to keep these as close to the horizontal as was practicable, for purposes of controlling the flow and preventing undue scouring of the canal beds and breaches in the furrow walls and, equally important, to increase the area of the plain, and of the sides of hills standing in it, which could be reached by the furrowed water. Despite the unavoidable rapid descent or small cascade here and there on the steep and rocky escarpment, for most of their lengths these canals fall at angles less steep than 1:20 and in places as gently as 1:100. The existing examples of Marakwet and Sonjo illustrate how such engineering and levelling perfection can be achieved through a combination of experience and trial-and-error. The longest of these artery canals is that running northwards from the gorge of the Engaruka river: it measures 1–2 m wide between its stone edges, although, since there was no laid bottom, the actual water flow over the gravel and silt bed would doubtless have been narrower. It is traceable up to 3 km from take-off, with divisions at various points. Along the canal’s upper stretch as it descends the escarpment (at a relatively steep angle between 1:15 and 1:20: Adams, 1986), its lower side is substantially supported by gravel embanking. When it reaches the foot of the escarpment scree, it swings at a right angle towards a hill standing separately in the plain. But in order to maintain as much height as possible, the canal is carried across this narrow valley on an embanked causeway (an aqueduct in effect) up to 3 m high (Fig. 11.7). By these means it achieves an advantage when it reaches the hill, where it divides to run as contour furrows constructed round each side. It appears from the levels of the latter that the effect of the embanking was augmented by a wooden scaffolding device to carry the water higher still on hollowed logs. There are other instances of stretches of embanked canal in the field system, with suggestions of wooden superstructures, or at least split and hollowed logs, to carry the water over the porous gravel. These belong to an evolved stage, when the highest artery canals were constructed, and clearly represent an effort to gain the maximum advantage from the available water, regardless of the correspondingly intensive demands this placed on the hydraulic ingenuity of the community and the sheer labour required for construction and constant maintenance of the works. It appears, moreover, that large areas of the fields were relaid to accord with these long embanked canals: this is indicated where series of older field divisions and feeder furrows are superimposed by, or incorporated into, later grids with variant alignments.
AN INTEGRATED AND CIRCUMSCRIBED SYSTEM UNDER STRAIN At some point the limits of feasible improvements and of the community’s technological resource would have been reached. Since it was not possible to increase the area of reliable cultivation beyond that to which water could be carried by gravity through the
The archaeology of drylands
208
highest and longest canals, a crisis must have been faced as the population attained the maximum that this finite amount of land and water could feed, despite all the complexity and ingenuity of the irrigation devices and other specialized elements of this integrated system. In fact these, doubtless together with the operation of the most productive crop rotations available (sorghum and varieties of pulses, in particular), combined with ratooning of the sorghum to obtain a supplementary harvest, and the application of cattle manure to maintain fertility and improve yields, may be seen more as reactions to the limitations of the situation rather than methods devised to achieve agricultural efficiency and increased productivity for their own sake. Equally likely, in so intricate and specialized a set of arrangements, there would have been a danger of trying to intensify too far in reaction to stress: in particular, shortening the fallow would have exacerbated soil-exhaustion and erosion. Indeed, despite all the effort of levelling and terracing to counteract these tendencies, the upper fields became denuded through heavy watering, the field divisions and furrows standing remarkably prominently here, while, as noted, the lower ones are overlaid with redeposited soil, with the old field lines there being visible only in recent gully sides. At the same time, this population was having to contend with hydrological decline, with less water flowing off the escarpment. This process is strikingly illustrated by the existence of canals leading off the gorges of the seasonal and occasional streams, whose flows are now far too inadequate to reward such labour. It is not necessary to conclude from this that those streams were perennial at the time when the canals were constructed, together with the laying out of grids of levelled fields irrigated from them. It is clear, nevertheless, that they must have enjoyed longer flows than now, with sufficient volumes of water in their catchments to ensure their persisting some time after a period of rain. This argument applies especially to Olemelepo (no. 1 on Figure 11.2), where any attempt to reopen the canals that led from that gorge, and to reactivate the archaeological fields that cover its outwash fan, would be pointless now. Even more striking is the case of the intermediate north gorge (no. 3)—a narrow cleft in the escarpment from which water issues in occasional years and then only for a few days following exceptional storms, when irrigation would be least needed. But at the time when the ancient settlement flourished, it was found worthwhile to construct short canals along the foot of the escarpment on either side of this gorge to irrigate a grid of fields (Fig. 11.5) on this small outwash (being too high for watering from the long northbound canal, led through the valley below from the main river). Immediately above those canals were built the two north-most villages—a further indication that there must have been a natural flow from this cleft for at least a few months of the year. Similarly at Makuyuni (no. 4), the positions of the top canals, and the large area of fields on the outwash served by these, seem to require more water and a longer season of reliable flow than obtains now. By implication, too, the volume of the Engaruka river itself would have been greater then, so that, when the declining trend set in, it may have become insufficient to irrigate adequately the whole basin dependent on its canals. Alternatively, as the performances of the seasonal streams and of the fields on their outwash fans became increasingly unreliable, the need would have arisen to channel water as broadly as possible from both sides of the main river. This was effected through the long cross-valley canals with take-offs at the gorge opening that were designed to
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
209
deliver water from the main river into the middle and lower parts of the Olemelepo and Makuyuni basins at seasons when those rivers had dried—projects that required relaying of feeder furrows and field grids, as already noted. From the plan (Fig. 11.2), these alternative sources of water and routes for channelling it may lend an impression of wonderful flexibility. In practice, however, there would usually have been little choice, the complex arrangements being dictated by the sheer necessity of carrying water to as large an area of fields as could possibly be reached from the main river. In years of low rainfall, and therefore heavy dependence on irrigation, this may have exhausted the whole volume of the river’s flow at certain seasons (as can happen nowadays, although the area of existing cultivation downriver is not as extensive as that of the ancient settlements at their prime). This hydrological decline must have been relative because, had Engaruka been much wetter at the time of the first settlement (about the fourteenth century apparently), the need to irrigate, or at least to devise such elaborate arrangements, would not have arisen. That notwithstanding, the archaeological evidence—the configuration of fields, canals and villages—demonstrates clearly enough a change in the performances of the escarpment streams, with their discharges now being definitely less than they were 500 years ago. Equally clearly, these changes occurred, or at least began, during the life of the old settlement—that is by the seventeenth century at latest—suggesting that, together with the strains imposed by the circumscribed situation, they constituted the main factor in the collapse and desertion of Engaruka no later than the eighteenth century. The dating is not perfectly precise, being based on a number of radiocarbon results from excavations in the villages, on the generally late iron age affinities of the pottery and other artefacts, as well as on the lack of any clear local memory about the former inhabitants (Sutton, 1998). It is presumed that the decline in the flows of the escarpment streams—so crucial at Engaruka—was due in large part to a drier climatic trend in the region at large. That implies that there would have been less rain falling at Engaruka itself, as well as in the highlands behind it where the escarpment streams rose, with the effect that dependence on the latter would have been increasing just as their flows were declining. Any such climatic trend ought to be detectable in the archaeological and geomorphological record of the broader region, and especially in lake deposits, although no consistent body of evidence can be cited at the present stage of research. In the broader region of the highlands and Rift Valley, in Kenya as well as northern Tanzania, there are individual instances of springs that are now dry and of late iron age settlements in marginal areas that are now deserted, but these have not yet been dated precisely enough for chronological comparison. Nevertheless, the essential question and the approximate dating are inescapably posed by the Engaruka experience. The further question is whether the decline in the escarpment streams may have been due, in greater or lesser measure, to very local factors, in particular environmental damage caused by a cultivating community of several thousand people over a period of three or more centuries. Wood requirements—for fuel, for fencing the villages, and for house building, and equally for the scaffolding and hollowed logs employed for carrying stretches of the canals—would have placed substantial demands on the forest resources on and above the escarpment. Arguably, such deforestation could have affected surface
The archaeology of drylands
210
moisture and the aquifers in the mountains and therefore the flows of the escarpment streams, rendering them more liable to sudden spates and equally sudden failure. However, the scale of the change suggests that a human factor of this sort can be only partly responsible, and that a decrease in rainfall must have occurred, beginning in about the sixteenth century. The size of that decrease in the highlands need not necessarily have been so substantial, but merely enough to diminish significantly the normal discharges of the streams descending the escarpment and their reliability for regular irrigation at the bottom. Whatever the cause or combination of factors, Engaruka was comprehensively abandoned at some point, around AD 1700 or perhaps somewhat after, on the (admittedly imprecise) dating indicators available. The state of the abandoned field and irrigation system, especially its upper part, and of the numerous homestead platforms in the seven villages immediately above the top canals, lends an impression of sudden desertion rather than slow decline and piecemeal abandonment. That is not easily testable and the impression may be illusory; but it invites one to speculate on other causes of desertion, even catastrophic ones. The possibilities are legion. Among those that have been suggested are a violent earthquake along the Rift fault, arguably upsetting the flows of the mountain streams and the canal take-offs; an unusually heavy eruption of the nearby volcano, Oldonyo Lengai, coating the fields with sulphurous ash; or a devastating attack by expanding Maasai pastoralists (or by Tatoga before them) anxious to secure the water of the Engaruka river and the adjacent grazing. There is no direct evidence to support any of these speculations, and the last would appear unlikely since any pastoral group in the area would have been outnumbered and would also have benefited from exchange of products with an agricultural community in its midst—assuming that the cattle that the latter kept for essential manuring as well as milking did not provoke insuperable jealousy. More likely, the central cause of the collapse of old Engaruka and its highly specialized and integrated irrigation agriculture was inherent in that system, which, being physically circumscribed by the lie of the land and the volume of water in the escarpment streams (the latter, moreover, declining), could not, in the long run, cope with the strains it inevitably generated, in particular its own demographic success. While this general explanation does not rule out other possible contributory factors, it is supported by the existence of several lesser sites in the district, situated similarly by streams (or springs) issuing from the escarpment base, with stone-lined canals and field divisions identical to those of Engaruka (see Figure 11.1). These obviously belonged to the same cultural group and ethnicity, and were presumably abandoned about the same time as was the main settlement and cultivated area at Engaruka (although it is quite possible that some of these outlying sites, on the far side of the Crater Highlands by Lake Eyasi and in the northerly direction above Lake Natron, may have been abandoned earlier or, contrarily, have lingered on a little later). Whatever the exact chronology, this general phenomenon of abandonment of settlements over a radius of some 60 km suggests an inherent and underlying factor rather than a catastrophic event. The fact of desertion, in whatever manner it is to be explained, should not, however, be interpreted as failure in an historical sense. A system so accomplished and complex as that of Engaruka, which evolved over two or three centuries, was surely a story of success in adjusting so effectively to the peculiarities of its own special environment. It is
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
211
more important to understand how it worked and succeeded than to worry about why it collapsed eventually, or again (the common antiquarian reaction to stone ruins) where that population ‘went’. The answer to that last question is that the degree of specialization and the various details of the settlement and its agricultural system had become so specific, culturally as well as functionally, to the situation and community of Engaruka that they could not be transplanted: in other words, this ethnicity would have expired as those villages and their fields had to be abandoned. Remnants presumably took refuge and became assimilated among other peoples of the region, thereby losing their Engaruka identity. One of these may have been Sonjo—a group of compact villages, each situated by a spring or river above Lake Natron, 100 km or so to the north (Adams et al., 1994). It thus forms an ‘island’ of Bantu-speaking cultivators surrounded by Maasai pastures, the latter consisting of poor scrubland in the Rift to the east but also the extensive plateau grasslands of Serengeti to the west. Each of the main existing Sonjo villages—as well as some that have been abandoned for a while—has relied on a basin of irrigated fields. Sorghum has been the principal crop, together with some finger-millet and distinctive varieties of beans. There is also a fair amount of rainfed cultivation in most years, with the normal rainfall being slightly higher than at Engaruka, so that Sonjo’s dependence on its irrigation works is not as extreme as was that of Engaruka. Cattle manure is not applied as fertilizer to the fields, and Sonjo, as far back as reliable information goes, have not kept cattle for fear, it is said, of provoking neighbouring Maasai to raid. (This is to overlook some experiments—and mixed experiences in contending with both cattle diseases and neighbouring Maasai—in the 1970s and 1980s.) There is, moreover, little use of stonework in the field divisions and canal banks, as is so distinctive at Engaruka, the basic reason being the relative paucity of surface stone in the alluvial soils of the Sonjo basins. Stone is, however, used extensively in the concentrated villages of Sonjo, which are situated on hills above the fields, in particular for terracing and revetting the homestead platforms (Fig. 11.9), and also for public areas and some rather peculiar open-air fireplaces. Fairly close parallels for these have been revealed by excavations in the Engaruka villages (Fosbrooke, 1938; Sassoon, 1967). Pending excavations of old Sonjo sites, it is not known how far back these features may be dated there, but it seems likely that some of the Sonjo villages were settled before the collapse of Engaruka. Possibly Sonjo and Engaruka constituted the northern and southern wings respectively of a single cultural group, of which only the one has persisted to the present. In that case, a detailed archaeological study of Sonjo should be revealing about the regional history. This study should compare the existing Sonjo settlements, that is the ‘traditional’ ones destroyed in 1975 during the Tanzanian government’s ‘villagization’ (ujamaa) campaign, and those deserted at unspecified dates in the nineteenth or preceding centuries. Such an exercise should be expected to carry the sequence back towards the time of Engaruka, and thus illustrate the connection. Even if it transpires that the Sonjo people are not related linguistically or in a direct cultural sense to those formerly inhabiting Engaruka (see Nurse and Rottland, 1993), a fuller study of their villages, both existing and deserted, should contribute to an understanding of
The archaeology of drylands
212
Figure 11.9 Sonjo: wooden house with thatch dome on stonerevetted platform-terrace in Oldonyo Sambu village (Kura) Photograph: J.E.G.Sutton specialized—or what are commonly called ‘intensive’ (Sutton, 1984; Widgren, this volume Chapter 14; Widgren and Sutton, 1999)—agricultural practices in isolated situations, as raised by the archaeological record at Engaruka.
ACKNOWLEDGEMENT The author’s continuing field research on the agricultural settlement of East Africa is supported by an emeritus fellowship awarded by the Leverhulme Trust, held at the British Institute in Eastern Africa and the Pitt Rivers Museum in Oxford.
REFERENCES Adams, W.M. (1986) Observations on the Engaruka irrigation furrows and river discharges. Pp. 49–51 in J.E.G.Sutton, The irrigation and manuring of the Engaruka field system. Azania 21:26–51. Adams, W.M. (1989) Definition and development in African indigenous irrigation. Azania 24:21–7. Adams, W.M., Potkanski, T. and Sutton, J.E.G. (1994) Indigenous farmer-managed
Engaruka: farming by irrigation in Maasailand c.AD 1400–1700
213
irrigation in Sonjo. Geographical Journal 160:17–32. Anderson, D.M. (1989) Agriculture and irrigation technology at Lake Baringo in the nineteenth century. Azania 24:84–97. Fosbrooke, H.A. (1938) Rift Valley ruins. Tanganyika Notes and Records 6:58–60. Hennings, R.O. (1951) African Morning. London, Chatto & Windus. Murdock, G.M. (1959) Africa: Its Peoples and their Culture History. New York, McGraw-Hill. Nurse, D. and Rottland, F. (1993) The history of Sonjo and Engaruka: a linguists’ view. Azania 28:1–5. Robertshaw, P. (1986) Engaruka revisited: excavations of 1982. Azania 21:1–26. Sassoon, H. (1966) Engaruka: excavations during 1964. Azania 1:79–99. Sassoon, H. (1967) New views on Engaruka. Journal of African History 8:201–17. Soper, R.C. (1983) A survey of the irrigation systems of the Marakwet. In B.E. Kipkorir, R.C.Soper and J.W.Ssenyonga (eds) Kerio Valley: Past, Present and Future: 75–95. Nairobi, University of Nairobi, Institute of African Studies. Sutton, J.E.G. (1973) The Archaeology of the Western Highlands of Kenya. Nairobi, British Institute in Eastern Africa, Memoir 3. Sutton, J.E.G. (1978) Engaruka and its waters. Azania 13:37–70. Sutton, J.E.G. (1984) Irrigation and soil-conservation in African agricultural history. Journal of African History 25:25–41. Sutton, J.E.G. (1985) Irrigation and terracing in African agricultural history: intensification, specialisation or over-specialisation? In I.S.Farrington (ed.) Prehistoric Intensive Agriculture in the Tropics: 737–64. Oxford, British Archaeological Reports, International Series 232, Volume 2. Sutton, J.E.G. (1986) The irrigation and manuring of the Engaruka field system. Azania 21:26–51. Sutton. J.E.G. (1988) More on the Nyanga terraces: the case for cattle manure. Zimbabwean Prehistory 20:21–4. Sutton, J.E.G. (1989) Towards a history of cultivating the fields. Azania 24:98–112. Sutton, J.E.G. (1993) Becoming Maasailand. In T.Spear and R.Waller (eds) Being Maasai: Ethnicity and Identity in East Africa: 38–60. London, James Currey. Sutton, J.E.G. (1998) Engaruka: irrigation agriculture in the northern Tanzanian Rift Valley before the Maasai era. Azania 33:1–37. Thorp, C. (1986) Engaruka faunal remains. Pp. 21–26 in P.Robertshaw, Engaruka revisited: excavations of 1982. Azania 21:1–26. Watson, E.E., Adams, W.M. and Mutiso, S.K. (1998) Indigenous irrigation, agriculture and development, Marakwet, Kenya. Geographical Journal 164:67–84. Widgren, M. and Sutton, J.E.G. (1999) (eds) Islands of Intensive Agriculture in the East African Rift and Highlands: a 500-year Perspective. Stockholm, Stockholm University, Department of Human Geography, working paper 43.
12 The agricultural landscape of the Nyanga area of Zimbabwe ROBERT SOPER
INTRODUCTION With 750–1,200 m of rainfall a year, Nyanga, in the eastern highlands of Zimbabwe (Fig. 12.1), cannot pretend to be a dryland or even semi-arid environment, but it can be regarded as marginal in some other respects. Furthermore, its well-preserved field systems and evidence for water management practices represent parallel responses to many of the questions addressed in this volume, even if overall aridity was not the primary driving compulsion. The landscape of Nyanga and adjacent areas to the west is indelibly printed with the ‘landesque capital’ remains of past agricultural activities. These take the form of stone-faced terraces and lowland cultivation ridges, together with associated stone-built settlement structures, in all covering around 7,000 km2 (Soper, 1996). Whilst the agricultural features themselves are difficult to date, the settlement sites range from about AD 1400 to 1900, with the earlier sites having no direct association as yet with the agricultural features. The area south of Nyanga town consists of a broad dissected plateau at around 1,800 m above sea level, falling relatively gently to the southwest to the main watershed between the Zambezi and Sabi/Limpopo catchments. To the east it rises to Mount Nyangani at nearly 2,600 m, beyond which are steep mountains and valleys into Mozambique. For about 60 km north of Nyanga, the highlands narrow progressively to a high ridge at around 2,000 m with higher peaks, and with steep escarpments to east and west. To the west of this ridge, granite inselbergs form often substantial hills rising from a base level of around 1,200 m, while dolerite sills and dikes form lesser features. The highland range extends northwards at a lower level for another 20–30 km, while the surrounding lowlands decline to around 900 m. The underlying geology consists of various granites, overlain by sedimentary rocks and dolerites that cap the highlands. Drainage radiates from Mount Nyangani into major rivers such as the Gairezi and Nyangombe to the north-northeast and north-northwest and
The agricultural landscape of the Nyanga area of Zimbabwe
215
Figure 12.1 Location of the Nyanga area, Zimbabwe the Pungwe to the south. Annual rainfall is almost entirely between November and March, with the average ranging from c.750 mm in the northern lowlands to 1,200 mm or more in the highlands. Annual variation may be as much as +/−50 per cent.
THE AGRICULTURAL LANDSCAPE OF NYANGA Terraces Stone-faced terraces cover large areas of the highland escarpments and the slopes of foothills and detached hills and ridges mainly to the west. Some slopes have ranges of up to 100 terraces (Fig. 12.2). The altitudinal range is from about 900 m in the northern lowlands to around 1,700 m on the escarpments and in the highlands, with very little above this level, which is about the upper limit for the cultivation of traditional grain crops at the present day. Study of aerial photographs has identified a minimum area of 22,000 ha of terracing, excluding sporadic outlying occurrences. Distribution favours dolerite soils and rocks. On the geological map sheet covering 2,750 km2, within which the main concentrations of terracing occur (Stocklmayer, 1978),
The archaeology of drylands
216
Figure 12.2 Terraced hillsides in the Nyanga lowlands Photograph: R.Soper over 19,000 ha of terracing have been plotted, of which 42 per cent are on dolerite, 57 per cent on granite and less than 1 per cent on sedimentary rocks (not well represented in this area). However, 26 per cent of the dolerites below 1,675 m are terraced as against only 5.5 per cent of the granites, and most of the latter are adjacent to dolerite ocurrences. The dolerites weather to red clay loams or sandy clay loams of greater fertility than the sandy granite soils, but are heavily leached on the highland plateaux. The younger slope soils have more inherent fertility but are often thin and very stony, so that terracing is necessary to clear the stones and concentrate the soil for cultivation. Terracing also provides fairly level surfaces, protects against erosion and impedes drainage to allow water percolation. Terrace surfaces are generally narrow, commonly between 1.5 and 3 m, except on very gentle slopes where they may be up to 10 m wide. Fall between terraces is normally between 25 and 80 cm, except on the steepest slopes. Slopes of up to 30 degrees were regularly terraced, in some cases up to 40 degrees. Construction varies with geology, topography and the amount of stone to be disposed of, and possibly also with date, though the latter remains to be established. The best terraces have substantial walls around a metre in thickness, with a double facing of large stones and a fill of smaller stones. A low lip is usually present, but the wall may rise a metre or more above the upper terrace surface, where there was a large amount of stone to clear. Such terraces
The agricultural landscape of the Nyanga area of Zimbabwe
217
now have a more or less horizontal profile, which could be the result of soil movement since abandonment. Terraces are not precisely levelled on the contour, allowing for longitudinal drainage, so that it could not have been intended to flood them either artificially or by rainfall. Stone-lined drains carried excess run-off down slope, while in some cases upstanding walls were pierced by drain holes. This type of terrace is generally found on dolerite but occasionally also on granite. The soil is often shallow, from less than 20 cm up to a maximum of 50–60 cm against the lower wall face. It is relatively stoneless, so that it must have been worked over to remove even the smallest stones during construction. The substratum is of densely packed stones in a red clay matrix in the case of dolerite, or more or less decomposed rock on granite. In granite areas with less stone, and on sedimentary argillites in the northern part of the complex, terraces are generally lower and the stonework appears to consist of no more than a simple revetment, while terrace profiles are sloping, with gradients of up to 15 or even 20 degrees being recorded. The only excavated transect showed numerous stones remaining in the soil. This type may represent the rapid exploitation of less favourable but still fertile soils, and it is not known if it is contemporary with the former type. Most of the terraces do not appear to have been irrigated. There are a few cases where old water furrows do traverse ranges of terraces, and they may well have been used for irrigating those below, but no distribution channels have been observed and settlement sites also appear to have been served. In the case of the detached hills to the west, many of which also have extensive terracing, gravity irrigation would not have been feasible. The chronological range of terrace building is uncertain, but probably spanned at least the seventeenth to early ninetenth centuries. Dating and associations are discussed below under landscape development. The only direct radiocarbon date for a terrace (Pta-7601) is 200 +/−50 BP, calibrated to anywhere between 1618 and 1878 at one sigma. This date was obtained from tiny disseminated charcoal fragments in soil of a second phase of terrace construction in a granite area, adjoining a stone enclosure. A total of 537 sherds, mostly small and worn, was also obtained from some 3 m3 of soil. The sherds and charcoal may derive from manuring with domestic refuse from the enclosure, but the latter appears to date from the later nineteenth century and there are some differences in the pottery, so they may well derive from an earlier site, perhaps contemporary with the earlier terrace phase. In either case the date gives only a maximum age, which is not very useful in view of the wide calibration bracket. Cultivation ridges The second notable feature of the old agricultural landscape comprises extensive networks of ridges and ditches on the lower, less stony, slopes below the escarpments and extending some 60 km to the west. No quantification of these has been attempted, but the total area must equal or exceed that of the terracing. In the terminology of Denevan and Turner (1974), these are long, flat-topped, linear ridges. Some, especially in wetter situations, tend to be more cambered due to the greater height and somewhat closer spacing needed for effective drainage. The features are parallel or sub-parallel linear banks, usually 7–10 m wide, between ditches up to a metre or so deep. They often run for
The archaeology of drylands
218
several hundred metres with a more or less shallow longitudinal gradient. These occur both in areas of impeded drainage (termed ‘vleis’) and on the valley sides or interfluves. An example may be described at the base of the main escarpment near Maristvale some 40 km north of Nyanga town. Here there is a broad bay in the escarpment about 2 km wide between high projecting spurs, and a series of streams converges across the piedmont slope. Virtually the whole of the interfluves and most of the stream valleys are scored with ridges and ditches covering around 1,000 ha. The central interfluve (Fig. 12.3) provides an area some 1,750 m long and around 500 m wide with a longitudinal fall of c.60 m and a maximum lateral height of c.10–12 m. Almost all of this is occupied by ridges, except for a stonier crest towards the upper end, which is terraced, and a few minor areas of outcropping rock with stone enclosures. The ridges trend longitudinally down the interfluve with a broadly parallel alignment, sometimes rather braided. A similar pattern is seen on the other interfluves. At the head of this interfluve at the base of the escarpment is a furrow take-off from a small stream. There appears to be no main feeder furrow from this, but water could be directed down any of the ditches, or to the occupation sites. Towards the lower end of the interfluve a furrow did carry water from a deep set of ditches diagonally across the ridges, probably to a stone enclosure on the crest. Soils here are silty sands over a sheet of consolidated rounded quartz gravel. Other occurrences are in more specifically waterlogged areas. An example is a regularly waterlogged perched vlei on the piedmont slope a few km south of the above site. Here there is a dendritic pattern of banks and ditches for maximum drainage, and a section showed a metre or so of mottled sandy clay loam overlying dense black clay. The clay loam must derive by erosion from the terraced area immediately above, perhaps before terracing anchored the soil or perhaps from inefficient use of the terraces. It was then re-exploited after deposition. Ridge size, patterns and orientation to slope appear to vary even within a single localized drainage basin and must represent a flexible system of balancing drainage and water retention under varying conditions of soil, slope, rainfall and seasonal water table. In the first case described, the primary purpose would seem to be controlled drainage raising the cultivation beds above any actual or potential waterlogging without removing rainfall too directly. If necessary, supplementary water could have been introduced to the ditches, though the water available at present would seem
The agricultural landscape of the Nyanga area of Zimbabwe
219
Figure 12.3 Vertical aerial photograph of cultivation ridges crossed by an old trackway and a water furrow Photograph: Harare, Office of the Surveyor General inadequate for any extensive irrigation. In wetter areas drainage could be more direct. Denevan and Turner (1974) review the advantages of raised beds in general. Relevant points here may be: control of erosion; provision of drier cultivation conditions where there is permanent or periodic inundation or waterlogging but with some water retention in the ditches still available to crop roots; wide beds reducing the ditch area; aeration of the soil; and modification of microclimate if there is danger of frost. To these could be added the variation of moisture availability across the ridge and ditch appropriate to different crops. Moisture-loving traditional root plants such as Colocasia (taro) and Zantedeschia (calla lily) would be appropriate for the wetter ditches, while sorghum, millets and legumes could grow on the ridges, with Plectranthus (‘Livingstone potato’) perhaps somewhere in between. The ridging systems remain to be dated, since none of the related stone enclosures mentioned above has been excavated. They cannot be later than nineteenth century and they are different from recent mihomba cultivation ridges, which are shorter, narrower,
The archaeology of drylands
220
straighter and generally restricted to waterlogged environments such as wet stream banks. The intimate relationship to terraces in the Maristvale area suggests contemporaneity with at least some terracing. It seems likely, though, that the large labour demands for constructing and operating both systems simultaneously on a large scale would have been beyond the capacity of individual communities. Water management Water is a critical resource in African agriculture generally, and its management in favourable conditions can provide insurance against bad rainfall years and extended dry periods within a normal wet season, as well as giving the potential to extend the growing season before or after. The possibility of supplementary water supply to some ridge systems has been mentioned above, as has the general lack of evidence for widespread terrace irrigation. Terraces and cultivation ridges, even if not directly irrigated, reflect water management by controlled drainage to provide good infiltration. Permanent streams are common in the Nyanga highlands and descending the escarpments; the potential of these was clearly appreciated because numerous old furrows have been observed, mainly in the highlands where they are better preserved by perennial grass cover. A tentative classification of these furrows can be suggested: 1 small furrows of varying gradient and length associated with occupation sites; 2 generally well-graded furrows on relatively narrow revetted shelves traversing ranges of terraces, probably used for irrigating those below but also often serving occupation sites; 3 furrows assocated with ridging systems; 4 well-graded furrows involving more or less massive earthen banks with potentially irrigable land below, sometimes with recognisable branch furrows or ditches; and 5 furrows without major banks or stone work. Type 1 is the commonest and most widespread in the highlands and would have served domestic requirements, livestock and homestead gardens. Some could be diverted to flush out stone-lined pits used for livestock, and provision was often made to impound the resultant slurry. Only a few cases of Type 2 have been recorded, both in the highlands and on the lower escarpment slopes, while the only case of Type 3 known to date is that described above. Type 4 appears to be restricted to a limited area centred on the northern part of Nyanga National Park and must have been for irrigation of unterraced fields, since no terraces are associated below them and only very rarely are settlement sites served. Type 5 is thought to belong to the colonial period. The others belong at least to the nineteenth century and probably earlier, while some examples of Type 1 are likely to be associated with seventeenth-century sites. Authorship The authorship of the agricultural works can almost certainly be attributed to the ancestors of the present indigenous inhabitants (that is, before the relocation of populations consequent on colonial land policies). These are the Unyama people for the
The agricultural landscape of the Nyanga area of Zimbabwe
221
area north of Nyanga town, the Manyika to the south, and the Maungwe west of the Nyangombe river. Genealogies and traditions of the chiefly families (Beach, 1995) go back well into the eighteenth century at least, and more in the case of the Manyika, and it is surprising that more oral traditions have not survived on the construction and use of terraces and ridges. It would seem that knowledge and use of these specialized agricultural techniques were common to a number of political and dialect groupings in northeastern Zimbabwe and should not be attributed to a single group.
AGRICULTURAL SYSTEMS The agricultural systems of the terrace builders integrated crops and animals. Cattle were almost certainly penned in a sunken stone-lined pit or small stone enclosure within the homestead. In the case of pits, roofed or tunnel-entrance passages would have admitted only dwarf cattle, bones of which have been recovered from the only site with good bone preservation (Plug et al., 1997). The small enclosures in the northern part of the area, however, have open entrances and could have accommodated larger beasts. Pits and internal enclosures in the lowlands are relatively small, with an internal diameter normally around 3 m and a depth or height of around 1.2–1.5 m. Fairly small cattle holdings are thus indicated. Seasonal permanent stall-feeding has been suggested by Sutton (1988), as practised for instance in parts of Nigeria and Ethiopia (Hallpike, 1970; Netting, 1968), but the height/depth rules out any substantial accumulation of manure in situ. Pits in the highlands are larger and deeper: usually 5–9 m in diameter, and 1.80–3 m deep. More cattle are thus indicated above the terrace zone, where the depth could have accommodated the accumulation of manure but was more likely for protection from the cold winds of winter. Goats and possibly sheep were kept in the houses, many of which have a low dividing wall with one half paved with stones (Soper, 1996). Pits and internal enclosures rarely contain any deposits beyond leaf mould and a little silting, and no dung heaps or other substantial middens have been found. Dung was thus regularly removed and must have been used for manure, with some possibly being dried for fuel in the highlands where wood may have been at a premium. Pits were provided with drains and in many cases in the highlands would have been flushed out with water from furrows. Again in the highlands, small dams were often built below the homestead to catch the slurry, or ditches were dug to channel it to small hollows. Many such pit structures have radial walls, which are thought to have sheltered gardens on which the slurry could have been used. Where no furrow was available, as more particularly in the lowlands, dung must have been removed by hand and any flushing have relied on rainwater; goat dung must similarly have been removed by hand from the houses. Domestic refuse was doubtless added to the manure. It is unlikely that there would have been sufficient manure to fertilize the full range of cultivated land. On general ethnographic analogies, one would expect it to have been used on homestead gardens, irrigated where practicable, and on terraced or other plots in the vicinity, but rarely on more outlying fields. Results of phosphate analysis from the archaeological contexts are ambiguous regarding the extent of manuring. Terracing per se could thus be considered a specialized technique, implying only
The archaeology of drylands
222
relative ‘intensification’, but a higher degree of the latter may be postulated in an inner zone around the homesteads, probably dependent on available water supply. The cultivation of outlying terraces, even on the more fertile dolerite soils, would have been less sustainable, and a continuous process of terrace building can be envisaged, with older terraces being fallowed or abandoned as fertility declined. The lack of excavation of settlement sites associated with cultivation ridge systems inhibits any conclusions on their use as yet. The range of crops and cultivation methods might be expected to have varied over the altitudinal and rainfall range of the complex. Summers (1958) identified seeds from Ziwa ruins in the lowlands at around 1,300 m. These comprised mainly traditional grains and legumes, including Sorghum, Pennisetum (bullrush millet), Eleusine (finger millet), Vigna unguiculata (cow pea), Vigna subterranea, Ricinus and perhaps Citrullus; part of a maize cob was also found, but in a surface context. Seeds recovered by flotation in the present research have not added any cultivars to this list. Enquiries about traditional crops add a number of important root crops: Plectranthus esculenta (‘Livingstone potato’), Colocasia (taro) and probably Zantedeschia (calla lily), as well as pumpkins and cucumbers, and several semi-wild fruits, leaf plants and oil-seed plants as well as numerous wild fruits and other plants were also harvested. The traditional varieties of Colocasia and also Zantedeschia are toxic without extended boiling. Traditional cropping practices commonly involved interplanting of grains, legumes and cucurbits. It is probable that outlying terraces were devoted mainly to grain staples, but predation by wild animals and birds could have been a problem. Gardens and in-fields were probably used more for vegetables, roots and legumes; here, a more intimate familiarity with soil depth and quality would have enabled more attention being given to the individual requirements of different plants.
LANDSCAPE DEVELOPMENT The processes and sequence of landscape modification are not yet well understood, but it is unlikely that the terracing and ridging were the work of a large and dense population over a relatively short time period. The Unyama people, within whose territory the greatest concentrations of terracing occur, were not sufficiently important to attract any attention or record from the Portuguese, who interacted closely with the Mutapa state to the west and also with the Manyika immediately to the south. The settlement pattern represents loosely dispersed homesteads in village groupings. Although the stone-built homesteads are very numerous and may be locally concentrated, especially in lowland dolerite areas, none appears to represent prolonged occupation, and there are very few stratified sites or substantial middens. We must therefore see terrace construction as an ongoing process over many generations among the communities of a fairly limited overall population. There is some indication of the reoccupation of homesteads, suggesting that whole settlements and their fields may have been fallowed and resettled, taking advantage of the established capital infrastructure. The limited number of dated sites enables only a tentative interpretation of the process of development of the complex, with some notable lacunae that may be real or only
The agricultural landscape of the Nyanga area of Zimbabwe
223
apparent. Thirty radiocarbon dates are now available, all from the northern area from just south of Nyanga town. All those later than about 400 BP have very wide calibration ranges. The earliest dated stone ruins are all in the highlands. By the fourteenth and fifteenth centuries, relatively extensive sites were occupied on the highest peaks and ridges at altitudes over 2,000 m, followed in the seventeenth century by pit structures now completely ruined at slightly lower altitudes. These sites are all above the level of terracing, so there is no direct association. Later highland pit structures tend to be lower again and are relatively well preserved, some with surviving dhaka (clay) walls. Their construction must have continued well into the nineteenth century, and there may have been a hiatus in highland occupation from the earlier ruined pits, perhaps occasioned by the second severe phase of the ‘Little Ice Age’ (Tyson and Lindesay, 1992). Occupants of these sites must have been responsible for the terracing on the western escarpments. Further south, the banked furrows of the National Park area, with their implied irrigation of unterraced fields, are a local, perhaps relatively late, development, probably also the work of pit-structure occupants living more or less closely above them. In the lowlands, most of the dated sites are stone enclosures within the Ziwa ruins National Monument and range between 140 and 220 BP, calibrating anywhere between the second half of the seventeenth century and the early or even late nineteenth century. Earlier sites may exist here or elsewhere in the lowlands, but have not been dated or perhaps not recognized if not built in stone, so it is not known if there was any occupation contemporary with the earlier highland sites. The extensive terracing of the Ziwa area, with which the stone enclosures are associated, can probably be bracketed between the seventeenth and early nineteenth centuries; most of the western lowlands between the escarpment and the Nyangombe river were depopulated by the end of the nineteenth century, when the first European travellers passed through. Further north there is a different type of homestead design, with small, well-built, central livestock enclosures. Three dates from here are recent, at 100 BP or less, but a couple of dates probably from secondary contexts (including the terrace date quoted above) suggest occupation contemporary with Ziwa. In general, one may suggest a continuous process of terrace construction, with new terraces being built as older ones declined in fertility and were abandoned. Terracing would have concentrated initially on dolerite soils and then spread to adjacent granite areas. Ultimately, terraceable land may have run out and the fertility of homestead plots proved unsustainable, resulting in piecemeal or wholesale removals to new sites. In this way, the impressive agricultural landscape we now see could have been created with a relatively low overall population density. The position of the cultivation ridges in this development is uncertain pending the dating of associated settlement sites. It may be assumed that some wetter lands were always exploited by terrace builders where conveniently available and that ridge practitioners resorted to terracing of suitable land within their ambit, but a large-scale simultaneous use of both terraces and ridges by the same communities seems unlikely in terms of labour requirements. Either lowland practice switched from a concentration on terracing to one on ridging (or vice versa), or each local community emphasized one or the other system according to the type of land available. While little direct research has
The archaeology of drylands
224
been done in areas to the west and southwest, it may be noted that the ridging systems continue to the Rusape/Headlands area but terracing becomes more sporadic, probably concentrated mainly on the limited dolerite occurrences. Each system was probably a parallel exploitation by related communities.
