the archaeology of drylands: living at the margin (one world archaeology)

Administrator
File Attachment
200045d6coverv05bjpg

THE ARCHAEOLOGY OF DRYLANDS

The One World Archaeology (OWA) series stems from conferences organized by theWorld Archaeological Congress (WAC) an international non-profit making organization which provides a forum of debate for anyone who is genuinely interested in or has aconcern for the past All editors and contributors to the OWA series waive any fees theymight normally receive from a publisher Instead all royalties from the series are receivedby the World Archaeological Congress Charitable Company to help the wider work ofthe World Archaeological Congress The sale of OWA volumes provides the means forless advantaged colleagues to attend World Archaeological Congress conferencesthereby enabling them to contribute to the development of the academic debatesurrounding 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 1ndash37) Peter JUcko

Academic Series Editors (Volume 38 onwards) Martin Hall and Julian Thomas Executive Series Editor (Volume 38 onwards) Peter Stone

1 What is an Animal TIngold (ed)

2 The Walking Larder Patterns of domestication pastoralism and predation JClutton-Brock

3 Domination and Resistance DMiller MJRowlands and CTilley (eds)

4 State and Society The emergence and development of social hierarchy and political centralization JGledhill BBender and MTLarsen (eds)

5 Who Needs the Past Indigenous values and archaeology RLayton (ed)

6 The Meaning of Things Material culture and symbolic expression IHodder (ed)

7 Animals into Art HMorphy (ed)

8 Conflict in the Archaeology of Living Traditions RLayton (ed)

9 Archaeological Heritage Management in the Modern World HFCleere (ed)

10 Archaeological Approaches to Cultural Identity SJShennan (ed)

11 Centre and Periphery Comparative studies in archaeology TCChampion (ed)

12 The Politics of the Past PGathercole and DLowenthal (eds)

13 Foraging and Farming The evolution of plant exploitation DRHarris and GCHillman (eds)

14 Whatrsquos New A closer look at the process of innovation SE van der Leeuw and RTorrence (eds)

15 Hunters of the Recent Past LBDavis and BOKReeves (eds)

16 Signifying Animals Human meaning in the natural world RGWillis (ed)

17 The Excluded Past Archaeology in education PGStone and RMacKenzie (eds)

18 From the Baltic to the Black Sea Studies in medieval archaeology DAustin and LAlcock (eds)

19 The Origins of Human Behaviour RAFoley (ed)

20 The Archaeology of Africa Food metals and towns TShaw PSinclair BAndah and AOkpoko (eds)

21 Archaeology and the Information Age A global perspective PReilly and SRahtz (eds)

22 Tropical Archaeobotany Applications and developments JGHather (ed)

23 Sacred Sites Sacred Places DL Carmichael JHubert BReeves and ASchanche (eds)

24 Social Construction of the Past Representation as power GCBond and AGilliam (eds)

25 The Presented Past Heritage museums and education PGStone and BLMolyneaux (eds)

26 Time Process and Structural Transformation in Archaeology SEvan der Leeuw and JMcGlade (eds)

27 Archaeology and Language I Theoretical and methodological orientations RBlench and MSpriggs (eds)

28 Early Human Behaviour in the Global Context MPetraglia and RKorisettar (eds)

29 Archaeology and Language II Archaeological data and linguistic hypotheses RBlench and MSpriggs (eds)

30 Archaeology and Anthropology of Landscape Shaping your landscape PJUcko and RLayton (eds)

31 The Prehistory of Food Appetites for Change CGosden and JGHather (eds)

32 Historical Archaeology Back from the edge PPAFunari MHall and SJones (eds)

33 Cultural Resource Management in Contemporary Society Perspectives on managing and presenting the past FP MacManamon and AHatton (eds)

34 Archaeology and Language III Artefacts languages and texts RBlench and M Spriggs (eds)

35 Archaeology and Language IV Language change and cultural transformation RBlench and MSpriggs (eds)

36 The Constructed Past Experimental archaeology education and the public PGStone and PPlanel (eds)

37 Time and Archaeology TMurray (ed)

38 The Archaeology of Difference Negotiating cross-cultural engagements in Oceania RTorrence and AClarke (eds)

39 The Archaeology of Drylands Living at the margin GBarker and DGilbertson (eds)


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 amp Francis Group

This edition published in the Taylor amp Francis e-Library 2005

ldquoTo purchase your own copy of this or any of Taylor amp Francis or Routledges collection of thousands of eBooks please go to

wwweBookstoretandfcoukrdquo

copy 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 DesertsmdashHistory 5 Land settlement mdashHistory 6 Land settlement patterns PrehistoricmdashHistory

7 Arid regions agriculturemdashSocial aspectsmdashHistory 8 Climatic changesmdashHistory I Barker Graeme

II Gilbertson DD III Series CC724A735 2000

9301ndashdc21 00ndash038257

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 ix List of tables xiii List of contributors xv Series editorsrsquo foreword xxv Preface xxvii

Part I Introduction

1 Living at the margin themes in the archaeology of drylands Graeme Barker and David Gilbertson 3

2 The dynamic climatology of drylands Greg Spellman 18

Part II Southwest and Central Asia

3

The decline of desert agriculture a view from the classical period Negev Steven ARosen

44

4

Farmers herders and miners in the Wadi Faynan southern Jordan a 10000-year landscape archaeology Graeme Barker

62

5

Differing strategies for water supply and farming in the Syrian Black Desert Paul Newson

85

6

Irrigation agriculture in Central Asia a long-term perspective from Turkmenistan Mark Nesbitt and Sarah OrsquoHara

101

Part III Sahara and Sahel

7

Conquests and land degradation in the eastern Maghreb during classical antiquity and the Middle Ages Jean-Louis Ballais

121

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

133

9 Twelve thousand years of human adaptation in Fezzan (Libyan Sahara) David Mattingly 156

10 Farming and famine subsistence strategies in Highland Ethiopia Ann Butler and ACatherine DrsquoAndrea 174

Part IV Eastern and southern Africa

11 Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 John EGSutton 195

12 The agricultural landscape of the Nyanga area of Zimbabwe Robert Soper 214

13

Fifteenth-century agropastoral responses to a disequilibrial ecosystem in southeastern Botswana John Kinahan

227

14

Islands of intensive agriculture in African drylands towards an explanatory framework Mats Widgren

246

Part V North and Central America

15

Prehistoric agriculture and anthropogenic ecology of the North American Southwest Paul EMinnis

264

16

The role of maguey in the Mesoamerican tierra friacutea ethnographic historic and archaeological perspectives Jeffrey RParsons and JAndrew Darling

280

Part VI Europe

17 Traditional irrigation systems in dryland Switzerland Anne Jones and Darren Crook 307

18

Desertification land degradation and land abandonment in the Rhocircne valley France Sander van der Leeuw

327

Index 346

Figures

11 The world map of drylands 312 A Roman-period fortified farm northwest Libya 413 The location of the case studies in this volume 514 Drowning in drylandsmdashtwo vehicles sunk in a flash-flood 721 Thermal regimes in two dryland locations Aswan Egypt and Jacobabad

Pakistan 2222 Mean monthly relative humidity at four locations 2423 The rainshadow effect leading to aridity 2524 The Hadley Cell circulation of the tropical northern hemisphere 2625 The structure of the trade wind atmosphere 2826 The interaction between the subtropical westerly flow and the tropical

easterlies leading to the creation of Saharan depressions 3027 The monthly progression of the East African Low-Level Jet Core 3228 The tracks of Sudano-Saharan depressions over the Sahara 3431 Terraced dam system in the central Negev 4532 The wine press at Shivta (Subeita) 4633 Sketch of the Byzantine town of Shivta (Esbeita or Subeita) 4734 Map of the general settlement system of the central Negev during the

Late Byzantine and Early Islamic periods 4835 View of the Byzantine town of Avdat (looking north) 4936 Elaborate raised field and dam system on Nahal Lavan 5037 The early Islamic village of Sede Boqer in the central Negev 5241 The location of Wadi Faynan within its region 6342 Looking northeast across part of the ancient field system to Khirbet

Faynan 6443 The survey area of the Wadi Faynan Landscape Survey 6544 Ethnoarchaeological survey the typical site of a winter bedouin tent in

Wadi Faynan 6745 The settlement locations of the first farmers in the Wadi Faynan 7046 Part of the Wadi Faynan field system WF4 showing the early bronze age

and the classical landscapes 7247 A field system on the northern side of the Wadi Faynan 7348 A field map of part of the field system WF4 7649 The distribution of copper through sediments accumulated behind the

Khirbet Faynan barrage 77410 Section through a Roman-period water conduit channel 79

51 The Hauran and the Harra regions of Syria

8652 Plan of the Roman-period settlement of Ad-Diyatheh and its water

channels and field systems 8853 The Roman-period settlement of Ad-Diyatheh and its channel walls 8954 Plan of the main irrigated zone along the Wadi Sham by the Roman-

period settlement of al-Namara 9255 View of the main irrigated zone along the Wadi Sham at al-Namara 9356 Canal 3 at al-Namara viewed from the east 9357 The ancient reservoir at Qasr Burqursquo 9558 Air photograph of Qasr Burqursquo and its reservoir 9561 Turkmenistan showing locations mentioned in Chapter 6 10362 Bronze and iron age settlement in the Merv oasis Turkmenistan 10771 The eastern Maghreb showing locations mentioned in Chapter 7 12272 Flood deposits at Ksar Rhilane (Tunisia) 12473 The modern Roman aqueduct crossing Wadi Bou Jbib Carthage 12674 Holocene terrace of Wadi Cheacuteria-Mezeraa (Algeria) 12775 Holocene terraces in the Wadi el Akarit (Tunisia) 12881 Tripolitania northwest Libya showing the principal landforms and

settlements and the location of the UNESCO Libyan Valleys Survey 13482 RomanomdashLibyan floodwater farming in the Wadi Gobbeen 13483 Simplified distribution of early RomanomdashLibyan farms 13684 A RomanomdashLibyan fortified farm (gasr) and its satellite buildings at

Ghirza 14485 Model of RomanomdashLibyan agriculture 14686 Walls in the desert wall systems in the Wadi Mimoun 14791 Map showing the location of the Fezzan and the area of most detailed

survey around Germa 15792 The major climatic fluctuations of the Holocene in the Libyan Sahara 15993 The settlement of Germa (ancient Garama) the capital of the Garamantes 16194 Schematic cross-section of a foggara 16395 Model of the neolithic landscape around Germa 16496 Model of the evolved Garamantian landscape around Germa 16797 Model of the medieval landscape around Germa 16898 Model of hypothetical future direction of settlement and farming in

Fezzan 169101 Map of Ethiopia showing Adi Ainawalid in Tigrai province 175102 Residential compound near fields Adi Ainawalid 176103 Intercropped bread and durum wheats near Mai Kayeh Tigrai 179104 Harvesting grasspea by hand uprooting Adi Ainawalid 181105 First threshing of teff Adi Ainawalid 182106 Winnowing teff Adi Ainawalid 183

107 Grain storage jars Adi Ainawalid 184111 The Rift Valley and Crater Highlands of northern Tanzania 196112 Engaruka and the Rift Valley escarpment 197113 The Engaruka escarpment from the east 198114 Engaruka south fields 200115 Grid of feeder furrows and levelled field plots 200116 The support for the lsquogreat northernrsquo canal 201117 The embanked causeway of the lsquogreat northernrsquo canal 202118 An angular cairn 203119 Sonjo wooden house with thatch dome 212121 Location of the Nyanga area Zimbabwe 215122 Terraced hillsides in the Nyanga lowlands 216123 Vertical aerial photograph of cultivation ridges 219131 The regional setting of Letsibogo southeastern Botswana 229132 The distribution and linkage of Khami period sites at Letsibogo 233133 Plan and section of Letsibogo Site 125 234134 Distribution of soil nutrient values at Letsibogo Site 125 236141 Eastern and southern Africa showing sites mentioned in Chapter 14 248142 An intensively cultivated landscape at Kwermusl 251143 Preparing the field at Kwermusl 252144 Piles of manure from stalled cattle Mama Issara 252145 An irrigation channel above Tot in Marakwet Kenya 253146 An irrigation canal under repair above Chesoi Marakwet 254151 The North American Southwest 265152 Prehistoric Hohokam communities and irrigation systems in the Phoenix

basin of the Salt river 267153 Aerial photograph of prehistoric trincheras (checkdam) fields near Casas

Grandes Chihuahua Mexico 268154 Gridded gardens of fields near Safford Arizona 269155 A rock mulch field near Casas Grandes Chihuahua Mexico 270156 An excavated rock pile from the field shown in Figure 155 270161 Middle America showing the approximate extent of the tierra friacutea 281162 Field of cultivated magueys 282163 Castrating a mature maguey plant 286164 Spinning maguey fibre 290165 Pre-columbian spindle whorls used for spinning maguey fibre 291166 Use of modern iron scraper for extracting maguey fibre 293167 Examples of pre-columbian trapezoidal tabular basalt scrapers 294168 Experimental use of pre-columbian trapezoidal tabular basalt scraper 295169 A pre-columbian scraper plane 295

1610 Modern iron scraper and pre-columbian obsidian scrapers 296171 The Valais canton Switzerland showing places mentioned in Chapter 17 308172 Distribution of agricultural land in Vernamiegravege during the 1960s 309173 The Grand Bisse de Lens 313174 The irrigation sectors in Vernamiegravege 317175 Distribution of water during the first tour from the bisses of Vernamiegravege

5th May-8th June 1964 318176 A tessel used by members of the consortage of the Grand Bisse de Lens 320181 The middle and lower Rhocircne valley showing the progress of Roman

colonization 329182 Settlement trends in the middle and lower Rhocircne valley 50 BC-AD 600 331183 The persistence of settlements in the middle and lower Rhocircne valley

through different occupation periods 333184 GIS maps of the Haut Comtat 335185 The three levels of the investigation into modern-day urbanmdashrural

dynamics in southern France 338186 Relations between cities individual communes and their contexts 339187 Differences in context occurring among towns of similar andor different

sizes in the Haut Comtat 340

Tables

21 The regional distribution of world drylands 1922 Rainfall regimes at selected dryland stations 1923 Estimates of the land area of arid lands 2024 Mean relative humidity at various isobaric levels in the Sahara and the

Arabian peninsula 2325 Seven ITCZ zones 2926 The extent and severity of desertification 3561 Simplified chronological chart of prehistoric settlement in Turkmenistan 11262 Simplified chronological chart of settlement in the Merv oasis in the

historic period 11371 Morphoclimatic evolution in the eastern Maghreb during the later

Holocene 12281 Farm products of the Tripolitanian pre-desert first to fifteenth centuries

AD 141101 Crops cultivated at Adi Ainawalid Tigrai Ethiopia 178102 Crops no longer cultivated at Adi Ainawalid Tigrai Ethiopia 186131 Selected radiocarbon measurements from Letsibogo 231 132 Faunal taxa from Letsibogo Site 125 237161 The prehispanic chronology of central Mexico 283171 Approximate numbers of named irrigators using the Grand Bisse de Lens

lsquoaqueductis communirsquo in 1457 314172 Examples of tours with the number of droits and sequence of irrigation

hours 316173 Examples of bisse disputes 321181 Evolution of the lsquonature-culturersquo debate over the last thirty years 341182 The different approaches of the historical and natural sciences to the

reconstruction of the past 342183 The opposition between analytical and integrative approaches in research 342

Contributors

JEAN-LOUIS BALLAIS is Professor of Physical Geography at the Universiteacute de Provence Aix-en-Provence (France) His principal research interests focus onHolocene Mediterranean erosion (he co-directed the study of erosion and geosystemshistory in classical antiquity and the Middle Ages for the EU-funded Archaeomedesproject described in Chapter 18) and present-day erosion desertification and landdegradation in the south of France and in the Maghreb Relevant recent publicationsinclude lsquoAeolian activity desertification and the ldquoGreen Damrdquo in the Ziban range Algeriarsquo in ACMillington and KPye (eds) Environmental Change in DrylandsBiogeographical and Geomorphological Perspectives 177ndash98 Chichester John Wiley and Sons 1994 lsquoThe south of France and Corsicarsquo in AJConacher and MSala (eds) Land Degradation in Mediterranean Environments of the World 29ndash39 Chichester John Wiley and Sons 1998 and (with J-CMeffre) Le Plan de Dieu (Nord-Vaucluse) Geacuteoarcheacuteologie et Histoire drsquoun Paysage Anthropiseacute Etudes Vauclusiennes 15 1996 (Institutional address Institut de Geographic lsquoUniversiteacute de Provence (Aix-Marseiile I) 29 Avenue Robert Schuman 13621 Aix-en-Provence France)

GRAEME BARKER (BA PhD University of Cambridge) taught prehistoric archaeologyat the University of Sheffield (1972ndash84) and was then Director of the British School at Rome (1984ndash88) before taking up his appointment as Professor of Archaeology at theUniversity of Leicester (UK) where he is currently Dean of the University GraduateSchool His principal research interests have been in the archaeology of subsistenceand agriculture with a special focus first on archaeozoology but later in landscapearchaeology He has conducted fieldwork in Italy Mozambique and the formerYugoslavia and has directed inter-disciplinary field projects in Italy Libya andcurrently in Jordan His publications include Landscape and Society PrehistoricCentral Italy London Academic Press 1981 (with RHodges) 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 CSGamble) Beyond Domestication in Prehistoric Europe London Academic Press 1985 (with JLloyd) Roman Landscapes Archaeological Survey in the Mediterranean Region London British School at Rome 1991 A Mediterranean Valley Landscape Archaeology andAnnales History in the Biferno Valley London Leicester University Press 1985 (twovolumes) (with DGilbertson BJones and D Mattingly) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume One Synthesis VolumeTwo Gazetteer and Pottery Paris UNESCO 1996 (with TRasmussen) The Etruscans Oxford Blackwells 1998 The Companion Encyclopedia of ArchaeologyLondon Routledge 1999 and (General editor with DMattingly) Mediterranean Landscape Archaeology Oxford Oxbow 2000 (five volumes) He was elected a

Fellow of the British Academy in 1999 (Institutional address the Graduate SchoolUniversity of Leicester Leicester LE1 7RH UK)

ANN BUTLER is an Honorary Lecturer at the Institute of Archaeology UniversityCollege London with a BSc degree in Botany (London University) an MA inArchaeology (Manchester University) and a PhD in Archaeology (London University)Her research interests centre on legumes as a human resource in the temperate OldWorld and include ancient diet and nutrition plant domestication and crop dispersalstraditional agriculture and sustainable farming Her fieldwork has been conducted inEurope Southwest Asia and Highland Ethiopia Her current research focuses on theevidence for the domestication and exploitation of legume crops Her recentpublications include lsquoPulse agronomy traditional systems and implications for early cultivationrsquo in PCAnderson (ed) Preacutehistoire de lrsquoAgriculture 67ndash78 Paris CNRS 1998 and lsquoTraditional seed cropping systems in the temperate Old World models forantiquityrsquo in CGosden and JHather (eds) The Prehistory of Food 473ndash77 London Routledge 1999 (Institutional address Institute of Archaeology University CollegeLondon 31ndash34 Gordon Square London WC1H OPY UK)

A CATHERINE DrsquoANDREA is an Associate Professor of Archaeology at Simon Fraser University British Columbia Canada She completed a BSc in Anthropology at theUniversity of Toronto an MSc in Bioarchaeology at University College London and aPhD in Anthropology at the University of Toronto Her research interests includepalaeoethnobotany ethnoarchaeology and early agrarian societies in Africa and the FarEast She is currently conducting ethnoarchaeological and palaeoethnobotanicalresearch in northern Ethiopia as well as collaborating on an excavation in northernGhana Her recent publications include lsquoThe dispersal of domesticated plants into northeastern Japanrsquo in CGosden and JHather (eds) The Prehistory of Food 163ndash83 London Routledge 1999 and (with DELyons Mitiku Haile and EA Butler)lsquoEthnoarchaeological approaches to the study of prehistoric agriculture in the Ethiopian Highlandsrsquo in Mvan der Veen (ed) The Exploitation of Plant Resources inAncient Africa 101ndash22 New York Plenum Publishing Corporation 1999 (Institutional address Department of Archaeology Simon Fraser University BurnabyBritish Columbia Canada V5A 1S6)

DARREN CROOK is currently a Postdoctoral Fellow in the Department of GeographyUniversity of Liverpool working on historical impacts of land use and climate onhydrology in a pre-alpine landscape funded by the Leverhulme Foundation Hegraduated with a BSc in Human Ecology from the University of Huddersfield HisPhD also from the University of Huddersfield dealt with the sustainability of the bissemountain irrigation system in the Valais Switzerland Publications from this include(with AM Jones) lsquoTraditional irrigation and its importance to the tourist landscape of Valais Switzerlandrsquo Landscape Research 24 (19) 199949ndash65 and (with AM Jones) lsquoTraditional water management in a developed world context an example from the Valais Switzerlandrsquo Mountain Research and Development 19 (2) 199979ndash99 (Institutional address Department of Geography University of Liverpool RoxbyBuilding Liverpool L69 3BX UK)

JANDREW DARLING completed his PhD in Anthropology at the University ofMichigan 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 (1989ndash90) in southeastern Hungary (1987) and the NorthAmerican Southwest (1984ndash1987) His research interests include compositionalstudies exchange regional interaction and ritual in prehistoric and ethnographiccomplex societies including parallel archival investigations on the development ofAmerican archaeology in Mexico during the early twentieth century Significantpublications include lsquoAnasazi mass inhumation and the execution of witches in theAmerican Southwestrsquo American Anthropologist 100 19931ndash21 (with MGlascock) lsquoAcquisition and distribution of obsidian in the north-central frontier of Mesoamericarsquo in ECRattray (ed) Rutas de Intercambio en Mesoamerica III Coloquio Pedro Posch Gimpera 345ndash64 Mexico DF Universidad Nacional Autonoma de Mexico 1998lsquoTrace element analysis of the Huitzila and La Lobera obsidian sources in the southern Sierra Madre Occidential Mexicorsquo Journal of Radioanalytical and Nuclear ChemistryArticles 196 (2) 1995243ndash52 and lsquoNotes on obsidian sources of the southern Sierra Madre Occidentalrsquo Ancient Mesoamerica 4 1993245ndash53 (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 Universityof Bournemouth (UK) and Distinguished Visiting Scholar in the Department ofGeography and Environmental Studies at the University of Adelaide (Australia) Hegraduated in Environmental Sciences at the University of Lancaster and gained hisPhD and DSc in Quaternary and Archaeological Geology from the University ofBristol Previous appointments have included a Senior Fulbright Scholar at theUniversity of Arizona the Directorship of the MSc Programme in EnvironmentalArchaeology and then Head of the Research School in Archaeology andArchaeological Science at the University of Sheffield where he became Reader thenProfessor Director of the Institute of Geography and Earth Science at the University ofWales Aberystwyth and Professorial Research Fellow University CollegeNorthampton In addition to dryland environments past and present his researchinterests include coastal geomorphology environmental change and caves andenvironmental archaeology in general His principal publications include (withRDSJenkinson) In the Shadow of Extinction The Quaternary Geology andPalaeoecology of the Lake Fissures and Smaller Caves at Creswell Crags Sheffield University of Sheffield Monographs in Prehistory 1984 (with DJBriggs andGRCoope) 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 (withGBarker BJones and DMattingly) Farming the Desert The UNESCO LibyanValleys Archaeological Survey Volume One Synthesis Volume Two Gazetteer andPottery Paris UNESCO 1996 and (with MKent and JPGrattan) The Outer Hebrides The Last 14000 Years Sheffield Sheffield Academic Press 1996(Institutional addresses School of Conservation Sciences University of BournemouthBournemouth BH12 5BB UK and Department of Geography and EnvironmentalStudies University of Adelaide South Australia 5005)

GAVIN GILLMORE is Senior Lecturer in Earth Science at University CollegeNorthampton His research interests include the application of fossil studies topalaeoenvironmental analysis ecotoxicology of cave environments and microfossiland stratigraphic studies of Quaternary sedimentary basins He has worked extensivelyfor oil exploration companies producing many consultancy reports on JurassicCretaceous and Tertiary-Quaternary microfossil assemblages Some current papers include (with MSperrin PPhillips and ADenman) lsquoRadon hazards geology and exposure of cave users a case study and some theoretical perspectivesrsquo Ecotoxicology and Environmental Safety 2000 (with KASmith and SSinclair) lsquoPalaeoenvironmental and biostratigraphical significance of Ostracoda from the Milton Formation (Quaternary) Northamptonshire UKrsquo Proceedings of the Geologists Association 2000 and (with TKjennerud and RKyrkjeboslash) lsquoThe reconstruction and analysis of palaeowater depths a new approach and test of micropalaeontologicalapproaches in the post-rift (Cretaceous to Quaternary) interval of the Northern NorthSearsquo in OJMartinsen and TDreyer (eds) Sedimentary Environments OffshoreNorwaymdashPalaeozoic to Recent Oslo Norwegian Petroleum Society Special Publication 2000 (Institutional address School of Environmental Science UniversityCollege Northampton Northampton NN2 7AL UK)

CHRIS HUNT is currently a Senior Lecturer in the Division of Geographical Sciences atthe University of Huddersfield His academic education led him from a degree inGeogaphyGeology and an MSc in Micropalaeontology (Palynology) at the Universityof Sheffield to a PhD at Aberystwyth (Wales) on the Pleistocene history of an area inSomerset He has published extensively in the area of environmental archaeology inBritain Europe North Africa and the Middle East Recent publications include (withGBarker DDGilbertson and DMattingly) lsquoRomano-Libyan agriculture integrated modelsrsquo in GBarker DGilbertson BJones and DMattingly Farming the Desert The UNESCO Libyan Valleys Survey Volume One Synthesis 265ndash90 Paris UNESCO 1996 (with SCampbell JScourse and DHKeen) The Quaternary of South West England Chichester Chapman amp Hall Geological Conservation Review Series14 1998 and (with DDGilbertson) lsquoContext and impacts of ancient catchment management in Mediterranean countries implications for sustainable resource usersquo in DWheater and CKirby (eds) Hydrology in a Changing Environment Volume II 473ndash83 Chichester John Wiley and Sons 1998 (Institutional address Division ofGeographical Sciences University of Huddersfield Queensgate Huddersfield HD13DH UK)

ANNE JONES is currently a Senior Lecturer in the Division of Geographical Sciences atthe University of Huddersfield and Head of Division She graduated with a BSc inGeography from Queen Mary College (University of London) and then obtained herDPhil thesis at the University of Oxford studying immigrant communities in MarseilleFrance Subsequently she has held posts at the Open University Liverpool Universityand Cambridgeshire College of Arts and Technology (now Anglia PolytechnicUniversity) Her research interests focus around the inter-relationship between demography and the allocation of scarce resources in marginal environments in alpineEurope the Mediterranean and Africa Her principal publications include lsquoExploiting a marginal European environment population control and resource management under

the Ancien Regimersquo Journal of Family History 16 (4) 1991363ndash79 (with COHunt) lsquoWalls wells and water supply aspects of the cultural landscape of Gozo Maltese Islandsrsquo Landscape Issues 11 (1) 199424ndash9 and (with COHunt and DSCrook) lsquoTraditional irrigation strategies and their implications for sustainable livelihoods in semi-arid areas examples from Switzerland and the Maltese islandsrsquo in HWheater and C Kirby (eds) Hydrology in a Changing Environment Volume II 485ndash94 Chichester John Wiley and Sons 1998 (Institutional address Division ofGeographical Sciences University of Huddersfield Queensgate Huddersfield HD13DH UK)

JOHN KINAHAN (PhD 1989 Witwatersrand) is an independent consultant inarchaeological and palaeoenvironmental studies and has authored more than fortyscientific papers in diverse fields Recent publications include Pastoral Nomads of the Central Namib Desert The People History Forgot Windhoek New Namibia Books 1991 lsquoThe rise and fall of nomadic pastoralism in the central Namib desertrsquo in TShaw PSinclair BAndah and AOkpoko (eds) The Archaeology of Africa Food Metals and Towns 372ndash85 Routledge One World Archaeology 20 1993 and lsquoA new archaeological perspective on nomadic pastoralist expansion in south-western Africarsquo in JEGSutton (ed) The Growth of Farming Communities in Africa from the Equator Southwards 211ndash26 Nairobi British Institute in Eastern Africa 1996 He is currently attached to the University of Uppsala in Sweden while carrying out research on thelong-term environmental impacts of nomadic pastoralism in Namibia and Tanzania (Institutional address Quaternary Research Services PO Box 22407 WindhoekNamibia and Department of Archaeology and Ancient History St Erikstorg 5University of Uppsala Uppsala S75310 Sweden)

SANDER VAN DER LEEUW presently Professor in the History and Archaeology ofTechniques at the Sorbonne in Paris followed a university education in History andArchaeology at the University of Amsterdam (PhD 1976) and has taught at theUniversities of Leiden and Amsterdam in the Netherlands Reading and Cambridge inthe UK Michigan (Ann Arbor) Massachussetts (Amherst) and the Santa Fe Institute inthe USA and the Australian National University He has undertaken fieldwork inSyria Holland the Philippines and France His main research interests embrace thetechnology of ancient pottery making regional and spatial archaeology the relationsbetween people and their environment through time as well as in the present and theuse of dynamical systems modelling for understanding social systems Among hispublications are Studies in the Technology of Ancient Pottery Amsterdam University of Amsterdam Institute for Pre- and Protohistory 1976 (with ACPritchard) The Many Dimensions of Pottery Amsterdam University of Amsterdam Institute for Pre-and Protohistory 1982 (with RTorrence) Whatrsquos New A Closer Look at the Process of Innovation London Unwin Hyman One World Archaeology 14 1987 (withJMcGlade) Time Process and Structured Transformation in Archaeology London Routledge 1997 and The ARCHAEOMEDES Project Understanding the Natural andAnthropogenic Causes of Land Degradation and Desertification in the MediterraneanLuxemburg Publications Office of the European Union 1998 (Institutional addressBoit 05 Maison de lrsquoArcheacuteologie et de lrsquoEtnologie 21 Alice de lrsquoUniversiteacute 92023 Nanterre France)

DAVID MATTINGLY is Professor of Roman Archaeology in the School ofArchaeological Studies at the University of Leicester and has conducted fieldwork inNorth Africa the Near East the Mediterranean and Britain He took his BA and PhD atthe University of Manchester the latter a study of Roman Libya Following the tenureof a British Academy Post-Doctoral Fellowship at the University of Oxfordresearching 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 impacton people and environment and olive oil production and trade in the ancient world Inaddition 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 publicationsinclude (with JALloyd) Libya Research in Archaeology Environment History andSociety 1969ndash1989 London Society for Libyan Studies 1989 (with GDBJones) An Atlas of Roman Britain Oxford Blackwell 1993 Tripolitania London Batsford 1995 (with GBarker DGilbertson and BJones) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume One Synthesis Volume Two Gazetteer andPottery Paris UNESCO 1996 Dialogues in Roman Imperialism Power Discourseand Discrepant Experience in the Roman Empire Portsmouth RI 1997 and (with MGillings and Jvan Dalen) Mediterranean Landscape Archaeology 3 Geographical Information Systems and Landscape Archaeology Oxford Oxbow 2000 (Institutional address School of Archaeological Studies University of Leicester Leicester LE17RH UK)

PAUL EMINNIS is an Associate Professor at the University of Oklahoma His researchinterests include paleoethnobotany human ecology social evolution human responsesto food shortages the relationships between archaeology and biodiversity and theprehistory of the North American Southwest For the past decade he has co-directed an archaeological project in northwestern Chihuahua Mexico to understand the regionalsetting of Casas Grandes one of the most complex prehistoric polities in NorthAmerica His books include Social Adaptation to Food Stress Chicago University of Chicago Press 1985 (with CRedman) Perspectives on Southwestern PrehistoryBoulder CO Westview Press 1990 (with WElisens) Biodiversity and Native AmericaNorman University of Oklahoma Press 2000 and Ethnobotany A Reader Norman University of Oklahoma Press 2000 (Institutional address Department ofAnthropology University of Oklahoma Norman OK 73019 USA)

MARK NESBITT is an ethnobotanist at the Centre for Economic Botany Royal BotanicGardens Kew An undergraduate degree in agricultural botany at Reading Universitywas followed by postgraduate training in archaeobotany at the Institute ofArchaeology University College London Since 1985 he has been involved in a widerange of archaeological fieldwork in the Near and Middle East including Turkey IraqBahrain and Turkmenistan His research interests include the origins development andsustainability of crop husbandry in arid lands the evolution of Old World cereals andarchaeological and ethnographic evidence for wild plant foods in the temperate zonesPublications include lsquoArchaeobotanical evidence for early Dilmun diet at Saar Bahrainrsquo Arabian Archaeology and Epigraphy 4 199320ndash47 lsquoPlants and people in

ancient Anatoliarsquo Biblical Archaeologist 58 199568ndash81 (with DSamuel) lsquoFrom staple crop to extinction The archaeology and history of the hulled wheatsrsquo in SPadulosi KHammer and JHeller (eds) Hulled Wheats 41ndash100 Rome IPGRI 1997 (Institutional address Centre for Economic Botany Royal Botanic Gardens KewRichmond Surrey TW9 3AE UK)

PAUL NEWSON is a PhD student at the University of Leicester After a first career as agraphic designer he took his BA and MA in archaeology at University CollegeLondon Funded by the Natural Environment Research Council he is preparing hisPhD thesis at Leicester on water management strategies in Roman Arabia and theirimplications for understanding processes of Romanization combining field studies ofdryland 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 southernJordan using Geographical Information Systems In 1999 he was Acting AssistantDirector of the Council for British Research in the Levantrsquos British Institute at Amman for Archaeology and History (Institutional address School of Archaeological StudiesUniversity of Leicester Leicester LE1 7RH UK)

SARAH OrsquoHARA is a Reader in Environment and Society in the School of GeographyUniversity of Nottingham She completed a BSc in Physical Geography and Geologyat the University of Liverpool an MSc in Geography at the University of AlbertaCanada and a DPhil in Geography at the University of Oxford Her research interestsinclude environmental reconstruction humanenvironmental interactions and waterresource management in the worldrsquos drylands Her research has been carried out in Albania Canada Iran Mexico Turkmenistan and Uzbekistan More recently she hasbegun collaborating with Professor Julian Henderson on the Ancient Raqqa IndustrialProject in Syria Recent publications include (with DThomas MD Bateman andDMershahi) lsquoDevelopment age and environmental significance of a Late Quaternarysand ramp central Iranrsquo Quaternary Research 48 1997155ndash61 (with THannan) lsquoIrrigation and water management in Turkmenistan past systems present problems and future scenariosrsquo Europe-Asia Studies 51 199921ndash41 and lsquoLearning from the past water management in Central Asiarsquo Water Policy (forthcoming 2000) (Institutional address School of Geography University of Nottingham NottinghamNG7 2RD UK)

JEFFREY RPARSONS (PhD in Anthropology University of Michigan 1966) iscurrently Professor of Anthropology and Curator of Latin American Archaeology atthe University of Michigan Ann Arbor Michigan USA In addition to ongoingfieldwork in the Valley of Mexico since 1961 he has also worked in Guatemala (TikalPeten 1966) Peru (Chilca central coast 1969ndash1970 and Junin central highlands 1975ndash76) 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 onthe pre-hispanic utilization of lacustrine resources in the Valley of Mexico and pre-hispanic regional organization in the Peruvian central highlands Significantpublications include (with CMHastings and Ramiro Matos M) lsquoRebuilding the state in highland Peru herder-cultivator interaction during the Late Intermediate Period in

the Tarama-Chinchaycocha regionrsquo Latin American Antiquity 8 1997317ndash41 (with GMastache RSantley and MC Serra) Arqueologiacutea Mesoamericana Homenaje a William TSanders Mexico DF Institute Nacional de Antropologiacutea e Histoacuteria 1996 lsquoPolitical implications of pre-hispanic chinampa agriculture in the Valley of Mexicorsquo in HHarvey (ed) Land and Politics in the Valley of Mexico A Two Thousand-Year Perspective 17ndash42 Albuquerque University of New Mexico Press 1991 and (withMH Parsons) Maguey Utilization in Highland Central Mexico An ArchaeologicalEthnography Ann Arbor University of Michigan Museum of AnthropologyAnthropological Paper No 82 1990 (Institutional address Department ofAnthropology University of Michigan Ann Arbor Michigan MI 48104 USA)

STEVE AROSEN teaches archaeology in the Department of Bible and Ancient NearEast at Ben-Gurion University in Beer-Sheva Israel He received his Bachelorrsquos degree from the University of California at Berkeley in mathematics and anthropologyand his graduate degrees in anthropology from the University of Chicago Prior to hiscurrent position he worked for eight years as a survey archaeologist for theArchaeology Survey of Israel in the Negev His research interests include desertadaptations the archaeology of pastoral nomadism Levantine prehistory and lithicanalysis His major publications include Lithics After the Stone Age Walnut Creek Altamira Press 1997 (with GAvni) The lsquoOded Sites Investigations of Two EarlyIslamic Pastoral Camps South of the Ramon Crater Beer-Sheva Ben-Gurion University Press 1997 and Archaeological Survey of Israel Map of Bersquoerot OdedBeer-Sheva Ben-Gurion University Press 1994 (Institutional address ArchaeologyDivision Ben-Gurion University PO Box 3653 Beer-Sheva 84105 Israel)

ROBERT SOPER (MA Cambridge UK) is a Senior Research Fellow in Archaeology atthe University of Zimbabwe Between 1962 and 1985 he worked for the NigerianAntiquities Department the British Institute in Eastern Africa Ibadan University andthe University of Nairobi His principal research interests have included the laterprehistory of East Africa (especially the Early Iron Age and later ceramics) the site ofOyo Ile in Nigeria and Great Zimbabwe tradition sites in northern ZimbabweSignificant publications include lsquoA general review of the Early Iron Age in thesouthern half of Africarsquo Azania 6 19725ndash37 lsquoRoulette decoration on African potteryrsquo African Archaeological Review 3 198529ndash51 lsquoThe palace at Oyo Ile western Nigeriarsquo West African Journal of Archaeology 22 1993 and (with BEKipkorir and JWSsennyonga) The Kerio Valley Past Present and FutureNairobi University of Nairobi Institute of African Studies 1983 He has conductedresearch on the Nyanga terrace complex since 1993 and a monograph on this isforthcoming (Institutional address History Department University of Zimbabwe POBox MP 167 Mount Pleasant Harare Zimbabwe)

GREG SPELLMAN is a lecturer in the School of Environmental Science at UniversityCollege Northampton (UK) After a BA in Geography at the University of Sheffieldhe took a PGDip in Applied Meteorology and Climatology at the University ofBirmingham and an MA in Professional Education at the University of Leicester Hisresearch interests include synoptic climatology extreme hydrological events in theIberian peninsula and the meteorology of air pollution He is currently researching onthe synoptic climatology of Spain particularly the evaluation of various downscaling

methods in order to improve regional climate change scenarios Recent publicationsinclude lsquoAn application of artificial neural networks to the prediction of surface ozone concentrations in the United Kingdomrsquo Applied Geography 19 1999123ndash36 and lsquoInvestigating the synoptic climatology of precipitation in Mallorca Spainrsquo Journal of Meteorology 23 1998117ndash30 (Institutional address School of Environmental Science University College Northampton Northampton NN2 7AL UK)

JOHN EGSUTTON (MA Oxford PhD East Africa FSA) was Director of the BritishInstitute in Eastern Africa from 1983 to 1998 and was previously Professor ofArchaeology at the University of Ghana He has been mainly concerned with laterarchaeology and its contribution to African history with a special interest in fieldsystems and agricultural technology His first visit to Engaruka in the northernTanzanian Rift Valleymdashthe subject of his contribution to this volumemdashwas in 1963 while a research scholar of the British Institute Later as a lecturer at the University ofDar es Salaam he continued the investigation of the Engaruka irrigation-agricultural settlement with student assistants That study has been extended in more recent yearswithin a broader comparative project on African field systems and cultivationstrategies 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 (2930 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 HumanGeography He received his PhD in Stockholm in 1983 and has researched on thehistorical geography of agricultural landscapes from the Iron Age to the present in Sweden and in southern and eastern Africa Among his publications are lsquoIs landscape history possiblersquo in PUcko and RLayton (eds) The Archaeology and Anthropology of Landscape 94ndash103 London Routledge 1999 lsquoFields and field systems in Scandinavia during the Middle Agesrsquo in GAstill and JLangdon (eds) Medieval Farming and TechnologymdashThe Impact of Agricultural Change in Northwest Europe173ndash92 Leiden Brill 1997 lsquoStrip fields in an iron age context a case study from Vaumlstergoumltland Swedenrsquo Landscape History 12 19905ndash24 and Settlement and Farming Systems in the Early Iron Age A Study of Fossil Agrarian Landscapes inOumlstergoumltland Sweden Stockholm Almquist amp Wiksell 1983 His most importantworks in Swedish are his contribution to the first volume of the agrarian history ofSweden Jordbrukets foumlrsta femtusen aringr (1998) and a book on medieval field systems inBohuslaumln Sweden (1995) (Institutional address Department of Human GeographyStockholm University S-106 91 Stockholm Sweden)

Foreword

One World Archaeology is dedicated to exploring new themes theories and applicationsin archaeology from around the world The series of edited volumes began withcontributions that were either part of the inaugural meeting of the World ArchaeologicalCongress in Southampton UK in 1986 or were commissioned specifically immediatelyfollowing the meetingmdashfrequently from participants who were inspired to make their own contributions Since then the World Archaeological Congress has held three furthermajor international congresses Barquisimeto Venezuela (1990) New Delhi India(1994) and Cape Town South Africa (1999) It has also held a series of more specialisedlsquointercongressesrsquo 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 manycountries

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 thatcharts 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 workin producing these books We should also like to express our thanks to Peter Ucko theinspiration behind both the World Archaeological Congress and the One WorldArchaeology series Without him none of this would have happened

Martin Hall Cape Town South AfricaPeter Stone Newcastle UK

Julian Thomas Manchester UK June 2000

Preface

This book stems from a symposium organized by the editors on the archaeology ofdrylands held at the World Archaeological Congress at Cape Town in January 1999 TheCongress provided the opportunity to bring together scholars working on the archaeologyof different regions of the worldrsquos drylands to pool experiences and in particular toinvestigate the extent to which we could discern common themes

Although over a third of the worldrsquos population today lives in arid and semi-arid lands there are many gaps in our understanding about how fragile or resilient these regions arefor human settlement To fill these gaps we need to answer questions that are likely to beof very great significance for the global community in the twenty-first century Many dryland regions have abundant remains of ancient settlement and people have oftenspeculated that the actions of farmers and herders in the past must have been important increating the degraded landscapes of the present For decades the debate has beencharacterized more by speculation than informed debate and by a propensity to argue forsimple processes of cause and effect in terms of climatic change or humanly inducedenvironmental degradation In the past fifteen years or so however inter-disciplinary archaeological and palaeoecological studies (especially when employed within integratedresearch frameworks) have demonstrated their potential to move the debate forwards byproviding detailed case studies of how ancient societies actually exploited drylandlandscapes how they interacted with them and the complex environmental and socialcontexts in which they variously succeeded or failed Moreover as we conclude inChapter 1 the advancement of understanding about past dryland societies andenvironments and of the complexity of their interactions with each other has a particularurgency given the way in which changing political agendas have been prone to eitherdemonize or sentimentalize them

Nine papers were given at the Cape Town symposium and a series of common themesabout dryland settlement emerged from the discussion Other papers were thencommissioned 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 whichare shown in Figure 13) different periods of the past and different kinds of societies butall addressing the issues we had identified at Cape Town to give a comparativeperspective 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 thearchaeology of drylands is eloquent testimony perhaps most of all to peoplersquos 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 Archaeologysymposium and for every assistance from them during the conference We would alsolike to thank the Society for Libyan Studies for a generous grant towards our two fares toCape 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 thecontributors to the volume for their patience especially those who contributed to thesymposium whose debates helped frame the discussion document we then circulated tothem and to the authors of the papers commissioned afterwards and who have remainedcommitted to our idea of the integrated comparative volume despite the timelag since theconference Finally we would like to dedicate this book to the memory of Barri JonesBarri was Professor of Archaeology at the University of Manchester and introduced usboth to dryland archaeology in the UNESCO Libyan Valleys Survey He died on the eveof his retirement in the summer of 1999 leaving a legacy of a major scholarly output ofbooks and papers an army of professional and amateur archaeologists enthused with hispassion for the subject and for his desert companions in particular memories of hair-raising adventures in his company He was a frenetic personality who was bothenchanting and exasperating to work withmdashhe was notorious for doing too many things at once mostly while nominally in control of a Landrover Amongst his many talentsthough he had an extraordinary topographical eye he was liable to get us lost in somedesert waste to visit an archaeological site once noted by a traveller 100 years ago andthrash the vehicle in the process but he was also by far the best person to be with to getsafely 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 60million km2 of which about ten million km2 are hyper-arid deserts (Fig 11) Drylands support over one fifth of the worldrsquos population and arid and semi-arid lands together over a third Living conditions vary from the most affluent and profligate to thedesperately poormdashin some cases in close proximity The political stability and ecological economic and social sustainability of dryland settlement are among the mostdaunting challenges confronting the global community in the twenty-first century water seems likely to be a primary flashpoint for disputes between neighbouring states withdryland irrigation systems under strain from fast-growing populations

Figure 11 The world map of drylands Source After UNEP 1992

and with environmental refugees from global warming predicted to be in the order of 150million by the year 2050 under the business-as-usual scenario (Houghton 1997) thecatastrophic consequences will be particularly acute for dryland populations

Many dryland regions have archaeological remains suggesting that once upon a time

there must have been intensive phases of settlement in what are now dry and degradedenvironments (Fig 12) People have often speculated about what must have happened to turn past glories into present-day barrenness generally dividing in favour of climaticchange or human agency as the primary culprit Perhaps the climate shifted to greateraridity Or was it that people sowed the seeds of their own destruction through their follyfor example by developing irrigation systems that caused salinization or by stripping thelandscape 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 ecologicaltheory suggests that relations between dryland environments climate and people are byno means simple (Beaumont 1993) drylands can sometimes be remarkably resilient forexample recovering relatively quickly from over-exploitation and simple procedures byfarmers 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 thearchaeological literature about how climate and people may or may not have affecteddrylands in the past

Figure 12 A Roman-period fortified farm on the desert margins of Tripolitania northwest Libya

Photograph GBarker

Modern inter-disciplinary archaeology especially when working in conjunction withother social and environment sciences has the potential to move the debate forwardArchaeology deals with the entire human past its geographical scope is regionallyspecific but worldwide its scale of enquiry ranges from distributions and processes ofchange at the global scale and over millennia down to the actions of individuals We canuse the techniques of landscape archaeology to understand how different kinds ofsocieties whether recent or remote in time exploited the different dryland environments

The archaeology of drylands 4

of the world We can characterize the risks and opportunities confronting those societiesidentify the solutions they reached and often the reasons for them as well as monitoringthe 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 usestrategies and the reasons for their successes and failures archaeology can contributeeffectively to modern debates about desertification and the sustainability of drylandsettlement (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 scholarsworking on the archaeology of different regions of the worldrsquos 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 ofcommon themes about

Figure 13 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 wererewritten and further papers commissioned to address these common issues within acomparative perspective in this book with case studies drawn from different kinds ofdryland regions throughout the world (Fig 13) Common themes though as we discussbelow do not equate with similar solutions to dryland living or similar responses to risksand opportunities

Living at the margin themes in the archaeology of drylands 5

THEMES IN DRYLAND ARCHAEOLOGY

The term lsquodrylandsrsquo obviously fixes attention upon low precipitation Commonknowledge emphasizes that the climatic significance of this shortage depends upon theother aspects of the atmospheric environmentmdashthe radiation budget thermal regimewind regime the sources and pathways of moisture including fog as well as the manyother components of the biosphere and lithosphere that play significant parts in thehydrological cycle The meteorology and climatology that produce drylands are notsimple (Spellman Chapter 2) understanding them requires an appreciation of variability of precipitation and drought in both space and time Rainfall in many drylands istypically characterized as consisting of erratic short localized downpours of highintensity Low average precipitation totals are associated with notable variability Fierceand localized downpours creating sudden and dangerous floods are the primary resourcebase that many indigenous peoples have had to utilize to maintain themselves their cropsand their animals for millennia (as well as environmental hazards for archaeologistsworking in arid lands see Fig 14) However it is the prospect of prolonged and severe drought that dominates thinking about drylands Instrumental historical andpalaeoenvironmental records show that episodes of severe drought lasting decades ormore in length have not been uncommon in many drylands over the last few thousandyears (eg Bureau of Meteorology nd Fritts 1991 Lamb et al 1995 Nicholson 1994) Ingenuity flexibility and enterprise have been required from individuals communitiesand 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 climateoperating at the global scale in the remote past from major shifts in the worldrsquos oceanographic and atmospheric systems The period from approximately 18000 to10000 years ago the last phase of the Pleistocene (the lsquoIce Agersquo) saw the last major ice advances of glaciers and icesheets in the high latitudes Regions of the world that nowenjoy a temperate climate such as Europe and North America were cold and aridRegions 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 between10000 and 8000 years ago during


Figure 14 Drowning in drylandsmdashtwo vehicles of the UNESCO Libyan Valleys Survey sunk in a flash-flood

Photograph GBarker

the opening millennia of the Holocene or Postglacial the environment across the Saharawas notably wetter than occurs today with the development of lakes and woodland infavoured 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 andhunting not just steppe animals like gazelle but wetland species such as turtle hippo andcrocodile (Cremaschi and di Lernia 1998 Wendorf and Schild 1980) In the Near Eastwetter environments at this time were the context for the development of mixed farmingsystems 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 thantoday (Nesbitt and OrsquoHara Chapter 6)

Desiccation started to develop about 6000 years ago in North Africa and the NearEast with people responding differently In the Sahara people shifted to cattle andsheepgoat herding (Barich 1987 1998 Wendorf et al 1984 1989) whereas the farmers of Wadi Faynan started to experiment with methods of trapping and storingwater 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 smallfeeder 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 4500 years ago asseasonal streams replaced perennial streams salt pans replaced fertile lake floors and themodern 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


or part of the year the provision of an adequate and reliable supply of water in warm andsunny regions has been a goal for innumerable communities over time Rainwaterharvesting and floodwater farming are described for North America Africa and Asia in avariety of chapters Four major themes emerge from these case studies the first three ofwhich are closely inter-related

The first is the resilience of many farmers in antiquity to cope with harsh and risk-prone arid environments and their climatic vicissitudes over long periods of time Thislongevity of occupation points to an inherent robustness of many of these pastcommunities their attitudes and ways of life The second is the repeated evidence for thesimilarity of the lsquobuilding blocksrsquo or lsquotacticsrsquo employed by most ancient farmers indrylandsmdashbuilding 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 inthe overall systems that were put together from such building blocks and the way theyinvariably reflect detailed local knowledge of topography weather patterns and so onancient farmers knew from observation exactly how and where the water would flowafter 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 matterof commonsense observation by ancient farmers of what was the lsquobest fitrsquo to particular environmental or economic circumstances We can see today how dryland communitiesattempt to manage themselves and their habitats within the context of a whole nexus ofattitudes beliefs as well as economic social geographical educational agricultural andtechnological processes and whilst many of these details will elude archaeologists giventhe nature of our evidence the case studies illustrate how people took choices and notalways the right ones within a complex mix of factors including perceptions of risk theneed or desire for economic advantage and the institutional and regional context inwhich they were operating As Widgren comments (Chapter 14) models of agrarian development too often assume even developments of farming systems in response toparticular environmental social or economic pressures but the archaeological recordemphasizes above all the unevenness of development in both space and time We do notsee the kind of evolutionary development of land use systems so often assumed for thepast for example from simple to complex or from extensive to intensivemdashthey were lsquoformed and changed within specific place-bound social historical and ecological contextsrsquo (Widgren p 262)

One key influence on the character and scale of an irrigation system was notsurprisingly related to the extent to which the agricultural product was to serve only thelocal community Examples of such systems lsquoislandsrsquo of relatively intense landscape development are described for societies varying widely in time place and socialcomplexity for example in the North American Southwest in prehistoric times (MinnisChapter 15) the Achaemenid or Parthian periods in Turkmenistan (Chapter 6) the Libyan oases (Mattingly Chapter 9) and the Tricastin region of southern France (van derLeeuw 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


with labour-intensive agriculture and how both market orientation and subsistencefarming 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 economiescharacterized by the large-scale import and export of goods products and information tothe core areas of the Mediterranean but also to the somewhat similar relationship ofhighland Mexico to the rest of the Inca state (Parsons and Darling Chapter 16) lsquoPatchinessrsquo in distribution has also been identified in the studies by Jones and Crook (Chapter 17) of the Swiss bissesmdashcanal systems that have tapped and redistributed water within the surprisingly dry environment of the Canton of Valais for over a millenniumbut that remain a relatively unknown but vital component of the economy of one theworldrsquos richest and technologically sophisticated countries Clearly archaeological investigation of lsquomarginalrsquo 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 lsquothe past is a foreign countryrsquo things were thought and done differently there Then as now it seemsthat many individuals organizations and polities have behaved in relation to theirsituation and their environment in manners that do not appear rational to the modernexternal observer or to those with the wisdom of hindsight Then as now people madepoor decisions foolish decisions self-interested decisions carried out actions that they ortheir neighbours had cause to regret or more generally they misunderstood their landand 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 copperand 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 inthe Soviet period Chapter 6) as Minnis (Chapter 15) comments in the case of North America indigenous subsistence foragers and farmers equally have not always beenenvironmentally lsquocorrectrsquo lsquosoundrsquo or lsquoneutralrsquo

A good example of the importance of perception affecting decision making by drylandfarmers responding to adverse conditions occurred during the great drought that affectedthe 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 ofinstrumental records maintained by amongst others telegraph operators (Bureau ofMeteorology nd) The well-known account of this episode by Meinig (1962) basedupon parliamentary records and newspapers reveals that the human and economicimpacts of the first two yearsrsquo drought were devastating Many wheat farmers weresustained by their belief that the lsquorain followed the ploughrsquo more ploughing and tilling they thought would release further soil moisture into the atmosphere and so break thedrought Others farmers desperate to maintain overall production totals and to pay theirmortgages to the Government for their newly acquired lands (pastureland hitherto)ploughed and planted ever larger areas of bush believing that minimal returns from vastareas would compensate for the poor yields per acre Both strategies of course made


things far worsemdashas the drought grew worse and human distress grew ever moreprofound ever larger areas of land were being ploughed and farmed It is salutary forarchaeologists who rarely have access to such sophisticated documents or precisechronologies to reflect on the complete reversal in present thinking of the nature ofhumanmdashenvironment interactions that underpinned these farmersrsquo behaviour just over a century ago

The case studies demonstrate repeatedly that drought however pernicious and sustained is not necessarily the sole cause of lsquoabandonment phasesrsquo identified in the archaeological record of drylands Chapter 10 by Butler and DrsquoAndrea 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 asdrought or flood Discussing the Northern Highlands of Ethiopia a region almostsynonymous with drought and famine for most readers they emphasize the potency forunderstanding famine in the area of the following sometimes archaeologically invisiblefactors human smallpox cattle rinderpest plagues of predators such as locusts ants andarmy worms conflict and social and political circumstances In fact drought aloneseldom 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 againstrisk and adversity more generally their ability to organize themselves effectively indrought-prone habitats and in the last resort their willingness to relocate (Mortimore 1998122) We can see from the archaeological record that systems without suchflexibility did not have the necessary resilience for longterm survival as in the case of theproductive but short-lived systems of cash-crop farming that Romanized Libyansdeveloped 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 (BarkerChapter 4) or the massive centralist-administered irrigation systems of Soviet Turkmenistan (Nesbitt and OrsquoHara Chapter 6)

ARCHAEOLOGY AND DESERTIFICATION

During the last few decades many of the worldrsquos drylandsmdashthe hotter drylands especially but not exclusivelymdashhave been viewed as threatened by lsquodesertificationrsquo The term was coined by Aubreville (1949) in a report on the vegetation of Africa and itsmeaning has developed through time Thomas and Middleton (19949ndash10) defined it as lsquoland degradation in arid semiarid and dry sub-humid areas resulting mainly from adverse human impactrsquo and though some authors have also used the term to refer to landdegradation caused by a sustained aridification of climate most prefer to use the term torefer to the effects of human actions though climate and people can clearly work intandem to produce deterioration in dryland environments as may be the case in thecontext of global warming today (Barrow 1995 Millington and Pye 1994 SpellmanChapter 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


caused degradation may include decreased vegetation cover timber loss salinizationreduced water supplies lower crop yields outbreaks of disease accelerated erosion ofsoils and dust storms and induced regional climatic change all encapsulated in thepopular metaphor of the conversion of pastoral or arable lands within drylands intodesolate 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 (1993474) concludedhis book with a bleak prediction of the inevitability of this process for the worldrsquos poorest nations lsquoin certain cases land degradation may be a sacrifice which has to be paid inorder that local populations can survive future drought or faminersquo According to Tolba and el-Kholy (1992134) the current rate of desertification is about 60000 km2 per year amounting to 011 per cent per year of the total area of dryland On this calculationdesertification today threatens no less than 70 per cent of the worldrsquos drylands which represents over 25 per cent of the worldrsquos land surface Grainger (1990) and Spooner(1989) argued that such desertification can be recognized in Australia North Americaand South Africa in the first half of the twentieth century and that it was likely to havebeen 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 recentdecades is actually a global problem of such vast dimensions as opposed to localmanifestations of local problems of smaller significance There are also examples ofspirals of land degradation being reversed by indigenous technological adaptationsworking in combination with population growth and market opportunities The Machakosdistrict of Kenya was considered an environmental disaster in the 1930s because ofmassive soil erosion and famine but by 1990 terrace construction had protected arableland farmed and protected trees provided sufficient fuel-wood and agricultural production per person and per hectare had increased sustaining a population five timeslarger than that of the 1930s (Tiffen et al 1994) Deforestation and massive erosion on the Yatenga plateau in Burkino Faso were exacerbated by mechanization programmesfunded by ill-judged development aid but the reinstatement of traditional systems of terrace building stopped erosion and doubled crop yields (Lean 1994) Mortimore(1998149ndash56) described an examples of c 150 years of sustainable intensification bysmallholders in Kano Nigeria in the context of population growth and monetization

Archaeological remains in many drylands have been grist to the mill of thedesertification debate For example Hughes with Thirgood (198260 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 civilizationhellip 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


environment

Terms such as excessive land use population pressure loss of biological diversity andvegetation cover mis-use of water accelerated soil erosion and ever-larger human needsall recur in archaeological explanations (Gilbertson et al Chapter 8) often in associationwith reference to times of perceived political unrest military invasion conflict anddrought

In his review of the historical likelihood of humanly induced desertification Spooner(1989) cautioned against the simplistic tendency to assume that desertification faminedrought and poverty will inevitably be found together and several case studies in thisbook support those who argue for the complexity of desertification processes today Forexample 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 theedge of the northern Sahara seem to have neither produced nor experienced the kind ofself-induced environmental degradation described by Hughes with Thirgood (1982)Indeed the increase in human population numbers farming intensity and landmanagement probably promoted a greener more diverse and infinitely richer andproductive environment than has occurred since the great aridification in climate thatafflicted the region some 4500 years ago There are in fact good reasons to suspect thatRomano-Libyan farmers did have significant and deleterious impacts upon their aridenvironment but there are few reasons to believe that catastrophic long-term climaticchange short-term catastrophic drought or anthropogenically induced environmentaldegradation played a central role in the progressive abandonment of these settlementsmdashaprocess that is still underway after nearly 1500 years

Rosen (Chapter 3) also suggests that substantial cultural changes in the ancient Negevdesert are not best explained by climatic catastrophe invasion or the inabilities of peopleto manage their desert environment rather periods of cultural florescence were related toincreased economic and social input from or integration with the Mediterranean lsquocorearearsquo with desert pastoralism strongly geared to active markets in the settled zone andlikely to be sorely afflicted by the latterrsquos collapse Yet in the adjacent deserts of southernJordan the Wadi Faynan Landscape Survey (with many of the same members as theUNESCO Libyan Valleys Survey and using similar methodologies) has foundconvincing evidence for dramatic humanly-induced land degradation in the wake ofagricultural and industrial intensification in the context of Roman imperialism In theSaharan Fezzan Garamantian development of foggara irrigation systems may have beena key factor leading to the decline of their civilization as a result of over-extraction from anon-renewable groundwater source (Mattingly Chapter 9) Ballais (Chapter 7) argues thatincreases in soil erosion in the eastern Maghreb in classical antiquity reflected specificcombinations of climatic change and human activities and affected parts of the landscapein different ways In the Roman imperial centuries increased intensity of rains or theannual amount of precipitation badly affected regions already made vulnerable byvegetation degradation or unwise cultivation systems whereas the irrigated zones andterraced mountains were more resilient In the middle and lower Rhocircne valley in southernFrance episodes of climatic change are out of step with archaeologically visible episodesof human impact in the Tricastin region here accelerated erosion can be tied clearly with


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 ofapparently exhausted depleted and deserted drylands today The archaeological literatureon drylands is replete with references to the likelihood of lsquoover-grazingrsquo lsquoexcessive grazingrsquo and so on implying that there is today and that there was in times past someknowable and realizable intensity of pastoral activity in drylandsmdasha lsquocarrying capacityrsquo which if exceeded must have had dire consequences for the pastoralists and theirhabitats However the core of this viewmdashthat such a carrying capacity existsmdashis now being challenged by ecologists who consider that lsquoboom-or-bustrsquo models of animal population numbers may be more appropriate (Thomas and Middleton 1994) There maynever 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 intheir production of foragemdashthis variability itself a consequence of precipitation which is variable in space and time (Holling 1973 Noy-Meir 1978 Olsvig-Whittaker et al1993 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 lsquoadaptedrsquo to grazing stress and that pastoral disturbance could be regarded as a natural component of many arid environments Inbrief 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 byGilbertson (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 beperiods of environmental devastation wrought by nomadic pastoralists were probably infact characterized by less arable land a progressive development of lsquonaturalrsquo vegetation and pastures and so less soil erosion

Hollingrsquos (1973) ecological view that arid lands are lsquonon-equilibrium but persistentrsquo may have utility for many archaeologists working in drylands not least because it servesas a disincentive to extrapolate from the local diagnosis of an ancient cause and effect tothe inference of causality at regional or global scales The possibility of non-linear relationships within and between environmental processes and human activities must alsobe considered Relatively minor changes in the human or biophysical environment can inprinciple set in train self-sustaining sequences of events and processes that can cause theenvironment to transform from one state to another with cause and effect entangled Indrylands today relationships between individuals communities institutions and thelandscapes with which they interact are clearly neither simple nor linear in form (Ellis1995 Phillips 1993 Chapter 10) The principal argument of the case studies in thisvolume is that the same was certainly the case in the past even though the nature ofarchaeological and palaeoecological evidence is such that it may sometimes be difficultor impossible to identify the key players critical species or ideas the nature ofunderlying trends or pre-disposing factors the agencies of stability and the triggers of


change

CONCLUSION

The archaeology of drylands is one of the richest bodies of archaeological data andfrequently the most visible for archaeological surveyors though it is also often amongstthe most vulnerable to destruction by development programmes far from the eye ofnational archaeological services Moreover the deflated landscapes of most drylandsfrequently pose daunting challenges of archaeological analysis given the paucity of deepstratigraphies and organic remains though in some cases both the latter are present inrich abundance palaeoecologists often face similar challenges requiring similarpersistence and ingenuity in response However the practical (and often logistical)difficulties of dryland archaeology should not dissuade us from attempting to understandits significance whilst the details of structure and agency in past dryland settlement willoften be problematical to determine a better understanding of the complexity of peoplersquos 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 desertificationtheory 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 theirecosystems today (eg Barker et al 1996) Similar arguments apply to pastoraldevelopment programmes (Kinahan Chapter 13) We need a sophisticated understanding of the environmental and social contexts of ancient dryland farmers and herders detailedknowledge of modern dryland ecologies and sympathetic awareness of issues such as theownership 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 57ndash8) the advancement of understanding of past dryland societies and landscapes through the combination of goodarchaeological science and social archaeology is critical most of all to combat thepoliticization of much past theorizing on these matters Relations between the desert andthe sown underpin many origin myths of ethnicity and the lsquobiographyrsquo of arid lands has frequently been rewritten to changing political agendas The role of desert pastoralists inancient Palestine as told to us through the Old and New Testaments is an obvious case inpoint where it is repeatedly represented as a moral force for good in the history of theIsraelites the preferment of the shepherd Abel over his brother Cain the farmer thesubstitution of the ram for Abrahamrsquos son Isaac the commandment that lists the ox andthe ass before the wife the parable of the sheep and goats and the lost sheep and the roleof Christ himself as the lamb of God the Lord our Shepherd of Psalm 23 A pastoralideology with numerous parallels to the Biblical stories underpinned the origin myth ofthe Incas and their sense of their right to rule subject peoples (Brotherston 1989) Yet inboth the Near East and North Africa simplistic notions of Islamic pastoralist invaders asthe prime causes of environmental and cultural decline have stemmed primarily frompolitical agendas (Rosen Chapter 3 Ballais Chapter 7) and one of the planks of modern


Zionism has often been the contrast between the lsquogreening of the desertrsquo 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 drylandfarming systems of Native American peoples have been variously portrayed asenvironmentally destructive or in sympathy with the landscape according to changingcolonial and post-colonial perspectives (Minnis Chapter 15) At the root of the 1990s massacres in Rwanda was the lsquoTutsisrsquo and Hutusrsquo belief that they are derived respectivelyfrom Nilotic cattle-herders and Bantu farmersmdasha note left with a group of Europeantourists killed in Bwindi National Park in Uganda by Hutu guerillas in 1999 read inbroken French lsquohere is the fate of all the Anglo-Saxons who betray us to the Nilotics against the Bantu cultivatorsrsquo (Hannan 1999)mdashbut in fact there is very little to distinguish the two groups and the NiloticBantu dichotomy is almost certainly amistaken creation of nineteenth- and twentieth-century scholarship (Hall 1987 1996) Asdryland peoples face the uncertainties of the twenty-first century understanding the richness diversity and above all the complexity of the archaeology of their antecedentshas never been more urgent

REFERENCES

Aubreville A (1949) Climats Foragravets et Deacutesertification de lrsquoAfrique Tropicale Paris Societeacute drsquoEditions Geacuteographiques Maritimes et Coloniales

Barich BE (1987) Archaeology and Environment in the Libyan Sahara TheExcavations in the Tadrart Acacus 1978ndash1983 Oxford British Archaeological Reports International Series 368

Barich BE (1998) People Water and Grain The Beginnings of Domestication in the Sahara and the Nile Valley Rome lsquoLrsquoErmarsquo di Bretschneider

Barker G Gilbertson D Jones B and Mattingly D (1996) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume One Synthesis VolumeTwo Gazetteer and Pottery Paris UNESCO Publishing

Barrow CJ (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 JClutton Brock (ed)The Walking Larder 240ndash55 London Unwin Hyman

Bureau of Meteorology Commonwealth of Australia (nd) Results of Rainfall Observations Made in South Australia and Northern Territory 1839ndash1950 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 forpastoral development In IScoones (ed) Living with Uncertainty New Directions in


Pastoral Development in Africa 37ndash46 London International Institute for theEnvironment and Development

Fantechi R and Margaris NS (1986) (eds) Desertification in Europe Proceedings of the Information Symposium in the EEC Programme on Climatology Held in MytileneGreece 15ndash18 April 1984 Dordrecht DReidel Publishing Company

Fritts HC (1991) Reconstructing Large-scale Climatic Patterns from Tree-Ring Data Tucson University of Arizona Press

Gilbertson DD (1996) Explanations environment as agency In GBarker DGilbertson BJones and DMattingly Farming the Desert The UNESCO LibyanValleys Archaeological Survey Volume One Synthesis 291ndash317 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 CS (1973) Resilience and stability of ecological systems Annual Review of

Ecology and Systematics 41ndash23 Houghton J (1997) Global Warming The Complete Briefing Cambridge Cambridge

University Press second edition Hughes JD with Thirgood JV (1982) Deforestation erosion and forest management in

ancient Greece and Rome Journal of Forest History 2660ndash75 IFAD (International Fund for Agricultural Development) (1992) Soil and Water

Conservation in Sub-Saharan Africa Towards Sustainable Production by the RuralPoor Amsterdam Free University of Amsterdam Centre for Development Cooperation Services

Lamb HF Gasse F Benkaddour A el-Hamouti N van der Kaars S Perkins WT Pearce NJ and Roberts CN (1995) Relations between century-scale Holocene arid events in tropical and temperate zones Nature 373134ndash7

Lean G (1994) How stones can hold back the Sahara The Independent on Sunday 16 October 199416

Meinig DW (1962) On The Margins of the Good Earth The South Australian WheatFrontier 1869ndash1884 Adelaide Rigby Limited

Millington AC 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 SE (1994) Rainfall fluctuations in Africa and their relationship to pastconditions over the continent The Holocene 4(2)121ndash31

Noy-Meir I (1978) Grazing and production in seasonal pastures analysis of a simplemodel Journal of Applied Ecology 15809ndash35

Olsvig-Whittaker LS Hosten PE Marcus I and Schochat E (1993) Influence of grazing on sand field vegetation in the Negev desert Journal of Arid Environments


2481ndash93 Phillips JD (1993) Biophysical feedbacks and the risks of desertification Annals of the

Association of American Geographers 83(4)630ndash40 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 andDevelopment

Spooner B (1989) Desertification the historical significance In RHuss-Ashmore and SHKatz (eds) African Food Systems in Crisis Part One Micro-Perspectives 111ndash62 New York Gordon and Breach

Sutton JEG (1977) The African Aqualithic Antiquity 5125ndash34 Thomas DSG and Allison RJ (1993) (eds) Landscape Sensitivity Chichester John

Wiley and Sons Thomas DSG and Middleton NJ (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 MK and el-Kholy OA (1992) (eds) The World Environment 1972ndash1992

London Chapman Hall Walker BH Ludwig B Holling CS and Peterman RM (1981) Stability of semi-

arid savannah grazing systems Journal of Ecology 69473ndash98 Warren A and Khogali M (1992) Assessment of Desertification and Drought in the

Sudano-Sahelian Region 1985ndash1991 London International Institute for Environmentand Development

Wendorf F and Schild R (1980) Prehistory of the Eastern Sahara New York Academic Press

Wendorf F Schild R and Close AE (1984) (eds) Cattle-Keepers of the Eastern Sahara The Neolithic of Bir Kiseiba Dallas Southern Methodist University

Wendorf F Schild R and Close AE (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 ashortage 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 Earthrsquos land surface (UNEP 1992 see Fig 11) The absence of significant moisture is manifest in the characteristics of the soils vegetation and topographyConsequently Oliver (1973) and Nir (1974) have suggested ways of identifying aridlands by a variety of non-climatic criteria Straightforward classical approaches create regionalizations using isopleths of climatic elements with respect to associations withvegetation and agricultural conditions such as the 250 mm rainfall limit as the aridboundary (Oliver 1981) In contrast indexing methods delimit regions with differinglevels of aridity by the application of objective standard formulas

The best-known classical method is that of Koppen (1931) who defined drylandregions in terms of an annual precipitation and temperature index Assuming a meanannual temperature of 18degC 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 regioncovering 12 per cent and semi-desert and steppes the other 14 per cent The system wascriticized by Mather (1974) for failing to consider water supply and having no physicalmeaning or indication of the atmospheric processes involved Water-balance models were developed independently by Penman (1948) and Thornthwaite (1948) Penmanrsquos model is more sophisticated and considers turbulent transfer and energy balance approachesThornthwaitersquos 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[PPeminus1] In this system arid regions have an index value of under minus667 whereas a semi-arid region is defined where Im lies between minus333 and minus667 The method was criticized by Wallen (1967) fortending 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 11) divides the globe on degrees of bioclimatic aridity using the values ofthe ratio PPET that is P=the mean annual precipitation (mm) and PET=the mean annual potential evapotranspiration (mm) as calculated by Penmanrsquos formula Three categories are relevant here hyperaridity where the PETP ratio is less than 005 aridity from 005

to 020 and semi-aridity from 020 to 045 Some classifications of drylands also include the dry sub-humid regions (045ltPPETlt065) (Le Houerou 1996) The hyper-arid zone is characterized by true desert climates where precipitation is extremely low andirregular in occurrence Perennial vegetation is almost totally absent and neither pastoralnor arable farming is possible using rainfall The arid zone receives annual rainfall totalsof 80ndash350 mm with inter-annual rainfall of 50ndash100 per cent (Beaumont 1989) Scattered vegetation allows low-intensity grazing but rain-fed agriculture is unlikely The semi-arid region with precipitation totals of 200ndash700 mm is dominated by grassland and scrub providing relatively good grazing Table 21 gives the regional distribution for theclasses types of drylands and Table 22 gives some examples of each regime

These methods give precise definitions of the desert boundary but discrepancies occur estimates of the worldrsquos dryland areas based on climatic classifications range from 26 to 36 per cent the greatest difference being with respect to the location of hyper-arid areas (Table 23) A figure of 43 per cent was obtained by McGinnies (1988) using moistureshortage of moisture in the main criterion Yair and Berkowicz (1989) have suggestedthat aridity should be redefined to include the sensitivity of an area to low rainfall byassessing the intensity and duration of rainfall soil salinity and the ratio of soil cover tobare rock

The following dryland types can be identified with reference to rainfall and thermal regimes two rainy seasons (eg Venezuela) winter rainfall (eg the North Africancoast) summer rainfall (eg the Sahel) almost rainless (eg the Sahara) and fog andmist (eg the Atacama desert) Seasonality in rainfall reflects the dominant rain-producing mechanism summer rains are

Table 21 Regional distribution of world drylands (103km2) (After Le Houerou 1996)

Zone AfricaAsiaAustralasiaEurope North America

South America

Total

Hyper-arid 6720 2773 0 0 31 257 9781 75 Arid 5035 6257 3030 110 815 445 15692 121 Semi-arid 5138 6934 3090 1052 4194 2645 23053 177 Dry sub-humid

2687 352 513 1835 2315 2070 12947 99

Table 22 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 O N D Fiya Chad 225 18deg0primeN 18 0 0 0 0 0 0 0 18 0 0 0 0 19deg10primeE Khartoum 390 15deg37primeN 157 0 0 0 0 25 7 53 71 18 5 0 0 Sudan 32deg33primeE

The dynamic climatology of drylands 19

brought by the Inter Tropical Convergence Zone (ITCZ) so regions with summer rainfalllie on equatorward-margins of drylands whereas regions with winter rainfall under theinfluence of mid-latitude disturbances are on the poleward sides Dryland types based ontemperature classifications are tropical deserts exhibiting little change in monthlytemperatures (eg Somalia) subtropical deserts experiencing considerable temperaturechanges (eg the Thar and Australian deserts) temperate drylands with cold winters (egdrylands in Iran Syria and Mongolia) and cold highland areas (eg Tibet) Yet anotherscheme is that of Thomas (1989) hot arid lands (coldest month temperature 20ndash30degC) drylands with mild winters (coldest months 10ndash20degC) drylands with cool winters (coldest months 0ndash10degC) and drylands with cold winters (coldest month less than 0degC) Despite these varieties of definition however the overall purpose of drylandclassification is the same to identify their global significance to examine the processesthat operate to create them and to assess whether any major changes are occurringAgnew and Anderson (1992) remark that there is a grave danger that arid lands aretreated by water resource managers as homogeneous entities with similar environmentsand similar problems when this is clearly not the case

CLIMATIC CHARACTERISTICS

Precipitation

By definition all dryland areas receive low annual precipitation and in most drylandareas as rainfall amounts diminish there is a corresponding increase in variability andunreliability (Le Houerou 1996) Mean values therefore do not adequately describe thetrue nature of the precipitation regime because annual totals will show significant year-

Kashgar 1309 39deg24PrimeN 86 15 3 13 5 8 5 10 8 3 3 5 8 China 76deg07primeE Amman 777 31deg57primeN 278 69 74 31 15 5 0 0 0 0 5 33 46 Jordan 35deg57primeE Lima Peru 120 12deg05primeS 43 3 0 0 0 5 5 8 8 8 3 3 0 77deg03primeW

Table 23 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 UNESCOHyper-arid 120 153 205 195 Semi-arid 143 152 158 133 Total 263 305 363 328


to-year departures from long-term norms Nir (1974) for instance mentions a rain eventfrequency of once every eight years at certain sites in the Sahara and once every eighteenyears in Peru The interquartile or 10ndash90 percentile ranges are more useful indicators(Beaumont 1989) The variability of rainfall in arid areas is greater than that of temperateregimes 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 rainfallwill inevitably record the highest variation even though the magnitude of that variabilityaway from the mean is smaller The variability in absolute terms may not be much greaterthan that of temperate regions (Cooke and Warren 1973) but for areas with low rainfalleven small variations are extremely significant (Agnew and Anderson 1992)

When rainfall events do occur in dryland areas it is often when rainbearing frontalsystems or tropical cyclones penetrate the region Incursions are therefore more frequentat the margins of dryland areas The usual mechanism in poleward areas is the southwardmovement of cold lsquoupper lowsrsquomdashareas of cold air in the upper atmosphere that have beencut off from the prevailing westerly circulation under conditions of low zonal flow in themid-latitude index cycle (Barry and Chorley 1998) On the equatorward margins theremnants of tropical cyclones and the seasonal advance of the ITCZ are important givingvery localized short-lived and often high-intensity rainfall events (Fig 13) Examples in Algeria listed by Barry and Chorley (1998) include 87 mm in three minutes (El Golea)385 mm in 25 minutes (Beni Abes) and 46 mm in 63 minutesmdashthough such catastrophic events are not a universal characteristic of drylands (Gordon and Lockwood 1970) Whatis 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 OrsquoHare 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 Egyptby El Dessouky and Jenkinson (1975) has been adapted for investigating the role ofatmospheric circulation pattern on rainfall in dry areas of Spain surface index values canbe correlated with rainfall amounts and statistically significant associations have beenidentified (Spellman 2000) There are therefore distinct opportunities for the analysis ofhistoric rainfall events and drought

Drought

The World Meteorological Organization (1975) defines drought as ldquoa deficit of rainfall in respect to the long term mean affecting a large area for one or several seasons or yearsthat drastically reduces primary production in natural ecosystems and rain-fed agriculturerdquo Drought is commonly defined in terms of its impacts rather than its causeshence the terms lsquoagricultural droughtrsquo and lsquohydrological droughtrsquo have been proposed (Smith 1992) Drought and aridity are not the same thing aridity refers to a negativeratio between mean annual rainfall and mean annual potential evapotranspiration Thedegree of aridity is inversely related to the magnitude of this ratio but drought is more orless related to aridity because arid regions experience frequent droughts Drought impactsare worst in dryland areas because the low mean annual variability of rainfall is


associated with high variability and drought duration is greatermdashthe drought in the Sahel began in the late 1960s and continues rainfall still not reaching the 1931ndash1960 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 Annualtemperature ranges are greatly affected by altitude and the distance from the sea but ingeneral high summer temperatures as a consequence of high radiation loads are commonto all regions (eg Fig 21) In the Sahara maximum average daily temperatures of more than 45degC are recorded in the interior and July temperatures of 375degC are recorded elsewhere except in the highlands

Figure 21 Thermal regimes in two dryland locations Aswan Egypt (112m 24ordm 02prime N 32deg 53prime E) and Jacobabad Pakistan (57m 28deg 17prime N 68deg 29prime E)

Relative humidity

Atmospheric moisture content is typically very low above dryland areas (Table 24) dryness is a response partly to the lack of local evapotranspiration and partly to the lackof horizontal moisture advection Coastal drylands have high humiditiesmdash60 per cent in parts of Western Australia for example compared with under 20 per cent 150 km inlandContrasting humidity regimes are shown in Figure 22 Walvis Bay Namibia in coastal drylands has high humidity In Saleh in the interior of Algeria has similarly low rainfallbut a much drier atmosphere Damascus Syria and Timbuktu Mali respectively northand south of the subtropical anticyclones have annual regimes that mirror each otherdepending on the timing of the rainy season


Wind

Persistent strong winds are common in drylands often a consequence of extensive flatareas with little vegetation cover to disturb air movement in the boundary layer Wind is amajor agent of erosion lifting significant quantities of dust from the dry soil hazyatmosphere with low visibilities is commonplace Wind has an important indirect climaticrole because the amount of dust influences the surface energy balance aiding the processof desertification (Le Houerou et al 1993)

ATMOSPHERIC PROCESSES CAUSING ARIDITY

Condensation of moisture in clouds occurs when moist air is cooled to the point wherebysaturation is reached This occurs through ascent mixing radiation-cooling or contact-cooling with a colder underlying surface The clouds that form must then grow to asufficient depth in order that drops of water can grow to a size to overcome air resistanceand fall to the ground The

Table 24 Mean relative humidity at various isobaric levels for radiosonde stations in the Sahara and the Arabian peninsula (After Lockwood 1974)

Isobaric level January July Fort Trinquet (25deg14primeN 11deg35primeW 360m) 850 33 16 700 22 22 500 15 31 Aulef-el-Arab (27deg04primeN 1deg06primeE 275m) 850 28 14 700 22 17 500 15 21 Habbaniya (33deg22primeN 43deg34primeE 45m) 850 53 15 700 33 18 500 26 7


Figure 22 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 completelack of atmospheric moisture or (2) inhibit cooling of the air through the prevention ofconvection or the creation of inversions or (3) reduce humidity by warming theatmosphere Climatological processes that cause aridity operate at both global andregional scales Thompson (1975) outlines four main processes that help to explain thedistribution 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 slowlycompressed and subsequently adiabatically warmed leading to a dry stable atmosphere(Fig 21) 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 ofrelatively cool air The properties of this boundary layer can be completely different tothat of the sinking air and are usually maintained by a source outside that of the mainanticyclone If the air forming this surface layer originated over the sea it may be moistand contain layer cloud which can result in light rain or drizzle Sinking air will alsoprevent significant depth of thermal convection despite high radiation receipt andsubsequent strong surface heating under clear skies Dryland areas are centred beneaththe subtropical anticyclones in both hemispheres

Wind direction is also important air flowing over the interior of a continent has areduced opportunity to absorb moisture at its base so strong stability and low humiditieswill develop in the lower levels In the northern hemisphere dry northeasterly winds (thereturning flow of the Hadley Cell circulation) contribute to much of the aridity of


Southwest Asia and the Middle East The third factor is topography natural obstaclesacross the path of prevailing winds can cause aridity on their leeward side Thus as moistair is forced to rise over a mountain range (Fig 23) 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 dryadiabatic lapse rate (C to D) and will be warmer than the ascending air at eachcorresponding altitude The air stream will arrive at the other side as a dry desiccatingwind as for example occurs on the leeside of the Sierra Nevada in North America or theAndes 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 lowerlayers (up to 500 m) This induces atmospheric stability which then reduces the potentialfor rainfall production by promoting thin extensive sheets of stratiform cloud cover andpersistent coastal mists and fogs At higher altitudes the air will be warm thus creating astrong inversion that further prevents convection Examples where this effect is importantinclude the Atacama and Namib deserts under the influence respectively of the Peru andBenguela currents

Figure 23 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 athermally driven rising limb of air in the equatorial zone a poleward-moving flow in the


upper atmosphere a sinking limb in the region of the subtropics and a returning tradewind flow at the surface that converges with corresponding winds from the oppositehemisphere at the ITCZ (Fig 24) On the poleward side of the upper atmosphere abovethe return branch of the cell is the subtropical jet a relatively narrow band of high-velocity westerly winds encircling the Earth The Hadley Cells exhibit marked seasonalvariation in intensity geographical extent and latitudinal position

Subtropical anticyclones

Between about 20deg to 40deg mean surface pressure patterns are dominated by adiscontinuous belt of subtropical high (STH) pressure areas broadly elliptical in shapeand oriented in an east-west direction On average they dominate the ocean basins inthese latitudes The geographical positions of the centres of the subtropical highsfluctuate In winter in the southern hemisphere they intensify and spread over theadjacent continental areas and an almost closed belt of high pressure can be formed Insummer thermally produced low-pressure centres over land masses (Australia southern Africa) disrupt the pattern In the northern hemisphere higher central pressure isexhibited 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 onlyone third of a degree of latitude (about 35 km) in the position of the Atlantic high (almostunobservable) causes a one degree

Figure 24 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 theSahel (Oliver 1981) For reasons that are unclear (Hastenrath 1985) the STH centresalso migrate longitudinallymdashin winter in the northern hemisphere all subtropical highsare centred over the eastern regions of their respective ocean basins whereas duringsummer they migrate to the west


Changes in the intensity of STHs (as measured by sea level pressure) are clearlydisplayed in the southern hemisphere The South Pacific anticyclone tends to be strongestin the southern hemisphere spring Jones (1991) has shown that the centre of the SouthPacific anticyclone has declined in strength over the period 1951ndash1985 yet the northern flanks have strengthened In the northern hemisphere temporal variations in the STHintensity have also been identified the North Atlantic anticyclone for instance showed asignificant 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 Earthrsquos surface and maximum pressure to the west at altitude Consequently the circulation around the centre is not parallel to the Earthrsquos surface but slopes gently towards the west with subsidence dominating the eastern half and rising air currentsmore frequent in the west Thus western air masses are more unstable and humidwhereas 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 attributetheir existence to the lsquopiling uprsquo or convergence of the poleward-flowing upper air lsquoanti-tradesrsquo at about 20deg inducing a downward movement of air and high pressure at thesurface Alternatively the main cause may be the movement of polar air (McIlveen1992) As a result of changes in the Coriolis Force with latitude anticyclonic cells nearthe polar front have a tendency to move equatorwards while low-pressure centres generally migrate towards the poles (Rossby 1947) These travelling cold anticyclonesfrequently rejuvenate the subtropical highs Polar outbreaks would therefore prefer theeastern parts of the ocean basins where cold ocean currents prevail and where frictionalong the continental coasts gives a strong meridional influence on these movementsPulses in the intensity of subtropical highs on a daily time scale might be explained bythis idea (McGregor and Nieuwolt 1998) In addition the interaction between cold polaroutbreaks and surface ocean currents maintains the observed higher pressure over theeastern oceans at low levels and the stronger development of STHs over the southernoceans Thermal explanations have also been proposed involving cooling in the upper airand cooling at the Earthrsquos 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 atall levels Cooling at the Earthrsquos surface is seen as a response to cold ocean currents andthe cool continental land areas in winter These features may correlate with the cellularpattern of the subtropical highs and their extension over continents

An explanation of the existence of subtropical highs can also be found in theconsideration of both thermal and dynamic mechanisms According to McIlveen (1992)the Coriolis effect may impose dynamic constraints on the flow of air in anticyclonicsystems once upper level convergence occurs the Coriolis Force prohibits the outflow ofair 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 airpressure falls more slowly with increasing altitude Isobaric surfaces therefore tend to


lsquodomersquo rising to greater heights than in the surrounding air which produces deep andwarm 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 lsquotrade windsrsquo These have a distinct three-layer structure the heights of which increase towards the equator (Fig 25) The height and intensity of the inversion layer will have a strong influence onprecipitation mechanisms These features are generally dependent on latitude or distancefrom the STH centre yet lower more intense inversions (and subsequent weakerconvection) are associated with the east side of ocean basins on the coast of West Africafor example the intensity can be between 5ndash8degC markedly reducing rainfall potential

Figure 25 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 globewhere the two Hadley circulations meet but Waliser and Gautier (1993) have identifiedseven separate ITCZ zones (Table 25) differentiated by structure and behaviour In the northern hemisphere dry conditions are associated with hot continental tropical airwhich moves in behind the ITCZ as it migrates southwards during the winter In Africaat its most southerly extent in January or February the ITCZ lies at about 8deg north of the equator in the west but about 15ndash20deg 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 theprogression (up to about 20degN) shows considerable year-to-year variability with latitudinal departures of up to 6deg (800 km) for some regions and up to 2deg for the global


average These departures which can last from 3 to 18 months may produce lengtheneddrought periods

UPPER AIR CIRCULATION

Explanations of the spatial variability of precipitation conditions can be afforded byreference to upper atmospheric flow particularly the position and intensity of thesubtropical jet stream During the winter dry season when anticyclonic conditions prevailover North Africa the jet stream becomes convergent towards the equator and produces adownward shift of air to feed high pressure at the surface In the summer months upperair divergence results from the convergence of southwesterly and northeasterly winds atthe surface This tends to draw in moist air and increase the likelihood of precipitationLow rainfall totals in dryland areas are explained by weak easterly jet streams associatedwith weaker circulation in the middle and upper troposphere

Subtropical Westerly Jet

Classical models commonly portray the upper air poleward-moving section of the Hadley Cell as a meridional flow (the lsquoanti-tradesrsquo) yet in reality this will be strongly redirected by the Coriolis Force as soon as it moves away from the equator resulting in a narrowband of high-velocity westerly winds known as the Subtropical Westerly Jet which isfound on average at around 30deg from the equator in both hemispheres Palmen andNewton (1969) describe the SWJ as a persistent long-wave pattern encircling the globe with wave troughs at 20degW 150degW and 90degE and wave crests at 70degW 40degE and 150degE Maximum wind speeds of up to 100 m per second are found in the vicinity of the wavecrests

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 maximumintensity 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 SWJwill 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 canextend southwards to interact with low-level (850 mb) underlying tropical easterlies (Fig

Table 25 Seven ITCZ zones (after Waliser and Gautier 1993) Zone Longitude limits (deg)Africa 10ndash40E Indian 60ndash100E West Pacific 100ndash150E Central Pacific 160Endash160W East Pacific 100ndash140W South America 45ndash75W Atlantic 10ndash40W


26) In the central parts of the Sahara rainfall occurs under the variable northward penetration of the West African monsoon trough which allows tongues of moistsouthwesterly 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 lsquorained outrsquo 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 40mm of precipitation

Tropical Easterly Jet

The Tropical Easterly Jet extends from Southeast Asia (80degE) to North Africa (50degE) at approximately latitude 15degN It spirals out clockwise from the subtropical high pressure centres and flows in the northern summer months (June to September) because it isrelated to the seasonal heating cycle (Hastenrath 1985) The strongest intensity is atabout 15 km altitude where maximum speeds are in the region of 40 m per second thereis a second weaker easterly flow at about 5 km The TEJ owes its existence to the strongsurface heating in summer over the land masses in Africa and Asia where very intense

Figure 26 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


heat lows promote the ascent of air to the upper atmosphere Mass convergence in theupper high-pressure system is so intense that pressure surfaces bulge upwards creatingan atmospheric thickness difference between the subtropics and the equatorial and midlatitudes resulting in a reversal in the normal equator-to-subtropics temperature gradient with warmer temperatures recorded in the subtropical upper atmosphere (McGregor andNieuwolt 1998) Convergence in the jet over Africa induces subsidence over the Saheland may be responsible for preventing the advancement of the West African monsoonrains Generally rainfall is greatest north of the jet entrance in the southern Asian regionand 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 hasits core at 15degN at about 4500 m (600 mb) and is located in a region south of the centralarea of anticyclonic outflow Maximum intensity occurs in the northern winter whenhemispheric temperature gradients are steepest flow can reach 10 m per second Rainfalloccurs 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 thewestern Indian Ocean (Fig 27) It exists in all months but its


Figure 27 The monthly progression of the East African Low-Level 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 northernsummer monsoon circulation in the African-Indian area Maximum coolness moistureand cloudiness coincide with the jet core and its eastern regions over the coast of easternAfrica where maximum ascent of air occurs (Kamara 1986) Minimum cloudinessoccurs above the jet core and to the west in the direction of the footslopes of the EastAfrican plateau where the air is descending creating the warmest driest and most stableair (Findlater 1972) The core of this jet occurs at about 1500m where velocities can


reach 25ndash50 m per second Branches of the jet can penetrate inland over eastern Africathrough topographic breaks in the East African plateau It has been related to rainfalloccurrences in the northern parts of Ethiopia and the tracks of so-called Sudano-Sahelian depressions (Fig 28)

PREVAILING WINDS AND MOUNTAIN BARRIERS

Where topographical factors are added to those caused by the general circulation aridityis greatly increased and it is in these areas that the most severe deserts are foundMoisture available for precipitation is trapped in a shallow layer beneath the subtropicalinversion The depth of this moist layer varies but if a mountain barrier projectsthroughout this moist layer it interrupts the surface flow and the surface moist layer willnot penetrate behind the mountain range Even if the range does not completely block themoist layer the reduction in moisture advection to the lee of the range can still besubstantial of the order of 60ndash70 per cent (Lockwood 1974) Dryness can further beenhanced by subsidence of air near the inversion down the lee slopes Such mountain-enclosed inland basins can be extremely arid Death Valley in California is a primeexample

OCEAN TEMPERATURES

Ocean temperature has a considerable influence on climate particularly in coastalregions cool ocean currents moving towards the equator stabilize the atmosphere andreduce atmospheric instability When cold water along the equator is well developed theair above will be too cold even to take part in the ascending motion of the Hadley Cellcirculation Along the coast of Peru the surface moist layer is less than 800 m deep andnormally only drizzle will fall from a deck of stratus The coasts of South America andSouthwest Africa are sheltered respectively by the Andes and Namib escarpment fromthe dynamically stable easterly trades allowing shallow tongues of cold air to roll in fromthe west These are capped by strong inversions at c600ndash1500 m which reinforce the trade wind inversions precluding the development of intense convective cells exceptwhere orographic ascent occurs Precipitation from fog may also result from oceancurrents 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 Nieuwolt1998) as the process by which dryland conditions are brought into


Figure 28 The tracks of Sudano-Saharan depressions over the Sahara


areas where such conditions did not previously exist According to Le Houerou (1996) ifhyper-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 (LeHouerou 1996 Table 26) However it remains debatable how extensive desertification is or how fast it is proceeding (Thomas and Middleton 1994) Two major factors areinvolved 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 1960sin spite of the fact that rainfall was well above the long-term average Le Houerou (1996146) concluded that lsquodesertification may therefore result from land abuse alonersquo Most meteorological models for dryland expansion or the occurrence of episodes ofdrought point ultimately to local changes to the surface energy balance or to large-scale shifts in atmospheric circulation lsquoHuman impactrsquo theories generally focus on areas withsparse vegetation which will commonly have surface air temperatures that are lower thantheir surroundings due to the increased amounts of surface reflectivity of solar radiation(Otterman 1974) Charney (1975) suggested a mechanism (lsquobiogeophysical feedbackrsquo) whereby over-grazing of desert margins can increase surface albedo decreasing the total energy absorbed at surface and reducing thermal convection thereby enhancing stabilityand reducing rainfall potential this in turn provides a positive feedback because lowermoisture availability leads to even less surface vegetation amounts Since much drylandrainfall comes from re-evaporated rainfall not from advected moisture from elsewhere


declining soil moisture may intensify drought conditions (Hulme 1989 Laval 1986)though the theory is disputed (Courel et al 1984 Idso 1977 Williams and Balling1994) A large artificial body of water such as Lake Nasser in Egypt does not increaserainfall in the Nubian desert despite its low albedo which extends over an area of 5000km2 and which contrasts with the very high reflectivity of the surrounding desert (LeHouerou 1996)

Bryson and Murray (1977) suggested that surface desiccation would lead to largeamounts of soil particles being entrained and then lifted aloft by the wind increasing theatmospheric albedo cooling the air aloft and causing it to subside warm adiabaticallyand form an inversion hence preventing convection and cloud development Theysuggested that this was illustrated by the Rajputana Desert on the borders of India andPakistan where extreme dustiness stifles rain processes even though atmospherichumidity is as high as in humid tropical forests

Large-scale climatic explanations for very dry episodes have recently focused onteleconnections Of importance are the possible impacts of anomalous patterns inenvironmental variables particularly sea surface temperature anomalies (SSTAs) whichinfluence 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 Nintildeo (ENSO) events Attempts have been made to link SSTAs in the tropical Atlantic to rainfall in the Sahel Owen and Ward (1989) have linked recurringSSTA patterns to notably wet and dry conditions in sub-Saharan Africa Another example of teleconnections is suggested by Gray (1990) who identified a positive associationbetween rainfall patterns in West Africa and the frequency of intense hurricanes reachingthe Atlantic coast of the United States During the period of drought in the western Sahel(1979ndash1987) 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

Table 26 The extent and severity of desertification (after Le Houerou 1996) Region Light Moderate Strong Severe Total area area area area area Africa 1180 9 1272 10 707 50 35 02 12860 560 Asia 1567 9 1701 10 430 30 5 01 16718 420 Australasia 836 13 24 4 11 02 4 01 6633 320 North America 134 2 588 8 73 01 0 00 7324 500 South America 418 8 311 6 62 02 0 00 5160 290 Total 4273 8 4703 9 1301 25 75 01 51691 397 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


instance the strong 1982ndash1983 ENSO showed good correlations with drought inAustralia (Nicholls 1987) Indonesia (Malingreau 1987) and western South America(Serra 1987) In other areas the relationship was dubious or had very low statisticalsignificance for example northeast Brazil (Gasques and Magalhes 1987) and southernAfrica (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 betweenthe drought and positive SSTAs in the Gulf of Guinea which are in turn related to theBenguela current There seems to be a South Atlantic Oscillation (SAO) comparable toENSO with many similarities between the Humboldt and Benguela currents theirupwellings 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 64 years yet Mediterraneandroughts are totally acyclical and unpredictable especially in North Africa and the NearEast (Le Houerou 1996) A considerable amount of work has been carried out on thissubject (eg Folland et al 1986 Kane 1999 Kiladis and Diaz 1989) Trenberth (1993)describes El Nintildeo as having lsquodifferent flavoursrsquo Consequently finer classifications havebeen attempted Kane (1999) for instance has identified lsquounambiguousrsquo 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 insummer 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 inrainfall in any dryland area no change must be assumed for the not too distant future Incontrast to the predictions about rainfall however GCMs agree (at a 50 per centconfidence level) that the twenty-first century is likely to be characterized by an increasein temperature of 2ndash3degC in the subtropics and 1ndash2degC in the tropics Statistical analysis oftemperature and mean annual evapotraspiration (PET) shows that each degree oftemperature corresponds to 72 mm of PET a year using the Penman equation (LeHouerou 1996) A temperature rise of 1ndash3degC would therefore correspond to a PET increase of 72ndash232 mm a yearmdasha significant increase in climatic aridity Furthermore effects on the movement of the ITCZ and patterns in the westerlies will have an impacton 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 ofdrylands can be recognized The reasons for the shortage of precipitation are many andcomplex The physical processes inherent in the maintenance of drylands involvesynergies and subtleties at a variety of time and spatial scales Similarly it is clear thatdifficult problems remain to be resolved before the magnitude and significance of human-


environmentmdashclimate 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 RG and Chorley RJ (1998) Atmosphere Weather and Climate London Routledge

Barry RG and Perry RJ (1973) Synoptic Climatology Methods and Applications London Methuen


Beaumont P Blake GH and Wagstaff JM (1988) The Middle East A Geographical Study London Fulton

Bryson RA and Murray TJ (1977) Climates of Hunger Mankind and the Worldrsquos 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 andTechnical Analysis 171ndash90 Cambridge Cambridge University PressIntergovernmental Panel of Climate Change

Charney J (1975) Dynamics of deserts and drought in the Sahel Quarterly Journal of The Royal Meteorological Society 101193ndash202

Cooke RU and Warren A (1973) Geomorphology in Deserts London Batsford Courel MF Kandel RS and Rasool SI (1984) Surface albedo and the Sahel drought

Nature 307528ndash31 El Dessouky TM and Jenkinson AF (1975) An Objective Daily Catalogue of Surface

Pressure Flow and Vorticity Indices for Egypt and its Use in Monthly RainfallForecasting Bracknell Meteorological Office Synoptic Climatology Branch Memorandum 46

Findlater J (1972) Aerial explorations of the low level cross equatorial current overeastern Africa Quarterly Journal of the Royal Meteorological Society 98274ndash89

Folland CKPalmer TN and Parker DE (1986) Sahel rainfall and worldwide seasurface temperatures Nature 320602ndash7

Gasques JG and Magalhes AR (1987) Climatic anomalies and their impact in Brazilduring the 1982ndash83 ENSO event In MGlantz RKatz and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 30ndash6 Boulder CO National Center for Atmospheric Research

Glantz M (1987) Impacts of the 1982ndash83 climate anomalies in the West African Sahel In MGlantz RKatz and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 62ndash4 Boulder CO National Center forAtmospheric Research

Gordon AH and Lockwood JG (1970) Maximum one day falls of precipitation inTehran Weather 252ndash8


Gray WM (1990) Strong association between West African rainfall and US landfall ofintense hurricanes Science 2491251ndash6

Hastenrath S (1985) Climate and Circulation of the Tropics Dordrecht DReidel Hills ES (1966) Arid Lands London Methuen Hulme M (1989) Is environmental degradation causing drought in the Sahel An

assessment from recent empirical research Geography 7438ndash46 Hulme M and Kelly M (1993) Exploring links between desertification and climate

change Environment 354ndash11 39ndash45 Idso SB (1977) A note on some recently proposed mechanisms of the genesis of

deserts Quarterly Journal of the Royal Meteorological Society 103369ndash70 Inoue M and Bigg GR (1995) Trends in wind and sea level pressure in the tropical

Pacific Ocean for the period 1950ndash1979 International Journal of Climatology 15 35ndash52

Jones JAA (1997) Global Hydrology Processes Resources and Environmental Management Harlow Longman

Jones PD (1991) Southern hemisphere sea level pressure data an analysis andreconstruction back to 1951 and 1911 International Journal of Climatology 11 585ndash608

Kamara SI (1986) The origins and types of rainfall in West Africa Weather 41 48ndash56 Kane RP (1999) Rainfall extremes in some selected parts of central and South America

ENSO and other relationships re-examined International Journal of Climatology19423ndash55

Kiladis GN and Diaz HF (1989) Global climatic anomalies associated with extremesof the Southern Oscillation Journal of Climate 21069ndash90

Kodama Y (1992) Large scale common features of subtropical precipitation zones (theBaiu Front The South Pacific Convergence Zone the South Atlantic Convergencezone) Part 1mdashCharacteristics of the Subtropical frontal zones Journal of the Meteorological Society of Japan 70813ndash35

Kodama Y (1993) Large scale common features of subtropical precipitation zones (theBaiu Front The South Pacific Convergence Zone the South Atlantic Convergencezone) Part IImdashConditions for generating the subtropical convergence zones Journal of the Meteorological Society of Japan 71581ndash610

Koppen W (1931) Die Klimate der Erde Berlin Lamb HH (1982) Climate History and the Modern World London Routledge Laval K (1986) General circulation model experiments with surface albedo change

Climatic Change 991ndash102 Le Houerou HN (1977) Biological recovery vs desertization Economic Geography

53413ndash20 Le Houerou HN (1996) Climate change drought and desertification Journal of Arid

Environments 34133ndash85 Le Houerou HN Popov GF and See L (1993) Agrobioclimatic Classification of

Africa Rome Food and Agriculture Organization Agrometeorology Series Working Paper No 6

Lockwood JG (1974) World Climatology An Environmental Approach London Edward Arnold


Malingreau VP (1987) The 1982ndash83 drought in Indonesia Assessment and monitoring In MGlantz RKatz and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 11ndash18 Boulder CO National Center forAtmospheric Research

Mather JR (1974) Climatology Fundamentals and Applications New York McGraw-Hill New York

McGinnies WG (1988) Climatic and biological conditions of arid lands a comparisonIn EEWhitehead CFHutchinson BNTimmerman and RGVardy (eds) Arid Lands Today and Tomorrow 61ndash8 Boulder CO Westview Press

McGregor GR 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 1203ndash10 Middleton NJ (1991) Desertification Oxford Oxford University Press Morel R (1992) Atlas Agroclimatique de Pays de la Zone de CILSS Niamey

AGRHYMET Musk LF (1988) Weather Systems Cambridge Cambridge University Press Nicholls N (1987) The El NineSouthern Oscillation phenomenon In MGlantz RKatz

and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 2ndash10 Boulder CO National Center for Atmospheric Research

Nicholson SE (1993) An overview of African rainfall fluctuations of the last decadesJournal of Climate 61463ndash6

Nicholson SE 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 JE (1973) Climate and Manrsquos Environment Chichester John Wiley and Sons Oliver JE (1981) Climatology Selected Applications London Edward Arnold Otterman J (1974) Baring high albedo soils by over-grazing Science 86531ndash3 Owen JA and Ward MN (1989) Forecasting Sahel rainfall Weather 4457ndash64 Palmen E and Newton CW (1969) Atmospheric Circulation Systems New York

Academic Press Pearce EA and Smith CG (1984) World Weather Guide London Hutchinson Penman H (1948) Natural evaporation from open water bare soil and grass

Proceedings of the Royal Society A193120ndash45 Rossby CG (1947) On the general circulation of the atmosphere in the middle latitudes

Bulletin of the American Meteorological Society 28255ndash80 Serra RB (1987) Impact of the 1982ndash83 ENSO on the southeastern Pacific fisheries

with emphasis on Chilean fisheries In MGlantz RKatz and MKrenz (eds) The Societal Impacts Associated with the 1982ndash83 Worldwide Climate Anomalies 24ndash9 Boulder CO National Center for Atmospheric Research

Sharon D (1972) The spottiness of rainfall in a desert area Journal of Hydrology 17 161ndash75

Sharon D (1981) The distribution in space of local rainfall in the Namib desert Journal of Climatology 169ndash75


Smith K (1992) Environmental Hazards Assessing Risk and Reducing Disaster London Routledge

Soliman KH (1953) Rainfall over Egypt Quarterly Journal of the Royal Meteorology79389ndash401

Spellman G (2000) An objective weather type method for the Iberian peninsulaWeather (in press)

Sweeney JC and OrsquoHare GP (1992) Geographical variations in the precipitationyields and circulation types in Britain and Ireland Transactions of the Institute of British Geographers 17448ndash63

Thomas DG (1989) (ed) Arid Zone Geomorphology London Bellhaven Press Thomas DG and Middleton NJ (1994) Desertification Exploding the Myth

Chichester John Wiley and Sons Thompson RD (1975) The Climatology of the Arid World Reading University of

Reading Department of Geography Paper No 35 Thornthwaite CW (1948) An approach towards a rational classification of climate

Geographical Review 3855ndash94 Trenberth KE (1993) The different flavours of El Nintildeo Proceedings of the 18th Annual

Climate Diagnostics Workshop 50ndash3 Boulder CO National Center for AtmosphericResearch

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 DE and Gautier C (1993) A satellite-derived climatology of the ITCZ Journal of Climate 6 2162ndash74

Wallen CC (1967) Aridity definitions and their applicability Geografiska Annaler 49a 367ndash84

Wilby RL and Wigley TML (1997) Downscaling general circulation model output areview of methods and limitations Progress in Physical Geography 21 530ndash48

Williams MAJ and Balling RC (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 SM (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 AROSEN

INTRODUCTION

The presence of sophisticated large-scale agricultural systems dating to classical timesin the arid regions of the central Negev southern Jordan and Sinai has long served bothto illustrate the ingenuity of the ancient peoples of the region and as an inspiration tomodern peoples as to the potential of wise exploitation of the desert Archaeologicalsurvey 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 198232 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 31)

The amazing efficacy of these systems has been repeatedly demonstrated Both texts (Bruins 198687 Kraemer 1958 Document 82 Mayerson 1960224ndash69) and experimental archaeology (Evenari et al 1982191ndash219) have indicated that yields from the desert zone using run-off water catchment systems could in fact approximate thoseof the Mediterranean zone (Bruins 198687 Evenari et al 1982191ndash219) Excavations and surveys have revealed the existence of large and numerous wine presses (Mazor1981 Rubin 199654 Shershefski 1991198ndash200 Fig 32) suggesting industriallevel production of grapes and wine The reconstruction and operation of some of thesesystems over several decades have demonstrated that in some ways they constitute anagricultural 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 togetherwith their village and homestead hinterland comprised an urban system proper with apopulation of over 20000 people whose subsistence was based on this agriculturalregime (Broshi 1979 Elliot 1982103ndash14 Shershefski 1991200ndash14 Fig 34)

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 centralNegev had been abandoned All previously occupied

Figure 31 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 SRosen

sites including towns villages farmsteads and even nomadic encampments had beendeserted and there is no evidence for any alternative settlements either permanent ornomadic (for example Avni 1996 Nachlieli 1992 Rosen 1987a Rosen and Avni1993) 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 33) and there has been no shortage of attempts to explain this apparent lsquotriumph of the desertrsquo Two general factors have been suggested as primary causes for Negev desertification (1) the Moslem or Arabconquests and the ensuing destruction of Byzantine civilization (for exampleLowdermilk 1945136 Negev 198815 Palmer 1872243 Reifenberg 195598Sharon 1969) and (2) climatic deterioration rendering habitation impossible due toshifting sands increased erosion and reduced water for agriculture (for exampleHuntington 1911 Issar 1995 Issar and Govrin 1991) Additional subfactors haveincluded the negative effects of over-grazing by the flocks and herds of bedouin(Reifenberg 195598) the destructive effects of earthquakes (Fabian 1994) andincreased marauding by nomads (Sharon 1976 cf also Lowdermilk 1945129)

Critical examination of these factors in the light of recent intensive archaeologicalresearch carried out in the Negev indicates that each of these explanations isfundamentally flawed as a prime mover in the desertification

The decline of desert agriculture a view from the classical period Negev 45

Figure 32 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 SRosen

of the Negev although each plays a role within a larger perspective The key issue rarelydiscussed in reviews of the decline of classical civilization in the Negev is that periods ofcultural florescence can usually be tied to increased economic and social input from orintegration with the Mediterranean core area The collapse of the economic core willinevitably result in the collapse of its dependants unless alternative economic paths areavailable

ARCHAEOLOGICAL OVERVIEW

The central Negev in the sixth century AD the Byzantine period in local terms was thewell-integrated frontier province of Palestina Tertia of the Late Roman empire (Mayerson 1994 Rubin 1997 Shershefski 1991 also see Isaac 1992) Although thelucrative trade route of the Nabatean period had long since been eclipsed by alternativetrade systems (Crone 1987 Negev 1988) the province functioned both as a strategicsouthern buffer zone protecting the Levantine heartland (Gihon 1980 Mayerson 19861990 also Isaac 1992 for a differing view) and as a gateway to both the holy pilgrimagedestinations of the Sinai and to the mineral-rich desert regions farther east and south (Mayerson 1982 1983)


Figure 33 Palmerrsquos 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 1872314

Archaeologically the region is marked by two complementary settlement systems(Avni 1996 Elliott 1982 Haiman 1995a Mayerson 1989 Negev 1988 Rubin 1990Rosen 1987b Rosen and Avni 1993 Shersehfski 1991 Fig 34) First in the north and in the higher mountains both better watered than areas farther south large towns such asAvdat (Fig 35) supported by intensive run-off agricultural systems (Fig 36) 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 deepinto the desert The design and construction of these towns are dominated by anarchitecture whose roots are undeniably in the Mediterranean zone with little adjustmentfor local conditions (Shershefski 1991228) excepting the use of local raw materials(Negev 1980) Christianity is the only religion represented at these sites in this pre-Islamic period and classic basilica-style churches are present in the plural at each townThe wealth of the towns is especially evident in these churches which showed suchfeatures as wall facings and furniture of marble imported from Anatolia elaboratemosaics and vaults of large wooden beams imported from the Mediterranean zone(Negev 1974)


Figure 34 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)


Figure 35 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 SRosen

Although defensively postured defence does not seem to have been a primaryconsideration in the settlements Aside from the isolated nature of many of the villagesand farmsteads only Mampsis shows a circumference wall although Avdat shows afortification wall on one side of the settlement Neither is especially massive Both Avdatand Nessana show internal forts indicating military presence Subeita presents a limitednumber of access gates to the town but these gates are in fact breaks in the continuum ofattached structures and not the gates of a city wall (Shershefski 1991184ndash8)

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 damsdrainage channels sluice gates and support walls Hill slopes are covered with tuleiliot el anabmdashrows of stone mounds and stone linesmdashwhose function was presumably connected to either ground clearance for run-off enhancement or some other form of agricultural activity (Evenari et al 1982127ndash47) Calculations based on aerial photography pedestrian survey and farm reconstruction demonstrate that the averageratio of drainage catchment to farmed area was


Figure 36 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 SRosen

approximately 211 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 forthe farmed fields of more than 400 mm (Evenari et al 198295ndash119) 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 winepresses at each town sometimes at an industrial scale demonstrates clearly the practiceof arboriculture and viticulture dates figs and even pomegranates were also grown(Mayerson 1960 Mazor 1981 Rubin 1996) Rubin (1996) characterizes this system asthe 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 thesignificantly lower rainfall associated with these southern areas the region is marked bythe general scarcity of agricultural remains and the presence of the larger-scale pastoral encampments The remains of pastoral encampments are found throughout the desert andsteppe zones but the larger aggregate camps are located only south of the agriculturalareas These camps are obviously smaller than the Byzantine towns and villages but they also differ in their basic architecture and organization In essence the structuresrevealed at such encampments are to be interpreted as ephemeral tent bases or in somecases as hut foundations that carried brush or tent superstructures The settlements alignalong secondary and tertiary drainages in patterns dictated by topography Analyses of


material culture also support the interpretation of these settlements as basicallypastoralnomadic (Rosen and Avni 199762ndash81) and textual references (Mayerson 1989)accord well with this A key point in analyses of these pastoral systems is their essentialdependence on the settled system to the north for both their subsistence and their materialculture 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 perhapsoccasionally tense must have been essentially stable for the nomadic system to havethrived

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 theeconomic and social stability that enabled the desert to bloom

THE ISLAMIC CONQUESTS AS CAUSE FOR DESERTIFICATION

The battle of Gaza in AD 633ndash4 marks the beginning of the political end of the Byzantine empire in the Negev Although the events leading up to that battle and thecauses behind the Byzantine collapse have been much discussed and are beyond thescope 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 ofthe Negev towns (per contra Negev 198815) In fact the processes of urban declineseem to have been initiated well before the Islamic period Mampsis (Negev 198815)does not appear to show an occupation in the seventh century at all Avdat showsevidence for a major earthquake at the beginning of the seventh century after which thecity seems to have been abandoned for two centuries and eventually reoccupied duringthe Islamic period (Fabian 1994) Significantly an earlier fourth- or fifth-century earthquake resulted in repairs and lsquoretro-fittingrsquo of various structures against furtherearthquake damage Nessana shows continued occupation at least into the late seventhcentury and probably well into the eighth both in the archaeology (Shershefski 1991550) and in the archives recovered from the site (Kraemer 1958213) with little obviousdisruption although a clear decline can be traced At Subeita the presence of a mosquewedged into an open space next to a church (Baly 1935 Segal 1983 Shershefski199174) indicates both clear continuity of occupation well into the eighth century and itscontemporaneity with at least one church on the site indicating the peaceful coexistence of the two religions during the Ummayad period Ruheiba (Shershefski 199195) alsoseems to show continued occupation into the early Islamic period Recent excavations atElusa have not revealed any evidence for Islamic occupation but nor is there anyevidence for destruction (although it must be admitted that the areas excavated are stillquite 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 eitherabandonment 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


Sassanids Although it is unlikely that the Persian armies that devastated the Levantactually came as far south as the central Negev the havoc wreaked on the Mediterraneanheartland 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 earlyAbassid periods seem to show a rural florescence The central village and satellite farmsat Sede Boqer (Nevo 1985 1991 Fig 37) are the best example of this phenomenonAnother example is the farmstead at Nahal Mitan (Haiman 1995b) Avni (1994) hasindicated the presence of at least thirteen mosques in the Negev highlands in this periodsome of

Figure 37 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 SRosen

which are clearly associated with farming settlements and others with pastoralencampments Finally a series of large homesteads that were colonized during theUmmayad and early Abassid periods has been excavated recently around the outskirts ofBeer-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 ninthcenturies 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 intothe pastoral subsistence system (Rosen and Avni 1993) The continued import and use of


typologically Byzantine ceramics (and other elements of material culture) from thesettled regions into the pastoral sites demonstrate underlying economic continuitiesThere was no break in relations between the nomads and the farmers in the transition tothe Ummayad administration Importantly there is no incursion of nomadic settlementtypes into the agricultural zones in this period Although erosion is a dominant feature inthe 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 conquestsmdasha problematic concept in itself for the Negevmdashdid not bring any desertification Whilst the late Byzantine period saw an urban decline in theNegev 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 threedistinct steps First one must establish the reality of the climatic change itself Secondthe suggested climatic change must be correlated chronologically with the culturaltransformation Third a reasonable scenario or mechanism for causality must beestablished beyond the mere fact of correlation it is not enough to establish a climaticchange 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 extensiveterraces sometime during the classical period (Bruins 1986189 Goldberg 1994)followed by their erosion and wadi downcutting (Ben-David 1997 Bruins 1986189 Reifenberg 1955) It is clear that there has been landscape degradation but it is not cleareither when this degradation occurred or whether it was the result of climatic changes orof other factors such as microtectonics or human intervention Certainly acceleratederosion 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 ofbreached 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 afactor The pastoral encampment of Nahal Oded (Fig 34) 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 cAD 800 is the result of an increased supply of Nile sands on the Levantine littoral to be correlatedwith increased monsoon rains in East Africa

Dead Sea water levels as established from salt cave evolution and analysis of coresediments have also been used to reconstruct climatic sequences (Frumkin et al 1991


1994 Geyh 1994 Issar 1995 Neev and Emery 199562) Summarized briefly higherDead 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 inthe middle of the first millennium BC that was not ameliorated until the beginning of thesecond millennium AD

Analyses of oxygen isotope ratios from cave speleothems and marine molluscs (Gatand Magaritz 1980 Geyh 1994) show high 18O ratio peaks of c2300 BP and c1500 BP indicating cooler temperatures (and presumably higher humidity) with cooler (andmoister) periods between and following These analyses accord well with the studies ofthe Dead Sea water level Of further interest is the apparently significantly warmer (anddrier) period prior to 2300 BP so that although it was not especially cool or moist on anyabsolute scale the c2300 BP episode is a relatively significant amelioration Laterepisodes 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 climateremained stable during the first millennium AD (per contra Rubin 1989) The next issues are whether the climatic fluctuations outlined above do indeed correspond withand 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 intrusionand 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 thedesertification of the Negev Thus the Nabatean and early Roman periods in the finalcenturies BC and first two centuries AD when agriculture was incipient at best (Bruins1986189 Mayerson 1963) seem to have been at a climatic optimum whereas thecultural peak in the succeeding Byzantine period seems to have been climatically dryThe Byzantine collapse and rise of the early Islamic empire seem to have been eitherstable climatically or marked by only minor fluctuations Although sand dunes did indeedbury 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 isimportant to stress here that the gradual abandonment of the Byzantine cities is notequivalent to either the abandonment of the Negev or desertification for as indicatedearlier 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 Negevprobably in the tenth or perhaps even eleventh centuries AD may in fact even beassociated with the beginning of climatic amelioration In short climatic change does notadequately explain the decline of classical civilization in the desert or the reversion of thedesert to desert

THE RISE OF THE DESERT

To understand the rise of the desert we must understand first its domestication Theessence of the classical period lsquoGreen Revolutionrsquo in the Negev was the transplantation of a Mediterranean-zone agricultural complex into the arid zone This complex in the


Mediterranean zone consists of cereal (wheat and barley) farming the cultivation of fruitcrops including grapes olives figs and dates and animal husbandry based especially onsheep goat and cattle with significantly less emphasis on pig Landscape managementin the form of hill slope terracing and various forms of irrigation is integral to thecomplex as well (Grigg 1974123ndash8 132ndash4 Stager 1985 compare also with Braudel 197256 59 423) Despite claims concerning the inappropriateness and instability ofMediterranean-zone farming systems in the New World and other non-Mediterranean environments (see Butzer 1996) the expansion of the Mediterranean zone into thedesert in terms of culture society and subsistence in fact proved a remarkably stablephenomenon enduring for at least half a millennium The stability of this system is evenmore marked given the political perturbations that occurred during this periodperturbations that included the rise and decline of urban centres the rise of Christianitythe 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 oftrade and subsistence into the larger state Even beyond the fact of active imperialsubsidy 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 isthat the Mediterranean economy itself should be seen as a flexible strategy fluctuatingbetween emphasis on cash crops and subsistence staples depending on the historical andeconomic contexts Within the Mediterranean zone during periods of social collapse thecomplex shifts towards subsistence mode whereas during times of economic prosperitycash 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 withoutreference to climatic change the desert environment exerts pressures on settlementsystems 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 labourinput than further north for example in the construction and constant maintenance ofterrace dam systems Subsistence is more difficult in a desert and therefore the raison drsquoecirctre of permanent settlement in the Negev has always been its integration with someother core region The decline of the core-region economy resulted in a reversion towardsthe subsistence end of the Mediterranean complex spectrum one that may not have beensustainable 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 spiteof its rural character still shows a high degree of social and economic integration withthe Mediterranean core area This is most obvious in the material culture continuitiesbetween the core and the periphery However it is especially impressive in theideological integration such that Negev rock inscriptions from this period follow verystandardized Islamic formulae (Sharon 1990) burials are typically Moslem (Rosen andAvni 199713) and mosques follow standard definitions (Avni 1994) It was in the laterAbbassid period following the political and economic shift of the Caliphate to Baghdadthat the Levant itself declined With that decline the means for the integration of thedesert and the sown were no longer available and the entire desert system both settled


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 thecentral Negev reverted from being a productive and integrated component of aMediterranean state system to its original desert state the processes reviewed here areindeed those of desertification

FINAL NOTE

The history of research on the rise of Near Eastern deserts is one inextricably tied to thepolitical and ideological struggles of the region For example the nineteenth-century British Orientalist Edward Palmer (1872241ndash3) viewed the decline of civilization and the rise of the desert as the result of invasion and indigence on the part of the localinhabitants Ellsworth Huntingdonrsquos (1911) environmental determinism in which heclaimed that settlement and the rise and decline of civilization were dictated by thecarrying capacity of a region in turn determined by climate and environment was inantithesis to attitudes like Palmerrsquos it was adopted as state policy by the British Foreign Office in its administration of Palestine and used as a rationale for limiting Jewishimmigration 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 ofnegligent administration discounting the role of climatic change (Troen 1989) IndeedBen-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 afactor in the reduction of vegetation and in the consequent rise in erosion has often beenstressed (Orni and Efrat 1980470 Reifenberg 195598 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 deniedtraditional pastoral nomadism and grazing as a significant factor in landscape degradation(Thomas and Middleton 199413 67ndash73)

Desertification is the result of a complex chain of causality On its most simple level itis clear that land degradation is the result of physical processes However these physicalprocesses are often set in motion by human activities (Glantz 1994) such that the issuesare social and historical (Blaikie and Brookfield 1987) The historical causalities are alsocomplex it is accepted knowledge that over-grazing by pastoralists causes erosion (Orniand Efrat 1974470 Reifenberg 195598) but Danin (198317) notes that lsquoduring the few years that several Negev and Sinai areas were closed to bedouin and their domesticanimals no substantial changes in the list of species and plant communities could bediscernedrsquo 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 almostall exploited for agriculture during the classical period and subsequently abandoned tobe exploited by pastoralists only later The simplistic notions of Islamic invasion orclimatic catastrophe as prime causes in the decline of the Negev in fact mask politicalagendas It is the historical complexities in all their richness and texture that need to be


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 inthe 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 valuablecomments Arlene Miller Rosen shared her knowledge of climate and climate-change freely and happily The photographs were developed from slides by Alter Fogel and themap was drafted by Patrice Kaminsky both of Ben-Gurion University

REFERENCES

Avni G (1994) Early mosques in the Negev highlandsmdashnew archaeological evidence on Islamic penetration of southern Palestine Bulletin of the American Schools of Oriental Research 29483ndash100

Avni G (1996) Nomads Farmers and Town-Dwellers Jerusalem Israel Antiquities Authority

Baly C (1935) Sbaita Palestine Exploration Fund Quarterly Statement 62171ndash81 Bar-Ziv H and Katz O (1993) Beer ShevamdashNahal Ashan Archaeological News 100

100ndash01 (In Hebrew) Ben-David R (1997) The geology and geomorphology of the Nahal lsquoOded site area In

SARosen and GAvni The lsquoOded Sites Investigations of Two Early Islamic Pastoral Camps South of the Ramon Crater 109ndash18 Beersheva Ben-Gurion University

Press Beersheva XI Studies by the Department of Bible and Ancient Near East Ben-Gurion D (1961) Introduction In YMorris Masters of the Desert 11ndash20 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 PBlaikie

and HBrookfield (eds) Land Degradation and Society 1ndash7 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 2361ndash10 Bruins HJ (1986) Desert Environment and Agriculture in the Central Negev and

Kadesh Barnea during Historical Times Nijkirk Netherlands Midbar Foundation Butzer KW (1974) Accelerated soil erosion a problem of man-land relationships In

IRManners and MWMikesell (eds) Perspectives on Environment 57ndash77 Washington DC Association of American Geographers

Butzer KW (1996) Ecology in the long view settlement histories agrosystemicstrategies and ecological performance Journal of Field Archaeology 23141ndash150


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 HistoricData Mississippi State Mississippi State University Cobb Institute of ArchaeologyOccasional Papers 82ndash01

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 urbandevelopment of Avdat Paper presented at the 20th Archaeological Congress inJerusalem Israel

Figueras P (1979) The Roman-Byzantine period In YGrades and EStern (eds) Beersheva 39ndash52 Jerusalem Keter (In Hebrew)

Frumkin A Magaritz M Carmi I and Zak I (1991) The Holocene climatic record ofthe salt caves of Mount Sedom Israel The Holocene 1191ndash200

Frumkin A Carmi I Zak I and Magaritz M (1994) Middle Holocene environmentalchange determined from the salt caves of Mount Sedom Israel In O Bar-Yosef and RKra (eds) Late Quaternary Chronology and Paleoclimates of the EasternMediterranean 315ndash22 Tucson University of Arizona Radiocarbon Press

Gat JR and Magaritz M (1980) Climatic variations in the eastern Mediterranean seaarea Naturwissenschaften 6780ndash7

Geyh MA (1994) The paleohydrology of the eastern Mediterranean In OBar-Yosef and RKra (eds) Late Quaternary Chronology and Paleoclimates of the EasternMediterranean 131ndash45 Tuscon University of Arizona Radiocarbon Press

Gihon M (1980) Research on the Limes Palaestina a stocktaking In WSHanson and LJFKeppie (eds) Roman Frontier Studies 1919 843ndash64 Oxford British Archaeological Reports International Series 71

Gilead I Rosen SA and Fabian P (1993) Horvat Beter (Bersquoer Matar) 1990ndash1991 Archaeological News 9988ndash89 (In Hebrew)

Glantz MH (1994) Drought desertification and food production In MHGlantz (ed)Drought Follows the Plough 7ndash32 Cambridge Cambridge University Press

Goldberg P (1994) Interpreting late Quaternary continental sequences in Israel InOBar-Yosef and RKra (eds) Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean 89ndash102 Tucson University of Arizona Radiocarbon Press

Grigg DB (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 theByzantine and early Islamic periods Bulletin of the American Schools of Oriental Research 29729ndash54

Haiman M (1995b) An early Islamic period farm at Nahal Mitnan Atiqot 261ndash13 Huntington E (1911) Palestine and Its Transformation Boston Houghton amp Mifflin Isaac B (1992) The Limits of Empire Oxford Clarendon Press Issar A (1995) Climatic change and history of the Middle East American Scientist

83350ndash5


Issar A and Govrin Y (1991) Climatic changes and the desertification of the Negev atthe end of the Byzantine period Katedra 6167ndash81 (In Hebrew)

Katz O (1993) Beer ShevamdashNahal Bekarsquo 1 Archaeological News 9987ndash8 (In Hebrew)

Katz O and May V (1996) Beer Sheva Ramot B Archaeological News 106162ndash4 (In Hebrew)

Kedar Y (1967) Ancient Agriculture in the Negev Highlands Jerusalem Bialik Institute

Khazanov AM (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 OBar-Yosef and AMKhazanov (eds) Pastoralism in the Levant Archaeological Materials in Anthropological Perspectives 11ndash18 Madison Prehistory Press

Kraemer CJ (1958) Non-Literary Papyri Excavations at Nessana Volume III Princeton Princeton University Press

Lowdermilk WC (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 107160ndash72 Mayerson P (1982) The pilgrim routes to Mount Sinai and the Armenians Israel

Exploration Journal 3244ndash57 Mayerson P (1983) The city of Elusa in the literary sources of the fourth-sixth centuries

Israel Exploration Journal 33247ndash53 Mayerson P (1986) The Saracens and the Limes Bulletin of the American Schools of

Oriental Research 26235ndash47 Mayerson P (1989) Saracens and Romans micro-macro relationships Bulletin of the

American Schools of Oriental Research 27471ndash9 Mayerson P (1990) Toward a comparative study of a frontier Israel Exploration

Journal 40267ndash79 Mayerson P (1994) Monks Martyrs Soldiers and Saracens Papers on the Near East in

Late Antiquity ( 1962ndash1993 ) Jerusalem Israel Exploration Society Mazor G (1981) The wine presses of the Negev Qadmoniot 53ndash5451ndash60 (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 ShevamdashNahal Kovshim

Archaeological News 9984ndash5 (In Hebrew) Neev D and Emery KO (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 81400ndash22 Negev A (1980) House and city planning in the ancient Negev and Provincia Arabia In

GGolany (ed) Housing in Arid Lands Design and Planning 3ndash32 New York John


Wiley and Sons Negev A (1988) The Nabatean Cities in the Negev Jerusalem Ariel 62ndash63 Negev N (1993) Beer ShevamdashKiryat HaUniversita (Mizrach) Archaeological News

9985ndash6 (In Hebrew) Nevo YD (1985) Sede Boqer and the Central Negev in the 7ndash8th 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 EH (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 SA (1987a) Demographic trends in the Negev highlands preliminary results

from the Emergency Survey Bulletin of the American Schools of Oriental Research26645ndash58

Rosen SA (1987b) Byzantine nomadism in the Negev results from the EmergencySurvey Journal of Field Archaeology 1429ndash42

Rosen SA (1994) The nomadic periphery archaeology of pastoralists in the southcentral Negev during late antiquity Aram 6295ndash309

Rosen SA and Avni G (1993) The edge of empire the archaeology of pastoral nomadsin the southern Negev highlands in late antiquity Biblical Archaeologist 56 189ndash99

Rosen SA and Avni G (1997) The lsquoOded 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 12171ndash8

Rubin R (1990) The Negev as Settled Land Jerusalem Hebrew University Press Rubin R (1996) Urbanization settlement and agriculture in the Negev desertmdashthe

impact of the Roman-Byzantine empire on the frontier Zeitschrift des Deutschen Palstini-Vereins 11249ndash60

Rubin R (1997) The Romanization of the Negev Israel geographical and culturalchanges in the desert frontier in late antiquity Journal of Historical Geography 23 267ndash83

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(633ndash1099) In MAvi-Yonah (ed) A History of the Holy Land 185ndash220 London Weidenfeld amp Nicolson

Sharon M (1976) Processes of destruction and nomadization in Eretz Israel underIslamic rule (633ndash1517 CE) In MSharon (ed) Notes and Studies on the History of theHoly Land Under Islamic Rule 7ndash32 Jerusalem Yad Ben Zvi Press

Sharon M (1990) Arabic rock inscriptions from the Negev In MSharon and MHalloun(eds) Supplement to the Map of Har Nafha (196) 9ndash35 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


Ancient Near East Stager LE (1985) The first fruits of civilization In JNTubb (ed) Palestine in the

Bronze and Iron Ages Papers in Honor of Olga Tufnell 172ndash88 London University of London Institute of Archaeology

Thomas DSG and Middleton NJ (1994) Desertification Exploding the Myth New York John Wiley and Sons

Troen I (1989) Calculating the economic absorptive capacity of Palestine a study of thepolitical uses of scientific research Contemporary Jewry 1019ndash38


4 Farmers herders and miners in the Wadi Faynan

southern Jordan a 10000-year landscape archaeology

GRAEME BARKER

INTRODUCTION

The Wadi Faynan Landscape Survey is a study of the landscape evolution of WadiFaynan in southern Jordan from prehistoric times to the present day as a contribution tothe 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 societiesin marginal environments learn to cope with risk What solutions did they develop andhow successful were they Why did they take the choices they took To what extent didtheir actions affect their landscape and for good or ill The rationale of the project hasbeen to bring together an inter-disciplinary team of archaeologists geographers and environmental biologists in the investigation of the landscape history of the chosen studyarea within a single integrated research framework (The Acknowledgements at the endof the chapter detail the numerous colleagues working in the project whose researches aresummarized 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 Romanconquest of the region (Fig 41) The catchment of the wadi forms a transect about five km wide running for some 15 km westwards from the rim of the Jordanian plateauc1500 m above sea level to the floor of the Wadi Arabah rift valley at about sea levelThe main wadi today is a bleak landscape arid and largely denuded of vegetation (Figs42 45) though where they cut through the plateau escarpment the channels of the threemain feeder tributaries (the Dana Ghuwayr and Shayqar) are in places well watered andcomparatively well vegetated from ground springs The Wadi Faynan today (part of theDana Nature Reserve of Jordanrsquos Royal Society for the Conservation of Nature) is usedmainly by nomadic bedouin herders but is also well known for its abundantarchaeological remains

The principal archaeological monuments of the Wadi Faynan long known to early travellers are the Khirbet Faynan (Fig 42) a major settlement of

Figure 41 The location of Wadi Faynan within its region

Farmers herders and miners in the Wadi Faynan southern Jordan 63

Figure 42 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 GBarker

Nabatean Roman and late Roman (Byzantine) date located at the head of the WadiFaynan near the confluence of the three main tributaries and nearby an aqueductreservoir and water mill of RomanByzantine date To the west of this complex is asubstantial (c5 km long) field system of rubble walls its surface pottery indicatingprimary use contemporary with that of the Khirbet Faynan settlement (Fig 43) Before our project began in 1996 reconnaissance surveys had also located a variety ofprehistoric sites both in the main wadi and in its tributaries some of which are beingexcavated by other teams Wadi Faynan and its environs are also characterized by richmineral deposits and from the work especially of the Bochum Mining Museum thehistory of copper exploitation here is comparatively well documented (Hauptmann 19891992 Hauptmann and Weisgerber 1987 Hauptmann et al 1992) Although Faynan copper was used by neolithic and chalcolithic societies the first intensive exploitationseems to date to the Early Bronze Age c3500ndash1900 BC There was a second significant episode in the first part of the first millennium BC the Edomite Iron Age Copper wasthen extracted on a major scale in Nabatean and especially Roman and Byzantine timesit 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 andEgypt were transported in the third and early fourth centuries AD to work the coppermines it controlled


Figure 43 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 lsquoarchaeological historyrsquo 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 verydifferent from settlement and land use today

METHODOLOGIES

The project began in 1996 and the fieldwork ends in 2000 ongoing results have beenreported in annual papers in Levant (Barker et al 1997 1998 1999 2000)

Geomorphological mapping and palaeoecological analysis of exposures and coredsediments are establishing an environmental sequence for the past 200000 years with aparticular focus on the past 10000 years The resulting event sequence is being datedvariously by radiocarbon and Optical Spin Luminescence (OSL) dates and bystratigraphic association with dated archaeological sites The changes that are beingobserved in the palaeoenvironmental sequence can be linked with increasing confidencevariously to natural processes of change such as climatic shifts and phases of tectonicactivity and to cultural processes of change such as arable pastoral and industrialactivities Geochemical analysis of sediments using EDMA (Energy Dispersive X-Ray Microanalysis) is also being used to monitor the bioaccumulation of heavy metals as a


result of metalliferous pollution providing an invaluable independent indicator of thechanging scales technologies and environmental impacts of mining and smeltingactivities to compare with the Bochum teamrsquos 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 thesurvey catalogue) that covers the lower terraces on the southern side of the present-day wadi channel (Figs 42 and 43) This involved the verification on the ground of walls shown on an earlier photogrammetric survey the systematic collection and counting ofartefacts 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 oradjacent to the fields A series of smaller satellite field systems on the northern side of themain channel has also been recorded in the same way These studies combined with testexcavations and identification of the prolific lithic and pottery collections haveconfirmed that the main phase of construction and use of the field systems is broadlylsquoclassicalrsquo (NabateanRomanByzantinemdasha period of some 1000 years) but have also established that the evidence is at the same time a complex palimpsest of reuse andadaptation of agricultural activities and systems of land management spanning the past6000 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 inFigure 42 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 usinghand-held Garmin 12 GPS units within a grid based on UTM (Universal TransverseMercator) coordinates This survey located over 1000 lsquositesrsquo varying from lithic scatters to settlement structures and enclosures dating to all archaeological periods fromprehistoric times to the recent past The investigation of a representative sample of thesesites to attempt to refine our understanding of their chronological and functionalpatterning 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 projectmdasha programme of ethnoarchaeological research (Fig 44) 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 theyhave done so in the recent past Within the


Figure 44 Ethnoarchaeological survey the typical site of a winter bedouin tent (beit sharsquoar) in Wadi Faynan

Note Gullies to direct rainwater away from the cleared menrsquos and womenrsquos sections are clearly visible The menrsquos hearth is under the tent poles to the left of the photographic scale The womenrsquos 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 CPalmer

study area planning recently abandoned bedouin structures and analysing their floorsediments combined with interviewing the families who used the structures isestablishing archaeological signatures of seasonality and different age and gender groupsto inform our interpretations of the settlement archaeology

The final major component of the projectrsquos methodology is the development of a Geographical Information System integrating all the above data This is attempting torefine further our understanding of changing relationships between arable pastoral andindustrial activities between the lsquoeconomicrsquo lsquosocialrsquo and lsquoritualrsquo landscapes that are being defined and between all these cultural activities and the development of the naturallandscape which is the core issue of the projectrsquos research agenda

THE NEOLITHIC EARLY FARMING c9500ndash4000 BC

The transition from the Pleistocene to the Holocene or Postglacial the modern climaticera occurred at approximately 9500 BC Our geomorphological studies have established


that as elsewhere in the region the Pleistocene late glacial environment was cold anddry We then have strong and widespread evidence from sediment sequences and thefauna and flora within them that the early Holocene was characterized by a significantlywetter environment than today which lasted until about 6000 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 Holocenemdashthe culmination of trends in settlement subsistence and social change that can be observed amongst the Natufianpeoples of the late Pleistocene following the peak of glacial conditions c20000 years ago (Sherratt 1997) After about 15000 years ago there was a sudden dramatic warmingand the Natufians were able to develop semi-permanent settlements by lakes and springsin the Jordan valley (the lsquoLevantine corridorrsquo) Excavations show that Natufian settlements such as Jericho and Abu Hureyra were sustained by a combination of fishingfowling hunting (especially gazelle) collecting forest foods in the valley and gatheringwild cereals and other grasses on the steppelands above (Hillman 1996) With the returnto cold and dry conditions termed the Younger Dryas (11000ndash9500 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 thatpeople 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 wehave only lithic evidence (flint implements found on the surface) for the presence ofNatufian hunter-fisher-gatherers in the Wadi Faynan but significantly most of it has been found in the upper tributaries near the springs (Fig 45)

With the beginning of the Holocene c9500 BC there was a sudden return to warmerconditions coinciding with the first generally accepted evidence in the Levantine corridorthat 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 staplefoods augmented by fishing fowling and hunting a variety of game (Bar-Yosef and Kislev 1989 Byrd 1992) Sedentary mixed farming in which goat herding wascombined with cereal cultivation then developed throughout the Near East about 1000years later (the Pre-Pottery Neolithic B or PPNB c8500ndash6500 BC) coinciding with major changes in architecture (the appearance of substantial square or rectangulardwellings 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 latterprovided 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 butis also clear from current investigations of PPNA and PPNB settlements in the WadiFaynan A PPNA site of simple rubble shelters and pits has evidence for mixed huntinggathering and plant cultivation (Finlayson and Mithen 1998) The inhabitants of asubstantial PPNB settlement of well-built stone houses were mixed farmers growingwheat barley and legumes and herding domestic sheep and goats (Simmons and al-Najjar 1996) The two sites are only 100 m apart in the Wadi Ghuwayr at the junctionbetween the mountains and the main wadi by the spring where Natufians also camped(Fig 45 upper photograph) We have also found traces of similar settlements in the


upper Wadi Dana by the main spring there Although these first agricultural communitiesclearly preferred the well-watered upper tributaries for their primary settlements otherlithic scatters indicate that they also used the main wadi presumably for hunting andherding

By the sixth and fifth millennia BC the zone of principal arable settlement hadexpanded out into the main wadi Excavations a few years ago revealed a lateneolithicearly chalcolithic settlement of simple rectangular drystone houses at Tell WadiFaynan 1 km west of Khirbet Faynan (al Najjar et al 1990 Fig 45 lower photograph) Our geomorphological investigations show that when these people settled at Tell WadiFaynan the climate was significantly wetter than either before or afterwards there was amore or less perennial stream by the sitemdashthe archaeological sediments contain for example frustules of the diatom Navicula a freshwater organism and the pottery andmortar contain fragments of reeds and grass as well as straw The likelihood is that theprimary farming zone was able to expand from the tributaries to the main wadi floor atthis time because the climatic amelioration allowed farmers to exploit the seasonalfloodwaters of the main wadi for their crops with methods akin to those used by theearlier neolithic farmers in the upper tributaries Presumably they sowed their crops oneither side of what was then the Faynan stream after spring floods had soaked the soilson either side of its course


Figure 45 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 GBarker


THE BRONZE AGE c4000ndash1200 BC THE BEGINNINGS OF METALLURGY AND DRY FARMING

Aridification began to develop in the fifth millennium BC leading to the development ofa relatively steppic landscape by the fourth and third millennia BC the period of theEarly Bronze Age This was a period of immense social change in the Near Eastcharacterized 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 inWadi Faynan presumably for ornamental purposes and Faynan copper was exploited bychalcolithic people and traded with the surrounding region However the first clearevidence for the systematic mining of the copper ores and their processing at Faynansettlements is in the Early Bronze Age (Adams and Genz 1995 Hauptmann 1989Wright et al 1998) This was the context for the emergence of local elites who controlledcopper production and the exchange to other regions of smelted copper ores and finishedartefacts The research by the Bochum Mining Museum suggests that at first ores visibleat 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 thensmelted in smelting ovens located on the windward side of ridges near the settlementsOur geochemical analyses of sediments at Tell Wadi Faynan indicate that these smeltingactivities 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 agesettlements associated with field systems both on the southern side of the wadi within thearea demarcated by the later classical field system and in the small tributary wadis on thenorthern side One zone of the classical field system encompasses the most substantial ofthese settlements where excavations by Dr Karen Wright have revealed evidence forirregularly built drystone structures together with enclosures middens pits and storagebins and evidence of working smelted copper into ingots and finished artefacts (Wrightet 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 ofbronze age structures and field boundaries some of the latter terraced (Fig 46) They are associated with circular or oval cisterns 30ndash50 cm deep fed by short feeder walls The northern settlement zones include sequences of roughly built terrace walls and checkdams built across the shallow floors of tributary wadis with pottery in associatedsediment sequences indicating a bronze age date (Fig 47) 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 haveanalogies with the pastoral encampments of the Negev Chapter 3) and a few larger more complex