DISCUSSION Although the resources of Nyanga would have been less critical than in the more arid situations that are the focus of most studies in this volume, the tactics of soil and water management show many parallels. The agricultural systems represent a range of specialized responses tailored to the potentialities of the local environment. The relative fertility of the younger slope soils was clearly appreciated and their potential was realized by the development of appropriate terracing technology for stone clearance, soil conservation and control of drainage. The cultivation ridges enabled the exploitation of the more leached and often waterlogged valley soils. The alternative options of ridging and terracing complemented each other and provided a risk strategy for coping with short-term climatic fluctuations, emphasizing either terrace cultivation in wet years or the valley soils in dry years. This might be within a single community where both resources were available, or by reciprocal co-operation between local communities. A similar cooperative relationship may be envisaged between highland and lowland communities, with greater concentration on cattle and cultivation respectively. Water resources were exploited by furrow technology for domestic convenience and garden irrigation to extend the growing season and lessen the effects of dry spells. Integration with livestock management produced manure to extend the fertility span of at least part of the cultivated land and to maximize returns from labour investment. Exotic items are extremely rare or absent, apart from a few glass beads. This and the lack of Portuguese references to the area indicate little participation in trading networks and little differentiation in relative wealth. Production for basic subsistence is thus indicated. Design and construction of homesteads appear to go beyond purely functional necessities, reflecting no great economic stress, while ample storage facilities show an adequate level of food production. Some sites such as ‘forts’ with evidence of regular occupation suggest some degree of local authority but no marked social stratification. Ethnographic parallels for terracing and irrigation in East Africa in general are consistently associated with acephalous, kin-based, social organization, within which the agricultural systems are integrated for land allocation, labour mobilization and the settlement of disputes (Håkansson, 1989). Something similar may be suggested here: the various chiefships within which the complex fell are unlikely to have had any significant function in directing subsistence activities or extracting undue tribute. Terrace building as an on-going piecemeal process is feasible within the labour resources of a family group, perhaps assisted by mutual working parties within the local community. Most of the water furrows would also be within the capabilities of the family, with the exception of Type 4, which must have required community co-operation for the substantial earth movement involved. The stimulus to the very labour-intensive cultivation practices would not seem to
The agricultural landscape of the Nyanga area of Zimbabwe
225
derive directly from serious environmental constraints. While political constraints are uncertain for the earlier centuries, they do not appear to have been particularly pressing for the eighteenth century: defensive structures indicate the need for temporary refuge, probably in response to more or less local raiding, but lowland settlement at least would have been vulnerable to any consistent outside threats. For explanation one may perhaps look more to the opportunities offered by local circumstances, as suggested by Brookfield (1986), whereby innovations adopted for the exploitation of particular niches, in this case the fertility of dolerite slope soils, offered a ‘quantum leap’ in productivity. Although overall population was low, initial relative local density induced by the preference for the dolerite areas provided the necessary labour resources and would have been enhanced by the resultant success. Reasons may be suggested for the decline and abandonment of the systems, but they remain to be tested. Declining fertility in the long term may have reduced the population below a critical level. Drastic drought could have been a factor: for instance, a catastrophic drought occurred in the lower Zambezi and coastal area in the 1820s, though it did not necessarily affect the Nyanga region. In Unyama, at least, persistent struggles for the chiefship between two factions from the late eighteenth century contributed to the depopulation of large areas of the western lowlands by the time of European penetration in the 1890s, but this should not have affected areas in the neighbouring chiefdoms. The systems had already survived the last drier cold phase of the Little Ice Age—may indeed even have been a response to it—and perhaps the subsequent climatic amelioration from the first half of the nineteenth century made them unnecessary.
ACKNOWLEDGEMENTS The research on which this chapter is based was carried out under a joint project of the British Institute in Eastern Africa and the History Department, University of Zimbabwe, in close co-operation with the National Museums and Monuments of Zimbabwe. Gratitude is acknowledged to these institutions and to various agencies of the government of Zimbabwe for facilitating the work. Particular thanks are due to John Sutton for initiating the project and advising on all stages of the research.
REFERENCES Beach, D. (1995) Archaeology and History in Nyanga, Zimbabwe. Harare, University of Zimbabwe, unpublished seminar paper. Brookfield, H.C. (1986) Intensification intensified. Archaeology in Oceania 21, 3:177– 80. Denevan, W. and Turner, B. (1974) Forms, functions, and associations of raised fields in the Old World Tropics. Journal of Tropical Geography 39:24–33. Håkansson, T. (1989) Social and political aspects of intensive agriculture in East Africa: some models from cultural anthropology. Azania 24:12–20. Hallpike, C.R. (1970) Konso agriculture. Journal of Ethiopian Studies 8, 1:31–43.
The archaeology of drylands
226
Netting, R.M. (1968) Hill Farmers of Nigeria: Cultural Ecology of the Kofyar of the Jos Plateau. Seattle, University of Wisconsin Press. Plug, I., Soper, R. and Chirawu, C. (1997) Pits, tunnels and cattle in Nyanga: new light on an old problem. South African Archaeological Bulletin 52, 166:89–94. Soper, R. (1996) The Nyanga terrace complex of eastern Zimbabwe: new investigations. Azania 31:1–35. Stocklmayer, V.R. (1978) The Geology of the Country around Inyanga. Salisbury (Harare), Rhodesian Geological Survey Bulletin 79. Summers, R. (1958) Inyanga: Prehistoric Settlements in Southern Rhodesia. Cambridge, Cambridge University Press. Sutton, J. (1988) More on the cultivation terraces of Nyanga: the case for cattle manure. Zimbabwean Prehistory 20:21–4. Tyson, P.D. and Lindesay, J.A. (1992) The climate of the last 2000 years in southern Africa. The Holocene 2, 3:271–8.
13 Fifteenth-century agropastoral responses to a disequilibrial ecosystem in southeastern Botswana JOHN KINAHAN
INTRODUCTION For centuries, rural communities in many parts of sub-Saharan Africa have relied on dryland cereal cultivation and livestock production, combined according to the limitations of rainfall and soils. Traditional agropastoralism is characterized by its relatively simple technology and high labour demands, with small farming settlements spaced by social and environmental circumstance (Niamir, 1991). This lack of modernization, together with widespread evidence of land degradation, is responsible for the negative perception that has guided successive development plans and conservationist interventions over the last few decades (Leach and Mearns, 1996). The conventional view that subsistence agropastoralism is environmentally destructive has, however, begun to change as fundamental concepts of savannah ecology are reconsidered in the light of new research (e.g. Behnke et al., 1993). It is, for example, no longer accepted that a state of natural equilibrium would exist were it not for the supposed effects of agropastoral settlement (Lamprey, 1983; Sinclair and Fryxell, 1985). On the contrary, the fact that dryland environments are prone to climatic variability in the form of unpredictable rainfall events better explains the vicissitudes of agropastoral production (Ellis and Swift, 1988; Nicholson, 1996; Rasmussen, 1985:119). Very high livestock densities are required to effect significant vegetation change under these conditions, even under sustained drought conditions (Pratt and Gwynne, 1977). Given the difficulty of maintaining high numbers on insecure resources, together with the labour demands of herding and cereal cultivation, dryland agropastoralism should have relatively little long-term environmental impact. The fact that such impacts do occur, to the extent that drylands are visibly altered as a result, suggests that the long-term environmental consequences of agropastoral production are not yet fully understood. Our unfortunate lack of detailed historical information on African drylands (Little, 1996) is due, in part, to the fact that equilibrium models relied on environmental indicators to estimate the degree of disturbance in a given system (Behnke and Scoones, 1993; Scoones, 1996). There is now an increasing interest in the temporal persistence of what were hitherto considered intrinsic vegetation features and soil conditions (Fairhead and Leach, 1996; Frost et al., 1986; Hoffman, 1997). Indeed, the notion of a simple disequilibrial relationship between agropastoralism and dryland environments has itself attracted criticism, partly for its failure to explain the differential impact of seasonal land
The archaeology of drylands
228
use practices (Illius and O’Connor, 1999). The adoption of disequilibrium perspectives (cf. Behnke and Scoons, 1993) by social scientists concerned with the ecology of subsistence farming in Africa (e.g. Powell, 1998; Sullivan, 1996) may have drawn attention away from the long-term impacts of subsistence farming in Africa by emphasizing the apparent sustainability of such systems (Mortimore, 1998). Nonetheless, there is a clear need for time series data, although it is acknowledged that direct monitoring and experimental simulation are not always practicable, especially in view of the certainty that some environmental processes would operate on the scale of decades, if not centuries (Coppock, 1993). In the circumstances, it is not surprising that the potential value of archaeological evidence has been raised in discussions of sustainable dryland management (Blackmore et al., 1990; Dennell, 1982; Leach and Mearns, 1996:5; Rapp, 1985:110; Stiles, 1995:16). My purpose here is to review the archaeological evidence of agropastoral settlement in one particular environment, that of southeastern Botswana, and to apply to it some of the more recent findings and concepts in dryland ecology. In doing so I hope to show that archaeological research in dryland environments could, by adopting a broader approach, make a useful contribution to contemporary issues such as land degradation. I also hope to alert environmental scientists to some of the major limitations of the archaeological record and the tenuous nature of inferences concerning past land use practices. The first of the following two sections sketches the archaeological and environmental characteristics of southeastern Botswana, and the second gives an outline of results from the excavation of a fifteenth-century AD Khami period settlement. I conclude with a discussion of some general implications of the archaeological evidence for dryland environmental history.
AGROPASTORALISM IN THE MOTLOUTSE RIVER ENVIRONMENT The Letsibogo area in southeastern Botswana (Fig. 13.1) lies more or less in the centre of the Shashe—Limpopo basin, which is an environment characterized by dry savannah woodland in a generally subdued terrain with well-developed drainage (Thomas and Shaw, 1991). The area is bisected by the Motloutse (tlou=elephant)—a major episodic river course with a narrow fringe of riparian bush on either bank and dependable supplies of water in
Fifteenth-century agropastoral responses
229
Figure 13.1 The regional setting of Letsibogo in the Shashe— Limpopo basin of southeastern Botswana numerous shallow wells. The present local population of up to ten persons per hectare (Campbell, 1990) is scattered among farmsteads and cattleposts, with the one large village, Mmadinare, having been established as recently as 1912 by Mphoeng, brother of
The archaeology of drylands
230
the Bamangwato paramount (Campbell et al., 1996). With an average annual precipitation of 350 mm or less, the agricultural potential of Letsibogo is low, for the minimum requirements of maize, sorghum and millet (FAO, 1978) are not met every year. Rainfed cultivation of cereals is nonetheless an important, if risky, component of the subsistence economy, together with well-established vegetable gardens at many farmsteads. In addition, present-day farming emphasizes a combination of small stock and cattle, which is appropriate to the dense bush conditions with their abundant browse and sparse perennial grass cover (Abel, 1993). Letsibogo clearly exemplifies a disequilibrial ecosystem in the sense of Ellis and Swift (1988), where a low and erratic rainfall induces wide fluctuations in primary productivity and livestock numbers, leading to the adoption of highly opportunistic land use practices (Behnke and Scoones, 1993:11; Westoby et al., 1989). However, the evidence of land cleared for cultivation, as well as advanced soil erosion and the encroachment of dense thornbrush, show that unreliable rainfall is not in itself an effective limitation on the impacts of agropastoralism at Letsibogo (cf. Illius and O’Connor, 1999). The present combination of marginal farming conditions and relatively high population density has not always existed on the Motloutse. Recent research points to apparent correspondences between climatic perturbations over the last 2,000 years and both the distribution and intensity of agropastoral settlement in the Shashe—Limpopo basin (Huffman, 1996a). The relevant archaeological evidence for pre-colonial farming in Botswana is discussed in detail by van Waarden (1999). Here it is sufficient to state that, after the initial appearance of Zhizo farming settlement early in the first millennium AD, a more complex pattern arose in about AD 600, with an apparent hierarchy indicated by the varying extent of dung deposits in areas of livestock enclosure (Denbow, 1986). These Toutswe chiefdoms formed part of large regional entities with a high level of social complexity, as is evident from the rise of major centres like Mapungubwe in the Limpopo valley (Hall, 1987). The arid conditions that affected much of southern Africa towards the end of the first millennium AD seem to have been less severe in the Shashe-Limpopo basin, where the density of farming settlement remained relatively high (Whitelaw, 1997:448). Huffman (1996a) argues that until about AD 1300, the end of the ‘Medieval Warm Epoch’ (cf. Tyson and Lindsay, 1992), annual precipitation would have had to be at least 150 mm higher than at present to permit cultivation of sorghum, which was a major staple at that time. The decrease in rainfall after AD 1300 therefore inevitably led to the abandonment of the capital at Mapungubwe in the Limpopo valley. As the limits of productive agriculture retreated to the north, a powerful new centre arose at Great Zimbabwe (Huffman, 1996b). Under these conditions, the Shashe—Limpopo basin would have been largely deserted, at least by agropastoralists. However, by AD 1450 the Zimbabwe empire collapsed and broke in two, with one of the new entities centred further west at Khami, possibly in response to a slight climatic amelioration, which in turn allowed some reoccupation of the Shashe-Limpopo basin (Huffman, 1996b). As conditions improved, the first Sotho— Tswana people, known archaeologically as Moloko (Evers, 1984), spread from the South African interior to establish themselves in the Shashe-Limpopo basin (Maggs and Whitelaw, 1991; van Waarden, 1989; Whitelaw, 1997). Although these successive and
Fifteenth-century agropastoral responses
231
contemporaneous cultural traditions had recognizably distinct ceramic assemblages (Huffman, 1980; Phillipson, 1977), all were patrilineal agropastoral economies with a common Southern Bantu social organization (Huffman, 1996b; Kuper, 1982). Until recently, the archaeology of Letsibogo was unexplored and little was known of the relationship between major pre-colonial centres and this somewhat remote and marginal area. Detailed surveys and test excavations in the vicinity of the MotloutseSedibe confluence near Mmadinare (Campbell et al., 1995) have revealed widespread agropastoral occupation in the first millennium AD, with the evidence suggesting a pattern of short-term shifting cultivation involving localized groups of small homesteads clustered around rocky outcrops. These indications of Zhizo farming settlement tend to be highly visible due to their exposure by severe sheet erosion and the development of deep gullies accompanied by significant root exposure over large areas of woodland, especially in the deep sandy loam soils close to the hills (Kinahan, 1999). Present evidence from Letsibogo indicates that Zhizo settlement was abruptly curtailed at the end of the first millennium AD and that occupation of the area only resumed almost 400 years later with a rapid influx of Khami settlement. The disparity between this evidence of local settlement, summarized in Table 13.1, and the regional pattern outlined by Whitelaw (1997), may be due to the relatively marginal situation of Letsibogo itself, although this has to be confirmed. Nonetheless, oral tradition identifies the Motloutse as the southern boundary of the Khami state known as Butua (Campbell et al., 1996), implying that this could have been the environmental limit for
Table 13.1 Selected radiocarbon measurements from Letsibogo Site no. Lab. no. AD cal. C14 yrs BP Moloko 79a 2 127 26 Khami 86 4 79b 125 125 Zhizo 106 109 30a 19
Range 1∂
Beta-80094 Beta-80092 Beta-81224 Beta-81225
400+/−70 360+/−70 360+/−70 280+/−70
1505, 1595, 1620 1530, 1545, 1635 1530, 1545, 1635 1660
1450–1645 1475–1655 1475–1655 1640–1680
Beta-80983 Beta-80979 Beta-80982 Beta-80986 Pta-7774
550+/−70 480+/−60 450+/−50 710+/−60 520+/−40
1425 1445 1460 1300 1434
1400–1425 1425–1485 1440–1505 1285–1325 1421–1447
Beta-80095 Beta-80984 Beta-81196 Beta-29951
1360+/−50 1220+/−60 1220+/−50 1100+/−50
685 880 879 981
665–770 785–960 790–905 899–1013
a subsistence economy dependent on both rainfed cereal cultivation and livestock husbandry. On the other hand, the evidence for an influx of Moloko settlement from
The archaeology of drylands
232
southern Africa after AD 1500 (Campbell et al., 1996) suggests that there may have been major political developments from which it is not possible to separate the environmental conditions of farming settlement. Khami pottery was found at more than fifty Letsibogo sites, but only ten of these clearly showed the settlement layout described by van Waarden (1989), with the stone granary supports arranged in a wide arc, open to the west and to the rear of the huts, which faced inward to the site of the cattle enclosure. Many of the Khami sites were found in localities with dense thornbush encroachment, and although this may have negatively influenced the survey results, the pattern of site distribution is suggestive. The settlements vary in size, although in terms of granary numbers there are only two general classes: those with twenty or less and those with more than forty. Although none of the sites exhibited stone walling consistent with elite status (Huffman and Hanisch, 1987), these disparities in size could indicate some functional differentiation among commoner settlements. A hierarchical clustering of the ten selected sites using Ward’s minimum variance of distance method (JMP, 1995:330–1) identifies three groups with roughly equidistant centres. All three central sites are from among those with larger numbers of granaries, as well as being the only Khami sites with confirmed cattle enclosures. The dating of the sites in Table 13.1 suggests that they could have formed a contemporaneous group. The linkages of the sites, together with a hypothetical farming settlement model, are shown in Figure 13.2.
ARCHAEOLOGICAL EVIDENCE FROM A KHAMI PERIOD VILLAGE Detailed information is available from the northernmost of the three central settlements at Letsibogo, Site 125. A radiocarbon date of AD 1434 (Pta-7774), with a 1∂ range of AD 1421–47, places the occupation of the site at the beginning of the Khami period (AD 1450–1800). The earlier date (Table 13.1) reported by Campbell et al. (1995) is not reliably associated with the evidence discussed here. The site, which measures c.500 m2, is situated 600 m from the Motloutse, on the western slopes of a low rocky ridge, with good access to water, arable soil, building timber and grazing. Although the immediate vicinity is deeply dissected by erosion gullies, the site itself shows little evidence of erosion. The surroundings of the site were thickly overgrown with thornbush, with the only archaeological indications at the surface comprising an arc of granary supports and a lobate area of soil discoloured by ashed dung. Excavation (Fig. 13.3) revealed a substantial dung deposit and yielded quantities of bone fragments of small stock and cattle, as well as abundant pottery and evidence of both metallurgy and cotton spinning.
Fifteenth-century agropastoral responses
233
Figure 13.2 The distribution (above) and linkage (below) of Khami period sites at Letsibogo, according to Ward’s minimim variance of distance method Key: Solid circle—settlement thought to be a focus of a settlement cluster; open circle—satellite site
The archaeology of drylands
234
Figure 13.3 Plan (above) and section (below) of Letsibogo Site 125
Fifteenth-century agropastoral responses
235
The area between the granaries and the stock enclosure, where the wattle and daub (dhaka) huts would have stood (cf. Kinahan et al., 1998; van Waarden, 1989), was generally poor in archaeological materials and, unexpectedly, contained no substantial hut remains. However, detailed granulometric analyses showed that, whereas soil from the perimeter of the site and the surrounding area contained very little fine sand and claysized material (fraction less than 60 µm), soil from the putative hut area had the same particle size distribution as freshly puddled dhaka (Kinahan, 1999)—a material the villagers of Mmadinare customarily obtain from old termitaries. Apparently, natural disaggregation of the hut structures has created a sealed datum surface in an area where sheet erosion has, over the intervening centuries, both lowered the surface and removed the lighter soil fraction. Soil nutrient analyses strongly confirmed these observations on the layout of the site (Fig. 13.4). Samples from a transect through the site showed high phosphate concentrations only in the area of the stock enclosure. A steep decrease in phosphate concentrations at the downslope edge of the discoloured soil area suggests that animal dung was retained by means of a palisade fence, although there is no surviving trace of the structure itself. By comparison, soil nitrogen levels are higher in the area outside the stock enclosure, possibly representing an accumulation from the relative concentration of nitrogen in building timber in the huts and fuelwood consumption in cooking fires. The apparent contrast between the hut area and the stock enclosure would be partly due to the volatility of nitrogen in dung, as well as the concentration of phosphorus as a result of burning. Excavations yielded almost twice as much animal bone from the granary area as from the stock enclosure, and very little from the hut area (Table 13.2). Whereas some wild species were represented in the huts and granaries, the bone from the stock enclosure was exclusively of either confirmed or probable sheep/goat and cattle. Cranial bones of domestic livestock were recovered from all parts of the site, but those from the stock enclosure were more fragmented and consisted mainly of loose teeth. A clear contrast could be seen in the distribution of post-cranial bone, with the greatest amount and range of skeletal parts, including small terminal limb elements, being found among the granaries. This suggests that the granary area rather than the stock enclosure was the main focus of domestic animal butchery and the disposal of bone, although the slaughter of stock was in all likelihood carried out somewhere on the perimeter of the settlement. Among the few cattle bones from the hut area was a scapula fragment bearing puncture marks attributable to the canine teeth of a domestic dog. It is conceivable that significant amounts of bone were redistributed in this way. Other evidence of post-depositional processes acting on the animal bone sample includes rodent gnawing (cf. Brain, 1981) and rootlet etchmarks, the latter mainly affecting bone in the granary area. The fact that soil in the granary area was mildly acidic compared with that of the stock enclosure may explain this difference (Fisher, 1995). Microscopic examination of soil
The archaeology of drylands
236
Figure 13.4 Distribution of soil nutrient values at Letsibogo Site 125
Fifteenth-century agropastoral responses
237
Table 13.2 Faunal taxa from Letsibogo Site 125 Fauna Granaries Huts Stores Totals Reptilia Unid. snake 1/1 1/1 Tortoise (cf. Geochelone) 2/1 2/1 Aves Unid. gamebird 1/1 1/1 Mammalia Unid. rodent 1/1 1/1 Hare (cf. Lepus) 5/1 5/1 Procavia capensis 6/2 2/1 8/3 Unid. Bovid size class I 1/1 1/1 2/2 Sheep/goat (Ovis aries/Capra hircus) 3/2 2/1 3/1 8/4 Unid. Bovid size class II 5/1 2/1 4/1 11/3 Cattle (Bos taurus) 9/1 2/1 2/1 13/3 Unid. Bovid size class IV 38/1 3/1 18/1 59/1 Note: The data are listed as NISP/MNI (number of identifiable specimens/minimum number of individuals) after Klein and Cruz-Uribe (1984); bovid size classes are after Brain (1974). samples from the stock enclosure, following the methods of Brochier et al. (1992), revealed evidence of fibro-radial spherulites only in the small northeastern lobe of the deposit (Fig. 13.3). Since spherulites are produced in the digestive tract of sheep rather than goats, and not at all in cattle, this confirms the presence of sheep on the site and indicates that livestock was segregated within a single enclosure complex. The distribution of pottery on the site paralleled that of food remains, with forty-five of the fifty-two vessels being found in the granary area. Most of these were high-necked jars, such as would have been used for fetching and carrying water from nearby wells, and globular cooking vessels. Utilitarian vessels of this kind probably dominate the pottery assemblage because they were subject to frequent breakage. Very large storage vessels, which were probably never moved from beneath the granaries or eaves of the huts, and bowls that would have been used only in and around the huts, make up a very small part of the assemblage. Pottery was most abundant around the midpoint of the granary area (cf. Fig. 13.3), that is to say at the highest and hindmost part of the site. This, according to the conventional layout, would have been occupied by the most senior man of the village. It is therefore significant that evidence of metallurgy, in the form of ore, slag and tuyère fragments, was most strongly associated ‘with this area, as were all finds of clay spindle whorls, since cotton spinning was the traditional preserve of men in Khami society (van Waarden, 1989). The archaeology of Site 125 at Letsibogo provides several important insights into Khami period settlement in the Motloutse River (Fig. 13.2). The site forms the centre of a settlement cluster, the study area as a whole having three such clusters, which were probably coeval. They represent a land use strategy that combined animal husbandry and
The archaeology of drylands
238
cereal cultivation. There can be no doubt that these activities formed the mainstay of the economy, for in the case of the cereal crops numerous storage granaries were required, and the substantial dung deposits must be the result of keeping domestic herds. The positioning of these major components of the site, as well as the distribution of small finds, conforms to the structural principles of the Southern Bantu settlement pattern (Huffman, 1996b; Kinahan et al., 1998; van Waarden, 1989). It is indeed apparent that an archaeological sampling strategy that was not informed by these principles could yield biased and perhaps misleading results. There are, nonetheless, considerable shortcomings in the archaeological evidence. In the first instance, there is no indication as to the length of occupation, and the number of inhabitants is not established. Although sorghum is likely to have been the main staple crop, the species of grain cultivated by this community is not known, and neither is the type of garden vegetables, which would almost certainly have formed part of the diet. Although no direct botanical evidence was found, wild plant foods were probably important here on the analogy of recent studies in Zimbabwe (Jonsson, 1998). The stock enclosure confirms the social importance of cattle, and the animal bone establishes the presence of small stock, but this evidence does not provide any means to estimate the size of the herds or the dietary importance of cattle as opposed to small stock, or wild species for that matter. Wild fauna may have been more important than the evidence suggests, especially if game was butchered and eaten away from the settlement. Finally, there is no evidence beyond that of cultural affinity to reflect on the nature of the relationship between this and other Khami period sites at Letsibogo and further afield. These are important limitations on the extent to which archaeological evidence can usefully contribute to dryland environmental history.
HUMAN IMPACTS AND DRYLAND ENVIRONMENTAL HISTORY In general, there is a satisfactory match between the archaeological data from Letsibogo and the palaeoclimatic model of Tyson and Lindsay (1992). With the added precision of Huffman’s (1996a) calibrated radiocarbon dates, a good correspondence is achieved between the onset of the AD 1425–1675 period of increased rainfall and the first appearance of Khami settlement at Letsibogo. However, the span of the calibrated Khami dates is but thirty-five years, with a maximum 1∂ range of 100 years (Table 13.1). Taken together with the hiatus of about 400 years following the curtailment of the earlier Zhizo period, this points to more variable conditions than the available palaeoclimatic data reveal. It would appear that within the Shashe-Limpopo basin, described as a ‘rainfall trough’ by Jackson (1961), Letsibogo is relatively marginal as far as farming conditions are concerned. By themselves, the radiocarbon dates from Letsibogo suggest that the Khami presence was an opportunistic response to a short-lived climatic amelioration in the Shashe—Limpopo. On this basis it may be argued that the short span of Khami settlement is a predictable consequence of the inverse relationship between variability of rainfall and long-term low average annual rainfall (Nicholson, 1996). Local rainfall anomalies are not unusual in these conditions, and it is possible that the Khami
Fifteenth-century agropastoral responses
239
occupation at Letsibogo represents a short-lived expansion at the margins of the agropastoral environment. In situations such as at Letsibogo, where rainfall is characterized by its variation rather than its average, models based on averages (such as that of Bryson and Bryson, 1996) will not reflect the short-term oscillations on which agropastoralism depends, and for this reason they will be less useful than in regions where relatively mesic conditions prevail. Archaeological proxy data can therefore help to indicate temporal variations that lie beyond the resolution limits of climatic models. The gradient from highly variable, low annual average, to less variable, high annual average rainfall effectively separates non-equilibrial, event-driven ecosystems from more stable equilibrial ecosystems (Behnke and Scoones, 1993; Frost et al., 1986). To Coppock (1993), these conditions produce functionally different ecosystems, with the more sustained impact on equilibrial systems resulting in potentially more rapid degradation. The non-equilibrial systems may, of course, be equally vulnerable if there is insufficient recovery time between episodes of impact. For purposes of agropastoral settlement, the threshold between equilibrial and non-equilibrial ecosystem dynamics probably lies at an average annual rainfall of around 350 mm, as now prevails at Letsibogo. Under such conditions, droughts are more frequent and severe, although they are interspersed by periods of above average rainfall that may extend over several years (Nicholson, 1996). Evidence of cereal cultivation at Letsibogo therefore does not necessarily imply a higher average annual rainfall (pace Huffman, 1996a). If the Khami occupation of Letsibogo may be assumed to represent an opportunistic, event-driven, episode, it is necessary to consider the extent to which the impact of this settlement has shaped the environment as it appears today. The immediate effects of clearing, tilling and weeding fields, together with those of livestock impact in the near vicinity of settlements, would have been highly visible but short-lived, as they are today. More persistent would have been the effects of soil nutrient redistribution and the creation of a patchy vegetation mosaic reflecting differential pressures of usage, on the one hand, and favourable germination and regeneration conditions, on the other. Both Coppock (1993) and Hoben (1996) have pointed to the effect of heavy grazing and the concentration of nutrients in the dung deposits of stock enclosures. Indeed, colonization of these deposits by lime-tolerant Cenchrus ciliaris grass is a notable characteristic of ancient stock enclosures in Botswana and is clearly visible on aerial photographs Denbow, 1979). In these environments, cattle are attracted to the pioneer grasses at abandoned settlements and thus play an important role in maintaining such open pastures (Homewood, 1992). At the same time, the nutrient status of areas immediately adjacent to settlements is lowered by high grazing and browsing pressure (Botkin et al., 1981), which tends to exacerbate the patchiness resulting from nutrient concentration in the stock enclosures. Coppock (1993:56) has remarked on the markedly higher fertility of soils in bush-encroached areas, demonstrating the beneficial effects of an encroachment phase following a period of heavy livestock utilization. Similar observations were reported by Reid and Ellis (1995), who recorded higher nutrient levels in the vicinity of abandoned pastoral encampments and thornbush seed density up to eighty-five times higher than the norm. Very dense stands of thornbush may also become established on abandoned stock enclosures through the germination of seed in, especially, goat dung, resulting in
The archaeology of drylands
240
characteristic age cohort patches (Kiyiapi, 1994). At this stage, the duration of the encroachment phase is not known, other than from anecdotal evidence (Kempff, 1994), although it does appear to extend beyond documented events and into the archaeological record. Aerial photographs of the Letsibogo area clearly show patchy thornbush cover in the vicinity of the Khami settlements. On the ground, these patches often include scattered specimens of Boscia albitrunca—a species that would have been conserved for its veterinary medicinal properties (Coates-Palgrave, 1981:187). Observable correspondences between the Khami site distribution and the physiognomic characteristics of the vegetation at Letsibogo suggest that the impact of agropastoralism under disequilibrial conditions has long-term consequences. Other studies indicate that changes in soil chemistry and vegetation are highly persistent (Blackmore et al., 1990; van der Koppel et al., 1997; Skarpe, 1991; Turner, 1998). Although it is possible that changes in the vegetation at Letsibogo were initiated at an earlier stage that the Khami period, the first-millennium Zhizo settlement pattern has different locational characteristics. Nonetheless, it is important to consider the effect of more recent land use practices: in the case of the severe erosion visible on the Zhizo sites, this is in some instances attributable to the development of gully systems on cattle paths originating in the modern village of Mmadinare (Kinahan, 1999). A recent contribution to the range ecology debate by Illius and O’Connor (1999) argues that disequilibrial dynamics would govern that part of the land use strategy in which livestock grazing was limited by the availability of water, whereas that in which livestock were limited by the availability of food would be subject to density-dependent, or equilibrial, dynamics. In this view, as suggested by Behnke and Scoones (1993), agropastoral impact would be minimal only in the area of rainy season grazing, while key resource areas used in the dry season would register greater impact. At the regional scale of rainfall distribution, domestic crop requirements and vegetation dynamics, Letsibogo therefore presents the characteristcis of a disequilibrial system. However, the evident long-term impacts of agropastoral settlement suggest that equilibrial dynamics would have placed definite limits on livestock numbers, even if a cattlepost system such as that of the modern Tswana (Shaw, 1974:88) was employed to lessen the degradation of key resource areas. Archaeological evidence may be highly relevant to the refinement and testing of soil loss estimates in such environments. As Biot (1993) has shown, field-based estimates of soil loss in eastern Botswana indicate that present stocking rates could be sustained for the next four centuries, in contrast with a more alarmist view that radical destocking should commence immediately to avoid irreversible land degradation. Securely dated archaeological settlement patterns, integrated with vegetation distribution, density and age cohort estimates, would provide essential baseline data for modelling recent environmental changes. The precision of such data is unavoidably problematic, but when there are widely variant competing estimates, as discussed by Biot (1993), the archaeological data could greatly reduce the uncertainty involved. Environmental scientists should take note, however, that there are several pitfalls in the application of archaeological evidence relating to agropastoral land use in Africa, two of which I should describe in conclusion. Archaeologists often draw broad regional
Fifteenth-century agropastoral responses
241
inferences from very limited, even ambiguous, field data, and this may easily conceal local variation, which is the essential basis of a particular land use strategy. Large data bases are uncommon, due to the time-consuming nature of archaeological sampling, and while archaeological observations are testable in a broad sense, they are not repeatable in the narrow sense employed by most natural scientists (cf. Hempel, 1966). This leads to the second pitfall: that of using archaeological data as if they were neutral observations. The Letsibogo evidence very clearly illustrates the social context of nearly all material aspects of southern Bantu settlement. It would be regrettable if, in the need to consider historical evidence, environmental scientists neglected to consider the social dimensions of dryland agropastoralism in Africa.
ACKNOWLEDGEMENTS Site 125 was found, mapped and tested by C. van Waarden in an initial phase of work at Letsibogo, reported in Campbell et al. (1995). I am indebted to A.C.Campbell, T.Hoffmann, A.Illius, A.Reid and C.van Waarden for critical comments on the manuscript. The excavations reported here were commissioned and funded by the Botswana Government Department of Water Affairs, to whom I am grateful for permission to publish this research.
REFERENCES Abel, N.O.J. (1993) Reducing cattle numbers on southern African range: is it worth it? In R.H.Behnke, I.Scoones and C.Kerven (eds) Range Ecology at Disequilibrium: New Models of Natural Variability and Pastoral Adaptation in African Savannas: 173–95. London, Overseas Development Institute. Behnke, R.H. and Scoones, I. (1993) Rethinking range ecology: implications for rangeland management in Africa. In H.Behnke, I. Scoones and C.Kerven (eds) Range Ecology at Disequilibrium: New Models of Natural Variability and Pastoral Adaptation in African Savannas: 1–30. London, Overseas Development Institute. Behnke, R.H., Scoones, I. and Kerven, C. (1993) (eds) Range Ecology at Disequilibrium: New Models of Natural Variability and Pastoral Adaptation in African Savannas. London, Overseas Development Institute. Biot, Y. (1993) How long can high stocking densities be sustained? In H.Behnke, I.Scoones and C.Kerven (eds) Range Ecology at Disequilibrium: New Models of Natural Variability and Pastoral Adaptation in African Savannas: 153–72. London, Overseas Development Institute. Blackmore, A.C., Mentis, M.T. and Scholes, R.J. (1990) The origin and extent of nutrient-enriched patches within a nutrient-poor savanna in South Africa. Journal of Biogeography 17:463–70. Botkin, D.B., Mellilo, J.M. and Wu, L.S.Y. (1981) How ecosystem processes are linked to large mammal population dynamics. In C.W.Fowler and T.D.Smith (eds) Dynamics of Large Mammal Populations: 18–34. New York, John Wiley.
The archaeology of drylands
242
Brain, C.K. (1974) Some suggested procedures in the analysis of bone accumulations from southern African Quaternary sites. Annals of the Transvaal Museum 29 (1):1–8. Brain, C.K. (1981) The Hunters or the Hunted? An Introduction to African Cave Taphonomy. Chicago, University of Chicago Press. Brochier, J.E., Villa, P. and Giacomarra, M. (1992) Shepherds and sediments: geoethnoarchaeology of pastoral sites. Journal of Anthropological Archaeology 11: 47–102. Bryson, R.A. and Bryson, R.U. (1996) Site-specific high-resolution archaeoclimatic modelling for Africa. Unpublished conference paper, Poznan, Society of Africanist Archaeologists, September 1996. Campbell, A.C. (1990) The Nature of Botswana: a Guide to Conservation and Development. Gland, IUCN. Campbell, A.C., Kinahan, J. and van Waarden, C. (1995) Letsibogo Dam and Reservoir: Mitigation of Archaeological Sites, Phase II, Final Report. Gaberone, Department of Water Affairs, unpublished report. Campbell, A.C., Kinahan, J. and van Waarden, C. (1996) Archaeological sites at Letsibogo Dam. Botswana Notes and Records 28:47–53. Coates-Palgrave, K. (1981) Trees of Southern Africa. Cape Town, C.Struik. Coppock, D.L. (1993) Vegetation and pastoral dynamics in the southern Ethiopian rangelands: implications for theory and management. In R.H.Behnke, I.Scoones and C.Kerven (eds) Range Ecology at Disequilibrium: New Models of Natural Variability and Pastoral Adaptation in African Savannas: 42–61. London, Overseas Development Institute. Denbow, J. (1979) Cenchrus ciliaris: an ecological indicator of iron age middens using aerial photography in eastern Botswana. South African Journal of Science 75:405–8. Denbow, J. (1986) A new look at the later prehistory of the Kalahari. Journal of African History 27:3–28. Dennell, R.W. (1982) Archaeology and the study of desertification. In B.Spooner and H.S.Mann (eds) Desertification and Development: Dryland Ecology in Social Perspective: 43–60. London, Academic Press. Ellis, J.E. and Swift, D.M. (1988) Stability of African pastoral ecosystems: alternate paradigms and implications for development. Journal of Range Management 41:450– 9. Evers, T.M. (1984) Sotho—Tswana and Moloko settlement patterns and the Bantu Cattle Pattern. In M.G.Hall, D.M.Avery, M.L.Wilson and A.J.B.Humphreys (eds) Frontiers: Southern African Archaeology Today: 236–47. Oxford, British Archaeological Reports, International Series 207. Fairhead, J. and Leach, M. (1996) Misreading the African Landscape: Society and Ecology in a Forest-Savanna Mosaic. Cambridge, Cambridge University Press. FAO (1978) Report on the Agroecological Zones Project, Vol. 1: Methodology and Results for Africa. Rome, Food and Agriculture Organization of the United Nations, World Soil Resources Reports No. 48. Fisher, J. (1995) Bone surface modifications in zooarchaeology. Journal of Archaeological Method and Theory 2 (1):7–68. Frost, P.G.H., Medina, E., Menaut, J.-C., Solbrig, O. and Walker, B. (1986) Responses of
Fifteenth-century agropastoral responses
243
savannas to stress and disturbance: a proposal for a collaborative programme of research. Report of IUBS working group on Decade of Tropics/Tropical savanna ecosystems. Biology International, Special Issue 10:1–78. Hall, M. (1987) The Changing Past: Farmers, Kings and Traders in Southern Africa, 200–1860. Cape Town, David Phillip. Hempel, C.G. (1966) Philosophy of Natural Science. Engelwood Cliffs, Prentice Hall. Hoben, A. (1996) Paradigms and politics in Ethiopia. In M.Leach and R.Mearns (eds) The Lie of the Land: Challenging Received Wisdom on the African Environment 186– 208. London, International African Institute. Hoffman, M.T. (1997) Human impacts on vegetation. In R.M.Cowling, D.M. Richardson and S.M.Pierce (eds) Vegetation of Southern Africa: 507–34. Cambridge, Cambridge University Press. Hoffman, M.T., Todd, S., Ntshona, Z. and Turner, S. (1999) Land Degradation in South Africa. Cape Town, Department of Environmental Affairs and Tourism. Homewood, K.M. (1992) Patch production by cattle. Nature 359:109–10. Huffman, T.N. (1980) Ceramics, classification and Iron Age entities. African Studies 39 (2):123–74. Huffman, T.N. (1996a) Archaeological evidence for climatic change during the last 2000 years in southern Africa. Quaternary International 33:55–60. Huffman, T.N. (1996b) Snakes and Crocodiles: Power and Symbolism in Ancient Zimbabwe. Johannesburg, Witwatersrand University Press. Huffman, T.N. and Hanisch, E.O.M. (1987) Settlement hierarchies in the northern Transvaal: Zimbabwe ruins and Venda history. African Studies 46 (1):79–116. Illius, A.W. and O’Connor, T.G. (1999) On the relevance of non-equilibrium concepts to arid and semi-arid grazing systems. Ecological Applications 9 (3). Jackson, S.P. (1961) Atlas Climatologique de L’Afrique. Lagos, Scientific Council for Africa. JMP (1995) Statistics and Graphics, Version 3.1. Cary, SAS Institute. Jonsson, J. (1998) Early Plant Economy in Zimbabwe. Uppsala, Studies in African Archaeology 16. Kempff, J. (1994) Probleme der Land-Degradation in Namibia: Ausmaβ, Ursachen und Wirkungsmuster. Würzburg, Würzburger Geographische Manuskripte, Heft 31. Kinahan, J. (1999) One Thousand Years of Agropastoral Settlement on the Motloutse River: Phase III Mitigation of Three Archaeological Sites Affected by the Letsibogo Dam near Mmadinare, Eastern Botswana. Gaborone, Aquatech Groundwater Consultants, unpublished report. Kinahan, J., Kinahan, J.H.A. and van Waarden, C. (1998) The archaeology and symbolic dimensions of a thirteenth century village in eastern Botswana. Southern African Field Archaeology 7 (2):63–71. Kiyiapi, J.L. (1994) Structure and characteristics of Acacia tortilis woodland on the Njemps Flats. In R.B.Bryan (ed.) Soil Erosion, Land Degradation and Social Transition: Geoecological Analysis of a Semi-arid Tropical Region, Kenya: 47–70. Advances in Geoecology 27. Klein, R.G. and Cruz-Uribe, K. (1984) The Analysis of Animal Bones from Archaeological Sites. Chicago, University of Chicago Press.