Figure 46 Part of the Wadi Faynan field system WF4 showing (above) the early bronze age and (below) the classical landscape in unit WF413

settlements near ridge-top spreads of bronze age slag and furnace waste presumed tohave hadmdashat least in partmdashan 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 butdiverse economic orientations agricultural pastoral and metallurgical Whereas neolithicfarmers in the Wadi Faynan were able to exploit well-watered locations bronze age


farmers were having to develop

Figure 47 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 thirdmillennia BC such as building walls to collect and trap seasonal floodwaters in storagecisterns 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 farmingyet found in the Near East making this another indicator of the social and economictransformations that characterized this phase of settlement in the region A degree ofpastoral specialization may have been another way in which bronze age societies wereable to respond to aridity whilst also being like metallurgy an indicator of complexeconomic structures of production and exchange What is also interesting is that we haveevidence for strong soil erosion through the second and first millennia BC andpalynological indicators that this reflects the impact of human activities on the landscapesuch as clearance of fuelwood for smelting and intensification in systems of cultivationand herding rather than climatic change

What happened in the Wadi Faynan during the second millennium BC (the LaterBronze Age) remains unclear as throughout the entire region The end of the millenniumwas marked by the widespread disintegration of urban polities throughout the EastMediterranean and Levant There is some evidence for climatic deterioration this is the


period of the Thera or Santorini eruption in the Aegean Seamdashand Egyptian scribes make references to raiding by lsquoSea Peoplesrsquo so there has been a lively debate aboutwhether lsquoexternalrsquo environmental or cultural factors such as these caused economic collapse or whether (more likely) they exacerbated internal processes of social changethat were already in train In the Levant it has commonly been argued that societiesturned to nomadic pastoralism at this time (Finkelstein 1995 LaBianca 1990) thoughconvincing archaeological signatures for such pastoralism are unclear (Bar-Yosef and Khazanov 1992) The debate is further obscured by Biblical archaeologists andhistorians looking for the nomadic peoples of Old Testament origin myths In theethnohistorical record furthermore specialized pastoralism is a highly complexeconomic system that invariably operates not in isolation but in close relationship withadjacent agricultural and urban systems as it may have operated in articulation witharable and metallurgical activities at Faynan in the Early Bronze Age It is thereforedifficult to separate absence of settlement evidence from evidence of settlement absencein 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 secondmillennium BC

THE IRON AGE c1200ndash300 BC TRANSFORMATIONS IN SETTLEMENT FARMING AND MINING

The early first millennium BC saw the emergence of iron age states west of the Jordanand 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 Younker1995) The State of Edom in which Wadi Faynan was situated was the least denselysettled of the three Although historians have accepted Josephusrsquo 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 aroundMadama 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 builtimmediately below its successor Khirbet Faynan at the strategic centre of the Faynanregion at the point where the three major tributary wadis come together to form the mainchannel of the Faynan We have also found zones of iron age settlement along thesouthern margins of the field system and on the northern side of the wadi The evidencesuggests therefore that the Edomite settlement system consisted of a few large anddiscrete habitation units probably organized hierarchically with WF424 the dominantsite Recent work in the neighbouring Wadi Fidan indicates that there may have beenother 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 asWF424 where we found thick deposits of slags charcoal-rich unlike the bronze age slags suggesting experimentation with new technologies to deal with the far largerquantities of ore being processed for the Edomite economy Geochemical analysesconfirm the increasing scale of smelting pollution


WF424 was associated with a field system of boulder-built walls often set orthostatically and also with substantial boundary walls built upstream of these fieldsalong the junction between the hill and the wadi floor These boundary walls collectedwater from the surrounding slopes and channelled it to exit sluices above the terracedfields so that maximum water flow could be directed down the central part of the fieldsystem Similar boundary walls enclosed iron age field systems on the northern side ofthe Wadi Faynan and in part at least they had a water-diversion function (Fig 47) We cannot be sure of the dating of these boundary walls but the fact that we have only foundthem enclosing field systems with significant iron age material and the constructionalsimilarities between the walls and the fields they enclose suggest an iron age date If thisdating is correct it implies that whereas bronze age farmers built simple terrace walls atright angles across wadi beds to check floodwater flow and to try to spread it laterallyover surrounding fields together with small catchments to collect water in cisterns ironage farmers in Faynan had learned to construct substantial and rather sophisticated wallsto divert the flow of floodwaters sometimes hundreds of metres from their natural line sothat far greater quantities of water could be collected and sent down a field system thanwas possible with bronze age technology

NABATEAN SETTLEMENT c300ndash63 BC

The Nabatean state with its capital Petra developed at the time that Romersquos power was expanding across the eastern Mediterranean in the last three centuries BC and flourisheduntil Palestine was annexed by Rome in 63 BC The Nabatean settlement system in thewadi like that of the Iron Age was dominated by one central settlement the communityat Khirbet Faynan This settlement presumably controlled copper production which onthe evidence of both mining archaeology and sediment geochemistry increaseddramatically in scale in the later centuries of the first millennium BC The landscape nowconsisted of a series of adjacent settlement units organized in some kind of hierarchicalrelationship with respect to Khirbet Faynan a series of large farmsteads of broadlyNabatean date on the southern slopes overlooked the classical field system excavationsby Wright et al (1998) found small buildings of Nabatean date in the area of majorbronze age settlement within the classical field system (Fig 46) 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 thatrun parallel to the main wadi along its southern side (Fig 48) Water was dammed as it issued from the adjacent hills diverted westwards by boulder walls along the contour ofthe slope and then channelled through simple sluices (gaps) and spillways (steppedstructures) onto terraced fields below Nabatean technology on the southern slopes mayalso have included channels formed by parallel walls that fed water directly into the fieldson either side through sluice gaps


THE ROMAN IMPERIAL LANDSCAPE 63 BC-cAD 600

With the Roman annexation of Palestine Faynan became one of the principal suppliers ofcopper 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 toChristian slaves being sent to work the mines as punishment (according to some sourcesthey were crippled to prevent their escape by blinding having a hand removed or havingtheir Achilles tendon severed) though most of the mining and smelting was probablydone by a workforce paid well for skilled and dangerous work The quantities of orebeing 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 metalsfound in a 2500-year long sequence of sediments behind a barrage at Khirbet Faynanindicate extraordinary levels of air pollution in Roman and Byzantine times with levelsseveral times lethal in terms of modern pollution criteria (Fig 49)

Figure 48 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 42 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


The satellite farms were abandoned leaving Khirbet Faynan as the single dominant settlement Our studies of field layout and construction and of the surface materialindicate that the entire agricultural landscape was now managed as a more or less integralunit or estate Systems of long parallel walls were built to divert water from the mainwadi into adjacent fields on low-elevation terrain and from the southern tributary wadis (Fig 410) Further down the tributary wadis similar channels were built at c45 degrees to water flow to collect any water that had bypassed the higher diversion walls or drainedback into the wadis from the higher terraced fields to force it once more onto adjacentcultivable land The effectiveness of the system is in part explained by the uniformly lowlevels of water infiltration we have found at sample sites from the upper slopes to thelowest fields but organizational factors were also important The field evidence supportsthe hypothesis of cooperation between areas of the field system fed by the parallelchannels rather than farmers with land upslope having exclusive access to thefloodwaters of particular wadis at the expense of other farmers with land

Figure 49 The distribution of copper (in parts per millionmdashppm) through the sediments that accumulated behind the Khirbet Faynan barrage

Note The sequence extends from c2500 years ago (far right) to the present day (far left)

further down the direction of flow the internal linkages between the system imply thatwater resources were shared down the length of the field system The construction ofmajor parallel channels to carry floodwater through the system demonstrates the sameengineering skills in moving water relatively long distances over gentle gradients as aredisplayed 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 thereservoir and ore-crushing mill near Khirbet Faynan


This imperial landscape was highly organized with large-scale industrial processing sustained by an integrated agricultural and hydraulic system But it was also increasinglybarren fast eroding and grossly polluted Charcoal samples from the smelting sitesstudied by the German team show that whereas Nabatean miners cut local firewood fortheir smelting activities by the Roman period timber was having to be brought downfrom the plateau because local supplies had been exhausted (Engel 1993 Hauptmann1992) Our pollen evidence indicates the same process of humanly induceddesertification by the time of Christ the landscape consisted of very degraded steppelandand this degradation then accelerated significantly through the first millennium AD Bythe end of the Roman period the steppic component of the pollen diagrams collapsesevidence of olive cultivation disappears and signs of cereal cultivation drasticallyreducemdashthe flora at this time was analogous to the modern pollen rain in the Dead SeaWe have also found widespread evidence that Roman farmers were trying to combat theeffects of wadi-downcutting in


Figure 410 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 GBarker

their alterations to the floodwater farming systems though these were ultimatelyineffective (the main wadi now flows at least 5 m below the parallel channel systems thatdiverted the Faynan floodwaters into the ancient fields)


THE POST-BYZANTINE LANDSCAPE

The desertic environment has persisted to the present day though there is evidence for anepisode of even greater aridity in the period cAD 1600ndash1850 The nature of the Islamic and later settlement systems following the eventual abandonment of Khirbet Faynanremains unclear but it is at least evident from our survey work so far that it was notcharacterized by a renewal of substantial settlement within the field system zone andthere are indications of ephemeral settlements on the surrounding slopes akin to the sitesof recently abandoned Bedouin encampments Small-scale increases in smelting pollutants in levels above the post-Byzantine collapse (such as the peak at c90 cm depth in Figure 49) indicate a revival of industrial activity at some time probably in the early second millennium AD (radiocarbon dates are awaited) and the range of pollutantssuggest the reworking of Roman and Byzantine slag deposits rather than renewed miningon 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 inthe region today

CONCLUSION

As described above we are beginning to detect oscillations in environmental changesettlement forms and agricultural and industrial activity over the past 10000 years If wecan understand how they do and do not inter-relate we should be able to write alandscape history in the Braudelian sense of complex interactions between short-term processes medium-term processes and the longue dureacutee 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 inextremely 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 ofland use from the Late Neolithic onwards a number of increasingly sophisticated systemsof water control can be discerned but again it is clear that there is no simple progressionin land use from simple to complex but rather oscillations between the two Anothercomplex 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 nosignificant impact on the landscape but it is possible that the erosion we can detectduring and after the Early Bronze Age whilst probably mainly a response to aridificationpartly reflects poor land management techniques Given the evidence of the geochemistryfor the beginning of smelting pollution at this time wood cutting for metallurgicalprocessing may also have been a factor However that may be it is clear that the demandsof Nabatean and in particular RomanByzantine mining in parallel with the intensiveagricultural practices developed to feed the workforce had an ultimately devastatingimpact on the landscape Whether or not climatic change was also a factor it is clear that


large-scale stripping of the landscape of vegetation made it extremely vulnerable toerosional forces

The geochemical evidence also demonstrates that the effects of Roman and Byzantine mining and smelting are still felt todaymdashthe milk urine and faeces of the bedouinsrsquo goats today have significant levels of heavy metals from grazing the polluted ground andcereal growth is also badly affected around the smelting sites Does the extraordinarydensity 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 strongpossibility (currently being tested by skeletal analysis) that the health of the Roman andByzantine population in particular was directly affected from inhalation skincontamination and bioaccumulation of polluted animal and plant foods Whilst thecollapse of intensive farming and mining in Late RomanByzantine times no doubt in partreflects changing economic relations between Faynan and the wider world it seemsinescapable that the activities of these farmers and miners had a profound impact on theirlandscape (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 oftheir inter-relationships indicates something of the potential complexity of the interplaybetween 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 throughGeographical Information Systems (GIS) we should be able to integrate our findings onhow the landscape has changed and the role of people in this to investigate also howdifferent populations in the past perceived their changing landscape and their place withinit It should be possible for example to model the spatial characteristics of air pollutionat different distances from the major smelting settlements and its different effects onsurrounding populations There is much still to be learned about the extent to whichfarmers using the field system operated independently or collaborated and in the lattercase whether they did so by cooperation or coercion There is also the question of howthe sacred and the secular related to one another in different periods There is intriguingevidence for example that whereas bronze age people kept their cemeteries and fieldsapart classical farmers deliberately constructed water diversion systems so that theyincorporated pre-existing burial cairns at key nodal points Reconstructing landscapehistories needs natural scientists to analyse changing forms of landscape andarchaeologists to analyse changing settlement morphologies and systems We hope thatthe effective partnership of disciplines in our project will allow us ultimately not just todescribe these two landscape histories for Faynan but also to understand theirinteractions looking at the perceptions and choices that underpinned human actions inthis landscape and shaped the latter with ultimately devastating consequences

ACKNOWLEDGEMENTS

This chapter represents the work of the Wadi Faynan Landscape Survey team especiallyRAdams (University of Bristol prehistoric pottery analysis) OCreighton (University ofExeter archaeological survey) PDaly (University of Oxford archaeological survey and


GIS analysis) DDGilbertson (University of Bournemouth geomorphology) JPGrattan(University of Wales Aberystwyth geomorphology geochemistry) COHunt(University of Huddersfield palynology) DJMattingly (University of Leicesterarchaeological survey) SJMcLaren (University of Leicester geomorphology)HMohammed (University of Benghazi Libya palynology) PNewson (University ofLeicester archaeological survey and GIS analysis) CPalmer (University of Leicesterethnoarchaeology) HParton (British Institute at Amman for Archaeology and HistoryRoman and post-Roman pottery analysis) FBPyatt (Nottingham Trent University environmental biology geochemistry) TEGReynolds (Cambridgeshire CountyCouncil lithic analysis) HSmith (University of Bournemouth ethnoarchaeologyenvironmental archaeology) RTomber (Museum of London Roman and post-Roman pottery analysis) and ATruscott (University of Wales Aberystwyth OSL dating)mdashtogether with the archaeological field team that has done most of the foot work Gratefulthanks are also due especially to the Arts and Humanities Research Board the Councilfor British Research in the Levant the Natural Environment Research Council theSociety of Antiquaries of London and the Universities of Leicester Huddersfield andAberystwyth for funding the project and to the Jordanian Department of AntiquitiesCBRLrsquos British Institute at Amman for Archaeology and History and the Royal Societyfor the Conservation of Nature for other essential support

REFERENCES

Adams R and Genz H (1995) Excavations at Wadi Fidan 4 a chalcolithic villagecomplex in the copper ore district of Feinan southern Jordan Palestine Exploration Quarterly 1278ndash20

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 OH Gilbertson DD Hunt CO Mattingly DJ McLaren SJand Thomas DC (1997) The Wadi Faynan Project southern Jordan a preliminaryreport on geomorphology and landscape archaeology Levant 29 19ndash40

Barker G Adams R Creighton OH Gilbertson DD Grattan JP Hunt COMattingly DJ McLaren SJ Mohammed HA Newson P Reynolds TEG andThomas DC (1998) Environment and land use in the Wadi Faynan southern Jordanthe second season of geoarchaeology and landscape archaeology (1997) Levant 305ndash26

Barker G Adams R Creighton OH Crook D Gilbertson DD Grattan JP HuntCO Mattingly DJ McLaren SJ Mohammed HA Newson P Palmer C PyattFB Reynolds TEG and Tomber R (1999) Environment and land use in the WadiFaynan southern Jordan the third season of geoarchaeology and landscapearchaeology (1997) Levant 31255ndash92

Barker G Adams R Creighton OH Daly P Gilbertson DD Grattan JP HuntCO Mattingly DJ McLaren SJ Newson P Palmer C Pyatt FB ReynoldsTEG Smith H Tomber R and Truscott AJ (2000) Archaeology and


desertification in the Wadi Faynan the fourth (1999) season of the Wadi FaynanLandscape Survey Levant 3227ndash52

Bar-Yosef O (1995) Earliest food producersmdashpre-pottery neolithic (8000ndash5500 BC) In TLevy (ed) The Archaeology of Society in the Holy Land 190ndash204 London Leicester University Press

Bar-Yosef O and Belfer-Cohen A (1992) From foraging to farming in theMediterranean Levant In ABGebauer and TDPrice (eds) Transitions to Agriculture in Prehistory 21ndash48 Madison Prehistory Press Monographs in World Archaeology4

Bar-Yosef O and Khazanov A (1992) (eds) Pastoralism in the Levant Archaeological Materials in Anthropological Perspectives Madison Prehistory Press Monographs inWorld Archaeology 10

Bar-Yosef O and Kislev ME (1989) The pre-pottery neolithic B period in eastern Jordan Paleacuteorient 15 (2)150ndash6

Byrd B (1992) The dispersal of food production across the Levant In ABGebauer andTDPrice (eds) Transitions to Agriculture in Prehistory 49ndash61 Madison Prehistory Press Monographs in World Archaeology 4

Byrd D (1994) Public and private domestic and corporate the emergence of thesouthwest Asian village American Antiquity 59 (4)639ndash66

Engel T (1993) Charcoal remains from an iron age copper smelting slag heap at FeinanWadi Arabah (Jordan) Vegetation History and Archaeobotany 2205ndash11

Finkelstein I (1995) Living on the Fringe The Archaeology and History of the NegevSinai 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) EpipalaeolithicProject report on reconnaissance survey 14ndash22 April 1996 Levant 30 27ndash32

Gophna R (1995) Early bronze age Canaan some spatial and demographic observationsIn TLevy (ed) The Archaeology of Society in the Holy Land 269ndash80 London Leicester University Press

Hauptmann A (1989) The earliest periods of copper metallurgy in Feinan Jordan InAHauptmann EPernicka and GAWagner (eds) Archaemetallurgie det Alten WelttOld World Archaeometallurgy 119ndash36 Bochum Deutsches Bergbau-Museum Der Anschnitt Beiheft 7

Hauptmann A (1992) FeinanWadi Feinan American Journal of Archaeology 96510ndash12

Hauptmann A and Weisgerber G (1987) Archaeometallurgical and mining-archaeological investigations in the area of Fainan Wadi lsquoArabah (Jordan) ADAJ31419ndash37

Hauptmann A Begemann F Heitkemper E Pernicka E and Schmitt-Strecker S (1992) Early copper produced at Feinan Wadi Araba Jordan the composition of oresand copper Archaeomaterials 61ndash33

Hillman GC (1996) Late Pleistocene changes in wild plant-foods available to hunter-gatherers of the northern Fertile Crescent possible preludes to cultivation In DRHarris (ed) The Origins and Spread of Agriculture and Pastoralism in Eurasia 159ndash203 London UCL Press


Kirkbride D (1966) Five seasons at the pre-pottery neolithic village of Beidha in Jordan Palestine Exploration Quarterly 988ndash72

LaBianca OslashS (1990) Hesban 1 Sedentarization and Nomadization Berrien Springs (MI) Andrews University Press

LaBianca OslashS and Younker RW (1995) The kingdoms of Ammon Moab and Edomthe archaeology of society in late bronze ageiron age Transjordan (ca1400ndash500 BCE) In TLevy (ed) The Archaeology of Society in the Holy Land 399ndash415 London Leicester University Press

Levy T Adams R and Shafiq R (1999) The Jebel Hamrat Fidan Project excavationsat the Wadi Fidan 40 cemetery Jordan (1997) Levant 31293ndash308

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 3427ndash56

Sherratt AG (1997) Climatic cycles and behavioural revolutions the emergence ofmodern humans and the beginning of farming Antiquity 71271ndash87

Simmons AH and al-Najjar M (1996) Test excavations at Ghwair I a neolithic settlement in the Wadi Feinan ACOR Newsletter 827ndash8

Wright K Najjar M Last J Moloney N Flender M Gower J Jackson NKennedy A and Shafiq R (1998) The Wadi Faynan Fourth and Third MillenniaProject 1997 report on the first season of test excavations at Wadi Faynan 100 Levant3033ndash60


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 largescalesettlement 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 (Tate199755 Fig 51) This region is within the 200 mm rainfall isohyet which is theaccepted limit for dry farming without irrigation Especially intriguing however isevidence for apparently permanent settlements of the same antiquity on the desert plateaubeyond in the region long called the al-Harra (lsquoBurnt Landrsquo in Arabic) by the bedouin and the Black Desert by European travellers (Braemer et al 1996b1) The water management features associated with some of these sites provide the focus of thischapter A number of reasons to explain the development of permanent settlement in theHarra in the Roman period has been suggested ranging from significant changes inclimate (that is increased rainfall compared with today) to the sedentarization ofpreviously nomadic tribes Certainly the new pattern of settlement did not endure by theLate Byzantine and Early Islamic periods (the seventh to ninth centuries AD) most of thesettlements 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)mdasha huge area of arid plateau (Hamada) forming the northern part of the Arabian steppe bounded by more fertile regions to the west (the Mediterraneanlittoral) the north (the Taurus foothills) and the east (the River Euphrates) and by theNafudh desert to the south The average annual rainfall within this region declines withlatitude from c150 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 kmwide and reaching 250 km in length towards the southeast making it difficult to traverse(Helms 198119) The lava emanated from a number of fissures in at least six successiveflows over a relatively short period of time These flows solidified into layers of hardbasalt

Figure 51 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 offissure cones and dissected by a series of wadis generally flowing in an easterly directionThe 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


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 thethird 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 thatcurrent evidence suggests were used at dryland settlements in the Harra during theRoman period taking three sites as case studies where I have conducted fieldwork Thefirst of these sites ad-Diyatheh is located on the western edge of the Harra near the steepdescent 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 wadiThe third site Qasr Burqursquo is on the eastern edge of the Harra near its junction with the Ruhba the large fertile alluvial plain that in season lsquohas provided highly-prized grazing for nomads from time immemorialrsquo (Braemer et al 1996b1)

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 52) Ad-Diyatheh lies at the junction between thesettled Hauran in the west and the Hamada steppelands to its east This demarcation linealso echoes the important 100 mm precipitation isohyet which follows the relief edge ofthe 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 wascarefully chosen being a small flat plateau above the north bank of the wadi withcommanding views across the Harra

Clustered around the fort and along the edge of the north bank of the wadi are theremains of c100 stone-built structures all of a similar construction and with ahomogeneous layout Many of these houses are in a very good state of preservation somehave underground rooms with extensions excavated into the wadi bank It has beenconvincingly suggested that this collection of structures forms a village ofcontemporaneous construction (Villeneuve 1986) Apart from a walled meeting orfunction area no public building of any other sort has been identified and there seems tobe 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 suchreservoirs However in the wadi bed south of the village and fort and continuing up ontothe right bank can be seen the entrances to some

Differing strategies for water supply and farming 87

Figure 52 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 53)

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 cattlerearing as in the areas of the limestone massif in northern Syria (Tate 1992 Villeneuve1986) As the French survey of the village was coming to a close it was realized that theagricultural operation was on a very large scale Initially it had seemed like a simplediversion of wadi floodwater onto the Harra plain below the fort and village but furtherinvestigation revealed the remains of a complex floodwater farming system together withother significant features such as watermills built onto leats extending out onto the Harraplain on the northern side of the wadi course

Sadler (1993) outlined the main features of these field systems in the following terms (Fig 52) 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 2and 3 km downstream (Sadler 1993428) The first is by far the best preservedconstituting a long low barrage of medium-sized stones crossing the wadi bed at anoblique angle to the flow The barrage leads directly into a long straight canalconstructed on a shallow gradient which would have allowed the momentum ofwaterflow to overcome the height difference between the Harra plain and the wadi bedseveral


Figure 53 The Roman-period settlement of Ad-Diyatheh and its channel walls in the Wadi Sham viewed from the west

Photograph PNewson

metres below Although the networks have been eroded somewhat by the action offloodwaters and neglect the course of the main canal can be identified as running moreor less in a parallel direction to the main wadi As it approaches the plain a succession oftributary canals leads off it to feed water into the irrigated zones This process is repeatedwith the other diversion barrages and their associated primary secondary and tertiarycanals further downstream Two secondary canals in the first network feed their own self-contained 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 feeda 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 asfar as the next wadi coming down from the Jebel the Wadi Gharaz Running from thiscanal in an easterly direction and at regular intervals along its course is a succession ofparallel 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 primarysecondary and tertiary canals amounts to around 1200 ha (Sadler 1993431) At the endof 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 wastedor allowed to erode the canals or fields by ponding

Around the slopes to the north of the village along the edges of a shallow valleyleading off from the Jebel evidence for another water-capture strategy has been recorded


This valley is at too high an altitude to be irrigated by the canals and received too littlefloodwater to allow cereals to be cultivated Nonetheless there are remains of stone-cleared fields and the vestiges of long low parallel walls some leading down the slopesat regular intervals and others at lower levels forming low terraces (Sadler 1993433)This complex appears to represent a different strategy of water management concernedwith 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 beendocumented 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 1993434)

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 thewinter cereals that were grown here on a large scale in Roman times and are stillcultivated by the local bedouin when the conditions allow The remains of eight suchmills have been located so far All display a similar construction and are usually locatedalong 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 canbe found situated amongst the canal networks of the Harra plateau have led Sadler(1993435) to suggest that the field system and the village must be contemporary From acursory assessment of the pottery from the remains of the fort and village Villeneuve(1986713) concluded that the main phase of occupation lasted from the late third centuryAD to the fifth and perhaps even into the seventh centuries but this date range is by nomeans certain

Al-Namara

The site of al-Namara (Fig 54) 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 SaadThe name al-Namara refers to the large basin etched into the plateau by the confluence ofthese two wadis at the basinrsquos centre is an lsquoislandrsquo of resistant rock the remains of a volcanic (basalt) plug Water flowing down from the Jebel in the spring floods poolsnaturally at points in the wadi beds around this island and lasts well into the summerTherefore the combination of ample water in an otherwise arid region and the strategicvantage point of the island has long attracted local pastoralists as well as the attention ofregimes trying to control them On the flat top of the island or lsquocitadelrsquo are the few remains of reused structures of the Roman army and a later Arab occupation whilst in thesurrounding basin and beyond are the extensive remains of water catchment systems andencampments (Figs 55 and 56) Little archaeological investigation has taken place at the site apart from an initial assessment and topographical plan of the citadel and itsimmediate environs within a 2times1 km area completed while work was being done to


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 smallalluvial plains 500ndash1000 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 mlong by 400 m wide The island intrudes into this plain rising to around 10 m in heightThe wadi has to curve its course round the island to the south as a result of which thesouth bank of the curve has gentle slopes whereas the north bank forms a cliff some 2ndash8 m high Downstream the banks lower to around 1 m in height and the distance betweenthem widens into a plain up to 1 km wide The northern terrace of the plain hassubstantial evidence for simple irrigation systems up to 1 km long whereas the southernside has more limited evidence (Braemer et al 1996b4)


Figure 54 Plan of the main irrigated zone along the Wadi Sham by the Roman-period settlement of al-Namara

Source After Braemer et al 1996b


Figure 55 View of the main irrigated zone along the Wadi Sham at al-Namara looking north from the lsquocitadelrsquo

Photograph PNewson

Figure 56 Canal 3 at al-Namara viewed from the east Photograph PNewson


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 asimilar construction consisting of a line of large othostatic boulders positioned on anoutcrop of hard basalt placed at an oblique angle to the direction of flow of wadifloodwaters The boulders were bound together with smaller stones placed around themThe low dams thus formed obstructed the wadi course only partially in order to capturewater in a controlled manner without being liable to being destroyed by the force of theflood (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 1996b9) The canals that lead off from these barrages seem to have served two purposes The twocanals 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 southof the island where the water naturally pools Downstream of the island two barragescapture water from the wadi and their associated canals lead it to areas where fields havebeen cleared of stones and laid out for irrigation

The first canal (Canal 3 on Figure 54) lies on the northern terrace of the wadi and isaround 25 km long The length of the other canal (Canal 4) cannot be measuredaccurately because it has been partially obscured by the new reservoir The area irrigatedby the first canal lies immediately adjacent to the wadi forming a small terrace up to 100m in width and roughly 2 km long This canal follows a similar pattern in its constructionto all the other canals with a channel way 1ndash2 m wide and a wall 05ndash1 m high of medium-sized boulders edging the downslope of the channel which follows the contourof the terrace bank (Fig 56) The wall was made waterproof with the packing of smaller stones and earth in the gaps between the larger stones The floodwaters flowed behindthis low wall and were let into the fields below at certain points by the means of simplespillways At the end of the main canal is a small secondary stretch which stops theremaining water from entering back into the wadi redirecting it back into the irrigatedarea There are two further canals along the course of the Wadi Saad these are short inlength (250 m and 800 m respectively) and seem to have been constructed to capturewater to irrigate small stone-cleared fields immediately adjacent to this wadi

Qasr Burqursquo

The site of Qasr Burqursquo 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 basinwhere water from surface and subsurface run-off from the surrounding slopes naturally ponds (Betts et al 19906 Fig 57) The most significant structure is a tall rectangularstone tower rising to a height of c5 m (Fig 58) the study of its arches and floor levels suggested that it may have reached a height of 13 m (Betts et al 199116) The tower is surounded by a series of rooms forming a rough square which


Figure 57 The ancient reservoir at Qasr Burqursquo Source After Kennedy and Riley 1990

Figure 58 Air photograph of Qasr Burqursquo and its reservoir looking southeast

Photograph BBewley


appears to have been constructed at a later date An analysis of the architectural details ofthe tower and surrounding structures along with the evidence from two burialinscriptions written in Greek led Svend Helms to suggest the site to be a monasticfoundation dating to the Late Antique and Early Islamic periods (Betts et al 199116) Potsherds found within the structure for the most part date to the Late Antique and EarlyIslamic periods and initially it was thought that the tower was probably constructed atthis time as the residence for a recluse (Gaube 1974) However earlier Roman potteryhas also been found in the area (Betts and Helms 19898 Betts et al 199122) and there is a strong possibility that the Roman army established themselves at the site earlierincluding perhaps constructing the tower (Kennedy and Riley 1990) Later this towercould have served as the focus for a monastery that grew up around it and changed itsfunction to that of a watch tower and possible refuge from hostile nomads Alternativelythe tower and enclosure buildings may have had a military purpose for the Arabgovernment in the sixth-century AD Ghassanid period (Betts et al 199117)

Whatever the function of these buildings it was vital for the occupants to secure anadequate 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 valleyPresumably the quantity of water collected behind this natural barrier was not sufficientto 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 topof the low ridge that was responsible for the natural pooling of floodwater within thewadi bed (Fig 57) This dam the lower courses of which still survive was composed oflarge-sized roughly dressed basalt boulders laid in a series of stretcher courses cappedby one of headers Two such walls were built about 10 m apart following the top of theridge the space between probably being filled with earth and rubble Along the sidefacing the water is evidence for plastering which together with the form of wallconstruction indicates that the dam is probably of a similar date to the tower and itssurrounding enclosure (Betts et al 199112) The edges of the reservoir are lined withlow roughly-coursed stone walls from which two short stone staircases lead down to thereservoir one on each side of it

No evidence has been found for a sluice gate of any type within the wall of the damwhich suggests that the water was not used to irrigate a network of fields in the manner ofthe impressive dam and field network further to the north at Qasr el-Gherbi near Palmyra (Schlumberger 1986) However one of the lsquoroomsrsquo of the enclosure (room 11) has been tentatively identified as a windmill (Betts et al 199117) an interpretation that if correct clearly implies that cereals were being grown in the vicinity of the settlementAround the dammed lake and its associated buildings are huge numbers of encampmentsburial cairns and corral remains along with scatters of artefacts from all periods in amanner similar to that at al-Namara Given the location of the settlement and its verylarge facility for water storage the likelihood is that the main function of Qasr Burqursquo was as an oasis settlement serving a series of northmdashsouth and eastmdashwest desert routes (Betts et al 19906)

No evidence for field systems has been found by the recent survey However in thenearby more fertile and stone-free Ruhba there are lsquofields of barley planted for animal fodder and watered by flood irrigationrsquo (Lancaster 1981 Helms 1989) beside the fixed


bedouin encampments of Feytha arRisha al-Fawq and ar-Risha al-Taht It is likely that any permanent population living at Qasr Burqursquo whether Roman or Arab was served byequivalent field systems to those situated in the Ruhba

DISCUSSION

On first consideration all three sites exhibit an attempt to impose Roman control on theperipheral regions adjacent to the settled provinces of Arabia and Syria The exercising ofcontrol would have been specifically aimed at the local populations of pastoralists thetraffic of caravans and the activities of occasional groups of bandits (Isaac 1990) Thistook the form of strategically-placed military structures at all three sites where watercould be collected and which would have attracted the travellers and pastoralists of theregion 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 substantialsettlements did occur At ad-Diyatheh this took the form of a village with substantialpermanent stone buildings At al-Namara and Qasr Burqursquo the settlements mostly consist of stone-cleared corrals for temporary structures such as tents but this does notnecessarily 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 locationof their campsites or they could have been occupied for substantial periods of the year bydifferent family groups of nomadic pastoralists

All the sites exhibit measures for the control and use of limited supplies of water forboth drinking and agricultural purposes but the methods utilized were on different levelsand scales At ad-Diyatheh varied and sophisticated techniques were employed forfloodwater farming The systems for floodwater farming at al-Namara are simpler and smaller At Qasr Burqursquo even though floodwater farming systems have not yet been identified cereal cultivation is implied by the probable presence of a mill and the sitecertainly displays an impressive scale of planning for the storage of much largerquantities 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 orboth were dealing with different issues in building the settlements and their associatedstructures Local environmental considerations were undoubtedly one important factoraffecting 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 othertwo sites requiring substantial and carefully designed structures to allow some of thiswater 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 theconstruction of a large extensive network of irrigation channels and a large dispersedspread 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 especiallyimpeded by the fact that at this point the wadi bed lies up to 15 m below much of thesurrounding plateau so irrigation is really effective only along areas of adjacent


floodplain On the other hand Qasr Burqursquo lies in a natural dip within the surroundingplateau with the majority of water collecting here through groundwater seepage(Lancaster and Lancaster 1999132)

A second area of differences though one harder to identify with the archaeologicalevidence available relates to the social and economic conditions of the inhabitants of thethree 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 uponthe 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 theprovision of accommodation for livestock below the family rooms (Villeneuve 1986)The reliance on crops for subsistence would have been much less for the pastoralpopulations postulated for al-Namara and Qasr Burqursquo and so the need for large farming systems proportionally less though the surrounding Harra could have supported onlylow-density populations

At both al-Namara and Qasr Burqursquo it may well have been only the residents of thesubstantial permanent structures be they soldiers or monks who attempted to grow cropson any scale The pastoral populations at these settlements may have grown cereals onany scale only when favourable climatic conditions allowed so would not have beeninclined to spend valuable time in investing in the substantial infrastructure requiredbeforehand using the locations simply as convenient watering places (Macdonald 1993)Caution is needed in advancing such an hypothesis however for the modern bedouin ofthe region in fact try to irrigate areas when settled in a particular location over a period oftime (Lancaster and Lancaster 1999132ndash66) It is probably reasonable to assumethough that the occupants of the substantial structures were generally more reliant onprovisions provided from the immediate area given that pastoralists would have had thecapacity to move to new areas when they had exhausted supplies at any particularlocation

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 watercontrol 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 atad-Diyatheh compared with the more refined architecture of the adjacent Jebel villagessuggests that ad-Diyatheh was built by lsquosedentarized nomadsrsquo (Villeneuve 1986710) 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 QasrBurqursquo may well be a later enhancement of a much older catchment system

Whatever the origins of these systems though the water management techniquesemployed 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 stronglysuggesting that these new approaches to agriculture came as a direct consequence ofpolitical and social changes brought about by the imposition of Roman hegemony Theextent to which these new agricultural regimes were imposed created and operated by theRoman army or reflect a variety of complex responses by indigenous populations to theopportunities of Romanization still remains unclear However as the precedingdiscussion has shown the likelihood is that the archaeology of the Syrian Black Desert


reflects both the external imposition of and indigenous adaptions to the forces of Romanimperialism

ACKNOWLEDGEMENT

I would like to thank Bob Bewley for permission to reproduce as Figure 58 his airphotograph of Qasr Burqursquo and its reservoir from his Aerial Archaeology in Jordan Project

REFERENCES


Betts A (1993) The BurqursquoRuweishid Project preliminary report on the 1991 fieldSeason Levant 251ndash11

Betts AVG and Helms SW (1989) A water harvesting and storage system at Ibn al-Ghazzi in eastern Jordan Levant 213ndash11

Betts A Helms S Lancaster W Jones E Lupton L Martin L and Matsaert F(1990) The BurqursquoRuweishid Project preliminary report on the 1988 field season Levant 221ndash20

Betts A Helms S Lancaster W and Lancaster F (1991) The BurqursquoRuweishid Project preliminary report on the 1989 field Season Levant 237ndash28

Braemer F Echallier J-C and Taraqji A (1996a) Khirbet el Umbashi (Syrie) Rapport preacuteliminaire sur les campagnes 1993 et 1994 Syria 73117ndash29

Braemer F Echallier J-C Hatoum H and Macdonald MCA (1996b)Archaeological and Epigraphic Rescue Survey at Al-Nemara Report on the First Season Sept-Oct 1996 Damascus Department of Antiquities and Museums of Syriaunpublished report

Evenari M Shanan L and Tadmor N (1971) The Negev The Challenge of a DesertCambridge Mass Harvard University Press

Evenari M Shanan L and Tadmor N (1982) The Negev The Challenge of a DesertCambridge Mass Harvard University Press second edition

Gaube H (1974) An examination of the ruins of Qasr Burqursquo Annual of the Department of Antiquities of Jordan 1993ndash100

Helms SW (1981) Jawa Lost city of the Black Desert London Methuen Helms SW (1989) Jawa at the beginning of the Middle Bronze Age Levant 21 141ndash68 Isaac B (1990) The Limits of Empire The Roman Army in the East Oxford Oxford

University Press Kennedy D and Riley D (1990) Romersquos Desert Frontier From the Air London

Batsford Lancaster W (1981) The Ruwala Bedouin Today Cambridge Cambridge University

Press


Lancaster W and Lancaster F (1999) People Land and Water in the Arab Middle EastEnvironments and Landscapes in the Bilacircd ash-Shacircm Studies in Environmental Anthropology Volume 2 Amsterdam Harwood Academic Publishers

Macdonald MCA (1993) Nomads and the Hawran in the Late Hellenistic and Romanperiods a reassessment of the epigraphic evidence Syria 70303ndash413

Sadler S (1993) Le terroir agricole de Diyateh lrsquoirrigation comme condition drsquoexistance de ce terroir In BGeyer (ed) Techniques et Pratiques Hydro-Agricoles Traditionnelles en Domaine Irrigueacute Approche Pluridisciplinaire des Modes de Culture avant la Motorisation en Syrie (Actes du Colloque de Damas 27 juin-1 juillet 1987) 421ndash51 Paris PGeuthner Bibliothegraveque Archeacuteologique et Historique 136

Schlumberger D (1986) Qasr el-Heir el-Gharbi Paris PGeuthner Bibliothegraveque Archeacuteologique et Historique 120

Tate G (1992) Les Campagnes de la Syrie du Nord di IIe au VIIe Siegravecle Un Exemple drsquoExpansion Demographique et Economique a la Fin de lrsquoAntiquiteacute Paris PGeuthner Bibliothegraveque Archeacuteologique et Historique 133

Tate G (1997) The Syrian countryside during the Roman era In SEAlcock (ed) The Early Roman Empire in the East 55ndash70 Oxford Oxbow Books

Villeneuve F (1986) Ad-Diyatheh village et castellum romains et byzantins a lrsquoest du Jebel Druze (Syrie) In PWMFreeman and DLKennedy (eds) Defence of the Roman and Byzantine East Volume II 697ndash715 Oxford British Archaeological Reports International Series 297


6 Irrigation agriculture in Central Asia a long-

term perspective from Turkmenistan MARK NESBITT AND SARAH OrsquoHARA

INTRODUCTION

Agriculture in the newly independent republics of the former Soviet Central Asia isalmost entirely dependent on irrigation Consequently access to water is essential and ithas long played an important role in the social environmental economic and politicalsituation of the region Today as in the past agriculture represents the single mostimportant economic activity throughout the region and currently over 40 per cent of thepopulation is employed in the commercial agricultural sector with the vast majority ofCentral Asians either partially or wholly dependent on subsistence agriculture Theagricultural sector throughout Central Asia however is under threat because of the rapiddeterioration in the water distribution and irrigation since the collapse of the SovietUnion (OrsquoHara in press)

Central Asia boasts a long history of irrigated agriculture but the exploitation of theregionrsquos 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 irrigationprojects were constructed to satisfy Moscowrsquos continual demands for cotton In order to maximize agricultural output water was taken from areas of surplus to those of deficitoften involving transfers over considerable distances and in some case from otherrepublics Today however this huge highly integrated network serves five independentstates each following its own agenda for reform The implications for the regionrsquos water resources are immense and it is becoming increasingly difficult to reach a consensus onhow the water distribution and irrigation system should be managed and maintained(Bedford 1996 OrsquoHara in press) Further complicating the matter is the fact that CentralAsiarsquos irrigation zones are plagued by secondary salinization and high water tables (OrsquoHara 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 indeedcould be exacerbated by changing land and agricultural policies coupled with anincreased demand for water as population rises Should the system fail the consequences would be enormous and could ultimately undermine regional security The question ofsustainable irrigation is therefore urgent Given that Central Asia not only has a longhistory of irrigated agriculture but has witnessed the rise and fall of a number of majorempires over the last few thousand years it may well be that lessons can be learned fromthe past An assessment of former irrigation and water management practices mayhighlight whether sustainable irrigation is a feasible option and if so how it might be

achieved Here we review the history of settlement agriculture and irrigation over some 8000

years in southern Turkmenistan (Fig 61) A large body of archaeological evidence isavailable for this region much of it resulting from the establishment of the SouthTurkmenistan Multi-Disciplinary Archaeological Expedition (YuTAKE) in 1946 Many of its publications were not widely distributed even within the former Soviet Union butwe have been able to draw on a wide range of useful syntheses published in westernjournals A more recent phase of fieldwork involving a number of international researchteams has resulted in a series of renewed excavations at several important sites includingJeitun Anau Gonur Depe and Merv Although many of these projects are ongoingimportant papers pertaining to the area have emerged (Harris et al 1993 1996 Herrmann 1997 Herrmann et al 1998 Hiebert 1994) providing valuable informationon changes in environment and society over this period Historical sources are moreproblematic Although literate civilizations have existed in the region since theAchaemenid period there is no systematic body of texts comparable to the clay tablets ofMesopotamia For the medieval period we are largely dependent on short descriptions inaccounts by Arab or Chinese travellers or Arab historians Some Sasanian records havesurvived through their use by the Arab historians Prior to this we are again dependent onbrief travellersrsquo accounts and histories compiled far away to the west in classical Greece and Rome Our understanding of the political dynamics underlying the increasingly well-documented settlement archaeology is therefore currently less sophisticated than in theNear East proper

ENVIRONMENT

Turkmenistan covers an area of 480000 km2 90 per cent of which is covered by thevirtually uninhabited Kara Kum Desert (Babaev 1996) Most of Turkmenistan compriseslowlands with mountains being confined to the southern and western parts of thecountry It lies within the temperate desert zone (Babaev 1994) and has a markedcontinental climate (Orlovsky 1994) Precipitation mainly falls as snow or rain in winterwith almost none in the agriculturally-active summer months of June through to September Average annual precipitation varies from 90 mm in Dashouz to nearly 400mm in the southwest highlands of the Kopet Dagh but in much of the country it


Figure 61 Turkmenistan showing locations mentioned in Chapter 6

is less than 200 mm per annum and therefore insufficient for dryland agriculture Averagetemperatures are high varying from 12 to 18degC The coldest months are December toFebruary with temperatures frequently falling below 0degC and the hottest months June to August when temperatures often exceed 45degC Potential evaporation rates vary accordingly from 1ndash2 mm per day in winter to 10ndash15 mm per day in summer Total annual potential evaporation rates are of the order of 2500ndash3000 mm which are far higher than the precipitation rates

The hydrological network is weakly developed and all major sources of water rise outside the countryrsquos borders (Fig 61) The headwaters of the Amu Darya the largest river in Central Asia are in the Pamirs and the river flows through a number of countriesbefore discharging into the Aral Sea It displays two periods of peak discharge oneduring the spring associated with snow melt the other later in the summer when ice meltincreases its flow The other main rivers all rise in the mountains to the south the Atrekflowing into the Caspian Sea and the Murgab and Tejen draining into the Kara KumDesert Although small when compared with the Amu Darya they are an importantsource of water and have long been used by people occupying the region Fed by winterrains and snowmelt they have only one period of peak discharge during the spring Inaddition to these rivers there is a number of smaller intermittent rivers and springs mostof which cease to flow during the summer Today as in the past human settlement inTurkmenistan is concentrated in two zones the piedmont at the foot of the steep slopes ofthe Kopet Dagh mountains and the desert oases

Irrigation agriculture in Central Asia 103

THE NEOLITHIC EARLY FARMING IN THE PIEDMONT

The beginning of agriculture in Turkmenistan is best documented by the importantexcavations at Jeitun (Djeitun) located some 25 km north of Ashgabat in the piedmontbetween the Kopet Dagh mountains and the Kara Kum Desert First discovered in theearly 1950s Jeitun has been the subject of a number of detailed excavations that haveproduced one of the bestknown archaeological sequences in Central Asia Todayfourteen 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 theKopet Dagh and discharges into the Kara Kum Desert The settlement covers less than07 ha and estimates of the cultivated land surrounding it in the neolithic period arebetween 15 and 33 ha First excavated by Masson in the 1950s and 1960s the site wasdated to c6000 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 British-RussianmdashTurkmen excavations of 1989ndash94 which indicates that the site was occupiedbetween c6300 and 5600 cal BC (Harris et al 1993 1996)

Recovery of animal bones and charred plant remains from these new excavations hasallowed a reassessment of the sitersquos subsistence base The results confirm earlier evidence for a primarily agricultural system of subsistence based on cereals and domesticsheep and goat augmented by hunting primarily of gazelle The cereals are dominatedby einkorn wheat (Triticum monococcum) with small amounts of emmer (Tdicoccum)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 sickleblades accounting for 37 per cent of all tools found in Massonrsquos early investigation of the site (Masson and Sarianidi 1972) In addition to cereal cultivation the inhabitants ofJeitun herded goats and sheep faunal analysis shows that although raised primarily formeat these animals could also have been an important source of milk wool hair andskins The dominance of domesticated plants and animals from the very bottom of theJeitun sequence together with the absence of wild progenitors of wheat and sheep inCentral Asia supports the view that agriculture and its attendant domesticated species didnot evolve independently in the region but rather reached it from the Fertile Crescent ofsouthwest 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) forexample assumed that cultivation at Jeitun was entirely dependent on run-off from the Kopet Dag with Lewis (1966) suggesting that Jeitunrsquos location on the distal reaches of the Kara Su was due to the fact that neolithic farmers were better able to control andmanipulate flows in this part of the river system However Kohl (1981) argued thatJeitun was in fact located on the distal reaches of the Tejen Delta which at this timedischarged further into the Kara Kum Desert than today The presence of many seeds ofthe weeds club-rush (Scirpus maritimus) and goat-face grass (Aegilops tauschii) in


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 (lsquotakyrsrsquo) rather than on stream sides irrigated by less saline floodwaters Takyrs are highlyimpermeable almost flat clay surfaces that retain water and are of considerableimportance to communities living in the desert today All these different scenarios wouldlargely draw on naturally irrigated land with only relatively small-scale channels or embanking required No definite evidence of such features has been discovered thoughthis 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 todayThere is some evidence to suggest that this assumption may be unfounded for the base ofthe dunes overlying fluvial deposits at Jeitun has yielded an OSL date of c4500ndash5000 years BP so it is possible that arid conditions similar to today developed somewhat laterthan previously thought (just as wetter environments characterized the early Holocene inthe Levant Chapter 4) Further support for this hypothesis is provided by a recentanalysis of plant remains from the site suggesting that cultivation may have beenpossible without irrigation (MCharles pers comm) Despite this uncertainty though it is evident that there is a long history of agriculture in this region and that by the fifthmillennium BC agricultural settlements were spread along the piedmont from KyzalArvat 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 wasfollowed by several millennia of continuing settlement largely in the piedmont of theKopet Dagh during the Eneolithic or Chalcolithic (c4800ndash3000 BC) Early Bronze Age (c3000ndash2500 BC) and Middle Bronze Age (c2500ndash1900 BC) Sites grew significantly in sizemdasheneolithic settlements such as Altyn Depe Anau and Namazga cover up to 25 hamdashand there were two key changes in settlement pattern the expansion into the Geoksyuroasis in the Eneolithic and the emergence of state-level urbanism in the piedmont zone in the terminal Early Bronze Age

The Geoksyur oasis (Fig 61) is situated on the Tejen River delta and unlike otheroases 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 oasisdating from the earliest Eneolithic (Kohl 1984) but the oasis appears to have beenabandoned by the end of the Eneolithic Earlier eneolithic settlements are at the end ofbranches of the river delta suggesting that a modified form of floodwater irrigation waspractised Later in the Eneolithic well-developed artificial irrigation systems aredocumented for the first time in Turkmenistan (Namazga III period c3500ndash3000 BC) Aerial photographs and excavations have shown that land around the site of Geoksur Iwas irrigated by three parallel canals each up to 3 km long and 5 m wide possiblyirrigating 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


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 areasfortification walls around settlements increased status differentiation in burials andevidence of much interaction between settlements throughout the Kopet Dag region allconsistent with a state-level society (Hiebert 1994) These trends continued into theMiddle Bronze Age and by its terminal phase (2200ndash1900 BC) the foothills contained a number of very large sites such as Namazga (50 ha) and Altyn Depe (25 ha) This periodof expansion came to an end in the Late Bronze Age The settlements at Anau andNamazga for example were considerably smaller now covering only a few hectaresRelatively little is known about agriculture in the piedmont zone in the Eneolithic andBronze Age Irrigation canals have not been located in the piedmont but this may reflectdeposition and erosion in this geomorphologically-active zone The presence of bread wheat and six-row hulled barley in lencolithic samples from Anau dated to c4500ndash3000 BC has been cited as possible evidence for irrigation (Miller 1999) but both cereals weregrown 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 Agealthough this occurred while some sites such as Altyn Depe were still very large Anumber 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 thatpeople were forced to migrate whilst some authors have highlighted the potentialimpacts of climate change Lewis (1966) argued that there is no evidence of a major shiftin climate during this period but as mentioned above evidence is emerging for a shift todrier conditions c5000ndash4500 years ago coinciding with the rise of agriculture in theMurgab oasis It is possible therefore that conditions became sufficiently dry toprecipitate change

The bronze age settlements of the Merv oasis covered an area of 100 km northmdashsouth by 50 km eastmdashwest which is almost five times larger than the later medieval andclassical oasis to the south Hiebertrsquos recent re-analysis of the ceramic chronology and survey data suggests that the colonization of the oasis was rapid (Hiebert 1994) Thesites cluster in lsquomicro-oasesrsquo forming linear patterns that presumably followed old river branches (Fig 62) Settlements are characterized by large fortified building complexes with intervening fields which as Hiebert points out typify Central Asian oasisarchitecture of the time Initially settlements were located on the northern margins of theoasis with the system expanding southwards some 400 years later (Hiebertrsquos 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 thesmaller 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 andanimals were domesticated By the Bronze Age the variety of crops grown had increasedsignificantly compared with in the Neolithic samples from the middle bronze age site ofGonur Depe in the Merv oasis for instance are dominated by hulled and naked barley


with free-threshing wheat lentils peas chickpeas and grape also present (Miller 1993

Figure 62 Bronze and iron age settlement in the Merv oasis (adapted from Hiebert 1994)


Moore et al 1994) These finds are consistent with those from the neighbouring Geoksyur oasis (Lisitsina 1969 Lisitsina and Prishchepenko 1976) and are typical ofbronze age settlements throughout the Near East Late bronze age samples from TahirbajTepe 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 lessmaterial complexity (Kohl 1984) In the Merv oasis iron age sites are concentrated infour lsquomicro-oasesrsquo The northernmost two Takhirbai and Togolok contain bronze agesettlements while the southernmost two Yaz depe and Aravali represent newoccupation thus forming part of the pattern of southward movement of settlements thatcontinues until the Achaemenid period (Bader et al 1996 see below) This shift in settlements is most plausibly explained by increased extraction of water upstream bysettlements using more sophisticated canal systems collecting water near the head of thedelta However early sites in the upper part of the oasis may have been masked byalluvial deposition accounting in part for this pattern

ACHAEMENID TO MEDIEVAL URBAN SOCIETIES

The Achaemenid period (530ndash330 BC) marked two important transitions for the Merv oasis it was the first of several periods when Merv came under the control of an empirebased to the south and for the first time a series of urban centres emerged in the oasisFrom this time onwards Merv was also militarily important as a frontier city at thenortheastern part of firstly the Achaemenid and later the Seleucid (330ndash140 BC) Parthian (140 BCndashAD 220) and Sasanian (AD 220ndash651) empires Surveys of the magnificentruins of Mervrsquos 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 adjoiningSeleucid city of Gyaur Kala (400 ha) (Fig 62) which continued to be occupied for a period of over a millennium even after the construction of the nearby city of Sultan Kalain the eighth century AD Survey work in rural areas in the north of the oasis confirmsthis basic pattern of expansion with increasing residential areas from Achaemenid toSeljuk times (Bader et al 199394 Gubaev et al 1998) At its greatest extent the oasis covered c700 km2 The area cultivated appears to have fluctuated with a decrease inrural settlement in the Hellenistic period and a marked increase in cultivation andprobably the first construction of a large central dam and canal network in the Parthianperiod

Although written sources state that Merv was destroyed by Mongol invasions in AD1221ndash2 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 TimuridsNotwithstanding this the oasis declined in importance and the Timurid city wasabandoned by the nineteenth century Overall therefore changes in settlement patternsuggest three key phases in the occupation of the Merv oasis the initial colonization by dispersed but numerous bronze age settlements c2200 BC urban development in theAchaemenid period c600 BC and the gradual abandonment of intensive settlement in


most of the oasis in the centuries after the Mongol invasions of AD 1221ndash2 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 bestarchaeobotanical evidence for the Sasanian period It indicates that during the LateSasanian period (the fifth to seventh centuries AD) cereals consisted as before ofabundant hulled barley and free-threshing wheat and rarer broomcorn millet lentils(chickpea seems to disappear after the Bronze Age) very abundant cotton seed and awide range of fruits and vegetables including cucumbermelon grape almond peach andnuts Two changes are apparent in comparison with the Bronze Age first an increase incrop diversity particularly in the fruits and second and more importantly the addition ofcotton which is a source of both textile and oil and like millet a crop that expands thegrowing season through the summer after the wheat and barley harvest Overall therange of crops seems similar to that mentioned in Islamic times in the tenth centuryMervrsquos famously soft cotton textiles were exported as far as Africa and Spain and thereare thirteenth-century references to Mervrsquos fine grapes and other fruits (Serjeant 1972)

Of the range of crops grown in the Sasanian period barley and cotton are relativelytolerant of soil salinity (though not of course of heavily salinized soils) bread wheat ismoderately tolerant melon and grape are moderately sensitive and almond and peach aresensitive (Maas 1987) That crops sensitive to salinity are present throughout the lateSasanian sequence from Merv and that the full suite of crops is present in broadly similarquantities throughout the sequence strongly suggest that irrigation agriculture wassustained through this period without occurrence of catastrophic salinization This isconsistent with evidence for unbroken intensive settlement in the oasis from the Parthianto Seljuk periods Although salinization has often been viewed as an inherent andimminent threat in ancient irrigation systems in the Near East particularly inMesopotamia (Jacobsen and Adams 1958) there is increasing evidence that soilmanagement practices that avert salinization (and which are ethnographicallydocumented 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 innovationsduring 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 mostwidely used in the highland area of Iran Qanats are difficult to date directly butassociations with sites suggest that they became widespread in the highlands of Iran andneighbouring 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 199394 1996 Gubaev et al 1998) like the contemporary transformationsoccurring further south in Susiana (Wenke 1975ndash76) and Mesopotamia (Adams 1981)Major changes in irrigation technology in the Merv oasis are therefore later (if thedating is correct) than the first urbanization at Erk Kala and Gyaur Kala but coincide withthe 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 centuriesmdashduring which time the


oasis nourishedmdashmay give a good parallel The Arab historians and geographers such asMuqaddasi Al-Biruni and Yakut provide valuable accounts of water distribution and irrigation systems (see Bartold 1914 1928 and Le Strange 1905 for translations anddiscussions of their works) It is evident from these that the administration of scarcewater resources was central to the way in which the social and political hierarchy ofsettlements operated water was viewed as a lsquogift from Godrsquo that could not be owned or controlled by an individual The city of Merv had access to only one source of watermdashthe Murgab river which rises in the Afghan mountains and drains northward into the KaraKum Desert The riverrsquos annual discharge is about 12 km3 which is approximately 5 per cent of the total amount of water available for use in Turkmenistan at present (OrsquoHara 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 Petersen1997) The regionrsquos agricultural success was in part due to the land and watermanagement strategies of the time Land for example was divided into small plots thatwere intensively cultivated receiving water on a regular basis The amount of landcultivated in any given year depended on water availability Muqaddasi writing in thetenth century AD described how a depth gauge situated at the Razik Dam to the south ofthe city was used to determine whether there would be a surplus or deficit of water thatyear If the level reached the 60th point water would be plentiful that year and the orderwould be given to increase the amount of land cultivated whereas in years of low wateravailability the area was reduced and only the best lands were cultivated The dam wasextremely important and was in effect the only water storage facility for the city Itsmaintenance was assured by 400 divers employed around the clock each diver having todeliver a specified amount of wood and mud to the dam each day (Bartold 1914) Yakutwho resided in the city at its zenith in the early thirteenth century provides furtherdetails He described how water gauges were installed at the head of every canalthroughout 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 officeso 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 12000 workers paid by the water users who were also expected to take part in major construction schemes and in theannual maintenance programme

EXPANDING THE IRRIGATION SYSTEM THE TSARIST AND SOVIET PERIODS

When Central Asia finally came under Tsarist control in the late 1880s the newadministration attempted to introduce reforms in the irrigation sector These failedhowever and the authorities declared that irrigation would be run lsquoby customrsquo Notwithstanding this a number of subtle changes was made most important irrigationofficials became part of the Tsarist civil service and as such were no longer controlled bywater users This act severed the link between water users and providers so effectivelyundermining the traditional system of water management State salaries for officials were


low and there was no longer any incentive to control the system The situation wasexacerbated by the imposition of irrigation officials unaccustomed to the traditionalmethod of management resulting in increased problems within the system whichbecame subject to corruption and abuse by the wealthy and more powerful water users

More significant than Tsarist interventions in water management however were thechanges in agricultural policies The authorities in Moscow keen to end their reliance onAmerica for cotton (particularly following the American Civil War when supplies almostceased) recognized that Central Asia had the potential to become a major cotton growingregion in fact the main factor behind initiatives to increase the amount of land irrigatedwas cotton production (Lipovsky 1995) The subtle but nonetheless important changesin water management coupled with increased demand and use of water appear to havecaused widespread land degradation In the Merv oasis for example the irrigationnetwork was expanded by some 33000 ha (Zaharchenko 1990) but poor management ofthe system caused local water tables to rise resulting in salinization and widespreadsurface 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 theSoviet administration decreed that water management was to be taken lsquoout of the hands of traditional elders and councils with whom it residedrsquo (Black et al 1991) and like land was to become a common resource for the benefit of all Various governmentbodies were established to be responsible for the development of a regional watermanagement strategy that would allow centrally-determined production targets to be met With cotton production the priority for Moscow huge sums of money were invested inthe region in the development of massive highly integrated systems of water distributionand 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 Canalfor example considered to be one of the engineering feats of the Soviet era now transfersin excess of 129 km3 of water from the Amu Darya along its 1400 km length every yearirrigating an area of c1 million ha (Hannan and OrsquoHara 1998)

There has been much criticism of the management and maintenance of Soviet irrigation systems and the inefficiency of water use (eg Micklin 1991) Losses occurredthroughout the system with problems of seepage and evaporation from the manythousands of kilometres of unlined irrigation canals creating huge problems withwaterlogging and soil salinization Within a few years of the Kara Kum Canal beingconstructed 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 (OrsquoHara 1997) Water use at the field level also rose as field size increased to accommodateincreasingly bigger agricultural machinery not only increasing the amount of time that ittook 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 todominate with large and poorly levelled fields creating huge problems for irrigatorsMoreover unlike in the past access to water was not a problem with diversion schemesbringing what to many seemed an infinite supply of free water people who had long


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 theirrigation infrastructure dredging canals and ensuring that the drainage system was cleanAt the farm level maintenance became the responsibility of a few collective workers Inall cases the bulk of the work was done using heavy equipment Consequently waterusers had little if anything to do with the management or maintenance of the waterdistribution and irrigation system Despite Soviet successes in expanding the irrigationnetwork and increasing agricultural output the systems they built were (and still are)inflexible and highly inefficient By the 1980s agricultural land in Turkmenistan wasbeing abandoned at a rate of over 50000 ha per annum (Zaharchenko 1994) cleartestimony to the fact that this huge irrigation system is not sustainable

CONCLUSION

In Tables 61 and 62 we summarize the major trends in settlement and agriculture insouthern Turkmenistan It is evident that there is a strong correlation between the degreeof urbanization and population size (themselves correlates of centralized political control)and the sophistication of irrigation technology The range of crops likewise increasesthrough time Although

Table 61 Simplified chronological chart of prehistoric settlement in Turkmenistan

Archaeological period and date (cal BCAD)

Settlement Irrigation systems Crops

Neolithic 6300ndash4800 BC

Small farming villages on piedmont of Kopet Dagh Key site Jeitun

Crops cultivated in areas of high water-table possibly also simple diversion of streams

Main crop einkorn also emmer hulled and naked barley

Eneolithic (Chalcolithic) 4800ndash3000 BC

Larger complex settlements to 15 ha with shrines and fortifications in Middle Eneolithic (4000ndash3500 BC) spread of settlements to Geoksyur oasis in Middle Eneolithic but abandonment of oasis by EBA Key sites Altyn-depe Anau Geoksyur Namazga (phases IndashIII)

Simple irrigation assumed for piedmont and early occupation of Geoksyur large irrigation canals identified in Geoksyur oasis in late Eneolithic (Namazga phase III)

Hulled barley free-threshing wheat

Early Bronze Sites to 25 ha restricted to Irrigation assumed Hulled barley


Age (EBA) 3000ndash2500 BC

piedmont zone Key sites Altyn-depe Namazga (IV)

for large settlements in piedmont but no direct evidence

free-threshing wheat grape

Middle Bronze Age (MBA) 2500ndash1900 BC

Sites to 50 ha monumental architecture Abandonment of piedmont sites at end of MBA Major fortified settlements appear in Merv oasis in terminal period (2200ndash1900 BC) Key sites Altyn-depe Namazga (V) Gonur depe

Smaller-scale irrigation at northern fringe of Merv oasis

Main crop hulled barley also free-threshing wheat lentil pea chickpea grape

Late Bronze Age (LBA) 1900ndash1500 BC

More dispersed smaller sites (to 2 ha) in piedmont Period of Bactrian-Margiana Archaeological Complex abundant large sites in Merv oasis Key sites Namazga (VI) Gonur depe

Sophisticated canal irrigation in Merv oasis using water from main channels of rivers

Early Iron Age 1500ndash550 BC

Abundant settlements (to 15 ha) in piedmont and oases Key sites Tahirbaj Yaz-depe (Merv oasis)

Introduction of qanat(kiariz) irrigation to piedmont In Merv oasis settlement continues to shift to south

Broomcorn millet

Table 62 Simplified chronological chart of settlement in the Merv oasis in the historic period

Historical period and date (cal BCAD)


Achaemenid 550ndash330 BC

Founding of Achaemenid city at Erk Kala c500 BC dispersed settlement centred on large buildings throughout the oasis and continuing to Seljuk period

Seleucid (Hellenistic) 330ndash140 BC

Construction of the city of Antiochia (Gyaur Kala) incorporating Erk Kala as citadel

Marked reduction in rural settlement

Parthian New settlements in north Expansion of Main crops


140 BCndashAD 220 of the oasis fortifications on perimeter such as Gobekli

cultivated area possible construction of central Murghab dam and extensive canal network

cotton hulled barley free-threshing wheat also lentil grape almond

Sasanian AD 220ndash651

Peak of settlement in the Merv oasis much continuity with Parthian period Possible construction of Wall of Antiochus (usually dated to the Hellenistic period) which marks northern limit of most post-bronze age settlement

Cultivated area stable

Main crops cotton hulled barley free-threshing wheat also lentil pea melon grape almond peach broomcorn millet

Umayyad Abbasid 8thndash9th centuries Samanid Seljuk and Post Seljuk 11thndash13th centuries Mongol conquest 1221

Continuity in settlement after Arab conquest of AD 651 Merv is capital of Seljuk empire in 11th and 12th centuries Sultan Kala established 9th century fortified 12th century Devastating conquest of city by Mongols but archaeological evidence suggests some post-conquest occupation

Continuity in area cultivated until Mongol conquest results in destruction of dam system Abundant textual evidence for function of irrigation system in 10thndash13th centuries



Timurid 14thndash15th centuries Safavid 1502ndash1736

New city built 1409 at Abdullah Khan Kala but decline continues

Central dam rebuilt in Timurid period destroyed in war of 1727

Turkmen 18thndash19th centuries

Dispersed settlement with semi-independent landlords

Irrigation system functioning but small-scale cultivation

Russian conquest 1890 Soviet Union 1919ndash1991 Republic of Turkmenistan 1991ndash

Establishment of modern settlement at Mary planned villages and communal farms

Introduction of large-scale irrigation systems for cotton Karakum canal

American cotton species


we would hesitate to identify simple cause and effect it would appear that increasedpopulation was linked through a complex sequence of interactions with the expansion ofirrigated agriculture and increased centralization of authority The increase in populationnot only required an increase in the amount of land irrigated but also provided theworkforce necessary for this expansion to take place Irrigation flourished during periodsof political stability often when a single polity ruled over the area and declined inperiods of invasion or unstable internal political conditions

The decline of Merv can clearly be traced to the Mongol destruction of AD 1221ndash2 The Mongols took advantage of the fact that Merv like most other settlementsthroughout Central Asia was reliant on a single water source In their rapid conquest ofthe region the Mongols frequently forced communities to capitulate by disrupting watersupplies and damaging irrigation structures and all they needed to do at Merv was todestroy the main dam that controlled water in the oasis Whilst the city was in partrebuilt the irrigation systems were never fully reconstructed until the region once againcame under the influence of another empiremdashthat 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 pastsuggesting that sustainable irrigation in Turkmenistan is not only feasible but has beenthe norm Traditional irrigation systems were generally localized and often dependent ona single water supply that was not only limited but also liable to fluctuate considerablyfrom year to year Water management required considerable skill hence the mirab bashiresponsible for highly important and often contentious decisions on water allocation anddistribution was one of the most senior officials in central governmentmdashindeed the success of many political officials often hinged on their skill at managing local waterresources Yet whilst water was managed centrally all water users were responsible forthe upkeep of the system with those gaining more being expected to contribute moreThe fact that individuals could benefit as a result of their efforts gave all users a vestedinterest in ensuring that the irrigation network was maintained and that water was usedefficiently The Soviet system effectively broke this link with the system managedcentrally but from afar Together with the collectivization of land the imposition ofcentral planning meant that benefits were no longer linked to duty water users had no sayin how the system was managed nor were they responsible for its maintenance Theestablishment of myriad agencies to oversee different parts of the network resulted inunnecessary bureaucracy and waste In sum traditional irrigation and water distributionsystems tended to be small highly productive well managed extremely efficient andsustainable 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 declinewill be managed and what can be done to ensure the future sustainability ofTurkmenistanrsquos (and indeed Central Asiarsquos) water distribution and irrigation networkThe Central Asian Republics have inherited a Soviet-built system and must learn to work

present throughout oasis completed 1967


with the system and resources that are available While we cannot revert to the pastCentral Asiarsquos water managers would do well to look to the past for some of the answersto their current and future problems

REFERENCES

Adams RM (1981) Heartland of Cities Surveys of Ancient Settlement and Land Use onthe Central Floodplain of the Euphrates Chicago University of Chicago Press

Babaev AG (1994) Landscapes of Turkmenistan In VFet and KIAtamuradov (eds)Biogeography and Ecology of Turkmenistan 5ndash22 Dordrecht Kluwer

Babaev AG (1996) Problems of Aridland Development Moscow Moscow University Press

Bader A Gaibov V and Koshelenko G (199394) The northern periphery of the Mervoasismdashfrom the Achaemenid period to the Mongol conquest Silk Road Art and Archaeology 351ndash70

Bader AN Gaibov VA and Koselenko GA (1995) Walls of Margiana In AInvernizzi (ed) In the Land of the Gryphons Papers on Central Asian Archaeology in Antiquity 39ndash50 Florence Le Lettere

Bader A Gaibov V Gubaev A and Koshelenko G (1996) The oasis of Merv thedynamics of its settling and irrigation Ancient Civilizations from Scythia to Siberia349ndash60

Bartold VV (1914) K istorii Orosheniya Turkestana St Petersburg Selrsquoskago Vestnika (In Russian)

Bartold VV (1928) Turkestan down to the Mongol Invasion London Luzac second edition

Bedford DP (1996) International water management in the Aral Sea basin Water International 2163ndash9

Black C Dupree L Endicott-West E Naby E Matuszewski DC and WaldronAN (1991) The Modernization of Inner Asia Armonk NY MESharpe

Boardman S (1997) Plant use in the Merv oasis Iran 3529ndash31 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 1990ndash95 Rome ISIAO Hannan T and OrsquoHara SL (1998) Managing Turkmenistanrsquos Kara Kum canal

problems and prospects Post-Soviet Geography and Economics 39225ndash35 Harris DR and Gosden C (1996) The beginnings of agriculture in western Central

Asia In DRHarris (ed) The Origins and Spread of Agriculture and Pastoralism in Eurasia 370ndash89 London UCL Press

Harris DR Masson VM Berezin YE Charles MP Gosden C Hillman GCKasparov AK Korobkova GF Kurbansakhatov K Legge AJ and Limbrey S(1993) Investigating early agriculture in Central Asia new research at JeitunTurkmenistan Antiquity 67324ndash38

Harris DR Gosden C and Charles MP (1996) Jeitun recent excavations at an early


Neolithic site in southern Turkmenistan Proceedings of the Prehistoric Society 62423ndash42

Herrmann G (1997) Early and medieval Merv a tale of three cities Proceedings of the British Academy 941ndash43

Herrmann G and Petersen A (1997) The Ancient Cities of Merv TurkmenistanLondon University College London Institute of Archaeology International MervProject

Herrmann G Kurbansakhatov K and Simpson SJ (1998) The International MervProject Preliminary report on the sixth season (1997) Iran 3653ndash75

Hiebert FT (1994) Origins of the Bronze Age Oasis Civilization in Central AsiaCambridge MA Harvard University Peabody Museum of Archaeology andEthnology American School of Prehistoric Research Bulletin 42

Jacobsen T and Adams RM (1958) Salt and silt in ancient Mesopotamian agricultureScience 1281251ndash8

Kohl PL (1984) Central Asia Palaeolithic Beginnings to the Iron Age Paris Editions Recherche sur les Civilisations Synthegravese 14

Kornilov BA and Timoshinka VA (1975) The impact of the Kara Kum canal on theenvironment Soviet Geography 15308ndash14

Le Strange G (1905) The Lands of the Eastern Caliphate Cambridge Cambridge University Press

Lewis RA (1966) Early irrigation in West Turkestan Annals of the Association of American Geographers 56467ndash91

Lipovsky I (1995) The central Asian cotton epic Central Asian Survey 14529ndash42 Lisitsina GN (1969) The earliest irrigation in Turkmenia Antiquity 43279ndash88 Lisitsina GN (1981) The history of irrigation agriculture in southern Turkmenia In PL

Kohl (ed) The Bronze Age Civilization of Central Asia 350ndash8 Armonk NY MESharpe

Lisitsina GN and Prishchepenko LV (1976) The significance of paleoethnobotanicalremains for the reconstruction of early farming in the arid regions of the USSR Folia Quaternaria 4783ndash8

Maas EV (1987) Salt tolerance of plants In BRChristie (ed) CRC Handbook of Plant Science in Agriculture Volume II 57ndash71 Boca Raton FL CRC Press

Maier U (1996) Morphological studies of free-threshing wheat ears from a Neolithic sitein southwest Germany and the history of the naked wheats Vegetation History and Archaeobotany 539ndash55

Masson VM (1957) Jeitun and Kara-tepe Sovetskaya Arkheologiya 1143ndash60 (In Russian)

Masson VM and Sarianidi VI (1972) Central Asia Turkmenia before the Achaemenids London Thames amp Hudson

Micklin PP (1991) The Water Management Crisis in Soviet Central Asia Pittsburgh University of Pittsburgh Center for European and Russian Studies Carl Beck Papers inRussian and East European Studies

Miller NF (1993) Preliminary archaeobotanical results from the 1989 excavation at thecentral Asian site of Gonur Depe Turkmenistan International Association for the Study of the Cultures of Central Asia Information Bulletin 19149ndash63


Miller NF (1999) Agricultural development in western Central Asia in the Chalcolithicand Bronze Ages Vegetation History and Archaeobotany 813ndash19

Moore KM Miller NF Hiebert FT and Meadow RH (1994) Agriculture andherding in the early oasis settlements of the Oxus Civilization Antiquity 68 418ndash27

Nesbitt M (1994) Archaeobotanical research in the Merv Oasis Iran 3271ndash3 OrsquoHara SL (1997) Irrigation and land degradation implications for agriculture in

Turkmenistan central Asia Journal of Arid Environments 37165ndash79 OrsquoHara SL (in press) Central Asiarsquos water resources contemporary and future

management issues International Journal of Water Resources Development Orlovsky NS (1994) Climate of Turkmenistan In FFet and KIAtamuradov (eds)

Biogeography and Ecology of Turkmenistan 23ndash48 Dordrecht Kluwer Academic Press Monographicae Biologicae 72

Pierce RA (1960) Russian Central Asia 1867ndash1917 A Study in Colonial RuleBerkeley University of California Press

Powell MA (1985) Salt seed and yields in Sumerian agriculture A critique of thetheory of progressive salinization Zeitschrift fuumlr Assyriologie und VorderasiatischeArchaumlologie 757ndash38

Serjeant RB (1972) Islamic Textiles Material for a History up to the Mongol ConquestBeirut Librairie du Liban

Smith DR (1992) Salinization in Uzbekistan Post-Soviet Geography 3321ndash33 Wenke RJ (1975ndash76) Imperial investments and agricultural developments in Parthian

and Sasanian Khuzestan 150 BC to AD 640 Mesopotamia 10ndash1131ndash221 Zaharchenko BT (1990) Voda v Turmenskoy Zhizni (Water in Turkmen Life)

Ashgabat (In Russian) Zaharchenko BT (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 Khaldounrsquos 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 Sincethen many historians have tended to believe that the Arab invaders were responsible forthe land degradation that is so visible today in many parts of north Africa More thantwenty years ago after French decolonization controversy was particularly strong(Berque 1970 1972 Cahen 1968 Idris 1968a 1968b Poncet 1967 1968) thoughinterest has since decreased Since this period however advances in geoarcheologicalresearch allow a reassessment of the role of Arab nomads in land degradation in northAfrica the basis for which was more ideological than factual The purpose of this chapteris to discuss the geoarchaeological record of the eastern Maghreb (Fig 71) and to compare it with the historical record of conquests invasions and occupations in order toassess the respective roles of climate and people in shaping this regionrsquos 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 theLate Holocene a phase characterized throughout the region by a trend to aridificationfollowing the Middle Holocene climatic optimum (Ballais 1991a) This aridification wasthe reason for the reappearance of aeolian deflation on the great lsquochottsrsquo or lsquosebkhasrsquo (salt flats) of southern Tunisia as well as the incision of the lower prehistoric holoceneterrace especially in the present-day semi-arid subzone (Ballais and Benazzouz 1994Table 71) The phase is contemporaneous with the erosional crisis at the transitionbetween the Bronze and Iron Ages in the northern Mediterranean

Figure 71 The eastern Maghreb showing locations mentioned in Chapter 7

Key 1 Ksar Rhilane 2 chott Rharsa 3 Ksar Rheriss 4 Wadi Keacutebir-Miliane 5 Henchir Rayada 6 Wadi Cheacuteria-Mezeraa 7 Wadi el Akarit 8 Haiumldra Stippling denotes major areas of sand dunes horizontal dashed lines denote salt flats (lsquosebkhasrsquo or lsquochottsrsquo)

Table 71 Morphoclimatic evolution in the eastern Maghreb during the later Holocene dates in radiocarbon years BP

Isotopic chronology Morphogenesis Shifting sands incision610 +minus 110 Terrace aggradation Incision 1350 +minus 70 Terrace aggradation 1470 +minus 190 Terrace aggradation 1730 +minus 185 Pedogenesis 2380 +minus 155 Flood deposits 2420 +minus 70 Gyttja deposits 2590 +minus 90 Pedogenesis Deflation


(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 occurrenceof strong floods due to intense rainfalls the latter probably highly concentrated andepisodic Thus the climate was probably characterized by greater seasonality than beforeand a hydric balance less favourable than today As far as we can tell Phoeniciancolonization in the eastern Maghrebmdashfor long limited to a few coastal sitesmdashdoes 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 threemorphogenic phases can be discerned in the region the end of an aggradation period anincision period and a second alluvial aggradation period (Table 71)

The end of an aggradation period

Between c2400 and 2200 BP aggradation became widespread once more The evidenceincludes 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 71 site 1 Fig 72) gyttja accumulation in the Rharsa chott in the arid subzone (Ballais 1992 Fig 71 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 NemenchaMountains the fine alluvium has been organized in continuous beds by slow streams in alarge channel The mean annual accumulation rate was 14ndash22 mm which is very close to the rate calculated for the low prehistoric Holocene terrace (Ballais 1991b) The fewpollen grains taken from those deposits show that a woodland could have colonized theslopes All these characteristics are compatible with a more positive hydric balance in thisphase than either during the previous climatic phase or today especially in the arid andSaharan subzones In comparison with the climatic optimum period (Ballais 1991a) thedegradation of climate in the semi-arid subzone is shown by a probable increase insummer evaporation though winters remained sufficiently cool to permit the growth ofCedrus on the summits of Nemencha 1800 m above sea level

During this period Roman penetration of the interior seems to have been very limitedno evidence for Roman agricultural works has yet been observed

Incision 3680 +minus 160 Terrace aggradation 4220 +minus 50 Terrace aggradation

Conquests and land degradation in the eastern Maghreb 123

An incision period

No deposit has been identified for the period from c2200 BP to c1650 BP (the second and third centuries AD) but there is widespread evidence for

Figure 72 Flood deposits at Ksar Rhilane (Tunisia) bedded silts and sands (dated to 2380 +minus 155 BP) with ripple-marks in the upper part with the present-day dunes behind

Photograph J-LBallais

watercourse downcutting This period was thus characterized by streams having thecapability to incise their channels and to transport their alluvial sediments up to the baselevels It is possible precisely to measure neither the scale of the incision nor its annualrate owing to the lack of chronological markers Nevertheless it should be rememberedthat the mean annual rate of downcutting of the low prehistoric Holocene terrace was 12mm between 40003500 BP and 17001600 BP (Ballais 1991b) Thus during this periodslopes furnished very few colluvial sediments indicating either that they were wellprotected by rather dense vegetation or that the agricultural techniques

did not produce soil erosion The beginning of this period was marked by the almostcontinuous expansion of Roman colonization both westwards and southwards (Feacutevrier 1989) Even though they cannot be dated precisely numerous dams were built at aboutthis time on the watercourses of the Nemencha and Auregraves mountains (Ballais 1976 Birebent 1964 Leveau 1974ndash75) and in southern Tunisia (Ballais 1990 Trousset 1974) Sometimes their construction can be shown to be associated with the constructionof the Roman frontier works (limes) at the boundary of the arid and Saharan subzones


(Baradez 1949 Trousset 1974) Presumably it was the combination of favourableconditions of climate vegetation and soils and well-organized agricultural activities that was responsible for such limited soil erosion through the second and third centuriesAD

An aggradation period

Evidence for an aggradation terrace dating to historic times is widespread along most ofthe watercourses from the north to the south and from the humid subzone to the upperSaharan subzone The exception is the lower Saharan subzone so far the historicaggradation terrace unlike the late prehistoric terrace has not been identified in the farsouth of Tunisia As with the late prehistoric terrace the later feature developed for themost part in those catchments formed in loose erodible rocks The sediments aregenerally fine in texture beige in colour in the south and greyer in the north This terracealso has an extensive surface area particularly in the north and forms a sediment unitthat varies in thickness from 1 m (along small watercourses) to 5ndash6 m (along major rivers)

On the largest rivers detailed studies of sedimentation patterns show variations withlatitude Thus in the Wadi Leben terrace at Ksar Rheriss towards 35degN in the arid subzone (Ballais 1991b Fig 71 site 3) desiccation cracks appear in thin beds formedby washed clay and have been filled by sand during later flooding These patterns whichare characteristic of intermittently-flowing wadis disappear in the more humid subzoneswith the exception of the Keacutebir-Miliane wadi (Fig 71 site 4) which is today a perennial stream Conversely further to the north the terraced sediments of many watercoursescontain dark hydromorphic silt-like facies rich in organic matter or in manganese oxideIn eighteen different cases it is possible to correlate the presence of these facies with theperennial nature of the watercourse or inversely their absence with the intermittentnature of the watercourse In three other cases this correlation is not apparent Thesediments nearly always contain fragments of Roman or even earlier pottery (more thantwenty recorded examples have been noted) and often charcoal hearths or other artefactsThey cover the base of Roman bridges or aqueduct piles (Fig 73) or fill in low-volume dams built during the second and third centuries AD In another case along the Wadi esSgniffa a whole ancient settlement is covered by alluvial sediments There are still veryfew isotopic dates for the very low main historic Holocene terrace but two examples ofdated sediments are in the Wadi Cheacuteria-Mezeraa (Fig 71 site 6) which was radiocarbon-dated using samples of land molluscs to 1370 +minus 70 years BP (Fig 74) and in the Wadi el Akarit (Fig 71 site 7) where the terrace is dated to 1470 +minus 190 years BP (Ballais 1995 Fig 75) The rate at which the low terrace sediments of the historicperiod were accumulating became considerable at the eighteen locations examined theaverage reached 74 mm per year which is five times the rate of accumulation of thelower (prehistoric) Holocene terrace suggesting that the geosystems in which the historicterrace 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 aresult of a lengthy process of political and economic


Figure 73 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-LBallais

evolution Generally speaking the northeast of Tunisia near Carthage was cultivatedfrom perhaps the fifth century BC onwards and the cereal plains in the northwest ofTunisia and some parts of eastern Algeria from the third and second centuries BC Thesteppes of Algeria Tunisia and Libya were cultivated from the first and second centuriesAD and the borders of the Sahara during the second century AD at the time of theconstruction 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 wemight expect to find the imbalance of the geosystems resulting from this developmentoccurring at the same time in all places with the threshold producing the change fromincision to aggradation taking place at the same time in the north and south after eightcenturies of sedentary settlement in the former region and after a few decades of suchsettlement in the latter However this coincidence though not impossible seems highlyunlikely particularly if the tremendous differences in mean annual rainfall between thenorth and the south are taken into account In addition it has now been shown thatcomparable terrace deposits were deposited throughout the Mediterranean particularlytowards the end of classical antiquity and in the early Middle Ages (Ballais and Crambes1992 Bruumlckner 1986 Vita-Finzi 1969) The extensive occurrence of this feature can best be explained by an


Figure 74 Holocene terrace of Wadi Cheacuteria-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 +minus 70 BP) and a Roman drainage ditch filled with pebbles Photograph J-LBallais

underlying climatic shift across the region In view of the characteristics of the alluvialdeposits it seems likely that they originated from erosion of soils particularly those thatdeveloped during the Holocene Climatic Optimum and at the beginning of the Romanperiod

However land use systems in classical antiquity probably exacerbated these erosionaltrends Presumably the spread of cultivation and ploughing destroyed a large part of thevegetation on the watersheds reducing the cohesion of soils and superficial formations Itthen required only a small change in rainfall characteristics perhaps in the annual totalor at least in intensity and periodicity to produce considerable soil erosion and the startif not the return of water in the stream channels and increased discharge though thisincrease was not enough to carry the large sediment load from the slopes to the baselevels

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 isdifficult to compare the climatic situation with that of Sahara However the fact thatAcacia and Tamarix could be found in central Serir Tibesti at around 1700 BP and 1400BP indicates that at least in tropical Sahara the mean annual rainfall was more than thepresent day rainfall of 5 mm (Pachur 1974)


Figure 75 Holocene terraces in the Wadi el Akarit (Tunisia) A very low terrace of post-Islamic date (610 +minus 110 BP) is fitted into a gypseous terrace of late prehistoric date


THE VANDAL CONQUEST AND OCCUPATION

The Vandals are not one of the beloved peoples of eastern Maghreb history Too often itseems that their brief passage in the fifth and sixth centuries AD has no longer left traces(Courtois 1955) In fact despite new studies (Modeacuteran 1988) the Vandal period remains badly known for two main reasons The first one was the lack of interest ofFrench archeologists of the colonial period in the post-Roman civilizations (Feacutevrier 1989) The second one is a consequence of the first one the destruction of the Vandalsites 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 andland use are well known There is insufficient evidence on which to base any detaileddiscussion of climate soils vegetation and peoplersquos possible effects on the landscape at this time We can note however that alluvium continued to accumulate as the very lowhistoric terrace (Table 71)

BYZANTINE CONQUEST AND COLONIZATION

The Byzantines have been less neglected by French archeologists and historians of thecolonial period because they presented themselves as the legitimate heirs of the Roman


emperors (Ducellier 1988) They occupied the eastern Maghreb from the sixth to theseventh centuries AD As for the Vandal phase there is great uncertainty regardingmodes of rural settlement and land use Most information is available for the wars againstBerbers (Modeacuteran 1989) In particular most of the numerous small forts still visibletoday were attributed to Byzantine colonization but it now appears that some are laterthan the Arab conquest (Mahjoubi 1978) According to isotopic datings the end of theaggradation of the very low historic alluvial terrace coincides more or less with theByzantine period This can be confirmed at Haiumldra (Ammaedara Fig 71 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 easternMaghreb during the second part of the seventh century AD (Marccedilais 1946) presaging the high Islamic period which was a time of general economic prosperity (Vanacker1973) The last wave of Arab invaders the well-known Hilalian arrived in the middle ofthe eleventh century They are described as nomadic shepherds coming from UpperEgypt the lsquoplague of locustsrsquo in Ibn Khaldounrsquos memorable phrase who lsquopushed their flocks into the middle of the fields devastated the gardens stripped and ill-treated the country persons plundered the hamletsrsquo (Marccedilais 1946) In theory the consequences ofsuch devastation would have been so disastrous that they would have provokedcatastrophic and long-lived decline in the Tunisian economy (Al-Idrisi 1983 Ibn Khaldoun 1968 Marccedilais 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 71) 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 smallsize With only one exception (Wadi Keacutebir-Miliane) this terrace covers very small areas in particular in convex meander lobes and its height above the major bed rarely exceeds2 m Occasionally it appears as a rocky terrace that was breached in the previous build-up elsewhere the facies can sometimes be compared to that of the previous terracethough it is sometimes considerably coarser at least at the base The age of the terrace isstill rather uncertain because appropriate means of dating are not available but atHenchir Rayada (Fig 71 site 5) it contains Islamic pottery from the tentheleventhcenturies and in Wadi el Akarit (Fig 71 site 7) it was dated by radiocarbon using collagen to 610 +minus 110 years BP (Fig 75) This terrace is thus much younger than theperiod of the presumed Hilalian invasions

As for the previous terrace the widespread presence of a terrace of the same age can beseen 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 fifthcenturies AD) population was probably low at the beginning of the Hafside period inTunisia during the twelfth century AD This period marked the end of the medievalclimatic optimum and the beginning of the Little Ice Age in Europe (Lamb 1977 Le


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 furtherstudies 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 Evenif those shepherds did in fact cut down trees to any extent their main effect would havebeen to substitute pastures for cultivated fields The consequences would have been asfollows an increase in the rate of vegetation cover over the soil a consequent decrease inthe direct splash impact of rain drops on the soil and thus of pluvial erosion and sheeterosion and an increase in the lsquoroughnessrsquo of the terrain and thus a diminution in winderosion Within such a model it is necessary to moderate the intensity of such processesaccording to the climatic subzones and different grazing intensities but it seems realisticto suppose that in general such a move from arable to pastoral land use is likely toproduce less rather than more soil erosion

CONCLUSION

In the eastern Maghreb during classical antiquity and the Middle Ages fluctuations ingeosystems and in particular increases in soil erosion can be seen to have reflectedspecific combinations of climatic change and human activities A climatic fluctuation thatincreases the intensity of rains or the annual amount of precipitation affects slopes onlyif they have been made vulnerable by vegetation degradation or by cultivation systemsthat have not been designed to counteract erosion In other words phases of massiveagricultural colonization and phases of extension of the cultivated surface are veryfavourable to such erosion This was the case in the study area as in many parts of theMediterranean during the Roman period On the other hand periods of conquestgenerally seem to have been characterized by a contraction of the cultivated surface and aprogressive development of lsquonaturalrsquo vegetation or of pastures that limited soil erosion This may have been the situation in the case of the nomadic Hilalian shepherds of theArab conquest However there were exceptions to these trends in particular in theirrigated zones and in the terraced mountains

REFERENCES

Ballais J-L (1976) Morphogenegravese holocene dans la region de Cheacuteria (Nementchas-Algeacuterie) Actes du Symposium sur les Versants en Pays Meacutediterraneacuteens 127ndash30 Aix-en-Provence CEGERM 5

Ballais J-L (1990) Terrasses de culture et jessours du Maghreb oriental Meacutediterraneacutee3451ndash3

Ballais J-L (1991a) Evolution holocene de la Tunisie preacutesaharienne et saharienne Meacutediterraneacutee 431ndash8

Ballais J-L (1991b) Vitesse drsquoaccumulation et drsquoentaille des terrasses alluviales


holocegravenes et historiques au Maghreb oriental Physio-Geacuteo 22ndash2389ndash94 Ballais J-L (1991c) Les terrasses historiques de Tunisie Zeitschrift fur Geomorphologie

Suppl Bd 83221ndash6 Ballais J-L (1992) Le climat au Maghreb oriental apports de la geacuteomorphologie et de

la geacuteochimie Les Nouvelles de lrsquoArcheacuteologie 5027ndash31 Ballais J-L (1995) Alluvial Holocene terraces in eastern Maghreb climate and

anthropogenic controls In JLewin MGMacklin and JCWoodward (eds)Mediterranean Quaternary River Environments 183ndash94 Rotterdam Balkema

Ballais J-L and Benazzouz MT (1994) Donneacutees nouvelles sur la morphogenegravese et les paleacuteoenvironnements tardiglaciaires et holocegravenes dans la valleacutee de lrsquooued Cheacuteria-Mezeraa (Nemencha Algeacuterie orientale) Meacutediterraneacutee 3459ndash71

Ballais J-L and Crambes A (1992) Morphogenegravese holocene geacuteosystegravemes et anthropisation sur la montagne Sainte-Victoire Meacutediterraneacutee 1229ndash41

Baradez J (1949) Vue Aeacuterienne de lrsquoOrganisation Romaine dans le Sud Algeacuterien Fossatum Africae Paris Arts et Meacutetiers Graphiques

Baratte F and Duval F (1974) Les Ruines drsquoAmmaedara-Haiumldra Tunis Socieacuteteacute Tunisienne de Diffusion

Berque J (1970) Les Hilaliens repentis ou lrsquoAlgeacuterie rurale au XVIe s drsquoapregraves un manuscrit jurisprudentiel Annales Economic Socieacuteteacute Civilisation 51325ndash53

Berque J (1972) Du nouveau sur les Banucirc Hilacircl Studia Islamica 3699ndash113 Birebent J (1964) Aquae Romanae Recherches drsquoHydraulique Romaine dans lrsquoEst

Algeacuterien Alger Baconnier fregraveres Bruumlckner H (1986) Manrsquos impact on the evolution of the physical environment in the

Mediterranean region in historical times GeoJournal 13 (1)7ndash17 Brun A (1989) Microflores et paleacuteoveacutegeacutetations en Afrique du Nord depuis 30 000 ans

Bulletin de la Socieacuteteacute Geacuteologique de France 8(1)25ndash33 Cahen C (1968) Quelques mots sur les Hilaliens et le nomadisme Journal of Economic

and Social History of the Orient 11 (1)130ndash2 Courtois C (1955) Les Vandales et lrsquoAfrique Paris Arts et Meacutetiers Graphiques Decret F (1977) Carthage ou lrsquoEmpire de la Mer Paris Editions du Seuil Diehl C (1896) LrsquoAfrique Byzantine Histoire de la Domination Byzantine en Afrique

(533ndash709) Paris Imprimerie Nationale Dore JN and van der Veen M (1986) ULVS XV radiocarbon dates from the Libyan

Valleys Survey Libyan Studies 1765ndash8 Ducellier A (1988) Les Byzantins Histoire et Culture Paris Editions du Seuil El Bouch A and Ballais J-L (1997) Travertinisation deacutetritisme et anthropisation a Fegraves

(Maroc) Wuumlrzburger Geographische Arbeiten 92213ndash24 Fegravevrier P-A (1989) Approches du Maghreb Romain Aix-en-Provence Edisud Ibn Khaldoun A (1968) Muqqadima Beirut Commission Libanaise pour la Traduction

des Chefs drsquoOeuvre translated by VMonteil Idris HR (1968a) Lrsquoinvasion hilalienne et ses consequences Cahiers de Civilisation

Meacutedieacutevale 3353ndash71 Idris HR (1968b) De la reacutealiteacute de la catastrophe hilalienne Annales Economic Socieacuteteacute

Civilisation 23390ndash6 Al-Idrisi A (1983) Le Maghrib au 6e siegravecle de lrsquoHeacutegire Paris Publisud translated by


Hadj Sadok Jorda M Parron C Provansal M and Roux M (1993) Erosion et deacutetritisme holocene

en Basse Provence calcaire Lrsquoimpact de lrsquoanthropisation Travaux du Centre Camille Jullian 14225ndash33

Lamb HF Eicher U and Switsur VR (1989) An 18000-year record of vegetation lake-level and climatic change from Tigalmamine Middle Atlas Morocco Journal of Biogeography 1665ndash74

Lamb HH (1977) Climate Present Past and Future London Methuen Le Roy-Ladurie E (1967) Histoire du Climat depuis lrsquoAn Mil Paris Flammarion Leveau P (1974ndash75) Une valleacutee agricole des Neacutemenchas dans lrsquoAntiquiteacute romaine

lrsquooued Hallail entre Djeurf et Aiumln Mdila Bulletin Archeacuteologique du Comiteacute des Travaux Historiques et Scientifiques 10ndash11b103ndash21

Mahjoubi A (1978) Recherches drsquoHistoire et drsquoArcheacuteologie agrave Henchir El-Faouar (Tunisie) Tunis Publications de lrsquoUniversiteacute de Tunis

Marccedilais G (1946) La Berbeacuterie Musulmane et lrsquoOrient au Moyen Age Paris Aubier Modeacuteran Y (1988) Les premiers raids des tribus sahariennes en Afrique et la Johannide

de Corippus Histoire et Archeacuteologie de lrsquoAfrique du Nord 2479ndash90 Modeacuteran Y (1989) La deacutecouverte des Maures Reacuteflexions sur la lsquoreconquecirctersquo byzantine

de lrsquoAfrique en 533 Cahiers de Tunisie 43155ndash6 211ndash38 Pachur HJ (1974) Geomorphologische Untersuchungen im Raum des Serir Tibesti

(Zentrasahara) Berliner Geographische Abhandlungen 176ndash58 Poncet J (1967) Le mythe de la lsquocatastrophersquo hilalienne Annales Economie Socieacuteteacute

Civilisation 23660ndash2 Poncet J (1968) Encore agrave propos des Hilacircliens la mise au point de RIdris Annales

Economie Socieacuteteacute Civilisation 23660ndash2 Stockton CW 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 agrave la

Frontiegravere Tuniso-Libyenne Paris CNRS Trousset P (1986) Limes et frontiegravere climatique Histoire et Archeacuteologie de lrsquoAfrique du

Nord 55ndash84 Paris CTHS Vanacker C (1973) Geographic eacuteconomique de lrsquoAfrique du Nord selon les auteurs

arabes du IXe au milieu du XIIe siegravecle Annales Economie Socieacuteteacute Civilisation 3 659ndash80

Vita-Finzi C (1969) The Mediterranean Valleys Geological Changes in HistoricalTimes Cambridge Cambridge University Press


8 Success longevity and failure of arid-land

agriculture RomanomdashLibyan floodwater farming in the Tripolitanian pre-desert

DAVID GILBERTSON CHRIS HUNT AND GAVIN GILLMORE

INTRODUCTION

This chapter examines the successes and failures of RomanomdashLibyan and later floodwater farmers in the Tripolitanian pre-desert in Libya (Fig 81) This vast region of rocky plateaux and incised wadis lies between the higher better watered Gebel Nafusaand 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 wereused to manage occasional storm-water to sustain agricultural settlement with manyimpacts on soils geomorphology and biogeography (Fig 82) The vast scale of the ancient settlement stands in stark contrast to the depopulated modern landscape As longago as 1857 the explorer Heinrich Barth recorded that the landscape displayed a lsquosea-like level of desolationrsquo (1857125) Today the region remains empty and inhospitableexcept for a few pastoralists with mixed herds of sheep and goats The pastoralists exploitbore water rare springs and small wells In the hotter and drier parts of the year herdsmay be taken north to the better watered and cooler Gebel The modern towns of BeniUlid and Mizda are the only significant settlements in the region The area around BeniUlid is a dense mixture of modern development and remains of ancient buildingsevidencing substantial occupation from RomanomdashLibyan times to as recently as only 400 years ago (Gilbertson and Hunt 1988) The Wadi Merdum through Beni Ulid is alsoone of the last if not the last wadis in this region where active floodwater farmingcontinues Date palms figs plums and ancient olive trees can still be found growingalong 4ndash5 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

Figure 81 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 82 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 GBarker


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 fromthe Universities of Leicester Huddersfield Manchester and Sheffield in England Themain archaeological evidence is presented in a Gazetteer based upon 2437 site records(Barker et al 1996b) Many individual sites are themselves complex for example asingle entry deals with the complex networks of hundreds of substantial wadi walls thatwere 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 cursoryexamination of the distribution of large open (that is unenclosed or undefended) farmsand 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 83) The

Success longevity and failure of arid-land agriculture 135

Figure 83 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 1996162 165


latter are mainly attributed to the third to fourth centuries AD though many examplesalso 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-Punic-speaking Macae tribespeople here referred to as Romano-Libyan (Mattingly 1989) Initially transhumant pastoralists they took advantage of the extension of Romaninfluence into Tripolitania rapidly to develop a robust long-lasting mixed farming economy and a substantial increase in population in a desert environment broadly similarto 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 doubtcontinued through the RomanomdashLibyan period and it is transhumant pastoralism thatcharacterizes the human geography of the modern landscape A substantial trade tookplace 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 somewere exported to the wider Roman world Products produced by better-watered regions were imported into the pre-desert including even lsquoexoticrsquo foodstuffs such as deep-water sea fish

The density of RomanomdashLibyan and later settlement in the pre-desert varied significantly through both space and time (Flower and Mattingly 1995 Mattingly withFlower 1996) A dramatic transformation of pre-desert settlement with a rise in population to about 20000 in perhaps 2000 farms occurred in the study region duringthe first century AD The longevity of this occupation is rarely securely known Theestimated 1000 gsur built from the third century AD were perhaps fortified farms Thesemust have provided massive and secure storage of crops and food as buffers againstsequences of adverse drought years By the sixth to seventh centuries AD the entirenetwork had probably become consolidated into major lsquoagricultural estatesrsquo controlled by powerful rich local elites or warlords (Mattingly 1996) These lsquoestatesrsquo 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 thisintensification and reordering of settlement in the pre-desert were associated with a decreasing import of goods from the Mediterranean countries and decreasing quantitiesof olive oil sent to the coast (Mattingly 1996) GIS-based analyses suggest both a general lsquothinningrsquo of settlement and a slight northward and westward shift from the early extensive phase through this late RomanomdashLibyan period followed by a notable shiftnorthwards and a trend towards clustering of settlement during the Islamic period(Mattingly with Flower 1996) Eventually the only remaining large settlement withsignificant modern floodwater farming was Beni Ulid on the northern edge of the pre-desert

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 andis often incapable of sustaining sophisticated theoretical enquiry For example thelogistical reality of access by off-road vehicle in the difficult terrain and generally arduous circumstances forced a concentration of surveys upon the more accessible wadisat the expense of other wadis and the vast plateaux between them This survey pattern islikely to have under-represented features such as ancient farms in basins or on ancient


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 al1984) 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 al1996a) probably locating only 70ndash80 per cent of the walls initially present (Barker withGilbertson 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 forfield workers in this remote and politically sensitive area

COMPETING EXPLANATIONS

Reasons for the initiation growth stability and eventual decline of settlement andfarming in this pre-desert region remain remarkably unclear Some possible explanationsare set out below

First we consider the possible explanationsmdashsingly or in combinationmdashfor the abandonment of Romano-Libyan settlements and farms in the Tripolitanian pre-desert

Human processes

bull 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

bull Political social and economic changes at the coast and in the wider Roman empire brought about the loss of the lsquoRomanrsquo technocrats who made the system work

bull The arduous life did not give sufficient rewards to the farmers bull It was too unrewarding and unexciting for young peoplemdashthe opportunities and

good life in the coastal cities were too attractive bull Insecuritymdashraids and menace from the desert tribes to the south could no longer be

managed bull The first and second Arab invasions prompted the abandonment of the farms bull The quelling of nomadism by the development of farmssettlements 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

bull 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

bull 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

bull The region was drained by the activities of the equivalent of itinerant lsquocarpetbaggersrsquo who drained this region of its vitality and wealth

bull The question is based upon a misinterpretation the people were never fully


Natural processes

lsquoInduced environmental changersquo

Now we turn to the possible explanations for the expansion of these same settlements

sedentary and failed to return rather than leftmdashit was a threat to or loss of their mobility that was critical

bull A disease epidemic removed the capacity of people crops or livestock to continue in this demanding desert environment

bull The lsquolong-termrsquo climate became too arid bull 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

bull The inherent instability of the biophysical systems in dryland environments led to the growth of mutually reinforcing links between any of many possible cause-and-effect 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

bull There was a local version of the lsquoCharney Effectrsquo 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

bull 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

bull Accelerated soil erosion made arable and pastoral farming too difficult on the plateaux

bull Goats and sheep lsquoravagedrsquo the pastures bull Excessive trapping of water in soil produced soil salinization in plateau-basins and

on wadi floors bull 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

bull Excessive cropping caused the loss of soil fertility and unexpected crop failure bull The supply of timber for fuel and other domestic purposes was effectively

exhausted bull The inherent instability of the biophysical systems in dryland environments led to

the growth of mutually reinforcing links between any of many possible cause-and-effect 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


Human processes

Environmental processes

Several explanations appear to be contradictory some are counter-intuitive whilst others probably exceed the strength of the present evidence Several possible explanationscannot be investigated with the methods currently available Indeed many explanationswould be difficult to explore even in arid lands that have been the subject of sustained

bull 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 andor settlers

bull The southern lsquofrontierrsquo of the Roman empire was secured by a lsquodefence in depthrsquo made up of farming communities and some military installations

bull 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 lsquofrontierrsquo moved south bringing with it farmers and settlers

bull The arrival of Roman lsquoknow-howrsquo and lsquocan-dorsquo

bull The climate was lsquowetterrsquo (caused by cloud cover in greater quantity differently distributed more reliable andor more frequent) so agriculture prospered and settlement extended deeper into the pre-desert

bull 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

bull lsquoPioneerrsquo or lsquoeccentricrsquo people had experimented with small-scale cultivation (perhaps experimental in outlook) using ideas from indigenous people or elsewhere and started a lsquofashionrsquo that was thought worthy or otherwise good for personal development

bull The farms were started and maintained as a tax avoidance or tax mitigation measure bull 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

bull 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 andor the intensification of farming developments


intensive and extensive systematic studymdashwhich is far from the case here Through thefilter of nearly 2000 years it is difficult to separate cause and effect and lsquotriggersrsquo 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 ofprocesses will have operated at different times leading to a variety of outcomes Themost critical information and key ideas are set out below together with new evidenceproduced since the publication of the UNESCO survey in 1996

AGRICULTURE

The products of RomanomdashLibyan agriculture in the pre-desert are summarized in Table 81 (Gilbertson et al 1994 van der Veen et al 1996) They are essentially similar to theproducts of modern intensive mixed farming in many Mediterranean lands including theLibyan coastal plain and its hinterlands One noteworthy archaeozoological pattern wasthe increasing proportion of wild animals consumed with increasing distance south intothe 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 andmanure that domestic animals produced

There were many similarities to but also some differences from the more lsquonormalrsquo agricultural economy of the Mediterranean lands to the north For example the cropsproduced are similar to those of rain-fed agriculture to the north but the details of theagricultural practices used must have been notably different since precipitation in thepre-desert is both minimal and unreliablemdashless than 25ndash100 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 81 Farm products of the Tripolitanian pre-desert first to fifteenth centuries AD

Farm products Centuries ADPLANT CROPS 1ndash5 10ndash16 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 Sheepgoat + +


harvesting rainwater with particular emphasis upon managing overland flow andcontrolling floodwater in the wadi floorsmdashpractices originally termed lsquofloodwater farmingrsquo by Bryan (1929) in his studies in the American Southwest

Unfortunately the palaeoeconomic analyses that underpin these ideas derive from aminute subset of sites in the study region Nevertheless the ULVS survey design didattempt to ensure that material was available from the primary types of agriculturalbuildings located during the project The studies of seeds and animal bones that underpinpresent understanding derive from excavations at middens or in buildings at only fouropen 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 mid-Holocene to the period of instrumental records is vital to understanding its human historyUnfortunately such understanding is often rudimentary

Palaeoenvironment

Only four palaeoenvironmental studies have been reported previously Three of thesesuggested the cultivation of olives in the RomanomdashLibyan period A study of the sedimentary fill of a karstic plateau-basin north of Beni Ulid indicated the presence of awetter 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 4000 to5000 years ago creating an environment essentially similar to the modern arid steppe Astudy of cave deposits near Beni Ulid indicated the essential similarity of Romano-Libyan and modern conditions (Gale et al 1993) A third study analyzed pollen fromsediments infiltrated into a RomanomdashLibyan cross-wadi wall in the Wadi Mansur (Hunt et al 1986) suggesting a degraded steppe flora very similar to that of the modern Wadi

Gazelle + + Bovid + + Pig + Canid + + Camel + + Harerabbit + + Equid + + Antelope + + The present archaeobotanical evidence suggests that there were no fundamental differences between the agricultural economies of the RomanomdashLibyan Late Antique and Islamic periods In general hunted as opposed to herded animals became increasingly more important further south into the desert


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 thanoccurs today Cereals were cultivated The high frequency of charcoal recovered suggestsburning 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 RomanomdashLibyan deposits further south at Ghirza (van der Veen et al 1996 Fig 84) During the nineteenth century in South Australia a system of mediccereal rotation was developed by dryland farmers to improve nitrogen levels intheir soilsmdasha 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 aregrown in the second seeds from the first medic pasture then germinate to create anotherpasture in the third season The adoption of this lsquoAustralianrsquo system in parts of the Tripolitanian and Cyrenaican Gebel in the 1970s and 1980s increased cereal yields by asmuch as 50 per cent and allowed stocking rates to rise dramatically Chatterton andChatterton (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 haveimproved soil fertility and grazing Such a scenario is probable because in many areasrainfall would not have been sufficient every year to justify planting a cereal crop Overtime 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 hereThis 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 degradedenvironment with low local


Figure 84 A RomanomdashLibyan fortified farm (gasr) and its satellite buildings at Ghirza

Kite photograph GDBJones

pollen productivities and the local flora dominated by drought-resistant species In contrast the samples from contexts dating to the RomanomdashLibyan and Arab periods contain pollen of grasses and a diverse steppic flora with abundant pollen of cereals andolives reflecting crop plants Critically also the taphonomic patterns suggest that animalswere fed on monoculture cropsmdashgrasses cereal waste and chenopods Evidence from


Lm4 had previously suggested the stalling of animals at this site (Gilbertson and Hunt1990) This is a very different type of husbandry pattern than occurs today in thepredesert where goats forage widely and indiscriminately

Palaeoclimatology

Palaeoclimatic evidence from Morocco and the Nile basin suggests severe prolongedLate Holocene drought events The Late Holocene was notably drier than much of theEarly and Middle Holocene (Hassan 1981 1996 1997 1998 Lamb et al 1994 1995) Significant arid phases were identified for 4600ndash4000 BP 3800ndash3600 BP 2500 BP 2000 BP 1500 BP and approximately 700 BP (radiocarbon years) The flood record of theRiver Nile is especially interesting for the last 1500 years indicating low to very lowflows from AD 760 to 1070 with especially low flows between AD 930 and 1070 andbetween 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 thestudied sites The interior of Libya is heterogeneous and environmentally complex andclimate changes occurring elsewhere in North Africa may not necessarily havemanifested themselves there in quite the same way Gilbertson and Hunt (1996) andNicholson (1989 1994) describe the regional climatology The quantity variability andreliability of precipitation are not well known In general annual precipitation averagesbelow 100 mm north of Beni Ulid to less than 25 mm in the south Thus Wadi Umm el-Kharab and Wadi el-Amud can be anticipated to have an unreliable and variableprecipitation regime averaging about 30 mm a year Nowadays drought may occur inmany consecutive years A yearrsquos 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 lsquodesert farmingrsquo was not sustained by harvesting rainwater or floodwater Rather it was supported by a reliableunderground water supply perhaps a spring as at Lemasba Algeria (Shaw 1982) oroases as in the Fezzan (Mattingly this volume Chapter 9 van der Veen 1992) Spring-fed 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 alluvialfans may well have developed on several occasions during the Holocene Overall in theperiod during and after extensive farming it is evident that there were several episodes ofslope erosion fluvial aggradation incision and aeolian reworking Anthropogenicdepositsmdashlarge middens layers of ash or dungmdashoccur notably at the olive farm at Wadiel-Amud (Gilbertson and Hunt 1990)


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 floodplainsediments whilst intervening gullies can be scoured over 1 m deep leaving the trees oneroded pinnacles Elsewhere modern barley is successfully grown on the level surface ofrecently deposited flood-loams Cross-wadi walls promote fine-grained sedimentation and resultant increases in soil moisture seed catch and shrub growth on their upstreamside Downstream from walls waterfall-effects during flood promote gullies Later long after the walls are over-topped subsequent subsurface flow may promote temporarysprings sapping and piping These observations led to the development of a spatial modelto explain agricultural practice on wadi floors (Fig 85) The model also predicts where browse and shelter would have been available for stockmdashand thus an immediate source of the manure necessary to sustain intensive cereal cultivation (Barker et al 1996a Chatterton and Chatterton 1984)

Figure 85 Model of RomanomdashLibyan 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 animalsSubsurface processes are less securely known Field and laboratory evidence indicatesthat near-surface water is sometimes saline and it is not unreasonable to question whethersoil salinization may have been locally important in the past especially given the


deliberate introduction of large quantities of water on to wadi floors At present there isno evidence for large-scale salinization of wadi floors in RomanomdashLibyan 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 ancientfarming in the pre-desert (Fig 86) The vital role of walls in facilitating ancient farmingin drylands by trapping water and sediment has been recognized by numerous

Figure 86 Walls in the desert wall systems in the Wadi Mimoun near Gasr Lebr (Mm10)

Source After Hunt et al 1987


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 ofrain-storm in the American Southwest (Bryan 1929 Nabhan 1986a 1986b Nabhan andSheridan 1977 and see Minnis Chapter 15) However fences for water control have notbeen detected in the pre-desert although thorn scrub is still in widespread use to corralanimals

The antiquity of walls

Inevitably the age relationships of walls are often unknown They are inferred fromnearby archaeological features whose antiquity was typically determined from associatedceramicsmdasha necessary first assumption but one that may often be incorrect It is alsoclear that walls are likely to have been frequently rebuilt reused repaired repositioned orreformed In some areas most walls were perhaps associated with the second majorphase of Romano-Libyan settlement characterized by the construction of gsurElsewhere many walls relate to the open farms of the earlier RomanomdashLibyan settlement phase Some may even be older and associated with the modest numbers oflater prehistoric settlements other walls may post-date the RomanomdashLibyan 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 quantitymdashand perhaps the ratemdashat which run-off was delivered to the wadi floor Often water was led from the plateau or hill slopes into cisterns manywith sediment traps or into caves at the wadi edge On the wadi floor the primaryobjective was apparently to capture floodwater causing it to sink into the long-term storage provided by the wadi-floor alluvium Occasionally water was conducted intocisterns adjacent to ancient settlements Many cisterns remain in use today or at leastthey still function Erosion appears to have been understood and managed by theRomano-Libyan farmers Numerous wadi-floor walls contain lsquodrop structuresrsquo 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 overwhelmedand breached during flood peaks the reinforced surfaces downstream prevent scour andgully erosion These features still appear to be operating effectively with few displayingevidence of damage

A substantial literature describes the role of wall-managed floodwater for contemporary dryland management and developmentmdashnotably 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) Eventhough 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 harvestand channel run-off and storm water with marked ecological and biogeographic consequences (Gilbertson et al 1994)


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 desertrainfall Substantial permanent investment of time and human energy in the constructionand maintenance of walls would have been necessary People had to sow seed on wettedsoil 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 suchlocations 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 earlierfloods

WALL FUNCTIONS

Six hypotheses have been proposed that singly or in combination might explain thefunction of the wall systems observed in the pre-desert (Gilbertson et al 1984) Walls may

These functions are not necessarily exclusive The same wall may have had one or moreuses when first constructed and later may have acquired or lost other roles

Walls whose primary purpose was clearly to delineate ownership have not been foundin 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 atvarious points along the length of the wadi (Gilbertson and Chisholm 1996) It is notknown how widespread this practice was Overall most walls appear related tohydrologicalgeomorphic 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 tenureor managementmdashif it was marked on the groundmdashoperated within the hydrological constraints of the wall systems The archaeological consequence is that it is very difficultto distinguish factors such as past community ownership or social groupings from thepresent information on wall networks

bull capture store and redistribute surface water for human and animal consumption and irrigation

bull control fluvial erosion sediment entrainment transport and deposition bull 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

bull delineate areas of different land use bull represent the by-products of stone clearance to ease cultivation bull define parcels of land owned or controlled by different individuals or groups


ENVIRONMENTAL INFLUENCES

Within the long periods in which the wall systems were in use many droughts and otherenvironmental vicissitudes must have occurred Indeed the evidence describedpreviously from the Nile basin and Morocco suggests that droughts of 10ndash50 yearsrsquo duration probably occurred several times during the last two millennia Large parts of thepre-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 temporaryabandonment yet existsmdashat present both the environmental and archaeological data aretoo coarse to distinguish such events Similarly there are no clear indications that thedevelopment and major shifts in the form and distribution of RomanomdashLibyan settlement or its subsequent decline were associated with some form of climatic changeor fluctuation Nevertheless the existing palaeoecological and palaeoclimatic informationfrom the study area suggests that the climate during much of the period of RomanomdashLibyan settlement was not dissimilar to that which prevails today though vegetation wasgenerally 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 modernecological 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 desertificationdoes not seem to have played a central role in transforming the widely farmed and settledRomano-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 socialworlds (Barker et al 1996a Mattingly 1996) In brief the prime factor encouraging the Macae tribal pastoralists to become sedentary floodwater farmers appears to have beenregional stabilization resulting from the expansion of Roman influence into TripolitaniaEffective incorporation into the wider imperial economy produced different patterns ofland use greater stability access and trade with the vast new market and a majorincrease in population There are no grounds for believing that a widespread militarycolonization by soldier-farmers (limitanei) or a frontier army ever played a significantrole as was once suspected (Goodchild and Ward-Perkins 1949) Neither is there any indication that the expertise ideas or technology of floodwater farming were introducedfrom outside Probably the desert dwellers developed these approaches indigenously Asdemonstrated in many chapters in this volume the essence of this technology wasrepeatedly invented in antiquity in very different places


It remains unproven whether the replacement of open undefended farms by gsurshould be related to a greater sense of insecurity or whether the rich and powerful of thetime adopted these imposing enclosed structures to follow fashion as a display ofprestige or because the shade size and airiness of such buildings were well adapted tothe rigours of desert life (Fig 84) The progressive abandonment of settlement and farms in the southern part of the pre-desert region is perhaps best attributed to the widerpolitical and economic changes throughout the Mediterranean at the end of the Romanera with the development of smaller more regionalized group identities (Mattingly1996) There are no grounds for suspecting that the arrival of the Arab armies in the AD640s and later brought about major changes in pre-desert settlement or farming Floodwater farming was to continue at smaller scales for another thousand years indeedit 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 farmingsystems as the result of climatic economic or political change are not congruent with thehistory of the pre-desert as presently understood It is also clear that ecologicaldegradation of the landscape for example at the Lm4 farm post-dated the end of RomanomdashLibyan floodwater farming Floodwater farming seems to have come to an endgradually on a piecemeal basis in some areas though there may have been rapid earlyretreats from the southernmost outposts such as Wadi el-Amud as these became uneconomic with the collapse of long-distance trade networks in late RomanomdashLibyan times It is clear that partial use of systems such as at Mm10 and Lm4 continued afterformal use of the RomanomdashLibyan buildings ended People continued to grow cereal crops and keep stalled animals though they often no longer lived in the RomanomdashLibyan buildings At this stage the landscape still had a distinctly steppic aspect

The population of the pre-desert was never very large For maximum efficiencylabour-intensive maintenance of the wall networks is essential One might envisage that as Roman influence waned and the political landscape became unstable intensiveinvestment in farming complexes became a risky strategy People began to readoptlsquobedouinrsquo ways of life which are flexible and in many ways less arduous than living infixed settlements in this region As people abandoned buildings for tents a transhumantlifestyle became possible and people started to move to where rain had fallen mostintensively each year to grow their crops Because of increased mobility it was no longerimportant that walls were rigorously maintained Systems would be abandoned as they became inefficient The end of animal stalling as seen at Lm4 would have placedadditional stress on the landscape because grazing removed the steppe vegetation and ledto the modern pre-desert ecology Rapid alluviation events in the medieval or early post-medieval periods may perhaps have been linked with landscape degradation of this type(Barker with Gilbertson 1996 Gilbertson and Hunt 1996a)


CONCLUSION

The systems of floodwater farming developed by Romano-Libyan farmers in the Tripolitanian pre-desert seem to have been lsquosustainablersquo in the true sense of the word They have persisted in some localities for two millennia surviving the fall of empiresmajor economic catastrophes climate fluctuations and changes and other adversitiesMuch of the resilience of the floodwater farming systems is clearly the result of theexploitation 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 ofpatchy and unreliable storms Their farming systems seem to have been well adjusted tolocal conditions Details are however still sparse The hypothesis of ageomorpbiologically-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 RomanomdashLibyan farmers stalled their stock is significant becauseanimals 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 linkedto most of the cited lsquopush-factorsrsquo such as the fall of Rome which were relatively rapidIt may be that the bedouin lifestyle simply became more attractive to the small populationof the Tripolitanian pre-desert

REFERENCES

Anketell MJ Ghellali SM Gilbertson DD and Hunt CO (1995) Quaternaryfloodplain and wadi floor infill deposits in northeastern Libya and their implicationsfor landscape development In JLewin MMacklin and JMWoodward (eds)Quaternary Mediterranean River Environments 231ndash44 Amsterdam AA Balkema

Barker G with Gilbertson DD (1996) Farming the desert retrospect and prospect InGBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis 343ndash63 Paris UNESCO Publishing

Barker G and Gilbertson DD with Hunt CO and Mattingly DJ (1996) RomanomdashLibyan agriculture integrated models In GBarker DDGilbertson GDBJones andDJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis 265ndash90 Paris UNESCO Publishing

Barker G Gilbertson DD Jones GDB and Welsby DA (1991) The UNESCOLibyan Valleys Survey XXIII the 1989 season Libyan Studies 2231ndash60

Barker G Gilbertson DD Jones GDB and Mattingly DJ (1996a) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 SynthesisParis UNESCO Publishing

Barker G Gilbertson DD Jones GDB and Mattingly DJ (1996b) Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 2 Gazetteer and


Pottery Paris UNESCO Publishing Barth H (1857) Travels and Discoveries in North and Central Africa London

Longman Bryan RK (1929) Floodwater farming Geographical Review 19 (3)444ndash56 Chatterton BA and Chatterton L (1984) Medicagomdashits possible role in Romanomdash

Libyan dry farming and its positive role in modern dry farming Libyan Studies15157ndash60

Dorsett JE Gilbertson DD Hunt CO and Barker G (1984) The UNESCO LibyanValleys Survey IX image analysis of Landsat data and its application to environmentaland archaeological Surveys Libyan Studies 1571ndash80



Flower CP and Mattingly DJ (1995) ULVS XXVII mapping and spatial analysis ofthe Libyan Valleys Data using GIS Libyan Studies 2649ndash78

Gale SJ Gilbertson DD Hoare PG Hunt CO Jenkinson RDS Lamble APOrsquoToole C van der Veen M and Yates G (1993) Late Holocene environmental change in the Libyan pre-desert Journal of Arid Environments 241ndash15

Gilbertson DD (1986) (ed) Run-off Farming in Rural Arid Lands Theme Volume 6 (1 and 2) of Applied Geography

Gilbertson DD (1996) Explanations environment as agency In GBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCOLibyan Valleys Archaeological Survey Volume 1 Synthesis 291ndash318 Paris UNESCO Publishing

Gilbertson DD and Chisholm NT (1996) Manipulating the desert environmentancient walls floodwater farming and territoriality in the Tripolitanian pre-desert of Libya Libyan Studies 2717ndash52

Gilbertson DD and Hunt CO (1988) The UNESCO Libyan Valleys Survey XIX theCenozoic geomorphology of the Wadi Merdum Beni Ulid in the Libyan pre-desert Libyan Studies 1995ndash121

Gilbertson DD and Hunt CO (1990) The UNESCO Libyan Valleys Survey XXIgeomorphological studies of the Romano-Libyan Farm its floodwater control structures and weathered building stone at site Lm4 at the confluence of the Wadi elAmud and the Wadi Umm el Bagul in the Libyan pre-desert Libyan Studies 21 25ndash42

Gilbertson DD and Hunt CO (1996a) Quaternary geomorphology and palaeoecologyIn GBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis 49ndash82 Paris UNESCO Publishing

Gilbertson DD and Hunt CO (1996b) RomanomdashLibyan agriculture walls and floodwater farming In GBarker DDGilbertson GDBJones and DJ MattinglyFarming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1Synthesis 191ndash216 Paris UNESCO Publishing

Gilbertson DD Hayes PP Hunt CO and Barker G (1984) The UNESCO LibyanValleys Survey VII a classification and functional analysis of ancient irrigation and


wall systems in the Libyan pre-desert Libyan Studies 1545ndash70 Gilbertson DD Hunt CO and Fieller NRJ (1993) ULVS XXVI sedimento-logical

and palynological studies of Holocene environmental changes from a plateau basininfill sequence at Grerat Drsquonar Salem near Beni Ulid in the Tripolitanian pre-desert Libyan Studies 241ndash19

Gilbertson DD Hunt CO Fieller NRJ and Barker G (1994) The environmentalconsequences and context of ancient floodwater farming in the Tripolitanian pre-desert In ACMillington and KEPye (eds) Environmental Change and GeomorphicProcesses in Arid Lands 229ndash51 Chichester John Wiley and Sons

Goodchild RG and Ward-Perkins JB (1949) The Limes Tripolitanus in the light of recent discoveries Journal of Roman Studies 3981ndash95

Hassan FA (1981) Historical floods and their implications for climatic change Science2121142ndash5

Hassan FA (1996) Abrupt Holocene climatic events in Africa In GPeti and R Soper(eds) Aspects of African Archaeology 83ndash9 Harare University of ZimbabwePublications

Hassan FA (1997) Holocene palaeoclimates of Africa African Archaeological Review14 (4)213ndash30

Hassan FA (1998) The archaeology of North Africa at Kiekrz 1997 African Archaeology Review 15 (1)85ndash93

Hunt CO Mattingly DJ Gilbertson DD Barker G Dore JN Burns JRFleming AM and van der Veen M (1986) The UNESCO Libyan Valleys SurveyXIII interdisciplinary approaches to ancient farming in the Wadi Mansur TripolitaniaLibyan Studies 177ndash47

Hunt CO Gilbertson DD van de Veen M Jenkinson RDS Yates G andBuckland PC (1987) The UNESCO Libyan Valleys Survey XVII the palaeoecologyand agriculture of the abandonment Phase at Gasr Mm10 Wadi Mimoun in theTripolitanian pre-desert Libyan Studies 181ndash14

Lamb HF Duigan CA Gee JHR Keits K Lister G Maxted RW Merzouk ANiessen F Tahri M Whittington RJ and Zeroual A (1994) Lacustrinesedimentation in a high altitude semi-arid environment the palaeo-limnological record of Lake Isli High Atlas Morocco In ACMillington and KEPye (eds)Environmental Change and Geomorphic Processes in Arid Lands 229ndash51 Chichester John Wiley and Sons

Lamb HH Gasse F Benkaddour A El Hamouti N van der Kaar S Perkins WTPearce NJ and Roberts CN (1995) Relations between century-scale Holocene arid intervals in tropical and temperate zones Nature 373134ndash7

Mattingly DJ (1989) Farmers and frontiers exploiting and defending the countryside ofRoman Tripolitania Libyan Studies 20135ndash53

Mattingly DJ (1996) Explanation people as agency In GBarker DDGilbertsonGDBJones and DJMattingly Farming the Desert The UNESCO Libyan ValleysArchaeological Survey Volume 1 Synthesis 319ndash42 Paris UNESCO Publishing

Mattingly DJ with Flower C (1996) RomanomdashLibyan settlement site distribution and trends In GBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis


159ndash90 Paris UNESCO Publishing Nabhan GP (1986a) Papago Indian desert agriculture and water control in the Sonoran

Desert 1697ndash1934 Applied Geography 66 (1)42ndash3 Nabhan GP (1986b) lsquoAk-cintilde lsquoarroyo-mouthrsquo and the environmental setting of Papago

Indian fields in the Sonoran Desert Applied Geography 6 (1)61ndash75 Nabhan GP and Sheridan TE (1977) Living fencerows of the Rio San Miguel

Sonora Mexico traditional technology of floodplain management Human Ecology597ndash111

Nicholson SE (1989) Long term changes in African rainfall Weather 4446ndash56 Nicholson SE (1994) Rainfall fluctuations in Africa and their relationship to past

conditions over the continent The Holocene 4(2)121ndash31 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 BD (1982) Lamasba an ancient irrigation community Antiauiteacutes Africaines 18

61ndash103 van der Veen M (1992) Garamantian agriculture the plant remains from Zinchecra

Fezzan Libyan Studies 237ndash39 van der Veen M Grant A and Barker G (1996) RomanomdashLibyan agriculture crops

and animals In GBarker DDGilbertson GDBJones and DJMattingly Farming the Desert The UNESCO Libyan Valleys Archaeological Survey Volume 1 Synthesis227ndash64 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 Inmany cases pioneer studies have not been followed up in recent decades (Ball andBeadnell 1903 Fakhry 1974 RSGI 1937 Scarin 1934 1937) Yet there isundoubtedly much to learn from the manner in which desert people have exploitedresources and mastered the limitations of their environment A better understanding ofhuman adaptation to the desert environment has clear relevance for modern concernsabout 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 (1997ndash2000) The project is investigating the archaeology of a region of the Libyan Sahara c1000 km south of Tripoli (Fig 91) and follows on from earlier British work carried out by the late Charles Daniels Hisexploration and excavations from 1958 to 1977 accumulated a vast dossier of informationon one of the most important Saharan peoples of classical antiquitymdashthe 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 c500 BC to cAD 500 and at the height of their influence they controlled a vast desert territory ofc250000 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 theGaramantian capital They were several times defeated by Roman armies sent againstthem but their territory was never annexed to the Roman empire and for much of theRoman period they seem to have thrived on a combination of oasis agriculture and trade(Mattingly 199533ndash7 68ndash77 on relations with Rome)

Figure 91 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 regionmdashbroadly 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 aconcern with human interaction with the environment and a study of how this has variedover an extended period of time and changing conditions The Fezzan Project then hasrelevance to wider debates than parochial Libyan ones though the evolution of Libyanculture and society in the desert is of major importance in its own right (Bates 1914Brett and Fentress 1996 Camps 1980) The transition to farming the emergence ofsocial complexity and the formation of a distinctive Saharan culture were all achieved ina 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 atiny populationmdashLibya itself has under five million people most concentrated in thecoastal cities The desert is an extraordinarily difficult environment to live well in and asin most desert regions water is a critically scarce resource in large parts of the countryMost of Fezzan has negligible annual rainfall and today depends entirely on subterraneanfossil water sources for sustaining its human population its livestock and its areas ofcultivation In this respect the region has very different characteristics to the Libyan pre-desert zone between Fezzan and the Mediterranean coast studied in an earlier project

Twelve thousand years of human adaptation in Fezzan 157

(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 isone of people evolving strategies to utilize more effectively the huge groundwaterresources at the limited number of locations where they are relatively accessible at orclose to the surface Over the last few millennia the groundwater levels appear to havefallen significantly leading to small lakes drying up spring-lines ceasing irrigation systems being abandoned wells being deepened and so on As described below thestrategies evolved to tap into the groundwater were at certain times highly sophisticatedand 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 thematerial prosperity of the region also reached its apogee

The landscapes of Fezzan are very variablemdasha mixture of great sand seas gravel andboulder-strewn wastelands and hyper-arid rock plateaux (Fig 91) 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 aseries of aquifers The traditional pattern of cultivation involves scattered palm groves with intensively irrigated plots of wheat barley and sorghum Duveyrier (1864147ndash216 439) and Lyon (1821270ndash78) both give good accounts of Fezzanese plants and cultivation systems in the nineteenth century There are no perennial springs here todaythough there are hints in the landscape that at some point in the past there was an activespring-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 waterlevels in many wells to fall by up to 100 m in the last century As we shall see there areimportant lessons for the present to be learnt from the past history of human activity andover-exploitation of this resource

A vast sand sea (ergedeyen) 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 andClapperton 1826177ndash85) 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 Inthe neolithic and classical periods there were undoubtedly more of these lakesfacilitating 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 ofthe 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 lesserextent in the mesolithic and neolithic periods There are no perennial water sources onthe plateau itself the main aquifer lies deep beneath it but this is the one part of theFezzan to receive rain with any regularity at certain times of year pools of water can befound in the wadi beds along with some rough grazing Engraved rock art dating to boththe later prehistoric and historic periods is abundant in many of the hamada wadisrepresenting seasonal exploitation of this forbidding landscape by hunters and mobilepastoralists (Lutz and Lutz 1995) Even at the peak periods of oasis cultivation it is clearthat pastoral groups operated alongside cultivators in exploiting the potential of the desert


landscapes (Nicolaisen and Nicolaisen 1997 UNESCO Nomades 1962) The consensus scientific view is that the Saharan region in the middle palaeolithic

period (say 250000ndash40000 BP) was very much wetter than today with an abundance of vegetation and wildlife flourishing around vast inland lakes (Petit-Maire 1982 Petit-Maire et al 1980 Ziegert 1995) Thereafter in the late Pleistocene and early Holoceneconditions became subject to a series of dramatic swings from wetter to drier conditionsand back again (Fig 92) The neolithic period (c6000ndash1000 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 92 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 Sadek1980) may have been more abundant and more readily accessible at that time (moresprings small lakes and shallow aquifers) with wild fauna more diverse as a resultNeolithic rock engravings show that at the start of this period the Sahara supported alarge and rich wild fauna including species like the crocodile which require permanentwater but that this was crucially changed with increasing aridity leading to extinction ofmany species north of the Sahara and to major changes in the lifestyle of the survivinghuman groups (Barker 1989 Encyclopeacutedie 1997 Le Quellec 1987 Lutz and Lutz 1995 Mori 1969 1988) In the rock art we see evidence of the domestication of animalsand an increasing emphasis on fertility and ritualmdashperhaps reflecting the social stress caused by environmental change (Encyclopeacutedie 19972791ndash96 2800ndash2 Lutz and Lutz 1995145ndash65 169ndash75) The same trends towards domestication of animals are evidentalso in the few well-excavated neolithic rock shelters (Barich 1987 Cremaschi and di Lernia 1998)


AIMS OF THE FEZZAN PROJECT

The project addresses a number of key questions

The project geomorphologists have been studying the regional hydrology using field dataand remote sensing techniques to map gypsum formations which are indicative ofancient springs and small dried-up lakes sampling palaeo-lake sediments cross-sectioning spring mounds and assessing dune morphology (Mattingly et al 1998b117ndash22 1999b129ndash31) 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 toestablish whether they are of Pleistocene or Holocene date Laboratory techniques being used in support of the programme of palaeoenvironmental reconstruction include stableisotope analysis particle size distributions and mineral magnetic analysis of the putativelake 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 theescarpment With these methodologies we are hoping to be able to track through well-dated contexts the shrinkage and disappearance of the lakes with the onset of desiccationperhaps 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 inantiquity) (Fig 93) 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 onanother to a depth of 4ndash5 m Some earlier clearance excavation (Ayoub 1967a) hasrevealed a group of Garamantian buildings at the core of the site Unlike most of the laterstructures these have stone walls and reflect the power and wealth of the site in itsheyday in the period between the first and fourth centuries AD The origins of the

bull the transition to farming in the Saharan region and in particular the origins of agriculture

bull the diffusion or invention of farminghydraulic technology and the spread of different cultivated plants

bull the response of human populations to the climatic and environmental changes bull the origins of urbanization in the Sahara and its evolution over time bull the construction of identity through material culture bull inter-regional contact across the Sahara (trade) bull processes of desiccation (and desertification) in the northern Sahara bull the recognition of palaeo-hydrological features in the landscape (spring-lines lakes

marsh) bull the dating of changes in the hydrology bull the identification and dating of evidence of climatic and environmental change


settlement go back until at least the fifth century BC again with a sequence of mud-brick buildings

Figure 93 The settlement of Germa (ancient Garama) the capital of the Garamantes

Photograph DMattingly

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 timemdashsome phases are unmistakably more impoverished than others For instance much of themedieval and early modern periods is characterized by relatively low numbers ofimported 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 ceramicfinewares 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 indesiccated form is significant because it means that a reasonably full range of cultivatedand weed species is represented in the samples This gives useful information about thelocal environment the cultivated plants all require irrigation and the weeds often reflectthe arid and salty background conditions We are also identifying a series of significantbotanical horizons including a lsquomaize horizonrsquo representing the coming of New World crops and a lsquosorghum horizonrsquo representing northward transference from the Sahel orSudan probably within the Garamantian period This study builds on earlier work by vander Veen (1992) and is extending our knowledge of changing patterns of plant cultivationback from the present to c900 BC (for the broader North African context see van der Veen 1995 1999) Analysis of the faunal remains is also indicating change over time


with the bones providing important information about not only what stock was raised butalso 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 Thisaims to build upon Danielsrsquo earlier survey which was successful in locating a large number of cemeteries of Garamantian date particularly in the form of cairn graves alongthe foot of the hamada escarpment He was less successful in locating Garamantiansettlements though he was aware of a few village-like sites in the oasis and a number ofhillforts along the escarpment Systematic fieldwalking has now revealed that theGaramantian settlement pattern was far denser than previously suspected with numeroussatellite villages all around Old Germa (Fig 91) The fieldwalking essentially logs thedensity of archaeological material (humanly-made or imported goodsmdashlithics ceramics ostrich shell beads and so on) and isolates significant concentrations of such material aslsquositesrsquo Topographic survey of a selection of these sites has added structural detail andconfirms that we are dealing with settlements and not simply rubbish disposal Thesurvey complements the evidence of a series of excavations by Daniels on additionalGaramantian sites which we are also preparing for publication Zinchecra an earlyGaramantian hillfort and cemetery (Daniels 1968) Saniat Ben Howedi a rich Romanperiod cemetery (Ayoub 1968 Daniels 1989) and Saniat Gebril an oasis village

Our fieldwork has discovered sites of many different phases of activity not simply theGaramantian phase On the hamada to the south of Germa we have recorded a series ofimportant lithic scatters of palaeolithic date comprising tools such as 100000-year old handaxes together with chunky waste flakes and chippings at the locations where toolswere produced In our 1999 season a series of neolithic occupation sites was alsoidentified close to the edge of the sand sea to the north of Germa These sites yieldingextremely finely worked lithics early pottery grindstones ostrich eggshell fragments andbeads were probably occupied in the last few millennia BC when climatic conditionswere rapidly deteriorating Their inhabitants seem to have exploited a shallow and nowvanished lake site

A gazetteer of ancient sites throughout the el-Agial is being compiled combining boththe Danielsrsquo material and the new work Transcription of a series of air photographs takenin the 1950s and 1960s is revealing a wealth of information now destroyed by moderndevelopment This work complements a programme of remote sensing using modernsatellite imagery comparison of the satellite imagery and the air photographs hasrevealed 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 regionalwater 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 importantand enigmatic classes of monumentmdashthe 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 Locirc 1953 1954 Mattingly et al 1998a 190ndash2 1998b137ndash42 1999b139ndash42 Nesson et al 1973) They are readily identifiable at the surface where traces survive from the regularly spaced vertical shafts


that were dug to facilitate construction and maintenance of the channels though theymust have added hugely to the labour involved The shafts can be up to 20 m deepgradually diminishing in depth until the channel emerges at the surface (Fig 94) The available dating evidence indicates that the foggara system was introduced to Fezzanduring the Garamantian period with their use probably extending into the early Islamicperiod It is clear that these structures were a key to ancient irrigation in the regionthough evidently they have been dry for many centuries now There are many hundredsof these structures visible on the air photographs most being at least several km inlength The labour involved in their construction and maintenance was on a significantscale (Mattingly et al 1999b140ndash1)

HISTORICAL RECONSTRUCTION

What does all this new evidence add up to It is clear that there has been dramatic changein environment climate and human activity over time What

Figure 94 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 attachedThe reconstruction proposed at this stage is essentially a series of models designed forfurther testing and elaboration The clear trend running through though is one of anoverall decrease in water availability over time Climatic change and the onset ofdesertification have reduced rainfall to negligible levels and caused old surface watersources such as lakes and springs largely to dry up

During the Upper Pleistocene the region is known to have been very different from thedesert environment it has become The hamada plateau is assumed to have been well-vegetated savanna with abundant rainfall supporting a large range of animals and hunter-gatherer human groupsmdashwhose tool assemblages occur in profusion across its surface It is generally agreed that last phase of the Pleistocenemdashthe period c40000ndash10000 BPmdashwas one of high aridity in North Africa reflected in the Fezzan in a dearth of evidence for


the characteristic lithic assemblages of this phase The reappearance of substantial humanpopulations in the Early Holocene after 10000 BP can be related to a new period ofincreased rainfall (Fig 95) The landscape was well vegetated in this phase supporting a wide range of wild animals which were initially exploited through hunting especially onthe plateau and wadis of the hamada However in successive phases of further climaticchangemdashwhether major oscillations as indicated on Figure 92 or a more step-like progression towards acute aridificationmdashhuman settlement became increasingly focusedon locations where water was to be found at shallow depth Thus many sites presumablyseasonal camp-sites have been located in the el-Agial depression and around small lakeson the edge of the sand sea Because the mesolithic and neolithic phases were far from auniform period climatically it is necessary to undertake more work on the phasing ofsites of these periods through further analysis of tool types rock art phases and the

Figure 95 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


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 95) 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 canbe dated only in relative terms at present) and from some of the excavated rock sheltersThe exploitation of plant resources is most clearly signalled by the abundant grindstonesat the neolithic campsites by the lakes and water sources What is particularly interestingabout this transition in Libya is that it seems to arise as a response to adversity rather thanto opportunity people turned to stock raising and cultivation here during the fifth andfourth millennia BC when a dramatic change in the availability of water made a hunter-gatherer existence increasingly more precarious (Barker 1989 1996) The Fezzan Projectwill hopefully make an important contribution to these debates It is likely though hasnot yet been demonstrated that neolithic farmers grew their crops in small patches of soilnaturally irrigated by higher groundwater levels in contrast with the floodwater farmingsystems 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 Theperiod of the Garamantes however (between 900 BC and AD 500) marked a dramaticdevelopment in farming technologies and systems associated with transformations incultural complexity These transformations included

The Garamantes represent in part a continuation of the local neolithic tradition as is clearfrom lithic and ceramic finds at their early settlements But they probably comprised agreat confederation of tribes and there are indications that some elements may havemigrated from oases further east nearer Egypt bringing with them knowledge ofimproved technology for oasis cultivationmdashnotably the foggara There are clear parallels

bull the rise of a major polity and civilization in the Sahara (Daniels 1970 Ruprechtsberger 1997)

bull the development of urbanism (Daniels 1971262ndash5) bull the evolution of a hierarchical and probably slave-using society (Daniels 197027ndash

35) bull the adoption of a written script for the Libyan language (Daniels 1975) bull the further development of agriculture to encompass a range of Mediterranean and

desert crops that require intensive irrigation (cereals grapes olives dates) (Daniels 198956ndash58)

bull the introduction of the horse the camel and wheeled transport to the Sahara (Camps 1989)

bull the creation of trade and political relations that extended north to the Mediterranean east to Egypt and south to sub-Saharan Africa (Bovill 19681ndash44 Fontana 1995) and

bull a massive demographic expansion to a level that was probably not equalled again until the last forty yearsmdashDaniels (198949) estimated that there were at least 120000 Garamantian burials in the el-Agial alone


for instance between the Libyan tribesmen on Egyptian reliefs and in rock art of southernLibya and Algeria (Lutz and Lutz 1995140ndash1 Ruprechtsberger 199766ndash9) Most of the early Garamantian settlements currently known are situated along the edge of theescarpment many in defensible positions such as the classic hillfort site of ZinchecraBotanical remains from sites like Zinchecra dating to the first half of the first millenniumBC demonstrate that irrigated cultivation had begun by that early date (Daniels 198956ndash8 van der Veen 1992)