The archaeology of drylands
244
van der Koppel, J., Rietkerk, M. and Weissing, F.J. (1997) Catastrophic vegetation shifts and soil degradation in terrestrial grazing systems. Tree 12 (9):352–6. Kuper, A. (1982) Wives for Cattle: Bridewealth and Marriage in Southern Africa. London, Routledge & Kegan Paul. Lamprey, H.F. (1983) Pastoralism yesterday and today: the over-grazing problem. In F.Bourliere (ed.) Tropical Savannahs: 112–45. Amsterdam, Elsevier, Ecosystems of the World Vol. 3. Leach, M. and Mearns, R. (1996) (eds) The Lie of the Land: Challenging Received Wisdom on the African Environment. London, International African Institute. Little, P.D. (1996) Pastoralism, biodiversity and the shaping of savannah landscapes in East Africa. Africa 66 (1):37–51. Maggs, T.M.O’C. and Whitelaw, G. (1991) A review of recent archaeological research on food-producing communities in southern Africa. Journal of African History 32: 3– 24. Mortimore, M. (1998) Roots in the African Dust: Sustaining the Drylands. Cambridge, Cambridge University Press. Niamir, M. (1991) Traditional African range management techniques: implications for rangeland management. London, Overseas Development Institute, Pastoral Development Network Paper 31d:1–11. Nicholson, S.E. (1996) Environmental change within the historical period. In W.M. Adams, A.S.Goudie and A.R.Orme (eds) The Physical Geography of Africa: 60–87. Oxford, Oxford University Press. Phillipson, D.W. (1977) The Later Prehistory of Eastern and Southern Africa. London, Heinemann. Powell, N. (1998) Co-Management in Non-Equilibrium Systems: Cases from Namibian Rangelands. Uppsala, Swedish University of Agricultural Sciences, Agraria 138. Pratt, D.J. and Gwynne, M.D. (1977) (eds) Rangeland Management and Ecology in East Africa. London, Hodder & Stoughton. Rapp, R. (1985) Why do we need a scientific analysis of dryland degradation in Africa? In A.Hjort (ed.) Land Management and Survival: 109–18. Uppsala, Scandinavian Institute for African Studies. Rasmussen, K. (1985) A holistic view of ecological imbalance in drylands. In A. Hjort (ed.) Land Management and Survival: 119–27. Uppsala, Scandinavian Institute for African Studies. Reid, R.S. and Ellis, J.E. (1995) Impacts of pastoralists on woodlands in South Turkana, Kenya: livestock-mediated tree recruitment. Ecological Applications 5 (4):978–92. Scoones, I. (1996) Politics, polemics and pasture in southern Africa. In M.Leach and R.Mearns (eds) The Lie of the Land: Challenging Received Wisdom on the African Environment: 34–53. London, International African Institute. Shaw, M. (1974) Material culture. In W.D.Hammond-Tooke (ed.) The Bantu-Speaking Peoples of Southern Africa: 85–136. London, Routledge & Kegan Paul. Sinclair, A.R.E. and Fryxell, J.M. (1985) The Sahel of Africa: ecology of a disaster. Canadian Journal of Zoology 63:987–94. Skarpe, C. (1991) Impact of grazing in savanna ecosystems. Ambio 20 (8):351–6. Stiles, D. (1995) An overview of desertification as dryland degradation. In D.Stiles (ed.)
Fifteenth-century agropastoral responses
245
Social Aspects of Sustainable Dryland Management: 3–20. New York, John Wiley and Sons. Sullivan, S. (1996) Towards a non-equilibrium ecology: perspectives from an arid land. Journal of Biogeography 23:1–5. Thomas, D.S.G. and Shaw, P.A. (1991) The Kalahari Environment. Cambridge, Cambridge University Press. Turner, M.D. (1998) Long-term effects of daily grazing orbits on nutrient availability in Sahelian West Africa: 1. Gradients in the chemical composition of rangeland soils and vegetation. Journal of Biogeography 25:669–82. Tyson, P.D. and Lindsay, J.A. (1992) The climate of the last 2000 years in southern Africa. The Holocene 2:271–8. van Waarden, C. (1989) The granaries of Vumba: structural interpretation of a Khami Period commoner site. Journal of Anthropological Archaeology 8:131–57. van Waarden, C. (1999) The later Iron Age. In P.Lane, A.Reid and A.Segobye (eds) Ditswa Mmung: The Archaeology of Botswana: 115–61. Gaborone, Botswana Society. Westoby, M., Walker, B.H. and Noy-Meir, I. (1989) Opportunistic management for rangelands not at equilibrium. Journal of Range Management 42:266–74. Whitelaw, G. (1997) Southern African Iron Age. In J.O.Vogel (ed.) Encyclopaedia of Precolonial Africa: Archaeology, History, Languages, Cultures and Environments: 444–56. London, Sage (Alta Mira).
14 Islands of intensive agriculture in African drylands: towards an explanatory framework MATS WIDGREN
INTRODUCTION The social and cultural diversity of populations in dryland Africa is vast, with population densities ranging from less than five to more than 300 per square kilometre. As Mortimore (1998:17) has emphasized, this range in population densities cannot be explained by differences in climate: ‘there is a weak relation between aridity and population density. While high densities are rare in the arid zone, the higher ones are found not in the moist, but in the dry semi-arid zone.’ It is evident that the distribution of different farming systems, especially in the semi-arid lands, reflects social, economic and political factors at least as much as environmental factors. The farming systems developed for coping with arid lands are thus many and varied and are the result of centuries and millennia of agricultural experience. No single formula for cultivating arid lands can be found—each farming system relies on its own mix of components to cope with the two main problems of farming drylands: water management and fertility maintenance. This is achieved through crop varieties, meticulous timescheduling of planting, farming practices aimed at restoring organic content, and construction works such as terraces, irrigation furrows and so on. Furthermore, interactions with pastoralism seem to be a sine qua non of agrarian societies in drylands. The ways in which these different components are combined vary throughout eastern and southern Africa, although the regional distribution of farming systems in the area is only vaguely known and documented (Ker, 1995). Temporary cultivation with little investment in land is often assumed to be the general rule, but several exceptions can be documented. In West Africa, for example, ‘ring cultivation systems’ (Fussel, 1992:494) are practised akin to European infield-outfield cultivation, with intensively farmed and manured fields close to the settlement and a zone of temporary fields beyond. In southern Africa, different types of temporary cultivation in the savanna zone are common, but high output and socially-sustainable production can also be achieved in such an extensive system through a social system that caters for redistribution between years and between cultivators. The Sotho—Tswana settlement system in the interior areas of southern Africa has been recognized as an agricultural and social adaptation to low and erratic rainfall. The Tswana (in present-day Botswana and in South Africa) have a history of large, concentrated settlements combined with widely dispersed areas for arable fields and a pastoral organization reaching more than 20 km from the main settlements. This pattern of settlement and land use was contrasted by Sansom (1974:138ff) with the
Islands of intensive agriculture in African drylands
247
settlement structure of the Nguni peoples on the eastern rim of South Africa, who, because of the higher rainfall and more dissected landscape there, were able to base their agriculture on a confined territory in each settlement. Sansom’s thesis has been criticized on good grounds for being environmentally deterministic (Huffman, 1986). The problem is that it operates in a historical and social vacuum, whereas research has shown that the highly concentrated settlements among the Sotho—Tswana, and among previous populations in the same area, reflect social and political hierarchies rather than simply an adaptation to a semi-arid climate. However, the environmental arguments cannot be dismissed totally on these grounds. Within the ecological context of semi-arid lands with few topographical variations, and hence few variations in precipitation, the Tswana type of exploitation pattern represents a production form that is able to produce a surplus for an elite and to sustain large populations through a spatial and temporal redistribution of the harvest. This chapter is focused on still another way of increasing production in semi-arid lands, based on investments in land, on permanency of fields, and on labour-intensive forms of land management. Such farming systems are not the rule, and probably never have been, but exist as small pockets or ‘islands’ of intensive agriculture surrounded by pastoral land use or temporary cultivation. They are known from Nigeria to South Africa. In a series of articles Sutton has drawn our attention to different areas in eastern and southern Africa where large systems of ancient fields and furrows bear witness to abandoned agrarian communities with the characteristics of such ‘islands of intensification’ (Grove and Sutton, 1989; Sutton, 1984, 1985, 1989, 1998; Fig. 14.1). In 1995 Maggs presented new documentation from Marateng in the Lydenburg area in Mpumalanga province, South Africa, placing it in the context of the previously known field systems in eastern and southern Africa (Maggs, 1995). In the present volume, reports are presented of the ongoing research on the ancient fields at Engaruka in Tanzania (Sutton, Chapter 11) and from the important comparative example of Nyanga in Zimbabwe (Soper, Chapter 12). These archaeological complexes share a general dating to the middle part of the second millennium AD; they were abandoned between 100 and 400 years ago, so that none has written documentation on their use. To judge from the investments in land evidenced in these field systems, the agrarian communities that built them were capable
The archaeology of drylands
248
Figure 14.1 Eastern and southern Africa, showing sites mentioned in Chapter 14 of solving the basic problems of water management and fertility maintenance. In discussing the causes of abandonment of similar systems in other parts of the world, Brookfield (1986:180) has argued that they ‘were almost all highly conservationist, and it was their breakdown and abandonment that was more likely to yield damage to the land’. As with all large archaeological complexes of this sort, the central problems are about the dates of emergence and desertion and the reasons for their rise and eventual abandonment. In the discussion of these matters, surviving agrarian communities sharing the same characteristics are important in providing comparative evidence. Terracing and irrigation can be found in several locally-developed farming systems in the region, such as Marakwet in Kenya, Sonjo in Tanzania and Konso in Ethiopia. An overview of the whole problem of intensive or specialized agriculture in Africa was given in a special issue of Azania (1989, volume 24), and full references can be found there. Since then, important contributions have been published on Sonjo (Adams et al., 1994; Potkanski and Adams, 1998) and Marakwet (Adams et al., 1997; Watson et al., 1998), and one is forthcoming on Konso (Watson, 1999a, 1999b). Our empirical understanding of how such agricultural systems in the past emerged, developed and decayed derives from different types of situations and source materials. At one extreme, there are the cases of Engaruka in Tanzania and of Nyanga in Zimbabwe, which are deserted field systems with poor-to-non-existent historical documentation but
Islands of intensive agriculture in African drylands
249
with reasonably well-dated archaeological features. At the other extreme are currentlysurviving active farming systems like Sonjo and Marakwet where, however, the historical origins are still unclear. Into that context must also be brought cases like the Machakos in Kenya, where a development of intensification and of increasing technological investment in land (‘landesque capital’) is currently taking place (Tiffen et al., 1994). The concluding discussion of the Machakos study serves to remind us of the importance of studying the present implications and development possibilities of historical cases. (A similar approach to intensive agriculture can be found in Bebbington’s [1997] discussion of the recent development of islands of sustainable agriculture in the rural Andes.)
ISLANDS OF INTENSIVE AGRICULTURE The current definitions of ‘islands’ and of ‘intensive’ may both be questioned. Recent examples of islands of intensive agriculture share some common characteristics. First, they are characterized by agricultural systems that have for a long period been able to support a larger population density than surrounding areas. The metaphor ‘islands’ is used to describe the fact that these areas may exist within a ‘sea’ of less-intensive land use such as shifting cultivation or pastoralism. Second, to judge from the history of the recent examples, they have been less fragile and more robust in the face of both drought and human disturbances, which are so characteristic of the semi-arid lands of Africa. Many of these areas are now poorer than their more expansive peripheries, but they still—through traditional networks of exchange—play an important role in the foodsecurity system. They thus represent lessons from the past for the urgent problem of food security. Furthermore, the high productivity of land and the robust nature of agricultural production in these areas depend on the application of different combinations of farming practices, including manuring, composting, terracing, cut-off drains, irrigation and crop diversity. In many of the areas there is also evidence of careful management of trees and woodlands. Irrigation and soil conservation are connected with ‘landesque capital’ investment: activities affecting land and vegetation that reach beyond the immediate needs of the coming cropping season. The latter fact is also of crucial importance for the archaeological identification of past agrarian societies of that kind. Our interest in these systems stems from the fact that they seem to provide historical and contemporary examples of locally-developed solutions to the critical problems in modern African agriculture: low output from traditional systems; threatened sustainability of the production systems and/or widespread degradation; and unreliable access to food.
CURRENT FIELD RESEARCH Through detailed studies in two living agrarian landscapes in eastern Africa, we (see Acknowledgements) are seeking to understand the ecological, historical and social contexts of this type of intensive farming. Two case studies are being carried out, in Tanzania and Kenya. These empirical studies are focused on work processes, social
The archaeology of drylands
250
institutions, land tenure, technology and their material expressions as physical features, i.e. fields and landscapes. The project is based in anthropology and geography and combines the methods of landscape history with participatory approaches. Though the empirical focus is on two cases, we are working in close contact with other researchers studying abandoned field systems or intensively-farmed areas in other parts of eastern Africa. The project has thus set itself the task of finding a common explanatory framework to embrace historical questions such as why areas like Engaruka, Nyanga and Marateng were abandoned and why areas like Marakwet, Sonjo and Konso persist. At the same time, the framework should be able to accommodate questions on the future potential of these areas and the mechanisms of ‘take-off’. The questions we are asking thus relate to different phases in the histories of the areas. We are first asking under what circumstances did intensive farming originally begin: what were the specific, place-bound, events and characteristics? The second set of questions relates to the social organization that makes possible the mobilization of labour and the investment or gradual build-up of landesque capital. Closely related to this are the social practices that serve to reproduce the farming system from generation to generation, and are at the same time flexible enough to cater for population increase and settlement expansion. The third set of questions, not treated here, relates to the present developmental possibilities of these different areas. To what extent can they continue to play an important role in the future, either as cores in a food-security system, or as a basis for a market-oriented development? It is not our object to study in detail, and for their own sake, all the different farming practices that are used in these areas, such as terracing, composting, manuring, irrigation and so on. It has long been acknowledged that such locally-developed solutions to the problems of nutrient deficiency, land degradation and lack of water have a long historical tradition in Africa, and we have little to add to that debate. Instead, we are trying to understand the process whereby such practices are put together in a farming and social system capable of increasing both land productivity and food security in a sustainable way.
CASE STUDIES Mama Issara, Mbulu District, Northern Tanzania Mama Issara is the core area of the Iraqw people. Agriculture is restrained by the dissected topography, and cultivation is done entirely with hand implements. The system of intensive farming is unique in the region in terms of its diversification and elaboration, and has a history that goes back some 200 years. The population has been estimated at around 20,000, with a density of around 100 people per square kilometre. Terracing, mulching, manuring and water harvesting are practised (Börjeson, 1998; Loiske, 1993, 1995:14–30; Tengö, 1999; Figs. 14.2, 14.3, and 14.4). Mama Issara is a prime example of how local institutions for natural resource management have been able to uphold an intensive farming system for a long time (Börjeson, 1998, 1999). Several factors are of importance, including strong social
Islands of intensive agriculture in African drylands
251
cohesion, efficient forms of decision making and a tradition of communal labour cooperation. Also, religious beliefs support the sustainable use of natural resources, in that the earth spirit is thought to punish over-use of land and trees. As Börjeson (1999) has shown, the systems for the transfer of land between and within generations are an important part of these institutions and play a central role in the reproduction and persistence of the farming system. Though Mama Issara is involved to only a limited extent in market production, there is a considerable exchange of products between Mama Issara and the Iraqw expansion areas. All the families participate in institutionalized food exchanges involving between five and twenty-five other families. These exchanges are based on ritual, economic and social networks, covering areas with varying ecology and varying production (Loiske, 1999). The islands of
Figure 14.2 An intensively cultivated landscape at Kwermusl (Mama Issara, Mbulu district, Tanzania) Photograph: L.Börjeson
The archaeology of drylands
252
Figure 14.3 Preparing the field at Kwermusl Photograph: L.Börjeson
Figure 14.4 Piles of manure from stalled cattle—an integral part of the farming system in Mama Issara Photograph: L. Börjeson intensive agriculture can thus be seen not in isolation but as manifestations of the
Islands of intensive agriculture in African drylands
253
geographical division of labour. Marakwet, Kenya In any discussion of intensified agriculture, the Marakwet area constitutes a particularly interesting case. In the dry Kerio Valley in western Kenya, we find a system of irrigated farming that, from a modest beginning some 200 years ago, has grown into a comprehensive system in which the total length of the furrows now reaches 250 km (Adams et al., 1997; Watson et al., 1998; Figs. 14.5 and 14.6). A centralized political system has not developed, however
Figure 14.5 An irrigation channel above Tot in Marakwet, Kenya Photograph: L.Börjeson
The archaeology of drylands
254
Figure 14.6 An irrigation canal under repair above Chesoi in Marakwet, Kenya Photograph: M.Widgren —in fact, no single individual or group of people has an overview of how the system works in its totality, although it encompasses more than forty major furrows. Each irrigation furrow is under the control of the lineage that originally constructed it, while other groups lack primary rights to water, without which reliable farming is not possible. This situation has not resulted in a hierarchical society, such as might have been expected in terms of Wittfogel’s classic theory of ‘oriental despotism’, which states that
Islands of intensive agriculture in African drylands
255
societies with large-scale irrigation will develop centralized orders of command, which in turn will lead to despotic political systems (Wittfogel, 1957). Though this hypothesis in its more pronounced form has been criticized and is becoming somewhat dated, it remains an interesting fact that a society with such a comprehensive irrigation system as the Marakwet is organized acephalously. Östberg (1999) has recently summarized some preliminary results on the origin of this farming system. He argues that the development of a geographically-based division of labour between the groups inhabiting the Kerio valley is the key to explaining how the Marakwet came to develop an intensive irrigation system. Co-operation and competition between the agriculturalist Marakwet and the pastoral East Pokot have been instrumental in shaping the present-day utilization of resources in the valley. Unlike an evolutionary explanation, this finding emphasizes inter-dependence between different groups and increasing variation, instead of unidirectional development.
EXPLANATIONS OF INTENSIFICATION The discussion of the origins and persistence of these intensive systems can be initiated by asking the elementary geographical question about location: why do we find these systems in these specific places rather than elsewhere? The cases mentioned here and a handful of others share some locational characteristics. They are all located along the East African rift valley, the sharp topographical variations of which provide good opportunities for intensive farming. Many of these farming systems make use of the variations in precipitation and climate within short distances that are characteristic of the high escarpments here, but their locations can in no way be said to be simply environmentally determined: there are examples of similar environments along the rift valley where neither present intensive farming nor any traces of former intensive agriculture exist. Furthermore, areas of intensive farming can also be found in other types of environments in the semi-arid parts of eastern and southern Africa. The distribution of intensive agriculture in the semi-arid parts of Africa is thus not a direct reflection of natural conditions but the result of a complex interaction of ecological, social and historical factors. There is also no simple relation to economically-defined geographical variables. The location theory developed for agricultural activity puts the distance to market in a central place when explaining the distribution of intensive farming. In the recent case of Machakos, the proximity to the market in Nairobi is one important explanation, but it is in no sense the only one. In the case of Baringo (Anderson, 1988, 1989), the market situation also seems to have been of vital importance for the development of the irrigated agriculture during the nineteenth century. Market conditions do not play the same role in Marakwet and Mama Issara, however, both of which are remote from markets and suffer from poor communications. A second explanation for the geographical distribution of intensive farming would be that islands of locally-developed intensive agriculture are the remains of a type of agriculture that formerly was much more widespread. This explanation makes colonialism the main force behind the de-intensification of African agriculture, with the
The archaeology of drylands
256
islands being seen as pockets that have survived these developments. The problem of precolonial farming systems in Tanzania has been the object of a debate that is outside the scope of this chapter, but though the advent of colonialism certainly led in many cases to the disruption of local farming societies, it would be too simple to advance it as the main force behind a general de-intensification of farming systems in eastern Africa. It has even been proposed that the migrations triggered by long-distance trade may have indirectly led to the establishment of some of the intensive farming systems in Tanzania (Koponen, 1988:240f). The above-mentioned models of agrarian development are all based on the idea of an even development of farming systems in response to markets and/or population pressures. I find it more challenging to start from the opposite assumption: that social systems and landscapes are the result of geographically and socially uneven development. The idea of the uneven development of farming systems is supported by the fact that both the emergence and the decay of systems of intensive farming seem to be general traits in the history of agriculture throughout the world. Farming systems do not evolve from simple to complex or from extensive to intensive according to some pre-set model, but are formed and changed within specific, place-bound, social, historical and ecological contexts. If we accept the idea of uneven development, it is also much easier to understand why the intensity of agriculture is not evenly or directly related either to markets or to natural conditions. The islands of agrarian intensity have their own logic of development, and simplistic explanatory models cannot reflect their distribution or their development. The questions of where and why remain, however, central for our understanding of the processes behind intensive agriculture.
POLITICAL ECONOMY AND THE DEVELOPMENT OF HIERARCHIES In his discussion of intensive agriculture in eastern Africa, Thomas Håkansson contrasted the Boserupian explanation of agricultural intensification (intensification as a response to population pressure: Boserup, 1965) with two other models, both of which were based on the idea that intensification could be more broadly understood as an effect of pressure on production rather than population pressure (Håkansson, 1989). He argued that locallydeveloped systems of intensive farming were likely to be the outcome of one or both of the following sets of processes: first, political economy and the development of hierarchies, and second, commercial development and increasing market production. The political-economy model, as Håkansson termed it, was based mainly on research carried out in central America and southeastern Asia. In both regions, competitive feasting and redistribution between chiefs created a need for agrarian surpluses. As has been shown in many other studies, the development and decay of such hierarchies are very dynamic processes and could indeed account for the uneven development and the uneven location pattern of islands of intensification. Furthermore, Håkansson argued, tribute labour controlled by chiefs and kings can be seen as one of the ways of mobilizing the labour needed for the large investments in land connected with this agriculture in order to construct features such as irrigation furrows and stone terraces.
Islands of intensive agriculture in African drylands
257
However, these models of social systems do not fit very well with intensive agricultural systems in the context of eastern Africa, and the evidence that Håkansson cites from Africa is a single case study. In Marakwet, with its large and thriving irrigation system, the mobilization of labour and the surveillance of the irrigation system are based on the decentralized power of the elders and on negotiations, rather than on chiefly authority and tribute labour. As far as I can gather, the same holds true for Mama Issara and Sonjo. In these systems, labour, land rights and water rights are embedded in a clanand lineage-based society, rather than in chiefly authority. In this connection, the ideas put forward by Shipton (1984a, 1984b) on the relations in eastern Africa between farming intensity and population density, on the one hand, and state- or chiefdom-oriented social structures, on the other, are of interest. Intensive farming in eastern Africa, according to Shipton, is usually associated not with a centralized control of land but rather with lineage- and clan-based land rights. In the field pattern, this is associated with land strips expressing the kinship structure, so that clans, minimal lineages and heirs have their definite shares of the land. He argues that a more patchy system of fields is usually associated with chiefly control of land in less intensive farming systems, which is a model more in accordance with the Tswana system discussed in the beginning of this chapter than with the intensive systems we know on the ground in eastern Africa. Shipton’s conclusions and our own observations from our study areas make the hierarchy model less valid for understanding such systems. The market arguments, which are also advanced by Håkansson, also seem to be short of explanatory power in relation to the systems that we have been studying. At least today, many of the areas with intensive farming are poor and located far away from markets. In the case of the Iraqw in Tanzania, one can even observe an inverted relationship between labour intensity and proximity to market. The less labour-intensive agriculture is located closer to the market and is also more involved in market production, while the labour-intensive core area has poor roads and only a small share of cash crops. However, this does not mean necessarily that the studied areas are closed entities relying solely on subsistence production: as Loiske (1999) has shown in the case of the Iraqw, under the surface there is in fact a considerable amount of exchange of agricultural products between the core and peripheral areas. Therefore, to judge from the existing literature and from the evidence brought forward in this project, we have the paradoxical situation in different parts of the world that both hierarchies and the absence of hierarchies can be associated with labour-intensive agriculture. The same seems to hold true of the role of the market: both market orientation and subsistence farming can be connected with labour-intensive farming. The common denominator between these different situations, however, is that there is a geographical division of labour: the islands do not exist in isolation but are based on production and resource utilization from a range of different economic zones, based on different climates and/or different production systems. The exchange of products between different zones thus seems to be an important pre-condition for the existence of intensive agriculture. In the case of Mama Issara, these exchanges take place within the same ethnic group. In other cases, of which Marakwet is an example, exchange between agriculturists and pastoralists of different ethnical backgrounds may form an important incentive for specialization. This was also the case in Baringo (Anderson, 1988, 1989)
The archaeology of drylands
258
and may play a certain role among the Sonjo, who are surrounded by Maasai. There is a similar paradox in the case of internal social organization, in terms of hierarchical and egalitarian systems: the connection seems to be with the different ways of mobilizing labour. The empirical material shows that labour mobilization need not be associated only with tribute labour and the social division of labour between kings and commoners, but can also be organized according to age sets and/or labour exchange within more egalitarian social structures. The comparison with the hierarchical model has brought into focus three important factors that must be studied if we are to understand the emergence and persistence of islands of intensive cultivation and high productivity. First, they all form part of a wider geographical division of labour, but that can take different forms, being based variously on commercial development, on exchange within the ethnic group along kinship networks, or on exchange between agriculturists and pastoralists of different ethnic groups. Second, mobilization of labour is indeed an integral part of intensive farming. Projects such as terraces and furrows need investments and repairs, and with an increased number of crops per year, preparing the land, sowing and harvesting also become potential bottlenecks. Our case studies show that traditional systems of labour exchange and/or work based on age sets can provide such an input of work. Thus, large systems of irrigation and field terracing do not necessarily indicate a hierarchical chiefdom structure. Finally, land and water rights can be incorporated in a clan-based system, and it seems that these property rights can provide both the stability needed for investments in land and at the same time the flexibility to cater for fluctuations in climate, as well as social and political changes. This flexible system of land and water rights is, furthermore, closely connected with the mechanisms for reproducing social organization and mobilizing labour.
ACKNOWLEDGEMENTS This chapter is a preliminary report on a project financed by the Swedish International Development Authority (SIDA) and the Swedish Council for the Planning and Coordination of Research (FRN), with links with the Institute of Resource Assessment in Dar es Salaam, the British Institute in Eastern Africa and the National Museums of Kenya. I have drawn heavily on informal and productive discussions during a workshop in the field in Marakwet and at the British Institute in Eastern Africa, Nairobi, in October 1998. The participants in the field were Andrew Cheptum, Johnstone Kassagam (both at the National Museums of Kenya), Naomi Mason, John Sutton (British Institute in Eastern Africa), Elisabeth Watson (University of Cambridge) and the Swedish research team: Lowe Börjeson, Vesa-Matti Loiske and Wilhelm Östberg. Bill Adams has commented on an earlier version. All are warmly thanked for their contributions.
REFERENCES Adams, W., Potkanski, T. and Sutton, J.E.G. (1994) Indigenous farmer-managed
Islands of intensive agriculture in African drylands
259
irrigation in Sonjo, Tanzania. Geographical Journal 160 (1):17–32. Adams, W., Watson, E.E. and Mutiso, S.K. (1997) Water rules and gender: water rights in an indigenous irrigation system, Marakwet, Kenya. Development and Change 28:707–30. Anderson, D. (1988) Cultivating pastoralists: ecology and economy among the II Chamus of Baringo, 1840–1980. In D.Johnson and D.Anderson (eds) The Ecology of Survival: Case Studies from Northeast African History: 241–60. London, Lester Crook. Anderson, D. (1989) Agriculture and irrigation technology at Lake Baringo. Azania 24:89–97. Bebbington, A. (1997) Social capital and rural intensification: local organisations and islands of sustainability in the rural Andes. Geographical Journal 163:189–97. Börjeson, L. (1998) Landscape, Land Use and Land Tenure in Mama Issara, Tanzania. Mapping a ‘Traditional’ Intensive Farming System . Uppsala, Swedish University of Agricultural Sciences, Minor Field Study No. 47. Börjeson, L. (1999) Listening to the land: the Iraqw intensive farming system as told by a hill and its inhabitants. In M.Widgren and J.E.G.Sutton (eds) ‘Islands’ of Intensive Agriculture in the East African Rift and Highlands: A 500-Year Perspective: 56–73. Stockholm, Stockholm University, Department of Geography, Working Paper from the Environment and Development Studies Unit 43. Boserup, E. (1965) The Conditions of Agricultural Growth. London, Allen & Unwin. Brookfield, H. (1986) Intensification intensified. Archaeology in Oceania 31:177–80 Fussel, L.K. (1992) Semi-arid cereal and grazing systems of West Africa. In C.J. Pearson (ed.) Field Crop Ecosystems: 485–518. Amsterdam, Elsevier. Grove, A.T. and Sutton, J.E.G. (1989) Agricultural terracing south of the Sahara. Azania 24:113–22. Håkansson, T. (1989) Social and political aspects of intensive agriculture in East Africa: some models from cultural anthropology. Azania 24:12–20. Huffman, T. (1986) Archaeological evidence and conventional explanations of Southern Bantu settlement patterns. Africa 56 (3):280–98. Ker, A. (1995) Farming Systems of the African Savanna: A Continent in Crisis. Ottawa, International Development Research Centre. Koponen, J. (1988) People and Production in Late Precolonial Tanzania. Helsinki, Finnish Society for Development Studies; Finnish Anthropological Society; Finnish Historical Society in cooperation with Scandinavian Institute of African Studies. Loiske, V.-M. (1993) Mama Isara: A Sustainable Agricultural System in Mbulu District, Tanzania. Stockholm, Stockholm University, Department of Geography, Working Paper from the Environment and Development Studies Unit 21. Loiske, V.-M. (1995) The Village That Vanished. The Roots of Erosion in a Tanzanian Village. Stockholm, Stockholm University, Department of Human Geography, Meddelanden series B 94. Loiske, V.-M. (1999) Persistent peasants: The case of the Iraqw in central Tanzania. In M.Widgren and J.E.G.Sutton (eds) ‘Islands’ of Intensive Agriculture in the East African Rift and Highlands: A 500-Year Perspective: 44–53. Stockholm, Stockholm University, Department of Geography, Working Paper from the Environment and Development Studies Unit 43.
The archaeology of drylands
260
Maggs, T. (1995) From Marateng to Marakwet. Islands of agricultural intensification in Eastern and Southern Africa. Paper presented at the Prehistoric African Association Congress, Harare. Mortimore, M. (1998) Roots in the African Dust: Sustaining the Drylands. Cambridge, Cambridge University Press. Östberg, W. (1999) The origins and expansion of Marakwet hill-furrow irrigation in the Kerio Valley, Kenya: an interpretation. In M.Widgren and J.E.G.Sutton (eds) ‘Islands’ of Intensive Agriculture in the East African Rift and Highlands: A 500-Year Perspective: 15–43. Stockholm, Stockholm University, Department of Geography, Working Paper from the Environment and Development Studies Unit 43. Potkanski, T. and Adams, W.M. (1998) Water scarcity, property regimes and irrigation management in Sonjo, Tanzania. Journal of Development Studies 14: 86–116. Sansom, B, (1974) Traditional economic systems. In W.D.Hammond-Tooke (ed.) The Bantu-Speaking Peoples of Southern Africa: 135–76. London, Routledge & Kegan Paul. Shipton, P.M. (1984a) Strips and patches: a demographic dimension in some African land-holding and political systems. Man: 616–20. Shipton, P.M. (1984b) Lineage and locality as antithetical principles in East African systems of land tenure. Ethnology 23:117–32. Sutton, J.E.G. (1984) Irrigation and soil conservation in African agricultural history: with a reconsideration of the Inyanga terracing (Zimbabwe) and Engaruka irrigation works (Tanzania). Journal of African History 25:25–41. Sutton, J.E.G. (1985) Irrigation and terracing in African agricultural history: intensification, specialisation, or overspecialisation? In I.S.Farrington (ed.) Prehistoric Intensive Agriculture in the Tropics: 737–64. Oxford, British Archaeologial Reports, International Series 232, Volume 2. Sutton, J.E.G. (1989) Towards a history of cultivating the fields. Azania 24:98–122. Sutton, J.E.G. (1998) Engaruka: an irrigation community in northern Tanzania before the Maasai. Azania 33:1–38. Tengö, M (1999) Integrated Nutrient Management and Farmers’ Practises in the AgroEcosystem of Mama Issara, Tanzania. Stockholm University, Department of Systems Ecology, unpublished honours thesis. Tiffen, M., Mortimore, M. and Gichuki, F. (1994) More People, Less Erosion: Environmental Recovery in Kenya. Chichester, John Wiley and Sons. Watson, E. (1999a) Ground Truths: Land and Power in Konso, Ethiopia. University of Cambridge, Department of Geography, unpublished PhD dissertation. Watson, E. (1999b) Konso integrated agriculture as social process: abstract. In M. Widgren and J.E.G.Sutton (eds) ‘Islands’ of Intensive Agriculture in the East African Rift and Highlands: A 500-Year Perspective: 74. Stockholm, Stockholm University, Department of Geography, Working Paper from the Environment and Development Studies Unit 43. Watson, E.E, Adams, W. and Mutiso, S.K. (1998) Indigenous irrigation, agriculture and development, Marakwet, Kenya. Geographical Journal 164:67–84. Wittfogel, K. (1957) Oriental Despotism. New Haven, Conn, Yale University Press.
Part V NORTH AND CENTRAL AMERICA
15 Prehistoric agriculture and anthropogenic ecology of the North American Southwest PAUL E.MINNIS
INTRODUCTION Many semi-arid to arid areas are the heartlands of agriculture, and the lessons learned from millennia of food production in these often difficult environments can provide critical information for understanding the past. Perhaps as importantly, we can use knowledge of the astounding range of prehistoric agricultural strategies and their ecological effects to build a more sustainable future, especially where food production expands into unfamiliar and unfavourable locations. Here I outline the types of agriculture used by the ancient peoples of the region now encompassed by the southwestern part of the United States and northwestern Mexico, for convenience termed here the North American Southwest (Fig. 15.1). This region is an excellent location in which to address issues of prehistoric human ecology, because it is one of the most intensely studied dryland regions in the world, so we have in some locations surprising precision in paleoenvironmental reconstruction and awareness of the region’s prehistory. The chapter’s focus then shifts to the anthropogenic effects of farming and, finally, to discussion of the role agriculture played in the historical dynamics of the region.
ENVIRONMENT AND CULTURE HISTORY BACKGROUND The North American Southwest is an environmentally and anthropologically diverse region. The two hot deserts of the southern part of the region, the Sonoran and Chihuahuan, are interspersed with isolated mountains, and major mountain ranges tower up to nearly 4,000 m above sea level. The northern part of the region is dominated by the Colorado Plateau, with cool deserts and semi-arid grasslands. Substantial rivers, such as the Gila, Colorado and Rio Grande/Rio Bravo del Norte, are infrequent, but they were foci of prehistoric human occupation. Annual rainfall ranges from 127 mm in the lowest deserts to 700 mm in the mid-level mountains (Sellers and Hill, 1974;
Prehistoric agriculture and anthropogenic ecology
265
Figure 15.1 The North American Southwest: the states of Arizona and New Mexico, and portions of surrounding states in both the United States and Mexico Tuan et al., 1973). Typically, precipitation is bimodally distributed, with large winter storms and more localized summer monsoons. Thus, crops often require supplemental water to yield adequate harvests. Deserts now support grasslands and shrub communities with occasional ribbons of riparian vegetation (see Brown, 1982, for the best summary of the region’s biotic communities; for Mexico, see also Rzedowski, 1986). Low-elevation montane vegetation is dominated by oak, pine and juniper woodlands in various combinations, with higher montane forests of gymnosperms such as firs, spruces and pines. Ecology is dynamic, and there is evidence of substantial environmental change, including during the historic period—in fact, substantial environmental changes have been noted even within the last century. The best documented historic change has been the expansion of desert shrubs such as mesquite (Prosopis) and montane juniper (Juniperus) at the expense of desert grasslands. The best explanation for these changes involves fire suppression, drought and intensive livestock grazing (Bahre, 1991; Hastings and Turner, 1965; Humphrey, 1987). Many millennia of human occupation preceded the use of cultivated plants in the region (for general accounts of the regional prehistory, see: Cordell, 1997; Plog, 1997). The first post-Pleistocene peoples seemed to have lived in small hunter-gatherer bands until about 2000–1000 BC. Starting around this time, more aggregated populations practising some agriculture appeared in at least two locations: around Tucson, Arizona, and in northwestern Chihuahua (Hard and Roney, 1998). The most important crops, such
The archaeology of drylands
266
as maize (Zea mays), various pulses/beans (mostly Phaseolus) and squashes (Cucurbita), originated to the south in Mesoamerica. Yet, sedentary village agriculture seems not to have become widespread throughout the region until AD 200–700. Occasionally, complex regional polities developed, the best known examples being Chaco Canyon in New Mexico, Casas Grandes in northwestern Chihuahua and the Hohokam of Arizona. While population size, degree of aggregation and settlement locations fluctuated through time, due in part to environmental perturbations, agriculture has been the economic mainstay until and after European contact in the late 1500s. Prehistoric domesticated animals were restricted to the turkey and dog; sheep, horses, cattle and goats were historic European introductions.
PREHISTORIC AGRICULTURAL STRATEGIES Prehistoric humans farmed the North American Southwest for millennia and, not surprisingly, they developed a wide range of techniques and strategies to grow crops under difficult circumstances. The most difficult problem they faced was insufficient precipitation. Adding to a large corpus of research on ancient farming in the region are some excellent ethnographic studies of indigenous farming, especially of the Hopi (Bradfield, 1971; Hack, 1942) and of Sonoran desert peoples (Castetter and Willis, 1942, 1952). Not wishing to become bogged down in unnecessary taxonomic complexities, I shall divide agricultural techniques into four simple general categories: irrigation; floodwater farming; rain-fed farming; and rock mulching. Irrigation Irrigation was widely practised. Its origins are earlier than previously thought (Doolittle, 1990), and the frequency of irrigation agriculture increased through time. The largest and most famous irrigation systems in the region were built by the prehistoric peoples of the Salt and Gila river basins (where the modern city of Phoenix is located): ‘in terms of complexity it simply had no rival anywhere in Mexico’ (Doolittle, 1990:79). Complicated sets of canals totalling over 500 km were constructed (Fig. 15.2), although the destruction of canals by modern agriculture and explosive urban development has obliterated most of them (Dart, 1989; Fish and Nabhan, 1991; Howard, 1993). Most other irrigation systems in the region, however, seem to have been quite small and were organized at a familial level of production (Fish and Fish, 1984; Toll, 1995). Floodwater farming Evidence of ancient systems of floodwater farming is commonly found throughout the region at locations that are still used by some communities today (Nabhan, 1979, 1986a; Nabhan and Sheridan, 1977). At times, floodwater strategies blend into irrigation systems, and there is no point in trying to make a sharp distinction between the two. Usually, temporary features divert surface water run-off immediately following rains. Ak chin, fields at the alluvial fan of a short drainage, are another common form of farming
Prehistoric agriculture and anthropogenic ecology
267
(Nabhan, 1986b). Again, most ancient floodwater systems are rather small, lacking evidence of substantial super-familial co-ordination. One possible well-known exception is a 9 ha field at Chaco Canyon, the centre of a remarkably complex regional polity (Vivian, 1991). Some of the best known and easily seen archaeological remains of floodwater farming are checkdams (trincheras) (Fig. 15.3): rock walls across the topographic contour that catch water and soil (Donkin, 1979; Toll, 1995; Woodbury, 1961). Well-known examples of trincheras are found from the northern sector of the region at Mesa Verde (Cordell, 1977) to the southern part, such as around Casas Grandes in Chihuahua (Di Peso, 1974; Herold, 1970; Howard and Griffiths, 1966; Schmidt and Gerald, 1988). I have co-directed a long-term archaeological project in the Casas Grandes area for nine years, and even though it is clear that the irrigated floodplains were the primary prehistoric farming locations, trincheras are common archaeological features—we have recorded hundreds (Minnis and Whalen, 1996; Whalen and Minnis, 1996). Most are quite small, with fields averaging about 2,500 m2, the largest of the main group being 8,000 m2, but there is one exception—a series of trincheras that covered at least 100,000 m2. Interestingly, this field system is next to a site that appears to have been an administrative/
Figure 15.2 Prehistoric Hohokam communities and irrigation systems in the Phoenix basin of the Salt river Map courtesy of Suzanne K. Fish
The archaeology of drylands
268
Figure 15.3 Aerial photograph of prehistoric trincheras (checkdam) fields near Casas Grandes, Chihuahua, Mexico Photograph: A.Heisey ritual centre within the Casas Grandes polity. Its exceptionally large agricultural system may be evidence of organization and surplus production beyond the household level. Rain-fed farming Many areas of the region can be farmed with only direct precipitation under optimal conditions, but it is difficult to detect prehistoric dryland farming unless soil is modified sufficiently to leave archaeological remains. Non-irrigated gridded gardens (Fig. 15.4)— small plots marked by checkerboards of low stone walls—are one such modification and have been found in many areas, such as in southeastern Arizona (Gilman and Sherman, 1975) and northern New Mexico (Ford, 2000; Maxwell and Anschuetz, 1992). Direct rain-fed agriculture is risky farming in the light of the region’s marginal precipitation for maize-based farming, the documented fluctuation in annual precipitation and the apparent vulnerability of some soils to nutrient depletion after sustained cropping (Kohler et al., in press; Sandor, 1992). In fact, dryland maize farming in eastern New Mexico at the beginning of the twentieth century suffered a failure rate of one out of four years (Staten et al., 1939). It is likely that the successes and failures of rain-fed farming were especially important in prehistoric cultural dynamics.