The main phase of occupation at Zinchecra ended around 500 BC (van der Veen 199212ndash13) at which point it appears that an urban site originated at Germa Over timeGarama emerged as the Garamantian capital and in the Roman period was adorned withsubstantial public buildings and temples utilizing stone on a scale and with a quality ofdressing not previously witnessed Since there is no evidence to suggest a Romanoccupation of Fezzan these must be the result of contact diplomacy and trade betweenthe Roman empire and the Garamantian kingdom Garamantian culture nowhere betterillustrated than in its extraordinary funerary architecture was extremely eclecticmdashthough the variety of tomb types in contemporary use may also reflect the maintenance ofdiscrete tribal identities within the structure of the polity (Ayoub 1967b 1968 Daniels1971265ndash8 el-Rashedy 1988 Ruprechtsberger 199751ndash65)

The evolved settlement pattern (Fig 96) reflects the increasing localization of farmingactivity in the oases along the base of the depression In addition to the large urban centreat 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 thousandsof graves have been recorded as noted above) Hundreds of foggaras facilitated the large-scale and extensive cultivation of the valley-floor oasis area A crucial question we arestill seeking an answer to is why these systems were abandoned perhaps it was becauseof falling water levels in the aquifer The settlement density the number and scale of thecemeteries and the foggara systems all combine to highlight the Garamantian period asone of peak population and oasis cultivation

Garamantian civilization was thus the result of raised population levels in the northernSahara following the development of advanced irrigation systems The concentration oftens of thousands of people in the largest of these oases allowed them to dominate a largeexpanse of the Saharamdashraiding and trading in equal measure to all points of the compassClassical sources speak of the Garamantes hunting the troglodytae and lsquoEthiopiansrsquo which gives a strong hint of slave raiding against neighbouring peoples (Herodotus4183 cf Tacitus Hist 450) Quite apart from the possibility of selling-on such captives north across the Sahara the intensive irrigated cultivation and the dangerous task offoggara construction could have absorbed large numbers of slaves The evidence for theexistence of trans-Saharan trade at this date is partial at best but the large quantities ofRoman trade goods found at Garamantian sites and in their burials indicate thatsomething of value must have been passing the other way Apart from slaves it ispossible 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 theemergence of a social hierarchy with a prominent elite order enjoying significantlygreater wealth than the majority of the population who were still buried in relativelysimple cairn graves


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 199777ndash81 for examples) built of mud brick

Figure 96 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 perhapslinked to a shift from foggara to well irrigation (Fig 97) The problem with irrigation based on wells is that water must be mechanically raised by bucket before being fed intoirrigation canals with the result that in general each well can irrigate only a limited areaof fields around it The late medieval and early modern pattern is thus of small clumps ofpalms and cultivated fields clustered around many scattered wells in contrast with theevidently more extensive areas that appear to have been cultivated whilst the foggaraswere operating


Garama was displaced as the regional capital by sites further east and south (MurzukTraghen Zuila) but its substantial walls and kasbah guaranteed it a role in the politicsand warfare of the period Nonetheless when the earliest

Figure 97 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 nineteenthcenturies they found the Wadi el-Agial a desperately impoverished region with many ofits villages underpopulated and crumbling and the bulk of its agricultural productiontaken as taxes and rents by absentee sheiks and Turkish officials (Barth 1857143ndash9 Bruce-Lockhart and Wright 1999 Denham and Clapperton 1826169ndash77)

Only in the last forty years have modern artesian wells reversed the trend of declineand revived the population and agricultural productivity However this has been at a cost


to the aquifer levels which have already fallen significantly In the long term (Fig 98) it is possible that agriculture will be forced to contract around a limited group ofagricultural settlements with very deep bore artesian wells serving clusters of individualirrigated crop

Figure 98 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 c300 m diameter This system developed elsewhere in Libya forexploiting fossil water supplies deep below the Sahara (cf Allan 1979) The FezzanProject cannot offer solutions to the problem of where water is to come from next but ithas graphically illustrated the human consequences of past changes in water availabilityin the desert Whilst we may take pride in human ingenuity in finding ways to live in thedesert we may also reflect on the environmental costs that such lsquomasteryrsquo brings in its wake Garamantian development of the foggara irrigation systems may in the long termhave been a key factor leading to the decline of their civilization as a result of over-extraction from a non-renewable groundwater source


ACKNOWLEDGEMENTS

The Fezzan Project is sponsored by the Society for Libyan Studies the Arts andHumanities 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 theLeverhulme Trust This chapter was written during the tenure of a Research Readershipaward from the British Academy The project involves the work of many individualswho are thanked as a group here but whose contributions are clearly acknowledged inthe multi-authored interim reports Thanks are also due to DMiles-Williams for Figure 92 and LFarr for Figures 93ndash97

REFERENCES

Allan JA (1979) Managing agricultural resources in Libya recent experience Libyan Studies 1017ndash28

Ayoub MS (1967a) Excavations in Germa between 1962 and 1966 Tripoli Ministry of Education

Ayoub MS (1967b) The Royal cemetery at Germa A preliminary report Libya Antiqua3ndash4213ndash19

Ayoub MS (1968) The Cemetery ofSaniat Ben Howedy Tripoli Ministry of Education Bagnall RA (1935) Libyan Sands Travel in a Dead World London Hodder amp

Stoughton Ball J and Beadnell HJL (1903) Banana Oasis Its Topography and Geology Cairo

National Printing Department Barich BE (1987) (ed) Archaeology and Environment in the Libyan Sahara The

Excavations in the Tadrart Acacus 1978ndash1983 Oxford British Archaeological Reports International Series 368

Barker GW (1989) From classification to interpretation Libyan prehistory 1969ndash1989 Libyan Studies 2031ndash43

Barker G (1996) Prehistoric settlement In GBarker DGilbertson BJones andDMattingly Farming the Desert The UNESCO Libyan Valleys ArchaeologicalSurvey Volume 1 Synthesis 83ndash109 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 2Gazetteer and Pottery Paris UNESCO London Society for Libyan Studies

Barth H (1857) Travels and Discoveries in North and Central Africa London Longman Brown amp Green (reprint Longman 1965)

Bates O (1914) The Eastern Libyans London Frank Cass (reprint 1970) Bousquet B (1996) Tell-Douch et sa Region Geographic drsquoune Limite de Milieu agrave une

Frontiegravere drsquoEmpire Cairo Institut Franccedilaise drsquoArcheacuteologie Orientale Bovill EW (1968) The Golden Trade of the Moors Oxford Oxford University Press


(second edition) Brett M and Fentress E (1996) The Berbers Oxford Blackwell Bruce-Lockhart J and Wright J (1999) Difficult and Dangerous Roads Hugh

Clappertonrsquos Travels in Sahara and Fezzan (1822ndash1825) London Sickle Moon Press Camps G (1980) Berbegraveres Aux Marges de lrsquoHistoire Toulouse Hespeacuterides Camps G (1989) Les chars sahariens Images drsquoune societeacute aristocratique Antiquiteacutes

Africaines 2511ndash40 Churcher CS and Mills AJ (1997) Reports from the Survey of Dakhleh Oasis Western

Desert of Egypt 1977ndash1997 Oxford Oxford Archaeological Monographs Cremaschi M (1998) Late quaternary geological evidence for environmental change in

south-western Fezzan (Libyan Sahara) In MCremaschi and Sdi Lernia (eds) Wadi Teshuinat Palaeoenvironment and Prehistory in South-western Fezzan (Libyan Sahara) 13ndash47 Florence Insegna del Giglio


Daniels CM (1968) Garamantian excavations Zinchecra 1965ndash1967 Libya Antiqua5113ndash94

Daniels CM (1969) The Garamantes In WHKanes (ed) Geology Archaeology and Prehistory of the Southwestern Fezzan Libya 31ndash52 Castelfranco-Veneto Petroleum Exploration Society of Libya

Daniels CM (1970) The Garamantes of Southern Libya London Oleander Daniels CM (1971) The Garamantes of Fezzan In FFGadallah (ed) Libya in History

Proceedings of a Conference Held at the Faculty of Arts University of Libya1968261ndash85 Benghazi University of Libya

Daniels CM (1975) An ancient people of the Libyan Sahara In JBynon and T Bynon(eds) Hamito-Semitica 249ndash65 The Hague Mouton

Daniels CM (1989) Excavation and fieldwork amongst the Garamantes Libyan Studies2045ndash61

Denham D and Clapperton H (1826) Narration of Travels and Discoveries in Northernand Central Africa in the Years 1822ndash1824 London John Murray (reprinted 1965 as Missions to the Niger III Haklyt Society second series CXIX edited by EW Bovill)

Duveyrier H (1864) Les Touaregs du Nord Paris Challamel Aineacute Edmunds WM and Wright EP (1979) Groundwater recharge and palaeoclimate in the

Sirte and Kufra basins Journal of Hydrology 11971ndash87 Edwards D Hawthorne J Dore J and Mattingly DJ (1999 The Garamantes of

Fezzan revisited publishing the CMDanielsrsquo archive Libyan Studies 30109ndash27 el-Rashedy F (1988) Les pratiques funeacuteraires des Garamantes et leurs relations avec

celles drsquoautres peuples drsquoAfrique du Nord In Libya Antiqua Histoire Geacuteneacuterale de lrsquoAfrique Etudes et Documents III 85ndash114 Paris UNESCO

Encyclopeacutedie (1997) Encyclopeacutedie Berbegravere fasc xviii sv Fezzan 2777ndash817 Aix-en-Provence 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 deicommerci e della cultura dei Garamanti (IndashIII secd C) In Productions et


Exportations Africaines Actualiteacutes Archeacuteologiques 405ndash20 Paris VI Colloque International sur lrsquoHistoire et lrsquoArcheacuteologie de lrsquoAfrique du Nord

Goblot H (1979) Les Qanats une Technique drsquoAcquisition de lrsquoEau Paris and New York Mouton eacutediteur Industrie et Artisanat 9

Kanes WH (1969) (ed) Geology Archaeology and Prehistory of the Southwestern Fezzan Libya Castelfranco-Veneto Petroleum Exploration Society of Libya

Klitzsch E and Baird DW (1969) Stratigraphy and palaeohydrology of the Germa(Jarma) area southwest Libya In WHKanes (ed) Geology Archaeology and Prehistory of the Southwestern Fezzan Libya 67ndash80 Castelfranco-Veneto Petroleum Exploration Society of Libya

Le Quellec J-L (1987) LrsquoArt Rupestre du Fezzan Septentrional (Libye) Widyan Zredaet Tarut (Wadi esh-Shati) Oxford British Archaeological Reports International Series 365

Liverani M (1999) Ultime scoperte nella terra dei Garamanti Archeo Attualitagrave del Passato 15830ndash9

Locirc Capitaine (1953) Les foggaras du Tidikelt Travaux de lrsquoInstitut de Recherches Sahariennes 10139ndash79

Locirc Capitaine (1954) Les foggaras du Tidikelt Travaux de lrsquoInstitut de Recherches Sahariennes 11 49ndash77

Lutz R and Lutz G (1995) The Secret of the Sahara Innsbruck Golf Verlag Mattingly DJ (1995) Tripolitania London Batsford Mattingly DJ al-Mashai M Balcombe P Chapman S Coddington H Davison J

Kenyon D Wilson AI and Witcher R (1997) The Fezzan Project 1997methodologies and results of the first season Libyan Studies 2811ndash25

Mattingly DJ al-Mashai M Balcombe P Chapman S Coddington H Davison J Kenyon D Wilson AI and Witcher R (1998a) The Fezzan Project I researchgoals methodologies and results of the 1997 season Libya Antiqua ns 3 175ndash99

Mattingly DJ al-Mashai M Aburgheba H Balcombe P Eastaugh E Gillings M Leone A McLaren S Owen P Pelling R Reynolds T Stirling L Thomas DWatson D Wilson AI and White K (1998b) The Fezzan Project 1998 preliminaryreport on the second season of work Libyan Studies 29115ndash44

Mattingly DJ al-Mashai M Aburgheba H Balcombe P Eastaugh E Gillings M Leone A McLaren S Owen P Pelling R Reynolds T Stirling L Thomas DWatson D Wilson AI and White K (1999a) The Fezzan Project II preliminaryreport on the 1998 season Libya Antiqua ns 463ndash93

Mattingly DJ al-Mashai M Balcombe P Drake N Knight S McLaren S Pelling R Reynolds T Thomas D Wilson A and White K (1999b) The FezzanProject 1999 preliminary report on the third season of work Libyan Studies 30129ndash45

Mori F (1969) Prehistoric cultures in Tadrart Acacus Libyan Sahara In WHKanes(ed) Geology Archaeology and Prehistory of the Southwestern Fezzan Libya 21ndash30 Castelfranco-Veneto Petroleum Exploration Society of Libya

Mori F (1988) Lrsquoart rupestre preacutehistorique dans le Sahara libyen comme aboutissementdrsquoun long processsus bioculturel In Libya Antiqua Histoire Geacuteneacuterale de lrsquoAfrique Etudes et Documents III 157ndash63 Paris UNESCO


Muzzolini A (1991) Proposals for updating the rock-drawing sequence of the Acacus (Libya) Libyan Studies 227ndash30

Nesson C Rouvillois-Brigol M and Vallet J (1973) Oasis du Sahara Algeacuterien Paris Institut Geacuteographique National

Nicolaisen J and Nicolaisen I (1997) The Pastoral Touareg Ecology Culture andSociety London Thames amp Hudson two volumes

Pace P Sergi S and Caputo G (1951) Scavi Sahariani Monumenti Antichi 41 150ndash549

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 JBower and D Lubell

(eds) Prehistoric Cultures and Environments in the Late Quaternary of Africa 19ndash42 Oxford British Archaeological Reports International Series 405

Petit-Maire N Delibrias G and Gaven C (1980) Pleistocene lakes in the Shati area Fezzan (27deg30primeN) In MSarnthein ESeibold and PRognon (eds) Sahara and the Surrounding Seas 289ndash93 Rotterdam Balkema Palaeoecology of Africa 12

RSGI (1937)=Real Societagrave Geographica Italiana Il Sahara Italiano Fezzan e Oasi diGat Rome Societagrave Italiana Arti Grafiche

Ruprechtsberger EM (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 29deg Parallelo Florence Sansoni Shaw BD (1976) Climate environment and prehistory in the Sahara World

Archaeology 82133ndash49 Trousset P (1986) Les oasis preacutesahariennes dans lrsquoantiquiteacute partage de lrsquoeau et division

du temps Antiquiteacutes Africaines 22161ndash91 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 237ndash39 van der Veen M (1995) Ancient agriculture in Libya a review of the evidence Acta

Palaeobotanica 35 (1)85ndash98 van der Veen M (1999) (ed) The Exploitation of Plant Resources in Ancient Africa

New York Plenum Zaluski M and Sadek KE (1980) Hydrogeology of mesozoic aquifers in the western

part of Wadi al Ajal Symposium on the Geology of Libya 2635ndash42 Ziegert H (1995) Das neue Bild des Umenschen UniHH Forschung Beitrage aus der

Universitaumlt Hamburg 309ndash15


10 Farming and famine subsistence strategies in

Highland Ethiopia ANN BUTLER AND A CATHERINE DrsquoANDREA

INTRODUCTION

The Highlands of Ethiopia have an environment that is governed by the high altitude andwithin relatively low longitudes they have a temperate climate This supports a widerange of crops which include both indigenous African domesticates and the cool-season grain crops developed in and introduced in antiquity from southwest Asia Cultivation israinfed 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 tounderstand some of the rationale that underlies these farming practices The results of anew ethnoarchaeological study in the Ethiopian Highlands are presented here integratedwith some published accounts of traditional agriculture in the region

FIELDWORK

Ethnobotanical studies were carried out between 1996 and 1998 in the EthiopianHighlands (Tigrai province) about 2000 m above sea level in the mid-altitude region This is the agro-climatic zone known as dry woina dega (Bekele-Tesemma 19936) The average temperature range is between 5 and 40degC (Gebremedin and Haile 1997) Fieldwork was concentrated on the northern edge of the Giba plateau in the Endertaadministrative 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 101) Supplementary records were made both at further kushets within the same tabia and also at others within the woredaof 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 mainharvest time between November and December (Butler in press Butler et al 1999 DrsquoAndrea et al 1997 1999)

Figure 101 Map of Ethiopia showing Adi Ainawalid in Tigrai province

During the political upheavals in Ethiopia between 1973 and the early 1990s there waslarge-scale compulsory resettlement Individually owned and managed farmland wastaken and incorporated into large co-operatives By 1996 these had been dispersed land holdings had only recently been reallocated to individual farmers and traditional farmingpractices had been resumed However several families at Adi Ainawalid were spared theturmoil throughout the conflict they occupied their original family homes and they nowretain 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 graincrops supplemented by a few resources such as salt and oil bought from the regionalmarket 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 199348 Konde 1993 18) Individual holdings are very small rangingfrom one to eight tsumdi commonly consisting of at least two plots usually one adjacent

Farming and famine subsistence strategies in Highland Ethiopia 175

to the dwelling and the other(s) up to an hourrsquos walk away towards the perimeter of thevillage This is similar to the situation described for other areas in Ethiopia (Tsegaye1997) 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 householdlacking the means to work it or to find an alternative to farming as a means of livelihoodPlot boundaries may be defined by stones shallow drainage channels or spinybrushwood

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 foranimals within the surrounding stone-walled compound (DrsquoAndrea et al 1997 Fig 102) Livestock is also kept inside houses especially donkeys horses and calves toprotect them from predation by hyenas Gardens are common in larger maturecompounds but are rarely found with small houses The soils are largely derived fromlimestones weathered to vertisols and cambisols which are clays and sandy clays and inthis region they are typically stony thin and eroded (Hunting Technical Services Ltd1973ndash4 Mitiku Haile pers comm) The natural vegetation in the region is described as Acacia savannah (Bekele-Tesemma 19936) 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 102 Residential compound near fields west end of Adi Ainawalid facing southwest November 1997

Photograph CDrsquoAndrea

water and promotes run-off (Butzer 1981) Attempts are made to catch and retainrainwater 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


terracing low walls are constructed on the surrounding slopes using stones dug out fromlocal 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 ofgenera such as Acacia Mill and Erythrina L In the settlement area small stands ofEucalyptus species have been introduced mainly for shade and fuel There are occasionalcompounds with small mainly leguminous trees planted for shade and supplementarylivestock feed Until recently water for household use has had to be carried from theriver up to two hoursrsquo walk away often twice daily but at Adi Ainawalid three wells arenow available for use Irrigation is rare and confined to plots near the river which arerented 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 McCann199548ndash56) Their availability is essential to cultivation although donkeys and morerarely camels and mules also supply labour A 2-year-old ox takes a yearrsquos training and can give up to five yearsrsquo 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 findingto other surveys in the region (eg FAO 1986) Animals are commonly loaned to makeup a ploughing team and for threshing when up to eight or more oxen may be used totrample the yield from a single harvest Also a man with a team will plough land forothers for a payment of half the harvested crop Occasionally donkeys mules or mixedteams 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 fitwhen their labour is most in demand (Konde 199370) Cattle small ruminants equidsand the few camels graze field edges and stubble Feed crops are not cultivated nor island 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 periodsand to particular animals mainly oxen and cattle By-products of cultivation such as weeds and crop-processing residues and food-preparation residues are the mostimportant feed resources the choicest of which are fed to the oxen Tree and shrubvegetation 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 ashouse 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 Africandomesticates members of the assemblage of southwest Asian founder crops and someintroductions from the New World (Table 101) Farmers possess much detailed local knowledge of the habit and environmental requirements of the different varieties of each


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 oftraditional 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 aparticular field under the environmental conditions anticipated during the followingseason A range of traits in each taxon is desirable such as both early and late maturingvarieties (Gebremedin and Haile 1997) Also crop types are selected for variables suchas the colour of resulting food products and the baking quality and storage quality ofgrains are more important than the grain yield and the size (Haile 1995 Webb and vonBraun 1994) To extend the range of crop types available to individual farmers locallydeveloped populations of grains are exchanged and further varieties or species may bepurchased from the regional market These measures help to perpetuate recognized land-races (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 103) In other areas mixtures such as pea and faba bean (ater-abie) or sorghum with chickpea are sownWhen possible the varieties of the crops are chosen for their synchroneity ofdevelopment 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 singleresource and prepared for food as one or the constituent grain types may be separated inthe home When a single species is planted several varieties may be mixed Theproportion of different grains in the mixture at harvest differs from that sown so themixtures have to be reassembled each sowing season Following periods of drought cropfailures can result in a reduction in the number of crop species and varieties harvested aswell as the total yield

In the few house gardens chilli (berbere) garlic (tarsquoeda 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 carbohydratefoods 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 (sarsquori awald)

Table 101 Crops cultivated at Adi Ainawalid Tigrai Ethiopia Crop species Common name and varieties Local name Sorghum bicolor (L) Moench sorghum (5 varieties) mashella Eragrostis tef (Zucc) Trotter teff (red white) taffTriticum turgidum conv durum (Desf) MacKey

durum (black with hexaploid characters)

tselimoi

Triticum aestivum subsp vulgare (Vill) MacKey

bread wheat (shahan Canada wheat)

sindai


Figure 103 Intercropped bread and durum wheats near Mai Kayeh Tigrai November 1997


Hrufa (Nees) Stapf and Eleusine floccifera (Forssk) Spreng (rigaha) are gathered to weave into baskets and the labiate Otostegia integnfolia Benth (chirsquoindogwee) has insecticidal properties the juices being smeared onto livestock to prevent damage to theirhides

Cultivation systems

In Ethiopia as a whole the environmental factor that has the greatest control over thefarming schedule is said to be rainfall (McCann 199528ndash31) 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

Hordeum vulgare L barley (burguda sarsquosaa) segemTriticum L spHordeum L sp wheatbarley intercrop mixture hanfetse Eleusine coracana (L) Gaertn finger millet (black red

white) dagousha

Lens culinaris sspculinaris Medikus lentil bersheem Lathy rus sativus L grasspea gwayya

seberoCicer arietinum L chickpea shimbra Zea mays L maize (arigo beraho) efoonTrigonella foenum-graecum L fenugreek abakaLinum usitatissimum L linseed indata


sometimes with heavy hailstorms damaging to crops Cyclical episodes of low rainfalloccur associated with the El Nintildeo Southerly Oscillation (ENSO) with frequenciesvarying 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 hasbeen decreasing Although a belg season still occurs in the central Highlands in the provinces of Wello and Shewa (Rahmato 199154) and even in parts of the northernHighlands (Emergencies Unit for Ethiopia 1997) over the past thirty years the totalannual rainfall in much of the northern province of Tigrai has been drastically reducedand the spring rains have virtually ceased (Emergencies Unit for Ethiopia 1994) Thissituation has largely restricted the cropping periods to a single season (meher) associated with the kremt (Adebo 199375 Konde 199379) Today despite an annual rainfall of between 450 and 900 mm (Butzer 1981) Tigrai is known as one of the most drought-prone 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 withploughing The seed bed for cereals usually receives several ploughings but the plots forpulses may be ploughed only once Stones and tree stumps are retained in the soil toreduce erosion from wind and water Grain crops are usually broadcast sown betweenMay and Julymdasha 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 evenfor this crop category it appears minimal Cereal fields may be manured with dung fromgrazing ruminants although government supplies of chemical fertilizers are sometimesavailable their application is usually precluded by their expense Pulses are very rarelyfertilized Weeding also is rare but is more common for cereals Cereal plots may beploughed again to aerate the soil and facilitate drainage and reduce the weeds The mainharvesting season falls between October and December Crop plants are commonlyuprooted individually by hand (Fig 104) or they may be either uprooted or cut bysickle Weeds may also be uprooted and harvested separately for feed Unpalatable orspiny weeds remain standing in the fields Cultivation of the different crops is commonlystaggered across the growing season to increase the breadth of the harvest season therebypreventing an excessive concentration of labour and resources at one period Fingermillet sorghum maize and lentil are usually sown earlier than wheats barleys andgrasspea chickpea is often the latest crop (DrsquoAndrea et al 1997)

Grain separation takes place on threshing floors of compacted soil constructed at theedges of fields or within the settlement area as described in detail by DrsquoAndrea and others (1997 1999) The harvested crop is carried to the edge of the floor where it ispiled to dry or threshed immediately A threshing team of up to eight oxen is drivenaround the floor over the crop and the crop fractions are sorted with forks and brushes(Fig 105) 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 arenormally amalgamated for feed (Fig 106) 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


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 areminimized by child scarers or scare-crows stationed in the fields Crop rotation is used tobreak cycles of plant parasites such as Striga asiatica (L) Kunze (selemi) on cereals particularly sorghum and Orobanche minor Sm (mrsquoandat tali) on legumes Barley and

Figure 104 Harvesting grasspea by hand uprooting Adi Ainawalid November 1996

Photograph AButler

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 especiallychickpeas and flax are planted to reinvigorate plots depleted of nutrients Because of theshortage of land fallowing is unusual A fallowed plot will often signal a shortage ofoxen or human labour (Adebo 199383 Konde 199380)


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 willbe hired Ploughing is undertaken by men and older boys In cases where no male familymember is available to work the land it may be rented out in return for half the yield atharvest It was reported that although not unknown it was very rare for a woman toplough The whole family may be engaged in weeding and harvesting Men

Figure 105 First threshing of teff Adi Ainawalid November 1997



Figure 106 Winnowing teff Adi Ainawalid November 1996 Photograph AButler

and children perform the threshing and winnowing stages of crop processing Women areconcerned with small-scale winnowing and fine grain cleaning within the household andthe preparation of food Importantly they also play a significant role in discussions on theannual farming schedule and on crop and seed selection

Grain storage

Storage is overseen by the women (Tsegaye 1997) Cleaned grain is stored inside thehouses in clay or bamboo vessels about 1 m tall and sealed with dung (Fig 107) Fumigants or insecticides are not added but it is believed that the dung acts as an insectrepellant Small crop yields may be kept in skin bags or sacks Unthreshed crops arestacked in the house compounds as are the threshing residues for animal feed Cats arekept to deter rodents


Figure 107 Grain storage jars Adi Ainawalid November 1996 Photograph AButler

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 ofgrain through plunder Sorghum is known to have survived such emergency storage for atleast five years (McCann 199567ndash8 Rahmato 199131 Worede and Mekbib 1993)However owing to its secret nature this storage system was not investigated during thisstudy

Farming and fuel

In the Ethiopian Highlands the paucity of trees is believed to be long-standing and is increasing (Staringhl 1993) Historical descriptions give varying accounts of the vegetation


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 thatthere was about 40 per cent tree cover in the country as a whole but by the late 1980s thishad fallen to 56 per cent (FAO 1986) In Tigrai the demands for fuelwood and fortimber for manufacturing appear to have reduced the number of mature trees mainly tothe wooded conservation areas around churches or to single specimens used ascommunity assembly points This loss has been accelerating due partly to continuinghousehold demands and also to the need to supplement income outside farming the saleof firewood has been a traditional supplementary source of income until the recent pastNowadays 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 supportsand farming implements such as plough beams yokes and winnowing forks Up to 55 percent of the fuel resources are provided by alternatives to timber (World Bank 1984) andare mainly farming by-products Dung is perhaps the most valued It accumulates inresidential compounds where livestock are penned overnight and is collected by childrenfrom the grazing areas it is then spread on walls to dry and be stacked The culms ofsorghum 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 usuallypermanent fixtures of clay and stone constructed inside the kitchen building but smallportable stoves are also used (DrsquoAndrea et al 1999)

CROP DIVERSITY

Many Ethiopian crops are noted for an impressive diversity of form under environmentalconditions that were described by Vavilov (1935347) as relatively uniform within thehigh altitude This morphological diversity incorporates traits adapted to various stressconditions and it seems to be maintained by both environmental and human agenciesRecent studies of Ethiopian wheats have shown that crop varieties with deeply pigmentedblack 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 Hillmanpers 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 fallingin 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) andpea (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 becauseof their performance under potentially stressful environmental conditions they continue tobe selected by farmers (Belay et al 1995)

However while a general diversity of crop varieties appears to be continuallymaintained by farmersrsquo selection the cultivation of some species of food plants is


becoming restricted due to the changing pattern of climate and land shortages This isaffecting 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 102 lists the crops that are no longer found at Adi Ainawalid or indeed in the tabia of Mahabere Genet as a whole However they arefamiliar to many farmers of Adi Ainawalid and they have been grown locally in the pastThese crops are still grown in areas of higher rainfall particularly further south in Tigraiand occasionally they can be found in the regional market in Mekelle As opposed to theblack wheat (tselimoi) mentioned above lsquoclassicrsquo durum wheat recently still said to beone of the major cereals in Ethiopia (Engels and Hawkes 1991) is an uncommon crop inthe study area one farmer in the kushet of Adi Akel immediately adjacent to AdiAinawalid had twice obtained some durum grain but this had produced only sterileplants All the rare grains are valued as traditional resources with special propertiesdurum 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 andfaba beans although expensive are still regularly bought in for meals on

Holy Days throughout the year The Ethiopian pea is particularly sought after and theaddition 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 beenough to support dryland agriculture (Butzer 1981) At the same time however thisrainfall 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 springand summer monsoon seasons have diminished but the total annual rainfall may beaugmented to near-normal levels by unpredictable showers often of heavy hailstorms destructive to crops The opportunistic use of watered soil is a feature of Ethiopianhighland agriculture

The study periods between 1996 and 1997 fell within an episode of low precipitation

Table 102 Crops no longer cultivated at Adi Ainawalid Tigrai Ethiopia Crop species Common name Local

name Triticum turgidum spp diccocum (Schrank) Thell

Emmer wheat ares

Tturgidum conv durum (Desf) MacKey Durum wheat kinkinai Pisum sativum ssp abyssinicum ABr Ethiopian pea dekoko Psativum ssp sativum var arvense L Field pea aterVicia faba L Faba bean abiePisum sativum LVicia faba L Peafaba bean

intercrop aterabie


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 spatiallydiffering amounts The reservoir at Adi Ainwalid was dry in both years although the oneshared 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 farmersof Adi Ainawalid although they had cultivated a restricted range of the most drought-tolerant crop species experienced severe crop failure with very reduced harvestsFollowing a very dry summer there were outbreaks of heavy rain in November thatcaused 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 extraend-of-season crops of barley and chickpeas which would have been ready forharvesting at the end of the following spring In other highland regions when the summerrains 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) yetin Tigrai a chronic food shortage prevails Between 1988 and 1992 three-quarters of the families produced insufficient food (Holt and Lawrence 199326ndash31) and in 1997 the regional food production was deficient by about 20 per cent (Gebremedin and Haile1997) This is in contrast with previous times documentation from the sixteenth centuryfor example describes southern Tigrai as a land of great abundance of production withgreat yields of cereals and pulses (Alvares 152031) The current shortfall is thought tobe 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 (Pankhurst1992318 Zewde 1991195ndash6)

On the local level in 1996 crop production throughout Tigrai was officially reportedby the aid agencies as being poor (Ahrens and Spiess 1997) and Enderta was singled outas being in particular need of food aid (Emergencies Unit for Ethiopia 1997) At AdiAinawalid some supplies of grain had been distributed although towards the end of 1997food aid had been discontinued These food shortages prompted surveys in the Highlandsthat have highlighted a number of farming problems with priorities that vary slightly atdifferent seasons and in different areas lack of rain small land holdings insufficientoxen lack of human labour shortage of fuel high market prices shortage of livestockfeed livestock disease and unclean drinking water (Adebo 199370ndash89 Konde 199376ndash86) 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 associatedwith both drought and famine It is now believed that drought alone seldom causesfamine more often a combination of factors is involved These include epidemics ofhuman and veterinary diseases such as smallpox and cattle rinderpest plagues on cropsof insect predators such as locusts ants and army worms and social and political


circumstances (McCann 199589ndash91) The recent situations of conflict appear to underliemany 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(eg Westphal 1975 Simoons 1960) was abruptly and severely altered followingresettlement and lsquovillagizationrsquo until the early 1990s (Rahmato 1985 Rock 1994)Families were split up relocated in regions of unfamilar ecology and expected tocultivate alien (to them) crop assemblages on collective farms Now reinstated traditionalfarming is demonstrating in the selected crops and technology an essential stability thathas 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 Cropgermplasm is carefully conserved to allow the best selection of crop types to suit theagricultural situation Within a wide repertoire agrarian systems are flexible andencorporate strategies to maximize production under whatever conditions pertainMechanisms have been developed to minimize erosion water is reserved at run-off soil-water is exploited grazing is controlled and alternative sources of fuel have been foundIt appears that periods of low rainfall and crop failure can be endured for several years bythe 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 ofshortage These strategies appear to be able to promote survival for at least two to threeyears of high aridity

When droughts last longer than a few years or when epidemics of disease or pests orbouts of conflict or political unrest are superimposed upon drought periods then itappears that the social and farming mechanisms described above may be insufficient toavert severe food shortage A better understanding of climatic perturbations could speedthe 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 ofCanada (Grant No 410ndash96ndash1520) and supported with the valuable assistance of Dr Mitiku Haile Dean of Mekelle University College (MUC) We are grateful to theCommittee for Research and Conservation of Cultural Heritage (CRCCH) Addis Ababaand the Tigrai Bureau of Culture Tourism and Information Mekelle for grantingpermission to undertake the study Field assistance was provided by Shewiaye BelayZelealem Tesfay Derege Asefa and Alemtsehay Tsegay of MUC Figure 101 was drawn by Shannon Wood of Simon Fraser University (SFU) Useful comments on the text byDiane Lyons (SFU) are acknowledged Our deepest thanks go to the kind generous andpatient farmers of Adi Ainawalid and neighbouring settlements in south-central Tigrai


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 515ndash6196Situation Report Emergencies Unit for Ethiopia (UNDP-EUE) (httpwwwsasupenneduAfrican_StudiesEUEamhra696htm)

Alvares F (1520) The Prester John of the Indies Volume I Cambridge Cambridge University Press translation by CFBuckingham and GWBHuntingford (eds) 1961edition Hakluyut Society vols 114ndash15

Annicchiarico P Pecetti L and Damania AB (1995) Relationships betweenphenotype variation and climatic factors at collecting sites in durum wheat landracesHereditas 122163ndash7

Bauer DF (1975) For want of an oxhellip land capital and social stratification in Tigre InHGMarcus (ed) Proceedings of the First United States Conference on Ethiopian Studies 235ndash48 East Lansing MI Michigan State University African Studies Center

Bejiga G Tsegaye S Tullu A and Erskine W (1996) Quantitative evaluation ofEthiopian landraces of lentil (Lens culinaris) Genetic Resources and Crop Evolution43 293ndash301

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 selectionfor purple-grain tetraploid wheats in the Ethiopian highlands Genetic Resources and Crop Evolution 42387ndash91

Butler EA (in press) Sustainable agriculture in a harsh environment an Ethiopianperspective In FHassan (ed) Drought Food and Culture Food Security in Africarsquos Later Prehistory New York Plenum Publishing Corporation

Butler EA Tesfay Z DrsquoAndrea AC and Lyons DE (1999) The ethnobotany of Lathyrus sativus L in Highland Ethiopia In M van der Veen (ed) The Exploitation of Plant Resources in Ancient Africa 123ndash36 New York Plenum Publishing Corporation

Butzer KW (1981) Rise and fall of Axum a geoarchaeological approach American Antiquity 46471ndash95

DrsquoAndrea AC Haile M Butler EA and Lyons DE (1997) Ethnoarchaeologicalresearch in the Ethiopian highlands Nyame Akuma 4719ndash26

DrsquoAndrea AC Lyons D Haile M and Butler A (1999) Ethnoarchaeologicalapproaches to the study of prehistoric agriculture in the Ethiopian highlands In M vander Veen (ed) The Exploitation of Plant Resources in Ancient Africa 101ndash22 New York Plenum Publishing Corporation

Demissie A and Bjornstad A (1996) Phenotypic diversity of Ethiopian barleys inrelation to geographical regions altitudinal range and agro-ecological zones as an aid


to germplasm collection and comservation strategy Hereditas 12417ndash29 Emergencies Unit for Ethiopia (1994) Situation Report for Region 1 (Tigray)

Emergencies Unit for Ethiopia (UN-EUE) (httpwwwsasupenneduAfrican_StudiesEUEtigray0494html)

Emergencies Unit for Ethiopia (1997) Field Trip to Amhara and Tigray National Regional States Emergencies Unit for Ethiopia (UN-EUE) Development Programme (httpwwwsasupenneduAfrican_StudiesEUEnorth0296html)

Engels JMM and Hawkes JG (1991) The Ethiopian gene centre and its geneticdiversity In JMMEngels JGHawkes and MWorede (eds) Plant Genetic Resources of Ethiopia 23ndash41 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) (httpwwwu-fondetnoengelsktemakonf1-3html)

Govorov LI (1930) The peas of Abyssinia A contribution to the problem of the originof cultivated peas Essay II Bulletin of Applied Botany Genetics and Plant Breeding (Leningrad) 24399ndash431

Haile M (1995) Indigenous knowledge and agricultural practices in Central TigrayUnpublished 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 (1973ndash4) Tigray Rural Development Studies Map of Landforms in Mekelle District Gradients Soil Depth and Soil Types (1ndash6) 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 JC (1995) People of the Plow Madison Wisconsin University of Wisconsin Press

Pankhurst RA (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 MJ (1994) Famine and Food Insecurity in Ethiopia A Critical Assessment of theNotion of ldquoCoping Strategiesrdquo University of Leeds unpublished PhD thesis

Salt H (1814) A Voyage to Abyssinia and Travels in the Interior of that CountryLondon FC and JRivington

Simoons FJ (1960) Northwest Ethiopia Madison University of Wisconsin Press Spiess H (1994) Report on Drought Animals under Drought Conditions Emergencies

Unit for Ethiopia httpwwwsasupenneduAfrican_StudiesEUEdrought 0794html Staringhl M (1993) Foreward In ABekele-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


JMMEngels JGHawkes and MWorede (eds) Plant Genetic Resources of Ethiopia288ndash95 Cambridge Cambridge University Press

Tsegaye B (1997) The significance of biodiversity for sustaining agricultural productionand role of women in the traditional sector Agriculture Ecosystems and Environment62215ndash27

Vavilov NI (1935) The phyto-geographical basis of plant breeding Theoretical Bases for Plant Breeding Moscow 1 Reprinted and translated in DLoumlve (1992) Origin and Geography of Cultivated Plants 316ndash66 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 Nintildeo and drought early warning in Ethiopia Internet Journal for African Studies 2

Worede M and Hailu Mekbib H (1993) Linking genetic resource conservation tofarmers in Ethiopia In Wde Boef Kojo Amanor and KWellard (eds) Cultivating Knowledge 78ndash84 London Intermediate Technology Publications

World Bank (1984) Ethiopia Recent Economic Developments and Future ProspectsWashington DC World Bank

Zewde B (1991) A History of Modem Ethiopia 1855ndash1974 London James Curry


Part IV EASTERN AND SOUTHERN

AFRICA

11 Engaruka farming by irrigation in Maasailand

cAD 1400ndash1700 JOHN EGSUTTON

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 wellas in their exploitation in recent centuries by hunters herders and cultivators Thecontrasts are especially sudden and striking at Engaruka situated at the foot of the east-facing Rift wall at three degrees south (Fig 111) At an altitude of 1000 m (which is low for this interior region) and with unreliable and variable rainfall estimated at not morethan 400 mm in an average year it is a relatively hot dry and dusty place with highevapotranspiration Despite the attraction of a permanent supply of clear water in theEngaruka river no cultivators would ever have contemplated settling here by relying onthe rain alone for their crops

Immediately behind Engaruka the escarpment rises to 2000 m above sea level and towering above that are the Crater Highlands with the wide dome of Lolmalasin reachingto some 3500 m These highlands catch two or three times the rainfall of the Rift floortheir vegetation ranges from montane forests to open grasslands The latter have in recenttimes supported wild herbivores and pastoral communities with cattle sheep and goatsSince the seventeenth or eighteenth centuries AD these pastoralists have been MaasaiHowever the history of pastoralism here stretches back some 3000 years during whichtime successive groups of which the Maasai are the most recent have replaced orassimilated those who preceded them (Sutton 1993) Although these cool highlands havenot attracted agricultural settlement the run-off which descends the escarpment indeeply 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 112 and 113) Nowadays only one of them carries waterpermanently and this is the Engaruka river itself (no 2 on Figure 112) This is a fast but shallow stream usually 3ndash4 m wide

Figure 111 The Rift Valley and Crater Highlands of northern Tanzania showing Engaruka and related sites


Figure 112 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)

Engaruka farming by irrigation in Maasailand cAD 1400ndash1700 197

Figure 113 The Engaruka escarpment from the east with the gorge of Engaruka river (central)

Photograph JEGSutton

as it descends the rocky scree it is easily fordable except during spates following stormson and behind the escarpment With its speed compensating for its small dimensions itsdischarge into the plain is considerable The other streams (nos 1 4 and 5) flowseasonally or in the case of certain escarpment gullies and clefts (notably no 3) openvery occasionally after exceptional rain Of the seasonal streams the Makuyuni (no 4) onthe northern side is the most reliable in good years flowing for six to nine months andvery occasionally lasting throughout That at the south end Olemelepo (no 1) may carrynearly as much water overall but is extremely temperamental liable to open in spate andthen 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 outwashfans These consist of soil mixed with water-worn lava boulders of all sizes which haveaccumulated immediately below the points where the gorges of the seasonal as well asthe main river open onto the escarpment foot As a result the streams enter the plain at acommanding level on the crests of these fans the land falling away not only towards theRift floor but also on either side of the stream beds This situation continues in the caseof 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 andadvantageous level has been exploited for irrigated agriculture but at two separateperiods The second of these persists the present community of Engaruka continues toexpand and enlarge its cultivation area This dates from the 1890s when a few farmersfrom different parts of what was then German East Africa became established a short


distance down the river (that is largely below the area of the archaeological field remainsof the earlier period) and began cultivating the soft soil with the help of furrows takenoff both banks This community has been reinvigorated by new settlers on occasions inthe 1920s and 1940s during the British mandateship of Tanganyika and again in the1970s within the Tanzanian lsquovillagizationrsquo (ujamaa) movement The latter involved the incorporation of numbers of Maasai who previously herded in the surrounding plain sothat the character of what used to be called the lsquoSwahilirsquo village of Engaruka has altered In a series of good years some of these farmers cultivate in the Makuyuni basin (on thenorth side) too by relying on a combination of rainfall and water furrowed from thatstream (no 4) while its flow lasts Others farm in the Olemelepo basin (to the south) byusing a long cross-valley furrow (following close to the line of an ancient one) taken offthe Engaruka gorge

Before the 1890s however Engaruka was according to available reports desertedexcept for some pastoral Maasai whose cattle and goats grazed and browsed the sparsepasture of the Rift floor within reach of the river that being the only permanent water inthe district Information about previous inhabitants gleaned from local Maasai early in thetwentieth century is vague and is probably not genuine tradition so much as guessesoffered in response to direct questions about the lsquoruinsrsquo (for discussion see Fosbrooke 1938 Sassoon 196680ndash81 Sutton 197867ndash68) This negative reaction indicates thatthe place was deserted before the nineteenth century at latest The recent and existingcultivating community does not appear to be descended in any way from the earlierirrigation farmers who lived here between approximately the fifteenth and theseventeenth centuries The evidence for that settlement and its fields is exclusivelyarchaeological

THE ANCIENT FIELDS AND IRRIGATION SYSTEM

These earlier fields and irrigation worksmdashwhich cover some 2000 ha at the base of theescarpment around the foothills and into the plain (Fig 112)mdashare distinguished from the modern ones by their use of stone for dividing and levelling the plots for irrigation(Figs 114 and 115) by means of revetments and mild terracing and for lining andembanking the artery canals (Figs 116 and 117) and feeder furrows Equally distinctiveamong these ancient fields are two other types of stone features The first of theseconsists of numerous square or angular cairns standing up to 2 m high with rubble coresretained by drystone casings of larger boulders (Fig 118) They are interpreted as stone clearance devices necessitated by the thinness of soil and abundance of surface stonethese conditions being doubtless exacerbated by intensive cultivation with irrigation overa considerable period Secondly there are round enclosures up to 10 m across consistingof thick stone walls


Figure 114 Engaruka south fields stone field divisions and feeder furrow


Figure 115 Grid of feeder furrows and levelled field plots below the intermediate north gorge (no 3 on Figure 112)



Figure 116 The support for the lsquogreat northernrsquo canal running along the escarpment foot



Figure 117 The embanked causeway of the lsquogreat northernrsquo canal


of similar construction faced both outside and in with a narrow entrance gap Mostprobably these were built not to contain houses but for cattle The latter would have beenvalued for providing manure for the fields as well as milk and meat and would haveneeded to be stall-fed owing to the intense cultivation all around and the lack of pasturein 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


(Thorp 1986)mdashthese villages being situated on the escarpment scree above the levelattainable by channelled watermdashconsidered alongside examples of certain recent and existing compact and integrated agricultural

Figure 118 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)


systems in Africa (Sutton 1986 1989) Of relevance also is the archaeological exampleof the Nyanga terraced fields and connected stone-walled farmsteads with sunken stock-


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 partdeduced from ethnographic examples of communities currently cultivating in some caseswith irrigation at comparable altitudes with medium to low rainfall in this interior regionof East Africa

Of particular value in this exercise are the Sonjo villages and irrigated basins 100 kmor so to the north close to the TanzaniamdashKenya border Here sorghummdashan ancient African grain and the principal crop of the savanna regions across the continentthroughout the Iron Agemdashmaintains (despite the progressive popularity of maize in thetwentieth century) its dominant position with varieties selected and developed locallyboth 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 beenrecovered Other crops suggested by the examples of Sonjo and drier parts of the RiftValley 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 farmingcommunity Finger-millet while less tolerant than sorghum of heavy irrigation (and perhaps less productive in grain harvested per hectare) has the advantage of ripening ona low rainfall and also of long storage qualities Probably it would have been sown totake advantage of the rains in the main and if the crop were successful stored in theroofs or homestead granaries as reserve against a famine year It would also have beenvalued for beer as would surplus sorghum too doubtless improved by honey obtained inthe forests above Engaruka

Much but not all of Engarukarsquos ancient field area lies closer to the escarpment thanthat now settled and cultivated so that it survives most unusually as an expanse offossilized fields and irrigation devices Despite the effects of subsequent erosion inplaces with gullies damaging and destroying features on cutting through the loose andstony soil the upper part of the field system is preserved in a remarkably pristine stateVisibility depends on the season and the amount of grass and greenery on the trees andbushes Moreover there has been a noticeable increase of thornbush over the last fortyyears so that certain photographsmdashnotably Figure 115 taken in 1971mdashcannot be repeated This vegetational change apparently relates to a reduction of grazing by variouswild herbivores and also by Maasai cattle as well as to a cessation of burning as theagricultural population of the new Engaruka villages has increased

The lower part of the old field area has been subject to the opposite experiencemdashthat of redeposition of soil eroded from the upper partmdashso that the stone features there tend tobe obscured But the typical field divisions can be seen in gully sides and the irrigationgrid pattern is very clear from air photographs taken in the 1960s before the recentexpanse of bush It is on this relatively level soil with less surface stones that the presentinhabitants 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 alsofor terracing and revetting the numerous homestead platforms in seven large villages that are situated immediately above the top canals (Fig 112) The sheer density of these


remains over such a wide extent combined with complete abandonment of this system atleast two more probably three centuries ago makes old Engaruka unique as anarchaeological field system and one that can be mapped and studied on the ground (Formore detailed description see Sutton 1998 for discussion of particular features seeSutton 1978 and 1986 the villages and excavations undertaken in them are furtherdescribed by Sassoon 1966 and 1967 and by Robertshaw 1986) This intense use ofstone which in older ethno-historical literature of eastern Africa (such as Murdock 1959) was labelled lsquomegalithicrsquo 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 useThe obvious solution was to utilize these stones in the field divisions and terraces and inthe furrow and canal sides with any remaining excess being piled in the enclosure wallsand especially the cairns which were built as neatly and vertically as possible in order tominimize the waste of cultivable ground As pressure on resources of both soil and waterincreased in this isolated situationmdashone circumscribed by the limits to which irrigated water could be led by gravitymdashit appears that this commitment to stonework (which is explicable in the first place in functional and environmental terms) developed into acultural attachment if not a hallmark of the old Engaruka community

However before becoming unduly enthralled by the stone lsquoruinsrsquo of Engaruka and the accidents of survivalmdashlsquothe tyranny of the monumentsrsquo in Ian Farringtonrsquos phrasemdashit is encumbent to consider Engaruka in its regional ethnographic and historical contextThere is in fact nothing very unusual about irrigated agriculture with lsquoindigenous rootsrsquo in the precolonial past along the western wall of the Rift in northern Tanzania and Kenyathere 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 TaitaKilimanjaro and Mount Meru (For a survey of these remains see Sutton 1973 19841989 Widgren and Sutton 1999 and see also Widgren this volume Chapter 14) Among these communities numerous varieties of field systems are found prepared fordifferent crops and combinations and depending in greater or lesser measure on artificialirrigation of the lsquohill furrowrsquo sort (Adams 1989) that is by constructing small gravity-fed canals off springs or mountain streams But since these present and recent fields areobliged 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 notrecognizable as such on the landscape Or rather where there is a strong suggestion ofcontinuity of settlement and of cultivation dependent on irrigation over a long period onemay as a historian have to be content with regarding the ancient and the existing fieldsas all one At best therefore in these favoured areas of concentrated agriculturalsettlement one rarely gains anything more than an impression of the cultivation andirrigation system that operated in the past and one cannot discern the ancient fields asphysical units

Engarukamdashfor the dual reasons of its being a deserted site and a conspicuous onebecause of its stoneworkmdashis different therefore and extremely valuable as a researchresource being (with a few minor related sites in the district) a rare example of anarchaeological field and irrigation system that can be studied directly on the ground It


differs from many of the existing East African examples moreover in the degree of itsdependence on irrigation Few if any of those cited are in quite so arid terrain and somelike those on the southern and eastern slopes of Kilimanjaro and other mountains receivea high rainfall adequate to support forest if cultivation is restricted In a number of thesecases therefore a fair amount of cultivation is possible without any artificial irrigationand in places the latter option may be barely activated in average years necessary thoughit may be for survival through droughts and bad runs But the more important point is thatirrigationmdashor the ability to turn to itmdashhas become an essential element in theseagricultural systems because of the success of the latter over time and the size to whichthe communities have grown This has necessitated more production per hectare than isafforded by the rain alone (at least in years of low rainfall) and therefore the extension ofplanting assisted by irrigation devices into the dry season and the adoption of specialcrops varieties or combinations to suit the complex regime In this way devices that mayat first have been considered optional or supplementary in difficult years would in timehave become permanent and essential complements to developing agricultural systemsand the communities dependent on them

A further factor in the nineteenth century at particular favoured locations adjoining dryplains such as Taveta South Pare and Baringo was the supplying of trading caravansThis required the production of more than a normal surplus or at least facilities forgrowing a fast second crop to restock the granaries Although trade-routes and transport methods have changed in the twentieth century new opportunities have arisen forproduction for local and more distant markets On the slopes of Kilimanjaro and the hillsof North Pare for instance the production on smallholdings of coffee for exportinterplanting it with subsistence food crops has encouraged farmers to maintain theirrigation systems to ensure watering around the year And when coffee prices are low achannelled water supply is still valued for domestic needs in a highly populated ruralarea with stalled cattle kept by these means in some locations

Besides its excessively stony terrain it is doubtless Engarukarsquos extreme situationmdashan impossible one in fact in the eyes of most cultivatorsmdashthat explains the exquisite layout and detail of its fields and irrigation system over so wide an area Moreover as arguedbelow this situation is demonstrably harsher now than it was when Engaruka was firstinhabited which further explains why most of the ancient field area has remaineduntouched since it was abandoned The first settlers doubtless began on a small scale withrudimentary irrigation works probably on the easier terrain some way downstream of thegorge In time however as the community increased in numbers on the success of acultivation system evolved to handle the peculiarities of the location it would have beenobliged to expand its cultivation area into the more broken and stony ground closer to theescarpment and eventually as high onto the scree as could be reached by waterchannelled 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 themaximum 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 thetime of their abandonment but one must imagine improvised structures of stone andtrash as in existing irrigation systems requiring annual maintenance if not complete


rebuilding The positions of these take-offs can be estimated fairly accurately by tracingthe visible upper stretches of the artery canals back to source The stone linings andembanking of these upper artery canals are preserved quite spectacularly in somestretches along the base of the escarpment scree (Fig 116) around the small hills on the edge of the plain and through the whole area of ancient field remains (Fig 112) Every effort was made to keep these as close to the horizontal as was practicable for purposesof controlling the flow and preventing undue scouring of the canal beds and breaches inthe furrow walls and equally important to increase the area of the plain and of the sidesof hills standing in it which could be reached by the furrowed water Despite theunavoidable rapid descent or small cascade here and there on the steep and rockyescarpment for most of their lengths these canals fall at angles less steep than 120 and inplaces as gently as 1100 The existing examples of Marakwet and Sonjo illustrate howsuch engineering and levelling perfection can be achieved through a combination ofexperience and trial-and-error

The longest of these artery canals is that running northwards from the gorge of the Engaruka river it measures 1ndash2 m wide between its stone edges although since therewas no laid bottom the actual water flow over the gravel and silt bed would doubtlesshave been narrower It is traceable up to 3 km from take-off with divisions at various points Along the canalrsquos upper stretch as it descends the escarpment (at a relatively steepangle between 115 and 120 Adams 1986) its lower side is substantially supported bygravel embanking When it reaches the foot of the escarpment scree it swings at a rightangle towards a hill standing separately in the plain But in order to maintain as muchheight as possible the canal is carried across this narrow valley on an embankedcauseway (an aqueduct in effect) up to 3 m high (Fig 117) By these means it achieves an advantage when it reaches the hill where it divides to run as contour furrowsconstructed round each side It appears from the levels of the latter that the effect of theembanking 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 thewater over the porous gravel These belong to an evolved stage when the highest arterycanals were constructed and clearly represent an effort to gain the maximum advantagefrom the available water regardless of the correspondingly intensive demands this placedon the hydraulic ingenuity of the community and the sheer labour required forconstruction and constant maintenance of the works It appears moreover that largeareas of the fields were relaid to accord with these long embanked canals this isindicated where series of older field divisions and feeder furrows are superimposed by orincorporated into later grids with variant alignments

AN INTEGRATED AND CIRCUMSCRIBED SYSTEM UNDER STRAIN

At some point the limits of feasible improvements and of the communityrsquos technological resource would have been reached Since it was not possible to increase the area ofreliable cultivation beyond that to which water could be carried by gravity through the


highest and longest canals a crisis must have been faced as the population attained themaximum that this finite amount of land and water could feed despite all the complexityand ingenuity of the irrigation devices and other specialized elements of this integratedsystem In fact these doubtless together with the operation of the most productive croprotations available (sorghum and varieties of pulses in particular) combined withratooning of the sorghum to obtain a supplementary harvest and the application of cattlemanure to maintain fertility and improve yields may be seen more as reactions to thelimitations of the situation rather than methods devised to achieve agricultural efficiencyand increased productivity for their own sake Equally likely in so intricate andspecialized a set of arrangements there would have been a danger of trying to intensifytoo far in reaction to stress in particular shortening the fallow would have exacerbatedsoil-exhaustion and erosion Indeed despite all the effort of levelling and terracing to counteract these tendencies the upper fields became denuded through heavy wateringthe field divisions and furrows standing remarkably prominently here while as notedthe lower ones are overlaid with redeposited soil with the old field lines there beingvisible only in recent gully sides

At the same time this population was having to contend with hydrological declinewith less water flowing off the escarpment This process is strikingly illustrated by theexistence of canals leading off the gorges of the seasonal and occasional streams whoseflows are now far too inadequate to reward such labour It is not necessary to concludefrom 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 clearnevertheless that they must have enjoyed longer flows than now with sufficient volumesof water in their catchments to ensure their persisting some time after a period of rainThis argument applies especially to Olemelepo (no 1 on Figure 112) where any attempt to reopen the canals that led from that gorge and to reactivate the archaeological fieldsthat cover its outwash fan would be pointless now Even more striking is the case of theintermediate north gorge (no 3)mdasha narrow cleft in the escarpment from which waterissues in occasional years and then only for a few days following exceptional stormswhen irrigation would be least needed But at the time when the ancient settlementflourished it was found worthwhile to construct short canals along the foot of theescarpment on either side of this gorge to irrigate a grid of fields (Fig 115) on this small outwash (being too high for watering from the long northbound canal led through thevalley below from the main river) Immediately above those canals were built the twonorth-most villagesmdasha further indication that there must have been a natural flow fromthis 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 ofreliable flow than obtains now By implication too the volume of the Engaruka riveritself would have been greater then so that when the declining trend set in it may havebecome insufficient to irrigate adequately the whole basin dependent on its canalsAlternatively as the performances of the seasonal streams and of the fields on theiroutwash fans became increasingly unreliable the need would have arisen to channelwater as broadly as possible from both sides of the main river This was effected throughthe long cross-valley canals with take-offs at the gorge opening that were designed to


deliver water from the main river into the middle and lower parts of the Olemelepo andMakuyuni basins at seasons when those rivers had driedmdashprojects that required relaying of feeder furrows and field grids as already noted From the plan (Fig 112) these alternative sources of water and routes for channelling it may lend an impression ofwonderful flexibility In practice however there would usually have been little choicethe complex arrangements being dictated by the sheer necessity of carrying water to aslarge an area of fields as could possibly be reached from the main river In years of lowrainfall and therefore heavy dependence on irrigation this may have exhausted the wholevolume of the riverrsquos flow at certain seasons (as can happen nowadays although the areaof existing cultivation downriver is not as extensive as that of the ancient settlements attheir 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) theneed to irrigate or at least to devise such elaborate arrangements would not have arisenThat notwithstanding the archaeological evidencemdashthe configuration of fields canals and villagesmdashdemonstrates clearly enough a change in the performances of theescarpment streams with their discharges now being definitely less than they were 500years ago Equally clearly these changes occurred or at least began during the life of theold settlementmdashthat is by the seventeenth century at latestmdashsuggesting that together with the strains imposed by the circumscribed situation they constituted the main factorin the collapse and desertion of Engaruka no later than the eighteenth century The datingis not perfectly precise being based on a number of radiocarbon results from excavationsin the villages on the generally late iron age affinities of the pottery and other artefactsas well as on the lack of any clear local memory about the former inhabitants (Sutton1998)

It is presumed that the decline in the flows of the escarpment streamsmdashso crucial at Engarukamdashwas 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 thehighlands behind it where the escarpment streams rose with the effect that dependenceon the latter would have been increasing just as their flows were declining Any suchclimatic trend ought to be detectable in the archaeological and geomorphological recordof the broader region and especially in lake deposits although no consistent body ofevidence can be cited at the present stage of research In the broader region of thehighlands and Rift Valley in Kenya as well as northern Tanzania there are individualinstances of springs that are now dry and of late iron age settlements in marginal areasthat are now deserted but these have not yet been dated precisely enough forchronological comparison Nevertheless the essential question and the approximatedating 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 environmentaldamage caused by a cultivating community of several thousand people over a period ofthree or more centuries Wood requirementsmdashfor fuel for fencing the villages and forhouse building and equally for the scaffolding and hollowed logs employed for carryingstretches of the canalsmdashwould have placed substantial demands on the forest resources on and above the escarpment Arguably such deforestation could have affected surface


moisture and the aquifers in the mountains and therefore the flows of the escarpmentstreams rendering them more liable to sudden spates and equally sudden failureHowever the scale of the change suggests that a human factor of this sort can be onlypartly responsible and that a decrease in rainfall must have occurred beginning in aboutthe sixteenth century The size of that decrease in the highlands need not necessarily havebeen so substantial but merely enough to diminish significantly the normal discharges ofthe streams descending the escarpment and their reliability for regular irrigation at thebottom

Whatever the cause or combination of factors Engaruka was comprehensivelyabandoned at some point around AD 1700 or perhaps somewhat after on the (admittedly imprecise) dating indicators available The state of the abandoned field and irrigationsystem especially its upper part and of the numerous homestead platforms in the sevenvillages immediately above the top canals lends an impression of sudden desertion ratherthan slow decline and piecemeal abandonment That is not easily testable and theimpression may be illusory but it invites one to speculate on other causes of desertioneven catastrophic ones The possibilities are legion Among those that have beensuggested are a violent earthquake along the Rift fault arguably upsetting the flows ofthe mountain streams and the canal take-offs an unusually heavy eruption of the nearbyvolcano Oldonyo Lengai coating the fields with sulphurous ash or a devastating attackby expanding Maasai pastoralists (or by Tatoga before them) anxious to secure the waterof the Engaruka river and the adjacent grazing There is no direct evidence to support anyof these speculations and the last would appear unlikely since any pastoral group in thearea would have been outnumbered and would also have benefited from exchange ofproducts with an agricultural community in its midstmdashassuming 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 specializedand integrated irrigation agriculture was inherent in that system which being physicallycircumscribed 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 itinevitably generated in particular its own demographic success While this generalexplanation does not rule out other possible contributory factors it is supported by theexistence 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 tothose of Engaruka (see Figure 111) These obviously belonged to the same cultural group and ethnicity and were presumably abandoned about the same time as was themain settlement and cultivated area at Engaruka (although it is quite possible that some ofthese outlying sites on the far side of the Crater Highlands by Lake Eyasi and in thenortherly direction above Lake Natron may have been abandoned earlier or contrarilyhave lingered on a little later) Whatever the exact chronology this general phenomenonof abandonment of settlements over a radius of some 60 km suggests an inherent andunderlying 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 asthat of Engaruka which evolved over two or three centuries was surely a story ofsuccess in adjusting so effectively to the peculiarities of its own special environment It is


more important to understand how it worked and succeeded than to worry about why itcollapsed eventually or again (the common antiquarian reaction to stone ruins) wherethat population lsquowentrsquo The answer to that last question is that the degree of specializationand 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 Engarukathat they could not be transplanted in other words this ethnicity would have expired asthose villages and their fields had to be abandoned Remnants presumably took refugeand became assimilated among other peoples of the region thereby losing their Engarukaidentity

One of these may have been Sonjomdasha 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 lsquoislandrsquo of Bantu-speaking cultivators surrounded by Maasai pastures the latterconsisting of poor scrubland in the Rift to the east but also the extensive plateaugrasslands of Serengeti to the west Each of the main existing Sonjo villagesmdashas well as some that have been abandoned for a whilemdashhas 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 withthe normal rainfall being slightly higher than at Engaruka so that Sonjorsquos dependence on its irrigation works is not as extreme as was that of Engaruka Cattle manure is notapplied as fertilizer to the fields and Sonjo as far back as reliable information goes havenot kept cattle for fear it is said of provoking neighbouring Maasai to raid (This is tooverlook some experimentsmdashand mixed experiences in contending with both cattlediseases and neighbouring Maasaimdashin the 1970s and 1980s) There is moreover littleuse of stonework in the field divisions and canal banks as is so distinctive at Engarukathe basic reason being the relative paucity of surface stone in the alluvial soils of theSonjo 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 homesteadplatforms (Fig 119) and also for public areas and some rather peculiar open-air fireplaces Fairly close parallels for these have been revealed by excavations in theEngaruka villages (Fosbrooke 1938 Sassoon 1967) Pending excavations of old Sonjosites it is not known how far back these features may be dated there but it seems likelythat some of the Sonjo villages were settled before the collapse of Engaruka PossiblySonjo and Engaruka constituted the northern and southern wings respectively of a singlecultural group of which only the one has persisted to the present In that case a detailedarchaeological study of Sonjo should be revealing about the regional history This studyshould compare the existing Sonjo settlements that is the lsquotraditionalrsquo ones destroyed in 1975 during the Tanzanian governmentrsquos lsquovillagizationrsquo (ujamaa) campaign and those deserted at unspecified dates in the nineteenth or preceding centuries Such an exerciseshould be expected to carry the sequence back towards the time of Engaruka and thusillustrate the connection Even if it transpires that the Sonjo people are not relatedlinguistically or in a direct cultural sense to those formerly inhabiting Engaruka (seeNurse and Rottland 1993) a fuller study of their villages both existing and desertedshould contribute to an understanding of


Figure 119 Sonjo wooden house with thatch dome on stone-revetted platform-terrace in Oldonyo Sambu village (Kura)


specializedmdashor what are commonly called lsquointensiversquo (Sutton 1984 Widgren this volume Chapter 14 Widgren and Sutton 1999)mdashagricultural practices in isolated situations as raised by the archaeological record at Engaruka

ACKNOWLEDGEMENT

The authorrsquos continuing field research on the agricultural settlement of East Africa is supported by an emeritus fellowship awarded by the Leverhulme Trust held at the BritishInstitute in Eastern Africa and the Pitt Rivers Museum in Oxford

REFERENCES

Adams WM (1986) Observations on the Engaruka irrigation furrows and riverdischarges Pp 49ndash51 in JEGSutton The irrigation and manuring of the Engarukafield system Azania 2126ndash51

Adams WM (1989) Definition and development in African indigenous irrigationAzania 2421ndash7

Adams WM Potkanski T and Sutton JEG (1994) Indigenous farmer-managed


irrigation in Sonjo Geographical Journal 16017ndash32 Anderson DM (1989) Agriculture and irrigation technology at Lake Baringo in the

nineteenth century Azania 2484ndash97 Fosbrooke HA (1938) Rift Valley ruins Tanganyika Notes and Records 658ndash60 Hennings RO (1951) African Morning London Chatto amp Windus Murdock GM (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 linguistsrsquo view

Azania 281ndash5 Robertshaw P (1986) Engaruka revisited excavations of 1982 Azania 211ndash26 Sassoon H (1966) Engaruka excavations during 1964 Azania 179ndash99 Sassoon H (1967) New views on Engaruka Journal of African History 8201ndash17 Soper RC (1983) A survey of the irrigation systems of the Marakwet In BE Kipkorir

RCSoper and JWSsenyonga (eds) Kerio Valley Past Present and Future 75ndash95 Nairobi University of Nairobi Institute of African Studies

Sutton JEG (1973) The Archaeology of the Western Highlands of Kenya Nairobi British Institute in Eastern Africa Memoir 3

Sutton JEG (1978) Engaruka and its waters Azania 1337ndash70 Sutton JEG (1984) Irrigation and soil-conservation in African agricultural history

Journal of African History 2525ndash41 Sutton JEG (1985) Irrigation and terracing in African agricultural history

intensification specialisation or over-specialisation In ISFarrington (ed) Prehistoric Intensive Agriculture in the Tropics 737ndash64 Oxford British Archaeological ReportsInternational Series 232 Volume 2

Sutton JEG (1986) The irrigation and manuring of the Engaruka field system Azania2126ndash51

Sutton JEG (1988) More on the Nyanga terraces the case for cattle manureZimbabwean Prehistory 2021ndash4

Sutton JEG (1989) Towards a history of cultivating the fields Azania 2498ndash112 Sutton JEG (1993) Becoming Maasailand In TSpear and RWaller (eds) Being

Maasai Ethnicity and Identity in East Africa 38ndash60 London James Currey Sutton JEG (1998) Engaruka irrigation agriculture in the northern Tanzanian Rift

Valley before the Maasai era Azania 331ndash37 Thorp C (1986) Engaruka faunal remains Pp 21ndash26 in PRobertshaw Engaruka

revisited excavations of 1982 Azania 211ndash26 Watson EE Adams WM and Mutiso SK (1998) Indigenous irrigation agriculture

and development Marakwet Kenya Geographical Journal 16467ndash84 Widgren M and Sutton JEG (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 750ndash1200 m of rainfall a year Nyanga in the eastern highlands of Zimbabwe (Fig 121) 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 tomany of the questions addressed in this volume even if overall aridity was not theprimary driving compulsion The landscape of Nyanga and adjacent areas to the west isindelibly printed with the lsquolandesque capitalrsquo remains of past agricultural activities These take the form of stone-faced terraces and lowland cultivation ridges together withassociated stone-built settlement structures in all covering around 7000 km2 (Soper 1996) Whilst the agricultural features themselves are difficult to date the settlement sitesrange from about AD 1400 to 1900 with the earlier sites having no direct association asyet with the agricultural features

The area south of Nyanga town consists of a broad dissected plateau at around 1800 m above sea level falling relatively gently to the southwest to the main watershed betweenthe Zambezi and SabiLimpopo catchments To the east it rises to Mount Nyangani atnearly 2600 m beyond which are steep mountains and valleys into Mozambique Forabout 60 km north of Nyanga the highlands narrow progressively to a high ridge ataround 2000 m with higher peaks and with steep escarpments to east and west To thewest of this ridge granite inselbergs form often substantial hills rising from a base levelof around 1200 m while dolerite sills and dikes form lesser features The highland rangeextends northwards at a lower level for another 20ndash30 km while the surrounding lowlands decline to around 900 m The underlying geology consists of various granitesoverlain by sedimentary rocks and dolerites that cap the highlands

Drainage radiates from Mount Nyangani into major rivers such as the Gairezi andNyangombe to the north-northeast and north-northwest and

Figure 121 Location of the Nyanga area Zimbabwe

the Pungwe to the south Annual rainfall is almost entirely between November andMarch with the average ranging from c750 mm in the northern lowlands to 1200 mm ormore in the highlands Annual variation may be as much as +minus50 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 upto 100 terraces (Fig 122) The altitudinal range is from about 900 m in the northern lowlands to around 1700 m on the escarpments and in the highlands with very littleabove this level which is about the upper limit for the cultivation of traditional graincrops at the present day

Study of aerial photographs has identified a minimum area of 22000 ha of terracing excluding sporadic outlying occurrences Distribution favours dolerite soils and rocks Onthe geological map sheet covering 2750 km2 within which the main concentrations ofterracing occur (Stocklmayer 1978)

The agricultural landscape of the Nyanga area of Zimbabwe 215

Figure 122 Terraced hillsides in the Nyanga lowlands Photograph RSoper

over 19000 ha of terracing have been plotted of which 42 per cent are on dolerite 57 percent on granite and less than 1 per cent on sedimentary rocks (not well represented in thisarea) However 26 per cent of the dolerites below 1675 m are terraced as against only55 per cent of the granites and most of the latter are adjacent to dolerite ocurrences Thedolerites weather to red clay loams or sandy clay loams of greater fertility than the sandygranite soils but are heavily leached on the highland plateaux The younger slope soilshave more inherent fertility but are often thin and very stony so that terracing isnecessary to clear the stones and concentrate the soil for cultivation Terracing alsoprovides fairly level surfaces protects against erosion and impedes drainage to allowwater percolation

Terrace surfaces are generally narrow commonly between 15 and 3 m except on very gentle slopes where they may be up to 10 m wide Fall between terraces is normallybetween 25 and 80 cm except on the steepest slopes Slopes of up to 30 degrees wereregularly terraced in some cases up to 40 degrees Construction varies with geologytopography and the amount of stone to be disposed of and possibly also with datethough the latter remains to be established The best terraces have substantial wallsaround a metre in thickness with a double facing of large stones and a fill of smallerstones A low lip is usually present but the wall may rise a metre or more above theupper terrace surface where there was a large amount of stone to clear Such terraces


now have a more or less horizontal profile which could be the result of soil movementsince abandonment Terraces are not precisely levelled on the contour allowing forlongitudinal drainage so that it could not have been intended to flood them eitherartificially 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 50ndash60 cm against the lower wall face It is relatively stoneless so that it must have been worked over toremove even the smallest stones during construction The substratum is of denselypacked stones in a red clay matrix in the case of dolerite or more or less decomposedrock 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 morethan a simple revetment while terrace profiles are sloping with gradients of up to 15 oreven 20 degrees being recorded The only excavated transect showed numerous stonesremaining in the soil This type may represent the rapid exploitation of less favourablebut 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 whereold water furrows do traverse ranges of terraces and they may well have been used forirrigating those below but no distribution channels have been observed and settlementsites also appear to have been served In the case of the detached hills to the west manyof 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 belowunder landscape development The only direct radiocarbon date for a terrace (Pta-7601) is 200 +minus50 BP calibrated to anywhere between 1618 and 1878 at one sigma This datewas obtained from tiny disseminated charcoal fragments in soil of a second phase ofterrace construction in a granite area adjoining a stone enclosure A total of 537 sherdsmostly 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 thelatter appears to date from the later nineteenth century and there are some differences inthe pottery so they may well derive from an earlier site perhaps contemporary with theearlier terrace phase In either case the date gives only a maximum age which is not veryuseful in view of the wide calibration bracket

Cultivation ridges

The second notable feature of the old agricultural landscape comprises extensivenetworks 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 thetotal area must equal or exceed that of the terracing In the terminology of Denevan andTurner (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 closerspacing needed for effective drainage The features are parallel or sub-parallel linear banks usually 7ndash10 m wide between ditches up to a metre or so deep They often run for


several hundred metres with a more or less shallow longitudinal gradient These occurboth in areas of impeded drainage (termed lsquovleisrsquo) and on the valley sides or interfluves

An example may be described at the base of the main escarpment near Maristvalesome 40 km north of Nyanga town Here there is a broad bay in the escarpment about 2km wide between high projecting spurs and a series of streams converges across thepiedmont slope Virtually the whole of the interfluves and most of the stream valleys arescored with ridges and ditches covering around 1000 ha The central interfluve (Fig 123) provides an area some 1750 m long and around 500 m wide with a longitudinal fallof c60 m and a maximum lateral height of c10ndash12 m Almost all of this is occupied byridges except for a stonier crest towards the upper end which is terraced and a fewminor areas of outcropping rock with stone enclosures The ridges trend longitudinallydown the interfluve with a broadly parallel alignment sometimes rather braided Asimilar pattern is seen on the other interfluves At the head of this interfluve at the base ofthe 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 theoccupation sites Towards the lower end of the interfluve a furrow did carry water from adeep set of ditches diagonally across the ridges probably to a stone enclosure on thecrest Soils here are silty sands over a sheet of consolidated rounded quartz gravel

Other occurrences are in more specifically waterlogged areas An example is aregularly 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 asection showed a metre or so of mottled sandy clay loam overlying dense black clay Theclay loam must derive by erosion from the terraced area immediately above perhapsbefore terracing anchored the soil or perhaps from inefficient use of the terraces It wasthen re-exploited after deposition

Ridge size patterns and orientation to slope appear to vary even within a singlelocalized drainage basin and must represent a flexible system of balancing drainage andwater retention under varying conditions of soil slope rainfall and seasonal water tableIn the first case described the primary purpose would seem to be controlled drainageraising the cultivation beds above any actual or potential waterlogging without removingrainfall too directly If necessary supplementary water could have been introduced to theditches though the water available at present would seem


Figure 123 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 wherethere is permanent or periodic inundation or waterlogging but with some water retentionin the ditches still available to crop roots wide beds reducing the ditch area aeration ofthe soil and modification of microclimate if there is danger of frost To these could beadded the variation of moisture availability across the ridge and ditch appropriate todifferent crops Moisture-loving traditional root plants such as Colocasia (taro) and Zantedeschia (calla lily) would be appropriate for the wetter ditches while sorghummillets and legumes could grow on the ridges with Plectranthus (lsquoLivingstone potatorsquo) 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 andthey are different from recent mihomba cultivation ridges which are shorter narrower


straighter and generally restricted to waterlogged environments such as wet stream banksThe intimate relationship to terraces in the Maristvale area suggests contemporaneitywith at least some terracing It seems likely though that the large labour demands forconstructing and operating both systems simultaneously on a large scale would have beenbeyond the capacity of individual communities

Water management

Water is a critical resource in African agriculture generally and its management infavourable conditions can provide insurance against bad rainfall years and extended dryperiods within a normal wet season as well as giving the potential to extend the growingseason before or after The possibility of supplementary water supply to some ridgesystems has been mentioned above as has the general lack of evidence for widespreadterrace irrigation Terraces and cultivation ridges even if not directly irrigated reflectwater 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 furrowshave been observed mainly in the highlands where they are better preserved by perennialgrass cover A tentative classification of these furrows can be suggested

Type 1 is the commonest and most widespread in the highlands and would have serveddomestic requirements livestock and homestead gardens Some could be diverted toflush out stone-lined pits used for livestock and provision was often made to impoundthe resultant slurry Only a few cases of Type 2 have been recorded both in the highlandsand on the lower escarpment slopes while the only case of Type 3 known to date is thatdescribed above Type 4 appears to be restricted to a limited area centred on the northernpart of Nyanga National Park and must have been for irrigation of unterraced fields sinceno terraces are associated below them and only very rarely are settlement sites servedType 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 beassociated with seventeenth-century sites

Authorship

The authorship of the agricultural works can almost certainly be attributed to theancestors of the present indigenous inhabitants (that is before the relocation ofpopulations consequent on colonial land policies) These are the Unyama people for the

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


area north of Nyanga town the Manyika to the south and the Maungwe west of theNyangombe river Genealogies and traditions of the chiefly families (Beach 1995) goback well into the eighteenth century at least and more in the case of the Manyika and itis surprising that more oral traditions have not survived on the construction and use ofterraces and ridges It would seem that knowledge and use of these specializedagricultural techniques were common to a number of political and dialect groupings innortheastern Zimbabwe and should not be attributed to a single group

AGRICULTURAL SYSTEMS

The agricultural systems of the terrace builders integrated crops and animals Cattle werealmost certainly penned in a sunken stone-lined pit or small stone enclosure within thehomestead In the case of pits roofed or tunnel-entrance passages would have admittedonly dwarf cattle bones of which have been recovered from the only site with good bonepreservation (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 andinternal enclosures in the lowlands are relatively small with an internal diameternormally around 3 m and a depth or height of around 12ndash15 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 1970Netting 1968) but the heightdepth rules out any substantial accumulation of manure in situ Pits in the highlands are larger and deeper usually 5ndash9 m in diameter and 180ndash3 m deep More cattle are thus indicated above the terrace zone where the depth could haveaccommodated the accumulation of manure but was more likely for protection from thecold winds of winter Goats and possibly sheep were kept in the houses many of whichhave 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 thusregularly removed and must have been used for manure with some possibly being driedfor fuel in the highlands where wood may have been at a premium Pits were providedwith 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 homesteadto catch the slurry or ditches were dug to channel it to small hollows Many such pitstructures have radial walls which are thought to have sheltered gardens on which theslurry could have been used Where no furrow was available as more particularly in thelowlands dung must have been removed by hand and any flushing have relied onrainwater goat dung must similarly have been removed by hand from the housesDomestic 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 beenused on homestead gardens irrigated where practicable and on terraced or other plots inthe vicinity but rarely on more outlying fields Results of phosphate analysis from thearchaeological contexts are ambiguous regarding the extent of manuring

Terracing per se could thus be considered a specialized technique implying only


relative lsquointensificationrsquo but a higher degree of the latter may be postulated in an inner zone around the homesteads probably dependent on available water supply Thecultivation of outlying terraces even on the more fertile dolerite soils would have beenless sustainable and a continuous process of terrace building can be envisaged witholder 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 Ziwaruins in the lowlands at around 1300 m These comprised mainly traditional grains andlegumes 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 thepresent research have not added any cultivars to this list Enquiries about traditional cropsadd a number of important root crops Plectranthus esculenta (lsquoLivingstone potatorsquo) Colocasia (taro) and probably Zantedeschia (calla lily) as well as pumpkins andcucumbers 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 ofColocasia and also Zantedeschia are toxic without extended boiling Traditional cropping practices commonly involved interplanting of grains legumes and cucurbits It isprobable that outlying terraces were devoted mainly to grain staples but predation bywild 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 soildepth and quality would have enabled more attention being given to the individualrequirements of different plants

LANDSCAPE DEVELOPMENT

The processes and sequence of landscape modification are not yet well understood but itis unlikely that the terracing and ridging were the work of a large and dense populationover a relatively short time period The Unyama people within whose territory thegreatest concentrations of terracing occur were not sufficiently important to attract anyattention or record from the Portuguese who interacted closely with the Mutapa state tothe west and also with the Manyika immediately to the south The settlement patternrepresents loosely dispersed homesteads in village groupings Although the stone-built homesteads are very numerous and may be locally concentrated especially in lowlanddolerite areas none appears to represent prolonged occupation and there are very fewstratified sites or substantial middens We must therefore see terrace construction as anongoing process over many generations among the communities of a fairly limitedoverall population There is some indication of the reoccupation of homesteadssuggesting that whole settlements and their fields may have been fallowed and resettledtaking 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


apparent Thirty radiocarbon dates are now available all from the northern area from justsouth of Nyanga town All those later than about 400 BP have very wide calibrationranges

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 ataltitudes over 2000 m followed in the seventeenth century by pit structures nowcompletely ruined at slightly lower altitudes These sites are all above the level ofterracing so there is no direct association Later highland pit structures tend to be loweragain and are relatively well preserved some with surviving dhaka (clay) walls Their construction must have continued well into the nineteenth century and there may havebeen a hiatus in highland occupation from the earlier ruined pits perhaps occasioned bythe second severe phase of the lsquoLittle Ice Agersquo (Tyson and Lindesay 1992) Occupants ofthese sites must have been responsible for the terracing on the western escarpmentsFurther south the banked furrows of the National Park area with their implied irrigationof unterraced fields are a local perhaps relatively late development probably also thework 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 betweenthe second half of the seventeenth century and the early or even late nineteenth centuryEarlier sites may exist here or elsewhere in the lowlands but have not been dated orperhaps not recognized if not built in stone so it is not known if there was any occupationcontemporary with the earlier highland sites The extensive terracing of the Ziwa area with which the stone enclosures are associated can probably be bracketed between theseventeenth and early nineteenth centuries most of the western lowlands between theescarpment and the Nyangombe river were depopulated by the end of the nineteenthcentury when the first European travellers passed through Further north there is adifferent 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 fromsecondary contexts (including the terrace date quoted above) suggest occupationcontemporary 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 Terracingwould have concentrated initially on dolerite soils and then spread to adjacent graniteareas Ultimately terraceable land may have run out and the fertility of homestead plotsproved unsustainable resulting in piecemeal or wholesale removals to new sites In thisway the impressive agricultural landscape we now see could have been created with arelatively 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 werealways exploited by terrace builders where conveniently available and that ridgepractitioners 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 interms of labour requirements Either lowland practice switched from a concentration onterracing to one on ridging (or vice versa) or each local community emphasized one orthe other system according to the type of land available While little direct research has


been done in areas to the west and southwest it may be noted that the ridging systemscontinue to the RusapeHeadlands area but terracing becomes more sporadic probablyconcentrated mainly on the limited dolerite occurrences Each system was probably aparallel exploitation by related communities

DISCUSSION

Although the resources of Nyanga would have been less critical than in the more aridsituations that are the focus of most studies in this volume the tactics of soil and watermanagement show many parallels The agricultural systems represent a range ofspecialized responses tailored to the potentialities of the local environment The relativefertility of the younger slope soils was clearly appreciated and their potential was realizedby the development of appropriate terracing technology for stone clearance soilconservation and control of drainage The cultivation ridges enabled the exploitation ofthe more leached and often waterlogged valley soils The alternative options of ridgingand terracing complemented each other and provided a risk strategy for coping with short-term climatic fluctuations emphasizing either terrace cultivation in wet years or thevalley soils in dry years This might be within a single community where both resourceswere available or by reciprocal co-operation between local communities A similar co-operative relationship may be envisaged between highland and lowland communitieswith greater concentration on cattle and cultivation respectively Water resources wereexploited by furrow technology for domestic convenience and garden irrigation to extendthe growing season and lessen the effects of dry spells Integration with livestockmanagement produced manure to extend the fertility span of at least part of the cultivatedland 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 networksand little differentiation in relative wealth Production for basic subsistence is thusindicated Design and construction of homesteads appear to go beyond purely functionalnecessities reflecting no great economic stress while ample storage facilities show anadequate level of food production Some sites such as lsquofortsrsquo with evidence of regular occupation suggest some degree of local authority but no marked social stratificationEthnographic parallels for terracing and irrigation in East Africa in general areconsistently associated with acephalous kin-based social organization within which the agricultural systems are integrated for land allocation labour mobilization and thesettlement of disputes (Haringkansson 1989) Something similar may be suggested here the various chiefships within which the complex fell are unlikely to have had any significantfunction in directing subsistence activities or extracting undue tribute Terrace building asan 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 thewater furrows would also be within the capabilities of the family with the exception ofType 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


derive directly from serious environmental constraints While political constraints areuncertain for the earlier centuries they do not appear to have been particularly pressingfor the eighteenth century defensive structures indicate the need for temporary refugeprobably in response to more or less local raiding but lowland settlement at least wouldhave been vulnerable to any consistent outside threats For explanation one may perhapslook more to the opportunities offered by local circumstances as suggested by Brookfield(1986) whereby innovations adopted for the exploitation of particular niches in this casethe fertility of dolerite slope soils offered a lsquoquantum leaprsquo in productivity Although overall population was low initial relative local density induced by the preference for thedolerite areas provided the necessary labour resources and would have been enhanced bythe resultant success

Reasons may be suggested for the decline and abandonment of the systems but theyremain to be tested Declining fertility in the long term may have reduced the populationbelow a critical level Drastic drought could have been a factor for instance acatastrophic drought occurred in the lower Zambezi and coastal area in the 1820s thoughit did not necessarily affect the Nyanga region In Unyama at least persistent strugglesfor the chiefship between two factions from the late eighteenth century contributed to thedepopulation of large areas of the western lowlands by the time of European penetrationin the 1890s but this should not have affected areas in the neighbouring chiefdoms Thesystems had already survived the last drier cold phase of the Little Ice Agemdashmay indeed even have been a response to itmdashand 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 theBritish Institute in Eastern Africa and the History Department University of Zimbabwein close co-operation with the National Museums and Monuments of ZimbabweGratitude is acknowledged to these institutions and to various agencies of the governmentof Zimbabwe for facilitating the work Particular thanks are due to John Sutton forinitiating 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 HC (1986) Intensification intensified Archaeology in Oceania 21 3177ndash80

Denevan W and Turner B (1974) Forms functions and associations of raised fields inthe Old World Tropics Journal of Tropical Geography 3924ndash33

Haringkansson T (1989) Social and political aspects of intensive agriculture in East Africa some models from cultural anthropology Azania 2412ndash20

Hallpike CR (1970) Konso agriculture Journal of Ethiopian Studies 8 131ndash43


Netting RM (1968) Hill Farmers of Nigeria Cultural Ecology of the Kofyar of the JosPlateau Seattle University of Wisconsin Press

Plug I Soper R and Chirawu C (1997) Pits tunnels and cattle in Nyanga new lighton an old problem South African Archaeological Bulletin 52 16689ndash94

Soper R (1996) The Nyanga terrace complex of eastern Zimbabwe new investigationsAzania 311ndash35

Stocklmayer VR (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 manureZimbabwean Prehistory 2021ndash4

Tyson PD and Lindesay JA (1992) The climate of the last 2000 years in southernAfrica The Holocene 2 3271ndash8


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 thelimitations of rainfall and soils Traditional agropastoralism is characterized by itsrelatively simple technology and high labour demands with small farming settlementsspaced by social and environmental circumstance (Niamir 1991) This lack ofmodernization together with widespread evidence of land degradation is responsible forthe negative perception that has guided successive development plans and conservationistinterventions 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 arereconsidered in the light of new research (eg Behnke et al 1993) It is for example no longer accepted that a state of natural equilibrium would exist were it not for thesupposed effects of agropastoral settlement (Lamprey 1983 Sinclair and Fryxell 1985)On the contrary the fact that dryland environments are prone to climatic variability in theform of unpredictable rainfall events better explains the vicissitudes of agropastoralproduction (Ellis and Swift 1988 Nicholson 1996 Rasmussen 1985119) Very highlivestock densities are required to effect significant vegetation change under theseconditions even under sustained drought conditions (Pratt and Gwynne 1977) Given thedifficulty of maintaining high numbers on insecure resources together with the labourdemands of herding and cereal cultivation dryland agropastoralism should haverelatively 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 (Little1996) 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 Scoones1993 Scoones 1996) There is now an increasing interest in the temporal persistence ofwhat were hitherto considered intrinsic vegetation features and soil conditions (Fairheadand Leach 1996 Frost et al 1986 Hoffman 1997) Indeed the notion of a simple disequilibrial relationship between agropastoralism and dryland environments has itselfattracted criticism partly for its failure to explain the differential impact of seasonal land

use practices (Illius and OrsquoConnor 1999) The adoption of disequilibrium perspectives(cf Behnke and Scoons 1993) by social scientists concerned with the ecology ofsubsistence farming in Africa (eg Powell 1998 Sullivan 1996) may have drawnattention away from the long-term impacts of subsistence farming in Africa byemphasizing the apparent sustainability of such systems (Mortimore 1998) Nonethelessthere is a clear need for time series data although it is acknowledged that directmonitoring and experimental simulation are not always practicable especially in view ofthe certainty that some environmental processes would operate on the scale of decades ifnot centuries (Coppock 1993) In the circumstances it is not surprising that the potentialvalue of archaeological evidence has been raised in discussions of sustainable drylandmanagement (Blackmore et al 1990 Dennell 1982 Leach and Mearns 19965 Rapp 1985110 Stiles 199516)

My purpose here is to review the archaeological evidence of agropastoral settlement inone particular environment that of southeastern Botswana and to apply to it some of themore recent findings and concepts in dryland ecology In doing so I hope to show thatarchaeological research in dryland environments could by adopting a broader approachmake a useful contribution to contemporary issues such as land degradation I also hopeto alert environmental scientists to some of the major limitations of the archaeologicalrecord and the tenuous nature of inferences concerning past land use practices The firstof the following two sections sketches the archaeological and environmentalcharacteristics of southeastern Botswana and the second gives an outline of results fromthe excavation of a fifteenth-century AD Khami period settlement I conclude with adiscussion of some general implications of the archaeological evidence for drylandenvironmental history

AGROPASTORALISM IN THE MOTLOUTSE RIVER ENVIRONMENT

The Letsibogo area in southeastern Botswana (Fig 131) lies more or less in the centre of the ShashemdashLimpopo 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)mdasha major episodic river course with a narrow fringe of riparian bush on either bank and dependable suppliesof water in


Figure 131 The regional setting of Letsibogo in the ShashemdashLimpopo 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 largevillage Mmadinare having been established as recently as 1912 by Mphoeng brother of

Fifteenth-century agropastoral responses 229

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 importantif 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 theirabundant browse and sparse perennial grass cover (Abel 1993) Letsibogo clearlyexemplifies a disequilibrial ecosystem in the sense of Ellis and Swift (1988) where a lowand erratic rainfall induces wide fluctuations in primary productivity and livestocknumbers leading to the adoption of highly opportunistic land use practices (Behnke andScoones 199311 Westoby et al 1989) However the evidence of land cleared for cultivation as well as advanced soil erosion and the encroachment of dense thornbrushshow that unreliable rainfall is not in itself an effective limitation on the impacts ofagropastoralism at Letsibogo (cf Illius and OrsquoConnor 1999)

The present combination of marginal farming conditions and relatively high population density has not always existed on the Motloutse Recent research points to apparentcorrespondences between climatic perturbations over the last 2000 years and both thedistribution and intensity of agropastoral settlement in the ShashemdashLimpopo 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 thatafter the initial appearance of Zhizo farming settlement early in the first millennium ADa more complex pattern arose in about AD 600 with an apparent hierarchy indicated bythe 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 ofsocial complexity as is evident from the rise of major centres like Mapungubwe in theLimpopo 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 1997448) Huffman(1996a) argues that until about AD 1300 the end of the lsquoMedieval Warm Epochrsquo (cf Tyson and Lindsay 1992) annual precipitation would have had to be at least 150 mmhigher than at present to permit cultivation of sorghum which was a major staple at thattime The decrease in rainfall after AD 1300 therefore inevitably led to the abandonmentof the capital at Mapungubwe in the Limpopo valley As the limits of productiveagriculture retreated to the north a powerful new centre arose at Great Zimbabwe(Huffman 1996b)

Under these conditions the ShashemdashLimpopo basin would have been largely desertedat least by agropastoralists However by AD 1450 the Zimbabwe empire collapsed andbroke in two with one of the new entities centred further west at Khami possibly inresponse to a slight climatic amelioration which in turn allowed some reoccupation ofthe Shashe-Limpopo basin (Huffman 1996b) As conditions improved the first SothomdashTswana people known archaeologically as Moloko (Evers 1984) spread from the SouthAfrican interior to establish themselves in the Shashe-Limpopo basin (Maggs and Whitelaw 1991 van Waarden 1989 Whitelaw 1997) Although these successive and


contemporaneous cultural traditions had recognizably distinct ceramic assemblages(Huffman 1980 Phillipson 1977) all were patrilineal agropastoral economies with acommon Southern Bantu social organization (Huffman 1996b Kuper 1982)

Until recently the archaeology of Letsibogo was unexplored and little was known ofthe relationship between major pre-colonial centres and this somewhat remote and marginal area Detailed surveys and test excavations in the vicinity of the Motloutse-Sedibe confluence near Mmadinare (Campbell et al 1995) have revealed widespread agropastoral occupation in the first millennium AD with the evidence suggesting apattern of short-term shifting cultivation involving localized groups of small homesteadsclustered around rocky outcrops These indications of Zhizo farming settlement tend tobe highly visible due to their exposure by severe sheet erosion and the development ofdeep gullies accompanied by significant root exposure over large areas of woodlandespecially 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 almost400 years later with a rapid influx of Khami settlement The disparity between thisevidence of local settlement summarized in Table 131 and the regional pattern outlined by Whitelaw (1997) may be due to the relatively marginal situation of Letsibogo itselfalthough this has to be confirmed Nonetheless oral tradition identifies the Motloutse asthe southern boundary of the Khami state known as Butua (Campbell et al 1996) implying that this could have been the environmental limit for

a subsistence economy dependent on both rainfed cereal cultivation and livestockhusbandry On the other hand the evidence for an influx of Moloko settlement from

Table 131 Selected radiocarbon measurements from Letsibogo Site no Lab no C14 yrs BP AD cal Range 1part Moloko 79a Beta-80094 400+minus70 1505 1595 1620 1450ndash1645 2 Beta-80092 360+minus70 1530 1545 1635 1475ndash1655 127 Beta-81224 360+minus70 1530 1545 1635 1475ndash1655 26 Beta-81225 280+minus70 1660 1640ndash1680 Khami 86 Beta-80983 550+minus70 1425 1400ndash1425 4 Beta-80979 480+minus60 1445 1425ndash1485 79b Beta-80982 450+minus50 1460 1440ndash1505 125 Beta-80986 710+minus60 1300 1285ndash1325 125 Pta-7774 520+minus40 1434 1421ndash1447 Zhizo 106 Beta-80095 1360+minus50 685 665ndash770 109 Beta-80984 1220+minus60 880 785ndash960 30a Beta-81196 1220+minus50 879 790ndash905 19 Beta-29951 1100+minus50 981 899ndash1013


southern Africa after AD 1500 (Campbell et al 1996) suggests that there may have beenmajor political developments from which it is not possible to separate the environmentalconditions of farming settlement

Khami pottery was found at more than fifty Letsibogo sites but only ten of theseclearly 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 hutswhich faced inward to the site of the cattle enclosure Many of the Khami sites werefound in localities with dense thornbush encroachment and although this may havenegatively influenced the survey results the pattern of site distribution is suggestive Thesettlements vary in size although in terms of granary numbers there are only two generalclasses those with twenty or less and those with more than forty Although none of thesites exhibited stone walling consistent with elite status (Huffman and Hanisch 1987)these disparities in size could indicate some functional differentiation among commonersettlements A hierarchical clustering of the ten selected sites using Wardrsquos minimum variance of distance method (JMP 1995330ndash1) identifies three groups with roughly equidistant centres All three central sites are from among those with larger numbers ofgranaries as well as being the only Khami sites with confirmed cattle enclosures Thedating of the sites in Table 131 suggests that they could have formed a contemporaneousgroup The linkages of the sites together with a hypothetical farming settlement modelare shown in Figure 132

ARCHAEOLOGICAL EVIDENCE FROM A KHAMI PERIOD VILLAGE

Detailed information is available from the northernmost of the three central settlements atLetsibogo Site 125 A radiocarbon date of AD 1434 (Pta-7774) with a 1part range of AD 1421ndash47 places the occupation of the site at the beginning of the Khami period (AD1450ndash1800) The earlier date (Table 131) reported by Campbell et al (1995) is not reliably associated with the evidence discussed here

The site which measures c500 m2 is situated 600 m from the Motloutse on thewestern slopes of a low rocky ridge with good access to water arable soil buildingtimber and grazing Although the immediate vicinity is deeply dissected by erosiongullies the site itself shows little evidence of erosion The surroundings of the site werethickly overgrown with thornbush with the only archaeological indications at the surfacecomprising an arc of granary supports and a lobate area of soil discoloured by asheddung Excavation (Fig 133) revealed a substantial dung deposit and yielded quantities of bone fragments of small stock and cattle as well as abundant pottery and evidence ofboth metallurgy and cotton spinning


Figure 132 The distribution (above) and linkage (below) of Khami period sites at Letsibogo according to Wardrsquos minimim variance of distance method

Key Solid circlemdashsettlement thought to be a focus of a settlement cluster open circlemdashsatellite site


Figure 133 Plan (above) and section (below) of Letsibogo Site 125


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 substantialhut remains However detailed granulometric analyses showed that whereas soil fromthe perimeter of the site and the surrounding area contained very little fine sand and clay-sized material (fraction less than 60 microm) soil from the putative hut area had the sameparticle size distribution as freshly puddled dhaka (Kinahan 1999)mdasha material the villagers of Mmadinare customarily obtain from old termitaries Apparently naturaldisaggregation of the hut structures has created a sealed datum surface in an area wheresheet erosion has over the intervening centuries both lowered the surface and removedthe lighter soil fraction

Soil nutrient analyses strongly confirmed these observations on the layout of the site (Fig 134) Samples from a transect through the site showed high phosphate concentrations only in the area of the stock enclosure A steep decrease in phosphateconcentrations at the downslope edge of the discoloured soil area suggests that animaldung was retained by means of a palisade fence although there is no surviving trace ofthe structure itself By comparison soil nitrogen levels are higher in the area outside thestock enclosure possibly representing an accumulation from the relative concentration ofnitrogen in building timber in the huts and fuelwood consumption in cooking fires Theapparent contrast between the hut area and the stock enclosure would be partly due to thevolatility of nitrogen in dung as well as the concentration of phosphorus as a result ofburning

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 132) Whereas some wild species were represented in the huts and granaries the bone from the stock enclosure wasexclusively of either confirmed or probable sheepgoat and cattle Cranial bones ofdomestic livestock were recovered from all parts of the site but those from the stockenclosure were more fragmented and consisted mainly of loose teeth A clear contrastcould 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 thegranaries This suggests that the granary area rather than the stock enclosure was themain focus of domestic animal butchery and the disposal of bone although the slaughterof 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 puncturemarks attributable to the canine teeth of a domestic dog It is conceivable that significantamounts 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 thatsoil in the granary area was mildly acidic compared with that of the stock enclosure mayexplain this difference (Fisher 1995) Microscopic examination of soil


Figure 134 Distribution of soil nutrient values at Letsibogo Site 125


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 thedeposit (Fig 133) 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 andindicates 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 wellsand globular cooking vessels Utilitarian vessels of this kind probably dominate thepottery assemblage because they were subject to frequent breakage Very large storagevessels which were probably never moved from beneath the granaries or eaves of thehuts and bowls that would have been used only in and around the huts make up a verysmall part of the assemblage Pottery was most abundant around the midpoint of thegranary area (cf Fig 133) 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 seniorman of the village It is therefore significant that evidence of metallurgy in the form ofore slag and tuyegravere fragments was most strongly associated lsquowith this area as were all finds of clay spindle whorls since cotton spinning was the traditional preserve of men inKhami society (van Waarden 1989)

The archaeology of Site 125 at Letsibogo provides several important insights into Khami period settlement in the Motloutse River (Fig 132) The site forms the centre of a settlement cluster the study area as a whole having three such clusters which wereprobably coeval They represent a land use strategy that combined animal husbandry and

Table 132 Faunal taxa from Letsibogo Site 125 Fauna Granaries Huts Stores Totals Reptilia Unid snake 11 11 Tortoise (cf Geochelone) 21 21 Aves Unid gamebird 11 11 Mammalia Unid rodent 11 11 Hare (cf Lepus) 51 51 Procavia capensis 62 21 83 Unid Bovid size class I 11 11 22 Sheepgoat (Ovis ariesCapra hircus) 32 21 31 84 Unid Bovid size class II 51 21 41 113 Cattle (Bos taurus) 91 21 21 133 Unid Bovid size class IV 381 31 181 591 Note The data are listed as NISPMNI (number of identifiable specimensminimum number of individuals) after Klein and Cruz-Uribe (1984) bovid size classes are after Brain (1974)


cereal cultivation There can be no doubt that these activities formed the mainstay of theeconomy for in the case of the cereal crops numerous storage granaries were requiredand the substantial dung deposits must be the result of keeping domestic herds Thepositioning of these major components of the site as well as the distribution of smallfinds 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 anarchaeological sampling strategy that was not informed by these principles could yieldbiased 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 ofinhabitants is not established Although sorghum is likely to have been the main staplecrop the species of grain cultivated by this community is not known and neither is thetype of garden vegetables which would almost certainly have formed part of the dietAlthough no direct botanical evidence was found wild plant foods were probablyimportant here on the analogy of recent studies in Zimbabwe (Jonsson 1998) The stockenclosure confirms the social importance of cattle and the animal bone establishes thepresence of small stock but this evidence does not provide any means to estimate the sizeof the herds or the dietary importance of cattle as opposed to small stock or wild speciesfor that matter Wild fauna may have been more important than the evidence suggestsespecially if game was butchered and eaten away from the settlement Finally there is noevidence beyond that of cultural affinity to reflect on the nature of the relationshipbetween this and other Khami period sites at Letsibogo and further afield These areimportant limitations on the extent to which archaeological evidence can usefullycontribute to dryland environmental history

HUMAN IMPACTS AND DRYLAND ENVIRONMENTAL HISTORY

In general there is a satisfactory match between the archaeological data from Letsibogoand the palaeoclimatic model of Tyson and Lindsay (1992) With the added precision ofHuffmanrsquos (1996a) calibrated radiocarbon dates a good correspondence is achieved between the onset of the AD 1425ndash1675 period of increased rainfall and the first appearance of Khami settlement at Letsibogo However the span of the calibrated Khamidates is but thirty-five years with a maximum 1part range of 100 years (Table 131) Taken together with the hiatus of about 400 years following the curtailment of the earlier Zhizoperiod this points to more variable conditions than the available palaeoclimatic datareveal

It would appear that within the Shashe-Limpopo basin described as a lsquorainfall troughrsquo by Jackson (1961) Letsibogo is relatively marginal as far as farming conditions are concerned By themselves the radiocarbon dates from Letsibogo suggest that the Khamipresence was an opportunistic response to a short-lived climatic amelioration in the ShashemdashLimpopo On this basis it may be argued that the short span of Khamisettlement is a predictable consequence of the inverse relationship between variability ofrainfall and long-term low average annual rainfall (Nicholson 1996) Local rainfallanomalies are not unusual in these conditions and it is possible that the Khami


occupation at Letsibogo represents a short-lived expansion at the margins of theagropastoral environment In situations such as at Letsibogo where rainfall ischaracterized by its variation rather than its average models based on averages (such asthat 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 regionswhere relatively mesic conditions prevail Archaeological proxy data can therefore helpto 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 themore sustained impact on equilibrial systems resulting in potentially more rapiddegradation The non-equilibrial systems may of course be equally vulnerable if there is insufficient recovery time between episodes of impact For purposes of agropastoralsettlement the threshold between equilibrial and non-equilibrial ecosystem dynamics probably lies at an average annual rainfall of around 350 mm as now prevails atLetsibogo Under such conditions droughts are more frequent and severe although theyare interspersed by periods of above average rainfall that may extend over several years(Nicholson 1996) Evidence of cereal cultivation at Letsibogo therefore does notnecessarily 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 ofclearing tilling and weeding fields together with those of livestock impact in the nearvicinity 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 thecreation of a patchy vegetation mosaic reflecting differential pressures of usage on theone hand and favourable germination and regeneration conditions on the other BothCoppock (1993) and Hoben (1996) have pointed to the effect of heavy grazing and theconcentration of nutrients in the dung deposits of stock enclosures Indeed colonizationof these deposits by lime-tolerant Cenchrus ciliaris grass is a notable characteristic of ancient stock enclosures in Botswana and is clearly visible on aerial photographsDenbow 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 tosettlements is lowered by high grazing and browsing pressure (Botkin et al 1981) which tends to exacerbate the patchiness resulting from nutrient concentration in the stockenclosures Coppock (199356) has remarked on the markedly higher fertility of soils inbush-encroached areas demonstrating the beneficial effects of an encroachment phase following a period of heavy livestock utilization Similar observations were reported byReid and Ellis (1995) who recorded higher nutrient levels in the vicinity of abandonedpastoral 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 stockenclosures through the germination of seed in especially goat dung resulting in


characteristic age cohort patches (Kiyiapi 1994) At this stage the duration of theencroachment phase is not known other than from anecdotal evidence (Kempff 1994)although it does appear to extend beyond documented events and into the archaeologicalrecord

Aerial photographs of the Letsibogo area clearly show patchy thornbush cover in thevicinity of the Khami settlements On the ground these patches often include scatteredspecimens of Boscia albitruncamdasha species that would have been conserved for its veterinary medicinal properties (Coates-Palgrave 1981187) Observable correspondences between the Khami site distribution and the physiognomiccharacteristics of the vegetation at Letsibogo suggest that the impact of agropastoralismunder disequilibrial conditions has long-term consequences Other studies indicate thatchanges 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 thatchanges in the vegetation at Letsibogo were initiated at an earlier stage that the Khamiperiod the first-millennium Zhizo settlement pattern has different locationalcharacteristics Nonetheless it is important to consider the effect of more recent land usepractices in the case of the severe erosion visible on the Zhizo sites this is in someinstances attributable to the development of gully systems on cattle paths originating inthe modern village of Mmadinare (Kinahan 1999)

A recent contribution to the range ecology debate by Illius and OrsquoConnor (1999) argues that disequilibrial dynamics would govern that part of the land use strategy inwhich livestock grazing was limited by the availability of water whereas that in whichlivestock 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 keyresource areas used in the dry season would register greater impact At the regional scaleof rainfall distribution domestic crop requirements and vegetation dynamics Letsibogotherefore presents the characteristcis of a disequilibrial system However the evidentlong-term impacts of agropastoral settlement suggest that equilibrial dynamics wouldhave placed definite limits on livestock numbers even if a cattlepost system such as that of the modern Tswana (Shaw 197488) was employed to lessen the degradation of keyresource areas

Archaeological evidence may be highly relevant to the refinement and testing of soilloss 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 forthe next four centuries in contrast with a more alarmist view that radical destockingshould commence immediately to avoid irreversible land degradation Securely datedarchaeological settlement patterns integrated with vegetation distribution density andage cohort estimates would provide essential baseline data for modelling recentenvironmental changes The precision of such data is unavoidably problematic but whenthere are widely variant competing estimates as discussed by Biot (1993) thearchaeological 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 ofwhich I should describe in conclusion Archaeologists often draw broad regional


inferences from very limited even ambiguous field data and this may easily conceallocal variation which is the essential basis of a particular land use strategy Large databases are uncommon due to the time-consuming nature of archaeological sampling andwhile archaeological observations are testable in a broad sense they are not repeatable inthe narrow sense employed by most natural scientists (cf Hempel 1966) This leads tothe second pitfall that of using archaeological data as if they were neutral observationsThe Letsibogo evidence very clearly illustrates the social context of nearly all materialaspects of southern Bantu settlement It would be regrettable if in the need to considerhistorical evidence environmental scientists neglected to consider the social dimensionsof dryland agropastoralism in Africa

ACKNOWLEDGEMENTS

Site 125 was found mapped and tested by C van Waarden in an initial phase of work atLetsibogo reported in Campbell et al (1995) I am indebted to ACCampbellTHoffmann AIllius AReid and Cvan Waarden for critical comments on themanuscript The excavations reported here were commissioned and funded by theBotswana Government Department of Water Affairs to whom I am grateful forpermission to publish this research

REFERENCES

Abel NOJ (1993) Reducing cattle numbers on southern African range is it worth it InRHBehnke IScoones and CKerven (eds) Range Ecology at Disequilibrium New Models of Natural Variability and Pastoral Adaptation in African Savannas 173ndash95 London Overseas Development Institute

Behnke RH and Scoones I (1993) Rethinking range ecology implications forrangeland management in Africa In HBehnke I Scoones and CKerven (eds) Range Ecology at Disequilibrium New Models of Natural Variability and PastoralAdaptation in African Savannas 1ndash30 London Overseas Development Institute

Behnke RH Scoones I and Kerven C (1993) (eds) Range Ecology at Disequilibrium New Models of Natural Variability and Pastoral Adaptation in African SavannasLondon Overseas Development Institute

Biot Y (1993) How long can high stocking densities be sustained In HBehnkeIScoones and CKerven (eds) Range Ecology at Disequilibrium New Models of Natural Variability and Pastoral Adaptation in African Savannas 153ndash72 London Overseas Development Institute

Blackmore AC Mentis MT and Scholes RJ (1990) The origin and extent ofnutrient-enriched patches within a nutrient-poor savanna in South Africa Journal of Biogeography 17463ndash70

Botkin DB Mellilo JM and Wu LSY (1981) How ecosystem processes are linkedto large mammal population dynamics In CWFowler and TDSmith (eds) Dynamics of Large Mammal Populations 18ndash34 New York John Wiley


Brain CK (1974) Some suggested procedures in the analysis of bone accumulationsfrom southern African Quaternary sites Annals of the Transvaal Museum 29 (1)1ndash8

Brain CK (1981) The Hunters or the Hunted An Introduction to African CaveTaphonomy Chicago University of Chicago Press

Brochier JE Villa P and Giacomarra M (1992) Shepherds and sedimentsgeoethnoarchaeology of pastoral sites Journal of Anthropological Archaeology 11 47ndash102

Bryson RA and Bryson RU (1996) Site-specific high-resolution archaeoclimatic modelling for Africa Unpublished conference paper Poznan Society of Africanist Archaeologists September 1996

Campbell AC (1990) The Nature of Botswana a Guide to Conservation andDevelopment Gland IUCN

Campbell AC 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 AC Kinahan J and van Waarden C (1996) Archaeological sites atLetsibogo Dam Botswana Notes and Records 2847ndash53

Coates-Palgrave K (1981) Trees of Southern Africa Cape Town CStruik Coppock DL (1993) Vegetation and pastoral dynamics in the southern Ethiopian

rangelands implications for theory and management In RHBehnke IScoones andCKerven (eds) Range Ecology at Disequilibrium New Models of Natural Variabilityand Pastoral Adaptation in African Savannas 42ndash61 London Overseas Development Institute

Denbow J (1979) Cenchrus ciliaris an ecological indicator of iron age middens usingaerial photography in eastern Botswana South African Journal of Science 75405ndash8

Denbow J (1986) A new look at the later prehistory of the Kalahari Journal of African History 273ndash28

Dennell RW (1982) Archaeology and the study of desertification In BSpooner andHSMann (eds) Desertification and Development Dryland Ecology in SocialPerspective 43ndash60 London Academic Press

Ellis JE and Swift DM (1988) Stability of African pastoral ecosystems alternateparadigms and implications for development Journal of Range Management 41450ndash9

Evers TM (1984) SothomdashTswana and Moloko settlement patterns and the Bantu CattlePattern In MGHall DMAvery MLWilson and AJBHumphreys (eds) Frontiers Southern African Archaeology Today 236ndash47 Oxford British ArchaeologicalReports International Series 207

Fairhead J and Leach M (1996) Misreading the African Landscape Society andEcology 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)7ndash68

Frost PGH Medina E Menaut J-C Solbrig O and Walker B (1986) Responses of


savannas to stress and disturbance a proposal for a collaborative programme of researchReport of IUBS working group on Decade of TropicsTropical savanna ecosystemsBiology International Special Issue 101ndash78

Hall M (1987) The Changing Past Farmers Kings and Traders in Southern Africa200ndash1860 Cape Town David Phillip

Hempel CG (1966) Philosophy of Natural Science Engelwood Cliffs Prentice Hall Hoben A (1996) Paradigms and politics in Ethiopia In MLeach and RMearns (eds)

The Lie of the Land Challenging Received Wisdom on the African Environment 186ndash208 London International African Institute

Hoffman MT (1997) Human impacts on vegetation In RMCowling DM Richardsonand SMPierce (eds) Vegetation of Southern Africa 507ndash34 Cambridge Cambridge University Press

Hoffman MT Todd S Ntshona Z and Turner S (1999) Land Degradation in South Africa Cape Town Department of Environmental Affairs and Tourism

Homewood KM (1992) Patch production by cattle Nature 359109ndash10 Huffman TN (1980) Ceramics classification and Iron Age entities African Studies 39

(2)123ndash74 Huffman TN (1996a) Archaeological evidence for climatic change during the last 2000

years in southern Africa Quaternary International 3355ndash60 Huffman TN (1996b) Snakes and Crocodiles Power and Symbolism in Ancient

Zimbabwe Johannesburg Witwatersrand University Press Huffman TN and Hanisch EOM (1987) Settlement hierarchies in the northern

Transvaal Zimbabwe ruins and Venda history African Studies 46 (1)79ndash116 Illius AW and OrsquoConnor TG (1999) On the relevance of non-equilibrium concepts to

arid and semi-arid grazing systems Ecological Applications 9 (3) Jackson SP (1961) Atlas Climatologique de LrsquoAfrique Lagos Scientific Council for

Africa JMP (1995) Statistics and Graphics Version 31 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 Wuumlrzburg Wuumlrzburger 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 LetsibogoDam near Mmadinare Eastern Botswana Gaborone Aquatech Groundwater Consultants unpublished report

Kinahan J Kinahan JHA and van Waarden C (1998) The archaeology and symbolicdimensions of a thirteenth century village in eastern Botswana Southern African Field Archaeology 7 (2)63ndash71

Kiyiapi JL (1994) Structure and characteristics of Acacia tortilis woodland on the Njemps Flats In RBBryan (ed) Soil Erosion Land Degradation and Social Transition Geoecological Analysis of a Semi-arid Tropical Region Kenya 47ndash70 Advances in Geoecology 27

Klein RG and Cruz-Uribe K (1984) The Analysis of Animal Bones from Archaeological Sites Chicago University of Chicago Press


van der Koppel J Rietkerk M and Weissing FJ (1997) Catastrophic vegetation shiftsand soil degradation in terrestrial grazing systems Tree 12 (9)352ndash6

Kuper A (1982) Wives for Cattle Bridewealth and Marriage in Southern AfricaLondon Routledge amp Kegan Paul

Lamprey HF (1983) Pastoralism yesterday and today the over-grazing problem In FBourliere (ed) Tropical Savannahs 112ndash45 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 PD (1996) Pastoralism biodiversity and the shaping of savannah landscapes inEast Africa Africa 66 (1)37ndash51

Maggs TMOrsquoC and Whitelaw G (1991) A review of recent archaeological researchon food-producing communities in southern Africa Journal of African History 32 3ndash24


Niamir M (1991) Traditional African range management techniques implications forrangeland management London Overseas Development Institute Pastoral Development Network Paper 31d1ndash11

Nicholson SE (1996) Environmental change within the historical period In WMAdams ASGoudie and AROrme (eds) The Physical Geography of Africa 60ndash87 Oxford Oxford University Press

Phillipson DW (1977) The Later Prehistory of Eastern and Southern Africa London Heinemann

Powell N (1998) Co-Management in Non-Equilibrium Systems Cases from NamibianRangelands Uppsala Swedish University of Agricultural Sciences Agraria 138

Pratt DJ and Gwynne MD (1977) (eds) Rangeland Management and Ecology in East Africa London Hodder amp Stoughton

Rapp R (1985) Why do we need a scientific analysis of dryland degradation in AfricaIn AHjort (ed) Land Management and Survival 109ndash18 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 119ndash27 Uppsala Scandinavian Institute forAfrican Studies

Reid RS and Ellis JE (1995) Impacts of pastoralists on woodlands in South TurkanaKenya livestock-mediated tree recruitment Ecological Applications 5 (4)978ndash92

Scoones I (1996) Politics polemics and pasture in southern Africa In MLeach andRMearns (eds) The Lie of the Land Challenging Received Wisdom on the AfricanEnvironment 34ndash53 London International African Institute

Shaw M (1974) Material culture In WDHammond-Tooke (ed) The Bantu-Speaking Peoples of Southern Africa 85ndash136 London Routledge amp Kegan Paul

Sinclair ARE and Fryxell JM (1985) The Sahel of Africa ecology of a disasterCanadian Journal of Zoology 63987ndash94

Skarpe C (1991) Impact of grazing in savanna ecosystems Ambio 20 (8)351ndash6 Stiles D (1995) An overview of desertification as dryland degradation In DStiles (ed)


Social Aspects of Sustainable Dryland Management 3ndash20 New York John Wiley and Sons

Sullivan S (1996) Towards a non-equilibrium ecology perspectives from an arid land Journal of Biogeography 231ndash5

Thomas DSG and Shaw PA (1991) The Kalahari Environment Cambridge Cambridge University Press

Turner MD (1998) Long-term effects of daily grazing orbits on nutrient availability inSahelian West Africa 1 Gradients in the chemical composition of rangeland soils andvegetation Journal of Biogeography 25669ndash82

Tyson PD and Lindsay JA (1992) The climate of the last 2000 years in southernAfrica The Holocene 2271ndash8

van Waarden C (1989) The granaries of Vumba structural interpretation of a KhamiPeriod commoner site Journal of Anthropological Archaeology 8131ndash57

van Waarden C (1999) The later Iron Age In PLane AReid and ASegobye (eds)Ditswa Mmung The Archaeology of Botswana 115ndash61 Gaborone Botswana Society

Westoby M Walker BH and Noy-Meir I (1989) Opportunistic management for rangelands not at equilibrium Journal of Range Management 42266ndash74

Whitelaw G (1997) Southern African Iron Age In JOVogel (ed) Encyclopaedia of Precolonial Africa Archaeology History Languages Cultures and Environments444ndash56 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 populationdensities ranging from less than five to more than 300 per square kilometre AsMortimore (199817) has emphasized this range in population densities cannot beexplained by differences in climate lsquothere is a weak relation between aridity andpopulation density While high densities are rare in the arid zone the higher ones arefound not in the moist but in the dry semi-arid zonersquo 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 formulafor cultivating arid lands can be foundmdasheach farming system relies on its own mix ofcomponents to cope with the two main problems of farming drylands water managementand fertility maintenance This is achieved through crop varieties meticulous time-scheduling of planting farming practices aimed at restoring organic content andconstruction works such as terraces irrigation furrows and so on Furthermoreinteractions 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 andsouthern Africa although the regional distribution of farming systems in the area is onlyvaguely known and documented (Ker 1995) Temporary cultivation with littleinvestment in land is often assumed to be the general rule but several exceptions can bedocumented In West Africa for example lsquoring cultivation systemsrsquo (Fussel 1992494) are practised akin to European infield-outfield cultivation with intensively farmed andmanured 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 betweenyears and between cultivators The SothomdashTswana settlement system in the interior areasof southern Africa has been recognized as an agricultural and social adaptation to low anderratic 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 fieldsand a pastoral organization reaching more than 20 km from the main settlements Thispattern of settlement and land use was contrasted by Sansom (1974138ff) with the

settlement structure of the Nguni peoples on the eastern rim of South Africa whobecause of the higher rainfall and more dissected landscape there were able to base theiragriculture on a confined territory in each settlement

Sansomrsquos thesis has been criticized on good grounds for being environmentallydeterministic (Huffman 1986) The problem is that it operates in a historical and socialvacuum whereas research has shown that the highly concentrated settlements among theSothomdashTswana and among previous populations in the same area reflect social andpolitical hierarchies rather than simply an adaptation to a semi-arid climate However the environmental arguments cannot be dismissed totally on these grounds Within theecological context of semi-arid lands with few topographical variations and hence fewvariations in precipitation the Tswana type of exploitation pattern represents aproduction form that is able to produce a surplus for an elite and to sustain largepopulations 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 neverhave been but exist as small pockets or lsquoislandsrsquo 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 toabandoned agrarian communities with the characteristics of such lsquoislands of intensificationrsquo (Grove and Sutton 1989 Sutton 1984 1985 1989 1998 Fig 141) In 1995 Maggs presented new documentation from Marateng in the Lydenburg area inMpumalanga province South Africa placing it in the context of the previously knownfield systems in eastern and southern Africa (Maggs 1995) In the present volumereports are presented of the ongoing research on the ancient fields at Engaruka inTanzania (Sutton Chapter 11) and from the important comparative example of Nyanga inZimbabwe (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 400years ago so that none has written documentation on their use To judge from theinvestments in land evidenced in these field systems the agrarian communities that builtthem were capable

Islands of intensive agriculture in African drylands 247

Figure 141 Eastern and southern Africa showing sites mentioned in Chapter 14

of solving the basic problems of water management and fertility maintenance Indiscussing the causes of abandonment of similar systems in other parts of the worldBrookfield (1986180) has argued that they lsquowere almost all highly conservationist and it was their breakdown and abandonment that was more likely to yield damage to the landrsquo

As with all large archaeological complexes of this sort the central problems are aboutthe dates of emergence and desertion and the reasons for their rise and eventualabandonment In the discussion of these matters surviving agrarian communities sharingthe same characteristics are important in providing comparative evidence Terracing andirrigation can be found in several locally-developed farming systems in the region suchas Marakwet in Kenya Sonjo in Tanzania and Konso in Ethiopia An overview of thewhole problem of intensive or specialized agriculture in Africa was given in a specialissue 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 emergeddeveloped and decayed derives from different types of situations and source materials Atone extreme there are the cases of Engaruka in Tanzania and of Nyanga in Zimbabwewhich are deserted field systems with poor-to-non-existent historical documentation but


with reasonably well-dated archaeological features At the other extreme are currently-surviving active farming systems like Sonjo and Marakwet where however the historicalorigins are still unclear Into that context must also be brought cases like the Machakos inKenya where a development of intensification and of increasing technologicalinvestment in land (lsquolandesque capitalrsquo) is currently taking place (Tiffen et al 1994) The concluding discussion of the Machakos study serves to remind us of the importance ofstudying the present implications and development possibilities of historical cases (Asimilar approach to intensive agriculture can be found in Bebbingtonrsquos [1997] discussion of the recent development of islands of sustainable agriculture in the rural Andes)

ISLANDS OF INTENSIVE AGRICULTURE

The current definitions of lsquoislandsrsquo and of lsquointensiversquo may both be questioned Recent examples of islands of intensive agriculture share some common characteristics Firstthey are characterized by agricultural systems that have for a long period been able tosupport a larger population density than surrounding areas The metaphor lsquoislandsrsquo is used to describe the fact that these areas may exist within a lsquosearsquo of less-intensive land use such as shifting cultivation or pastoralism Second to judge from the history of therecent examples they have been less fragile and more robust in the face of both droughtand 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 theystillmdashthrough traditional networks of exchangemdashplay an important role in the food-security system They thus represent lessons from the past for the urgent problem of foodsecurity Furthermore the high productivity of land and the robust nature of agriculturalproduction in these areas depend on the application of different combinations of farmingpractices 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 andwoodlands Irrigation and soil conservation are connected with lsquolandesque capitalrsquo investment activities affecting land and vegetation that reach beyond the immediateneeds of the coming cropping season The latter fact is also of crucial importance for thearchaeological identification of past agrarian societies of that kind

Our interest in these systems stems from the fact that they seem to provide historicaland contemporary examples of locally-developed solutions to the critical problems in modern African agriculture low output from traditional systems threatenedsustainability of the production systems andor widespread degradation and unreliableaccess to food

CURRENT FIELD RESEARCH

Through detailed studies in two living agrarian landscapes in eastern Africa we (seeAcknowledgements) are seeking to understand the ecological historical and socialcontexts of this type of intensive farming Two case studies are being carried out inTanzania and Kenya These empirical studies are focused on work processes social


institutions land tenure technology and their material expressions as physical featuresie fields and landscapes The project is based in anthropology and geography andcombines the methods of landscape history with participatory approaches Though theempirical focus is on two cases we are working in close contact with other researchersstudying 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 explanatoryframework to embrace historical questions such as why areas like Engaruka Nyanga andMarateng were abandoned and why areas like Marakwet Sonjo and Konso persist At thesame time the framework should be able to accommodate questions on the futurepotential of these areas and the mechanisms of lsquotake-offrsquo

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 beginwhat were the specific place-bound events and characteristics The second set of questions relates to the social organization that makes possible the mobilization of labourand 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 generationand are at the same time flexible enough to cater for population increase and settlementexpansion The third set of questions not treated here relates to the presentdevelopmental possibilities of these different areas To what extent can they continue toplay 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 irrigationand 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 historicaltradition in Africa and we have little to add to that debate Instead we are trying tounderstand the process whereby such practices are put together in a farming and socialsystem capable of increasing both land productivity and food security in a sustainableway

CASE STUDIES

Mama Issara Mbulu District Northern Tanzania

Mama Issara is the core area of the Iraqw people Agriculture is restrained by thedissected 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 elaborationand has a history that goes back some 200 years The population has been estimated ataround 20000 with a density of around 100 people per square kilometre Terracingmulching manuring and water harvesting are practised (Boumlrjeson 1998 Loiske 1993 199514ndash30 Tengouml 1999 Figs 142 143 and 144)

Mama Issara is a prime example of how local institutions for natural resourcemanagement have been able to uphold an intensive farming system for a long time(Boumlrjeson 1998 1999) Several factors are of importance including strong social


cohesion efficient forms of decision making and a tradition of communal labour co-operation Also religious beliefs support the sustainable use of natural resources in thatthe earth spirit is thought to punish over-use of land and trees As Boumlrjeson (1999) has shown the systems for the transfer of land between and within generations are animportant part of these institutions and play a central role in the reproduction andpersistence 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 expansionareas All the families participate in institutionalized food exchanges involving betweenfive and twenty-five other families These exchanges are based on ritual economic andsocial networks covering areas with varying ecology and varying production (Loiske1999) The islands of

Figure 142 An intensively cultivated landscape at Kwermusl (Mama Issara Mbulu district Tanzania)

Photograph LBoumlrjeson


Figure 143 Preparing the field at Kwermusl Photograph LBoumlrjeson

Figure 144 Piles of manure from stalled cattlemdashan integral part of the farming system in Mama Issara

Photograph L Boumlrjeson

intensive agriculture can thus be seen not in isolation but as manifestations of the


geographical division of labour

Marakwet Kenya

In any discussion of intensified agriculture the Marakwet area constitutes a particularlyinteresting 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 acomprehensive system in which the total length of the furrows now reaches 250 km(Adams et al 1997 Watson et al 1998 Figs 145 and 146) A centralized political system has not developed however

Figure 145 An irrigation channel above Tot in Marakwet Kenya Photograph LBoumlrjeson


Figure 146 An irrigation canal under repair above Chesoi in Marakwet Kenya

Photograph MWidgren

mdashin fact no single individual or group of people has an overview of how the systemworks 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 notpossible This situation has not resulted in a hierarchical society such as might have beenexpected in terms of Wittfogelrsquos classic theory of lsquooriental despotismrsquo which states that


societies with large-scale irrigation will develop centralized orders of command which inturn will lead to despotic political systems (Wittfogel 1957) Though this hypothesis inits 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 asthe Marakwet is organized acephalously

Oumlstberg (1999) has recently summarized some preliminary results on the origin of thisfarming 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 theMarakwet came to develop an intensive irrigation system Co-operation and competition between the agriculturalist Marakwet and the pastoral East Pokot have been instrumentalin shaping the present-day utilization of resources in the valley Unlike an evolutionaryexplanation 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 initiatedby asking the elementary geographical question about location why do we find thesesystems 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 variationsof which provide good opportunities for intensive farming Many of these farmingsystems make use of the variations in precipitation and climate within short distances thatare characteristic of the high escarpments here but their locations can in no way be saidto be simply environmentally determined there are examples of similar environmentsalong the rift valley where neither present intensive farming nor any traces of formerintensive agriculture exist Furthermore areas of intensive farming can also be found inother types of environments in the semi-arid parts of eastern and southern Africa Thedistribution 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 ecologicalsocial 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 centralplace when explaining the distribution of intensive farming In the recent case ofMachakos the proximity to the market in Nairobi is one important explanation but it isin no sense the only one In the case of Baringo (Anderson 1988 1989) the marketsituation also seems to have been of vital importance for the development of the irrigatedagriculture during the nineteenth century Market conditions do not play the same role inMarakwet and Mama Issara however both of which are remote from markets and sufferfrom poor communications

A second explanation for the geographical distribution of intensive farming would bethat islands of locally-developed intensive agriculture are the remains of a type of agriculture that formerly was much more widespread This explanation makescolonialism the main force behind the de-intensification of African agriculture with the


islands being seen as pockets that have survived these developments The problem of pre-colonial farming systems in Tanzania has been the object of a debate that is outside thescope of this chapter but though the advent of colonialism certainly led in many cases tothe disruption of local farming societies it would be too simple to advance it as the mainforce 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 (Koponen1988240f)

The above-mentioned models of agrarian development are all based on the idea of an even development of farming systems in response to markets andor population pressures I find it more challenging to start from the opposite assumption that social systems andlandscapes 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 theemergence and the decay of systems of intensive farming seem to be general traits in thehistory of agriculture throughout the world Farming systems do not evolve from simpleto complex or from extensive to intensive according to some pre-set model but are formed and changed within specific place-bound social historical and ecologicalcontexts If we accept the idea of uneven development it is also much easier tounderstand why the intensity of agriculture is not evenly or directly related either tomarkets or to natural conditions The islands of agrarian intensity have their own logic ofdevelopment and simplistic explanatory models cannot reflect their distribution or theirdevelopment 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 Haringkansson contrasted the Boserupian explanation of agricultural intensification (intensification as a response topopulation pressure Boserup 1965) with two other models both of which were based onthe idea that intensification could be more broadly understood as an effect of pressure onproduction rather than population pressure (Haringkansson 1989) He argued that locally-developed systems of intensive farming were likely to be the outcome of one or both ofthe following sets of processes first political economy and the development ofhierarchies and second commercial development and increasing market production Thepolitical-economy model as Haringkansson termed it was based mainly on research carried out in central America and southeastern Asia In both regions competitive feasting andredistribution between chiefs created a need for agrarian surpluses As has been shown inmany other studies the development and decay of such hierarchies are very dynamicprocesses and could indeed account for the uneven development and the uneven locationpattern of islands of intensification Furthermore Haringkansson argued tribute labour controlled by chiefs and kings can be seen as one of the ways of mobilizing the labourneeded for the large investments in land connected with this agriculture in order toconstruct features such as irrigation furrows and stone terraces


However these models of social systems do not fit very well with intensiveagricultural systems in the context of eastern Africa and the evidence that Haringkansson cites from Africa is a single case study In Marakwet with its large and thriving irrigationsystem the mobilization of labour and the surveillance of the irrigation system are basedon the decentralized power of the elders and on negotiations rather than on chieflyauthority and tribute labour As far as I can gather the same holds true for Mama Issaraand Sonjo In these systems labour land rights and water rights are embedded in a clan-and 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 farmingintensity 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 toShipton is usually associated not with a centralized control of land but rather withlineage- 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 theirdefinite shares of the land He argues that a more patchy system of fields is usuallyassociated with chiefly control of land in less intensive farming systems which is amodel more in accordance with the Tswana system discussed in the beginning of thischapter than with the intensive systems we know on the ground in eastern AfricaShiptonrsquos 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 Haringkansson also seem to be short of explanatory power in relation to the systems that we have been studying At leasttoday many of the areas with intensive farming are poor and located far away frommarkets In the case of the Iraqw in Tanzania one can even observe an invertedrelationship 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 productionwhile 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 relyingsolely on subsistence production as Loiske (1999) has shown in the case of the Iraqwunder the surface there is in fact a considerable amount of exchange of agriculturalproducts 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 bothhierarchies 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 marketorientation and subsistence farming can be connected with labour-intensive farming The common denominator between these different situations however is that there is ageographical division of labour the islands do not exist in isolation but are based onproduction and resource utilization from a range of different economic zones based ondifferent climates andor different production systems The exchange of products betweendifferent 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 sameethnic group In other cases of which Marakwet is an example exchange betweenagriculturists and pastoralists of different ethnical backgrounds may form an importantincentive for specialization This was also the case in Baringo (Anderson 1988 1989)


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 waysof mobilizing labour The empirical material shows that labour mobilization need not beassociated only with tribute labour and the social division of labour between kings andcommoners but can also be organized according to age sets andor labour exchangewithin 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 ofislands of intensive cultivation and high productivity First they all form part of a widergeographical division of labour but that can take different forms being based variouslyon commercial development on exchange within the ethnic group along kinshipnetworks or on exchange between agriculturists and pastoralists of different ethnicgroups Second mobilization of labour is indeed an integral part of intensive farmingProjects such as terraces and furrows need investments and repairs and with an increasednumber of crops per year preparing the land sowing and harvesting also becomepotential bottlenecks Our case studies show that traditional systems of labour exchangeandor work based on age sets can provide such an input of work Thus large systems ofirrigation and field terracing do not necessarily indicate a hierarchical chiefdom structureFinally 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 landand at the same time the flexibility to cater for fluctuations in climate as well as socialand political changes This flexible system of land and water rights is furthermoreclosely connected with the mechanisms for reproducing social organization andmobilizing labour

ACKNOWLEDGEMENTS

This chapter is a preliminary report on a project financed by the Swedish InternationalDevelopment Authority (SIDA) and the Swedish Council for the Planning andCoordination of Research (FRN) with links with the Institute of Resource Assessment inDar es Salaam the British Institute in Eastern Africa and the National Museums ofKenya I have drawn heavily on informal and productive discussions during a workshopin the field in Marakwet and at the British Institute in Eastern Africa Nairobi in October1998 The participants in the field were Andrew Cheptum Johnstone Kassagam (both atthe National Museums of Kenya) Naomi Mason John Sutton (British Institute in EasternAfrica) Elisabeth Watson (University of Cambridge) and the Swedish research teamLowe Boumlrjeson Vesa-Matti Loiske and Wilhelm Oumlstberg Bill Adams has commented onan earlier version All are warmly thanked for their contributions

REFERENCES

Adams W Potkanski T and Sutton JEG (1994) Indigenous farmer-managed


irrigation in Sonjo Tanzania Geographical Journal 160 (1)17ndash32 Adams W Watson EE and Mutiso SK (1997) Water rules and gender water rights

in an indigenous irrigation system Marakwet Kenya Development and Change28707ndash30

Anderson D (1988) Cultivating pastoralists ecology and economy among the II Chamusof Baringo 1840ndash1980 In DJohnson and DAnderson (eds) The Ecology of Survival Case Studies from Northeast African History 241ndash60 London Lester Crook

Anderson D (1989) Agriculture and irrigation technology at Lake Baringo Azania2489ndash97

Bebbington A (1997) Social capital and rural intensification local organisations andislands of sustainability in the rural Andes Geographical Journal 163189ndash97

Boumlrjeson L (1998) Landscape Land Use and Land Tenure in Mama Issara TanzaniaMapping a lsquoTraditionalrsquo Intensive Farming System Uppsala Swedish University of Agricultural Sciences Minor Field Study No 47

Boumlrjeson L (1999) Listening to the land the Iraqw intensive farming system as told by ahill and its inhabitants In MWidgren and JEGSutton (eds) lsquoIslandsrsquo of Intensive Agriculture in the East African Rift and Highlands A 500-Year Perspective 56ndash73 Stockholm Stockholm University Department of Geography Working Paper from theEnvironment and Development Studies Unit 43

Boserup E (1965) The Conditions of Agricultural Growth London Allen amp Unwin Brookfield H (1986) Intensification intensified Archaeology in Oceania 31177ndash80 Fussel LK (1992) Semi-arid cereal and grazing systems of West Africa In CJ Pearson

(ed) Field Crop Ecosystems 485ndash518 Amsterdam Elsevier Grove AT and Sutton JEG (1989) Agricultural terracing south of the Sahara Azania

24113ndash22 Haringkansson T (1989) Social and political aspects of intensive agriculture in East Africa

some models from cultural anthropology Azania 2412ndash20 Huffman T (1986) Archaeological evidence and conventional explanations of Southern

Bantu settlement patterns Africa 56 (3)280ndash98 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 FinnishHistorical Society in cooperation with Scandinavian Institute of African Studies

Loiske V-M (1993) Mama Isara A Sustainable Agricultural System in Mbulu DistrictTanzania Stockholm Stockholm University Department of Geography WorkingPaper 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 GeographyMeddelanden series B 94

Loiske V-M (1999) Persistent peasants The case of the Iraqw in central Tanzania In MWidgren and JEGSutton (eds) lsquoIslandsrsquo of Intensive Agriculture in the East African Rift and Highlands A 500-Year Perspective 44ndash53 Stockholm Stockholm University Department of Geography Working Paper from the Environment andDevelopment Studies Unit 43


Maggs T (1995) From Marateng to Marakwet Islands of agricultural intensification inEastern and Southern Africa Paper presented at the Prehistoric African AssociationCongress Harare


Oumlstberg W (1999) The origins and expansion of Marakwet hill-furrow irrigation in the Kerio Valley Kenya an interpretation In MWidgren and JEGSutton (eds) lsquoIslandsrsquo of Intensive Agriculture in the East African Rift and Highlands A 500-Year Perspective 15ndash43 Stockholm Stockholm University Department of Geography Working Paper from the Environment and Development Studies Unit 43

Potkanski T and Adams WM (1998) Water scarcity property regimes and irrigationmanagement in Sonjo Tanzania Journal of Development Studies 14 86ndash116

Sansom B (1974) Traditional economic systems In WDHammond-Tooke (ed) The Bantu-Speaking Peoples of Southern Africa 135ndash76 London Routledge amp Kegan Paul

Shipton PM (1984a) Strips and patches a demographic dimension in some Africanland-holding and political systems Man 616ndash20

Shipton PM (1984b) Lineage and locality as antithetical principles in East Africansystems of land tenure Ethnology 23117ndash32

Sutton JEG (1984) Irrigation and soil conservation in African agricultural history witha reconsideration of the Inyanga terracing (Zimbabwe) and Engaruka irrigation works(Tanzania) Journal of African History 2525ndash41

Sutton JEG (1985) Irrigation and terracing in African agricultural historyintensification specialisation or overspecialisation In ISFarrington (ed) Prehistoric Intensive Agriculture in the Tropics 737ndash64 Oxford British Archaeologial ReportsInternational Series 232 Volume 2

Sutton JEG (1989) Towards a history of cultivating the fields Azania 2498ndash122 Sutton JEG (1998) Engaruka an irrigation community in northern Tanzania before the

Maasai Azania 331ndash38 Tengouml M (1999) Integrated Nutrient Management and Farmersrsquo Practises in the Agro-

Ecosystem of Mama Issara Tanzania Stockholm University Department of SystemsEcology 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 MWidgren and JEGSutton (eds) lsquoIslandsrsquo of Intensive Agriculture in the East AfricanRift and Highlands A 500-Year Perspective 74 Stockholm Stockholm UniversityDepartment of Geography Working Paper from the Environment and DevelopmentStudies Unit 43

Watson EE Adams W and Mutiso SK (1998) Indigenous irrigation agriculture anddevelopment Marakwet Kenya Geographical Journal 16467ndash84

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 EMINNIS

INTRODUCTION

Many semi-arid to arid areas are the heartlands of agriculture and the lessons learnedfrom millennia of food production in these often difficult environments can providecritical information for understanding the past Perhaps as importantly we can useknowledge of the astounding range of prehistoric agricultural strategies and theirecological effects to build a more sustainable future especially where food productionexpands into unfamiliar and unfavourable locations Here I outline the types ofagriculture used by the ancient peoples of the region now encompassed by thesouthwestern part of the United States and northwestern Mexico for convenience termedhere the North American Southwest (Fig 151) This region is an excellent location in which to address issues of prehistoric human ecology because it is one of the mostintensely studied dryland regions in the world so we have in some locations surprisingprecision in paleoenvironmental reconstruction and awareness of the regionrsquos prehistory The chapterrsquos 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 diverseregion The two hot deserts of the southern part of the region the Sonoran andChihuahuan are interspersed with isolated mountains and major mountain ranges towerup to nearly 4000 m above sea level The northern part of the region is dominated by theColorado Plateau with cool deserts and semi-arid grasslands Substantial rivers such asthe Gila Colorado and Rio GrandeRio Bravo del Norte are infrequent but they werefoci of prehistoric human occupation Annual rainfall ranges from 127 mm in the lowestdeserts to 700 mm in the mid-level mountains (Sellers and Hill 1974

Figure 151 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 winterstorms and more localized summer monsoons Thus crops often require supplementalwater to yield adequate harvests

Deserts now support grasslands and shrub communities with occasional ribbons ofriparian vegetation (see Brown 1982 for the best summary of the regionrsquos biotic communities for Mexico see also Rzedowski 1986) Low-elevation montane vegetation is dominated by oak pine and juniper woodlands in various combinations with highermontane forests of gymnosperms such as firs spruces and pines Ecology is dynamic andthere is evidence of substantial environmental change including during the historicperiodmdashin fact substantial environmental changes have been noted even within the lastcentury The best documented historic change has been the expansion of desert shrubssuch as mesquite (Prosopis) and montane juniper (Juniperus) at the expense of desert grasslands The best explanation for these changes involves fire suppression drought andintensive livestock grazing (Bahre 1991 Hastings and Turner 1965 Humphrey 1987)

Many millennia of human occupation preceded the use of cultivated plants in theregion (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 2000ndash1000 BC Starting around this time more aggregated populations practising some agriculture appeared in at least two locations around Tucson Arizonaand in northwestern Chihuahua (Hard and Roney 1998) The most important crops such

Prehistoric agriculture and anthropogenic ecology 265

as maize (Zea mays) various pulsesbeans (mostly Phaseolus) and squashes (Cucurbita)originated to the south in Mesoamerica Yet sedentary village agriculture seems not tohave become widespread throughout the region until AD 200ndash700 Occasionally complex regional polities developed the best known examples being Chaco Canyon inNew Mexico Casas Grandes in northwestern Chihuahua and the Hohokam of ArizonaWhile population size degree of aggregation and settlement locations fluctuated throughtime due in part to environmental perturbations agriculture has been the economicmainstay until and after European contact in the late 1500s Prehistoric domesticatedanimals were restricted to the turkey and dog sheep horses cattle and goats werehistoric European introductions