Prehistoric agriculture and anthropogenic ecology
269
Figure 15.4 Gridded gardens of fields outlined by low rock walls, near Safford, Arizona Photograph: P.Minnis Rock mulching Rock mulching involves planting crops in piles of stones, or covering the field’s surface with stones, and is used worldwide (Lightfoot, 1996). The rock mulch conserves moisture and can have other benefits such as protecting roots from rodent predation. Like the other agricultural types mentioned here, rock mulching is found in many areas of the region. Examples are known in the north near Santa Fe, New Mexico (Anschuetz, 1995; Ford, 2000; Lightfoot, 1996; Maxwell, 1995; Maxwell and Anschuetz, 1992), but they are best known from the Sonoran desert of central Arizona, where Suzanne and Paul Fish and their collaborators have documented the widespread use of rock-mulch piles for the cultivation of maguey, the century plant (Agave sp.) (Fish et al., 1985). They estimate that up to 50,000 such piles are present in the foothills north of Tucson, indicative of the substantial cultivation of a plant previously thought to have been gathered only from naturally propagated stands (for the importance of maguey, see also Chapter 16). We recently discovered similar rock-mulch fields in Chihuahua (Minnis and Whalen, 1996), which are the first evidence of agave cultivation in the Chihuahuan desert (Figs. 15.5 and 15.6). As in the case of other forms of agriculture in the region that have been studied, production seems to have been small scale: each field consisted of a little less that 100 stone piles.
The archaeology of drylands
270
Figure 15.5 A rock mulch field near Casas Grandes, Chihuahua, Mexico Note: Presumably the century plant or maguey (Agave sp.) was grown here; note the intact rock pile in the foreground Photograph: P.Minnis
Figure 15.6 An excavated rock pile from the field shown in Figure 15.5 Photograph: P.Minnis
Prehistoric agriculture and anthropogenic ecology
271
Despite being concerned with a small area of the world, a century of intensive archaeological research combined with an excellent ethnographic record have led to the documentation of tremendous diversity in agro-ecological strategies. The research suggests that prehistoric people may well have been able to farm much of the region except for higher elevations and the most desolate desert plains. The sophisticated suite of agricultural techniques allowed people to farm a wide range of locations. Yields of irrigated flood plain with permanent water and fertile soils would, not surprisingly, have been the economic foundation for communities with the highest population densities, but elsewhere other techniques seem to have overcome low and erratic precipitation and occasionally poor soil fertility.
ANTHROPOGENIC EFFECTS OF FARMING All humans affect their natural environments. Despite the many claims to the contrary, this is as true for indigenous North Americans as for peoples elsewhere (e.g. Denevan, 1992; Krech, 1999; Minnis and Elisens, 2000). Examples of the small-scale alterations from the region in prehistoric times include expanding the range of some plants such as Parry’s agave (Agave parryi) (Minnis and Plog, 1976), pruning the Douglas fir (Pseudotsuga menziesii) to yield beams at Mesa Verde (Nichols and Smith, 1965) and the manipulation of squawbush (Rhus trilobata) to produce unusually elongated stems for basketry (Bohrer, 1983). Fire is one of the most widely documented ethnographic examples in North America of anthropogenic ecology (e.g. Denevan, 1992; Dobyns, 1981; Krech, 1999; Mills 1986). It has been presumed that the suppression of both naturally and humanly set fires was a major factor leading to the modern invasion of shrubs into desert grasslands (Hastings and Turner, 1965; Humphrey, 1987). While I suspect that this model is correct and that prehistoric peoples did, in fact, set fires for a variety of reasons, the evidence of burning in the archaeological record is modest. Bohrer (1992) discusses small-scale burning by prehistoric Hohokam in the Sonoran desert. Except for fire, most effects of anthropogenic ecology in the prehistory of the region appear to have been very limited. By its very nature, however, agriculture alters environments, and such alterations have the potential to affect ecological patterning widely. Three potential ecological consequences of farming are briefly outlined here: deforestation; an increase in weeds; and soil modification. Deforestation Humans use wood, and often lots of it, for fuel and for their material culture. In addition, woodland agriculturalists remove tree cover for fields. While deforestation in prehistory seems not to have been as severe an ecological problem in the region as in some areas today (such as in Nepal, for example), there are some documented cases here of woodland reduction by prehistoric peoples. Wyckoff (1977), as an early case, noted a significant increase in arboreal pollen, particularly pine, juniper and oak (Quercus), following the abandonment of Mesa Verde in southwestern Colorado by prehistoric
The archaeology of drylands
272
peoples. This change, he suggested, was best explained as woodland recovery once human wood-harvesting pressures were relaxed or ended, though it could as easily have been due to the successional re-establishment of woodlands on abandoned fields. I documented a dramatic decline in riparian wood (mostly cottonwood/ willow, Populus/Salix) during the Classic Mimbres period (AD 1000–1150), which was the time of highest population density in the Mimbres valley of southwestern New Mexico (Minnis, 1985). The frequency of these woods then recovered when there were less dense prehistoric occupations. This small valley had a limited floodplain—the location for the most productive and reliable farming—and estimates of field requirements for various time periods indicate that the population of the Classic Mimbres expanded beyond the ability of the floodplain to support it. The increased presence of small villages and agricultural features in upland, secondary farming locations at this time is consistent with these estimates. Therefore, it seems that the riparian trees were removed for field clearance during the period of highest population density. The Chaco Canyon area of northwestern New Mexico offers another possible example of deforestation. Betancourt (1990) noted a clear reduction of piñon pine (Pinus edulis) wood from packrat middens during the height of the human population in the Chaco Canyon area of northwestern New Mexico. He interprets this pattern as decimation of local woodlands through human wood harvest; unlike in the previous cases, he argues that there was no documented recovery of piñon after the human abandonment of the region. Hall (1985) reviewed the pollen records from Chaco Canyon, suggesting that the Chacoan area of northwestern New Mexico was shrub and grasslands with only scattered, low-density piñon and juniper populations, which were species already growing in suboptimal conditions. While humans may well have reduced the woody plants, these conifers were not major components of the vegetation. Furthermore, Hall sees a slight increase in pine pollen after the prehistoric abandonment of the region. While further research is needed better to understand the human ecology of the Chaco Canyon area, both studies provide evidence of woodland reduction, perhaps due to field clearance. Increase in weeds A second likely environmental effect of prehistoric farming is an increase in ‘weedy’ species. Agriculture can increase the abundance of these plants in two ways. First, soil preparation in fields often presents ideal settings for such plants. Second, and less directly, as long as agriculture encourages sedentism, more soil will be disturbed by daily activities beyond farming. Seeds of weedy genera, particularly goosefoot (Chenopodium), pigweed (Amaranthus) and purslane (Portulaca), are some of the most ubiquitous remains found by flotation in archaeological sites of prehistoric villages in the region. These genera, together with the groundcherry (Physails), are some of the most common remains from prehistoric faeces from the northern part of the Southwest (Minnis, 1989). Seeds of weeds are also common constituents of paleoethnobotanical assemblages from the Sonoran desert (e.g. Gasser and Kwiatkowski, 1991). In fact, as Ford (1981) and others have pointed out, these weed seeds can constitute an important and welcome garden resource for human consumption.
Prehistoric agriculture and anthropogenic ecology
273
Soil modification Soil modification is the third anthropogenic effect of agriculture considered here. One of the best known deleterious effects of agriculture in arid areas is salinization, a topic discussed elsewhere in this volume (for example, Chapter 6). If salinization was a problem in the prehistory of the region, it would most likely have occurred in the large irrigation systems of the hottest desert near Phoenix, Arizona, where there were intensive crop production and very high evaporation rates. There has been some speculation that sporadic fields in the Sonoran desert were affected by salinization, largely based on historic records of such problems in a few locations (Susan Fish, pers. comm.). However, there is no compelling archaeological evidence that salinization was an important contributing factor to the abandonment of these systems (Krech, 1999). There is, in contrast, evidence of smaller-scale soil modification due to agriculture. Scholars working in the Dolores area of southwestern Colorado have used settlement locations to argue that dryland farming was especially important here for determining population and settlement dynamics through time, even though soil modifications have not been observed (e.g. Kohler et al., in press; Van West, 1994). Sandor (1990, 1992, 1995) found that soils behind trincheras in southwestern New Mexico still seem to show the effects of nutrient depletion after hundreds of years since their last use.
PREHISTORIC AGRICULTURE AND HISTORICAL DYNAMICS Although it is perhaps unfashionable now to view environmental conditions and fluctuations as important considerations in understanding the historical dynamics of ancient groups, there is sufficient research in this region to demonstrate such linkages. As expected in a semi-arid to arid area, variations in precipitation seem to have had the most profound effects on prehistoric farmers (e.g. Dean et al., 1985; Euler et al., 1979; Gumerman, 1988; Minnis, 1985; Petersen, 1988; Tainter and Tainter, 1996; Van West, 1994). Again, we can turn to the Mimbres valley of southwestern New Mexico for an example (Minnis, 1985). As outlined previously, human populations grew from at least AD 200 through to AD 1150, with a dramatic population peak during the Classic Mimbres period (AD 1000–1150). Analysis of demography and field requirements suggests that farmers of this period needed to utilize non-floodplain fields, usually in upland settings that were not only less productive but were also more vulnerable to precipitation variation than the floodplain fields. Consistent with this argument is the fact that there was an increased occupation of upland settlement during the Classic Mimbres period. The increased reliance on secondary field locations worked for a time because (according to dendroclimatological records) the first part of the Classic Mimbres period enjoyed an unusually favourable climatic regime. During the latter part of the Classic Mimbres period, the climate returned to a historically more typical pattern, so that populations dependent on upland farming had serious problems provisioning themselves. These problems were exacerbated by the fact that the society seems to have been characterized
The archaeology of drylands
274
by intensified local economic inter-relationships. More intense interdependence increased the social, political and economic impacts of the deterioration in the farming system. This may well have then reverberated throughout the population and contributed to the collapse of their regional system around AD 1130–1150. Low precipitation is, however, only one environmental factor in understanding the role of farming in the ancient history of the region. Graybill and Nials (1989), for example, argue that too much, rather than too little, water caused the destruction of the canals around Phoenix by flooding in the mid-1400s, and that this may have been a significant contributor to the collapse of the political structure. Numerous scholars have also noted the relationship between the organization of the irrigation systems and the socio-political landscape: those who controlled flow presumably had some power or at least advantage over downstream villages. This was certainly the case for the Hohokam (Gumerman, 1991) and was probably also so for Casas Grandes (Lekson, 1999). Finally, Cordell (1999) suggests that, in the final analysis, the ability of the Anasazi of Arizona to move over the landscape was the critical characteristic that allowed them to farm and survive for centuries in environments not especially benign for plant cultivation. This mobility would have been a, if not the, critical means by which the prehistoric peoples dealt with changing agricultural conditions and the anthropogenic effects of their activities.
CONCLUDING THOUGHTS Understanding the prehistoric human ecology of the North American Southwest has valuable lessons. It is obvious that there is a great diversity of prehistoric agricultural strategies in the prehistoric record, adapted to a wide range of environmental conditions. What is less obvious is how these data might be of practical use in the area where industrial-scale agriculture is now articulated with a capitalist economy, since, with rare exception, indigenous prehistoric farmers in the region were not as concerned with surplus production. It is unlikely that the prehistoric techniques will fit directly into the modern context, although the principles underlying traditional agriculture may be useful. One could conceive, for example, of how rock mulching—a relatively environmentallybenign activity—might be used in modern arid-land farming. More likely, indigenous farming strategies may well find some use in household gardening or ‘boutique farming’, even in densely urban settings within the region where mechanization is essential. And, of course, the techniques practised by the indigenous prehistoric farmers of the region might be transferable to other arid and semi-arid areas of the world where smaller-scale crop production is economically viable and where food production continues to expand in previously unused or underutilized and often marginal locations. Painting in the broadest strokes, I have argued that the prehistoric populations of this arid region affected their biotic environments. As severe as these impacts may have been for the indigenous peoples and for the local ecology of the time—and no doubt there were serious problems on occasion—no lasting ecological alterations occurred. I say this with the caveat that more study of desert grassland fire frequency, and of its causes, would be useful. Therefore, modern environmental planners in the region will be served better by studies of possible small-scale anthropogenic ecology, rather than of
Prehistoric agriculture and anthropogenic ecology
275
widespread, general changes due to prehistoric humans: like politics, anthropogenic ecology is local. Still we should not conclude, as some would like, that indigenous peoples were environmentally neutral. Within the region, Dobyns (1981) points out that the ecologically-harmful effects of livestock occurred among indigenous peoples once they acquired exotic domestic livestock. From the wider geographic focus on North America, Krech (1999) argues that the ‘Indian-as-ecologist’ image is misleading and unjustified, which is a point also made by Denevan (1992). I agree, but suggest that this misses the most important point. Whether one characterizes Native Americans as preservationists, conservationists or ecologists, this is less important than understanding how they interacted with their environment, including understanding how they farmed. There are real lessons to be learned: the evidence for less substantial ecological consequences in prehistory compared with today is due to relatively low population density, the infrequency of stratified societies with economies geared toward substantial surplus production, and the rather high level of residential relocation in prehistory. In short, few people, staying in locations for relatively short periods of time, with a familial mode of production, simply did not impact the environment as much as historical populations with relatively high population densities (rural as well as urban), industrial development, large-scale mechanized agriculture, exotic species introduced from elsewhere, and effective fire suppression. Human ecology is a matter neither of mystic and romantic ideology nor simply of indigenous cosmology: it must be grounded in an understanding of historical ecology and biology.
ACKNOWLEDGEMENTS I would like to acknowledge the assistance and support of Michael Whalen of the University of Tulsa, who is a good colleague and my co-director on a long-term project in Chihuahua, Mexico, and also thank Patricia Gilman of the University of Oklahoma for commenting on previous drafts of this text.
REFERENCES Anschuetz, K.F. (1995) Saving a rainy day: the integration of diverse agriculture technologies to harvest and conserve water in the Lower Chama Valley, New Mexico. In W.Tool (ed.) Soil, Water, Biology, and Belief in Prehistoric and Traditional Southwestern Agriculture: 25–40. Albuquerque, New Mexico Archaeological Council Special Publication 2. Bahre, C.J. (1991) A Legacy of Change: Historic Human Impact on Vegetation of the Arizona Borderlands. Tucson, University of Arizona Press. Betancourt, J.L. (1990) Late Quaternary biogeography of the Colorado Plateaus. In J.Betancourt, T.Van Devender and P.Martin (eds) Packrat Middens: The Last 40,000 Years of Biotic Change: 259–93. Tucson, University of Arizona Press. Betancourt, J.L., Van Devender, T.R. and Martin, P.S. (1990) Packrat Middens: The Last
The archaeology of drylands
276
40,000 Years of Biotic Change. Tucson, University of Arizona Press. Bohrer, V.L. (1983) New life from ashes: the tale of the burnt bush (Rhus trilobata). Desert Plants 5:122–4. Bohrer, V.L. (1992) New life from ashes II. Desert Plants 10:122–5. Bradfield, M. (1971) The Changing Pattern of Hopi Agriculture. London, Royal Anthropological Institute of Great Britain and Ireland Occasional Papers 30. Brown, D.E. (1982) Biotic communities of the American Southwest United States and Mexico. Desert Plants 4:1–342. Castetter, E.F. and Willis H.B. (1942) Pima and Papago Indian Agriculture. Albuquerque, University of New Mexico Press. Castetter, E.F. and Willis H.B. (1952) Yuman Indian Agriculture. Albuquerque, University of New Mexico Press. Cordell, L.S. (1977) Predicting site abandonment at Wetherill Mesa. The Kiva 40: 189– 202. Cordell, L.S. (1997) Prehistory of the Southwest. San Diego, Academic Press. Cordell, L.S. (1999) Succeeding in agriculture in the Anasazi way. New Mexico Journal of Science 39. Dart, A. (1989) Prehistoric Irrigation in Arizona. Tucson, Institute for American Research Technical Report 89–1. Dean, J.S., Euler, R.C., Gumerman, G.J., Plog, F., Hevly, R.H. and Karlstrom, T.N.V. (1985) Human behavior, demography, and paleoenvironment on the Colorado Plateaus. American Antiquity 50:537–54. Denevan, W.M. (1992) The pristine myth: the landscape of the Americas in 1942. Annals of the Association of American Geographers 82:369–85. Di Peso, C.C. (1974) Casas Grandes: A Fallen Trading Center of the Gran Chichimeca. Flagstaff, Northland Press. Dobyns, H.F. (1981) From Fire to Flood: Historic Human Destruction of the Sonoran Desert Riverine Oases. Socorro, Ballena Press Anthropological Papers 20. Donkin, R.A. (1979) Agricultural Terracing in the Aboriginal New World. New York, Viking Fund Publications in Anthropology 56. Doolittle, W.E. (1990) Canal Irrigation in Prehistoric Mexico: The Sequence of Technological Change. Austin, University of Texas Press. Euler, R.C., Gumerman, G.J., Karlstrom, T.N.V., Dean, J.S. and Hevly, R.H. (1979) The Colorado Plateaus: cultural dynamics and paleoenvironments. Science 205: 1089–101. Fish, S.K. and Fish, P.R. (1984) Prehistoric Agricultural Strategies in the Southwest. Tempe, Arizona State University, Anthropological Research Reports 20. Fish, S.K. and Nabhan, G.P. (1991) Desert as context: the Hohokam environment. In G.Gumerman (ed.) Exploring the Hohokam: Prehistoric Desert Peoples of the American Southwest: 29–60. Albuquerque, University of New Mexico Press. Fish, S.K., Fish, P.R., Miksicek, C. and Madsen, J. (1985) Prehistoric Agave cultivation in southern Arizona. Desert Plants 7:107–12. Ford, R.I. (1981) Gardening and farming before AD 1000: patterns of prehistoric cultivation north of Mexico. Journal of Ethnobiology 1:6–27. Ford, R.I. (2000) Human disturbance and biodiversity diversity: a case study from northern New Mexico . In P.Minnis and W.Elisens (eds) Biodiversity and Native
Prehistoric agriculture and anthropogenic ecology
277
America. Norman, University of Oklahoma Press (in press). Gasser, R.E. and Kwiatkowski, S.M. (1991) Food for thought: recognizing patterns in Hohokam subsistence. In G.Gumerman (ed.) Exploring the Hohokam: Prehistoric Desert Peoples of the American Southwest: 417–59. Albuquerque, University of New Mexico Press. Gilman, P.A. and Sherman, S. (1975) An Archaeological Survey of the Graham-Curtin Project: Phase II. Tucson, Arizona State Museum, Cultural Resource Management Section Report. Graybill, D.A. and Nials, F.L. (1989) Aspects of climate, streamflow and geomorphology affecting irrigation systems in the Salt River valley. In C.A.Heatherington and D.A.Gregory (eds) The 1982–1984 Excavations at Las Colinas: Environment and Subsistence: 39–58. Tucson, Arizona State Museum Archaeological Series No. 162. Gumerman, G.J. (1988) The Anasazi in a Changing Environment. Cambridge, Cambridge University Press. Gumerman, G.J. (1991) Exploring the Hohokam: Prehistoric Desert Peoples of the American Southwest. Albuquerque, University of New Mexico Press. Hack, J.T. (1942) The Changing Physical Environment of the Hopi Indians of Arizona. Cambridge Mass., Harvard University, Peabody Museum Papers 31 (1). Hall, S.A. (1985) Quaternary pollen analysis and vegetational history of the Southwest. In V.Bryant and R.Holloway (eds) Pollen Record of Late-Quaternary North American Sediments: 95–123. Dallas, American Association for Stratigraphic Palynologists. Hard, R.J. and Roney, J.R. (1998) A massive terraced village complex in Chihuahua, Mexico, 3000 years before present. Science 279:1661–4. Hastings, J.R. and Turner, R.M. (1965) The Changing Mile: An Ecological Study of Vegetation Change with Time in the Lower Mile of an Arid and Semiarid Region. Tucson, University of Arizona Press. Herold, L.C. (1970) Trincheras and Physical Environment Along the Rio Gavilan, Chihuahua, Mexico. Denver, University of Denver, Department of Geography, Technical Paper No. 65–1. Howard, J. (1993) A Paleohydraulic Approach to Examining Agricultural Intensification in Hohokam Irrigation Systems. Greenwich, JAI Press, Research in Economic Anthropology, Supplement 7. Howard, W.A. and Griffiths, T.M. (1966) Trinchera Distribution in the Sierra Madre Occidental, Mexico. Denver, University of Denver, Department of Geography, Technical Paper No. 66–1. Humphrey, R.R. (1987) 90 Years and 535 Miles: Vegetation Change Along the Mexican Border. Albuquerque, University of New Mexico Press. Kohler, T.A., Kresl, J., Van West, C., Carr, E. and Wilshusen, R.H. (in press) Be there then: a modeling approach to settlement determinants and spatial efficiency among late ancestral Pueblo populations of the Mesa Verde region, U.S. Southwest. In T.Kohler and G.Gumerman (eds) Dynamics of Human and Primate Societies: Agent-Based Modeling of Social and Spatial Processes. Oxford, Oxford University Press. Krech, S. (1999) The Ecological Indian: Myth and History. New York, W.W.Norton. Lekson, S.H. (1999) The Chaco Meridian: Centers of Political Power in the Ancient Southwest. Walnut Creek, Altamira Press.
The archaeology of drylands
278
Lightfoot, D.R. (1996) The nature, history and distribution of lithic mulch agriculture: an ancient technique of dryland agriculture. Agricultural History Review 44: 206–22. Maxwell, T.D. (1995) A comparative study of prehistoric farming strategies. In H. Toll (ed.) Soil, Water, Biology and Belief in Prehistoric and Traditional Southwestern Agriculture: 3–10. Albuquerque, New Mexico Archaeological Council, Special Publication 2. Maxwell, T.D. and Anschuetz, K.F. (1992) The southwestern ethnographic record and prehistoric agricultural diversity. In T.Killion (ed.) Gardens of Prehistory: The Archaeology of Settlement Agriculture in Greater Mesoamerica: 35–68. Tuscaloosa, University of Alabama Press. Mills, B.J. (1986) Prescribed burning and hunter-gatherer subsistence systems. Haliksa’i: UNM Contributions to Anthropology 5:1–26. Minnis, P.E. (1985) Social Adaptation to Food Stress: A Prehistoric Southwestern Example. Chicago, University of Chicago Press. Minnis, P.E. (1989) Prehistoric diet in the northern Southwest: macroplant remains from Four Corners feces. American Antiquity 54:543–63. Minnis, P.E. and Elisens, W.J. (2000) Biodiversity and Native America. Norman, University of Oklahoma Press (in press). Minnis, P.E. and Plog, S.E. (1976) A study of the site specific distribution of Agave Parryi in east central Arizona. The Kiva 41:299–308 Minnis, P.E. and Whalen, M.E. (1996) Prehistoric Upland Agriculture in the Casas Grandes Core. Washington, D.C., final project report submitted to the National Geographic Society. Nabhan, G.P. (1979) The ecology of floodwater farming in the arid southwestern North America. Agro-Ecosystems 5:245–55. Nabhan, G.P. (1986a) Papago Indian desert agriculture and water control in the Sonoran desert, 1697–1934. Applied Geography 6 (1):42–3. Nabhan, G.P. (1986b) ‘Ak-ciñ’ “arroyo-mouth” and the environmental setting of Papago Indian fields in the Sonoran desert. Applied Geography 6 (1):61–75. Nabhan, G.P. and Sheridan, T.E. (1977) Living fencerows of the Rio San Miguel, Sonora, Mexico: traditional technology of floodplain management. Human Ecology 5:97–111. Nichols, R.F. and Smith, D.G. (1965) Evidence of prehistoric cultivation of Douglas-Fir trees at Mesa Verde. American Antiquity Memoir 31 (2):57–64. Petersen, K.L. (1988) Climate and the Dolores River Anasazi. Salt Lake City, University of Utah Anthropological Papers 113. Plog, S. (1997) Ancient Peoples of the American Southwest. London, Thames & Hudson. Rzedowski, J. (1986) Vegetation de Mexico. Mexico, D.F., Editorial Limusa. Sandor, J.A. (1990) Prehistoric agricultural terraces and soils in the Mimbres area, New Mexico. World Archaeology 22:166–80. Sandor, J.A. (1992) Long-term effects of prehistoric agriculture on soils: examples of New Mexico and Peru. In V.Holliday (ed.) Soils in Archaeology: Landscape Evolution and Human Occupation: 217–45. Washington, D.C., Smithsonian Institution Press. Sandor, J.A. (1995) Searching soil for clues about Southwestern prehistoric agriculture. In H.W. Toll (ed.) Soil Water, Biology and Belief in Prehistoric and Traditional
Prehistoric agriculture and anthropogenic ecology
279
Southwestern Agriculture: 119–37. Albuquerque, New Mexico Archaeological Council Special Publication 2. Schmidt, R.H. Jr. and Gerald, R.E. (1988) The distribution of conservation type watercontrol systems in the northern Sierra Madre Occidental. The Kiva 53: 165–79. Sellers, W.D. and Hill, R.H. (1974) Arizona Climate. Tucson, University of Arizona Press. Staten, G., Burnham, D.R. and Carter, J. Jr. (1939) Corn Investigations in New Mexico. Las Cruces, New Mexico State University, Agricultural Experiment Station Bulletin 260. Tainter, J.A. and Tainter, B.B. (1996) Evolving Complexity and Environmental Risk in the Prehistoric Southwest. Reading, Addison-Wesley Publishing. Toll, H.W. (1995) Soil, Water, Biology and Belief in Prehistoric and Traditional Southwestern Agriculture. Albuquerque, New Mexico Archaeological Council Special Publications 2. Tuan, Y.-F., Everard, C.E., Widdison, J.G. and Bennett, I. (1973) The Climate of New Mexico. Santa Fe, State Planning Office. Van West, C.R. (1994) Modeling Prehistoric Agricultural Productivity in Southwestern Colorado: A CIS Approach. Pullman, Washington State University, Department of Anthropology Reports of Investigations 67. Vivian, R.G. (1991) Chacoan subsistence. In P.J.Crown and W.J.Judge, Chaco and Hohokam: Prehistoric Regional Systems in the American Southwest: 57–76. Santa Fe, School of American Research Press. Whalen, M.E. and Minnis, P.E. (1996) El Sistema Regional de Paquimé, Chihuahua, Mexico. Mexico D.F., Informe Téchnio Final presented to the Consejo de Arqueología, Institute Nacional de Antropología e Historia. Woodbury, R.B. (1961) Prehistoric Agriculture at Point of Pines, Arizona. Salt Lake City, Memoirs of the Society for American Archaeology 26 (3) part 2. Wyckoff, D.G. (1977) Secondary forest succession following abandonment of Mesa Verde. The Kiva 42:215–32.
16 The role of maguey in the Mesoamerican tierra fría: ethnographic, historic and archaeological perspectives JEFFREY R.PARSONS AND J.ANDREW DARLING
INTRODUCTION Pre-columbian civilizations in Mesoamerica flourished in three very different environments: the tierra caliente—warm, humid and thickly forested lowlands below 1,000 m above sea level along the Atlantic, Caribbean and Pacific coasts of Mexico and adjacent Central America; the tierra templada—subhumid to semi-arid, frost-free, temperate highlands between 1,000 and 1,800 m in Guatemala and southern Mexico; and the tierra fría—semi-arid and arid highlands, with average annual rainfall as low as 300 mm, and with severe winter frosts, at elevations above 1,800 m in central and northcentral Mexico (Fig. 16.1). Mesoamerica was one of the world’s hearths of early plant domestication, and agriculture provided the economic basis of the chiefdoms, states and empires that developed there after c.1500 BC (Flannery, 1973; MacNeish, 1991). Yet it was the only one of the world’s ancient primary civilizations that lacked a domestic herbivore. Through the use of domestic camelids (llamas and alpacas) in the central Andes, and sheep and goats in much of the Old World, food producers in virtually all other regions where ancient states and empires existed were able significantly to extend their productive landscapes into drier and colder zones and over a full annual cycle. Some of them became full- or part-time herders, and herder-cultivator relationships became important in the long-term development of socio-political complexity. In this chapter we address the question of how ancient Mesoamericans, with their seemingly more limited capacity to generate and manipulate energy, could have attained a level of organizational complexity on a par with that of the central Andes and several Old World regions where agriculture and pastoralism were combined in antiquity. This question becomes increasingly important because the largest polities of ancient Mesoamerica—Teotihuacan between AD 200 and 600, Tula between AD 900 and 1200 and the Aztecs of Tenochtitlan and their neighbours between AD 1300 and 1520—were all centred in the comparatively cold and dry tierra fría. We are particularly
The role of maguey in the Mesoamerican tierra fría
281
Figure 16.1 Middle America, showing the approximate extent of the tierra fría (shaded) interested in understanding how the resources of the tierra fría underwrote the development of Mesoamerica’s largest polities in the face of the winter frosts and low seasonal rainfall that limited seed-based agriculture to one crop per year, even in those comparatively few zones where irrigation was able to overcome the constraints of aridity. Our focus in this chapter is on several species of domestic agave cactus that have come to be known collectively as ‘maguey’ in the Mexican highlands. The most important maguey species include Agave salmiana, A. magisapa, A.atroviens, A.ferox, A.hookeri and A.americana (Gentry, 1982). Cultivated maguey is still an important component of agriculture throughout the Mexican tierra fría today (Fig. 16.2), and it is known to have played a significant role in the economy of this region for thousands of years (Parsons and Parsons, 1990). Other species of agave are cultivated in other parts of Mesoamerica, but these are invariably of secondary importance. Most archaeologists working in Mesoamerica have overlooked the full significance of maguey. With their interests dominated by the cultivation of annual seed crops (primarily maize, beans, amaranth and squash), archaeologists have tended to ignore or downplay the potential of other types of food production (cf Mangelsdorf et al., 1964; Puleston, 1968, 1973; Willey et al.,
The archaeology of drylands
282
Figure 16.2 Field of cultivated magueys; all the sap has been extracted from the plant in foreground Photograph: J.Parsons 1964). An extreme expression of this viewpoint is the assertion by Blanton et al. (1981:174) that ‘in the highland valleys [of Mesoamerica] the surest way of producing a large surplus was to plant maize everywhere’. By contrast, we argue that maguey and seed crops were fully complementary in the Mesoamerican tierra fría, and that maguey there made available some of the same kinds of coping strategies complementary to seed cultivation as did llamas and alpacas in the central Andes and sheep and goats in Mesopotamia and elsewhere in the Old World. It is important to note that, as sheep, goats, pigs and cattle became increasingly important as introduced sources of food and fibre in highland Mexico after European contact in the early sixteenth century (Crosby, 1972, 1986), so too there was a corresponding decline in the importance of maguey—a decline that has continued, at an accelerating pace, down to the present day. The purpose of this chapter is to develop the following inter-related hypotheses: • in the Mesoamerican tierra fría, the development of complex society during the Middle and Late Formative (Table 16.1) depended upon the domestication of maguey as a primary complement to seed crops for the production of food and fibre (an idea originally advanced by Sauer, 1941);
The role of maguey in the Mesoamerican tierra fría
283
Table 16.1 The prehispanic chronology of central Mexico Date Period Phase AD 1520 Aztec IV Late Postclassic Aztec III AD 1350 Middle Postclassic Aztec I–II AD 1150 Early Postclassic Mazapan AD 950 Epiclassic Coyotlatelco AD 700 Metepec Xolalpan Classic Tlamimilolpa AD 150 Miccaotli Tzacualli 50 BC Terminal Formative Patlachique 250 BC Late Formative Ticoman 500 BC Middle Formative La Pastora 900 BC El Arbolillo Early Formative Bomba 1200 BC Ixtapaluca
• the expansion of Mesoamerican civilization into the drier highland regions of central and north-central Mexico depended upon the full integration of seed-based and maguey-based agricultural production; • agricultural production in the drier highland regions of central and especially northcentral Mexico was based upon the generalized production of both seed crops and maguey in comparatively well-watered core areas (the irrigable river valleys) and more specialized production of maguey and probably nopal (Opuntia sp., another domesticated cactus) in the drier peripheral zones, i.e. the more extensive piedmonts and plains beyond the reach of effective irrigation; • the archaeological record hints at a major change in the technology of maguey production in central and north-central Mexico after the Classic period; this technological change is suggestive of some basic differences in the larger political economies of classic and Postclassic states in highland Mesoamerica. In developing and addressing these ideas, we are constrained by serious limitations of the known archaeological record of prehistoric maguey utilization. Few archaeologists have investigated maguey production, and many remain unaware of its key archaeological
The archaeology of drylands
284
correlates. Much of north-central Mexico remains archaeologically understudied, and so it is difficult to make good inferences on the basis of archaeological data from the region where maguey production was probably most critical in the prehistoric economy. Consequently, we are able to make very few definitive statements based upon archaeological remains about the specifics of how maguey was actually used at different times and places in the past. This chapter is thus very much an exercise in hypothesis building, in which we rely primarily on ethnographic observations and historical documentation and only secondarily on archaeological data. We begin by highlighting maguey’s importance as a source of food and fibre in contemporary highland Mexico. We use these contemporary data to quantify the potential need and availability of maguey sap, flesh and fibre in pre-columbian times. We also employ analogies from the technology of historic maguey utilization to infer some of the archaeological correlates of prehistoric maguey production. We conclude by combining ethnographic, historic and archaeological data to develop hypotheses for the role of maguey in the classic and Postclassic expansion of Mesoamerican civilization into the tierra fría from its Early and Middle Formative bases in warmer and more humid regions.