PREHISTORIC AGRICULTURAL STRATEGIES

Prehistoric humans farmed the North American Southwest for millennia and notsurprisingly they developed a wide range of techniques and strategies to grow cropsunder difficult circumstances The most difficult problem they faced was insufficientprecipitation Adding to a large corpus of research on ancient farming in the region aresome excellent ethnographic studies of indigenous farming especially of the Hopi(Bradfield 1971 Hack 1942) and of Sonoran desert peoples (Castetter and Willis 19421952) Not wishing to become bogged down in unnecessary taxonomic complexities Ishall divide agricultural techniques into four simple general categories irrigationfloodwater farming rain-fed farming and rock mulching

Irrigation

Irrigation was widely practised Its origins are earlier than previously thought (Doolittle1990) and the frequency of irrigation agriculture increased through time The largest andmost famous irrigation systems in the region were built by the prehistoric peoples of theSalt and Gila river basins (where the modern city of Phoenix is located) lsquoin terms of complexity it simply had no rival anywhere in Mexicorsquo (Doolittle 199079) Complicated sets of canals totalling over 500 km were constructed (Fig 152) although the destruction of canals by modern agriculture and explosive urban development has obliterated most ofthem (Dart 1989 Fish and Nabhan 1991 Howard 1993) Most other irrigation systemsin the region however seem to have been quite small and were organized at a familiallevel of production (Fish and Fish 1984 Toll 1995)

Floodwater farming

Evidence of ancient systems of floodwater farming is commonly found throughout theregion at locations that are still used by some communities today (Nabhan 1979 1986aNabhan and Sheridan 1977) At times floodwater strategies blend into irrigationsystems and there is no point in trying to make a sharp distinction between the twoUsually 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


(Nabhan 1986b) Again most ancient floodwater systems are rather small lackingevidence 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 153) rock walls across the topographic contour thatcatch 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 (Cordell1977) to the southern part such as around Casas Grandes in Chihuahua (Di Peso 1974Herold 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 farminglocations trincheras are common archaeological featuresmdashwe have recorded hundreds (Minnis and Whalen 1996 Whalen and Minnis 1996) Most are quite small with fieldsaveraging about 2500 m2 the largest of the main group being 8000 m2 but there is one exceptionmdasha series of trincheras that covered at least 100000 m2 Interestingly this field system is next to a site that appears to have been an administrative

Figure 152 Prehistoric Hohokam communities and irrigation systems in the Phoenix basin of the Salt river Map courtesy of Suzanne K Fish


Figure 153 Aerial photograph of prehistoric trincheras (checkdam) fields near Casas Grandes Chihuahua Mexico

Photograph AHeisey

ritual centre within the Casas Grandes polity Its exceptionally large agricultural systemmay 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 optimalconditions but it is difficult to detect prehistoric dryland farming unless soil is modifiedsufficiently to leave archaeological remains Non-irrigated gridded gardens (Fig 154)mdashsmall plots marked by checkerboards of low stone wallsmdashare one such modification and have been found in many areas such as in southeastern Arizona (Gilman and Sherman1975) and northern New Mexico (Ford 2000 Maxwell and Anschuetz 1992) Directrain-fed agriculture is risky farming in the light of the regionrsquos 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 thebeginning of the twentieth century suffered a failure rate of one out of four years (Statenet al 1939) It is likely that the successes and failures of rain-fed farming were especially important in prehistoric cultural dynamics


Figure 154 Gridded gardens of fields outlined by low rock walls near Safford Arizona

Photograph PMinnis

Rock mulching

Rock mulching involves planting crops in piles of stones or covering the fieldrsquos surface with stones and is used worldwide (Lightfoot 1996) The rock mulch conserves moistureand can have other benefits such as protecting roots from rodent predation Like the otheragricultural types mentioned here rock mulching is found in many areas of the regionExamples are known in the north near Santa Fe New Mexico (Anschuetz 1995 Ford2000 Lightfoot 1996 Maxwell 1995 Maxwell and Anschuetz 1992) but they are bestknown from the Sonoran desert of central Arizona where Suzanne and Paul Fish andtheir 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 50000 such piles are present in the foothills north of Tucson indicative of thesubstantial cultivation of a plant previously thought to have been gathered only fromnaturally 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 155 and 156) As in the case of other forms of agriculture in the region that have been studiedproduction seems to have been small scale each field consisted of a little less that 100stone piles


Figure 155 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 PMinnis

Figure 156 An excavated rock pile from the field shown in Figure 155

Photograph PMinnis


Despite being concerned with a small area of the world a century of intensivearchaeological research combined with an excellent ethnographic record have led to thedocumentation of tremendous diversity in agro-ecological strategies The research suggests that prehistoric people may well have been able to farm much of the regionexcept for higher elevations and the most desolate desert plains The sophisticated suiteof agricultural techniques allowed people to farm a wide range of locations Yields ofirrigated flood plain with permanent water and fertile soils would not surprisingly havebeen the economic foundation for communities with the highest population densities butelsewhere other techniques seem to have overcome low and erratic precipitation andoccasionally poor soil fertility

ANTHROPOGENIC EFFECTS OF FARMING

All humans affect their natural environments Despite the many claims to the contrarythis is as true for indigenous North Americans as for peoples elsewhere (eg Denevan1992 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 asParryrsquos agave (Agave parryi) (Minnis and Plog 1976) pruning the Douglas fir(Pseudotsuga menziesii) to yield beams at Mesa Verde (Nichols and Smith 1965) and themanipulation of squawbush (Rhus trilobata) to produce unusually elongated stems forbasketry (Bohrer 1983)

Fire is one of the most widely documented ethnographic examples in North America of anthropogenic ecology (eg Denevan 1992 Dobyns 1981 Krech 1999 Mills 1986) Ithas been presumed that the suppression of both naturally and humanly set fires was amajor factor leading to the modern invasion of shrubs into desert grasslands (Hastingsand Turner 1965 Humphrey 1987) While I suspect that this model is correct and thatprehistoric peoples did in fact set fires for a variety of reasons the evidence of burningin the archaeological record is modest Bohrer (1992) discusses small-scale burning by prehistoric Hohokam in the Sonoran desert Except for fire most effects of anthropogenicecology in the prehistory of the region appear to have been very limited By its verynature however agriculture alters environments and such alterations have the potentialto affect ecological patterning widely Three potential ecological consequences offarming are briefly outlined here deforestation an increase in weeds and soilmodification

Deforestation

Humans use wood and often lots of it for fuel and for their material culture In additionwoodland agriculturalists remove tree cover for fields While deforestation in prehistoryseems not to have been as severe an ecological problem in the region as in some areastoday (such as in Nepal for example) there are some documented cases here ofwoodland reduction by prehistoric peoples Wyckoff (1977) as an early case noted asignificant increase in arboreal pollen particularly pine juniper and oak (Quercus)following the abandonment of Mesa Verde in southwestern Colorado by prehistoric


peoples This change he suggested was best explained as woodland recovery oncehuman 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 PopulusSalix) during the Classic Mimbres period (AD 1000ndash1150) 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 denseprehistoric occupations This small valley had a limited floodplainmdashthe location for the most productive and reliable farmingmdashand estimates of field requirements for varioustime periods indicate that the population of the Classic Mimbres expanded beyond theability of the floodplain to support it The increased presence of small villages andagricultural features in upland secondary farming locations at this time is consistent withthese estimates Therefore it seems that the riparian trees were removed for fieldclearance 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 pintildeon pine (Pinus edulis)wood from packrat middens during the height of the human population in the ChacoCanyon area of northwestern New Mexico He interprets this pattern as decimation oflocal woodlands through human wood harvest unlike in the previous cases he arguesthat there was no documented recovery of pintildeon after the human abandonment of the region Hall (1985) reviewed the pollen records from Chaco Canyon suggesting that theChacoan area of northwestern New Mexico was shrub and grasslands with only scatteredlow-density pintildeon and juniper populations which were species already growing in suboptimal conditions While humans may well have reduced the woody plants theseconifers were not major components of the vegetation Furthermore Hall sees a slightincrease in pine pollen after the prehistoric abandonment of the region While furtherresearch is needed better to understand the human ecology of the Chaco Canyon areaboth 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 lsquoweedyrsquo species Agriculture can increase the abundance of these plants in two ways First soilpreparation in fields often presents ideal settings for such plants Second and lessdirectly as long as agriculture encourages sedentism more soil will be disturbed by dailyactivities 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 regionThese 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 fromthe Sonoran desert (eg Gasser and Kwiatkowski 1991) In fact as Ford (1981) andothers have pointed out these weed seeds can constitute an important and welcomegarden resource for human consumption


Soil modification

Soil modification is the third anthropogenic effect of agriculture considered here One ofthe best known deleterious effects of agriculture in arid areas is salinization a topicdiscussed 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 largeirrigation systems of the hottest desert near Phoenix Arizona where there were intensivecrop production and very high evaporation rates There has been some speculation thatsporadic fields in the Sonoran desert were affected by salinization largely based onhistoric records of such problems in a few locations (Susan Fish pers comm) However there is no compelling archaeological evidence that salinization was an importantcontributing factor to the abandonment of these systems (Krech 1999)

There is in contrast evidence of smaller-scale soil modification due to agricultureScholars working in the Dolores area of southwestern Colorado have used settlementlocations to argue that dryland farming was especially important here for determiningpopulation and settlement dynamics through time even though soil modifications havenot been observed (eg Kohler et al in press Van West 1994) Sandor (1990 1992 1995) found that soils behind trincheras in southwestern New Mexico still seem to showthe 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 andfluctuations as important considerations in understanding the historical dynamics ofancient groups there is sufficient research in this region to demonstrate such linkages Asexpected in a semi-arid to arid area variations in precipitation seem to have had the mostprofound effects on prehistoric farmers (eg Dean et al 1985 Euler et al 1979 Gumerman 1988 Minnis 1985 Petersen 1988 Tainter and Tainter 1996 Van West1994)

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 200through to AD 1150 with a dramatic population peak during the Classic Mimbres period (AD 1000ndash1150) 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 thefloodplain fields Consistent with this argument is the fact that there was an increasedoccupation of upland settlement during the Classic Mimbres period The increasedreliance on secondary field locations worked for a time because (according todendroclimatological records) the first part of the Classic Mimbres period enjoyed anunusually favourable climatic regime During the latter part of the Classic Mimbresperiod the climate returned to a historically more typical pattern so that populationsdependent on upland farming had serious problems provisioning themselves Theseproblems were exacerbated by the fact that the society seems to have been characterized


by intensified local economic inter-relationships More intense interdependence increasedthe social political and economic impacts of the deterioration in the farming system Thismay well have then reverberated throughout the population and contributed to thecollapse of their regional system around AD 1130ndash1150

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 exampleargue that too much rather than too little water caused the destruction of the canalsaround Phoenix by flooding in the mid-1400s and that this may have been a significantcontributor to the collapse of the political structure Numerous scholars have also notedthe relationship between the organization of the irrigation systems and the socio-political landscape those who controlled flow presumably had some power or at least advantageover downstream villages This was certainly the case for the Hohokam (Gumerman1991) 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 moveover the landscape was the critical characteristic that allowed them to farm and survivefor centuries in environments not especially benign for plant cultivation This mobilitywould have been a if not the critical means by which the prehistoric peoples dealt withchanging agricultural conditions and the anthropogenic effects of their activities

CONCLUDING THOUGHTS

Understanding the prehistoric human ecology of the North American Southwest hasvaluable lessons It is obvious that there is a great diversity of prehistoric agriculturalstrategies in the prehistoric record adapted to a wide range of environmental conditionsWhat is less obvious is how these data might be of practical use in the area whereindustrial-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 themodern context although the principles underlying traditional agriculture may be usefulOne could conceive for example of how rock mulchingmdasha relatively environmentally-benign activitymdashmight be used in modern arid-land farming More likely indigenous farming strategies may well find some use in household gardening or lsquoboutique farmingrsquo even in densely urban settings within the region where mechanization is essential Andof course the techniques practised by the indigenous prehistoric farmers of the regionmight 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 inpreviously 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 beenfor the indigenous peoples and for the local ecology of the timemdashand no doubt there were serious problems on occasionmdashno lasting ecological alterations occurred I say this with the caveat that more study of desert grassland fire frequency and of its causeswould be useful Therefore modern environmental planners in the region will be servedbetter by studies of possible small-scale anthropogenic ecology rather than of


widespread general changes due to prehistoric humans like politics anthropogenicecology 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 theecologically-harmful effects of livestock occurred among indigenous peoples once theyacquired exotic domestic livestock From the wider geographic focus on North AmericaKrech (1999) argues that the lsquoIndian-as-ecologistrsquo image is misleading and unjustified which is a point also made by Denevan (1992) I agree but suggest that this misses themost important point Whether one characterizes Native Americans as preservationistsconservationists or ecologists this is less important than understanding how theyinteracted with their environment including understanding how they farmed There arereal lessons to be learned the evidence for less substantial ecological consequences inprehistory compared with today is due to relatively low population density theinfrequency of stratified societies with economies geared toward substantial surplusproduction and the rather high level of residential relocation in prehistory In short fewpeople staying in locations for relatively short periods of time with a familial mode ofproduction simply did not impact the environment as much as historical populations withrelatively high population densities (rural as well as urban) industrial developmentlarge-scale mechanized agriculture exotic species introduced from elsewhere andeffective fire suppression Human ecology is a matter neither of mystic and romanticideology nor simply of indigenous cosmology it must be grounded in an understandingof historical ecology and biology

ACKNOWLEDGEMENTS

I would like to acknowledge the assistance and support of Michael Whalen of theUniversity 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 forcommenting on previous drafts of this text

REFERENCES

Anschuetz KF (1995) Saving a rainy day the integration of diverse agriculturetechnologies to harvest and conserve water in the Lower Chama Valley New MexicoIn WTool (ed) Soil Water Biology and Belief in Prehistoric and Traditional Southwestern Agriculture 25ndash40 Albuquerque New Mexico Archaeological Council Special Publication 2

Bahre CJ (1991) A Legacy of Change Historic Human Impact on Vegetation of the Arizona Borderlands Tucson University of Arizona Press

Betancourt JL (1990) Late Quaternary biogeography of the Colorado Plateaus InJBetancourt TVan Devender and PMartin (eds) Packrat Middens The Last 40000 Years of Biotic Change 259ndash93 Tucson University of Arizona Press

Betancourt JL Van Devender TR and Martin PS (1990) Packrat Middens The Last


40000 Years of Biotic Change Tucson University of Arizona Press Bohrer VL (1983) New life from ashes the tale of the burnt bush (Rhus trilobata)

Desert Plants 5122ndash4 Bohrer VL (1992) New life from ashes II Desert Plants 10122ndash5 Bradfield M (1971) The Changing Pattern of Hopi Agriculture London Royal

Anthropological Institute of Great Britain and Ireland Occasional Papers 30 Brown DE (1982) Biotic communities of the American Southwest United States and

Mexico Desert Plants 41ndash342 Castetter EF and Willis HB (1942) Pima and Papago Indian Agriculture

Albuquerque University of New Mexico Press Castetter EF and Willis HB (1952) Yuman Indian Agriculture Albuquerque

University of New Mexico Press Cordell LS (1977) Predicting site abandonment at Wetherill Mesa The Kiva 40 189ndash

202 Cordell LS (1997) Prehistory of the Southwest San Diego Academic Press Cordell LS (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 89ndash1 Dean JS Euler RC Gumerman GJ Plog F Hevly RH and Karlstrom TNV

(1985) Human behavior demography and paleoenvironment on the ColoradoPlateaus American Antiquity 50537ndash54

Denevan WM (1992) The pristine myth the landscape of the Americas in 1942 Annals of the Association of American Geographers 82369ndash85

Di Peso CC (1974) Casas Grandes A Fallen Trading Center of the Gran ChichimecaFlagstaff Northland Press

Dobyns HF (1981) From Fire to Flood Historic Human Destruction of the Sonoran Desert Riverine Oases Socorro Ballena Press Anthropological Papers 20

Donkin RA (1979) Agricultural Terracing in the Aboriginal New World New York Viking Fund Publications in Anthropology 56

Doolittle WE (1990) Canal Irrigation in Prehistoric Mexico The Sequence ofTechnological Change Austin University of Texas Press

Euler RC Gumerman GJ Karlstrom TNV Dean JS and Hevly RH (1979) TheColorado Plateaus cultural dynamics and paleoenvironments Science 205 1089ndash101

Fish SK and Fish PR (1984) Prehistoric Agricultural Strategies in the SouthwestTempe Arizona State University Anthropological Research Reports 20

Fish SK and Nabhan GP (1991) Desert as context the Hohokam environment InGGumerman (ed) Exploring the Hohokam Prehistoric Desert Peoples of theAmerican Southwest 29ndash60 Albuquerque University of New Mexico Press

Fish SK Fish PR Miksicek C and Madsen J (1985) Prehistoric Agave cultivationin southern Arizona Desert Plants 7107ndash12

Ford RI (1981) Gardening and farming before AD 1000 patterns of prehistoriccultivation north of Mexico Journal of Ethnobiology 16ndash27

Ford RI (2000) Human disturbance and biodiversity diversity a case study fromnorthern New Mexico In PMinnis and WElisens (eds) Biodiversity and Native


America Norman University of Oklahoma Press (in press) Gasser RE and Kwiatkowski SM (1991) Food for thought recognizing patterns in

Hohokam subsistence In GGumerman (ed) Exploring the Hohokam PrehistoricDesert Peoples of the American Southwest 417ndash59 Albuquerque University of New Mexico Press

Gilman PA and Sherman S (1975) An Archaeological Survey of the Graham-Curtin Project Phase II Tucson Arizona State Museum Cultural Resource ManagementSection Report

Graybill DA and Nials FL (1989) Aspects of climate streamflow and geomorphologyaffecting irrigation systems in the Salt River valley In CAHeatherington andDAGregory (eds) The 1982ndash1984 Excavations at Las Colinas Environment andSubsistence 39ndash58 Tucson Arizona State Museum Archaeological Series No 162

Gumerman GJ (1988) The Anasazi in a Changing Environment Cambridge Cambridge University Press

Gumerman GJ (1991) Exploring the Hohokam Prehistoric Desert Peoples of theAmerican Southwest Albuquerque University of New Mexico Press

Hack JT (1942) The Changing Physical Environment of the Hopi Indians of ArizonaCambridge Mass Harvard University Peabody Museum Papers 31 (1)

Hall SA (1985) Quaternary pollen analysis and vegetational history of the SouthwestIn VBryant and RHolloway (eds) Pollen Record of Late-Quaternary North American Sediments 95ndash123 Dallas American Association for Stratigraphic Palynologists

Hard RJ and Roney JR (1998) A massive terraced village complex in ChihuahuaMexico 3000 years before present Science 2791661ndash4

Hastings JR and Turner RM (1965) The Changing Mile An Ecological Study ofVegetation Change with Time in the Lower Mile of an Arid and Semiarid RegionTucson University of Arizona Press

Herold LC (1970) Trincheras and Physical Environment Along the Rio Gavilan Chihuahua Mexico Denver University of Denver Department of GeographyTechnical Paper No 65ndash1

Howard J (1993) A Paleohydraulic Approach to Examining Agricultural Intensification in Hohokam Irrigation Systems Greenwich JAI Press Research in EconomicAnthropology Supplement 7

Howard WA and Griffiths TM (1966) Trinchera Distribution in the Sierra Madre Occidental Mexico Denver University of Denver Department of Geography Technical Paper No 66ndash1

Humphrey RR (1987) 90 Years and 535 Miles Vegetation Change Along the MexicanBorder Albuquerque University of New Mexico Press

Kohler TA Kresl J Van West C Carr E and Wilshusen RH (in press) Be therethen a modeling approach to settlement determinants and spatial efficiency among lateancestral Pueblo populations of the Mesa Verde region US Southwest In TKohlerand GGumerman (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 WWNorton Lekson SH (1999) The Chaco Meridian Centers of Political Power in the Ancient

Southwest Walnut Creek Altamira Press


Lightfoot DR (1996) The nature history and distribution of lithic mulch agriculture anancient technique of dryland agriculture Agricultural History Review 44 206ndash22

Maxwell TD (1995) A comparative study of prehistoric farming strategies In H Toll(ed) Soil Water Biology and Belief in Prehistoric and Traditional SouthwesternAgriculture 3ndash10 Albuquerque New Mexico Archaeological Council SpecialPublication 2

Maxwell TD and Anschuetz KF (1992) The southwestern ethnographic record andprehistoric agricultural diversity In TKillion (ed) Gardens of Prehistory The Archaeology of Settlement Agriculture in Greater Mesoamerica 35ndash68 Tuscaloosa University of Alabama Press

Mills BJ (1986) Prescribed burning and hunter-gatherer subsistence systems Haliksarsquoi UNM Contributions to Anthropology 51ndash26

Minnis PE (1985) Social Adaptation to Food Stress A Prehistoric SouthwesternExample Chicago University of Chicago Press

Minnis PE (1989) Prehistoric diet in the northern Southwest macroplant remains fromFour Corners feces American Antiquity 54543ndash63

Minnis PE and Elisens WJ (2000) Biodiversity and Native America Norman University of Oklahoma Press (in press)

Minnis PE and Plog SE (1976) A study of the site specific distribution of Agave Parryi in east central Arizona The Kiva 41299ndash308

Minnis PE and Whalen ME (1996) Prehistoric Upland Agriculture in the Casas Grandes Core Washington DC final project report submitted to the National Geographic Society

Nabhan GP (1979) The ecology of floodwater farming in the arid southwestern NorthAmerica Agro-Ecosystems 5245ndash55

Nabhan GP (1986a) Papago Indian desert agriculture and water control in the Sonorandesert 1697ndash1934 Applied Geography 6 (1)42ndash3

Nabhan GP (1986b) lsquoAk-cintildersquo ldquoarroyo-mouthrdquo and the environmental setting of Papago Indian fields in the Sonoran desert Applied Geography 6 (1)61ndash75

Nabhan GP and Sheridan TE (1977) Living fencerows of the Rio San MiguelSonora Mexico traditional technology of floodplain management Human Ecology597ndash111

Nichols RF and Smith DG (1965) Evidence of prehistoric cultivation of Douglas-Fir trees at Mesa Verde American Antiquity Memoir 31 (2)57ndash64

Petersen KL (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 amp Hudson Rzedowski J (1986) Vegetation de Mexico Mexico DF Editorial Limusa Sandor JA (1990) Prehistoric agricultural terraces and soils in the Mimbres area New

Mexico World Archaeology 22166ndash80 Sandor JA (1992) Long-term effects of prehistoric agriculture on soils examples of

New Mexico and Peru In VHolliday (ed) Soils in Archaeology Landscape Evolution and Human Occupation 217ndash45 Washington DC Smithsonian Institution Press

Sandor JA (1995) Searching soil for clues about Southwestern prehistoric agricultureIn HW Toll (ed) Soil Water Biology and Belief in Prehistoric and Traditional


Southwestern Agriculture 119ndash37 Albuquerque New Mexico Archaeological Council Special Publication 2

Schmidt RH Jr and Gerald RE (1988) The distribution of conservation type water-control systems in the northern Sierra Madre Occidental The Kiva 53 165ndash79

Sellers WD and Hill RH (1974) Arizona Climate Tucson University of Arizona Press

Staten G Burnham DR and Carter J Jr (1939) Corn Investigations in New MexicoLas Cruces New Mexico State University Agricultural Experiment Station Bulletin260

Tainter JA and Tainter BB (1996) Evolving Complexity and Environmental Risk inthe Prehistoric Southwest Reading Addison-Wesley Publishing

Toll HW (1995) Soil Water Biology and Belief in Prehistoric and Traditional Southwestern Agriculture Albuquerque New Mexico Archaeological Council SpecialPublications 2

Tuan Y-F Everard CE Widdison JG and Bennett I (1973) The Climate of New Mexico Santa Fe State Planning Office

Van West CR (1994) Modeling Prehistoric Agricultural Productivity in SouthwesternColorado A CIS Approach Pullman Washington State University Department of Anthropology Reports of Investigations 67

Vivian RG (1991) Chacoan subsistence In PJCrown and WJJudge Chaco and Hohokam Prehistoric Regional Systems in the American Southwest 57ndash76 Santa Fe School of American Research Press

Whalen ME and Minnis PE (1996) El Sistema Regional de Paquimeacute Chihuahua Mexico Mexico DF Informe Teacutechnio Final presented to the Consejo de Arqueologiacutea Institute Nacional de Antropologiacutea e Historia

Woodbury RB (1961) Prehistoric Agriculture at Point of Pines Arizona Salt Lake City Memoirs of the Society for American Archaeology 26 (3) part 2

Wyckoff DG (1977) Secondary forest succession following abandonment of MesaVerde The Kiva 42215ndash32


16 The role of maguey in the Mesoamerican tierra friacutea ethnographic historic and archaeological

perspectives JEFFREY RPARSONS AND JANDREW DARLING

INTRODUCTION

Pre-columbian civilizations in Mesoamerica flourished in three very differentenvironments the tierra calientemdashwarm humid and thickly forested lowlands below 1000 m above sea level along the Atlantic Caribbean and Pacific coasts of Mexico andadjacent Central America the tierra templadamdashsubhumid to semi-arid frost-free temperate highlands between 1000 and 1800 m in Guatemala and southern Mexico andthe tierra friacuteamdashsemi-arid and arid highlands with average annual rainfall as low as 300mm and with severe winter frosts at elevations above 1800 m in central and north-central Mexico (Fig 161)

Mesoamerica was one of the worldrsquos hearths of early plant domestication and agriculture provided the economic basis of the chiefdoms states and empires thatdeveloped there after c1500 BC (Flannery 1973 MacNeish 1991) Yet it was the onlyone of the worldrsquos ancient primary civilizations that lacked a domestic herbivore Through the use of domestic camelids (llamas and alpacas) in the central Andes andsheep and goats in much of the Old World food producers in virtually all other regionswhere ancient states and empires existed were able significantly to extend theirproductive landscapes into drier and colder zones and over a full annual cycle Some ofthem 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 limitedcapacity to generate and manipulate energy could have attained a level of organizationalcomplexity on a par with that of the central Andes and several Old World regions whereagriculture and pastoralism were combined in antiquity This question becomesincreasingly important because the largest polities of ancient MesoamericamdashTeotihuacan between AD 200 and 600 Tula between AD 900 and 1200 and the Aztecs ofTenochtitlan and their neighbours between AD 1300 and 1520mdashwere all centred in the comparatively cold and dry tierra friacutea We are particularly

Figure 161 Middle America showing the approximate extent of the tierra friacutea (shaded)

interested in understanding how the resources of the tierra friacutea underwrote the development of Mesoamericarsquos 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 cometo be known collectively as lsquomagueyrsquo in the Mexican highlands The most important maguey species include Agave salmiana A magisapa Aatroviens Aferox Ahookeriand Aamericana (Gentry 1982) Cultivated maguey is still an important component ofagriculture throughout the Mexican tierra friacutea today (Fig 162) and it is known to have played a significant role in the economy of this region for thousands of years (Parsonsand Parsons 1990) Other species of agave are cultivated in other parts of Mesoamericabut these are invariably of secondary importance

Most archaeologists working in Mesoamerica have overlooked the full significance ofmaguey With their interests dominated by the cultivation of annual seed crops (primarilymaize beans amaranth and squash) archaeologists have tended to ignore or downplaythe potential of other types of food production (cf Mangelsdorf et al 1964 Puleston 1968 1973 Willey et al

The role of maguey in the Mesoamerican tierra friacutea 281

Figure 162 Field of cultivated magueys all the sap has been extracted from the plant in foreground

Photograph JParsons

1964) An extreme expression of this viewpoint is the assertion by Blanton et al(1981174) that lsquoin the highland valleys [of Mesoamerica] the surest way of producing alarge surplus was to plant maize everywherersquo By contrast we argue that maguey and seed crops were fully complementary in the Mesoamerican tierra friacutea and that maguey there made available some of the same kinds of coping strategies complementary to seedcultivation as did llamas and alpacas in the central Andes and sheep and goats inMesopotamia and elsewhere in the Old World It is important to note that as sheepgoats pigs and cattle became increasingly important as introduced sources of food andfibre in highland Mexico after European contact in the early sixteenth century (Crosby1972 1986) so too there was a corresponding decline in the importance of magueymdasha 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

bull in the Mesoamerican tierra friacutea the development of complex society during the Middle and Late Formative (Table 161) 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)


In developing and addressing these ideas we are constrained by serious limitations of theknown archaeological record of prehistoric maguey utilization Few archaeologists haveinvestigated maguey production and many remain unaware of its key archaeological

Table 161 The prehispanic chronology of central Mexico Date Period PhaseAD 1520 Aztec IV Late Postclassic Aztec III AD 1350 Middle Postclassic Aztec IndashII 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

bull 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

bull agricultural production in the drier highland regions of central and especially north-central 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 ie the more extensive piedmonts and plains beyond the reach of effective irrigation

bull 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


correlates Much of north-central Mexico remains archaeologically understudied and soit is difficult to make good inferences on the basis of archaeological data from the regionwhere maguey production was probably most critical in the prehistoric economyConsequently we are able to make very few definitive statements based uponarchaeological remains about the specifics of how maguey was actually used at differenttimes and places in the past This chapter is thus very much an exercise in hypothesisbuilding in which we rely primarily on ethnographic observations and historicaldocumentation and only secondarily on archaeological data

We begin by highlighting magueyrsquos importance as a source of food and fibre in contemporary highland Mexico We use these contemporary data to quantify the potentialneed 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 thearchaeological correlates of prehistoric maguey production We conclude by combiningethnographic historic and archaeological data to develop hypotheses for the role ofmaguey in the classic and Postclassic expansion of Mesoamerican civilization into thetierra friacutea 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 theimportance of maguey as a source of food and fibre for thousands of years in highlandMexico (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 magueyand the use of its sap fibre and flesh Historians have found references to magueycultivation and use in written documents that extend back to the early sixteenth centuryIn their midden excavations and surface surveys archaeologists have found the physicalremains of agave fibre plus spindle whorls used in the spinning of maguey fibre andscraping tools used in the extraction and processing of sap and fibre Nevertheless thetechnology and organization of pre-columbian maguey utilization have remained poorlyunderstood Recent ethnographic research on maguey use in central and north-central Mexico (Parsons and Parsons 1990 Patrick 1985 Rangel 1987 Ruvalcaba 1983Salinas and Bernard 1983 Sanchez 1980) has provided some new insights intoprehistoric maguey utilization In the next few paragraphs we shall briefly highlight someaspects of this research

Maguey sap and flesh

The maguey plant provides a rich store of both sap and edible flesh Maguey sap isacquired for human use by means of procedures that interrupt the final stage of a plantrsquos normal seven to twenty-five year maturation process in order to extract the sap throughan initial lsquocastrationrsquo (a procedure that halts the natural flow of sap to an emergent seed-bearing stalk Fig 163) followed by daily scraping and extraction operations over a period of three to six months Individual plants in cultivated fields typically approach


maturity continuously throughout the year The timing of their planting and replacementis often explicitly managed so as to ensure continuous productivity with no more than 5ndash10 per cent of a fieldrsquos 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 severalhundred litres of sap and a hectare of land typically yields 5000ndash9000 litres of sap per year (Parsons and Parsons 1990338) The sap may be allowed to ferment to form a beer-like 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 areunstable and cannot remain unused for more than about a week As syrup or sugarhowever maguey sap is much more durable and in these forms sap surpluses can bestored 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 plantrsquos leaves and squeezing out the liquid in a single operation (Bye et al 1975) As we shall note below there is reason to thinkthat something analogous to this less-efficient Tarahumara procedure (lsquoless efficientrsquo in the sense that not all the plantrsquos sap can be extracted in this manner and the plantrsquos fibre and flesh are usually not utilized at all) may have characterized maguey sap extractionduring the Formative and Classic periods in central Mexico prior to the


Figure 163 Castrating a mature maguey plant Photograph JParsons

implementation in the Postclassic period of the more efficient techniques observedethnographically 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 TheTarahumara for example prepare cakes of baked maguey flesh which can be stored forup to six months (Bye et al 1975)

Maguey sap and flesh are rich in both nutrients and calories Ruvalcaba (198389) citesanalyses showing that one litre of pulque contains 574 calories Davidson and Ortiz deMontellano (1983155) report that one tablespoon of maguey sap contains (among otherthings) 008 g of protein 535 g of carbohydrates 20 calories 033 mg of Vitamin C002 mg of calcium 503 mg of phosphorous 127 mg of potassium 300 micro-grams of iron 170 micro-grams of magnesium 90 micro-grams of selenium 60 micro-grams of


chromium and 40 micro-grams of zinc In the early 1940s Anderson et al (1946888) 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 percent of total riboflavin 23 per cent of total niacin 48 per cent of total Vitamin C 8 percent of total calcium and 20 per cent per cent of total iron Ross (1944 cited in Fish et al1986) found that 100 g of cooked agave flesh contains 347 calories and 45 g of protein

It appears that in most tierra friacutea contexts maguey can produce approximately asmany calories and essential nutrients per hectare as the standard seed crops and thatwhen the plantrsquos flesh and sap are both consumed maguey can potentially produce morecalories than seed crops on a given unit of land (Parsons and Parsons 1990337 338345) Only on irrigated land are seed crops significantly more productive than magueyCritically though maguey can be interplanted with seed crops in virtually all agriculturalsettings and when this is done (as it commonly has been throughout the historic period intierra friacutea 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 subsistencesecurity for pre-hispanic agriculturalists in the tierra friacutea annual energy productivity on most kinds of cultivated land could have been doubled agricultural productivity couldhave been extended over a full annual cycle agricultural productivity could have beenextended into nearby drier colder and less fertile areas which are marginal for seedcrops and the year-round productivity of maguey could have been combined with thelong-term storability of seed crops Recent ethnographic studies (Parsons and Parsons199031) also reveal that maguey-sap exploitation can easily be deferred to the winter agricultural off-season (because the collection of the matured plantrsquos sap can be postponed for up to six months after the initial castration operation without any apparentloss in productivity) thereby reinforcing the complementarity between maguey and seedcrops

Furthermore because of its resistance to drought frost and hailmdashall common causes of seed-crop failure in the tierra friacuteamdashmaguey stands out as an ideal highland famine foodFor example elderly people living in agriculturally-marginal parts of central Mexico vividly recall that during the most violent years of the Mexican revolution (1913ndash17) when normal access to market produce (including maize and beans) was frequentlydisrupted by military hostilities their families survived for weeks and months at a timeon maguey and nopal products which were readily available at all times of the year intheir own fields (Parsons and Parsons 199011)

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 forsedentary cultivators during times of serious crop scarcity as was commonly done byprehistoric colonists with certain types of introduced wild or semi-wild plants in ancient Polynesia (Kirch 1984131ndash2) The extensive stands of wild maguey and nopal thattoday occur throughout the most marginal parts of the arid highlands in central and north-central Mexico might be relicts of such pre-hispanic practicesmdashthe 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


flesh and fibre can be handled on any level from a nuclear family household up to thelarge commercial plantation (hacienda) employing several hundred workers organizedwithin a hierarchical administration There appear to be no inherent qualities of this plantthat might require or select for either diffuse or centralized organization for its growthcultivation harvest or for the extraction or processing of its products Nevertheless allour ethnographic and historic observations of maguey cultivation are from contextswhere the production of maguey is directly combined with that of seed crops This meansthat we lack historic analogies for fully specialized maguey agriculturalists that is wheremaguey cultivation might have been carried out separately from that of maize beanssquash or amaranth We shall need to remember this point in the concluding section ofthis chapter where we propose that specialized maguey cultivation may have played akey 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 thetierra friacutea 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 1964131) and so there would have been an evengreater 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 rottingso that its hard flesh (which usually comprises more than 97 per cent of the plantrsquos weight) can be easily separated from the encased fibre by scraping When properlymanaged both the sap and fibre of an individual plant can be extracted for human useThis same study also revealed the critical importance of dried maguey stumps as fuel inareas where firewood is scarce or absent Pre-hispanic highland populations living in drysparsely-forested terrain may have been as much interested in the fuel that magueyprovided as they were in the food and fibre that the plant produced De Sahagun (1969volume 3145) for example specifically mentions the sale for fuel of dried magueystumps and leaves in sixteenth-century market-places in Mexico City

Parsons and Parsons (1990157) found that an average maguey leaf provides roughly 75 g of dried fibre An average maguey plant has twenty to thirty leaves and thusprovides approximately 2000 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 Thusone maguey plant provides enough fibre for about 10 m2 of cloth More precise calculations would have to make allowance for variable thread thickness thread spacingfibre quality type of costume and so on Nevertheless these rough estimates suggest thatone maguey plant would have provided enough fibre for outfitting an average pre-columbian person with most of the maguey-fibre textile required for clothing over aperiod of a few years

On an average cultivated hectare of land in highland central Mexico about thirtymaguey plants can be exploited each year for both sap and fibre (Parsons and Parsons


1990336 338) Thus 1 ha of cultivated maguey could potentially outfit approximatelythirty people with the maguey cloth they would need for a few (say three) yearsAlternatively assuming each average person requires one third of hisher wardrobe to bereplaced each year then 1 ha of cultivated maguey would provide the annual maguey-cloth needs for some ninety people We can simplify our calculations by calling it aneven 100 On this basis a million peoplemdashapproximately the number of people living inthe Valley of Mexico (the core region of Aztec civilization in AD 1500)mdashwould annually have required the fibre production (c600000 kg) of the cultivated maguey fromsome 10000 hectares which was roughly 5 per cent of the total arable landscape in theValley of Mexico This same amount of land could potentially at the same time haveproduced annually about 50ndash90 million litres of aguamiel roughly 6000 metric tons ofcooked maguey flesh perhaps 8000ndash10000 metric tons of interplanted maize or beansand many tons of dried maguey stumps for use as household fuel (Parsons and Parsons1990337 338)

Obviously the above figures require extensive refinement For example overallproductivity of maguey and other crops is likely to have been significantly lower than theabove-cited figures which derive from central Mexico in the increasingly more aridterrain 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 considersthese figures and remembers that maguey production (of both sap and fibre) can bedeferred to the agricultural off-season and that household spinning and weaving can alsobe relegated to the winter off-season period then the complementarity of maguey andseed crop cultivation in the tierra friacutea 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 andceramic 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 specificproductive functions

Spinning

Maguey fibre continues to be spun into thread using traditional drop-spinning techniques that employ wooden spindles and ceramic spindle whorls (Fig 164) (Spindle whorls today are also sometimes made of stone bone or wood) We now have some goodarchaeological data on the nature and distribution of pre-columbian ceramic spindle whorls in central Mexico and we can distinguish between small whorls (weighing lessthan c7 gm) used for spinning thinner lighter cotton fibre and large whorls (weighing more than c11 gm) used for spinning thicker heavier maguey fibre (eg Norr 1987 Parsons 1972 Sejourne 1983 Smith and Hirth 1988 Fig 165) Studies of living spinners show that those whorls that weigh 20ndash30 gm can


Figure 164 Spinning maguey fibre showing wooden spindle and ceramic spindle whorl in use

Photograph JParsons


Figure 165 Pre-columbian spindle whorls used for spinning maguey fibre

be used to produce a wide range of fine to coarse maguey thread whereas lighter (c11ndash15 gm) and heavier (c35ndash140 gm) whorls could only have been used to produce respectively a much narrower range of fine or coarse maguey thread (Parsons andParsons 1990329 331)

Consequently we now have a sense about how we might eventually be able to identifygeneralized versus specialized spinners in the archaeological record once the right kindof archaeological information becomes available This prospect becomes especiallyinteresting in the light of historically-based discussions of the organization of spinning and weaving and of the importance of textiles in tribute market exchange ceremonialpresentations 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 spinningworkshops we should be in a much better position to infer the extent to which differentspinners were involved in either specialized or generalized spinning in the production ofeither cotton or maguey thread and in tributary market or domestic modes of productionWe also suspect that the elaborate stamped moulded and incised designs socharacteristic of Postclassic spindle whorls (Fig 165) 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 centralMexico are extremely scarce before the Postclassic period Numerous possibilities might


explain this such as spinning without whorls the use of perishable wooden whorls or theuse of simple perforated sherd disks that are not always recognized as spindle whorls byarchaeologists However this clear-cut difference between Classic and Postclassic spinning technology is so dramatic as to suggest a major reorganization of spinning afterthe Classic period This contrast may signify that in highland central Mexico spinning(and possibly weaving as well) became more specialized and more efficient during thePostclassic than it had been earlier This in turn suggests changes in the organization ofmaguey production and fibre processing in the tierra friacutea

Carrascorsquos (1976) discussion of different kinds of cloth production in commoner households and palace workshops in the early sixteenth century is certainly suggestive inthis regard as is Hicksrsquo (1987) emphasis on the importance of certain kinds of textiles innew forms of market-based redistribution in Late Postclassic times Both Carrasco andHicks have relied exclusively on sixteenth- and seventeenth-century historic sources to develop their ideas about textile production and distribution in the Aztec heartlandArchaeological 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 bothmaguey whorls and cotton whorls co-occur at many Postclassic sites in the Valley ofMexico (tierra friacutea where cotton cannot be grown locally) and in at least one MiddlePostclassic site in the nearby tierra templada (Norr 1987) where maguey has never beenproduced in historic times This co-occurrence of cotton and maguey spinning inecologically lsquoinappropriatersquo zones implies the existence of fairly complex redistributionalnetworks 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 mountedin a wooden handle (Fig 166) 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 167)mdasha 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 haveinterpreted these implements as hoes associated with maize cultivation recentexperimental work shows that these implements are admirably suited for scrapingmaguey fibre (Parsons and Parsons 1990175 Fig 168)

These trapezoidal scrapers are quite widespread throughout the highlands of central and north-central Mexico and in the Southwest United States


Figure 166 Use of modern iron scraper for extracting maguey fibre

Photograph JParsons

(Brumfiel 1976 Cabrero 1989 Fish et al 1986 Mastache et al 1990 Sanders et al1979 Sejourne 1983 Fig 137 Spence 1971 Tesch and Abascal 1974 Trombold1985 1989) Over time they tend to displace another distinctive tool the scraper plane orlsquoturtleback scraperrsquo (Tolstoy 1971 Fig 169) Experimental work with archaeologicalscraper-planes in the southern highlands of Mexico (Hester and Heizer 1972) has shownthat repeated downward blows with the rounded side of this tool (which typically weighsabout 400 gm) are effective to mash up raw maguey leaves while the flat bottom side ofthe same implement can serve to scrape the mashed flesh away from the fibres (using alateral 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 androtting leaves to soften the flesh The scraper plane would probably have predominated inthe context of earlier andor more generalized fibre production where high efficiencywas less important If so then increasing specialization and efficiency of maguey fibreprocessing (manifested archaeologically by a progressive shift from scraper planes totrapezoidal ground-stone scrapers) appear to have paralleled increasing efficiency in thespinning of maguey fibre (manifested archaeologically by a dramatic


Figure 167 Examples of pre-columbian trapezoidal tabular basalt scrapers

increase in quantities and variability of ceramic spindle whorls) during the Postclassictimes

The extraction of maguey sap

There is a very distinctive and highly specialized modern iron tool used for the twice-daily scraping of the surface of the sap-collecting cavity in the maguey plantrsquos interior The pre-hispanic analogue of this elliptical or circular iron scraper appears to be adistinctive plano-convex stone scraper (Fig 1610) This implement has a broad distribution in the highlands of central and north-central Mexico (Cabrero 1989234


238ndash41 Dibble and Anderson 1963 Fig 778 Gamio 1979 [1922] 214 Mastache et al 1990189 Meigham

Figure 168 Experimental use of pre-columbian trapezoidal tabular basalt scraper

Photograph JParsons

Figure 169 A pre-columbian scraper plane (width c12 cm) Source Adapted from Hester and Heizer 1972

1976 Michelet 1984 Parsons and Parsons 1990 Rodriguez 1985199 Sanders 19651966 Sanders et al 1979 Spence 1971 Trombold 1985 1989 Vaillant 1931417)


These studies indicate that this tool appears as early as the Late Formative and increasesmarkedly in frequency by the Postclassic There seems little doubt about its primaryfunction These distinctive scrapers apparently do not occur archaeologically in anysignificant numbers outside

Figure 1610 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 JParsons

the tierra friacutea which lends additional support to our belief that this artefact was usedexclusively 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 andthe sap squeezed out in a single operation to (2) something comparable to thehistorically-known process in central Mexico in which both the sap and fibre of individual plants are extracted through specialized procedures over a period of severalmonths Once again we suggest that this shift was in the interests of greater overallefficiency of plant use in increasingly specialized economies stimulated by both the


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 ancientagriculturally-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 theydeveloped and changed This chapter has addressed a large region from the perspective ofa particular type of agriculture in the context of inadequate archaeological informationOur conclusions are thus necessarily generalized and tentative Testing these hypotheseswill involve the archaeological identification of maguey production and processing andthe comparison of tool kits and midden contents from sites in different parts of the tierra friacutea 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 onregional settlement patterns in north-central Mexico in order to provide information onvariability 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 tenuremdashnone of which can presently be inferred in any satisfactory or credible way Our conclusions are presented below aseight 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 friacutea 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 friacutea with no significant shifts towards greater specialization or efficiency until the development of increasinglycomplex and urbanized society late in the first millennium BC With their increasedoverhead costs and greater spatial separation between food producers and foodconsumers urbanized states from the early first millennium AD onwards would haveneeded 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 specializedmaguey-nopal producers dispersed extensively in agriculturally more marginal landscapes and on the other more generalized seed cropmdashmaguey cultivators living in more nucleated settlements in restricted more productive river valleys where theirrigation of seed crops was feasible The effective integration of these agriculturally-generalized cores and agriculturally-specialized peripheries would have been dependentupon the existence of redistributional networks large enough to move staples oversignificant distances in a regular and predictable manner Because the scope and scale ofpre-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


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 expansionof pastoralists into the dry steppes of Inner Asia after the late second millennium BCSahlins for example once suggested (196833ndash4 following leads by Lattimore 1951 and Krader 1957) that an effective adaptation by specialized pastoralists to the sparselyoccupied grasslands of Inner Asia might not have occurred until there was enoughpressure from expanding states in better-watered and longer-settled regions to the south where generalized neolithic agriculturalists combining cultivation and herding had livedfor 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 pastoralistswho may have moved into the dry Inner Asian steppes after c1500 BC in order to escape the tribute and labour service demands imposed on them by increasingly large andpowerful Near Eastern and East Asian polities Did intensified and more efficientmaguey utilization provide some cultivators living in highland central Mexico during theera of state growth in the early first millennium AD with the means to escape thedemands of their would-be overlords by emigrating to and flourishing in the sparsely occupied drylands to the north Alternatively was the development of greater socio-political 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 greatersocio-political centralization in order more effectively to exploit the opportunities to acquire new wealth and new types of prestige-building exotica that were increasinglyavailable from developing state systems along their southern flanks in central and westernMexico (Barfield [1989] presents an intriguing Old World analogy that extends theearlier 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 ofthis shift might relate to the changing character of urbanism and the dynamics ofurbanizationmdashfor example the development of large centres inhabited predominantly bynon-food producers Another aspect of this change may relate to the development of newstatus roles and the need to distinguish them by implementing new sumptuary rules suchas regulating the production and use of pulque and certain types of clothing involved inpublic ritual performances and displays in which elites played different roles inincreasingly stratified societies Most important of all might have been the changingnature of tribute exchange and governance whereby for example different kinds ofcloth 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 innorth-central Mexico during the Classic period might have been subsequently lsquoimportedrsquo 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 itwas in the context of expansion into this driest northernmost part of Mesoamerica that the


most effective and efficient maguey exploitation first developed Weintraub (1992) forexample reports the presence of maguey fibre and leaf fragments in flotation samplesfrom late Classic-Epiclassic contexts at the northern centre of La Quemadamdashperhaps 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 andspace throughout Mesoamerica We suspect that the cold dry lands of central and north-central Mexico will show an unusually high level of such productive efficiency andspecialization because it was in these regions that ancient Mesoamericans were forced toconfront 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 Mesoamericantierra friacutea 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 theend of the Formative period Some of the changes in the technology of magueyproduction probably reflect the demands of state administrators for greater productiveefficiency and specialization in their domains Shall we discover notably lesstechnological 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 productionof 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 theinitial domestication and accompanying botanical diversification of maguey in the tierra friacutea of central Mexico during the Early and Middle Formative at a time when tribal lsquobig menrsquo and aspiring chiefs in developing ranked societies throughout southern and centralMexico were seeking higher levels of local productivity to sustain and enhance theirprestige

(8) Equally important should be the value to local elites in the tribal and emergent-chiefdom societies of north-central Mexico of prestige-enhancing materials such as decorated ceramics fancy gold and copper metalwork carved stone feather headdressesand fine cloth which were becoming increasingly available from the workshops ofskilled often state-sponsored craftsmen in central and western Mexico from the EarlyClassic 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 north-central Mexico that we should look for new perspectives on the northward expansion ofthe Mesoamerican frontier into the cold dry lands of the tierra friacutea

In sum the inhabitants of the Mesoamerican tierra friacutea in central and north-central Mexico were living at the colder drier edges of a civilization rooted in warmer wetter


lower lands to the south east and west The constraints of nature made tierra friacuteapopulations particularly dependent upon technological and organizational innovation fortheir survival as Mesoamericans Maguey cultivation was a key part of this innovationand survival From the Early Formative onward the presence of more complex societiesalong their peripheries provided both material benefits and socio-political problems for tierra friacutea peoples These benefits and problems in turn would have provoked technological and organizational innovationsmdashresponses that we perceive today as the long-term northward expansion of the Mesoamerican frontier into central and north-central Mexico

REFERENCES

Anawalt P (1980) Costume and control Aztec sumptuary laws Archaeology 3333ndash43 Anawalt P (1990) The emperorrsquos cloak Aztec pomp and Toltec circumstances

American Antiquity 55291ndash307 Anderson RK 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 Nationrsquos Health 368 883ndash903

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 161ndash84 Washington DC Dumbarton Oaks

Blanton R Kowalewski S Feinman G and Appel J (1981) Ancient MesoamericaCambridge Cambridge University Press

Blasquez P and Blasquez I (1897) Tratado del Maguey Puebla Mexico Narisco Bassols second edition

Braniff B (1989) Oscilacioacuten de la frontera norte Mesoamericana un ensayo nuevoArqueologiacutea 199ndash114 Mexico DF Institute Nacional de Antropologiacutea e Histoacuteria

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 womenrsquos production in Aztec Mexico In JGero and MConkey (eds) Engendering Archaeology 224ndash51 Oxford Basil Blackwell

Bye R Burgess D and Trias A (1975) Ethnobotany of the Western Tarahumara ofChihuahua Mexico I notes on the genus Agave Botanical Museum Leaflets Harvard University 24585ndash112

Cabrero MT (1989) Civilization en El Norte de Mexico Mexico DF Universidad Nacional Autoacutenoma de Mexico

Carrasco P (1976) La Sociedad Mexica antes de la Conquista Histoacuteria General de Meacutexico 1165ndash288 Mexico DF El Colegio de Mexico

Crosby A (1972) The Columbian Exchange Biological and Cultural Consequences of


1492 Westport Connecticut Greenwood Press Crosby A (1986) Ecological Imperialism The Biological Expansion of Europe 900ndash

1900 AD Cambridge Cambridge University Press Darling JA (1998) Obsidian Distribution and Exchange in the North-Central Frontier

of Mesoamerica AD 0ndash1500 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 Aztecwound remedy Journal of Ethnopharmacology 8149ndash61

Dibble C and Anderson A (1963) (translators) Florentine Codex General History of the Things of New Spain by Fray Bernardino de Sahagun Book 11mdashEarthly ThingsSanta Fe Monographs of the School of American Research and the Museum of NewMexico

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 Tumacacoriacute Dragoon Arizona The Amerind Foundation Inc Report No 7

DiPeso C Rinaldo J and Fenner G (1974) Casas Grandes A Fallen Trading Centerof 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 SpainNew York Orion Press translated by DHeyden and FHorcasitas

Ekholm G (1942) Excavations at Gusave Sinaloa Mexico Anthropological Papers of the American Museum of Natural History 38223ndash139

Evans S (1990) The productivity of maguey terrace agriculture in central Mexico duringthe Aztec period Latin American Antiquity 1117ndash32

Fish S Fish P Miksicek C and Madsen J (1986) Prehistoric agave cultivation insouthern Arizona Desert Plants 72107ndash12

Flannery K (1968) Archaeological systems theory and early Mesoamerica In BMeggers (ed) Anthropological Archaeology in the Americas 67ndash87 Washington DC Anthropological Society of Washington

Flannery K (1973) The origins of agriculture Annual Review of Anthropology 2 271ndash310

Flannery K (1986) (ed) Guila Naquitz Archaic Foraging and Early Agriculture inOaxaca Mexico New York Academic Press

Flannery K Marcus J and Kowalewski S (1981) The Preceramic and Formative of theValley of Oaxaca In JSabloff (ed) Supplement to the Handbook of Middle American Indians 149ndash83 Austin University of Texas Press

Foster M (1978) Loma San Gabriel A Prehistoric Culture of Northwest MexicoBoulder University of Colorado unpublished PhD thesis Ann Arbor UniversityMicrofilms

Foster M (1985) The Loma San Gabriel occupation of Zacatecas and Durango MexicoIn MFoster and PWeigand (eds) The Archaeology of West and NorthwestMesoamerica 327ndash52 Boulder CO Westview Press

Gamio M (1979) (ed) [1922] La Poblacioacuten del Valle de Teotihuacaacuten Tomo II Mexico DF Institute Nacional Indigenista


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 DF Fondo de Cultura Econoacutemica second edition

Guerrero R (1980) El Pulque Religion Cultura Folklore Pachuca Mexico Institute Nacional de Antropologiacutea e Histoacuteria Centre Regional de Hidalgo

Healan D (1977) Archaeological implications of daily life in ancient Tollan HidalgoMexico World Archaeology 9140ndash56

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 artifactsscraper planes from Mitla and Yagul Oaxaca Berkeley Contributions of the University of California Archaeological Research Facility 14107ndash23

Hicks F (1987) First steps toward a market-integrated economy in Aztec Mexico InHClassen and PVan de Velde (eds) Early State Dynamics Studies in Human SocietyVolume 291ndash107 Leiden Brill

Hough W (1908) The pulque of Mexico Washington DC Proceedings of the US National Museum 33577ndash592

Kelley JC (1990) The Retarded Formative of the northwest frontier of Mesoamerica InMCarmena Macias (ed) El Preclaacutesico o FormativomdashAvances y Perspectivas 405ndash23 Mexico DF Institute Nacional de Antropologiacutea e Histoacuteria Museo Nacional de Antropologiacutea

Kelly I (1945) The Archaeology of the Autlan-Tuxcacuesco Area of Jalisco Part IBerkeley 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 TheTuxcacuesco-Zapotitlan Zone Berkeley University of California Press Ibero-Americana 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 Ecology115ndash38 Washington DC 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 MiddleAmerica In RWauchope (ed) Handbook of Middle American Indians 1427ndash45 Austin University of Texas Press

Mastache G Cobean R Rees C and Jackson D (1990) Las Industrias Liacuteticas Coyotlatelco en el Area de Tula Mexico DF Institute Nacional de Antropologiacutea e Histoacuteria


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 DF Centre drsquoEtudes Mexicaines et Centroamericaines Etudes Mesoamericaines 9

Murra J (1962) Cloth and its functions in the Inca state American Anthropologist 64 710ndash28

Nelson B (1997) Chronology and stratigraphy at La Quemada Zacatecas MexicoJournal of Field Archaeology 2485ndash109

Norr L (1987) The excavation of a Postclassic house at Tetla In DGrove (ed) Ancient Chalcatzingo 400ndash8 Austin University of Texas Press

Parsons JR 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 45ndash79 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 ofMexico American Antiquity 40207ndash15

Patrick L (1985) Agave and Zea in highland central Mexico the ecology and history of the Metepantli In IFarrington (ed) Prehistoric Intensive Agriculture in the Tropics539ndash47 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 TikalGuatemala Implications for Subsistence Models Philadelphia University of Pennsylvania unpublished PhD thesis

Rangel S (1987) Etnobotaacutenica de los Agaves del Valle del Mezquital Mexico DF Universidad Nacional Autonoma de Mexico unpublished thesis

Rodriguez F (1985) Les Chichimeques Mexico DF Centre drsquoEtudes Mexicaines et Centrameacutericaines

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 Histoacuteria y Presente de Epazoyucan HidalgoTexcoco Mexico Universidad Autoacutenoma Chapingo Depto de Industrias Agriacutecolas Coleccioacuten Cuadernos Universitarios Serie Ciencias Sociales No 4

de Sahagun B (1969) Histoacuteria General de las Cosas de Nueva Espantildea Mexico DF Editorial Porrua edited by AMGaribay three volumes

Sahlins M (1968) Tribesmen Englewood Cliffs Prentice-Hall Salinas J and Bernard R (1983) Etnografiacutea Otomi Mexico DF Institute Nacional

Indigenista Sanchez J (1980) Etnografiacutea de la Sierra Madre Occidental Tepehuanes y

Mexicaneros Mexico DF Institute Nacional de Antropologiacutea e Histoacuteria Sanders WT (1965) The Cultural Ecology of the Teotihuacan Valley A Preliminary

Report University Park PA Pennsylvania State University Department of


Anthropology Sanders WT (1966) Life in a Classic village In Teotihuacan Onceava Mesa Redonda

1123ndash47 Mexico DF Institute Nacional de Antropologiacutea e Histoacuteria Sanders WT Parsons JR 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 31353ndash64 Sejourne L (1983) Arqueologiacutea e Histoacuteria del Valle de Meacutexico de Xochimilco a

Amecameca Mexico DF Siglo Veintiuno Editores Serra Puche M (1988) Los Recursos Lacustres de la Cuenca de Mexico durante el

Formativo Mexico DF Institute de Investigaciones Antropoloacutegicas Universidad Nacional Autoacutenoma de Mexico

Smith C (1967) Plant remains In DBeyers (ed) The Prehistory of the Tehuacan Valley1220ndash55 Austin University of Texas Press

Smith C and Kerr T (1968) Pre-Conquest plant fibres from the Tehuacan Valley Mexico Economic Botany 22354ndash8

Smith M and Hirth K (1988) The development of pre-hispanic cotton-spinning technology in western Morelos Mexico Journal of Field Archaeology 15349ndash58

Spence M (1971) Some Lithic Assemblages of Western Zacatecas and DurangoCarbondale Southern Illinois University University Museum Mesoamerican StudiesNo 8

Taylor W (1966) Archaic cultures adjacent to the northeastern frontiers of MesoamericaIn RWauchope (ed) Handbook of Middle American Indians 459ndash94 Austin University of Texas Press

Tesch M and Abascal R (1974) Azadas Comunicaciones 1137ndash40 Puebla Mexico Fundacioacuten Alemana para la Investigacioacuten Cientiacutefica

Tolstoy P (1971) Utilitarian artifacts of central Mexico In GEkholm and IBernal (eds)Archaeology of Northern Mesoamerica Pt 1 Handbook of Middle American Indians10270ndash96 Austin University of Texas Press

Trombold C (1985) A summary of the archaeology in the La Quemada region InMFoster and PWeigand (eds) The Archaeology of West and Northwest Mesoamerica237ndash67 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 DF Informe to the Institute Nacional de Antropologiacutea e Histoacuteria

Trombold C (1990) A reconsideration of chronology for the La Quemada portion of thenorthern Mesoamerican frontier American Antiquity 55308ndash24

Vaillant G (1931) Excavations at Ticoman New York Anthropological Papers of theAmerican 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 Buffalounpublished MA thesis

Willey G Ekholm G and Millon R (1964) The patterns of farming life andcivilization In RWauchope (ed) Handbook of Middle American Indians 1446ndash500 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 arein less developed world contexts or relate to prehistoric or historic episodes before thedevelopment of modern technology In this chapter we document an instance of an extanttraditional dryland irrigation system (termed bisse) in Switzerlandmdashone of the most developed and technologically sophisticated countries in Europe Although drylandirrigation systems are widespread in the semi-arid regions of the northern and centralparts of the Mediterranean basin these are comparatively little documented (Hunt andGilbertson 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 ofthis system It takes a historical perspective and deals mostly with the social and culturalstructures that have developed to control access to the water and which incidentallyaccount for much of the success of this system These typically are difficult to recoverfrom the archaeological record

A major area of uncertainty with research into abandoned systems is the problem oftheir environmental relationships Contemporary field measurements enable assessmentof the ways in which these systems interact with landscape processes These are criticalbecause in part they account for the robustness and longevity of some dryland irrigationsystems

THE VALAIS

The Valais is a mountainous canton in southwest Switzerland (Fig 171) Altitudes range from 372 m at Lake Geneva to 4634 m at the summit of Pointe Dufour Topographicallythe Valais can be divided into three regions

Figure 171 The Valais canton Switzerland showing places mentioned in Chapter 17

mdashthe Rhocircne valley the tributary valleys and the mountain zones The Rhocircne graben or trench divides the two main mountain zones the Bernese Oberland to the north and thePennine Alps to the south Settlement is concentrated in low-lying areas such as the Rhocircne valley and its major lateral valleys Today the canton is also divided culturally intotwo 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 Frenchspeaking This study focuses particularly on the commune of Vernamiegravege located in the Central Valais on the southern edge of the Rhocircne graben (Fig 172)

Climate

The main controls on mountain climates are altitude continentality latitude andtopography (Beniston 1994) The Valais lies within a ring of high alpine mountains andso is partly in rain shadow The reduced amounts of precipitation received together withhigh evapotranspiration as a result of the high summer temperatures and low humiditymean therefore that areas within the Canton can properly be described as semi-arid using the definitions of UNEP (1992) and Reynard (1995) Annual precipitation increaseswith altitude from around 580 mm per year on the Rhocircne valley floor to about 2100 mm in the high alps (Loup 1965 Reynard 1995) Aspect also controls humidity throughdifferent thermal regimes south-facing adret slopes receive 50 per cent more sunshine than north-facing ubac slopes (Loup 1965) Precipitation can vary considerably fromyear 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


extensive dry periods throughout the summer and indeed in spring and autumn as a resultof the foehn winds (Boueumlt 1972) High summer evapotranspiration leads to water deficitsof as much as 300 mm per month during the growing season (Michelet 1995 Primaultand Catzeflis 1966) particularly in the Central and Haut Valais Reynard (1995)suggests that during the summer months 23ndash30 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 1994Loup 1965) partly because of the high altitude and steep slopes of this alpine terrain Analtitudinally-sensitive pattern of agricultural land use (Netting 1972) incorporating asophisticated traditional irrigation culturemdashthe bissesmdashhas emerged in response to restricted land and water availability Pasture land is concentrated at high altitude (up to2600 m) and most arable activity occurs below 1500 m Pasture vines orchard cropsand 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 werelargely self-sufficient with a range of land types and

Figure 172 Distribution of agricultural land in Vernamiegravege during the 1960s

Source Modified from Berthoud 1967 figure 33

therefore products distributed throughout the commune (Fig 172) Most will have had access to vineyards on the lower slopes (700ndash900 m) with hayfields cereal crops and

Traditional irrigation systems in dryland Switzerland 309

vegetables being grown around the settlement (900ndash1600 m) and pasture and alpine summer grazing above (1900ndash2400 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 harvestthe payment of taxes and tithes and to provide for tools and so on families andcommunities would attempt to produce a surplus for saleexchange or an off-farm income from other activities (Jones 1991)

In a pastoral economy the quantity and quality of the productmdashfor example cheesemdashwill depend on the quantity and quality of the grass eaten by the cows Also the animalsneed to be supported throughout the year and in the Valais the harvest must support notonly the human population during the long snow-bound winter months but also their livestock Approximately 10000ndash12000 m3 per hectare of water are required during the growing season for successful hay meadows (Muller 1946) Michelet (1995) hascalculated that with the high evapotranspiration rates and a low growing season rainfallof 300 mm there is a water deficit of 7000ndash9000 m3 per hectare The hay meadows (mayens) are also important in the transhumance process providing cattle with interimgrazing on the way up and down to the high alpine pastures Whilst there is generallysufficient rainfall for alpine pastures to provide adequate grazing intensive exploitationdoes mean that the pastures need to be periodically improved particularly where they areunderlain 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 ofprecipitation is fairly even throughout the year most of the precipitation in the wintermonths is as snow and extensive areas of the canton are glaciated at high altitude Thismeans that there is a source of water that can be used for irrigationmdashglacial meltwatermdashbut 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 continuesto 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 slopeofftake 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 andor artificial or subterranean channels of any dimension that is or has been used to supply andor distribute water under gravitational flow primarily for locally governed and organized irrigation

(Crook 199778)

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


Crook and Jones (1999b) set out the design principles of the bisse system distinguishing between traditional and modern technologies and showing how innovation and adaptationhave taken place The comparatively simple technology enabled a quick response toperiodic and haphazard physical disruption The continuity of the system has beenachieved 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 LaTegravene 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 laterconstruction particularly in the fourteenth century The earliest surviving documentationof 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 anintensification of agriculture (Crook 1997) Any extension of agriculture particularly inthe drier areas would have required the exploitation of new water resources The earliestmention of the bisses is certainly in some of the driest areas such as Visp (Viegravege) Raron (Rarogne) and Sierre (Dubuis 1995) New economic opportunities may also haveprovided incentives to intensify agriculture through irrigation Following thedemographic impact of the plague in 1349 there was a reduction in the demand forcereals This meant that the surviving population could convert land to cattle productionand 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 andbisse irrigation technology provided that opportunity Inventories (Aufdereggen and Werlen 1993 Rauchenstein 1908) and analysis of archival records of first mentions ofbisses (Crook 1997) indicate that there was an expansion in construction during the fourteenth and fifteenth centuries The opportunities for agricultural extension were alsoprovided by the general retreat of the glaciers between cAD 1100 and 1400 enablingmore 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 therewas a need to intensify agriculture for economic reasonsmdashthe need for example to support a growing urban population At this time however the glaciers were at theirhistorical maxima (Aellen 1988 Chen 1990) Clearly the response to opportunity andstress 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

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


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 tothe water the amount of water that can be taken by any one individual at a particulartime and the provision of maintenance requirements Clearly access to irrigation waterswas an important economic determinant for mountain farmers attempting to exploit aniche advantage In response to the nature of the environment and the task of bringingwater from one locality to another over distances as much as 32 km (Crook 1997) bisseconstruction required resources far greater than any one individual could provide Thismeant that the farmers needed to work co-operatively (Fig 173 Table 171) Co-operation needed to continue after the basic system was constructed To this endassociations 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 corveacutee 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 waterrights are attached to most water sources Crook (1997) notes how this has led to bissescrossing each other and large torrents by-passing conduitsmdashhydrologically 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 ofSt-Maurice and Sion (Ammann 1995)


Figure 173 The Grand Bisse de Lens Key Sectors from which water was taken by irrigaters from the communes of Icogne (A) Lens (B) Chermignon drsquoen Haut (C) Chermignon drsquoen Bas (D) and DiognemdashMontana (E) Source Modified from Crook 1997


Conferred tenure has been transformed into a type of hydraulic tenure with the waterrights 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 lsquoprior usersquo (Stelling-Michaud 1956) Such claims led to disputes however which could ultimately lead to the destruction of the bisse by one of the parties (Beacuterard 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 tothe 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 drsquoeau 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 suchas duration or permanence flexibility the quality or title of security transferability anddivisibility (Scott and Coustalin 1995) Individual water rights have low exclusivityhowever since independent use is difficult without multilateral control and agreementupstream and down Any water right is dependent on natural variation in flow andsometimes on upstream users In times of drought or other stresses the notion ofexclusivity is adapted to reasonable shares in proportion to full rights between all bissesdown a slope profile (Stelling-Michaud 1956) The predetermined flow of water in mostbisses 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 preventaccess to the common resource by outsiders (Jones 1987) Examples of these restrictivepractices are found for instance in the consortages of the Grand Bisse de Lens Bisse deVercorin and Bisse Dessous (Crook 1997) Other measures to protect the access to waterrights included reclaiming rights from women who married men from outside thecommune (Netting 1972) Water rights are however divisible and have beencharacterized by fragmentary inheritance strategies and family agreements (Weigandt1977 Weinberg 1972) The sharing and renting of water rights among those eligibleenable 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

Table 171 Approximate numbers of named irrigators using the Grand Bisse de Lens lsquoaqueductis communirsquo in 1457

Commune No of irrigators Icogne 16 Lens 25 Chermignon drsquoen Haut (superieur) 23 Chermignon drsquoen Bas (inferieur) 12 Diogne-Montana 7 Total 83 Source Commune Archive of Chermignon and Lens 16


longer have any need of them For example the 1980 register of water rights for theGrand 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 Lensconsortage archive) This situation is resolved by these individuals being asked torelinquish rights where they have no practical use Outside agriculture these rights haveno monetary value (Grand Bisse de Salins consortage archive) though sentimental attachment means that not all are willing to do this Equally some families have acquiredmore water rights over the years through inheritance as a result of which they also holdmore 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 ofchanges (mutations) to water acquisition and allocation are described and recorded in theratement The ratement is a useful documentary tool in plotting the expansion andcontraction 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 ofconsorts is difficult to detect because often only one family member will be namedchanges 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 Vernamiegravege had five ratementsbetween 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 recordwhen water may be taken from the bisse for how long andor how much may be usedThe precise arrangements varied from locality to locality as did the terminologyemployed (Crook 1997) One droit drsquoeau on the Bisse de Clavoz would irrigate 3040 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 Forexample the Bisse Vieux receives a total discharge of 150 litres per second and there aresix water rights associated with this bisse hence each water right is equal to 25 litres persecond (source Bisse Vieux archive) Water rights are usually divided according to theday hour and rotation (tourskehrs) (Table 172) 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 nightmdash0400ndash1800 hours morning afternoon and nightmdash0500ndash1300 1300ndash2100 and 2100ndash0500 hours early morning morning afternoon and nightmdash0400ndash0900 0900ndash1400 1400ndash2000 and 2000ndash0400 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 determinedby the position


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

Table 172 Examples of tours with the number of droits and sequence of irrigation hours

a Bisse Vieux 1839Tour 2 Tiers No of droits Sequence of irrigation hours1 3 15 3 6 2 3 6 6 12 3 5 5 2 12 4 12 5 7 Total 11 72 Tour 4 Quarte No of droits Sequence of irrigation hours1 4 12 3 4 12 4 12 2 3 8 10 6 3 9 2 2 112 112 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 1865Tour 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 12 10 12 8 2 2 3 9

Source Bisse Vieux ratement 1 865 (Communal Archive of Nendaz P324) c Bisse Vieux 1924Tour 2 Droit No of people Sequence of irrigation hours1 5 9 3 5 12 1 1 12 2 9 2 2 2 1 12 1 12 1 12 1 12 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)


vary with the length and discharge of the bisse and the number of irrigators (Table 172) 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 anddemand The distribution of the water would normally take place in rotated sequencedown the bisse with each subsequent section of the bisse receiving water in turn (Figs 174 and 175) Sundays and feast days would normally be reserved for irrigating church lands The sequence would normally be repeated with each rotation although there couldbe 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 fieldsclosest to the village to reduce the risk of injury and to save effort This practice is stillcurrent for instance in Ausserburg (Crook 1997)

Figure 174 The irrigation sectors in Vernamiegravege Note The numbering of sectors is the same as in Figure 175 Source After Berthoud 1967 Figure 38


Figure 175 Distribution of water during the first tour from the bisses of Vernamiegravege 5th Mayndash8th June 1964

Note The numbering of the sectors is as in Figure 174 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 andincorporated the use of ocular tools (Marieacutetan 1948) each family would have adistinctive mark consisting of a series of dots and lines and for each bisse sector the


family marks of all eligible families were inscribed onto all sides of a wooden stick or onto a wooden block (termed tessel tesseln or wassertessle) together with symbols indicating the number and timing of water rights owned by each family (Briw 1961Lautenschlager 1965 Fig 176) 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 Thetessel system was still in use in some areas such as Mund Zeneggen and the Loumltschental 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 bedistributed The details concerning the rules and regulations of the consortages and of the process of water management and of distribution were contained in a documentknown as the reacuteglement These were first translated from the Latin into the vernacular inthe sixteenth century (Bratt 1995) and in many cases are still operative today althoughthey have been reviewed and modernizedmdashthe reacuteglement for the Grand Bisse de Lens for example underwent major revisions to the 1457 original in 1698 1914 and 1980 Thelarge temporal gaps between new statutes and reacuteglements hint at processes operating outside the rule books the resultant documents are reflections of the complexity of caselaw and of the careful preservation of institutional memory (Crook 1997 Mahdi 1986)

Monetary fines cautionary tales of ghostly processions exclusion and the threat ofpurgatory 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 forirrigation has resulted in bitter disputes between the controlling bodies (Table 173) Those communes with the greatest threat to their water security demonstrate some of thebest examples of ongoing disputes Many disputes arose at the time of construction oreven before and many operational disputes usually have their origins in these earlierevents Disputes of this nature were referred to the Bishop of Sion until 1627 andthereafter to the Cantonal civil courts although the local priests still played an importantpart 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 amillennium and as such has contributed to the distinctive Valaisian landscape


Figure 176 A tessel used by members of the consortage of the Grand Bisse de Lens at Chermignon drsquoen 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 dayrsquos 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


Table 173 Examples of bisse disputes Commune Dates Reason Settlement Arbitrators Lax v Martisburg1

1347 1367 1443 1554

Climate change leading to a water supply drying up

Sequential sharing of remaining sources

Bishop of Sion

Marisburg v Fiesch amp Fieschertal2

1351ndash1747 Water rights for a new suon

Agreement between communes

1811ndash1961 Dispute over water rights

Document3 Napoleon Bonaparte

Sion v Ayent amp Saviegravese4

1484 Claims of illegal use and sale of water

No official judgement

Arbaz v Grimisuat5

1686 Illegal diversion of water between two points

Construction of partition

Bishop of Sion

Vercorin Recircchy amp Chalias v Grocircne amp Loye6

13851390ndash1448

Insufficient water in dry spells because of excessive abstraction for the Grande Bisse Neuf at Grocircne

Sharing arrangement 13 Vercorin Recircchy amp Chalais 23 Grocircne amp Loye

Inhabitants of Lens Lords of Grange and Bishop of Sion

1548ndash1565 Construction of a new bisse without authorization or water rights

Compensation payment amp dry weather clause

Tribunal

Bagnes v Levron7

1443ndash1465 1515 1545 1626 162930 1839 1923

Opposition to a new bisse because of challenge to water rights and damage claims leading to sporadic vandalism and destruction

Compensation appointment of guards (largely ineffective)

Initially the Abbot of St-Maurice

Conthey (Savoy) v Saviegravese (Valais)8

C14-C17 Territorial dispute over alp and bisse source leading to murder assassination sackingburning of village

Improvement as from 1462 when Savoy was beaten at the Battle of La Plante

Overseers of Bern and Fribourg

Ayent v Sion9

1950 Dispute over usufructory and

Convention in favour of Sion

Riedmatten amp Zimmermann


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 inpart to the nature of social systems and technological responses but it also reflects thecomparatively low levels of environmental impact The range of contemporary waterquality in the bisses has no detrimental effect on soil alkalinity sodicity and salinitymdashfactors 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 insoils Gullying and sheet wash erosion on steep slopes resulting from the use of gravitydistribution techniques have been negated by concentrating ruissellement on resilient hay and grass meadows Terracing has also been used to reduce slope angles in pasturesorchards and arable fields The glacial water is carried over long distances enabling it towarm thus preventing plant damage and making the water safe for cattle to drink (Crook1997) The meltwaters can also carry large amounts of sediment which provide lining tobisse 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 particularlyon the Rhocircne 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 ofbiodiversity also follows abandonment as patterning imposed by the bisse disappears For this reason some bisses are now maintained as part of a general landscapemanagement strategy particularly in tourist areas (Crook and Jones 1999a)

ascribed water rights and tariffs refusal to sign the convention

solicitors

Commune Dates Reason Settlement ArbitratorsConsortage of the Grande Bisse de Lens v multiple interest group10

1989 to date

The Sarmona section of the bisse lost large quantities of water through infiltration The solution a concrete conduit used already on other sections without complaints has been objected to as being insensitive to the natural environment and ecology

Pendingmdashthe probable outcome will be a compromise

Tribunal with representation by Icogne and Crans Development Society environmental pressure groups and the public

Sources 1 2 Liniger (1980) 3 Deacutepartement des Simpelberges (1811) 4 5 9 Communal Archive at Sion 6 Stelling-Michaud (1956) 7 Beacuterard (1982) 8 Roten-Dumoulin (1990) 10 Grand Bisse de Lens consortage archive


DISCUSSION

This chapter has documented a traditional dryland irrigation system in one of the mostdeveloped 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 matrixalthough their lack of adverse environmental impact must also have been significantNotably 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 weatherand repeated crop failures (Crook 1997 Pfister 1994) The systems originated in earlymedieval times and have survived feudalism and its break-up the Napoleonic invasion the appearance of industrial society and the modern communications and economicrevolutions 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 modernmaterials and techniques have been selectively adopted (Crook and Jones 1999b) Thesocial structures show some similarities with other dryland systems OrsquoNeill (1987) demonstrated a complex and well-adjusted social matrix to Portuguese irrigation systemsA 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 individualwater rights functioning within a corporate setting and with an element of democracy indecision making Also significant is a collective memory which provides lsquocase lawrsquo and an effective mechanism for conflict resolution The systemsrsquo physical longevity can be ascribed to the flushing of salts from the fields by the application of water lsquoin excessrsquo 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 couldpossibly be generally applied to comparable systems of irrigation evidenced in thearchaeological record

ACKNOWLEDGEMENTS

DSCrook acknowledges a University of Huddersfield Research Studentship and a grantfrom the Dudley Stamp Memorial Fund The help of numerous Valaisan farmers andofficials was invaluable CO Hunt drafted the diagrams and suggested modifications tothe text


REFERENCES

Aellen M (1988) Fluctuations of Glaciers 1980ndash5 Volume 5 Paris IAHS-UNESCO Ammann HR (1995) Aperccedilu sur les documents relatifs aux canaux drsquoirrigation du Haut

Valais agrave lrsquoeacutepoque meacutedieacutevale Annales Valasiennes 70263ndash80 Aufdereggen J and Werlen C (1993) Rapport BissesSuonens Sion Canton du Valais

Service de lrsquoEnvironement et lrsquoAmeacutenagement du Territoire Beniston M (1994) Climate scenarios for mountain regions an overview of possible

approaches In MBeniston (ed) Mountain Environments in Changing Climates 136ndash53 London Routledge

Beacuterard C (1982) Bataille pour lrsquoEau Sierre Les Editions Monographic SA Berthoud G (1967) Changements Eacuteconomiques et Sociaux de la Montagne Vernamiegravege

en Valais Berne Franke Boueumlt M (1972) Climat et Meacuteteacuteorologie de la Suisse Romande Lausanne Payot

Edition Bratt G (1995) The Bisses of Valais Man-Made Watercourses in Switzerland Gerards

Cross Guy Bratt Briw VA (1961) Aus Geschichte und Brauchtum der pfarrgemeinde Fiesch Visp

Buckdruckerie Mengis Chen JY (1990) Changes of Alpine Climate and Glacier Water Resources Zurich

Eidgenoumlssische Technische Hochschule unpublished PhD thesis Cosinschi M (1994) Le Valais Cartoscopie drsquoun Espace Reacutegional Lausanne Editions

Payot Crook DS (1997) Sustainable Mountain Irrigation The Bisses of the Valais

Switzerland a Holistic Appraisal Huddersfield University of Huddersfield unpublished PhD thesis

Crook DS and Jones AM (1999a) Traditional irrigation and its importance to thetourist landscape of Valais Switzerland Landscape Research 2449ndash65

Crook DS and Jones AM (1999b) Design principles from traditional mountainirrigation systems (bisses) in the Valais Switzerland Mountain Research and Development 1979ndash99

Curdey P Mottet M Nicoud C Baudais D Lundstroumlm-Baudais K and Moulin B (1993) Brig-GlisWaldmatte un habitat alpin de lrsquoacircge du Fer fouilles archeacuteologiques N9 en Valais Archeacuteologie Suisse 16138ndash51

Daudry D and Daudry G-J (1995) Le ru de Mazod-Cheacutetoz (Valleacutee drsquoAoste Italie) Histoire techniques de construction importance agricole Annales Valaisannes 70 143ndash62

Deacutepartement des Simpelberges (1811) Franzoumlsisches Reich Simplon Deacutepartement de Simpelberges

Dubuis P (1995) Exposeacute introductif bisse et conjoncture eacuteconomique le cas du Valais aux XIVegrave et XVegrave sieacutecles Annales Valaisannes 7039ndash46

Farrington IS and Park CC (1978) Hydraulic engineering and irrigation agriculture inthe Moche valley Peru cAD 1250ndash1532 Journal of Archaeological Science 5255ndash


68 Grove AT and Grove JM (1990) Traditional montane irrigation systems in modern

Europe an example from Valais Switzerland Agriculture Ecosystems and Environment 33181ndash6

Harris B (1971) The Monte Moro pass and the Col drsquoHeacuterens Alpine Journal 76 127ndash32

Harris B (1972) Travel and trade in the Pennine Alps Alpine Journal 77175ndash82 Hunt CO and Gilbertson DD (1998) Context and impacts of ancient catchment

management in Mediterranean countries implications for sustainable resource use InHWheater and CKirby (eds) Hydrology in a Changing Environment Volume II 473ndash84 Chichester John Wiley and Sons

Jones AM (1987) Kin relations in a French alpine community a preliminaryinvestigation Sociologica Ruralis 27304ndash22

Jones AM (1991) Exploiting a marginal European environment population control andresource management under the Ancien Reacutegime Journal of Family History 16 363ndash79

Jones AM and Hunt CO (1994) Wells walls and water supply aspects of the culturallandscape of Gozo Maltese Islands Landscape Issues 1124ndash9

Jones AM Hunt CO and Crook DS (1998) Traditional irrigation strategies and theirimplication for sustainable livelihoods in semi-arid area examples from Switzerland and the Maltese Islands In HWheater and CKirby (eds) Hydrology in a Changing Environment Volume II 485ndash94 Chichester John Wiley and Sons

Jossen E (1989) Mund Das Safrandorf im Wallis Naters Commune de Mund Lautenschlager E (1965) Le systeme drsquoirrigation drsquoAusserberg en Valais Bulletin

Murithienne 889ndash16 Liniger M (1980) Bisses et autre raz des Alpes occidentales Les Alpes 5642ndash4 Loup J (1965) Pasteurs et Agriculteurs Valaisans Contribution agrave lrsquoEacutetude des

Problegravemes Montagnards Grenoble Imprimerie Allier Macheral C (1984) Lrsquoeau du glacier Eacutetudes Rurales 93ndash94205ndash38 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 181ndash98 Washington DC National Academy Press

Marieacutetan I (1948) Les Bisses La Lutte pour lrsquoEau en Valais Neuchatel Editions du Griffon

Michelet P (1995) Les techniques drsquoentretien les bisses Annales Valaisannes 70163ndash74

Muller H (1946) De quelques solutions nouvelles du problegraveme de lrsquoirrigation Bulletin Murithienne 6333ndash40

Netting RMcC (1972) Of men and meadows strategies of alpine land useAnthropological Quarterly 45132ndash44

Netting RMcC (1981) Balancing on an Alp Ecological Change and Continuity in a Swiss Mountain Community Cambridge Cambridge University Press

OrsquoNeill BJ (1987) Social Inequality in a Portuguese Hamlet Cambridge Cambridge University Press

Ostrorn E (1990) Governing the Commons The Evolution of Institutions of Collective


Action Cambridge Cambridge University Press Pfister C (1994) Spatial Patterns of Climatic Change in Europe in AD 1675ndash1715

Bern Gustav Fischer Verlag Primault B and Catzeflis J (1966) Du climat valaisan La Recherche Agronomique en

Suisse 5248ndash67 Quaglia CL (1984) Le Mont du Lens Lens Commune de Lens Rauchenstein F (1908) Les Bisses du Canton Valais Sion Departement de lrsquoInteacuterieur Reynard E (1995) Lrsquoirrigation par les bisses en Valais Approche geacuteographique Annales

Valaisannes 7047ndash64 Roten-Dumoulin R-M (1990) Saviegravese une Commune Rurale dans le Valais du XIXegrave

Siegravecle Brig Rotten-Verlags AG Ruedin P (1986) Le Bisse de Clavoz au fil de lrsquoeau et des ans 13 Eacutetoiles Reflets du

Valais 534ndash5 Scott A and Coustalin G (1995) The evolution of water rights Natural Resources

Journal 35821ndash979 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 3828ndash36 Walter-Coward E (1990) Property rights and network order the case of irrigation works

in the western Himalayas Human Organisation 4978ndash88 Weigandt E (1977) Inheritance and demography in the Swiss Alps Ethnohistory 24

133ndash48 Weinberg D (1972) Cutting the pie in the Swiss Alps Anthropological Quarterly 45

125ndash31


18 Desertification land degradation and land abandonment in the Rhocircne valley France

SANDER VAN DER LEEUW

INTRODUCTION

Since the mid-1980s the European Union has had a research programme on the causes ofdesertification and land degradation in southern Europe (Fantechi and Margaris 1986) Itinitially focused on climate (eg the HAPEX-SAHEL and EFEDA programmes) but from the late 1980s two other elements were introduced the study of atmosphere-geosphere interactions and their effect on land use and living conditions in the drylandsof Spain Italy and Greece (eg the MEDALUS and ERMES programmes) and the studyof long-term human-environmental interactions (eg the ARCHAEOMEDESprogramme the focus of this chapter van der Leeuw 1998a) This development reflectedtwo consecutive changes in perspective moving from the idea that people (reactively)adapt to their environment to focusing on their proactive role in modifying theirenvironment (to its detriment) and somewhat later to accentuating their interactive andmutually dependent relationship with it The shift offers an interesting opportunity forarchaeology because studying long-term natural processes without looking at the socio-cultural dynamics of human society makes little sense in this context Archaeology is theonly discipline that can do so

However to meet this challenge archaeologists have to overcome some importantintellectual difficulties these are either relicts from the history of our discipline (such asour tendency to consider the past for the pastrsquos sake) or are due to the wider context ofthe 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 hasbrought together a team of researchers from up to seven European countries representinga full range of academic and applied disciplines with the aim of improving ourunderstanding of desertification land degradation and land abandonment along thenorthern 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 Rhocircne valley Vera basin Empordagrave Isle of Braccedil) to the Roman and medieval periods (lower Rhocircne valley) and the present (Argolid Veneto Languedoc-Roussillon Midi-Pyreacuteneacutees Marina Baixa Baixo Mondego) These studies have beenundertaken on a range of spatial scales in different climate zones and focusing ondifferent aspects of human-environmental interaction degradation and desertification

This chapter summarizes the multidisciplinary research that was undertaken from 1992

to 1999 in southern France by one of the ARCHAEOMEDES teams consisting ofenvironmental and cultural archaeologists social and physical geographers statisticiansremote sensing and information scientists an economist and an ancient historian Initially(1992ndash94) one group focused on the archaeology of the lower and middle Rhocircne 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 Rhocircne valley over the last 10000 years and human-land relationships in the Roman period in the middle and lower valley (Fig 181)

CLIMATE ENVIRONMENT AND PEOPLE IN THE VALDAINE

In the Valdaine the region around Monteacutelimar in the middle valley the fact that 40 km oftrenches were being dug allowed us to investigate the exposed sections and takenumerous micromorphological samples in nested areas with spatial scales of 01 1 10and 100 km2 By correlating these columns we built up a detailed three-dimensional interpretation of the erosionmdashcumulation dynamics of this landscape throughout theHolocene Temporal calibration was based on a combination of archaeological andradiocarbon dating of the sections We were able to distinguish the traces left by differentkinds of socio-natural impact on the landscape such as

bull erosive crises regularly rejuvenating the soil (middle neolithic late neolithic middle iron age Roman [third century AD] and modern)

bull degradation from over-intensive agriculture (early Roman empire [first and second centuries AD] and modern period)

bull degradation of the drainage of the soil due to rising riverlake levels and water table (late neolithic chalcolithic middle iron age late antique early medieval)

bull drying out of the soils contemporaneous with incision of the rivers and a deficit in the annual water balance (earlymiddle mesolithic late bronze age late iron age)

bull 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)


Figure 181 The middle and lower Rhocircne valley southern France showing the progress of Roman colonization

It is thus not enough to present the long-term evolution of landscapes under changingclimatic conditions as lsquowetrsquo and lsquodryrsquo lsquofavourablersquo or lsquounfavourablersquo 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 seventhmillennium BP This first erosive cycle did not occur in an environment that washomogeneously subject to excess human pressure Where extensive degraded surfacesoccurred they were due to a combination of naturally unstable or metastable landscapesin the hills and the destabilization of vegetation on the colluvial deposits through humanpressure Second important demographic increases and extensive exploitation of the

Desertification land degradation and land abandonment 329

basin are not always linked to erosionmdashhuman impact and climate are regularly out ofstep 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 lsquofragilizationrsquo with lsquodelayed responsersquo brought a gradual change in the long-term dynamics whereas in the early Holocene erosion occurred only when climatic andanthropogenic dynamics were pushing in the same direction the slightest climatic oranthropogenic event today can cause erosion The cumulative effect of ten millennia ofsocio-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 hasbeen responsible for the fragility of many southern European landscapes that werebrought under anthropogenic influence relatively early and implies that the area wouldsuffer badly if a climatic oscillation even a minor one were to occur today That willneed to be taken into account in assessing the effects of potential climate changespredicted by Global Climatic Models Finally it explains why every year many moreacres of land are lost to agriculture by land abandonment than by degradation ordesertification the countryside cannot sustain the absence of human interference anymore than it can sustain excessive exploitation A comparison of the pedogenesiserosioncurve for the Valdaine with climatic indicators such as oxygen isotopes alpine glaciermovements and subalpine lake hydrology points not only to an overall correlation butequally to the urgent need to base our assessment of the impact of global change onregional research The complexity of the dynamics governing the European climatemakes this all the more important

SOCIO-NATURAL INTERACTIONS IN THE ROMAN PERIOD

The second axis of the archaeological research was as has been indicated spatiallyoriented with a focus on Roman settlement in the middle and lower Rhocircne basin We selected the Roman period for four reasons it represents between the Neolithic and thesub-recent period the principal period of long-term demographic expansion in the area itencompasses a complete cycle of socio-natural interaction from colonization toabandonment including the economic crisis that is regarded as having afflicted much ofthe empire in the second and third centuries AD it resembles our own epoch in that itconfronts an urban perspective driven by organizational rationalization with landscapesthat it does not have any experience with and finally there is an extraordinary wealth ofarchaeological and written data available We took as our basic premise that thefoundation of a settlement represents a spatial choice and thus an assessment of thelandscape 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 statisticallysignificant 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 economicactivities undertaken This resulted in a chronology of settlement foundations andabandonments (Fig 182) that allowed us to map the colonization of the lower andmiddle Rhocircne valley by the Romans (Fig 181) We then endeavoured to reconstruct the


ancient landscape

Figure 182 Settlement trends in the middle and lower Rhocircne valley 50 BCmdashAD 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=Haut-Comtat D=Valdaine L=Lunellois T=Tricastin U=Uzegravege V=Vaunage

combining variables dependent on relief (altitude slope slope orientation and receptionof solar energy) with the distance from a site to the road system andor to open waterThough 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 toboth the well-drained valley floors and the higher slopes and finally to the valley bottoms


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 agronomistColumella which is based on ease of tilling rather than fertility by comparing it with theresults of a semantic analysis in Pliny the Elderrsquos De Re Agraria of all associations between a word for soil the adjectives accorded to it and the plants mentioned asfavouring this soil Combining the results with existing soil maps we roughed outlsquoreclassifiedrsquo soil maps Comparison with a rare documentmdashthe Roman tax and property map known as the Cadastre B of the city of Orangemdashenabled us to establish that the relative agricultural suitability thus derived from the agronomists coincided quite wellwith the relative tax assessments of plots in the Tricastin the area north of OrangeMoreover the fact that valleys requiring drainage were among the last zones to be settledconcurs with Columellarsquos comments that such locations were the least favoured In a final multivariate analysis the archaeological landscape and pedological data were combinedto give us a sense of the principal socio-natural categories of settlements in our sample

The lsquocrisisrsquo of the second and third centuries AD

Having thus detailed the natural conditions of various aspects of the Roman colonizationof the valley we focused in particular on the lsquocrisisrsquo of the second and third centuries AD In the literature this crisis is ascribed to a wide range of causes from saturnism toinvasions by Asiatic horsemen and from bad government to lsquothe environmentrsquo We first investigated whether there was any correlation between the numbers of sites abandonedat that period and their environments As Figure 183 shows however there clearly is none the many sites abandoned towards the end of the second century are randomly andequally distributed over different landscapes and soils Moreover the slight increase inprecipitation at the time is documented only for the alpine climate system that feeds theRhocircne rather than for the Mediterranean system that governs the local precipitationSome lands along the banks of the Rhocircne were thus reclaimed by the river but in most of the sample areas we have no important traces of increased erosion Finally the sitesabandoned in any area are the smallest ones which were last established whereas theoverwhelming majority of the early sites seems to have outlived the problems probablybecause they were located in the best spots and well-connected to the road networkmdashmost of them are situated


Figure 183 The persistence of settlements in the middle and lower Rhocircne valley through different occupation periods (Occ 1ndash6) of 100 years in each of 11 categories of environment (Rel 0ndash10) (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 70ndash80 per cent of sites do not survive more than two centuries

at crossroads Tentatively therefore we explain the lsquocrisisrsquo as far as our region is concerned in terms of a restructuring of the exploitation system Was there any loss ofresilience 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 processsufficient information to deal with the dynamics that it encounters it is the ubiquity ofthings going wrong that is characteristic of such a crisis as seems in fact evident in thewritten 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


map enabled us to investigate such issues in more detail Here in the first century BCthe emperors instigated a large (10000 km2) drainage scheme so that land holdings could be allocated to retired army veterans who as smallholders could both ensure the peaceand maintain the drainage system It is clear from the Cadastre B map which dates to AD77 that by that time large parts of the area still remained unoccupied probably becausethe principal lsquopeace dividendrsquo of the time was a reduction in the number of legions As the Roman drainage system was oriented northmdashsouth at an angle of 45 degrees to thenatural drainage the lack of maintenance rendered the huge system dysfunctionalpromoting erosion and seriously compromising agriculture In this area thereforeeconomic crisis seems closely connected to earlier imperial megalomaniamdashthe emperors went lsquoa drain too farrsquo

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 thelsquoguesstimatedrsquo relative carrying capacity for each landscape unit the probability of newsettlements in the different landscape units in the next period The resulting maps (Fig 184) show some interesting patterns especially towards the end of the Roman periodwhen there is an increase in new settlement foundations in areas abandoned less than acentury 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 settlementrsquos location relative to other settlements in maintaining inter-settlement dynamics the dependence of individual settlements on the others in their neighbourhood theirresilience and so on To inform this thinking we tried to define the resilience ofindividual towns in the middle and lower Rhocircne valley in the modern period (betweenc1800 and the present) on the basis of census data to investigate how far the observedloss of resilience of the rural areas has been tied to the dynamics of the urban system Wefound through a series of statistical operations (cf ARCHAEOMEDES 1998) that thisloss of resilience depends in part on the local resources and accessibility of thesettlement but equally if not more so on its population profile (age professonaldiversity and so on) and its relative position in the urban hierarchy The latter expressesthe settlementrsquos size and also the number of functions it fulfils which in turn is related to the settlementrsquos attractiveness for people in the surrounding areas However althoughthese factors together define the lsquodynamismrsquo of the settlementmdashits capacity to achieve things (and thus to adapt) to attract new inhabitants and so onmdashthe potential to use them for predicting the viability of individual settlements is limited for two reasons First theposition of individual settlements was considered relative to the whole of the settlementsystem whereas the interaction between local neighbourhood and more distant dynamicsis an important determinant for a settlementrsquos chances for survivalmdasha multi-scalar approach is thus required Second the lsquorecentrsquo period we used is relatively short whenviewed against the slow dynamics of settlement systems allowing us to monitor only partof a single cycle of such dynamics


Figure 184 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 ANALYSIS OF LONG-TERM TRENDS

The second phase of the project (1995ndash6) developed as a study of settlement systems in southern France over the last 2000 years combining elements of the archaeological andgeographical approaches used already despite the difficulties inherent in working witharchaeological historical cartographic and demographic (census) data simultaneously Inthis study we tried to interpret the social dynamics of rural and urban interactions Thecore questions asked were

After ensuring that the different categories of data could be used as equivalent indicatorsof the same processes (a major challenge) we defined theoretical lsquobasins of attractionrsquo for the settlements of the LunelloisVaunage area in the western lower Rhocircne (Fig 181) over some twelve centuries at the beginning of our era based on the settlement hierarchyand on a gravity model of spatial interaction Testing them against the archaeological datagave a rather good fit Moreover the shape of these basins turns out to be related to fossilaspects 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 butthat the village structure is essentially medieval The overall spatial configuration and themain anchor points are spatially stable Neither colonization wars political disasters norepidemics have fundamentally changed the spatial organization of the area because theyoperate 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 theseroads At each spatial scale however one can observe different irreversible structuralchanges as a result of gradual processes At the micro-scale for example the iron age lsquooppidarsquo settlements of Ambrussum and Mauressip were replaced in the second or thirdcentury by Lunel Viel and Calvisson but Sommiegraveres was stable for 2000 years until the last few decades of the twentieth century At the local scale the development of tourismin 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 Beacuteziers and Nicircmes) but as soon as those three poles came into direct competition in the twentieth

bull once a settlement system is established what happens to it bull how far does the settlement pattern determine the further development of the

landscape bull if it does is that an incremental process or are there phases of sudden

transformation bull what determines spatial choices in a landscape where the main resources have been

identified bull how is settlement structure affected by demographic changes bull what are the primary factors determining the success (or failure) of an individual

settlementmdasheg the environment the transportation network the density of pre-existing settlement


century the equilibrium changed in the favour of Montpellier At the scale of the regionas a whole the following transformations can be noted first in the fifth century ADtowns lost control over their hinterlands probably due to the fact that an excessive degreeof centralization of power in the cities is not sustainable the resulting fragmentationcontinued until the twelfth and thirteenth centuries when a new and much lesshierarchical urban system took over with multiple links between all three levels thestructure is presently undergoing another fundamental transformation under the impact ofthe formation of mega-cities Over the centuries the dynamism of the overall settlementsystem has evolved in tandem with the intensity of an individual settlementrsquos 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 thesystem for a better understanding of a settlement system it is essential to take in terms ofall dimensions (time size and space) the whole range of scales into account and not limitoneself to the top and middle tiers of the settlement hierarchy The model of settlement asa 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 notonly lsquotop-downrsquo but equally lsquobottom-uprsquo It is obvious that we have to take the rural environment into account from the perspective of both population system and resourcesThe study of the changes occurring in the links between these settlements is still to beundertaken and they may well be more frequent than changes in the settlement structureitself

In the most recent (1996ndash9) phase of the project we have attempted to take these lessons into account linking three levels of investigation of modern-day urbanmdashrural dynamics in southern France (Fig 185) covering the study area as a whole that is theregions of Midi-Pyreacuteneacutees and Languedoc-Roussillon as well as adjacent parts ofProvence-Alpes-Cocircte drsquoAzur three areas composed of one or two departmentsmdashAveyron-Lozegravere Heacuterault-Gard and Comtat and a micro-region the Causse Meacutejan where we looked at the population and settlement dynamics of all individual communesincluding individual households over about the last fifty years in combination with theiruse of rural resources

In this research we have chosen an approach based on the following working hypotheses

bull the settlement structure reflects information processing rather than energy processing (contrary to the traditional tenets of urban studies and archaeology)

bull the information flows go up and down the hierarchy so we must approach our analysis from both the top-down and the bottom-up


Figure 185 The three levels of the investigation into modern-day urban-rural dynamics in southern France

This led us to choose the following proxy measures for aspects of these informationflows 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 asthe proxy for rate of information processing Over and beyond these of course we tookspatial variables into account (location environment accessibility) as well as resources(land use hydrology) and forms of resource exploitation (organization and structure offarms)

Much remains to be done and the following conclusions are both partial and

perspectives (the former is common in geography we have here focused on the latter)

bull innovation drives the system (Gueacuterin-Pace 1993) bull scaling (the rank-size distribution) should conform to the intensity of information

flows


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 settlementson the basis of spatial context (Fig 186) and to view the settlement system as a nestedset of interaction zones in which the dynamic effects of equivalent settlements aredifferent according to spatial level and scale (Fig 187) Taking this perspective enabled us to understand some of the variables and their thresholds and interactions Tounderstand

Figure 186 Schematic representation of the way we have constituted the relations between cities (ie 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 dynamics we must determine the quantitative effect of the mix of dimensions ofinteraction for example home-to-work commuter patterns follow a different dynamic from shopping or schooling interactions and tourism retirement and active populationsalso operate differentially On the other hand we must combine the demographicdynamics with the institutional and the agricultural dynamics administrative and civilservice jobs for example have other dynamic prospects than the private sector industryis in this respect different from services and so on lsquoHeritagersquo effects (differences in flexibility between matter energy and information flows) are much more important

Figure 187 Schematic representation of the differences in context occurring among towns of similar andor different sizes in the Haut Comtat


than first assumed a detailed comparison of the Comtat in the 1870s and 1970s forexample 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 changeswhereas in the 1970s the same inherited spatial infrastructure hampered change

CONCLUSION

This chapter has surveyed the methodological development of the ARCHAEOMEDESproject and summarized some of its findings regarding the co-evolution of social and natural dynamics in the middle and lower Rhocircne valley over the past 2000 years andhow 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 todeal together with the very long term including the present The fact that such a self-evident approach is not more widespread seems at least in part due to the fact that thestudy 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 traditionie between nature and culture (Table 181) and between ways to view the past and waysto view the present (Table 182) Another but more common opposition is the inevitableone between narrow-focus analysis and broad-focus integrative research (Table 183) 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 self-imposed constraints on the kinds of

Table 181 Evolution of the lsquonaturemdashculturersquo debate over the last thirty years Pre-1980s 1980s 1990sCulture is natural Nature is cultural The relationship is dualistic Humans are reactive to the environment

Humans are proactive in the environment

Humans are interactive with the environment

Environment is dangerous to humans

Humans are dangerous for the environment

Neither are dangerous if handled carefully both if that is not the case

Environmental crises hit humans

Environmental crises are caused by humans

Environmental crises are caused by socio-natural interaction

Adaptation Sustainability Resilience Apply technofixes No new technology Minimalist balanced use of

technology lsquoMilieu perspective dominates

lsquoEnvironnementrsquo perspective dominates

Attempts to balance both perspectives


issues and questions a community of scholars is interested in maintained by a number ofsocial and educational techniques Few researchers are comfortable with the recognition

Table 182 The different approaches of the historical and natural sciences to the reconstruction of the past

Historical approach Evolutionary approachInterest in past Interest in present Understanding of the present based on the past

Understanding of the past based on the 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 183 The opposition between analytical and integrative approaches in research

Attribute Analytical approach Integrative approachPhilosophy bull narrow and targeted bull broad and exploratory bull disproof by experiment bull multiple lines of converging

evidence bull parsimony the rule bull requisite simplicity the goal Perceived bull biotic interactions bull biophysical interactions organization bull fixed environment bull self-organization bull single scale bull multiple scales with cross scale

interactions Causation bull single and separable bull multiple and only partially

separable Hypotheses bull single hypotheses and nulls bull multiple competing hypotheses bull rejection of false hypotheses bull separation among competing

hypotheses Uncertainty bull eliminate uncertainty bull incorporate uncertainty Statistics bull standard statistics bull non-standard statistics bull experimental bull concern with Type I error bull concern with Type II error Evaluation goal

bull to reach ultimate unanimous agreement

bull to reach a partial consensus

The danger bull exactly right answer for the wrong question

bull exactly right question but useless answer

Source After Holling 1998


that what they discuss are subjective opinions concerning objective results obtained asanswers to equally subjective questions If we are to progress towards a more holisticperspective on socio-natural interactions over the long term it is essential that thedifferent disciplines together define the questions that they will address and that these arenot 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 subjectmdashdesertificationmdashthat 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 upour disciplinary kitchens to each other we have drastically changed our perspectives inthree ways First the team has evolved its perspective on the role of people in theirenvironment from seeing it as reactive via proactive to interactive Second the focus haschanged through the project from desertification to degradation to abandonment asinterest shifted increasingly (and necessarily) to the social dynamics Finally our centralconcepts used to define socio-natural co-evolution have evolved from adaptation to sustainability to resilience This was not achieved overnightmdashtrans-disciplinarity is shedding blood sweat and tears togethermdashbut the results of the project convince us thatthe exercise has been well worthwhile

ACKNOWLEDGEMENTS

This chapter summarizes the work of many people in the ARCHAEOMEDES team inparticular J-FBerger J-LFiches MGazenbeek JJGirardot H Mathian DPumainLSanders PhCour F-PTourneux Ph Verhagen ITounsi CJung ThOdiot GChouquer CRaynaud SThieacutebault and FMagnin The project is co-ordinated by FFavory and Svan der Leeuw

REFERENCES

ARCHAEOMEDES (1998) Des Oppida aux Meacutetropoles Paris Anthropos-Economica Fantechi R and Margaris NS (1986) (eds) Desertification in Europe Proceedings of

the Information Symposium in the EEC Programme on Climatology Held in MytileneGreece 15ndash18 April 1984 Dordrecht DReidel Publishing Company

Favory F and van der Leeuw SE (1998) ARCHAEOMEDES La dynamique spatio-temporelle de lrsquohabitat antique dans la valleacutee du Rhocircne bilan et perspectives Revue Archeacuteologique du Narbonnaise 31257ndash98

Favory F Girardot J-J van der Leeuw SE Tourneux F-P and Verhagen Ph (1994) Lrsquohabitat rural remain en basse valleacutee du Rhocircnemdashde lrsquoutilisation de la teacuteleedeacutetection et des SIG en archeacuteologie Nouvelles de lrsquoArcheacuteologie 5746ndash9

Gueacuterin-Pace F (1993) Deux Siegravecles de Croissance Urbaine La Population des VillesFranccedilaises de 1831 agrave 1990 Paris Anthropos-Economica

Holling CS (1998) Two cultures of ecology Conservation Ecology 2(2)4 [online]


van der Leeuw SE (1998a) The ARCHAEOMEDES ProjectmdashUnderstanding the Natural and Anthropogenic Causes of Land Degradation and Desertification in theMediterranean Basin Luxemburg Publications Office of the European Union

van der Leeuw SE (1998b) La nature serait-elle drsquoorigine culturelle Histoire archeacuteologie sciences naturelles et environnement In ADucros JDucros and F Joulian (eds) La Culture Est-Elle Naturelle 83ndash98 Paris Editions Errance

van der Leeuw SE Favory F and Fiches J-L (in press) Archeacuteologie et Systegravemes Socio-Environnementaux Etudes Multiscalaires sur la Valleacutee du Rhocircne dans le Programme ARCHAEOMEDES Valbonne CRA-CNRS


Index

Entries in bold denote figures Abu Hureyra 70 Achaemenid empire

See Turkmenistan ad-Diyatheh 88ndash92 90 95 99 100 field systems 90 settlement 89 90 topography 99

Adi Ainawalid 181ndash90 192 ndash4 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) 19ndash20 intensive farming in 252ndash64 Mama Issara case study 256ndash9 261 263 264 Marakwet case study 210 212 254 255 256 259ndash61 263 264 temperature index 19 23 topography 19 26 34 52 eastern and southern Africa 254 261 southern Africa savanna zone 27 37 237 252ndash3 Southwest Africa 34 Sub-Saharan Africa 9 37 171 233 Tripolitania 137ndash56 See also Fezzan

agrarian development models of 7 ndash8 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 92ndash6 98 99 100

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 246ndash7 stall-feeding 148 155ndash6 207ndash8 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 27ndash9 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 109ndash12

evidence 171 244

Index 347

See also plants archaeology of drylands themes 3 ndash ARCHAEOMEDES Project 341 archaeozoology 145 architecture 48 53 70 73 97 101 183 232 Argolid 341 arid zones 3 13 18 19 56 128ndash30 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 24ndash7 See also dessication drylands

Arizona 276 279 ar-Risha al-Fawq 98 ar-Risha al-That 98 Asia 8 30

southeastern 262ndash3 Southwest 26 107

Atlantic Ocean 35 292 atmospheric processes 6 22ndash33

See also climate winds

Atrek River 106 Aubreville A 10 Auregraves Mountains 129 Ausserburg 331 Australia 11 20 26 36 Avdat 48 51 53 56 Aveyron-Lozegravere 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 68ndash9 82 87 92 100

settlements 81 98 See also pastoralists

Beer-Sheba 54 55 Beidha 70

Index 348

Ben-Gurion D 58 Beni Abes 21 Beni Ulid 138 142 147 150 Berbers 134 Bernese Oberland 322 beverages 212 295 296 310 Beacuteziers 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 321ndash37

abandonment of 332 definition of 319ndash20 environmental impact of 328ndash32 maintenance of 321 technology of 320ndash1 332ndash3 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 235ndash49 255

agropastoralism Motloutse River 234ndash8 climate 233 236ndash7 245 Khami settlement 234 236 237 238ndash44 population density 235ndash6 precipitation levels 235ndash6 245

Bourgogne King of 326 Brazil 36 bridges 130 134 Brig 325 British-Russian-Turkmenian excavations 107 Bryan RK 147 buildings

huts wattle and daub 241 mudbrick 166 public 172 stone 99

burials 58 109 172 burial cairns 98 167 172

Index 349

burial inscriptions 97 Burkino Faso Yatenga plateau 11 burning 148 180 212 244

See also fire Butua 244 ndash39 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 108ndash9 173 214ndash5 216 217 219 264 277 285

underground 90 168 See also bisses

canals bronze age 113 construction 205ndash7 209ndash10 212ndash13 216 iron age 111 Roman 12 90ndash2

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 Meacutejan 351 cave deposits 56 148 cemeteries 83 168 173 Central America 2 Central Asia 104 ndash20 centralized authority 10 119 262ndash5 350 ceramics 55 106 109 153 172

See also pottery cereals 142 148 157 172 324

Index 350

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 290ndash1 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 191ndash2 wild 69 cultivation 70 79 82 91ndash2 98 99ndash100 107 147 150 153 155 167 187ndash8 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 78ndash80 81 262 See also canals drainage

lsquochaos theoriesrsquo 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 ndash80 lsquochottrsquo See salt Christianity 48 57 churches 49 53ndash4 190 cisterns 73 75 77 143 154 clay 130 183 225 231 climate

catastrophic 13 climatic change 6 67 75 81ndash2 climatic optimum period 127 Global Climatic Models 38 339 global warming 4 11 See also anticyclones Charney effect El Nintildeo evapotranspiration humidity hyperarid zones Inter Tropical Convergence Zones Little Ice Age monsoon rains paleoclimatology precipitation

Index 351

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 181ndash3 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 225ndash7 229 231 ndash2 cultural changes 12 15 67 currents

Peru current 26 Benguela current 26 37 Humboldt current 37

Index 352

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 274ndash6 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 11ndash12 34ndash5 area of 12ndash14 36 humanly-induced 11 12ndash14 36 79 154 162 rate of 11 recent 12

deserts 20 33 275ndash6 284 agriculture decline of 45ndash59 Atacama 20 26 Chihuahuan 271 273 277 Kara Kum 104 106 107 113 Libyan 162ndash5 margins 8 36 38 Nafudh 86 Namib 26 34 Near Eastern 57ndash8 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 6ndash7 35 143 163 165 166 See also aridification

development aid 11

Index 353

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 225ndash6 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 224ndash6 230ndash1 343 See also canals irrigation

droughts 5 7 9 10 11 12 13 20 28 34ndash6 45 142 144 150 155 184 187 195 133 235 248 276

definition of 23 192 dry farming (without irrigation) 73ndash5 87 drylands

archaeological evidence 12 14 archaeology difficulties of 15 biophysical systems 143 144 climatology of 6 19ndash38 233 coastal 24 definition of 19ndash21 diversity of 8 farming 7ndash8 233ndash4 fragility of 14 339 geographical extent of 3ndash4 19 20 36 human perceptions and decisions impact of 4 9ndash10 16 244ndash7 resilience of 4 risks and opportunities 5 sustainability of 3 5 temperate 21 22 themes 3ndash16 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 33ndash4

Index 354

Rift Valley 201 202 203 204 209 210 215 216 217 261 East Pokot people 265 ecology 3 13 economic conditions 8 75 100ndash 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 Nintildeo Southerly Oscillation (ENSO) 35 ndash6 elites 73 142 156 172 239 255 266 292 299 310ndash1

See also hierarchies Elusa 48 54 56 emmer See cereals Empordagrave 341 Enderta administrative region 181 Energy Dispersive X-Ray Microanalysis

See EDMA Engaruka River 203 206 214 216 218 Engaruka 203ndash19 256 258

climate 201 215 field systems abandonment of 210 211 215ndash18 irrigation systems 205ndash13 population 209 213 precipitation 211 214 215 topography 201ndash4

ENSO See El Nintildeo 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 150ndash1 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 181ndash95 229 257 agriculture 180 186ndash94

Index 355

climate 180 192 194 land-holding 181ndash2 Northern Highlands 10 population 193ndash4 precipitation 34 186ndash7 lsquovillagizationrsquo 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 56ndash7 195 258 260ndash 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 78ndash80 83 143 154 156 204ndash7

bronze age 73ndash4 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 252ndash64 337 irrigation-based 8 259 labour-intensive 231 2649 livestock 143 mixed 156 227ndash8 231 modern intensive mixed 7 70 75 141 145 sedentary 67 295 subsistence 9 57 100 103 107 152 211 233ndash4 264 surpluses 319 technology 145 165 wheat 10 See also animals cash-crop farming

Index 356

floodwater farming irrigation plants

farms 77 142ndash3 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 170ndash2 176

aims of 165ndash8 Fezzan the 12 150 162ndash77

classical 163 climate 164 169ndash70 early Holocene 162 164 169ndash70 early Islamic 168 early Modern 167 Garamantian 167 168 171ndash3 historical reconstruction 167ndash75 landscapes 162ndash5 late Pleistocene 163 169 later prehistoric and historic 163 medieval 167 mesolithic 163 neolithic 163ndash4 165 palaeolithic 163 population 162 171 172 precipitation 162 163 169 settlements 172 175 See also Sahara

fields clearance 90 91 280 divisions 72 209ndash10 213 216 217 maps 142 systems 51 67 72 73 74 76 78 81 89 90ndash2 98 99ndash100 205ndash13 220 253 274ndash6 surveys 142 walking 68 167

figs 57 138 145

Index 357

fire 283 286 suppression of 273 283 See also burning wood

fishing 7 70 142 flexibility 10 156 268 336 ndash7 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 277ndash8 See also irrigation risk management Tripolitanian pre-desert water run-off

flora 70 81 148 fogmist 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 gasrgsur forts 51 89 92 134 162 232

See also gasrgsur hillforts military structures

fowling 70 France 13

archaeologists 90 132ndash3 decolonization 125 Tricastin region 13 341 342ndash3 See also Rhocircne 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 ndash7 Garamantes people 162ndash3 172 173 175

development off oggara irrigation systems 13 168 172ndash3 174 175 early Islamic period 172ndash3 Garamantian civilizationculture 167 172 175

gardens See horticulture

Index 358

Gasr Lebr 153 gasrgsur (Romano-Libyan fortified farm) 140 142 147 148 153 156 173 194

agriculture 145ndash6 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 ndash9 geomorphology 67 70 138 157 165ndash6 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 GermaOld 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 359

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 185ndash6 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

Haiumldra 134 Haluza See Elusa Hamada al-Hamra 138 Hamada 87 handaxes 168 Hapex-Sahel 341 Harra plateau 87ndash9 90 92 100 Hauran 87 ndash9 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 Heacuterault-Gard 351 herding 70 76 107 146

See also grazing hierachies development of 8 108 172 265ndash8 294 301 309 310 348 350ndash6

See also elites

Index 360

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 278ndash9 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 9ndash10 11 14 45 64 70 100 144 156ndash7 162 163 165 176 219 259 285ndash6 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 stabilityinstability Inter Tropical Convergence Zones 19 20 26 27ndash8 36

See also climate inter-cropping

See cereals crops

Iran 20 107 113 Irano-Turanian desert steppe 45 Iraqw people 259ndash62 267

Index 361

irrigation systems 3 8 92 93 96 99 104 108 109 112 135 154 255 256 257 277 278 285 agriculture 56 103ndash20 112 259 maintenance of 114 115 119 212ndash13 spray 319 technology 113 115 212 See also bisses floodwater farming salinization

Islamic empire 47 56 57 Islamic pastoralist invaders 15ndash16

Isle of Braccedil 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 ndash9 Jebel 92 100 ndash Jeitun 105 107 ndash8 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 262ndash5

Machakos district 12 See also Africa

Kerio Valley 262 265 Khirbet el-Umbashi 89 Khirbet Faynan 64ndash5 70 71 76ndash80

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 362

Kurnub See Mampsis kushetts 183 Kwermusl 260 Kyzal Arvat 108

La Quemada 310 La Tegravene Iron Age 324 labour 181ndash91 194 215 256

communal 231 257 264 321 demands 57 145 168 226 233 mobilization of 231 256 263ndash4 tribute 231 263ndash4 306

Lake Eyasi 218 Lake Nasser 35 Lake Natron 218 219 lakes 7 70 164 165ndash6 168 169ndash70 147 341 344

See geomorphology sediments lacustrine soil

land abandonment 336 339 341ndash3 carrying capacity 13ndash14 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 ndash83 landscape degradation 11 55 58 80 115 157 235 249 258 341

human impact on 58 75 81ndash2 134 125 215ndash16 233ndash4 279ndash81 283 336 339 lsquomarginalrsquo 9

Languedoc-Roussillon 341 351 Lattimore O 310 lava 87 Le Houerou HN 34 36 lead 78 legumes 70 184 192ndash3 227 230 Lemasba 150 Letsibogo 236 ndash49 Levant 47 54 56 57 70 76

early Holocene 108 Levant 67 Lewis RA 107 109 Libya 8 92 131 146 150

foggaras 8

Index 363

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 207ndash8 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 LunelloisVaunage (western Lower Rhocircne) 350

Maasai tribespeople 203 207 218 219 268 Maasailand irrigation farming 203 ndash19 Macae tribespeople 142 156 157 Machakos 258 265 Madama 76 Maghreb 13 126

aggradation period 127 129ndash31 agriculture 128 climate 126 127 128 130ndash1 134 144 colonization Arab 133ndash4 Byzantine132ndash3 Phoenician 125ndash7 Roman 13 127ndash32 134 Vandal 132 133 conquests and land degradation 125ndash34 eastern 13 125ndash34 Holocene 125 incision period 127ndash8 Phoenicians 125ndash7 precipitation 13 127 131 134

Index 364

Roman colonization 13 127ndash32 134 settlements 130 Vandal colonization 132 133 See also ULVS

maguey 280 292ndash311 and seed cultivation 291ndash2 295ndash7 305ndash6 decline of 291 fibre use of 296ndash7 importance as food source 292ndash6 sap extraction 302ndash4 significance of as crop 290ndash2 technology of 292 297ndash304 292

Mahabere Genet 181 193 Mai Kayeh 186 maize See cereals Makuyuni river 203ndash5 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 266ndash7 303

expanding 144 154 211 320ndash1 loss of 142 Mekelle 181 184 192

marshes 165 Masson VM 107 109 Maungwe people 229 Mauretania 30 McGinnies WG 19 McIlveen R 27 meat 107 146 210 mechanization programmes 11 286 MEDALUS Project 341 lsquoMedieval Warm Epochrsquo 238

See also climate

Index 365

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 113ndash14 settlement pattern 111ndash14

Mesa Verde 278 283 Mesoamerica 277

tierra friacutea 288ndash308 civilization expansion of 292 299 305ndash8 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-Pyreacuteneacutees 341 350 migration 109 172 266

See also mobility settlement transhumance

military garrisonsstructures 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 ndash5 minerals 47 65

See also copper pollution

Index 366

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 Monteacutelimar 342 Montpellier 350 Morocco 135 150 155 mortar 71 mortgages 9 Mortimore M 11 255 mosaics 49 Moscow 104 115 mosques 54ndash5 58 Motloutse River 236ndash9 246 Mount Kilimanjaro 212 ndash3 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 ndash9

Index 367

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 45ndash59

Abassid 53 54 57 Byzantine 47ndash52 49 classical 54ndash5 56 climate 46 54ndash6 58 desertification 46ndash7 57ndash8 early Roman and Nabatean 47 55ndash6 63 65 Islamic conquests 52ndash4 58 population 45 57 precipitation 51 rise of desert 56ndash7 settlements 49 abandonment of 45ndash6 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 NiloticBantu dichotomy 15 Nicircmes 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 271ndash83 See also Arizona New Mexico Classic Mimbres period 280 281ndash2 population levels 273 280 281ndash2 283 precipitation levels 271ndash3 276 279 281ndash2 prehistoric agricultural practices 271 273ndash82 topography 271ndash3

Index 368

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 ndash8 OSL 67 166 108 Oumlstberg W 265 ostrich eggshells 168

shell beads 167 168 over-grazing 3 12ndash3 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 ndash8 paleoclimatology 150 155 247

See also climate paleoeconomic analyses 147 paleoenvironment 147ndash8 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 369

Pamirs mountains 106 Pare 212 ndash3 Parthian empire 112 pastoralism 12ndash3 203 255 292 310 323

pastoral encampments 51ndash2 72ndash3 92 98 pastoralists 15 16 92 98ndash9 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 187ndash8 remains 107 108 167 root 148 225 wild 184ndash6 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 57ndash8 114 143 156 173ndash5 194 240 255 285 292ndash3 346

See also instability stability state systems

political economy 265 ndash8 political stabilityinstability 12 119 156 182 195

Index 370

See also instability political systems 14 194 262ndash5 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 97ndash8 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 186ndash7

seasonality 20ndash1 summer rains 20ndash1 38 186 187 variability 22ndash3 30 winter rains 20 38 104 106 221 See also climate

predation 9 183 188 194 195 230 280 prestige items 310 ndash1 prickly pear 185 Provence-Alpes-Cocircte drsquoAzur 350 pulses 109 112 187 189 192ndash4 212 216 219 277 293 299 Pungwe River 222 Punic Wars 128

qanats 8 105 113 168 Qasr Burqursquo 89 96ndash8 99 ndash 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 371

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 356ndash7

See also ARCHAEOMEDES Project reservoirs 65 80 89 96 97 98 193 resettlement 9

compulsory 81 194 227 230 Rharsa chott 128 Rhocircne Valley 12 322 336 341ndash57

lower 337 338 settlement trends 340 341ndash6 Valdaine region 337 339 climate 337ndash9 Roman colonization 339ndash45 338 lsquocrisisrsquo period 341ndash5

Rio Bravo del Norte River 275 Rio Grande River 275 risk management 5 8 10 64 154 232 ritual 165 road networks 345ndash6 350 rock art 164 165 170 172

inscriptions 57 rock mulching 280ndash2 286 rock shelters 165 170 Roman empire 12 47 64 66 78 87 98ndash9 100 134 143 145 155ndash6 157 162 167 172 344ndash50

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 372

Sabi River 222 Safford Arizona 279 Sahara 12 19 21 30 113 131 132 172

culture 162 171 oases 160 precipitation 6ndash8 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 (lsquochottrsquo lsquosebkhsrsquo) 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 130ndash1 143 151 152 154 336

aggradation 126 127 129ndash31 133 149 fluvial deposits 107 incision 126 127ndash8 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

Index 373

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 12ndash13 86 late Byzantine 86 Libyan Pre-desert 12ndash13 Natufian 69 pastoral-nomadic 52 54 55 patterns of 4 67 247 252ndash64 67 Romano-Libyan 12ndash13 152 Southern Bantu 244 247

Shashe-Limpopo basin 236 247 ndash 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 PM 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 100ndash 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 230ndash1 255

Index 374

erosion 11 12 13 14 75 128 131 142 143 236 247 337 338ndash9 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 ndash8 Somalia 20 Sommiegraveres 350 Sonjo villages 212 215 219ndash 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 114ndash15 119 See also Turkmenistan

Spain 21 113 341 specialization 76 219 229 267 295 302 309ndash10 311 spices and herbs 185 spillways 78 96 spindle whorls 296 302 311

ceramicclay 243 298 300 302 spindles wooden 301 spinning and weaving 301ndash3

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 156ndash 268 349ndash50

See also environment and human activity land carrying capacity

state systems development of 108 309 ndash11

Index 375

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 205ndash7 209ndash10 211 217 226 230 238

storage facilities 142 232 246 247 grain storage 190ndash1 183 194 209 See also granaries water

streams 7 107 109 129 213 214 216ndash7 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 25ndash7 Sudan 168 Sudano-Sahelian depressions 33 34 Sultan Kala 112 summer rains See precipitation Summers R 230 Susiana 113 Sutton JE 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 86ndash100

Tahirbaj Tepe 112 Tahiti-Darwin Sea 35 Taita 213 Takhirbai 112 takyrs 108 Tanganyika British mandateship of 207 Tanzania 212 217 256 257 258ndash62 266

Mbulu District Northern Tanzania See Africa drylands See also villagization

Index 376

Tarahumara people 297 309 Tatoga tribespeople 218 Taurus Mountains 87 tax issues 145 175 324 Tejen River 106 107 108 Tejen 108 teleconnections 35 ndash6 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 222ndash5 228ndash9 230ndash1

hillslope terracing 56 maintenance 54ndash5 57 terrace erosion 54ndash5 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 DG and Middleton NJ 10 11 thorn scrub 18 153 212 238 240 248 Thornthwaite CW 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 193ndash4 Timbuktu 23 Timurid people 112 Togolok 112 tomatoes 184 tools 107 169 191 304

ceramic 297ndash8 flint 69 168 169 iron 302ndash3 scraping 293 300ndash4 stone 163 297ndash8

Tot 262 tourism 14 350 Toutswe people 238 towers 96 98 towns 45 48 56 133

Index 377

See also urbanism trade 57 142 156 162 166 172 266

routes 47 72 98 166ndash7 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 138ndash57

environment 146ndash50 population 141 154 155 156 precipitation 147 149 153 Romano-Libyan floodwater farming137 141 144 149ndash53 154 146 150 decline of 155ndash6 RomanomdashLibyan and later settlement 141 152 settlements abandonment of 142ndash4 147 154ndash5 settlements expansion of 144ndash5 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 129ndash31 134 Turkey 175 Turkmenistan 7 104ndash20

climate 106 107ndash8 109

Index 378

environment 104ndash6 irrigation agriculture 103ndash20 and see following periods Achaemenid period 9 104 111 chalcolithic 108 early bronze age 108 eneolithic iron age 108ndash11 middle bronze age 108 111 112 neolithic 106ndash8 Namazga III period 108 Parthian 111 112ndash13 Sasanian 112ndash13 Seleucid (Hellenistic) period 111 Seljukpost-Seljuk 112 Tsarist Soviet periods 9 11 114ndash15 population 103 109 113 119 precipitation 104ndash6 settlement 116 settlement abandonment 108 See also qanats

Tutsi people 14 Tyson PD and Lindsay JA 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 142ndash3 146 UNESCO 20 United States 36 114 265ndash6 304 310 Universal Transverse Mercator 68 universities 140 Unyama people 230 233 urbanism 108 112 115 165 166ndash7 172 173 175 277 325 353 354

urban decline 45ndash6 52 53 75 urbanrural interaction 345ndash50 urban systems 45 72 111ndash14 172 343 345ndash6 See also settlements towns

UTM See Universal Transverse Mercator

Valais canton 9 321ndash37 agricultural patterns 317ndash19 climate 317 320 population growth 320 precipitation 317 319 topography 315ndash17

Index 379

See also bisses Valdaine See Rhocircne 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 Vernamiegravege commune of 322 323 329

irrigation sectors of 326 327 See also Valais canton

Viegravege See Visp villages

Engaruka 207ndash8 215 Fezzan 167 168 172ndash3 174 Mesoamerican 273 Negev 53 Nyanga 229 Rhocircne Valley 345 Sonjo 209 210 217ndash18 218 Syrian Black Desert 88 92 99ndash100 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 Cheacuteria-Mezeraa 130 132 Wadi Dana 70 Wadi el Akarit 130 134 133 Wadi el-Agial (al-Hayat) 163ndash4 166 168 169 172 175 Wadi el-Amud 143 147 150 156 Wadi es Sgniffa 130 Wadi Faynan 7 9 10 64 ndash8364

bronze age 72ndash5 chalcolithic 65 lsquoclassicalrsquo 67 climate 70 field systems 65 66 68 77 Holocene 69 70 81

Index 380

iron age 75ndash6 Nabatean 77 81 neolithic 65 69ndash71 81 Pleistocene 69 population 82 Post-Byzantine 80ndash1 Roman Imperial 77ndash80 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 Keacutebir-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 ndash6 Wadi umm el-Kharab 140 143 150 154 wadis 12 45 55 70 73 77 78 89 138 143 151 154 164

downcutting 54ndash5 79ndash80 128 wadi floors 51 70 149ndash50 wadi walls 139 150 151ndash3 See also geomorphology ULVS

Waldmatte 325 walls 68 90 142 143 167 175 231 239

boundary 72 76 desert 151ndash3 functions 8 50 75 77 151 153 274 276 diversion 74 78 98 139 171 205 fortification 50 108 wall networks 137 151ndash4 maintenance of 155ndash6 See also floodwater farming geomorphology terrace systems walls

war 134 175 warlords 142 water

conflicts over 323 328 330ndash1 333 distribution 113 325ndash7 deterioration in 9 103 115 119ndash20 213ndash15 216 domestic uses 183 194 211 226 gauges 113

Index 381

harvesting 8 88 144ndash5 149 152 162 257 loss via seepage etc 78 99 115 mills 65 90 91ndash2 pumps diesel and electric 163 resource management 15 21 81 86 91 99 100 112 113ndash15 119ndash20 220 226ndash7 230 252 254 misuse of 12 115 175 rights 92 113 260 263 264 282 320 321ndash8 333 run-off 50 91 107 152 153 156 274 catchment systems 45 48 57 91 92 151ndash3 183 194 201 maintenance of 103 115 shortage 319 storage 98 99 100 152 153 supplies fossil 162 175 underground 149 163ndash4 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 ndash4 Weintraub P 311 Wello province (Ethiopia) 187 wells 90 138 164 168 175 184 237

artesian 137 174ndash5 West Africa 27 30 35

lsquoring cultivation systemsrsquo 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 25ndash6 East African Low Level Jet 32ndash4 Hadley Cell circulation 26 27 28ndash9 30 31 34 Subtropical Westerly Jet 31 trade winds 27 29 Tropical Easterly Jet 31ndash2 West African Mid-Tropospheric Jet 32 See also anti-cyclones climate

Index 382

cyclones hurricanes precipitation

wine 45 167 presses 45 51 47

winnowing 188 190 winter rains See precipitation winters 20 128 Wittfogel K lsquooriental despotismrsquo 264 ndash5 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 ndash5 World Meteorological Organization 21 Wright K 73 Wyckoff DG 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 238ndash9 247 248 Zimbabwe 212 238 247

Nyanga region 208 220ndash32 221 253 255 256 agriculture 227ndash8 landscape 221ndash7 229ndash31 population 229 230 231 232 precipitation 220 221 settlements 220 229ndash30 231 topology 220ndash1 RusapeHeadlands area 230

Zinchecra 168 172 Zionism 15 58 Ziwa 230

Ziwa ruins National Monument 229ndash30

Index 383

Zuila 175

Index 384


The One World Archaeology (OWA) series stems from conferences organized by theWorld Archaeological Congress (WAC) an international non-profit making organization which provides a forum of debate for anyone who is genuinely interested in or has aconcern for the past All editors and contributors to the OWA series waive any fees theymight normally receive from a publisher Instead all royalties from the series are receivedby the World Archaeological Congress Charitable Company to help the wider work ofthe World Archaeological Congress The sale of OWA volumes provides the means forless advantaged colleagues to attend World Archaeological Congress conferencesthereby enabling them to contribute to the development of the academic debatesurrounding 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












































Edited by


London and New York












British Library










Contents


Part I Introduction




3


44

4


62

5


85

6


101


7


121

8


133






13


227

14


246


15


264

16


280

Part VI Europe


18


327

Index 346

Figures


























Tables









Contributors






























Foreword






Preface







Part I INTRODUCTION



INTRODUCTION







Photograph GBarker













Photograph GBarker


























environment










change

CONCLUSION






REFERENCES


















































GREG SPELLMAN

DEFINING DRYLANDS









South America

Total


2687 352 513 1835 2315 2070 12947 99






Precipitation









Drought




Temperature



Relative humidity



Wind

































Trade winds

































OCEAN TEMPERATURES


DESERTIFICATION

















CONCLUSION




REFERENCES











































































ASIA



INTRODUCTION




































































FINAL NOTE





ACKNOWLEDGEMENTS


REFERENCES

































83350ndash5



















































GRAEME BARKER

INTRODUCTION







Photograph GBarker






METHODOLOGIES






































































CONCLUSION







ACKNOWLEDGEMENTS




REFERENCES




































INTRODUCTION









Ad-Diyatheh











Photograph PNewson







Al-Namara










Photograph PNewson





Qasr Burqursquo





Photograph BBewley








DISCUSSION












ACKNOWLEDGEMENT


REFERENCES














Press












INTRODUCTION





ENVIRONMENT




















































CONCLUSION





























Broomcorn millet









































REFERENCES
























































JEAN-LOUIS BALLAIS

INTRODUCTION


















An incision period





































CONCLUSION


REFERENCES
















































INTRODUCTION























Human processes












Natural processes


















Human processes















AGRICULTURE








ENVIRONMENT


Palaeoenvironment













Palaeoclimatology




Geomorphology


















Design principles






WALL FUNCTIONS












HUMAN AGENCIES








CONCLUSION


REFERENCES
























































INTRODUCTION






DESERT LANDSCAPES











































































ACKNOWLEDGEMENTS


REFERENCES










































































INTRODUCTION


FIELDWORK















Livestock




Plant resources










tselimoi



sindai





Cultivation systems



white) dagousha











Photograph AButler









Grain storage





Farming and fuel





CROP DIVERSITY












Emmer wheat ares


intercrop aterabie





CONCLUDING REMARKS







ACKNOWLEDGEMENTS



REFERENCES


















































AFRICA


















































































ACKNOWLEDGEMENT


REFERENCES

























INTRODUCTION







Terraces













Cultivation ridges















Water management




Authorship




























DISCUSSION







ACKNOWLEDGEMENTS


REFERENCES

















JOHN KINAHAN

INTRODUCTION

































































ACKNOWLEDGEMENTS


REFERENCES















































































INTRODUCTION
























CASE STUDIES
















Marakwet Kenya





Photograph MWidgren
























ACKNOWLEDGEMENTS


REFERENCES








































AMERICA


PAUL EMINNIS

INTRODUCTION












Irrigation


Floodwater farming








Photograph AHeisey


Rain-fed farming




Photograph PMinnis

Rock mulching






Photograph PMinnis






Deforestation






Increase in weeds



Soil modification









CONCLUDING THOUGHTS






ACKNOWLEDGEMENTS


REFERENCES

















































































INTRODUCTION









Photograph JParsons




































Maguey fibre










Spinning




Photograph JParsons















Photograph JParsons











Photograph JParsons











CONCLUSION




















REFERENCES

































































































Part VI EUROPE



INTRODUCTION



THE VALAIS




Climate















COPING STRATEGIES



(Crook 199778)










knowledge




Water rights






















5 12 8 2 4 1 12 10 12 8 2 2 3 9











Conflict resolution










1347 1367 1443 1554



Bishop of Sion









Arbaz v Grimisuat5



Bishop of Sion


13851390ndash1448






Tribunal

Bagnes v Levron7

1443ndash1465 1515 1545 1626 162930 1839 1923








Ayent v Sion9








solicitors


1989 to date






DISCUSSION



ACKNOWLEDGEMENTS



REFERENCES






















































125ndash31




INTRODUCTION
























ancient landscape











































flows










CONCLUSION




































ACKNOWLEDGEMENTS


REFERENCES












Index
















evidence 171 244

Index 347











Index 348










Index 349










cats 190 cattle 57 90 184 203 207 218 219 248 277 294




Index 350








Index 351











Index 352









development aid 11

Index 353



animal 10 194 217 human 10 114 143 154 192 194




droughts 5 7 9 10 11 12 13 20 28 34ndash6 45 142 144 150 155 184 187 195 133 235 248 276






plateau 33ndash4

Index 354











Index 355








Index 356









figs 57 138 145

Index 357

















Index 358







Index 359






grinding 91 gsur







See also elites

Index 360






See also climate






See cereals crops


Index 361











Index 362











foggaras 8

Index 363








Index 364









See also climate

Index 365














Index 366





Index 367








Index 368


oases 106 150 162 166 167 169




oil 12 141 144 167 presses 51 146





See also animals







Index 369








Index 370











radiation budget 6



Index 371











Index 372












Index 373









Index 374












Index 375











Index 376







Index 377






European 86 174 229 trees 132 138 183 184 191 258 342










Index 378








Index 379











Index 380




Waldmatte 325 walls 68 90 142 143 167 175 231 239




Index 381







Index 382


wine 45 167 presses 45 51 47










Index 383

Zuila 175

Index 384












































Edited by


London and New York












British Library










Contents


Part I Introduction




3


44

4


62

5


85

6


101


7


121

8


133






13


227

14


246


15


264

16


280

Part VI Europe


18


327

Index 346

Figures


























Tables









Contributors






























Foreword






Preface







Part I INTRODUCTION



INTRODUCTION







Photograph GBarker













Photograph GBarker


























environment










change

CONCLUSION






REFERENCES


















































GREG SPELLMAN

DEFINING DRYLANDS









South America

Total


2687 352 513 1835 2315 2070 12947 99






Precipitation









Drought




Temperature



Relative humidity



Wind

































Trade winds

































OCEAN TEMPERATURES


DESERTIFICATION

















CONCLUSION




REFERENCES











































































ASIA



INTRODUCTION




































































FINAL NOTE





ACKNOWLEDGEMENTS


REFERENCES

































83350ndash5



















































GRAEME BARKER

INTRODUCTION







Photograph GBarker






METHODOLOGIES






































































CONCLUSION







ACKNOWLEDGEMENTS




REFERENCES




































INTRODUCTION









Ad-Diyatheh











Photograph PNewson







Al-Namara










Photograph PNewson





Qasr Burqursquo





Photograph BBewley








DISCUSSION












ACKNOWLEDGEMENT


REFERENCES














Press












INTRODUCTION





ENVIRONMENT




















































CONCLUSION





























Broomcorn millet









































REFERENCES
























































JEAN-LOUIS BALLAIS

INTRODUCTION


















An incision period





































CONCLUSION


REFERENCES
















































INTRODUCTION























Human processes












Natural processes


















Human processes















AGRICULTURE








ENVIRONMENT


Palaeoenvironment













Palaeoclimatology




Geomorphology


















Design principles






WALL FUNCTIONS












HUMAN AGENCIES








CONCLUSION


REFERENCES
























































INTRODUCTION






DESERT LANDSCAPES











































































ACKNOWLEDGEMENTS


REFERENCES










































































INTRODUCTION


FIELDWORK















Livestock




Plant resources










tselimoi



sindai





Cultivation systems



white) dagousha











Photograph AButler









Grain storage





Farming and fuel





CROP DIVERSITY












Emmer wheat ares


intercrop aterabie





CONCLUDING REMARKS







ACKNOWLEDGEMENTS



REFERENCES


















































AFRICA


















































































ACKNOWLEDGEMENT


REFERENCES

























INTRODUCTION







Terraces













Cultivation ridges















Water management




Authorship




























DISCUSSION







ACKNOWLEDGEMENTS


REFERENCES

















JOHN KINAHAN

INTRODUCTION

































































ACKNOWLEDGEMENTS


REFERENCES















































































INTRODUCTION
























CASE STUDIES
















Marakwet Kenya





Photograph MWidgren
























ACKNOWLEDGEMENTS


REFERENCES








































AMERICA


PAUL EMINNIS

INTRODUCTION












Irrigation


Floodwater farming








Photograph AHeisey


Rain-fed farming




Photograph PMinnis

Rock mulching






Photograph PMinnis






Deforestation






Increase in weeds



Soil modification









CONCLUDING THOUGHTS






ACKNOWLEDGEMENTS


REFERENCES

















































































INTRODUCTION









Photograph JParsons




































Maguey fibre










Spinning




Photograph JParsons















Photograph JParsons











Photograph JParsons











CONCLUSION




















REFERENCES

































































































Part VI EUROPE



INTRODUCTION



THE VALAIS




Climate















COPING STRATEGIES



(Crook 199778)










knowledge




Water rights






















5 12 8 2 4 1 12 10 12 8 2 2 3 9











Conflict resolution










1347 1367 1443 1554



Bishop of Sion









Arbaz v Grimisuat5



Bishop of Sion


13851390ndash1448






Tribunal

Bagnes v Levron7

1443ndash1465 1515 1545 1626 162930 1839 1923








Ayent v Sion9








solicitors


1989 to date






DISCUSSION



ACKNOWLEDGEMENTS



REFERENCES






















































125ndash31




INTRODUCTION
























ancient landscape











































flows










CONCLUSION




































ACKNOWLEDGEMENTS


REFERENCES












Index
















evidence 171 244

Index 347











Index 348










Index 349










cats 190 cattle 57 90 184 203 207 218 219 248 277 294




Index 350








Index 351











Index 352









development aid 11

Index 353



animal 10 194 217 human 10 114 143 154 192 194




droughts 5 7 9 10 11 12 13 20 28 34ndash6 45 142 144 150 155 184 187 195 133 235 248 276






plateau 33ndash4

Index 354











Index 355








Index 356









figs 57 138 145

Index 357

















Index 358







Index 359






grinding 91 gsur







See also elites

Index 360






See also climate






See cereals crops


Index 361











Index 362











foggaras 8

Index 363








Index 364









See also climate

Index 365














Index 366





Index 367








Index 368


oases 106 150 162 166 167 169




oil 12 141 144 167 presses 51 146





See also animals







Index 369








Index 370











radiation budget 6



Index 371











Index 372












Index 373









Index 374












Index 375











Index 376







Index 377






European 86 174 229 trees 132 138 183 184 191 258 342










Index 378








Index 379











Index 380




Waldmatte 325 walls 68 90 142 143 167 175 231 239




Index 381







Index 382


wine 45 167 presses 45 51 47










Index 383

Zuila 175

Index 384