MAGUEY AS A SOURCE OF FOOD AND FIBRE Investigations over the past century have produced considerable information about the importance of maguey as a source of food and fibre for thousands of years in highland Mexico (Beals, 1932; Flannery, 1968, 1986, et al., 1981; Goncalves de Lima, 1978; Guerrero, 1980; Healan, 1977; Hough, 1908; MacNeish et al., 1967; Smith, 1967; Smith and Kerr, 1968; Taylor, 1966). Ethnographers have described the cultivation of maguey and the use of its sap, fibre and flesh. Historians have found references to maguey cultivation and use in written documents that extend back to the early sixteenth century. In their midden excavations and surface surveys, archaeologists have found the physical remains of agave fibre, plus spindle whorls used in the spinning of maguey fibre and scraping tools used in the extraction and processing of sap and fibre. Nevertheless, the technology and organization of pre-columbian maguey utilization have remained poorly understood. Recent ethnographic research on maguey use in central and north-central Mexico (Parsons and Parsons, 1990; Patrick, 1985; Rangel, 1987; Ruvalcaba, 1983; Salinas and Bernard, 1983; Sanchez, 1980) has provided some new insights into prehistoric maguey utilization. In the next few paragraphs we shall briefly highlight some aspects of this research. Maguey sap and flesh The maguey plant provides a rich store of both sap and edible flesh. Maguey sap is acquired for human use by means of procedures that interrupt the final stage of a plant’s normal seven to twenty-five year maturation process, in order to extract the sap through an initial ‘castration’ (a procedure that halts the natural flow of sap to an emergent seedbearing stalk: Fig. 16.3), followed by daily scraping and extraction operations over a period of three to six months. Individual plants in cultivated fields typically approach
The role of maguey in the Mesoamerican tierra fría
285
maturity continuously throughout the year. The timing of their planting and replacement is often explicitly managed so as to ensure continuous productivity, with no more than 5– 10 per cent of a field’s maguey plants producing sap at any particular point in time (Parsons and Parsons, 1990). Over its three to six month production period, a single maguey plant yields several hundred litres of sap, and a hectare of land typically yields 5,000–9,000 litres of sap per year (Parsons and Parsons, 1990:338). The sap may be allowed to ferment to form a beerlike liquid (pulque), or it may be consumed in its unfermented liquid form (aguamiel), or it may be boiled down to form thick syrup or solid sugar. Aguamiel and pulque are unstable, and cannot remain unused for more than about a week. As syrup or sugar, however, maguey sap is much more durable, and in these forms sap surpluses can be stored and redistributed over a period of many months, or even longer. The modern Tarahumara of northern Mexico extract agave sap, for the preparation of a fermented beverage, by simply mashing up the plant’s leaves and squeezing out the liquid in a single operation (Bye et al., 1975). As we shall note below, there is reason to think that something analogous to this less-efficient Tarahumara procedure (‘less efficient’ in the sense that not all the plant’s sap can be extracted in this manner, and the plant’s fibre and flesh are usually not utilized at all) may have characterized maguey sap extraction during the Formative and Classic periods in central Mexico, prior to the
The archaeology of drylands
286
Figure 16.3 Castrating a mature maguey plant Photograph: J.Parsons implementation in the Postclassic period of the more efficient techniques observed ethnographically in central Mexico. The leaves, heart and stalk of the maguey plant can also be cooked and eaten, as is still commonly done among more isolated groups in central and northern Mexico. The Tarahumara, for example, prepare cakes of baked maguey flesh, which can be stored for up to six months (Bye et al., 1975). Maguey sap and flesh are rich in both nutrients and calories. Ruvalcaba (1983:89) cites analyses showing that one litre of pulque contains 574 calories. Davidson and Ortiz de Montellano (1983:155) report that one tablespoon of maguey sap contains (among other things) 0.08 g of protein, 5.35 g of carbohydrates, 20 calories, 0.33 mg of Vitamin C, 0.02 mg of calcium, 5.03 mg of phosphorous, 12.7 mg of potassium, 30.0 micro-grams of iron, 17.0 micro-grams of magnesium, 9.0 micro-grams of selenium, 6.0 micro-grams of
The role of maguey in the Mesoamerican tierra fría
287
chromium and 4.0 micro-grams of zinc. In the early 1940s, Anderson et al. (1946:888) found that in the diets of their study group of rural highland villagers, pulque supplied 12 per cent of total calories, 6 per cent of total protein, 10 per cent of total thiamine, 24 per cent of total riboflavin, 23 per cent of total niacin, 48 per cent of total Vitamin C, 8 per cent of total calcium and 20 per cent per cent of total iron. Ross (1944, cited in Fish et al., 1986) found that 100 g of cooked agave flesh contains 347 calories and 4.5 g of protein. It appears that, in most tierra fría contexts, maguey can produce approximately as many calories and essential nutrients per hectare as the standard seed crops, and that when the plant’s flesh and sap are both consumed, maguey can potentially produce more calories than seed crops on a given unit of land (Parsons and Parsons, 1990:337, 338, 345). Only on irrigated land are seed crops significantly more productive than maguey. Critically, though, maguey can be interplanted with seed crops in virtually all agricultural settings, and when this is done (as it commonly has been throughout the historic period in tierra fría contexts where subsistence agriculture remains the norm), the overall nutritional and energetic output on a given unit of land is potentially doubled. Combining maguey and seed crops, therefore, would have maximized subsistence security for pre-hispanic agriculturalists in the tierra fría: annual energy productivity on most kinds of cultivated land could have been doubled; agricultural productivity could have been extended over a full annual cycle; agricultural productivity could have been extended into nearby drier, colder and less fertile areas, which are marginal for seed crops; and the year-round productivity of maguey could have been combined with the long-term storability of seed crops. Recent ethnographic studies (Parsons and Parsons, 1990:31) also reveal that maguey-sap exploitation can easily be deferred to the winter agricultural off-season (because the collection of the matured plant’s sap can be postponed for up to six months after the initial castration operation without any apparent loss in productivity), thereby reinforcing the complementarity between maguey and seed crops. Furthermore, because of its resistance to drought, frost and hail—all common causes of seed-crop failure in the tierra fría—maguey stands out as an ideal highland famine food. For example, elderly people living in agriculturally-marginal parts of central Mexico vividly recall that, during the most violent years of the Mexican revolution (1913–17), when normal access to market produce (including maize and beans) was frequently disrupted by military hostilities, their families survived for weeks and months at a time on maguey and nopal products, which were readily available at all times of the year in their own fields (Parsons and Parsons, 1990:11). It might even be useful to think about the extent to which stands of wild or semi-wild maguey and nopal may have been deliberately extended so as to provide food for sedentary cultivators during times of serious crop scarcity, as was commonly done by prehistoric colonists with certain types of introduced wild or semi-wild plants in ancient Polynesia (Kirch, 1984:131–2). The extensive stands of wild maguey and nopal that today occur throughout the most marginal parts of the arid highlands in central and northcentral Mexico might be relicts of such pre-hispanic practices—the self-perpetuated descendants of semi-managed ancestors? Ethnographic and historic studies show that the organization of maguey exploitation can be quite varied. The management of maguey cultivation and the production of its sap,
The archaeology of drylands
288
flesh and fibre can be handled on any level from a nuclear family household up to the large commercial plantation (hacienda) employing several hundred workers organized within a hierarchical administration. There appear to be no inherent qualities of this plant that might require, or select for, either diffuse or centralized organization for its growth, cultivation, harvest, or for the extraction or processing of its products. Nevertheless, all our ethnographic and historic observations of maguey cultivation are from contexts where the production of maguey is directly combined with that of seed crops. This means that we lack historic analogies for fully specialized maguey agriculturalists, that is, where maguey cultivation might have been carried out separately from that of maize, beans, squash or amaranth. We shall need to remember this point in the concluding section of this chapter, where we propose that specialized maguey cultivation may have played a key role in the northward expansion of the prehistoric Mesoamerican frontier. Maguey fibre In pre-columbian Mesoamerica there were only two major kinds of fibre for making textiles: cotton and maguey (Anawalt, 1980, 1990). Cotton could not be grown in the tierra fría, and so maguey was the only important source of textile fibre that could be locally produced in the highlands of central and north-central Mexico. There are also suggestions that fine cotton cloth was reserved for the elite during later Postclassic times (Anawalt, 1980; Berdan, 1987; Duran, 1964:131), and so there would have been an even greater need for large quantities of maguey-fibre textiles in highland Mesoamerica. In their recent ethnographic work, Parsons and Parsons (1990) observed a sequence of steps by means of which the massive maguey leaf is softened through heating and rotting so that its hard flesh (which usually comprises more than 97 per cent of the plant’s weight) can be easily separated from the encased fibre by scraping. When properly managed, both the sap and fibre of an individual plant can be extracted for human use. This same study also revealed the critical importance of dried maguey stumps as fuel in areas where firewood is scarce or absent. Pre-hispanic highland populations living in dry, sparsely-forested terrain may have been as much interested in the fuel that maguey provided as they were in the food and fibre that the plant produced. De Sahagun (1969, volume 3:145), for example, specifically mentions the sale for fuel of dried maguey stumps and leaves in sixteenth-century market-places in Mexico City. Parsons and Parsons (1990:157) found that an average maguey leaf provides roughly 75 g of dried fibre. An average maguey plant has twenty to thirty leaves and thus provides approximately 2,000 g of dried fibre. An average modern carrying cloth (ayate) made of woven maguey thread measures about 1 m square and weighs about 200 g. Thus, one maguey plant provides enough fibre for about 10 m2 of cloth. More precise calculations would have to make allowance for variable thread thickness, thread spacing, fibre quality, type of costume and so on. Nevertheless, these rough estimates suggest that one maguey plant would have provided enough fibre for outfitting an average precolumbian person with most of the maguey-fibre textile required for clothing over a period of a few years. On an average cultivated hectare of land in highland central Mexico, about thirty maguey plants can be exploited each year for both sap and fibre (Parsons and Parsons,
The role of maguey in the Mesoamerican tierra fría
289
1990:336, 338). Thus, 1 ha of cultivated maguey could potentially outfit approximately thirty people with the maguey cloth they would need for a few (say three) years. Alternatively, assuming each average person requires one third of his/her wardrobe to be replaced each year, then 1 ha of cultivated maguey would provide the annual magueycloth needs for some ninety people. We can simplify our calculations by calling it an even 100. On this basis, a million people—approximately the number of people living in the Valley of Mexico (the core region of Aztec civilization in AD 1500)—would annually have required the fibre production (c.600,000 kg) of the cultivated maguey from some 10,000 hectares, which was roughly 5 per cent of the total arable landscape in the Valley of Mexico. This same amount of land could potentially, at the same time, have produced annually about 50–90 million litres of aguamiel, roughly 6,000 metric tons of cooked maguey flesh, perhaps 8,000–10,000 metric tons of interplanted maize or beans and many tons of dried maguey stumps for use as household fuel (Parsons and Parsons, 1990:337, 338). Obviously, the above figures require extensive refinement. For example, overall productivity of maguey and other crops is likely to have been significantly lower than the above-cited figures, which derive from central Mexico, in the increasingly more arid terrain of north-central Mexico. However, when taken in the spirit of very rough approximation, they seem useful at this stage of hypothesis building. When one considers these figures and remembers that maguey production (of both sap and fibre) can be deferred to the agricultural off-season, and that household spinning and weaving can also be relegated to the winter off-season period, then the complementarity of maguey and seed crop cultivation in the tierra fría becomes even more fully apparent, as does the greatly improved economic security the two cultivation systems provide in combination.
THE TECHNOLOGY OF PRE-HISPANIC MAGUEY USE Ethnographic and archaeological studies indicate that several categories of stone and ceramic tools can be confidently associated with some aspects of pre-hispanic maguey cultivation and processing. This section of the chapter highlights some of the best insights we now have about which archaeological implements can be linked with specific productive functions. Spinning Maguey fibre continues to be spun into thread using traditional drop-spinning techniques that employ wooden spindles and ceramic spindle whorls (Fig. 16.4). (Spindle whorls today are also sometimes made of stone, bone or wood.) We now have some good archaeological data on the nature and distribution of pre-columbian ceramic spindle whorls in central Mexico, and we can distinguish between small whorls (weighing less than c.7 gm) used for spinning thinner, lighter cotton fibre and large whorls (weighing more than c.11 gm) used for spinning thicker, heavier maguey fibre (e.g. Norr, 1987; Parsons, 1972; Sejourne, 1983; Smith and Hirth, 1988; Fig. 16.5). Studies of living spinners show that those whorls that weigh 20–30 gm can
The archaeology of drylands
290
Figure 16.4 Spinning maguey fibre, showing wooden spindle and ceramic spindle whorl in use Photograph: J.Parsons
The role of maguey in the Mesoamerican tierra fría
291
Figure 16.5 Pre-columbian spindle whorls used for spinning maguey fibre be used to produce a wide range of fine to coarse maguey thread, whereas lighter (c.11– 15 gm) and heavier (c.35–140 gm) whorls could only have been used to produce, respectively, a much narrower range of fine or coarse maguey thread (Parsons and Parsons, 1990:329, 331). Consequently, we now have a sense about how we might eventually be able to identify generalized versus specialized spinners in the archaeological record, once the right kind of archaeological information becomes available. This prospect becomes especially interesting in the light of historically-based discussions of the organization of spinning and weaving, and of the importance of textiles in tribute, market exchange, ceremonial presentations, and as markers of social status in both pre-columbian Mesoamerican and Andean civilizations (Anawalt, 1980, 1990; Carrasco, 1976; Hicks, 1987; Murra, 1962). Once we have better control over spindle whorl weights at specific spinning workshops, we should be in a much better position to infer the extent to which different spinners were involved in either specialized or generalized spinning, in the production of either cotton or maguey thread, and in tributary, market or domestic modes of production. We also suspect that the elaborate stamped, moulded and incised designs, so characteristic of Postclassic spindle whorls (Fig. 16.5), may relate to specific social units associated with particular kinds of whorl, thread and textile production (Parsons, 1975). Archaeologists have discovered that ceramic spindle whorls in highland central Mexico are extremely scarce before the Postclassic period. Numerous possibilities might
The archaeology of drylands
292
explain this, such as spinning without whorls, the use of perishable wooden whorls, or the use of simple perforated sherd disks that are not always recognized as spindle whorls by archaeologists. However, this clear-cut difference between Classic and Postclassic spinning technology is so dramatic as to suggest a major reorganization of spinning after the Classic period. This contrast may signify that in highland central Mexico spinning (and possibly weaving as well) became more specialized and more efficient during the Postclassic than it had been earlier. This, in turn, suggests changes in the organization of maguey production and fibre processing in the tierra fría. Carrasco’s (1976) discussion of different kinds of cloth production in commoner households and palace workshops in the early sixteenth century is certainly suggestive in this regard, as is Hicks’ (1987) emphasis on the importance of certain kinds of textiles in new forms of market-based redistribution in Late Postclassic times. Both Carrasco and Hicks have relied exclusively on sixteenth- and seventeenth-century historic sources to develop their ideas about textile production and distribution in the Aztec heartland. Archaeological data will surely extend and amplify these insights once more high-quality information about spindle-whorl variability over time and space becomes available. It is already well known, for example (Parsons, 1972), that substantial numbers of both maguey whorls and cotton whorls co-occur at many Postclassic sites in the Valley of Mexico (tierra fría where cotton cannot be grown locally) and in at least one Middle Postclassic site in the nearby tierra templada (Norr, 1987), where maguey has never been produced in historic times. This co-occurrence of cotton and maguey spinning in ecologically ‘inappropriate’ zones implies the existence of fairly complex redistributional networks for raw fibre, spun thread and woven textiles in central and north-central Mexico during the Postclassic. Scraping maguey fibre Today, maguey fibres are detached from the encasing flesh with an iron scraper mounted in a wooden handle (Fig. 16.6). These scrapers are dull, even-edged tools designed to scrape away the flesh without cutting or shredding the fibres. We think the pre-hispanic analogue is a trapezoidal ground-stone tool made of tabular basalt (Fig. 16.7)—a tool that is particularly common in the Later Postclassic (Brumfiel, 1976; Sanders et al., 1979; Tesch and Abascal, 1974) but that also occurs in at least one Late-Terminal Formative context in the Valley of Mexico (Serra Puche, 1988). Although some archaeologists have interpreted these implements as hoes associated with maize cultivation, recent experimental work shows that these implements are admirably suited for scraping maguey fibre (Parsons and Parsons, 1990:175; Fig. 16.8). These trapezoidal scrapers are quite widespread throughout the highlands of central and north-central Mexico and in the Southwest United States
The role of maguey in the Mesoamerican tierra fría
293
Figure 16.6 Use of modern iron scraper for extracting maguey fibre Photograph: J.Parsons (Brumfiel, 1976; Cabrero, 1989; Fish et al., 1986; Mastache et al., 1990; Sanders et al., 1979; Sejourne, 1983: Fig. 137; Spence, 1971; Tesch and Abascal, 1974; Trombold, 1985, 1989). Over time they tend to displace another distinctive tool: the scraper plane or ‘turtleback scraper’ (Tolstoy, 1971; Fig. 16.9). Experimental work with archaeological scraper-planes in the southern highlands of Mexico (Hester and Heizer, 1972) has shown that repeated downward blows with the rounded side of this tool (which typically weighs about 400 gm) are effective to mash up raw maguey leaves, while the flat bottom side of the same implement can serve to scrape the mashed flesh away from the fibres (using a lateral motion while bearing down on the pulpy mass of mashed leaf). Trapezoidal scrapers were probably used in more specialized maguey fibre production, in which greater efficiency in fibre extraction was achieved by means of cooking and rotting leaves to soften the flesh. The scraper plane would probably have predominated in the context of earlier and/or more generalized fibre production, where high efficiency was less important. If so, then increasing specialization and efficiency of maguey fibre processing (manifested archaeologically by a progressive shift from scraper planes to trapezoidal ground-stone scrapers) appear to have paralleled increasing efficiency in the spinning of maguey fibre (manifested archaeologically by a dramatic
The archaeology of drylands
294
Figure 16.7 Examples of pre-columbian trapezoidal tabular basalt scrapers increase in quantities and variability of ceramic spindle whorls) during the Postclassic times. The extraction of maguey sap There is a very distinctive and highly specialized modern iron tool used for the twicedaily scraping of the surface of the sap-collecting cavity in the maguey plant’s interior. The pre-hispanic analogue of this elliptical or circular iron scraper appears to be a distinctive plano-convex stone scraper (Fig. 16.10). This implement has a broad distribution in the highlands of central and north-central Mexico (Cabrero, 1989:234,
The role of maguey in the Mesoamerican tierra fría
295
238–41; Dibble and Anderson, 1963: Fig. 778; Gamio, 1979 [1922]: 214; Mastache et al., 1990:189; Meigham,
Figure 16.8 Experimental use of pre-columbian trapezoidal tabular basalt scraper Photograph: J.Parsons
Figure 16.9 A pre-columbian scraper plane (width c.12 cm) Source: Adapted from Hester and Heizer, 1972 1976; Michelet, 1984; Parsons and Parsons, 1990; Rodriguez, 1985:199; Sanders, 1965, 1966; Sanders et al., 1979; Spence, 1971; Trombold, 1985, 1989; Vaillant, 1931:417).
The archaeology of drylands
296
These studies indicate that this tool appears as early as the Late Formative and increases markedly in frequency by the Postclassic. There seems little doubt about its primary function. These distinctive scrapers apparently do not occur archaeologically in any significant numbers outside
Figure 16.10 Modern iron scraper and pre-columbian obsidian scrapers Note: Modern iron scraper for sap extraction (left) and pre-columbian obsidian scrapers were probably used for the same purpose (two at right); the handles of the two obsidian scrapers have been partly broken off Photograph: J.Parsons the tierra fría, which lends additional support to our belief that this artefact was used exclusively in the production of maguey sap. From these indications we can infer that, over time, maguey sap processing in central and north-central Mexico shifted from (1) something akin to the previously-noted ethnographic Tarahumara procedure, in which agave leaves are simply mashed up and the sap squeezed out in a single operation, to (2) something comparable to the historically-known process in central Mexico, in which both the sap and fibre of individual plants are extracted through specialized procedures over a period of several months. Once again, we suggest that this shift was in the interests of greater overall efficiency of plant use in increasingly specialized economies, stimulated by both the
The role of maguey in the Mesoamerican tierra fría
297
higher population densities and the increased tributary demands of Postclassic societies.
CONCLUSION The contributors to this book seek to address a series of key issues relating to how ancient agriculturally-based societies adapted to the constraints of aridity, and how they coped with the diverse cultural forces that acted upon them in the arid settings in which they developed and changed. This chapter has addressed a large region from the perspective of a particular type of agriculture, in the context of inadequate archaeological information. Our conclusions are thus necessarily generalized and tentative. Testing these hypotheses will involve the archaeological identification of maguey production and processing, and the comparison of tool kits and midden contents from sites in different parts of the tierra fría, both with each other and with those from lower elevations in more humid zones. It will also involve collecting a great deal more systematic archaeological information on regional settlement patterns in north-central Mexico, in order to provide information on variability over time and space in population size, socio-political hierarchy, sedentary versus mobile occupation, inter-regional exchange patterns, migration from one zone to another, and agricultural field systems and land tenure—none of which can presently be inferred in any satisfactory or credible way. Our conclusions are presented below, as eight principal points. (1) Since at least the Middle Formative, maguey cultivation has been an equal partner with seed crops in agricultural production in the Mesoamerican tierra fría. We doubt that agriculture without domestic maguey could have sustained pre-hispanic state-level society in this comparatively cold, dry part of Mesoamerica. (2) Maguey production, and agricultural production in general, remained generalized throughout most of the Formative in the Mesoamerican tierra fría, with no significant shifts towards greater specialization or efficiency until the development of increasingly complex and urbanized society late in the first millennium BC. With their increased overhead costs and greater spatial separation between food producers and food consumers, urbanized states from the early first millennium AD onwards would have needed to intensify and expand all types of agricultural production. (3) The northward expansion of Mesoamerican civilization into north-central Mexico in the Classic period was underwritten by the integration of, on the one hand, specialized maguey-nopal producers dispersed extensively in agriculturally more marginal landscapes and, on the other, more generalized seed crop—maguey cultivators living in more nucleated settlements in restricted, more productive, river valleys where the irrigation of seed crops was feasible. The effective integration of these agriculturallygeneralized cores and agriculturally-specialized peripheries would have been dependent upon the existence of redistributional networks large enough to move staples over significant distances in a regular and predictable manner. Because the scope and scale of pre-state, Formative-period, redistributive networks were restricted owing to their personalized, kinship-based character, Mesoamerican civilization could not have expanded northwards into north-central Mexico until the development of large states in central Mexico during the Early Classic (Braniff, 1989; Darling, 1998; Kelley, 1990;
The archaeology of drylands
298
Nelson, 1997; Trombold, 1990). It may be useful to think about the post-Formative expansion of complex society into the arid lands of north-central Mexico as a process somewhat analogous to the expansion of pastoralists into the dry steppes of Inner Asia after the late second millennium BC. Sahlins, for example, once suggested (1968:33–4, following leads by Lattimore, 1951 and Krader, 1957), that an effective adaptation by specialized pastoralists to the sparsely occupied grasslands of Inner Asia might not have occurred until there was enough pressure from expanding states in better-watered and longer-settled regions to the south, where generalized neolithic agriculturalists combining cultivation and herding had lived for many centuries. Were the Classic period maguey-and-maize cultivators of arid north-central Mexico the Mesoamerican counterparts of the first substantial numbers of specialized pastoralists who may have moved into the dry Inner Asian steppes after c.1500 BC in order to escape the tribute and labour service demands imposed on them by increasingly large and powerful Near Eastern and East Asian polities? Did intensified and more efficient maguey utilization provide some cultivators living in highland central Mexico during the era of state growth in the early first millennium AD with the means to escape the demands of their would-be overlords, by emigrating to and flourishing in the sparsely occupied drylands to the north? Alternatively, was the development of greater sociopolitical complexity in north-central Mexico during the Classic period primarily a product of indigenous populations of marginal agriculturalists and hunter-gatherers who developed more intensive forms of agriculture (including maguey production) and greater socio-political centralization, in order more effectively to exploit the opportunities to acquire new wealth and new types of prestige-building exotica that were increasingly available from developing state systems along their southern flanks in central and western Mexico? (Barfield [1989] presents an intriguing Old World analogy that extends the earlier thinking of Lattimore [1951].) (4) The transition from the Classic to the Postclassic in central and north-central Mexico saw the development of increasingly specialized and efficient economies. Part of this shift might relate to the changing character of urbanism and the dynamics of urbanization—for example, the development of large centres inhabited predominantly by non-food producers. Another aspect of this change may relate to the development of new status roles and the need to distinguish them by implementing new sumptuary rules, such as regulating the production and use of pulque and certain types of clothing involved in public ritual performances and displays in which elites played different roles in increasingly stratified societies. Most important of all might have been the changing nature of tribute, exchange and governance, whereby, for example, different kinds of cloth and beverages assumed new functions as material symbols of new socio-economic and socio-political relationships (Hicks, 1987; Murra, 1962). (5) It could be useful, in connection with the transition just noted, to think about the extent to which some techniques and procedures developed for maguey exploitation in north-central Mexico during the Classic period might have been subsequently ‘imported’ from there back into central Mexico. If it was in arid north-central Mexico that maguey was especially critical in the domestic and political economy, then we might expect that it was in the context of expansion into this driest northernmost part of Mesoamerica that the
The role of maguey in the Mesoamerican tierra fría
299
most effective and efficient maguey exploitation first developed. Weintraub (1992), for example, reports the presence of maguey fibre and leaf fragments in flotation samples from late Classic-Epiclassic contexts at the northern centre of La Quemada—perhaps the earliest known examples of such material from agriculturally-based societies in northern Mesoamerica. In addition, some particularly early examples of well-documented spindle whorls derive from late first-millennium and early second-millennium AD contexts in north-central and northwestern Mexico (for example: DiPeso et al., 1974; Ekholm, 1942; Foster, 1978, 1985; Kelly, 1945, 1947, 1949; Meigham, 1976; Charles Trombold, pers. comm.), and in the adjacent Southwest United States (DiPeso, 1951, 1956). In future years, as more archaeological data accumulate, it will be interesting to compare the degree to which productive efficiency and specialization vary over time and space throughout Mesoamerica. We suspect that the cold, dry lands of central and northcentral Mexico will show an unusually high level of such productive efficiency and specialization, because it was in these regions that ancient Mesoamericans were forced to confront the most serious environmental constraints on seed-based agriculture. (6) On the other hand, even now we can sense that it was not environmental problems alone that caused the technological and organizational innovation in the Mesoamerican tierra fría. In north-central Mexico there appears to have been very little change in population density, organizational complexity, maguey-related technology or agricultural technology generally, prior to the development of large states in central Mexico at the end of the Formative period. Some of the changes in the technology of maguey production probably reflect the demands of state administrators for greater productive efficiency and specialization in their domains. Shall we discover notably less technological change or diversity in areas where such state-imposed demands were weak or absent? Does the apparently lower efficiency of Classic period maguey-related technology indicate that Classic states were less demanding on the labour and production of their subjects than those of the subsequent Postclassic? (7) Looking further back in time, it should also be useful to think about the relationships between the competitive arena of chiefly politics (Helms, 1979) and the initial domestication and accompanying botanical diversification of maguey in the tierra fría of central Mexico during the Early and Middle Formative, at a time when tribal ‘big men’ and aspiring chiefs in developing ranked societies throughout southern and central Mexico were seeking higher levels of local productivity to sustain and enhance their prestige. (8) Equally important should be the value to local elites in the tribal and emergentchiefdom societies of north-central Mexico of prestige-enhancing materials such as decorated ceramics, fancy gold and copper metalwork, carved stone, feather headdresses and fine cloth, which were becoming increasingly available from the workshops of skilled, often state-sponsored craftsmen in central and western Mexico from the Early Classic. It is most especially to the varied and changing processes of socio-political interaction between elites in different types of hierarchical societies in central and northcentral Mexico that we should look for new perspectives on the northward expansion of the Mesoamerican frontier into the cold, dry lands of the tierra fría. In sum, the inhabitants of the Mesoamerican tierra fría in central and north-central Mexico were living at the colder, drier edges of a civilization rooted in warmer, wetter,
The archaeology of drylands
300
lower lands to the south, east and west. The constraints of nature made tierra fría populations particularly dependent upon technological and organizational innovation for their survival as Mesoamericans. Maguey cultivation was a key part of this innovation and survival. From the Early Formative onward, the presence of more complex societies along their peripheries provided both material benefits and socio-political problems for tierra fría peoples. These benefits and problems in turn would have provoked technological and organizational innovations—responses that we perceive today as the long-term northward expansion of the Mesoamerican frontier into central and northcentral Mexico.
REFERENCES Anawalt, P. (1980) Costume and control: Aztec sumptuary laws. Archaeology 33:33–43. Anawalt, P. (1990) The emperor’s cloak: Aztec pomp and Toltec circumstances. American Antiquity 55:291–307. Anderson, R.K., Calvo, C., Serrano, G. and Payne, G. (1946) A study of the nutritional status and food habits of Otomi Indians in the Mezquital valley of Mexico. American Journal of Public Health and the Nation’s Health 36:8: 883–903. Barfield, T. (1989) The Perilous Frontier: Nomadic Empires and China. Oxford, Basil Blackwell. Beals, R. (1932) The Comparative Ethnology of Northern Mexico before 1750. Berkeley, University of California Press, Ibero-Americana No. 2. Berdan, F. (1987) The economics of Aztec luxury trade and tribute. In E. Boone (ed.) The Aztec Templo Mayor. 161–84. Washington, D.C., Dumbarton Oaks. Blanton, R., Kowalewski, S., Feinman, G. and Appel, J. (1981) Ancient Mesoamerica. Cambridge, Cambridge University Press. Blasquez, P. and Blasquez, I. (1897) Tratado del Maguey. Puebla, Mexico, Narisco Bassols, second edition. Braniff, B. (1989) Oscilación de la frontera norte Mesoamericana: un ensayo nuevo. Arqueología 1:99–114. Mexico, D.F., Institute Nacional de Antropología e História. Brumfiel, E. (1976) Specialization and Exchange at the Late Postclassic (Aztec) Community of Huexotla, Mexico. University of Michigan, Ann Arbor, unpublished PhD thesis. Brumfiel, E. (1991) Weaving and cooking: women’s production in Aztec Mexico. In J.Gero and M.Conkey (eds) Engendering Archaeology: 224–51. Oxford, Basil Blackwell. Bye, R., Burgess, D. and Trias, A. (1975) Ethnobotany of the Western Tarahumara of Chihuahua, Mexico. I: notes on the genus Agave. Botanical Museum Leaflets, Harvard University 24:5:85–112. Cabrero, M.T. (1989) Civilization en El Norte de Mexico. Mexico, D.F., Universidad Nacional Autónoma de Mexico. Carrasco, P. (1976) La Sociedad Mexica antes de la Conquista. História General de México, 1:165–288. Mexico, D.F., El Colegio de Mexico. Crosby, A. (1972) The Columbian Exchange: Biological and Cultural Consequences of
The role of maguey in the Mesoamerican tierra fría
301
1492. Westport, Connecticut, Greenwood Press. Crosby, A. (1986) Ecological Imperialism: The Biological Expansion of Europe, 900– 1900 A.D . Cambridge, Cambridge University Press. Darling, J.A. (1998) Obsidian Distribution and Exchange in the North-Central Frontier of Mesoamerica, AD 0–1500. Ann Arbor, University of Michigan, unpublished PhD thesis. Ann Arbor, University Microfilms. Davidson, J. and Ortiz de Montellano, B. (1983) The antibacterial properties of an Aztec wound remedy. Journal of Ethnopharmacology 8:149–61. Dibble, C. and Anderson, A. (1963) (translators) Florentine Codex: General History of the Things of New Spain, by Fray Bernardino de Sahagun. Book 11—Earthly Things. Santa Fe, Monographs of the School of American Research and the Museum of New Mexico. DiPeso, C. (1951) The Babocomari Village Site on the Babocomari River, Southeastern Arizona. Dragoon, Arizona, The Amerind Foundation, Inc., Report No. 5. DiPeso, C. (1956) The Upper Pima of San Cayetano del Tumacacorí. Dragoon, Arizona, The Amerind Foundation, Inc., Report No. 7. DiPeso, C., Rinaldo, J. and Fenner, G. (1974) Casas Grandes: A Fallen Trading Center of the Gran Chichimeca. Volume 8: Bone, Economy, Burials. Flagstaff, Northland Press. Duran, Fray Diego (1964) [1581] The Aztecs: The History of the Indians of New Spain. New York, Orion Press, translated by D.Heyden and F.Horcasitas. Ekholm, G. (1942) Excavations at Gusave, Sinaloa, Mexico. Anthropological Papers of the American Museum of Natural History 38:2:23–139. Evans, S. (1990) The productivity of maguey terrace agriculture in central Mexico during the Aztec period. Latin American Antiquity 1:117–32. Fish, S., Fish, P., Miksicek, C. and Madsen, J. (1986) Prehistoric agave cultivation in southern Arizona. Desert Plants 7:2:107–12. Flannery, K. (1968) Archaeological systems theory and early Mesoamerica. In B. Meggers (ed.) Anthropological Archaeology in the Americas: 67–87. Washington, D.C., Anthropological Society of Washington. Flannery, K. (1973) The origins of agriculture. Annual Review of Anthropology 2: 271– 310. Flannery, K. (1986) (ed.) Guila Naquitz: Archaic Foraging and Early Agriculture in Oaxaca, Mexico. New York, Academic Press. Flannery, K., Marcus, J. and Kowalewski, S. (1981) The Preceramic and Formative of the Valley of Oaxaca. In J.Sabloff (ed.) Supplement to the Handbook of Middle American Indians 1:49–83. Austin, University of Texas Press. Foster, M. (1978) Loma San Gabriel: A Prehistoric Culture of Northwest Mexico. Boulder, University of Colorado, unpublished PhD thesis. Ann Arbor, University Microfilms. Foster, M. (1985) The Loma San Gabriel occupation of Zacatecas and Durango, Mexico. In M.Foster and P.Weigand (eds) The Archaeology of West and Northwest Mesoamerica: 327–52. Boulder CO, Westview Press. Gamio, M. (1979) (ed.) [1922] La Población del Valle de Teotihuacán, Tomo II. Mexico, D.F., Institute Nacional Indigenista.
The archaeology of drylands
302
Gentry, H. (1982) Agaves of Continental North America. Tucson, University of Arizona Press. Goncalves de Lima, O. (1978) El Maguey y el Pulque en los Codices Mexicanos. Mexico, D.F., Fondo de Cultura Económica, second edition. Guerrero, R. (1980) El Pulque: Religion, Cultura, Folklore. Pachuca, Mexico, Institute Nacional de Antropología e História, Centre Regional de Hidalgo. Healan, D. (1977) Archaeological implications of daily life in ancient Tollan, Hidalgo, Mexico. World Archaeology 9:140–56. Helms, M. (1979) Ancient Panama: Chiefs in Search of Power. Austin, University of Texas Press. Hester, T. and Heizer, R. (1972) Problems in the functional interpretation of artifacts: scraper planes from Mitla and Yagul, Oaxaca. Berkeley, Contributions of the University of California Archaeological Research Facility 14:107–23. Hicks, F. (1987) First steps toward a market-integrated economy in Aztec Mexico. In H.Classen and P.Van de Velde (eds) Early State Dynamics. Studies in Human Society, Volume 2:91–107. Leiden, Brill. Hough, W. (1908) The pulque of Mexico. Washington, D.C., Proceedings of the U.S. National Museum 33:577–592. Kelley, J.C. (1990) The Retarded Formative of the northwest frontier of Mesoamerica. In M.Carmena Macias (ed.) El Preclásico o Formativo—Avances y Perspectivas: 405–23. Mexico, D.F., Institute Nacional de Antropología e História, Museo Nacional de Antropología. Kelly, I. (1945) The Archaeology of the Autlan-Tuxcacuesco Area of Jalisco, Part I. Berkeley, University of California Press, Ibero-Americana No. 26. Kelly, I. (1947) Excavations at Apatzingan, Michoacan. New York, Viking Fund Publications in Anthropology No. 7. Kelly, I. (1949) The Archaeology of the Autlan-Tuxcacuesco Area of Jalisco, Part II: The Tuxcacuesco-Zapotitlan Zone. Berkeley, University of California Press, IberoAmericana No. 27. Kirch, P. (1984) The Evolution of Polynesian Chiefdoms. Cambridge, Cambridge University Press. Krader, L. (1957) Culture and environment in interior Asia. Studies in Human Ecology: 115–38. Washington, D.C., Pan American Union, Social Science Monographs III. Lattimore, O. (1951) Inner Asian Frontiers of China. New York, American Geographic Society, Research Series No. 21. MacNeish, R. (1991) The Origins of Agriculture and Settled Life. Norman, University of Oklahoma Press. MacNeish, R., Nelken-Turner, A. and Johnson, I. (1967) The Prehistory of the Tehuacan Valley. Volume 2. Non-Ceramic Artifacts. Austin, University of Texas Press. Mangelsdorf, P., MacNeish, R. and Willey, G. (1964) Origins of agriculture in Middle America. In R.Wauchope (ed.) Handbook of Middle American Indians 1:427–45. Austin, University of Texas Press. Mastache, G., Cobean, R., Rees, C. and Jackson, D. (1990) Las Industrias Líticas Coyotlatelco en el Area de Tula. Mexico, D.F., Institute Nacional de Antropología e História.
The role of maguey in the Mesoamerican tierra fría
303
Meigham, C. (1976) The Archaeology of Amapa, Nayarit. Los Angeles, University of California, Institute of Archaeology, Monumenta Archaeologica No. 2. Michelet, D. (1984) Rio Verde, San Luis Potosi (Mexique). Mexico, D.F., Centre d’Etudes Mexicaines et Centroamericaines, Etudes Mesoamericaines 9. Murra, J. (1962) Cloth and its functions in the Inca state. American Anthropologist 64: 710–28. Nelson, B. (1997) Chronology and stratigraphy at La Quemada, Zacatecas, Mexico. Journal of Field Archaeology 24:85–109. Norr, L. (1987) The excavation of a Postclassic house at Tetla. In D.Grove (ed.) Ancient Chalcatzingo: 400–8. Austin, University of Texas Press. Parsons, J.R. and Parsons, M. (1990) Maguey Utilization in Highland Central Mexico: An Archaeological Ethnography. Ann Arbor, University of Michigan Museum of Anthropology, Anthropological Paper No. 82. Parsons, M. (1972) Spindle whorls from the Teotihuacan Valley, Mexico. Miscellaneous Studies in Mexican Prehistory: 45–79. Ann Arbor, University of Michigan Museum of Anthropology, Anthropological Paper No. 45. Parsons, M. (1975) The distribution of Late Postclassic spindle whorls in the Valley of Mexico. American Antiquity 40:207–15. Patrick, L. (1985) Agave and Zea in highland central Mexico: the ecology and history of the Metepantli. In I.Farrington (ed.) Prehistoric Intensive Agriculture in the Tropics: 539–47. Oxford, British Archaeological Reports, International Series 232, Part 2. Puleston, D. (1968) Brosimum Alicastrum as a Subsistence Alternative for the Classic Maya of the Central Southern Lowlands. Philadelphia, University of Pennsylvania, unpublished MA thesis. Puleston, D. (1973) Ancient Maya Settlement Patterns and Environment at Tikal, Guatemala: Implications for Subsistence Models. Philadelphia, University of Pennsylvania, unpublished PhD thesis. Rangel, S. (1987) Etnobotánica de los Agaves del Valle del Mezquital. Mexico, D.F., Universidad Nacional Autonoma de Mexico, unpublished thesis. Rodriguez, F. (1985) Les Chichimeques. Mexico, D.F., Centre d’Etudes Mexicaines et Centraméricaines. Ross, W. (1944) The Present-Day Dietary Habits of the Papago Indians. Tucson, University of Arizona, unpublished MS thesis. Ruvalcaba, J. (1983) El Maguey Manso: História y Presente de Epazoyucan, Hidalgo. Texcoco, Mexico, Universidad Autónoma Chapingo, Depto. de Industrias Agrícolas, Colección Cuadernos Universitarios, Serie Ciencias Sociales No. 4. de Sahagun, B. (1969) História General de las Cosas de Nueva España. Mexico, D.F., Editorial Porrua, edited by A.M.Garibay, three volumes. Sahlins, M. (1968) Tribesmen. Englewood Cliffs, Prentice-Hall. Salinas, J. and Bernard, R. (1983) Etnografía Otomi. Mexico, D.F., Institute Nacional Indigenista. Sanchez, J. (1980) Etnografía de la Sierra Madre Occidental: Tepehuanes y Mexicaneros. Mexico, D.F., Institute Nacional de Antropología e História. Sanders, W.T. (1965) The Cultural Ecology of the Teotihuacan Valley: A Preliminary Report. University Park PA, Pennsylvania State University, Department of
The archaeology of drylands
304
Anthropology. Sanders, W.T. (1966) Life in a Classic village. In Teotihuacan: Onceava Mesa Redonda 1:123–47. Mexico, D.F., Institute Nacional de Antropología e História. Sanders, W.T., Parsons, J.R. and Santley, R. (1979) The Basin of Mexico: Ecological Processes in the Evolution of a Civilization. New York, Academic Press. Sauer, C. (1941) The personality of Mexico. Geographical Review 31:353–64. Sejourne, L. (1983) Arqueología e História del Valle de México, de Xochimilco a Amecameca. Mexico, D.F., Siglo Veintiuno Editores. Serra Puche, M. (1988) Los Recursos Lacustres de la Cuenca de Mexico durante el Formativo. Mexico, D.F., Institute de Investigaciones Antropológicas, Universidad Nacional Autónoma de Mexico. Smith, C. (1967) Plant remains. In D.Beyers (ed.) The Prehistory of the Tehuacan Valley 1:220–55. Austin, University of Texas Press. Smith, C. and Kerr, T. (1968) Pre-Conquest plant fibres from the Tehuacan Valley, Mexico. Economic Botany 22:354–8. Smith, M. and Hirth, K. (1988) The development of pre-hispanic cotton-spinning technology in western Morelos, Mexico. Journal of Field Archaeology 15:349–58. Spence, M. (1971) Some Lithic Assemblages of Western Zacatecas and Durango. Carbondale, Southern Illinois University, University Museum, Mesoamerican Studies No. 8. Taylor, W. (1966) Archaic cultures adjacent to the northeastern frontiers of Mesoamerica. In R.Wauchope (ed.) Handbook of Middle American Indians 4:59–94. Austin, University of Texas Press. Tesch, M. and Abascal, R. (1974) Azadas. Comunicaciones 11:37–40. Puebla, Mexico, Fundación Alemana para la Investigación Científica. Tolstoy, P. (1971) Utilitarian artifacts of central Mexico. In G.Ekholm and I.Bernal (eds) Archaeology of Northern Mesoamerica, Pt. 1, Handbook of Middle American Indians 10:270–96. Austin, University of Texas Press. Trombold, C. (1985) A summary of the archaeology in the La Quemada region. In M.Foster and P.Weigand (eds) The Archaeology of West and Northwest Mesoamerica: 237–67. Boulder CO, Westview Press. Trombold, C. (1989) Comprehensive Summary of the 1986 Excavations of MV-138, a Village Outlier of La Quemada in Zacatecas, Mexico. Mexico, D.F., Informe to the Institute Nacional de Antropología e História. Trombold, C. (1990) A reconsideration of chronology for the La Quemada portion of the northern Mesoamerican frontier. American Antiquity 55:308–24. Vaillant, G. (1931) Excavations at Ticoman. New York, Anthropological Papers of the American Museum of Natural History , Volume 32, Part 2. Weintraub, P. (1992) The Use of Wild and Cultivated Plant Foods at the Site of La Quemada, Zacatecas, Mexico. Buffalo, State University of New York at Buffalo, unpublished MA thesis. Willey, G., Ekholm, G. and Millon, R. (1964) The patterns of farming life and civilization. In R.Wauchope (ed.) Handbook of Middle American Indians 1:446–500. Austin, University of Texas Press.
Part VI EUROPE
17 Traditional irrigation systems in dryland Switzerland ANNE JONES AND DARREN CROOK
INTRODUCTION Most dryland irrigation systems, including most of those documented in this volume, are in less developed world contexts or relate to prehistoric or historic episodes before the development of modern technology. In this chapter we document an instance of an extant traditional dryland irrigation system (termed bisse) in Switzerland—one of the most developed and technologically sophisticated countries in Europe. Although dryland irrigation systems are widespread in the semi-arid regions of the northern and central parts of the Mediterranean basin, these are comparatively little documented (Hunt and Gilbertson, 1998; Jones and Hunt, 1994; Jones et al., 1998; and see Chapter 18). In the Valais canton, Switzerland, the bisse system has a history that spans at least a millennium, and at the beginning of the twenty-first century is a vital component of the advanced Swiss economy. This chapter examines the factors underlying the longevity of this system. It takes a historical perspective and deals mostly with the social and cultural structures that have developed to control access to the water and which, incidentally, account for much of the success of this system. These, typically, are difficult to recover from the archaeological record. A major area of uncertainty with research into abandoned systems is the problem of their environmental relationships. Contemporary field measurements enable assessment of the ways in which these systems interact with landscape processes. These are critical because, in part, they account for the robustness and longevity of some dryland irrigation systems.
THE VALAIS The Valais is a mountainous canton in southwest Switzerland (Fig. 17.1). Altitudes range from 372 m at Lake Geneva to 4,634 m at the summit of Pointe Dufour. Topographically, the Valais can be divided into three regions
The archaeology of drylands
308
Figure 17.1 The Valais canton, Switzerland, showing places mentioned in Chapter 17 —the Rhône valley, the tributary valleys and the mountain zones. The Rhône graben, or trench, divides the two main mountain zones, the Bernese Oberland to the north and the Pennine Alps to the south. Settlement is concentrated in low-lying areas such as the Rhône valley and its major lateral valleys. Today the canton is also divided culturally into two linguistic zones and economically into three areas (the Bas, Central and Haut Valais). The Haut Valais is German speaking, whilst Bas Valais and Central Valais are French speaking. This study focuses particularly on the commune of Vernamiège located in the Central Valais on the southern edge of the Rhône graben (Fig. 17.2). Climate The main controls on mountain climates are altitude, continentality, latitude and topography (Beniston, 1994). The Valais lies within a ring of high alpine mountains and so is partly in rain shadow. The reduced amounts of precipitation received, together with high evapotranspiration as a result of the high summer temperatures and low humidity, mean, therefore, that areas within the Canton can properly be described as semi-arid using the definitions of UNEP (1992) and Reynard (1995). Annual precipitation increases with altitude from around 580 mm per year on the Rhône valley floor to about 2,100 mm in the high alps (Loup, 1965; Reynard, 1995). Aspect also controls humidity through different thermal regimes: south-facing adret slopes receive 50 per cent more sunshine than north-facing ubac slopes (Loup, 1965). Precipitation can vary considerably from year to year by more than 55 per cent of annual average rainfall (Reynard, 1995). Whilst precipitation is fairly constant during the year, it is not unusual for there to be
Traditional irrigation systems in dryland Switzerland
309
extensive dry periods throughout the summer, and indeed in spring and autumn as a result of the foehn winds (Bouët, 1972). High summer evapotranspiration leads to water deficits of as much as 300 mm per month during the growing season (Michelet, 1995; Primault and Catzeflis, 1966), particularly in the Central and Haut Valais. Reynard (1995) suggests that, during the summer months, 2.3–3.0 mm of water a day must be supplied by irrigation for successful agriculture. Agricultural patterns Just less than half the land area of the Valais has agricultural potential (Cosinschi, 1994; Loup, 1965), partly because of the high altitude and steep slopes of this alpine terrain. An altitudinally-sensitive pattern of agricultural land use (Netting, 1972), incorporating a sophisticated traditional irrigation culture—the bisses—has emerged in response to restricted land and water availability. Pasture land is concentrated at high altitude (up to 2,600 m), and most arable activity occurs below 1,500 m. Pasture, vines, orchard crops and some arable lands are irrigated. In common with other alpine areas, the main type of agricultural economy has been based on pastoralism. Before the twentieth century, families and communities were largely self-sufficient, with a range of land types and,
Figure 17.2 Distribution of agricultural land in Vernamiège during the 1960s Source: Modified from Berthoud, 1967: figure 33 therefore, products distributed throughout the commune (Fig. 17.2). Most will have had access to vineyards on the lower slopes (700–900 m), with hayfields, cereal crops and
The archaeology of drylands
310
vegetables being grown around the settlement (900–1,600 m), and pasture and alpine summer grazing above (1,900–2,400 m). Land was held either by families or by the community. The type of ownership determined the resource management practices (Jones, 1991; Netting, 1972; Ostrom, 1990). To be able to offset the risk of a bad harvest, the payment of taxes and tithes and to provide for tools and so on, families and communities would attempt to produce a surplus for sale/exchange or an off-farm income from other activities (Jones, 1991). In a pastoral economy, the quantity and quality of the product—for example, cheese— will depend on the quantity and quality of the grass eaten by the cows. Also, the animals need to be supported throughout the year, and in the Valais, the harvest must support not only the human population during the long, snow-bound winter months but also their livestock. Approximately 10,000–12,000 m3 per hectare of water are required during the growing season for successful hay meadows (Muller, 1946). Michelet (1995) has calculated that, with the high evapotranspiration rates and a low growing season rainfall of 300 mm, there is a water deficit of 7,000–9,000 m3 per hectare. The hay meadows (mayens) are also important in the transhumance process, providing cattle with interim grazing on the way up and down to the high alpine pastures. Whilst there is generally sufficient rainfall for alpine pastures to provide adequate grazing, intensive exploitation does mean that the pastures need to be periodically improved, particularly where they are underlain with impoverished soils. In the Valais, irrigation is utilized to enable the population to maintain a presence above very limited subsistence levels. As was noted above, although the distribution of precipitation is fairly even throughout the year, most of the precipitation in the winter months is as snow and extensive areas of the canton are glaciated at high altitude. This means that there is a source of water that can be used for irrigation—glacial meltwater— but not in the areas where it is required. The bisse or suonen irrigation system was developed as a response to the shortage of water during the growing season and continues to be practised despite technological advances in terms of spray irrigation.
COPING STRATEGIES The bisses are an indigenous response to water shortage in the Valais, similar to slope offtake systems found in other dryland areas (Vincent, 1995). A bisse can be defined as a linear water course constructed and maintained in the Valais canton of Switzerland, with natural and/or artificial or subterranean channels of any dimension, that is or has been used to supply and/or distribute water under gravitational flow, primarily for locally governed and organized irrigation. (Crook, 1997:78) Most bisses divert water from glacial meltwater streams during the high-flow summer months. As such, the construction of the bisses encountered the technical challenges identified by Vincent (1995) for mountain irrigation systems. These include:
Traditional irrigation systems in dryland Switzerland
311
1 the capturing of water and the maintenance of headworks in difficult hydrological environments; 2 the transport of water across rugged, steep, or unstable slopes from higher capture zones to lower use altitudes; 3 a high ratio of canal length to irrigated area; 4 the distribution of water over land of different gradients; 5 the integration of aspects of water tenure with water allocation arrangements; and 6 the availability of technology that can be sustained with available skills and knowledge. Crook and Jones (1999b) set out the design principles of the bisse system, distinguishing between traditional and modern technologies and showing how innovation and adaptation have taken place. The comparatively simple technology enabled a quick response to periodic and haphazard physical disruption. The continuity of the system has been achieved through material, technological and socio-cultural adaptation (Crook, 1997). There is little firm evidence to suggest why or when the system originated. It is considered that the presence of winter cereals at Waldmatte, near Brig, during the La Tène Iron Age indicates a cultural adaptation to the naturally dry environment (Curdey et al., 1993). The original traces of irrigation have been lost or overlaid by later construction, particularly in the fourteenth century. The earliest surviving documentation of a dispute over water rights is dated AD 1008 (Liniger, 1980). Some have argued that the bisses are a response to climate change (Grove and Grove, 1990; Tufnell, 1984), although there is little conclusive evidence to support this theory (Dubuis, 1995). Equally, population growth up to AD 1350 may have necessitated an intensification of agriculture (Crook, 1997). Any extension of agriculture, particularly in the drier areas, would have required the exploitation of new water resources. The earliest mention of the bisses is certainly in some of the driest areas, such as Visp (Viège), Raron (Rarogne) and Sierre (Dubuis, 1995). New economic opportunities may also have provided incentives to intensify agriculture through irrigation. Following the demographic impact of the plague in 1349, there was a reduction in the demand for cereals. This meant that the surviving population could convert land to cattle production and benefit from the emerging markets for Valasian cattle in northern Italy. To do this, however, they required access to the alpine pastures and improved hay production, and bisse irrigation technology provided that opportunity. Inventories (Aufdereggen and Werlen, 1993; Rauchenstein, 1908) and analysis of archival records of first mentions of bisses (Crook, 1997) indicate that there was an expansion in construction during the fourteenth and fifteenth centuries. The opportunities for agricultural extension were also provided by the general retreat of the glaciers between c.AD 1100 and 1400, enabling more land to be brought into production (Aellen, 1988; Harris, 1971, 1972; Pfister, 1994). A second period of expansion occurred during the late nineteenth century, when there was a need to intensify agriculture for economic reasons—the need, for example, to support a growing urban population. At this time, however, the glaciers were at their historical maxima (Aellen, 1988; Chen, 1990). Clearly, the response to opportunity and stress in both the socio-economic and physical environments led to the development of the bisses as part of the coping strategy at a community and population level.
The archaeology of drylands
312
SOCIAL DIMENSIONS OF CONTROL The bisses, like other irrigation systems (Daudry and Daudry, 1995; Vincent, 1995), can be characterized by the arrangements made by communities to control who has access to the water, the amount of water that can be taken by any one individual at a particular time, and the provision of maintenance requirements. Clearly, access to irrigation waters was an important economic determinant for mountain farmers attempting to exploit a niche advantage. In response to the nature of the environment and the task of bringing water from one locality to another over distances as much as 32 km (Crook, 1997), bisse construction required resources far greater than any one individual could provide. This meant that the farmers needed to work co-operatively (Fig. 17.3, Table 17.1). Cooperation needed to continue after the basic system was constructed. To this end, associations of water users were established, variously termed consortages, suonengenosseschaften or geteilschaft. The water rights were held collectively by the consortages; manual resources were supplied by corvée labour (communal labour parties) and materials were provided from local, often communally owned, supplies. Water rights Access to water was in the form of the possession of water rights. In the Valais, water rights are attached to most water sources. Crook (1997) notes how this has led to bisses crossing each other and large torrents by-passing conduits—hydrologically bizarre, but socio-economically rational. The nature of water rights has evolved over time. Originally, water rights were a form of conferred tenure, granted by the King of Bourgogne and delegated to the Bishopric of St-Maurice and Sion (Ammann, 1995).
Traditional irrigation systems in dryland Switzerland
313
Figure 17.3 The Grand Bisse de Lens Key: Sectors from which water was taken by irrigaters from the communes of Icogne (A), Lens (B), Chermignon d’en Haut (C), Chermignon d’en Bas (D) and Diogne—Montana (E) Source: Modified from Crook, 1997
The archaeology of drylands
314
Table 17.1 Approximate numbers of named irrigators using the Grand Bisse de Lens ‘aqueductis communi’ in 1457 Commune No. of irrigators 16 Icogne Lens 25 Chermignon d’en Haut (superieur) 23 Chermignon d’en Bas (inferieur) 12 Diogne-Montana 7 Total 83 Source: Commune Archive of Chermignon and Lens 16 Conferred tenure has been transformed into a type of hydraulic tenure, with the water rights associated with many bisses reflecting the underlying property grid at the time of construction (Walter-Coward, 1979, 1990). Other types of water right are based on the premise of ‘prior use’ (Stelling-Michaud, 1956). Such claims led to disputes, however, which could ultimately lead to the destruction of the bisse by one of the parties (Bérard, 1982). Initially, those who assisted in the construction of the bisse acquired water rights in accordance with the amount of land they owned, though this need not be adjacent to the bisse. This dependence of ascribed water rights on land-holding size vanished with inter-generational transfer and sale. The right to use bisse water (droit d’eau or a pose) operates on two levels: the consortages or commune and the individual. At the level of the consortages, a successful system of water rights maintains exclusivity without sacrificing other characteristics such as duration or permanence, flexibility, the quality or title of security, transferability and divisibility (Scott and Coustalin, 1995). Individual water rights have low exclusivity, however, since independent use is difficult without multilateral control and agreement upstream and down. Any water right is dependent on natural variation in flow and, sometimes, on upstream users. In times of drought or other stresses, the notion of exclusivity is adapted to reasonable shares in proportion to full rights between all bisses down a slope profile (Stelling-Michaud, 1956). The predetermined flow of water in most bisses can be altered according to weather conditions, enabling individual water rights to have flexibility. Water rights were generally attached to the consort rather than to the land, to prevent access to the common resource by outsiders (Jones, 1987). Examples of these restrictive practices are found, for instance, in the consortages of the Grand Bisse de Lens, Bisse de Vercorin and Bisse Dessous (Crook, 1997). Other measures to protect the access to water rights included reclaiming rights from women who married men from outside the commune (Netting, 1972). Water rights are, however, divisible and have been characterized by fragmentary inheritance strategies and family agreements (Weigandt, 1977; Weinberg, 1972). The sharing and renting of water rights, among those eligible, enable greater flexibility in the system. Increased mobility and a decline in dependence on agriculture for survival mean that many owners of water rights now live outside the commune to which they apply, or no
Traditional irrigation systems in dryland Switzerland
315
longer have any need of them. For example, the 1980 register of water rights for the Grand Bisse de Lens indicates that water rights for this bisse are held by individuals living in Geneva, Lausanne, St-Maurice and Zermatt (source: Grand Bisse de Lens consortage archive). This situation is resolved by these individuals being asked to relinquish rights where they have no practical use. Outside agriculture, these rights have no monetary value (Grand Bisse de Salins consortage archive), though sentimental attachment means that not all are willing to do this. Equally, some families have acquired more water rights over the years through inheritance, as a result of which they also hold more voting rights in the General Assembly of the consortages and carry more weight in decision making. Water rights of consorts, the sequence of irrigation turns, and the registration of changes (mutations) to water acquisition and allocation are described and recorded in the ratement. The ratement is a useful documentary tool in plotting the expansion and contraction of the consortages as a result of either demographic change or change in the amount of irrigated land and technical improvements. Whilst an increase in the number of consorts is difficult to detect because often only one family member will be named, changes to the number of time periods or sections of the bisse (poses, tours, tassets) can be more easily determined. The ratements were altered only after significant changes had occurred to the water rights. For example, the bisses of Vernamiège had five ratements between 1912 and 1954 (1912, 1923, 1935, 1946, and 1954: Berthoud, 1967). Not only do water rights identify those who have access to the water; they also record when water may be taken from the bisse, for how long and/or how much may be used. The precise arrangements varied from locality to locality, as did the terminology employed (Crook, 1997). One droit d’eau on the Bisse de Clavoz would irrigate 3,040 m2 and correspond to one third of the flow from the bisse (Ruedin, 1986). With meadow irrigation, traditional water rights related to the volume of water that could be taken. For example, the Bisse Vieux receives a total discharge of 150 litres per second and there are six water rights associated with this bisse, hence each water right is equal to 25 litres per second (source: Bisse Vieux archive). Water rights are usually divided according to the day, hour and rotation (tours/kehrs) (Table 17.2). The right to use water could be at any time of the day or night, according to the regulations established by the consortages. The twenty-four hours could be divided into specific time periods (Bisse Vieux archive; Crook, 1997; Grand Bisse de Lens consortage archive; Ruedin, 1986): day and night— 0400–1800 hours; morning, afternoon and night—0500–1300, 1300–2100 and 2100– 0500 hours; early morning, morning, afternoon and night—0400–0900, 0900–1400, 1400–2000, and 2000–0400 hours; or more finely up to eight three-hourly periods. In the past, indeed up until the 1950s in some areas, the scheduling of irrigation was determined by the position
The archaeology of drylands
316
Table 17.2 Examples of tours, with the number of droits and sequence of irrigation hours a. Bisse Vieux, 1839 Tour 2 Tiers No. of droits Sequence of irrigation hours 1 3 15, 3, 6 2 3 6, 6, 12 3 5 5, 2 1/2, 4 1/2, 5, 7 Total 11 72 Tour 4 Quarte No. of droits Sequence of irrigation hours 1 4 12, 3, 4 1/2, 4 1/2 2 3 8, 10, 6 3 9 2, 2, 11/2, 11/2, 3, 2, 4, 3, 9 4 5 3, 2, 4, 3, 9 Total 21 95 Source: Bisse Vieux ratement, 1839 (Communal Archive of Nendaz P259) b. Bisse Vieux, 1865 Tour 2 No. of people Division Mutations Total hours Droit sequence (hours) 11 Tiers 0 73 9, 16, 6, 6, 3, 9, 5, 3, 4, 6, 6 Tour 12 No. of people Division Mutations Total hours Droit sequence (hours) 21 Quarts 2 98 5, 3, 3, 2, 2, 3, 6, 3, 4, 5, 12, 8, 2, 4, 1 1/2, 10 1/2, 8, 2, 2, 3, 9 Source: Bisse Vieux ratement, 1 865 (Communal Archive of Nendaz P324) c. Bisse Vieux, 1924 Tour 2 Droit No. of people Sequence of irrigation hours 1 5 9, 3, 5 1/2, 1, 1 1/2 2 9 2, 2, 2, 1 1/2, 1 1/2, 1 1/2, 1 1/2, 3, 3 3 5 5, 7, 3, 3, 6 4 4 7, 5, 5, 7 5 2 12, 12 Total 25 120 Source: Bisse Vieux ratement, 1924 (Communal Archive of Nendaz P514) of the sun and shadows at particular locations (Bratt, 1995; Netting, 1981). A tour is the time taken for all the land served by a bisse to be irrigated, which can
Traditional irrigation systems in dryland Switzerland
317
vary with the length and discharge of the bisse and the number of irrigators (Table 17.2). Where the daily clock is divided into larger segments, a single tour will take longer than one where the daily schedule is in smaller parts. Much, however, depends on supply and demand. The distribution of the water would normally take place in rotated sequence down the bisse, with each subsequent section of the bisse receiving water in turn (Figs. 17.4 and 17.5). Sundays and feast days would normally be reserved for irrigating church lands. The sequence would normally be repeated with each rotation, although there could be different rights attached to each section (Grand Bisse de Lens consortage archive; Grand Bisse de Salins consortage archive). Thus farmers closest to the source were not advantaged over the tail-end users. Night irrigation was organized to cover the fields closest to the village, to reduce the risk of injury and to save effort. This practice is still current, for instance in Ausserburg (Crook, 1997).
Figure 17.4 The irrigation sectors in Vernamiège Note: The numbering of sectors is the same as in Figure 17.5 Source: After Berthoud, 1967: Figure 38
The archaeology of drylands
318
Figure 17.5 Distribution of water during the first tour from the bisses of Vernamiège, 5th May–8th June 1964 Note: The numbering of the sectors is as in Figure 17.4 Source: Modified from Berthoud, 1967 For the system of water rights to work effectively it needed to be recognized by all, as well as being registered in the ratement. Originally, recognition was through custom and incorporated the use of ocular tools (Mariétan, 1948): each family would have a distinctive mark, consisting of a series of dots and lines, and for each bisse sector, the
Traditional irrigation systems in dryland Switzerland
319
family marks of all eligible families were inscribed onto all sides of a wooden stick or on to a wooden block (termed tessel, tesseln or wassertessle), together with symbols indicating the number and timing of water rights owned by each family (Briw, 1961; Lautenschlager, 1965; Fig. 17.6). Every morning, during the irrigation period, the guard or erwin of the consortage (responsible for the day-to-day running of the bisse) would hang a tessel from each family chalet with the entitlement of water for that day. The tessel system was still in use in some areas such as Mund, Zeneggen and the Lötschental valley, until the 1920s (Jossen, 1989; Macheral, 1984; Quaglia, 1984). Conflict resolution It is inevitable that conflicts will arise when water, as with any scarce resource, has to be distributed. The details concerning the rules and regulations of the consortages, and of the process of water management and of distribution, were contained in a document known as the réglement. These were first translated from the Latin into the vernacular in the sixteenth century (Bratt, 1995) and in many cases are still operative today, although they have been reviewed and modernized—the réglement for the Grand Bisse de Lens, for example, underwent major revisions to the 1457 original in 1698, 1914 and 1980. The large temporal gaps between new statutes and réglements hint at processes operating outside the rule books: the resultant documents are reflections of the complexity of case law and of the careful preservation of institutional memory (Crook, 1997; Mahdi, 1986). Monetary fines, cautionary tales of ghostly processions, exclusion and the threat of purgatory were means by which individuals were censored for mis-demeanours. Equally, there were inter-communal disputes over rights to water. The threat to water sources for irrigation has resulted in bitter disputes between the controlling bodies (Table 17.3). Those communes with the greatest threat to their water security demonstrate some of the best examples of ongoing disputes. Many disputes arose at the time of construction, or even before, and many operational disputes usually have their origins in these earlier events. Disputes of this nature were referred to the Bishop of Sion until 1627, and thereafter to the Cantonal civil courts, although the local priests still played an important part in arbitration (Communal Archive at Sion; Grand Bisse de Lens consortage archive).
ENVIRONMENTAL IMPACTS The bisse system of the Valais has been in documented operation for almost a millennium, and as such has contributed to the distinctive Valaisian landscape
The archaeology of drylands
320
Figure 17.6 A tessel, used by members of the consortage of the Grand Bisse de Lens at Chermignon d’en Bas from the late nineteenth century until about 1920, when this practice was abandoned in this locality Note: Every morning during the irrigation period, the erwin (the person responsible for allocation of water) would hang the tessel outside the chalet of the family with that day’s entitlement to irrigate. A middle notch indicated morning at one end and afternoon at the other. The tessel could be hung from either end, the uppermost section indicating the entitlement for that day
Traditional irrigation systems in dryland Switzerland
Table 17.3 Examples of bisse disputes Commune Dates Reason Lax v 1347, 1367, Climate change Martisburg1 1443, 1554 leading to a water supply drying up
Settlement Sequential sharing of remaining sources Agreement between communes
321
Marisburg v 1351–1747 Water rights for a Fiesch & new suon Fieschertal2 1811–1961 Dispute over water Document3 rights Sion v Ayent 1484 Claims of illegal use No official and sale of water judgement & Savièse4 1686 Illegal diversion of Construction of Arbaz v water between two partition Grimisuat5 points Vercorin, 1385/1390– Insufficient water in Sharing Rêchy & 1448 dry spells because of arrangement 1/3 Chalias v Vercorin, Rêchy excessive Grône & abstraction for the & Chalais 2/3 Grande Bisse Neuf Grône & Loye Loye6 at Grône 1548–1565 Construction of a Compensation new bisse without payment & dry authorization or weather clause water rights Bagnes v 1443–1465, Opposition to a new Compensation, 7 appointment of 1515, 1545, bisse because of Levron challenge to water guards (largely 1626, rights and damage ineffective) 1629/30, 1839, 1923 claims leading to sporadic vandalism and destruction Conthey C14-C17 Territorial dispute Improvement as (Savoy) v over alp and bisse from 1462 when Savièse source leading to Savoy was murder, beaten at the (Valais)8 assassination, Battle of La sacking/burning of Plante village Ayent v 1950 Dispute over Convention in usufructory and favour of Sion Sion9
Arbitrators Bishop of Sion
Napoleon Bonaparte Bishop of Sion Inhabitants of Lens, Lords of Grange and Bishop of Sion Tribunal
Initially the Abbot of StMaurice
Overseers of Bern and Fribourg
Riedmatten & Zimmermann,
The archaeology of drylands
ascribed water rights and tariffs, refusal to sign the convention
322
solicitors
Commune Dates Reason Settlement Arbitrators Consortage of 1989 The Sarmona section of Pending—the Tribunal with representation by the bisse lost large probable the Grande to outcome will Icogne and Crans Bisse de Lens date quantities of water Development through infiltration. The be a v. multiple solution, a concrete compromise. Society, interest environmental conduit, used already on group10 other sections without pressure groups and complaints, has been the public objected to as being insensitive to the natural environment and ecology Sources: 1, 2 Liniger (1980); 3 Département des Simpelberges (1811); 4, 5, 9 Communal Archive at Sion; 6 Stelling-Michaud (1956); 7 Bérard (1982); 8 Roten-Dumoulin (1990); 10 Grand Bisse de Lens consortage archive of the present day: straight-line water-courses that dissect natural streams and torrents, and grid pattern reticulations in areas where the water is finally distributed onto the fields (Crook and Jones, 1999a). That such a system has not been totally abandoned is due in part to the nature of social systems and technological responses, but it also reflects the comparatively low levels of environmental impact. The range of contemporary water quality in the bisses has no detrimental effect on soil alkalinity, sodicity and salinity— factors that are known to hinder plant growth (Crook, 1997; Jones et al., 1998). The high levels of infiltration during traditional gravity irrigation (ruissellement), together with the leaching caused by rainfall after the irrigation season, help to prevent salt accumulation in soils. Gullying and sheet wash erosion on steep slopes, resulting from the use of gravity distribution techniques, have been negated by concentrating ruissellement on resilient hay and grass meadows. Terracing has also been used to reduce slope angles in pastures, orchards, and arable fields. The glacial water is carried over long distances, enabling it to warm, thus preventing plant damage and making the water safe for cattle to drink (Crook, 1997). The meltwaters can also carry large amounts of sediment, which provide lining to bisse channels when they pass over permeable bedrock. The deposition of sediments onto the fields is thought to have contributed to the maintenance of soil fertility, particularly on the Rhône valley floor and in waters draining from areas of metamorphic bedrock. Over the last thousand years, the bisses have become part of the overall management strategy for the very dynamic Valaisian landscape. The abandonment of bisses has led to landslides as slopes have become saturated with unmanaged water. A general loss of biodiversity also follows abandonment, as patterning imposed by the bisse disappears. For this reason some bisses are now maintained as part of a general landscape management strategy, particularly in tourist areas (Crook and Jones, 1999a).
Traditional irrigation systems in dryland Switzerland
323
DISCUSSION This chapter has documented a traditional dryland irrigation system in one of the most developed countries in the world. It is argued that the bisses have survived major climatic, economic and social change because of the nature of their social matrix, although their lack of adverse environmental impact must also have been significant. Notably, the bisses survived the full rigour of the Little Ice Age, which in Switzerland was characterized by episodes with very cold and sometimes extremely variable weather and repeated crop failures (Crook, 1997; Pfister, 1994). The systems originated in early medieval times and have survived feudalism and its break-up, the Napoleonic invasion, the appearance of industrial society and the modern communications and economic revolutions. It is clear that the bisses have survived these changes by a continual process of adjustment, which has been facilitated by the manner in which the consortages have been prepared to be flexible in their approach. The technology is in many ways comparable with that found in other mountain irrigation systems in less-developed regions today (Vincent, 1995), and in prehistoric contexts (for example: Farrington and Park, 1978), although in recent years modern materials and techniques have been selectively adopted (Crook and Jones, 1999b). The social structures show some similarities with other dryland systems. O’Neill (1987) demonstrated a complex and well-adjusted social matrix to Portuguese irrigation systems. A group approach to water management can also be seen in the Maltese Islands (Jones et al., 1998). These systems are all characterized by a high level of equity, with individual water rights functioning within a corporate setting and with an element of democracy in decision making. Also significant is a collective memory, which provides ‘case law’ and an effective mechanism for conflict resolution. The systems’ physical longevity can be ascribed to the flushing of salts from the fields by the application of water ‘in excess’ of simple irrigation requirements: in the case of the bisses, salts are carried away by leaching and run-off. Such systems, all relatively long-lived, have been able to cope in dynamic physical and socio-economic contexts most of all because of the equitable ways in which their social control structures have been formulated. This observation could possibly be generally applied to comparable systems of irrigation evidenced in the archaeological record.
ACKNOWLEDGEMENTS D.S.Crook acknowledges a University of Huddersfield Research Studentship and a grant from the Dudley Stamp Memorial Fund. The help of numerous Valaisan farmers and officials was invaluable. C.O. Hunt drafted the diagrams and suggested modifications to the text.
The archaeology of drylands
324
REFERENCES Aellen, M. (1988) Fluctuations of Glaciers, 1980–5, Volume 5. Paris, IAHS-UNESCO. Ammann, H.R. (1995) Aperçu sur les documents relatifs aux canaux d’irrigation du Haut Valais à l’époque médiévale. Annales Valasiennes 70:263–80. Aufdereggen, J. and Werlen, C. (1993) Rapport Bisses/Suonens. Sion, Canton du Valais, Service de l’Environement et l’Aménagement du Territoire. Beniston, M. (1994) Climate scenarios for mountain regions: an overview of possible approaches. In M.Beniston (ed.) Mountain Environments in Changing Climates: 136– 53. London, Routledge. Bérard, C. (1982) Bataille pour l’Eau. Sierre, Les Editions Monographic S.A. Berthoud, G. (1967) Changements Économiques et Sociaux de la Montagne: Vernamiège en Valais. Berne, Franke. Bouët, M. (1972) Climat et Météorologie de la Suisse Romande. Lausanne, Payot Edition. Bratt, G. (1995) The Bisses of Valais: Man-Made Watercourses in Switzerland. Gerards Cross, Guy Bratt. Briw, V.A. (1961) Aus Geschichte und Brauchtum der pfarrgemeinde Fiesch. Visp, Buckdruckerie Mengis. Chen, J.Y. (1990) Changes of Alpine Climate and Glacier Water Resources. Zurich, Eidgenössische Technische Hochschule, unpublished PhD thesis. Cosinschi, M. (1994) Le Valais: Cartoscopie d’un Espace Régional. Lausanne, Editions Payot. Crook, D.S. (1997) Sustainable Mountain Irrigation? The Bisses of the Valais, Switzerland, a Holistic Appraisal. Huddersfield, University of Huddersfield, unpublished PhD thesis. Crook, D.S. and Jones, A.M. (1999a) Traditional irrigation and its importance to the tourist landscape of Valais, Switzerland. Landscape Research 24:49–65. Crook, D.S. and Jones, A.M. (1999b) Design principles from traditional mountain irrigation systems (bisses) in the Valais, Switzerland. Mountain Research and Development 19:79–99. Curdey, P., Mottet, M., Nicoud, C., Baudais, D., Lundström-Baudais, K. and Moulin, B. (1993) Brig-Glis/Waldmatte, un habitat alpin de l’âge du Fer: fouilles archéologiques N9 en Valais. Archéologie Suisse 16:138–51. Daudry, D. and Daudry, G.-J. (1995) Le ru de Mazod-Chétoz (Vallée d’Aoste, Italie). Histoire, techniques de construction, importance agricole. Annales Valaisannes 70: 143–62. Département des Simpelberges (1811) Französisches Reich. Simplon, Département de Simpelberges. Dubuis, P. (1995) Exposé introductif: bisse et conjoncture économique: le cas du Valais aux XIVè et XVè siécles. Annales Valaisannes 70:39–46. Farrington, I.S. and Park, C.C. (1978) Hydraulic engineering and irrigation agriculture in the Moche valley, Peru: c.AD 1250–1532. Journal of Archaeological Science 5:255–
Traditional irrigation systems in dryland Switzerland
325
68. Grove, A.T. and Grove, J.M. (1990) Traditional montane irrigation systems in modern Europe: an example from Valais, Switzerland. Agriculture, Ecosystems and Environment 33:181–6. Harris, B. (1971) The Monte Moro pass and the Col d’Hérens. Alpine Journal 76: 127– 32. Harris, B. (1972) Travel and trade in the Pennine Alps. Alpine Journal 77:175–82. Hunt, C.O. and Gilbertson, D.D. (1998) Context and impacts of ancient catchment management in Mediterranean countries: implications for sustainable resource use. In H.Wheater and C.Kirby (eds) Hydrology in a Changing Environment. Volume II: 473– 84. Chichester, John Wiley and Sons. Jones, A.M. (1987) Kin relations in a French alpine community: a preliminary investigation. Sociologica Ruralis 27:304–22. Jones, A.M. (1991) Exploiting a marginal European environment: population control and resource management under the Ancien Régime. Journal of Family History 16: 363–79. Jones, A.M. and Hunt, C.O. (1994) Wells, walls and water supply: aspects of the cultural landscape of Gozo, Maltese Islands. Landscape Issues 11:24–9. Jones, A.M., Hunt, C.O. and Crook, D.S. (1998) Traditional irrigation strategies and their implication for sustainable livelihoods in semi-arid area: examples from Switzerland and the Maltese Islands. In H.Wheater and C.Kirby (eds) Hydrology in a Changing Environment. Volume II: 485–94. Chichester, John Wiley and Sons. Jossen, E. (1989) Mund: Das Safrandorf im Wallis. Naters, Commune de Mund. Lautenschlager, E. (1965) Le systeme d’irrigation d’Ausserberg en Valais. Bulletin Murithienne 88:9–16. Liniger, M. (1980) Bisses et autre raz des Alpes occidentales. Les Alpes 56:42–4. Loup, J. (1965) Pasteurs et Agriculteurs Valaisans: Contribution à l’Étude des Problèmes Montagnards. Grenoble, Imprimerie Allier. Macheral, C. (1984) L’eau du glacier. Études Rurales 93–94:205–38. Mahdi, M. (1986) Private rights and collective water management in a High Atlas tribe. In BOSTID (Board on Science and Technology for International Development) (ed.) Proceedings of the Conference on Common Property Resource Management: 181–98. Washington D.C., National Academy Press. Mariétan, I. (1948) Les Bisses: La Lutte pour l’Eau en Valais. Neuchatel, Editions du Griffon. Michelet, P. (1995) Les techniques d’entretien les bisses. Annales Valaisannes 70:163– 74. Muller, H. (1946) De quelques solutions nouvelles du problème de l’irrigation. Bulletin Murithienne 63:33–40. Netting, R.McC. (1972) Of men and meadows: strategies of alpine land use. Anthropological Quarterly 45:132–44. Netting, R.McC. (1981) Balancing on an Alp: Ecological Change and Continuity in a Swiss Mountain Community. Cambridge, Cambridge University Press. O’Neill, B.J. (1987) Social Inequality in a Portuguese Hamlet. Cambridge, Cambridge University Press. Ostrorn, E. (1990) Governing the Commons: The Evolution of Institutions of Collective
The archaeology of drylands
326
Action. Cambridge, Cambridge University Press. Pfister, C. (1994) Spatial Patterns of Climatic Change in Europe in A.D. 1675–1715. Bern, Gustav Fischer Verlag. Primault, B. and Catzeflis, J. (1966) Du climat valaisan. La Recherche Agronomique en Suisse 5:248–67. Quaglia, C.L. (1984) Le Mont du Lens. Lens, Commune de Lens. Rauchenstein, F. (1908) Les Bisses du Canton Valais. Sion, Departement de l’Intérieur. Reynard, E. (1995) L’irrigation par les bisses en Valais. Approche géographique. Annales Valaisannes 70:47–64. Roten-Dumoulin, R-M. (1990) Savièse: une Commune Rurale dans le Valais du XIXè Siècle. Brig, Rotten-Verlags AG. Ruedin, P. (1986) Le Bisse de Clavoz: au fil de l’eau et des ans. 13 Étoiles Reflets du Valais 5:34–5. Scott, A. and Coustalin, G. (1995) The evolution of water rights. Natural Resources Journal 35:821–979. Stelling-Michaud, S. (1956) Vercorin: Une Commune Valaisanne au Moyen Age. Sion, Commune de Vercorin, Extrait de Variesia. Tufnell, L. (1984) Glacier Hazards: Topics in Applied Geography. Harlow, Longman. UNEP (1992) World Atlas of Desertification. London, Edward Arnold. Vincent, L. (1995) Hill Irrigation: Water and Development in Mountain Agriculture. London, Intermediate Technology Publications Walter-Coward, E. (1979) Principles of social organisation in an indigenous irrigation system. Human Organisation 38:28–36. Walter-Coward, E. (1990) Property rights and network order: the case of irrigation works in the western Himalayas. Human Organisation 49:78–88. Weigandt, E. (1977) Inheritance and demography in the Swiss Alps. Ethnohistory 24: 133–48. Weinberg, D. (1972) Cutting the pie in the Swiss Alps. Anthropological Quarterly 45: 125–31.
18 Desertification, land degradation and land abandonment in the Rhône valley, France SANDER VAN DER LEEUW
INTRODUCTION Since the mid-1980s, the European Union has had a research programme on the causes of desertification and land degradation in southern Europe (Fantechi and Margaris, 1986). It initially focused on climate (e.g. the HAPEX-SAHEL and EFEDA programmes), but from the late 1980s two other elements were introduced: the study of atmospheregeosphere interactions and their effect on land use and living conditions in the drylands of Spain, Italy and Greece (e.g. the MEDALUS and ERMES programmes); and the study of long-term human-environmental interactions (e.g. the ARCHAEOMEDES programme, the focus of this chapter: van der Leeuw, 1998a). This development reflected two consecutive changes in perspective: moving from the idea that people (reactively) adapt to their environment to focusing on their proactive role in modifying their environment (to its detriment), and somewhat later to accentuating their interactive, and mutually dependent, relationship with it. The shift offers an interesting opportunity for archaeology, because studying long-term natural processes without looking at the sociocultural dynamics of human society makes little sense in this context. Archaeology is the only discipline that can do so. However, to meet this challenge, archaeologists have to overcome some important intellectual difficulties; these are either relicts from the history of our discipline (such as our tendency to consider the past for the past’s sake) or are due to the wider context of the western intellectual tradition, such as the nature-culture opposition and differences between naturalists and historians in their approaches to the past (van der Leeuw, 1998b). This philosophy has underpinned the ARCHAEOMEDES Project, which, since 1992, has brought together a team of researchers from up to seven European countries, representing a full range of academic and applied disciplines, with the aim of improving our understanding of desertification, land degradation and land abandonment along the northern Mediterranean rim. In selecting field sites, different time-frames were taken into account, from the later Palaeolithic (Epirus), via prehistoric cultures of the earlier Holocene (lower Rhône valley, Vera basin, Empordà, Isle of Braç), to the Roman and medieval periods (lower Rhône valley) and the present (Argolid, Veneto, LanguedocRoussillon, Midi-Pyrénées, Marina Baixa, Baixo Mondego). These studies have been undertaken on a range of spatial scales, in different climate zones, and focusing on different aspects of human-environmental interaction, degradation and desertification. This chapter summarizes the multidisciplinary research that was undertaken from 1992
The archaeology of drylands
328
to 1999 in southern France by one of the ARCHAEOMEDES teams, consisting of environmental and cultural archaeologists, social and physical geographers, statisticians, remote sensing and information scientists, an economist and an ancient historian. Initially (1992–94), one group focused on the archaeology of the lower and middle Rhône valley, and another on the settlement history of the same area over the last two centuries. The archaeological research attacked the topic in two ways, concentrating on climate change and its impact on degradation in the Valdaine region of the Rhône valley over the last 10,000 years and human-land relationships in the Roman period in the middle and lower valley (Fig. 18.1).
CLIMATE, ENVIRONMENT AND PEOPLE IN THE VALDAINE In the Valdaine, the region around Montélimar in the middle valley, the fact that 40 km of trenches were being dug allowed us to investigate the exposed sections and take numerous micromorphological samples in nested areas with spatial scales of 0.1, 1, 10 and 100 km2. By correlating these columns, we built up a detailed three-dimensional interpretation of the erosion—cumulation dynamics of this landscape throughout the Holocene. Temporal calibration was based on a combination of archaeological and radiocarbon dating of the sections. We were able to distinguish the traces left by different kinds of socio-natural impact on the landscape, such as: • erosive crises regularly rejuvenating the soil (middle neolithic, late neolithic, middle iron age, Roman [third century AD] and modern); • degradation from over-intensive agriculture (early Roman empire [first and second centuries AD] and modern period); • degradation of the drainage of the soil due to rising river/lake levels and water table (late neolithic, chalcolithic, middle iron age, late antique, early medieval); • drying out of the soils contemporaneous with incision of the rivers and a deficit in the annual water balance (early/middle mesolithic, late bronze age, late iron age); • wherever the soil was covered by trees or grasses and shrubs, regeneration of organic and mineral compounds and soil structure at the end of long periods of pedogenesis (early neolithic, late bronze age, high medieval (tenth to twelfth centuries AD).
Desertification, land degradation and land abandonment
329
Figure 18.1 The middle and lower Rhône valley, southern France, showing the progress of Roman colonization It is thus not enough to present the long-term evolution of landscapes under changing climatic conditions as ‘wet’ and ‘dry’, ‘favourable’ or ‘unfavourable’, or simply to speak of phases of degradation. The study of the overall dynamic, moreover, contributes a number of important insights. The first is that the area had already seen major erosion by the seventh millennium BP. This first erosive cycle did not occur in an environment that was homogeneously subject to excess human pressure. Where extensive degraded surfaces occurred, they were due to a combination of naturally unstable or metastable landscapes in the hills and the destabilization of vegetation on the colluvial deposits through human pressure. Second, important demographic increases, and extensive exploitation of the
The archaeology of drylands
330
basin, are not always linked to erosion—human impact and climate are regularly out of step. This alerts us to the non-linearities in the process and to the possibility that we, too, may be confronted with unexpected effects of past human impact. This ‘fragilization’ with ‘delayed response’ brought a gradual change in the long-term dynamics: whereas in the early Holocene erosion occurred only when climatic and anthropogenic dynamics were pushing in the same direction, the slightest climatic or anthropogenic event today can cause erosion. The cumulative effect of ten millennia of socio-natural interaction has been to reduce the resilience of these landscapes and make them dependent on human interference to maintain their present state. The process has been responsible for the fragility of many southern European landscapes that were brought under anthropogenic influence relatively early, and implies that the area would suffer badly if a climatic oscillation, even a minor one, were to occur today. That will need to be taken into account in assessing the effects of potential climate changes predicted by Global Climatic Models. Finally, it explains why every year many more acres of land are lost to agriculture by land abandonment than by degradation or desertification: the countryside cannot sustain the absence of human interference any more than it can sustain excessive exploitation. A comparison of the pedogenesis/erosion curve for the Valdaine with climatic indicators such as oxygen isotopes, alpine glacier movements and subalpine lake hydrology points not only to an overall correlation but equally to the urgent need to base our assessment of the impact of global change on regional research. The complexity of the dynamics governing the European climate makes this all the more important.
SOCIO-NATURAL INTERACTIONS IN THE ROMAN PERIOD The second axis of the archaeological research was, as has been indicated, spatially oriented, with a focus on Roman settlement in the middle and lower Rhône basin. We selected the Roman period for four reasons: it represents, between the Neolithic and the sub-recent period, the principal period of long-term demographic expansion in the area; it encompasses a complete cycle of socio-natural interaction, from colonization to abandonment, including the economic crisis that is regarded as having afflicted much of the empire in the second and third centuries AD; it resembles our own epoch in that it confronts an urban perspective, driven by organizational rationalization, with landscapes that it does not have any experience with; and finally, there is an extraordinary wealth of archaeological and written data available. We took as our basic premise that the foundation of a settlement represents a spatial choice, and thus an assessment of the landscape at the time the settlement was founded. This approach enabled us to work on a sufficiently large sample from the eight initial sample areas to have statistically significant results. First, we carried out a multivariate analysis of site characteristics such as the periods of foundation and of abandonment, size, relative wealth and, if any, the kind of economic activities undertaken. This resulted in a chronology of settlement foundations and abandonments (Fig. 18.2) that allowed us to map the colonization of the lower and middle Rhône valley by the Romans (Fig. 18.1). We then endeavoured to reconstruct the
Desertification, land degradation and land abandonment
331
ancient landscape,
Figure 18.2 Settlement trends in the middle and lower Rhône valley, 50 BC—AD 600 Key: (above) the number of settlements in active use during each period (Per) of 50 years; (below) the number of new settlement foundations (Imp) in each period. Regions: A=Alpilles, B=Beaucairois, C=HautComtat, D=Valdaine, L=Lunellois, T=Tricastin, U=Uzège, V=Vaunage combining variables dependent on relief (altitude, slope, slope orientation and reception of solar energy) with the distance from a site to the road system and/or to open water. Though problematical (Favory and van der Leeuw, 1998; Favory et al., 1994; van der Leeuw et al., in press), this exercise yielded a coherent picture of the spread of settlement within each of the sample areas: from the base of the foothills and the lower slopes, to both the well-drained valley floors and the higher slopes and finally to the valley bottoms
The archaeology of drylands
332
requiring drainage prior to exploitation. The other main environmental component in the decision making was in all likelihood the nature of the soils. We tested the soil classification of the Roman agronomist Columella, which is based on ease of tilling rather than fertility, by comparing it with the results of a semantic analysis, in Pliny the Elder’s De Re Agraria, of all associations between a word for soil, the adjectives accorded to it, and the plants mentioned as favouring this soil. Combining the results with existing soil maps, we roughed out ‘reclassified’ soil maps. Comparison with a rare document—the Roman tax and property map known as the Cadastre B of the city of Orange—enabled us to establish that the relative agricultural suitability thus derived from the agronomists coincided quite well with the relative tax assessments of plots in the Tricastin, the area north of Orange. Moreover, the fact that valleys requiring drainage were among the last zones to be settled concurs with Columella’s comments that such locations were the least favoured. In a final multivariate analysis, the archaeological, landscape and pedological data were combined to give us a sense of the principal socio-natural categories of settlements in our sample. The ‘crisis’ of the second and third centuries AD Having thus detailed the natural conditions of various aspects of the Roman colonization of the valley, we focused in particular on the ‘crisis’ of the second and third centuries AD. In the literature, this crisis is ascribed to a wide range of causes, from saturnism to invasions by Asiatic horsemen, and from bad government to ‘the environment’. We first investigated whether there was any correlation between the numbers of sites abandoned at that period and their environments. As Figure 18.3 shows, however, there clearly is none: the many sites abandoned towards the end of the second century are randomly and equally distributed over different landscapes and soils. Moreover, the slight increase in precipitation at the time is documented only for the alpine climate system that feeds the Rhône, rather than for the Mediterranean system that governs the local precipitation. Some lands along the banks of the Rhône were thus reclaimed by the river, but in most of the sample areas we have no important traces of increased erosion. Finally, the sites abandoned in any area are the smallest ones, which were last established, whereas the overwhelming majority of the early sites seems to have outlived the problems, probably because they were located in the best spots and well-connected to the road network— most of them are situated
Desertification, land degradation and land abandonment
333
Figure 18.3 The persistence of settlements in the middle and lower Rhône valley through different occupation periods (Occ 1–6) of 100 years, in each of 11 categories of environment (Rel 0–10): (above) in absolute numbers and (below) as a proportion of the total sites in that environment. The lower graph makes clear that, whatever the environment, 70–80 per cent of sites do not survive more than two centuries at crossroads. Tentatively, therefore, we explain the ‘crisis’, as far as our region is concerned, in terms of a restructuring of the exploitation system. Was there any loss of resilience in the social components of the co-evolving socio-natural dynamics? After all, from a social perspective, a crisis is a temporary incapacity of a society to process sufficient information to deal with the dynamics that it encounters; it is the ubiquity of things going wrong that is characteristic of such a crisis, as seems in fact evident in the written records of the time. The high quality of data in the Tricastin, which include reconstructions of Roman land divisions from ground observations and aerial photography, as well as the Cadastre B tax
The archaeology of drylands
334
map, enabled us to investigate such issues in more detail. Here, in the first century BC, the emperors instigated a large (10,000 km2) drainage scheme so that land holdings could be allocated to retired army veterans, who as smallholders could both ensure the peace and maintain the drainage system. It is clear from the Cadastre B map, which dates to AD 77, that by that time large parts of the area still remained unoccupied, probably because the principal ‘peace dividend’ of the time was a reduction in the number of legions. As the Roman drainage system was oriented north—south, at an angle of 45 degrees to the natural drainage, the lack of maintenance rendered the huge system dysfunctional, promoting erosion and seriously compromising agriculture. In this area, therefore, economic crisis seems closely connected to earlier imperial megalomania—the emperors went ‘a drain too far’! Finally, using Geographical Information Systems, we tried to make for each period and area a map predicting, on the basis of existing settlements, abandonments and the ‘guesstimated’ relative carrying capacity for each landscape unit, the probability of new settlements in the different landscape units in the next period. The resulting maps (Fig. 18.4) show some interesting patterns, especially towards the end of the Roman period, when there is an increase in new settlement foundations in areas abandoned less than a century before. In terms of understanding settlement shifts of the kind interpreted as evidence of the second-/third-century crisis, we clearly need to investigate the role of a settlement’s location relative to other settlements in maintaining inter-settlement dynamics, the dependence of individual settlements on the others in their neighbourhood, their resilience, and so on. To inform this thinking, we tried to define the resilience of individual towns in the middle and lower Rhône valley in the modern period (between c.1800 and the present) on the basis of census data, to investigate how far the observed loss of resilience of the rural areas has been tied to the dynamics of the urban system. We found, through a series of statistical operations (cf. ARCHAEOMEDES, 1998), that this loss of resilience depends, in part, on the local resources and accessibility of the settlement, but equally, if not more so, on its population profile (age, professonal diversity, and so on) and its relative position in the urban hierarchy. The latter expresses the settlement’s size and also the number of functions it fulfils, which in turn is related to the settlement’s attractiveness for people in the surrounding areas. However, although these factors together define the ‘dynamism’ of the settlement—its capacity to achieve things (and thus to adapt), to attract new inhabitants, and so on—the potential to use them for predicting the viability of individual settlements is limited, for two reasons. First, the position of individual settlements was considered relative to the whole of the settlement system, whereas the interaction between local, neighbourhood and more distant dynamics is an important determinant for a settlement’s chances for survival—a multi-scalar approach is thus required. Second, the ‘recent’ period we used is relatively short when viewed against the slow dynamics of settlement systems, allowing us to monitor only part of a single cycle of such dynamics.
Desertification, land degradation and land abandonment
335
Figure 18.4 GIS maps of the Haut Comtat indicating, for each period, the probable distribution of settlement foundations, settlement abandonments and functioning settlements Note: These maps are probabilistic assessments relating these changes in settlement pattern to the estimated carrying capacity of the different landscape units
The archaeology of drylands
336
THE ANALYSIS OF LONG-TERM TRENDS The second phase of the project (1995–6) developed as a study of settlement systems in southern France over the last 2,000 years, combining elements of the archaeological and geographical approaches used already, despite the difficulties inherent in working with archaeological, historical, cartographic and demographic (census) data simultaneously. In this study, we tried to interpret the social dynamics of rural and urban interactions. The core questions asked were: • once a settlement system is established, what happens to it? • how far does the settlement pattern determine the further development of the landscape? • if it does, is that an incremental process or are there phases of sudden transformation? • what determines spatial choices in a landscape where the main resources have been identified? • how is settlement structure affected by demographic changes? • what are the primary factors determining the success (or failure) of an individual settlement—e.g. the environment? the transportation network? the density of preexisting settlement? After ensuring that the different categories of data could be used as equivalent indicators of the same processes (a major challenge), we defined theoretical ‘basins of attraction’ for the settlements of the Lunellois/Vaunage area in the western lower Rhône (Fig. 18.1) over some twelve centuries at the beginning of our era, based on the settlement hierarchy and on a gravity model of spatial interaction. Testing them against the archaeological data gave a rather good fit. Moreover, the shape of these basins turns out to be related to fossil aspects of the landscape that were not known when we defined them. Then we followed the history of the principal settlements and their attraction basins through time. The principal conclusion was that the present-day structure of southern France originates, as far as its urban component is concerned, in the Roman period, but that the village structure is essentially medieval. The overall spatial configuration and the main anchor points are spatially stable. Neither colonization, wars, political disasters nor epidemics have fundamentally changed the spatial organization of the area, because they operate on different spatio-temporal scales. The road system also remained stable because roads link many settlements, of which some are always sufficiently active to need these roads. At each spatial scale, however, one can observe different, irreversible, structural changes as a result of gradual processes. At the micro-scale, for example, the iron age ‘oppida’ settlements of Ambrussum and Mauressip were replaced in the second or third century by Lunel Viel and Calvisson, but Sommières was stable for 2,000 years, until the last few decades of the twentieth century. At the local scale, the development of tourism in two waves, at the end of the nineteenth century and in the 1950s to 1960s, changed the coastal area without affecting the backbone of the urban system (Montpellier, Béziers and Nîmes), but as soon as those three poles came into direct competition, in the twentieth
Desertification, land degradation and land abandonment
337
century, the equilibrium changed in the favour of Montpellier. At the scale of the region as a whole, the following transformations can be noted: first, in the fifth century AD, towns lost control over their hinterlands, probably due to the fact that an excessive degree of centralization of power in the cities is not sustainable; the resulting fragmentation continued until the twelfth and thirteenth centuries, when a new and much less hierarchical urban system took over, with multiple links between all three levels; the structure is presently undergoing another fundamental transformation under the impact of the formation of mega-cities. Over the centuries, the dynamism of the overall settlement system has evolved in tandem with the intensity of an individual settlement’s contact with its hinterland and in the extent to which the size-hierarchy has been stretched, reflecting the effect of oscillations in the power-law structure of the settlement system. There are related methodological lessons from this investigation. Given that the hinterland of individual settlements plays an important role in the dynamics of the system, for a better understanding of a settlement system it is essential to take, in terms of all dimensions (time, size and space), the whole range of scales into account and not limit oneself to the top and middle tiers of the settlement hierarchy. The model of settlement as a dynamic system, in which upper-level structuring is the result of lower-level interactions, requires that we completely change our approach and analyze the system not only ‘top-down’ but equally ‘bottom-up’. It is obvious that we have to take the rural environment into account from the perspective of both population system and resources. The study of the changes occurring in the links between these settlements is still to be undertaken; and they may well be more frequent than changes in the settlement structure itself. In the most recent (1996–9) phase of the project we have attempted to take these lessons into account, linking three levels of investigation of modern-day urban—rural dynamics in southern France (Fig. 18.5) covering the study area as a whole, that is, the regions of Midi-Pyrénées and Languedoc-Roussillon, as well as adjacent parts of Provence-Alpes-Côte d’Azur; three areas composed of one or two departments— Aveyron-Lozère, Hérault-Gard and Comtat; and a micro-region, the Causse Méjan, where we looked at the population and settlement dynamics of all individual communes, including individual households, over about the last fifty years, in combination with their use of rural resources. In this research, we have chosen an approach based on the following working hypotheses: • the settlement structure reflects information processing rather than energy processing (contrary to the traditional tenets of urban studies and archaeology); • the information flows go up and down the hierarchy, so we must approach our analysis from both the top-down and the bottom-up
The archaeology of drylands
338
Figure 18.5 The three levels of the investigation into modern-day urban-rural dynamics in southern France perspectives (the former is common in geography; we have here focused on the latter); • innovation drives the system (Guérin-Pace, 1993); • scaling (the rank-size distribution) should conform to the intensity of information flows. This led us to choose the following proxy measures for aspects of these information flows: demography as proxy for the total information-processing capacity of a settlement; socio-professional diversity as the proxy for the diversity of information-processing capacity (and, hence, for the range of channels linking settlements); and age profiles as the proxy for rate of information processing. Over and beyond these, of course, we took spatial variables into account (location, environment, accessibility) as well as resources (land use, hydrology) and forms of resource exploitation (organization and structure of farms). Much remains to be done, and the following conclusions are both partial and
Desertification, land degradation and land abandonment
339
preliminary, but they show the interest of looking at settlement structure in this manner, first, it proved necessary (and possible) to differentiate the roles of individual settlements on the basis of spatial context (Fig. 18.6) and to view the settlement system as a nested set of interaction zones, in which the dynamic effects of equivalent settlements are different according to spatial level and scale (Fig. 18.7). Taking this perspective enabled us to understand some of the variables and their thresholds and interactions. To understand
Figure 18.6 Schematic representation of the way we have constituted the relations between cities (i.e. urbanized agglomerations of communes) and individual communes and their contexts Key: (above) scales and levels of analysis; (below) spatial entities and different scales of neighbourhood
The archaeology of drylands
340
the dynamics we must determine the quantitative effect of the mix of dimensions of interaction: for example, home-to-work commuter patterns follow a different dynamic from shopping or schooling interactions, and tourism, retirement and active populations also operate differentially. On the other hand, we must combine the demographic dynamics with the institutional and the agricultural dynamics: administrative and civil service jobs, for example, have other dynamic prospects than the private sector, industry is in this respect different from services, and so on. ‘Heritage’ effects (differences in flexibility between matter, energy and information flows) are much more important
Figure 18.7 Schematic representation of the differences in context occurring among towns of similar and/or different sizes in the Haut Comtat
Desertification, land degradation and land abandonment
341
than first assumed: a detailed comparison of the Comtat in the 1870s and 1970s, for example, shows how in the first of these two decades the inherited spatial infrastructure (reflecting the spatial structure of matter) helped the area to deal with rapid changes, whereas in the 1970s the same inherited spatial infrastructure hampered change.
CONCLUSION This chapter has surveyed the methodological development of the ARCHAEOMEDES project and summarized some of its findings regarding the co-evolution of social and natural dynamics in the middle and lower Rhône valley over the past 2,000 years and how their interactions have created the present-day landscape. However, its over-arching purpose is to argue that archaeologists and colleagues in cognate disciplines have to try to deal together with the very long term, including the present. The fact that such a selfevident approach is not more widespread seems at least in part due to the fact that the study of the long-term evolution of socio-natural systems is at the crossroads of two of the most profound disciplinary oppositions that exist in our western intellectual tradition, i.e. between nature and culture (Table 18.1), and between ways to view the past and ways to view the present (Table 18.2). Another, but more common, opposition is the inevitable one between narrow-focus analysis and broad-focus integrative research (Table 18.3). These oppositions have dogged many attempts at cross-disciplinary interaction, in part because of the structure of the academic world: after all, disciplines are by definition selfimposed constraints on the kinds of
Table 18.1 Evolution of the ‘nature—culture’ debate over the last thirty years Pre-1980s 1980s 1990s Nature is cultural The relationship is dualistic Culture is natural Humans are reactive to Humans are proactive in Humans are interactive with the the environment the environment environment Environment is Humans are dangerous Neither are dangerous if handled dangerous to humans for the environment carefully, both if that is not the case Environmental crises Environmental crises are Environmental crises are caused hit humans caused by humans by socio-natural interaction Adaptation Sustainability Resilience Apply technofixes No new technology Minimalist, balanced use of technology ‘Milieu perspective ‘Environnement’ Attempts to balance both dominates perspective dominates perspectives
The archaeology of drylands
342
Table 18.2 The different approaches of the historical and natural sciences to the reconstruction of the past Historical approach Evolutionary approach Interest in past Interest in present Understanding of the present based on Understanding of the past based on the the past present Time and process irreversible Time and process reversible, cyclical or reproducible Accentuates differences Accentuates similarities Stress on case studies Focus on generalizations No coherence between events Coherence between events Focus on inter-scale interaction Focus on intra-scale interaction Table 18.3 The opposition between analytical and integrative approaches in research Attribute Analytical approach Integrative approach • broad and exploratory Philosophy • narrow and targeted • disproof by experiment • multiple lines of converging evidence • parsimony the rule • requisite simplicity the goal Perceived • biotic interactions • biophysical interactions organization • fixed environment • self-organization • single scale • multiple scales with cross scale interactions Causation • single and separable • multiple and only partially separable Hypotheses • single hypotheses and nulls • multiple, competing hypotheses • rejection of false hypotheses • separation among competing hypotheses Uncertainty • eliminate uncertainty • incorporate uncertainty Statistics • standard statistics • non-standard statistics • experimental • concern with Type I error • concern with Type II error Evaluation • to reach ultimate unanimous • to reach a partial consensus goal agreement • exactly right question but The danger • exactly right answer for the useless answer wrong question Source: After Holling, 1998 issues and questions a community of scholars is interested in, maintained by a number of social and educational techniques. Few researchers are comfortable with the recognition
Desertification, land degradation and land abandonment
343
that what they discuss are subjective opinions concerning objective results obtained as answers to equally subjective questions. If we are to progress towards a more holistic perspective on socio-natural interactions over the long term, it is essential that the different disciplines together define the questions that they will address, and that these are not defined primarily in ways that suit one particular discipline and not another. In the case of the ARCHAEOMEDES project, the fieldwork very often enabled the researchers involved to build relationships that could withstand inter-disciplinary debate. A contributing factor was that we began the discussions around a subject— desertification—that was not familiar to most of the people involved (archaeologists) and in which they had no professional (career) stake. However that may be, by opening up our disciplinary kitchens to each other, we have drastically changed our perspectives in three ways. First, the team has evolved its perspective on the role of people in their environment from seeing it as reactive, via proactive, to interactive. Second, the focus has changed through the project from desertification to degradation to abandonment, as interest shifted increasingly (and necessarily) to the social dynamics. Finally, our central concepts used to define socio-natural co-evolution have evolved from adaptation to sustainability to resilience. This was not achieved overnight—trans-disciplinarity is shedding blood, sweat and tears together!—but the results of the project convince us that the exercise has been well worthwhile.
ACKNOWLEDGEMENTS This chapter summarizes the work of many people in the ARCHAEOMEDES team, in particular J.-F.Berger, J.-L.Fiches, M.Gazenbeek, J.J.Girardot, H. Mathian, D.Pumain, L.Sanders, Ph.Cour, F.-P.Tourneux, Ph. Verhagen, I.Tounsi, C.Jung, Th.Odiot, G.Chouquer, C.Raynaud, S.Thiébault and F.Magnin. The project is co-ordinated by F.Favory and S.van der Leeuw.
REFERENCES ARCHAEOMEDES (1998) Des Oppida aux Métropoles. Paris, Anthropos-Economica. Fantechi, R. and Margaris, N.S. (1986) (eds) Desertification in Europe. Proceedings of the Information Symposium in the EEC Programme on Climatology Held in Mytilene, Greece, 15–18 April 1984. Dordrecht, D.Reidel Publishing Company. Favory, F. and van der Leeuw, S.E. (1998) ARCHAEOMEDES. La dynamique spatiotemporelle de l’habitat antique dans la vallée du Rhône: bilan et perspectives. Revue Archéologique du Narbonnaise 31:257–98. Favory, F., Girardot, J.-J., van der Leeuw, S.E., Tourneux, F.-P. and Verhagen, Ph. (1994) L’habitat rural remain en basse vallée du Rhône—de l’utilisation de la téleedétection et des S.I.G. en archéologie. Nouvelles de l’Archéologie 57:46–9. Guérin-Pace, F. (1993) Deux Siècles de Croissance Urbaine. La Population des Villes Françaises de 1831 à 1990. Paris, Anthropos-Economica. Holling, C.S. (1998) Two cultures of ecology. Conservation Ecology 2(2):4 [online]
The archaeology of drylands
344
van der Leeuw, S.E. (1998a) The ARCHAEOMEDES Project—Understanding the Natural and Anthropogenic Causes of Land Degradation and Desertification in the Mediterranean Basin. Luxemburg, Publications Office of the European Union. van der Leeuw, S.E. (1998b) La nature, serait-elle d’origine culturelle? Histoire, archéologie, sciences naturelles et environnement. In A.Ducros, J.Ducros and F. Joulian (eds) La Culture, Est-Elle Naturelle?: 83–98. Paris, Editions Errance. van der Leeuw, S.E., Favory, F. and Fiches, J.-L. (in press) Archéologie et Systèmes Socio-Environnementaux. Etudes Multiscalaires sur la Vallée du Rhône dans le Programme ARCHAEOMEDES. Valbonne, CRA-CNRS.
Index Entries in bold denote figures. Abu Hureyra, 70 Achaemenid empire. See Turkmenistan ad-Diyatheh, 88–92, 90, 95, 99, 100; field systems, 90; settlement, 89, 90; topography, 99 Adi Ainawalid, 181–90, 192 –4 Adi Akel, 192, 193 Adi Gudem, 181 Aegean Sea, 75 aerial photography, 52, 108, 143, 168, 212, 223, 248, 348 Afghan Mountains, 114 Africa, 7, 10, 28, 29, 113, 211; African drylands: definition of (water-balance methods), 19–20: intensive farming in, 252–64: Mama Issara case study, 256–9, 261, 263, 264: Marakwet case study, 210, 212, 254, 255, 256, 259–61, 263, 264: temperature index, 19, 23: topography, 19, 26, 34, 52; eastern and southern Africa, 254, 261; southern Africa, savanna zone, 27, 37, 237, 252–3; Southwest Africa, 34; Sub-Saharan Africa, 9, 37, 171, 233; Tripolitania, 137–56. See also Fezzan agrarian development, models of, 7 –8 agricultural machinery, 115 aid agencies, 194 Al-Biruni, 12 Algeria, 21, 23, 131 al-Harra (Burnt Land), 87 al-Hayat. See Wadi el-Agial alluvium, 128, 130, 131, 133, 151; alluvial fans, 99, 149, 274; alluvial plains, 88, 92; alluvial terraces, 55, 133. See also floods; geomorphology; sedimentation; winds al-Namara, 89, 92–6, 98, 99, 100
Index
347
Altyn Depe, 108, 109 Ambrussum, 350 American Civil War, 114 American Southwest. See North America Ammon, 76 amphorae, 167 Amu Darya River, 106, 115 Anasazi people, 285 Anatolia, 49 Anau, 105, 108, 109 Andean civilisation, 303 Andes, 25, 33, 258, 292, 294 animal bones, 107, 146, 167, 229, 239, 244, 247 animals, 6, 149, 155, 170, 286; butchery techniques, 167, 241; fodder, 14, 98, 183, 184, 187, 190, 194; husbandry, 56, 148, 238, 244; population numbers, 14, 147, 183, 233, 236, 246–7; stall-feeding, 148, 155–6, 207–8, 211, 227. See also goats; grazing; livestock; over-grazing; oxen; pastoralism; sheep; wild animals anthropogenic deposits, 150, 282 anticyclones, 28; North Atlantic, 28; South Pacific, 28; subtropical, 24, 25, 27–9. See also climate; winds Antiochia. See Gyaur Kala aqueducts, 65, 80, 130, 131 aquifers, 164, 165, 168, 172, 175, 218 Arabia, 99 Arabs: armies, 126, 156; conquests, 46, 92, 133, 142; government, 97; historians, 104, 113. See also ULVS Aral Sea, 106 Aravali, 112 archaeobotany, 109–12; evidence, 171, 244.
Index See also plants archaeology of drylands, themes, 3 – ARCHAEOMEDES Project, 341 archaeozoology, 145 architecture, 48, 53, 70, 73, 97, 101, 183, 232 Argolid, 341 arid zones, 3, 13, 18, 19, 56, 128–30, 255, 286, 292; aridification, 7, 11, 13, 72, 81, 125, 147, 169; aridity, 20, 23, 25, 34, 55, 70, 143, 169, 236: atmospheric processes causing, 24–7. See also dessication; drylands Arizona, 276, 279 ar-Risha al-Fawq, 98 ar-Risha al-That, 98 Asia, 8, 30; southeastern, 262–3; Southwest, 26, 107 Atlantic Ocean, 35, 292 atmospheric processes, 6, 22–33. See also climate; winds Atrek River, 106 Aubreville, A., 10 Aurès Mountains, 129 Ausserburg, 331 Australia, 11, 20, 26, 36 Avdat, 48, 51, 53, 56 Aveyron-Lozère, 34 Azania, 257 Aztec civilization, 292, 301 Baghdad, 58 Baixo Mondego, 341 Bamangwato people, 238 bandits, 99 Bantu farmers, 14, 219 Baringo, 213, 265, 267 barley. See cereals barrages, 91, 96 Barth, H., 138 basalt, 87, 92, 96, 98 basketry, 283 bedouin, 46, 58, 68–9, 82, 87, 92, 100; settlements, 81, 98. See also pastoralists Beer-Sheba, 54, 55 Beidha, 70
348
Index Ben-Gurion, D., 58 Beni Abes, 21 Beni Ulid, 138, 142, 147, 150 Berbers, 134 Bernese Oberland, 322 beverages, 212, 295, 296, 310 Béziers, 350 Bible: archaeologists, 75; New Testament, 15; Old Testament, 15, 75; stories, 15 biogeography, 138 Biot, Y., 249 Bisse de Vercorin, 328 Bisse Dessous, 328 Bisse Vieux, 329 bisses, 9, 321–37; abandonment of, 332; definition of, 319–20; environmental impact of, 328–32; maintenance of, 321; technology of, 320–1, 332–3; tessel system, 328, 329 Black Desert. See Syria Blanton, R., 294 Bochum Mining Museum, 66, 68, 73, 80 Bolshevik Revolution, 115 Boserup, E., 266 Botswana, 235–49, 255; agropastoralism, Motloutse River, 234–8; climate, 233, 236–7, 245; Khami settlement, 234, 236, 237, 238–44; population density, 235–6; precipitation levels, 235–6, 245 Bourgogne, King of, 326 Brazil, 36 bridges, 130, 134 Brig, 325 British-Russian-Turkmenian excavations , 107 Bryan, R.K., 147 buildings: huts, wattle and daub, 241; mudbrick, 166; public, 172; stone, 99 burials, 58, 109, 172; burial cairns, 98, 167, 172;
349
Index
350
burial inscriptions, 97 Burkino Faso, Yatenga plateau, 11 burning, 148, 180, 212, 244. See also fire Butua, 244 –39 Byzantine empire, 54, 66; collapse of, 46, 52, 56; late Byzantine Period, 86; society, 48; stability of, 50 Calvisson, 350 camels, 172, 184 canal systems, 6, 96, 108–9, 173, 214–5, 216, 217, 219, 264, 277, 285; underground, 90, 168. See also bisses canals: bronze age, 113; construction, 205–7, 209–10, 212–13, 216; iron age, 111; Roman, 12, 90–2 Cape Town, 4 caravans, 99, 214 Caribbean, 292 carrying capacity. See land Carthage, 131 Casas Grandes, 277, 278, 281, 285 cash-crop farming, 10, 57, 184, 266 Caspian Sea, 106 catastrophic events, 4, 21, 113, 233. See also climate; earthquakes cats, 190 cattle, 57, 90, 184, 203, 207, 218, 219, 248, 277, 294; cattle rinderpest, 10, 194. See also grazing; livestock; over-grazing Causse Méjan, 351 cave deposits, 56, 148 cemeteries, 83, 168, 173 Central America, 2 Central Asia, 104 –20 centralized authority, 10, 119, 262–5, 350 ceramics, 55, 106, 109, 153, 172 See also pottery cereals, 142, 148, 157, 172, 324:
Index
351
barley, 51, 56, 70, 98, 107, 109, 111, 112, 134, 144, 150, 163, 172, 184, 187; emmer, 107; maize, 167, 183, 187, 209, 235, 273, 276, 290–1, 296, 297; millet, 111, 112, 187, 209, 217, 225, 228, 236; sorghum, 163, 167, 184, 187, 191, 209, 213, 217, 225, 228, 236, 244; wheat, 51, 56, 70, 107, 109, 112, 144, 163, 184, 187, 188, 191–2; wild, 69; cultivation, 70, 79, 82, 91–2, 98, 99–100, 107, 147, 150, 153, 155, 167, 187–8, 244; plains, 130. See also crops ceremonial presentations, 303 Chaco Canyon, New Mexico, 276, 277, 283 Chad, 162 chalcolithic: samples, 109; societies, 65 channels, water, 78–80, 81, 262. See also canals; drainage ‘chaos theories’. See environment and human activity charcoal, 80, 130, 148, 225 Charney effect, 34, 144. See also climate cheese, 324 Chesoi, 264 Chihuahua, 277, 278, 279 –80 ‘chott’. See salt Christianity, 48, 57 churches, 49, 53–4, 190 cisterns, 73, 75, 77, 143, 154 clay, 130, 183, 225, 231 climate: catastrophic, 13; climatic change, 6, 67, 75, 81–2; climatic optimum period, 127; Global Climatic Models, 38, 339; global warming, 4, 11. See also anticyclones; Charney effect; El Niño; evapotranspiration; humidity; hyperarid zones; Inter Tropical Convergence Zones; Little Ice Age; monsoon rains; paleoclimatology; precipitation;
Index
352
radiation; winds; Younger Dryas coastal cities, 142, 143, 145, 163 coffee, 214 colonialism, 14, 228, 265, 294 Colorado Plateau, 275, 283 Colorado River, 275 Columella, 346 communications, 265, 266 competition, 218, 219 Comtat, 351, 355 conflict, 3, 10, 11, 195, 233, 299, 345 conquest. See military invasion conservation projects, 235 Constantinois, 133 Continental Intercalaire aquifer system, 165 co-operation, 78, 232, 259, 348 copper, 9, 66, 73, 77, 80 corrals, 99. See also livestock enclosures corruption, 115 cotton, 104, 113, 115, 300; spinning, 238, 243; textiles, 112 Crater Highlands, 203, 218 Cremaschi, M., 170 crisis: economic, 75, 343; erosional, 125, 337, 341 crocodiles, 6, 165 crops, 212, 223; desert, 171; diversity, 112, 115, 180, 181–3, 184, 211, 252, 255; double-cropping, 187, 192; failures, 144, 184, 193, 276, 295, 332; inter-cropping, 184, 187, 228, 297; New World, 167, 184; over-cropping, 144; rotation, 147, 187, 188, 213; yields, 11, 12, 82, 184, 192, 213. See also cereals cultivation ridges, 222, 225–7, 229, 231 –2 cultural changes, 12, 15, 67 currents: Peru current, 26; Benguela current, 26, 37; Humboldt current, 37
Index
353
cyclones, tropical, 20. See also winds Dagh. See Kopet Dag Damascus, 23 dams, 52, 55, 77, 96, 98, 112, 120, 129, 130, 229; checkdams (trincheras), 72, 74, 274–6; Roman, 12; terrace dams, 46, 50 Dana Nature Reserve, 64 Dana tributary, 64 Daniels, C., 162, 168, 172 Dashouz, 105 date palms, 56, 138, 145, 151, 164, 172, 173 Dead Sea, 56, 81 Death Valley, California, 33 deforestation, 3, 11, 134, 144, 217, 283 degradation, 119, 143, 341, 344; long-term, 11, 13 Denevan W. and Turner, B., 225, 226, 286 deposition, 108, 109, 151 desertification, 5, 14, 23, 143, 156, 166, 169, 341, 344, 358; definition of, 11–12, 34–5; area of, 12–14, 36; humanly-induced, 11, 12–14, 36, 79, 154, 162; rate of, 11; recent, 12 deserts, 20, 33, 275–6, 284; agriculture, decline of, 45–59; Atacama, 20, 26; Chihuahuan, 271, 273, 277; Kara Kum, 104, 106, 107, 113; Libyan, 162–5; margins, 8, 36, 38; Nafudh, 86; Namib, 26, 34; Near Eastern, 57–8; Negev, 13, 91; Nubian, 36; Rajputana, 37; Sonoran, 271, 277, 279, 281; Syrian (Badiyat al-Sham), 86; Thar, 21. See also Fezzan; Sahara; Saharan Sahel, Syrian Black Desert dessication, 6–7, 35, 143, 163, 165, 166. See also aridification development aid, 11;
Index
354
development projects, 11, 15, 233; food-for-work programmes, 183 di Lernia, S., 169 disease, 11; animal, 10, 194, 217; human, 10, 114, 143, 154, 192, 194 ditches, 109, 225–6, 229. See also drainage Djeitun. See Jeitun dogs, 244, 276 Dolores, Colorado, 284 domestic animals, 107, 146, 155, 165, 170, 244, 277 drainage, 236, 257, 342, 346; drainage channels, 50, 131, 223, 224–6, 230–1, 343. See also canals; irrigation droughts, 5, 7, 9, 10, 11, 12, 13, 20, 28, 34–6, 45, 142, 144, 150, 155, 184, 187, 195, 133, 235, 248, 276; definition of, 23, 192 dry farming (without irrigation), 73–5, 87 drylands: archaeological evidence, 12, 14; archaeology, difficulties of, 15; biophysical systems, 143, 144; climatology of, 6, 19–38, 233; coastal, 24; definition of, 19–21; diversity of, 8; farming, 7–8, 233–4; fragility of, 14, 339; geographical extent of, 3–4, 19, 20, 36; human perceptions and decisions, impact of, 4, 9–10, 16, 244–7; resilience of, 4; risks and opportunities, 5; sustainability of, 3, 5; temperate 21, 22; themes, 3–16. See also arid zones; environment and human activity dung, 69, 147, 150, 184, 187, 190, 192, 229, 238, 240, 244, 247, 248. See also fertilizer; manure dust storms, 10, 23, 35, 144 Duveyrier, H., 164 earthquakes, 46, 53, 218 East Africa, 56, 203, 206, 211, 232; plateau, 33–4;
Index
355
Rift Valley, 201, 202, 203, 204, 209, 210, 215, 216, 217, 261 East Pokot people, 265 ecology, 3, 13 economic conditions, 8, 75, 100–, 143, 156, 238, 255, 285, 348 EDMA, 68, 78 Edom, 76 Edomites, 65; settlements, 65, 76 EFEDA, 341 Egypt, 21, 35, 66, 134, 172; Egyptian relief art, 171; Egyptian scribes, 75 El Golea (Algeria), 21 El Niño Southerly Oscillation (ENSO), 35 –6 elites, 73, 142, 156, 172, 239, 255, 266, 292, 299, 310–1. See also hierarchies Elusa, 48, 54, 56 emmer. See cereals Empordà, 341 Enderta administrative region, 181 Energy Dispersive X-Ray Microanalysis. See EDMA Engaruka River, 203, 206, 214, 216, 218 Engaruka, 203–19, 256, 258; climate, 201, 215; field systems, abandonment of, 210, 211, 215–18; irrigation systems, 205–13; population, 209, 213; precipitation, 211, 214, 215; topography, 201–4 ENSO. See El Niño Entalo-Wajeret, 181 environment and human activity, non-linear relationships between, 13 environmental determinism, 58, 255 Epirus, 341 Erk Kala, 112, 113 ERMES Project, 341 erosion, 46, 58, 80, 82, 92, 108, 109, 126, 150–1, 154, 212, 216, 226, 249, 346, 348; aeolian deflation, 125, 149; fluvial, 153; gullies, 238; pluvial, 134; sheet wash, 127, 134, 237, 241, 332; wind, 24, 134, 187. See also geomorphology; soil Ethiopia, 181–95, 229, 257; agriculture, 180, 186–94;
Index
356
climate, 180, 192, 194; land-holding, 181–2; Northern Highlands, 10; population, 193–4; precipitation, 34, 186–7; ‘villagization’, 181 ethnicity, 14, 218, 219, 267 ethnoarchaeological research, 68, 181 ethnobotanical research, 181 ethnography, 112, 282, 296 ethnohistorical record, 76 Euphrates River, 87 Europe, 6, 135, 341; climate, 339, 347 European Union, 341 evapotranspiration, 18, 21, 22, 36, 144, 203, 321, 324. See also climate exchange, 56–7, 195, 258, 260–, 267, 303, 309, 310 exports, 8, 213 falaj (aflaj), 168 fallow, 189, 216 famine, 10, 11, 194, 212 farm reconstruction, 51 farmers, 4, 55, 68, 78–80, 83, 143, 154, 156, 204–7; bronze age, 73–4; Nabatean, 78; neolithic, 171; Roman, 79, 171; Romano-Libyan, 13, 78; resilience of, 8; upland, 321. See also indigenous peoples farming, 265; arable, 143; intensive, 13, 252–64, 337; irrigation-based, 8, 259; labour-intensive, 231, 264;9; livestock, 143; mixed, 156, 227–8, 231; modern intensive mixed, 7, 70, 75, 141, 145; sedentary, 67, 295; subsistence, 9, 57, 100, 103, 107, 152, 211, 233–4, 264; surpluses, 319; technology, 145, 165; wheat, 10. See also animals; cash-crop farming;
Index
357
floodwater farming; irrigation; plants farms, 77, 142–3, 144, 147, 156, 181, 211; collective, 194; fortified, 4, 141 (see also gsur); Opus Africanum style, 139, 140; satellite, 53, 78. See also settlements fauna, 70, 107, 246, 247 fences, 152, 244 Fertile Crescent, 107 fertility, human, 233, 165 fertilizer, 184, 219, 229; chemical, 187. See also dung; manure Feytha, 99 Fezzan Project, 162, 163, 170–2, 176; aims of, 165–8 Fezzan, the, 12, 150, 162–77; classical, 163; climate, 164, 169–70; early Holocene, 162, 164, 169–70; early Islamic, 168; early Modern, 167; Garamantian, 167, 168, 171–3; historical reconstruction, 167–75; landscapes, 162–5; late Pleistocene, 163, 169; later prehistoric and historic, 163; medieval, 167; mesolithic, 163; neolithic, 163–4, 165; palaeolithic, 163; population, 162. 171, 172; precipitation, 162, 163. 169; settlements, 172, 175. See also Sahara fields: clearance, 90, 91, 280; divisions, 72, 209–10, 213, 216, 217; maps, 142; systems, 51, 67, 72, 73, 74, 76, 78, 81, 89, 90–2, 98, 99–100, 205–13, 220, 253, 274–6; surveys, 142; walking, 68, 167 figs, 57, 138, 145
Index
358
fire, 283, 286; suppression of, 273, 283. See also burning; wood fishing, 7, 70, 142 flexibility, 10, 156, 268, 336 –7 floods, 6, 9, 71, 75, 92, 96, 128, 129, 151, 154, 193, 285; flash-floods, 7, 45; floodplains, 280; floodwaters, 88, 95, 129. See also erosion floodwater farming, 7, 55, 70, 75, 77, 78, 82, 90, 91, 99, 100, 146, 156, 172, 277–8. See also irrigation; risk management; Tripolitanian pre-desert; water: run-off flora, 70, 81, 148 fog/mist, 6, 19, 25, 33 foggara, 12, 113, 168, 169, 172, 173, 175. See also irrigation systems food shortages, 193, 194, 195 foot surveys, 51, 143 fortified buildings, 109. See also gasr/gsur forts, 51, 89, 92, 134, 162, 232. See also gasr/gsur, hillforts; military structures fowling, 70 France, 13; archaeologists, 90, 132–3; decolonization, 125; Tricastin region, 13, 341, 342–3. See also Rhône valley frosts, 292, 293 fruits, 52, 56, 70, 113, 138, 144, 230 fuel wood, 4, 11, 76, 80, 184, 191, 244, 283. See also wood funerary architecture, 173 furrows, 228, 229, 231, 232, 255, 262, 266, 268; furrow irrigation, 115, 252 Gairezi River, 222 Garama, 162, 166, 167, 168, 172, 173 –7 Garamantes people, 162–3, 172, 173, 175; development off oggara irrigation systems, 13, 168, 172–3, 174, 175; early Islamic period, 172–3; Garamantian civilization/culture, 167, 172, 175 gardens. See horticulture
Index
359
Gasr Lebr, 153 gasr/gsur (Romano-Libyan fortified farm), 140, 142, 147, 148, 153, 156, 173, 194; agriculture, 145–6 Gaza, battle of, 53 gazelle, 7, 70, 107 GCMs. See General Circulation Models Gebel Nafusa, 138 General Circulation Models, 36, 344 geochemical analysis, 67, 73, 77, 83 Geographical Information Systems. See GIS-based analyses Geoksyur oasis, 108 –9 geomorphology, 67, 70, 138, 157, 165–6, 217; tectonic activity, 67. See also alluvium; anthropogenic deposits; deforestation; degradation; dust storms; EDMA; erosion; floods; geochemical analysis; glaciers; gullies; land; sand; sedimentation; soil; streams; wadis; wind geosystems, 130, 131, 135 Germa/Old Germa. See Garama Ghat, 162 Gheriat el-Gharbia, 150 Ghirza, 148 Ghuwayr tributary, 64 Giba plateau, 181 Gila River, 275, 277 Gilbertson, D., 12, 13, 150 GIS-based analyses, 69, 83, 142, 348, 349 glaciers, 6, 325, 344; glacial meltwater, 106, 319, 320, 332 glass ware, 167, 173 global warming. See climate goats, 56, 58, 69, 70, 83, 107, 138, 144, 150, 203, 207, 229, 244, 246, 277, 294 gold, 173 Gonur Depe, 104, 109
Index
360
granaries, 212, 214, 244. See also storage Grand Bisse de Lens, 328, 329, 332, 333 Grand Erg Oriental sand sea, 128 grapes, 45, 52, 57, 109, 113, 142, 145, 172 grasses, 70, 107, 148, 185–6, 248, 342; medic-enriched grasses, 147, 156 grasslands, 18, 219, 275, 276, 283, 310. See also hay meadows gravel embanking, 215 grazing, 19, 143, 151, 195, 212, 276, 321. See also over-grazing Great Zimbabwe, 238 Greece, 341 Green agendas, 11 grindstones, 168, 171; grinding, 91 gsur. See gasr Guatemala, 292 Guinea, Gulf of, 31, 36 gullies, 55, 69, 143, 151, 153, 207, 239, 248, 336. See also erosion; geomorphology Gyaur Kala, 112, 113 gypsum, 166 gyttja, 127, 128 Haïdra, 134 Haluza. See Elusa Hamada al-Hamra, 138 Hamada, 87 handaxes, 168 Hapex-Sahel, 341 Harra plateau, 87–9, 90, 92, 100 Hauran, 87 –9 Haut Comtat, 349, 355 hay meadows, 324, 325, 336. See also grasses hearths, 68, 130, 212, 219 heavy metals, 68, 78 Helms, S., 98 Henchir Rayada, 134 Hérault-Gard, 351 herding, 70, 76, 107, 146. See also grazing hierachies, development of, 8, 108, 172, 265–8, 294, 301, 309, 310, 348, 350–6. See also elites
Index
361
highland areas, 19 Hilalian peoples, 134 hillforts, 168, 172, 173. See also forts Hohokam people, 276, 278, 282, 285 honey, 184, 212 Hopi people, 276 horses, 172, 183, 277 horticulture, 70, 183, 184, 227, 229, 232, 374, 247, 278–9, 286 houses, 90; mudbrick, 106, 108; stone, 70, 88, 182 humidity, 22, 23, 25, 35, 56, 322 hunter-gatherers, 70, 170, 276, 310 hunting, 7, 70, 107, 164, 170, 146 Huntingdon, Ellsworth, environmental determinism, 58 hurricanes, 35. See also winds Hutu people, 14 hyenas, 183 hyper-arid zones, 3, 18, 19, 34, 164. See also climate Ibn Khaldoun, 126, 134 imports, 8, 48, 55, 142, 155, 167 Incas, 8, 14 incision. See sedimentation India, 35 Indian Ocean, 32 indigenous peoples and local technologies, 8, 9–10, 11, 14, 45, 64, 70, 100, 144, 156–7, 162, 163, 165, 176, 219, 259, 285–6, 324 See also farmers; farming Indonesia, 36 industrial development, 286 Inner Asia, 310 insecticides, 190 insecurity, 143, 156 instability, 75, 144, 233. See also political stability/instability Inter Tropical Convergence Zones, 19, 20, 26, 27–8, 36. See also climate inter-cropping. See cereals; crops Iran, 20, 107, 113 Irano-Turanian desert steppe, 45 Iraqw people, 259–62, 267
Index
362
irrigation systems, 3, 8, 92, 93, 96, 99, 104, 108, 109, 112, 135, 154, 255, 256, 257, 277, 278, 285; agriculture, 56, 103–20, 112, 259; maintenance of, 114, 115, 119, 212–13; spray, 319; technology, 113, 115, 212. See also bisses; floodwater farming; salinization Islamic empire, 47, 56, 57; Islamic pastoralist invaders, 15–16 Isle of Braç, 341 isotopic dating, 55, 130, 134, 166, 344 Israel, 9, 14, 92 Issar, A., 56 Italy, 325, 341 ITCZ. See Inter Tropical Convergence Zones Jawa, 89 Jebel al-Arab, 87 –9 Jebel, 92, 100 – Jeitun, 105, 107 –8 Jericho, 70 Jewish immigration, 58 Jordan River, 15 Jordan, 7, 9, 10, 12, 45, 70, 76; Royal Society for the Conservation of Nature, 63 Josephus, 76 Kara Kum canal, 115 Kara Su (Black Water), 106, 107 kasbah. See markets Kenya, 212, 217, 257, 258, 262–5; Machakos district, 12. See also Africa Kerio Valley, 262, 265 Khirbet el-Umbashi, 89 Khirbet Faynan, 64–5, 70, 71, 76–80; abandonment of, 81 kibbutz movement, 15 Kinahan, J., 13 kinship, 232, 267, 310 Konso, 257, 259 Kopet Dag, 105, 106, 107, 108, 109, 112 Koppen, W., 18, 20. See also climate Krech, S., 286 Ksar Rheriss, 130 Ksar Rhilane, 128, 129
Index
363
Kurnub. See Mampsis kushetts, 183 Kwermusl, 260 Kyzal Arvat, 108 La Quemada, 310 La Tène Iron Age, 324 labour, 181–91, 194, 215, 256; communal, 231, 257, 264, 321; demands, 57, 145, 168, 226, 233; mobilization of, 231, 256, 263–4; tribute, 231, 263–4, 306 Lake Eyasi, 218 Lake Nasser, 35 Lake Natron, 218, 219 lakes, 7, 70, 164, 165–6, 168, 169–70, 147, 341, 344. See geomorphology; sediments: lacustrine; soil land: abandonment, 336, 339, 341–3; carrying capacity, 13–14, 58, 343; clearance, 236, 245; collectivization of, 119; management, 13, 56, 81, 253; ownership, 153, 187, 257, 263, 264, 319; rights, 324; shortages, 192; use, 5, 12, 131 Landsat images, 143 landscape archaeology, 4, 64 –83 landscape degradation, 11, 55, 58, 80, 115, 157, 235, 249, 258, 341; human impact on, 58, 75, 81–2, 134, 125, 215–16, 233–4, 279–81, 283, 336, 339; ‘marginal’, 9 Languedoc-Roussillon, 341, 351 Lattimore, O., 310 lava, 87 Le Houerou, H.N., 34, 36 lead, 78 legumes, 70, 184, 192–3, 227, 230 Lemasba, 150 Letsibogo, 236 –49 Levant, 47, 54, 56, 57, 70, 76; early Holocene , 108 Levant, 67 Lewis, R.A., 107, 109 Libya, 8, 92, 131, 146, 150; foggaras, 8;
Index
364
language, 171; oases, 9; written scripts, 171. See also Fezzan; Tripolitanian Pre-Desert; ULVS Limes (Roman frontier works), 130, 131 Limes Palestina, 48 Limpopo River, 222, 238 Lisitsina, 107 lithic assemblages, 68, 69, 70, 168, 170, 172 Little Ice Age, 134, 230, 233 See also climate Liverani, M., 162 livestock, 4, 100, 142, 151, 183, 248; livestock pits, 226, 227; livestock enclosures, 152, 182, 191, 207–8, 209, 227, 230, 236, 238, 241, 243, 244, 245, 246. See also animals; cattle; goats; grazing; nomads; pastoralism; sheep llamas and alpacas, 292, 294 Lolmalasin Mountain, 203 Lunel Viel, 350 Lunellois/Vaunage (western Lower Rhône), 350 Maasai tribespeople, 203, 207, 218, 219, 268 Maasailand, irrigation farming, 203 –19 Macae tribespeople, 142, 156, 157 Machakos, 258, 265 Madama, 76 Maghreb, 13, 126; aggradation period, 127, 129–31; agriculture, 128; climate, 126; 127, 128, 130–1, 134, 144; colonization: Arab, 133–4, Byzantine,132–3, Phoenician, 125–7, Roman, 13, 127–32, 134, Vandal, 132, 133; conquests and land degradation, 125–34; eastern, 13, 125–34; Holocene, 125; incision period, 127–8; Phoenicians, 125–7; precipitation, 13, 127, 131, 134;
Index
365
Roman colonization, 13, 127–32, 134; settlements, 130; Vandal colonization, 132, 133. See also ULVS maguey, 280, 292–311; and seed cultivation, 291–2, 295–7, 305–6; decline of, 291; fibre, use of, 296–7; importance as food source, 292–6; sap extraction, 302–4; significance of as crop, 290–2; technology of, 292, 297–304, 292 Mahabere Genet, 181, 193 Mai Kayeh, 186 maize. See cereals Makuyuni river, 203–5, 206, 216 Mali, 23 Maltese Islands, 337 Mama Issara. See Africa: drylands Mampsis, 48, 51, 53, 56 Mamshit. See Mampsis manganese oxide, 130 manure, 83, 146, 151, 187, 210, 216, 219, 225, 229, 232, 257, 259, 261. See also dung; fertilizer Manyika people, 229, 230 Mapungubwe, 238 Marakwet. See Africa: drylands Marateng, 256, 259 marble, 49 Marina Baixa, 341 Maristvale, 226, 228 markets, 8, 12, 53, 173, 192, 194, 265, 266–7, 303; expanding, 144, 154, 211, 320–1; loss of, 142; Mekelle, 181, 184, 192 marshes, 165 Masson, V.M., 107, 109 Maungwe people, 229 Mauretania, 30 McGinnies, W.G., 19 McIlveen, R., 27 meat, 107, 146, 210 mechanization programmes, 11, 286 MEDALUS Project, 341 ‘Medieval Warm Epoch’, 238. See also climate
Index
366
Mediterranean zone, 9, 12, 45, 47, 48, 126, 131, 156, 172; basin, 134, 141, 160, 315: agricultural system 51, 56, 145; East Mediterranean, 75; Mediterranean cold fronts, 31 Meigs, P., 20 Mekelle, 181, 192, 193 Merv, 113; decline of, 119; International Merv Project, 112; oasis, 104, 109, 110, 111, 114; productivity, 113–14; settlement pattern, 111–14 Mesa Verde, 278, 283 Mesoamerica, 277, tierra fría 288–308; civilization, expansion of, 292, 299, 305–8; population, 307; precipitation levels, 288, 290; tierra caliente, 288; tierra templada, 288, 300. See also maguey Mesopotamia, 105, 113, 294 metallurgy, 73, 240, 246 Mexican Revolution, 299 Mexico City, 299 Mexico, 9, 275, 277, 292, 294, 295, 299, 310; highlands, 9, 291; Valley of, 297, 300 middens, 73, 147, 150, 229, 230, 283, 296 Middle East, 25 Midi-Pyrénées, 341, 350 migration, 109, 172, 266. See also mobility; settlement; transhumance military garrisons/structures, 51, 78, 98, 100, 112, 145. See also forts military invasion, 12, 13, 58, 156; Vandal conquest, 127 milk, 107, 146, 210, 218 millet. See cereals mills, grain, 78, 100 mills, ore-crushing, 80 Mimbres valley, New Mexico, 283, 284 –5 minerals, 47, 65. See also copper; pollution;
Index smelting mining, 10, 68, 73, 76, 77, 82, 83 Mirab bashi (chief water master), 114, 119 Mizda, 138 Mmadinare, 237, 244, 249 Moab, 76 mobility, 69, 100, 144, 157, 285, 286, 329. See also migration molluscs, 56, 128, 130 Moloko. See Sotho-Tswana people monastic foundations, 98 monetization, 11 Mongol invasions, 112, 119 Mongolia, 20 monks, 100 monsoon rains, 32, 56, 193, 276 Montélimar, 342 Montpellier, 350 Morocco, 135, 150, 155 mortar, 71 mortgages, 9 Mortimore, M., 11, 255 mosaics, 49 Moscow, 104, 115 mosques, 54–5, 58 Motloutse River, 236–9, 246 Mount Kilimanjaro, 212 –3 Mount Meru, 212 Mount Nyangani, 222 mountain barriers, 25, 33, 275 Mozambique, 222 Mpumalanga province, 256 Mukaddasi, 113 mulching, 259. See also rock mulching mules, 184 Murgab oasis, 109 Murgab River, 106, 109, 114 Murzuk, 174 Mutapa state, 230 Nabatean caravanserai, 48 Nahal Lavan, 52 Nahal Mitan, 54 Nahal Nizzana, 46 Nahal Oded, 56 Nairobi, 265 Namazga, 108 –9
367
Index
368
Namibia, Walvis Bay, 22 Native Americans, 15, 286 natron, 173 Natufian peoples, 70 nature-culture debate, 356 Near East, 7, 14, 36, 70, 73, 75, 112 Negev, 45–59; Abassid, 53, 54, 57; Byzantine, 47–52, 49; classical, 54–5, 56; climate, 46, 54–6, 58; desertification, 46–7, 57–8; early Roman and Nabatean, 47, 55–6, 63, 65; Islamic conquests, 52–4, 58; population, 45, 57; precipitation, 51; rise of desert, 56–7; settlements, 49: abandonment of, 45–6, 72 Nemencha Mountains, 128, 129 Neo-Babylonians, 76 Nepal, 283 Nessana, 53 New Mexico, 275, 276, 279, 283 Nguni people, 256 Niger, 162 Nigeria, 6, 229, 256; Kano, 12 Nile basin, 150, 155 Nile River, 150 Nilotic cattle-herders, 14 Nilotic/Bantu dichotomy, 15 Nîmes, 350 Nir, 18, 21 Nizzana. See Nessana nomadic pastoralism, 13, 58, 64, 76, 100, 134. See also bedouin; pastoralism nomads, 46, 55, 87, 89, 98, 126, 143 North Africa, 5, 11, 14, 19, 28, 29, 35, 126, 150, 170 North America, 5, 7, 9, 10; North American Southwest, 9, 146, 151, 271–83. See also Arizona; New Mexico: Classic Mimbres period, 280, 281–2; population levels, 273, 280, 281–2, 283; precipitation levels, 271–3, 276, 279, 281–2; prehistoric agricultural practices, 271, 273–82; topography, 271–3
Index
369
nuts, 113, 144 Nyanga National Park, 228, 231 Nyanga town, 226, 228 Nyanga. See Zimbabwe Nyangombe river, 222, 229, 231 oases, 106, 150, 162, 166, 167, 169; architecture, 109; cultivation and technology, 171; settlements, 98 Oboda. See Avdat ocean temperatures, 6, 33; sea surface temperature anomalies, 37 oil, 113, 182 Oldonyo Lengai, 218 Oldonyo Sambu, 220 Olemelepo River, 205, 206, 216 olives, 52, 57, 80, 138, 147, 149, 151, 172; oil, 12, 141, 144, 167; presses, 51, 146 Optical Spin Luminescence. See OSL oral traditions, 227, 237, 295 Orange, Cadastre B tax map, 345 –8 OSL, 67, 166, 108 Östberg, W., 265 ostrich eggshells, 168; shell beads, 167, 168 over-grazing, 3, 12–3, 34, 45, 55, 58, 143, 157, 184, 248. See also livestock oxen, 181, 184, 188, 194, 195. See also animals Pacific Ocean, 35, 292 Pakistan, 22, 35 palaeoecology, 14, 67, 147 –8 paleoclimatology, 150, 155, 247. See also climate paleoeconomic analyses, 147 paleoenvironment, 147–8, 166 paleoethnobotanical assemblages, 166 See also cereals; plants Palestina Tertia, 47, 53 Palestine: ancient, 15, 65, 77; British administration of, 58 Palmer, E., 48, 58 Palmyra, 99
Index
370
Pamirs mountains, 106 Pare, 212 –3 Parthian empire, 112 pastoralism, 12–3, 203, 255, 292, 310, 323; pastoral encampments, 51–2, 72–3, 92, 98; pastoralists, 15, 16, 92, 98–9, 100, 137, 163. See also bedouin; livestock; transhumance Penman, H., 18, 36 Pennine Alps, 322 Persian armies, 54 Peru, 20, 33 Petra, 64, 70, 77 Phaino settlement, 66, 78 Phoenix, Arizona, 277, 278, 284, 285 phosphate analysis, 230 pigs, 56, 294 pilgrimage sites, 48 pioneer culture, 12, 144 plague, 325 plants: cultivation, 7, 70; medicinal, 246; parasites, 187–8; remains, 107, 108, 167; root, 148, 225; wild, 184–6, 244. See also cereals; crops; fruits; maguey; pulses; trees; vegetables; vines; weeds plateaux, 64, 68, 87, 143, 138, 153, 164, 169, 222, 224 Pleistocene (Ice Age), 6 Pliny the Elder, De Re Agraria, 345 ploughing, 9, 132, 184, 187, 189 political agendas, impact of, 14, 58, 116, 233 political conditions, 9, 14, 57–8, 114, 143, 156, 173–5, 194, 240, 255, 285, 292–3, 346. See also instability; stability; state systems political economy, 265 –8 political stability/instability, 12, 119, 156, 182, 195.
Index
371
See also instability political systems, 14, 194, 262–5, 277 pollen evidence, 80, 128, 147, 148, 283 pollution, 77, 80, 83; metalliferous, 9, 72, 76, 81, 82. See also smelting Polynesia, 299 ponds, 96, 184 pools, 96, 97–8, 114, 164 population, 116; density, 252, 263, 283; expansion, 3, 12, 13, 119 Portugal, 232 pottery, 66, 68, 70, 74, 78, 92, 130, 168, 170, 217, 225, 240, 246; Islamic, 133; potsherds, 74, 82, 97. See also ceramics poverty, 12 precipitation, 13, 18, 20, 28, 29, 30, 33, 34, 36, 45, 150, 186–7; seasonality, 20–1; summer rains, 20–1, 38, 186, 187; variability, 22–3, 30; winter rains, 20, 38, 104, 106, 221. See also climate predation, 9, 183, 188, 194, 195, 230, 280 prestige items, 310 –1 prickly pear, 185 Provence-Alpes-Côte d’Azur, 350 pulses, 109, 112, 187, 189, 192–4, 212, 216, 219, 277, 293, 299 Pungwe River, 222 Punic Wars, 128 qanats, 8, 105, 113, 168 Qasr Burqu’, 89, 96–8, 99 – Qasr el-Gherbi, 98 radiation: budget, 6; loads, 23, 25, 28; solar, 36. See also climate radiocarbon dating, 67, 107, 130, 134, 225, 231, 239, 247 raiding, 76, 143, 173, 219, 233 rainfall. See precipitation rain-fed farming, 219, 238, 239, 278 rainshadow effect, 25 Ramon Crater, 56 Raron (Rarogne), 325 raw materials, local, 48
Index
372
Razik Dam, 113 refugees, 3 regional contexts, 8, 276 Rehovot. See Ruheiba religious beliefs, 260 rents, 175, 187, 188 research, interdisciplinary, 4, 126, 147, 258, 341, 356–7. See also ARCHAEOMEDES Project reservoirs, 65, 80, 89, 96, 97, 98, 193 resettlement, 9; compulsory, 81, 194, 227, 230 Rharsa chott, 128 Rhône Valley, 12, 322, 336, 341–57; lower, 337, 338: settlement trends, 340, 341–6; Valdaine region, 337, 339: climate, 337–9; Roman colonization, 339–45, 338: ‘crisis’ period, 341–5 Rio Bravo del Norte River, 275 Rio Grande River, 275 risk management, 5, 8, 10, 64, 154, 232 ritual, 165 road networks, 345–6, 350 rock art, 164, 165, 170, 172; inscriptions, 57 rock mulching, 280–2, 286 rock shelters, 165, 170 Roman empire, 12, 47, 64, 66, 78, 87, 98–9, 100, 134, 143, 145, 155–6, 157, 162, 167, 172, 344– 50; army, 92, 97, 100; drainage systems, 13, 79; economic needs of, 9, 143, 144; engineers, 79; farmers, 144; technocrats, 142, 143, 144 Roman-Byzantine Empire, 53 roofs, 183 Ruhba, 89, 99 Ruheiba, 48, 54, 56 run-off water catchment systems. See floodwater farming; irrigation; water: run-off Ruvalcaba, J., 298 Rwanda, 15
Index Sabi River, 222 Safford, Arizona, 279 Sahara, 12, 19, 21, 30, 113, 131, 132, 172; culture, 162, 171; oases, 160; precipitation, 6–8; subzones, 127, 128, 129, 131. See also climate; Fezzan Saharan depressions, 31 Saharan Sahel, 19, 22, 27, 31, 34, 35, 167 Sahlins, M., 310 Saleh, Algeria, 23 salinization, 4, 104, 113, 115. See also soil Salt River, 277, 278 salt, 128, 173, 182; salt flats (‘chott’, ‘sebkhs’), 125, 126; pans, 7 sand: dunes, 56, 127, 128, 129, 164, 165; movements, 46, 127, 129; seas, 127, 162, 163, 168, 169, 170 Saniat Gebril, 168 Santa Fe, New Mexico, 279 Santorini, 76 SAO. See South Atlantic Oscillation Sasanian empire, 112; records, 104 Sassanid people, 54 satellite imagery, 168 screes, 150 scrub. See thorn scrub Sede Boqer, 54 sedentary populations, 100, 131 Sedibe River, 131 sedimentation, 130–1, 143, 151, 152, 154, 336; aggradation, 126, 127, 129–31, 133, 149; fluvial deposits, 107; incision, 126, 127–8, 130, 133, 149, 337 sediments, 67, 69, 71, 73, 78, 80, 81, 129, 148, 166; colluvial, 128; core, 55; lacustrine, 166, 215. See also geomorphology; soil seeds, 70, 107, 147, 212, 230, 284 Seleucid empire, 112 Seljuks, 112
373
Index
374
semi-arid zones, 3, 18, 19, 28, 36, 126, 128, 255, 265, 286, 292 semi-precious stones, 173 Serengeti plains, 219 Serir Tibesti, 133 settlements: abandonment of, 5, 6, 10, 13, 53, 56, 57, 81, 86, 172, 236, 253, 256, 280; bronze age, 77; iron age, 76, 215; Islamic, 12–13, 86; late Byzantine, 86; Libyan Pre-desert, 12–13; Natufian, 69; pastoral-nomadic, 52, 54, 55; patterns of, 4, 67, 247, 252–64, 67; Romano-Libyan, 12–13, 152; Southern Bantu, 244, 247 Shashe-Limpopo basin, 236, 247 – Shayqar tributary, 64 sheep, 57, 70, 107, 138, 144, 203, 229, 244, 246, 277, 292, 294. See also animals shellfish gathering, 7 shells, 166 Shewa province (Ethiopia), 187 Shipton, P.M., 267 Shivta. See Subeita shrubs, 151, 184, 276, 283, 342 sickle blades, 107 Sierra Nevada, 25 Sierre, 325 silts, 129, 132 Sinai, 45, 48, 59 skins, 107, 184 slag, 77, 246 slaves, 65, 78, 173; slave societies, 171 sluice gates, 51, 77, 99 slurry, 228, 229 smelting, 68, 73, 76, 77, 80, 82 See also pollution snow, 105, 106, 324 social conditions, 3, 7, 9, 14, 57, 73, 75, 100–, 114, 143, 155, 165, 194, 238, 255, 259, 265, 267, 285, 337, 346. See also elites; hierarchies; instability; stability soil, 18, 129, 138, 145, 183, 235, 346; conservation, 230–1, 255;
Index
375
erosion, 11, 12, 13, 14, 75, 128, 131, 142, 143, 236, 247, 337, 338–9; fertility, 213, 223, 231, 241, 242, 245, 246, 252, 254, 256, 279, 281, 332: loss of, 144, 276; management, 112, 230; maps, 341; modification, 279, 281; nitrogen levels, 147, 241, 337; reclamation, 152; salinity of, 20, 107, 112, 332; salinization, 115, 143, 150, 281; waterlogging, 115, 143, 193, 209, 224, 226, 230, 332. See also geomorphology soldier-farmers, 100, 156, 346 –8 Somalia, 20 Sommières, 350 Sonjo villages, 212, 215, 219–, 257, 258, 267 Sonoran desert peoples, 276 sorghum. See cereals Sotho-Tswana people, 238, 239; settlement system, 253 South Africa, 10, 255 South America, 33, 35 South Atlantic Oscillation, 35 South Australia, 9, 147 South Turkmenistan Multi-Disciplinary Archaeological Expedition, 104 Southern Bantu settlement pattern, 247, 249 Soviet Union, break-up of, 104, 119; Soviet-built irrigation systems, 103, 114–15, 119. See also Turkmenistan Spain, 21, 113, 341 specialization, 76, 219, 229, 267, 295, 302, 309–10, 311 spices and herbs, 185 spillways, 78, 96 spindle whorls, 296, 302, 311; ceramic/clay, 243, 298, 300, 302 spindles, wooden, 301 spinning and weaving, 301–3. See also cotton springs, 64, 70, 106, 138, 164, 165, 166, 169, 213, 218 squashes, 277, 293, 299 SSTA. See ocean temperatures: sea surface temperature anomalies St Maurice and Sion, Bishopric of, 325, 332 stability, 3, 14, 52, 53, 57, 156–, 268, 349–50. See also environment and human activity; land: carrying capacity state systems, development of, 108, 309 –11
Index
376
STH centres. See subtropical high pressure centres stone: cairns, 205, 208, 210; clearance, 50, 91, 153, 205, 209, 222, 223, 230; troughs, 90; use of, 95, 98, 172, 182, 205–7, 209–10, 211, 217, 226, 230, 238 storage facilities, 142, 232, 246, 247; grain storage, 190–1, 183, 194, 209. See also granaries; water streams, 7, 107, 109, 129, 213, 214, 216–7, 218, 228. See also geomorphology Subeita, 47, 48, 51, 52, 54, 56 sub-humid regions, 18 Sub-Saharan Africa. See Africa subsistence farming. See farming subtropical high pressure centres, 25–7 Sudan, 168 Sudano-Sahelian depressions, 33, 34 Sultan Kala, 112 summer rains. See precipitation Summers, R., 230 Susiana, 113 Sutton, J.E., 229, 256 Switzerland, traditional irrigation techniques. See bisses Syria, 20, 23, 87, 90, 99, 100 Syrian Black Desert: population, 100; precipitation, 86; Roman period, 86, 88, 93; settlements, 86, 88, 97, 99; topography, 86; water supply and farming , 86–100 Tahirbaj Tepe, 112 Tahiti-Darwin Sea, 35 Taita, 213 Takhirbai, 112 takyrs, 108 Tanganyika, British mandateship of, 207 Tanzania, 212, 217, 256, 257, 258–62, 266; Mbulu District, Northern Tanzania. See Africa: drylands. See also villagization
Index
377
Tarahumara people, 297, 309 Tatoga tribespeople, 218 Taurus Mountains, 87 tax issues, 145, 175, 324 Tejen River, 106, 107, 108 Tejen, 108 teleconnections, 35 –6 Tell Wadi Faynan, 70, 71, 73 temples, 51, 173 Tenochtitlan, 292 tents, 53, 68, 99, 157 Teotihuacan people, 292 terrace systems, 7, 11, 45, 73, 207, 215, 219, 222–5, 228–9, 230–1; hillslope terracing, 56; maintenance, 54–5, 57; terrace erosion, 54–5; terrace walls, 72, 76, 91, 171, 183, 210; terraced dams, 50; terraced fields, 75, 77, 78. See also wadis; walls textiles, 300, 301, 303, 310, 311 Thera, 76 Thomas, D.G. and Middleton, N.J., 10, 11 thorn scrub, 18, 153, 212, 238, 240, 248 Thornthwaite, C.W., 18, 20 threshing, 184, 187, 190 Tibet, 20 tierra caliente. See Mesoamerica tierra fria. See Mesoamerica tierra templada. See Mesoamerica Tigrai province (Ethiopia), 181, 184, 185, 186, 187, 192; precipitation, 193–4 Timbuktu, 23 Timurid people, 112 Togolok, 112 tomatoes, 184 tools, 107, 169, 191, 304; ceramic, 297–8; flint, 69, 168, 169; iron, 302–3; scraping, 293, 300–4; stone, 163, 297–8 Tot, 262 tourism, 14, 350 Toutswe people, 238 towers, 96, 98 towns, 45, 48, 56, 133.
Index
378
See also urbanism trade, 57, 142, 156, 162, 166, 172, 266; routes, 47, 72, 98, 166–7, 211, 231. See also exports; imports Traghen, 175 transhumance, 138, 142, 156, 324. See also pastoralism transport, 172, 214, 266. See also road networks travellers, 64, 99, 105; European, 86, 174, 229 trees, 132, 138, 183, 184, 191, 258, 342; crops, 51, 141, 144, 150, 156, 317; planting, 12, 183. See also deforestation; floodwater farming; wood tribal identities; 173; tribal tribute, 299, 306 Tricastin. See France trincheras, 278, 284. See also dams Tripoli, 162; Department of Antiquities, 139 Tripolitania, 4, 92, 138 Tripolitanian Pre-desert, 7, 10, 138–57; environment, 146–50; population, 141, 154, 155, 156; precipitation, 147, 149, 153; Romano-Libyan floodwater farming;137, 141, 144, 149–53, 154, 146, 150: decline of, 155–6; Romano—Libyan and later settlement, 141, 152; settlements, abandonment of, 142–4, 147, 154–5; settlements, expansion of, 144–5, 154. See also geomorphology; ULVS tropical forests, 35 troughs. See stone Tswana people, 249, 255, 267 Tucson, Arizona, 276, 279 Tula, 292 tuleiliot el anab, 51 Tunisia, 126, 129–31, 134 Turkey, 175 Turkmenistan, 7, 104–20; climate, 106, 107–8, 109;
Index
379
environment, 104–6; irrigation agriculture, 103–20 and see following periods: Achaemenid period, 9, 104, 111: chalcolithic, 108: early bronze age, 108: eneolithic/ iron age, 108–11: middle bronze age, 108, 111, 112: neolithic, 106–8: Namazga III period, 108: Parthian, 111, 112–13: Sasanian, 112–13: Seleucid (Hellenistic) period, 111: Seljuk/post-Seljuk, 112: Tsarist/ Soviet periods, 9, 11, 114–15; population, 103, 109, 113, 119; precipitation, 104–6; settlement, 116; settlement abandonment, 108. See also qanats Tutsi people, 14 Tyson, P.D. and Lindsay, J.A., palaeoclimatic model, 247 Uganda, Bwindi National Park, 14 ujamaa movement. See villagization ULVS. See UNESCO Libyan Valleys Survey Ummayad empire, 53, 54, 55 UNEP, 18, 322 UNESCO Libyan Valleys Survey, 12, 138, 139, 140, 142–3, 146 UNESCO, 20 United States, 36, 114, 265–6, 304, 310 Universal Transverse Mercator, 68 universities, 140 Unyama people, 230, 233 urbanism, 108, 112, 115, 165, 166–7, 172, 173, 175, 277, 325, 353, 354; urban decline, 45–6, 52, 53, 75; urban/rural interaction, 345–50; urban systems, 45, 72, 111–14, 172, 343, 345–6. See also settlements; towns UTM. See Universal Transverse Mercator Valais canton, 9, 321–37; agricultural patterns, 317–19; climate, 317, 320; population growth, 320; precipitation, 317, 319; topography, 315–17.
Index
380
See also bisses Valdaine. See Rhône valley van der Veen, M., 167 vaults, 49 vegetables, 113, 230, 324. See also horticulture vegetation cover, 18, 34, 64, 129, 134, 156, 169, 183, 191, 203, 248, 276; loss of, 4, 11, 12, 13, 58, 82, 131, 143, 156. See also deforestation; geomorphology Veneto, 341 Venezuela, 19 Vera Basin, 341 Vernamiège, commune of, 322, 323, 329; irrigation sectors of, 326, 327. See also Valais canton Viège. See Visp villages: Engaruka, 207–8, 215; Fezzan, 167, 168, 172–3, 174; Mesoamerican, 273; Negev, 53; Nyanga, 229; Rhône Valley, 345; Sonjo, 209, 210, 217–18, 218; Syrian Black Desert, 88, 92, 99–100. See also villagization villagization, 181, 195, 207, 219 Vincent, L., 324 vines, 52, 323, 324 Visp, 325 volcanic activity, 218 Wadi Arabah rift valley, 64, 68, 71 Wadi Bou Jbib, 131 Wadi Chéria-Mezeraa, 130, 132 Wadi Dana, 70 Wadi el Akarit, 130, 134, 133 Wadi el-Agial (al-Hayat), 163–4, 166, 168, 169, 172, 175 Wadi el-Amud, 143, 147, 150, 156 Wadi es Sgniffa, 130 Wadi Faynan, 7, 9, 10, 64 –8364; bronze age, 72–5; chalcolithic, 65; ‘classical’, 67; climate, 70; field systems, 65, 66, 68, 77; Holocene, 69, 70, 81;
Index
381
iron age, 75–6; Nabatean, 77, 81; neolithic; 65, 69–71, 81; Pleistocene, 69; population, 82; Post-Byzantine, 80–1; Roman Imperial, 77–80; settlements, 68, 69, 71, 72 Wadi Faynan Landscape Survey, 12, 64, 67 Wadi Fidan, 76 Wadi Gharaz, 91 Wadi Ghuwayr, 70, 71, 80 Wadi Gobbeen, 140, 143 Wadi Kébir-Miliane, 130, 134 Wadi Leben, 130 Wadi Mansur, 143, 147, 153 Wadi Merdum, 138 Wadi Mimoun, 143, 147, 152 Wadi Saad, 92, 96 Wadi Sham, 89, 90, 92 –6 Wadi umm el-Kharab, 140, 143, 150, 154 wadis, 12, 45, 55, 70, 73, 77, 78, 89, 138, 143, 151, 154, 164; downcutting, 54–5, 79–80, 128; wadi floors, 51, 70, 149–50; wadi walls, 139, 150, 151–3. See also geomorphology; ULVS Waldmatte, 325 walls, 68, 90, 142, 143, 167, 175, 231, 239; boundary, 72, 76; desert, 151–3; functions, 8, 50, 75, 77, 151, 153, 274, 276: diversion, 74, 78, 98, 139, 171, 205: fortification, 50, 108; wall networks, 137, 151–4: maintenance of, 155–6. See also floodwater farming; geomorphology; terrace systems: walls war, 134, 175 warlords, 142 water: conflicts over, 323, 328, 330–1, 333; distribution, 113, 325–7: deterioration in, 9, 103, 115, 119–20, 213–15, 216; domestic uses, 183, 194, 211, 226; gauges, 113;
Index
382
harvesting, 8, 88, 144–5, 149, 152, 162, 257; loss via seepage etc., 78, 99, 115; mills, 65, 90, 91–2; pumps, diesel and electric, 163; resource management, 15, 21, 81, 86, 91, 99, 100, 112, 113–15, 119–20, 220, 226–7, 230, 252, 254: misuse of, 12, 115, 175; rights, 92, 113, 260, 263, 264, 282, 320, 321–8, 333; run-off, 50, 91, 107, 152, 153, 156, 274: catchment systems, 45, 48, 57, 91, 92, 151–3, 183, 194, 201: maintenance of, 103, 115; shortage, 319; storage, 98, 99, 100, 152, 153; supplies, fossil, 162, 175: underground, 149, 163–4, 90; tables, 103, 107, 114, 115, 162, 163, 168, 337. See also climate; floods; geomorphology; wells water-users, association of, 325 wealth, 49, 52, 167, 173, 232, 310 weaving, 301, 303 weeds, 167, 184, 187, 283 –4 Weintraub, P., 311 Wello province (Ethiopia), 187 wells, 90, 138, 164, 168, 175, 184, 237; artesian, 137, 174–5 West Africa, 27, 30, 35; ‘ring cultivation systems’, 252 West African monsoon trough, 29, 30 Western Australia, 22 wheat. See cereals Whitelaw, T., 22 Widgren, M., 8 wild animals, 146, 155, 170, 203, 212, 229, 244, 247. See also animals windmills, 98 winds, 5, 23, 25, 29, 33, 323; anti-trade winds, 31; direction, 25–6; East African Low Level Jet, 32–4; Hadley Cell circulation, 26, 27, 28–9, 30, 31, 34; Subtropical Westerly Jet, 31; trade winds, 27, 29; Tropical Easterly Jet, 31–2; West African Mid-Tropospheric Jet, 32. See also anti-cyclones; climate;
Index
383
cyclones; hurricanes; precipitation wine, 45, 167; presses, 45, 51, 47 winnowing, 188, 190 winter rains. See precipitation winters, 20, 128 Wittfogel, K., ‘oriental despotism’, 264 –5 woina dega, 181 women, 188, 191, 329 wood: forest foods, 70; timber, 12, 191, 227, 238, 241; uses of, 215; woodcutting, 81, 215, 280; woodlands, 6, 127, 134, 211, 237, 255, 273, 280. See also deforestation; degradation; geomorphology; trees wool, 107, 146 World Archaeological Congress, 4 –5 World Meteorological Organization, 21 Wright, K., 73 Wyckoff, D.G., 283 Yakut, 114 Yaz Depe, 112 Younger Dryas, 70. See also climate YuTAKE. See South Turkmenistan Multi-Disciplinary Archaeological Expedition Zagros Mountains, 107 Zambezi River, 222, 233 Zhizo farming settlement, 238–9, 247, 248 Zimbabwe, 212, 238, 247; Nyanga region, 208, 220–32, 221, 253, 255, 256; agriculture, 227–8; landscape, 221–7, 229–31; population, 229, 230, 231, 232; precipitation, 220, 221; settlements, 220, 229–30, 231; topology, 220–1; Rusape/Headlands area, 230 Zinchecra, 168, 172 Zionism, 15, 58 Ziwa, 230; Ziwa ruins National Monument, 229–30
Index Zuila, 175
384