bogaard, amy - neolithic farming in central europe, an archaeobotanical study of crop husbandry...

dxt dxt

iNEOLITHIC FARMING INCENTRAL EUROPE

Neolithic Farming in Central Europe examines the nature of the earliest cropcultivation, a subject that illuminates the lives of Neolithic farming familiesand the day to day reality of the transition from hunting and gathering tofarming.

Debate surrounding the nature of crop husbandry in Neolithic centralEurope has focused on the permanence of cultivation, its intensity and itsseasonality, variables that carry different implications for Neolithic society.Amy Bogaard reviews the archaeological evidence for four major competingmodels of Neolithic crop husbandry shifting cultivation, extensive ploughcultivation, oodplain cultivation and intensive garden cultivation andevaluates charred crop and weed assemblages. Her conclusions identify themost appropriate model of cultivation, and highlight the consequences ofthese agricultural practices for our understanding of Neolithic societies incentral Europe.

Amy Bogaard is Lecturer in Archaeological Science at the Departmentof Archaeology, University of Nottingham. Her main research interests areearly farming practices and archaeobotany.

iii

NEOLITHICFARMING IN

CENTRAL EUROPE

An archaeobotanical study of crophusbandry practices

Amy Bogaard

iv

First published 2004by Routledge

2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN

Simultaneously published in the USA and Canadaby Routledge

270 Madison Avenue, New York, NY 10016

Routledge is an imprint of the Taylor & Francis Group

2004 Amy Bogaard

All rights reserved. No part of this book may be reprinted orreproduced or utilised in any form or by any electronic, mechanical, orother 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 DataA catalogue record for this book is available from the British Library

Library of Congress Cataloging in Publication DataBogaard Amy, 1972

Neolithic Farming in Central Europe: an archaeobotanical study ofcrop husbandry practices.

p. cm.Includes bibliographical references.

1. Neolithic periodEurope, Central. 2. BandkeramikcultureEurope, Central. 3. AgricultureEurope,

CentralOrigin. 4. Agriculture, PrehistoricEurope, Central.5. Plant remains (Archaeology)Europe, Central.

6. Europe, CentralAntiquities. I. Title.GN776.2.A1B64 2004

930dc22

2003027487

ISBN 0-415-32485-8 (hbk)ISBN 0-415-32486-6 (pbk)

This edition published in the Taylor & Francis e-Library, 2005.

ISBN 0-203-35800-7 Master e-book ISBN

ISBN 0-203-66978-9 (Adobe eReader Format)

To purchase your own copy of this or any of Taylor &Francis or Routledge scollection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.

vT O M Y E X T E N D E D FA M I LY

vii

CONTENTS

List of gures viiiList of tables xiAcknowledgements xiii

Introduction 1

1 The study area and its archaeological background 10

2 Models of crop husbandry in Neolithic central Europe 21

3 The key variables of permanence, intensity andseasonality and their wider implications 50

4 Archaeobotanical, ecological and statistical methodology 60

5 Testing the four major crop husbandry models 96

6 Identication of separate ecological gradients andspecic crop husbandry practices 115

7 Conclusions: Neolithic farming in central Europe 154

Notes 171Bibliography 173Index 202

viii

FIGURES

1.1 Map showing the study area, which encompasses muchof the loess belt (dark shading) and the Alpine Foreland 10

4.1 Map showing the location of archaeological sites fromwhich the selected archaeobotanical samples derive 69

5.1 Histograms showing proportions of perennials in thearchaeobotanical samples 97

5.2 Results of the discriminant analysis separating autumn-and spring-sown weed associations in Germany basedon semi-quantitative data 100

5.3 The relationship of autumn-sown and spring-sownweed associations from Germany, Asturias plots andarchaeobotanical samples to the discriminant functionextracted to distinguish autumn- and spring-sownweed associations from Germany based on semi-quantitative data 102

5.4 The relationship of archaeobotanical samples from differentprocessing stages to the discriminant function based onsemi-quantitative data 103

5.5 Results of the discriminant analysis separating Evviapulse gardens and elds based on semi-quantitative data 106

5.6 The relationship of Evvia pulse gardens and elds,Asturias plots and archaeobotanical samples to thediscriminant function extracted to distinguish Evviagardens and elds based on semi-quantitative data 107

5.7 The relationship of archaeobotanical samples fromdifferent processing stages to the discriminant functionbased on semi-quantitative data 108

6.1 Correspondence analysis plot of weed species in theglume wheat ne sieve by-product samples 117

6.2 Correspondence analysis plot of glume wheat ne sieveby-product samples showing the contribution ofPhleum pratense L. 118

ix

6.3 Correspondence analysis plot of glume wheat ne sieveby-product samples showing the archaeological sites fromwhich samples derive 119

6.4 Correspondence analysis plot of glume wheat ne sieveby-product samples showing the geographical regionsfrom which samples derive 119

6.5 Correspondence analysis plot of glume wheat ne sieveby-product samples showing the period from whichsamples derive 120

6.6 Correspondence analysis plot of glume wheat ne sieveby-product samples showing life history 121

6.7 Correspondence analysis plot of glume wheat ne sieveby-product samples showing germination time 122

6.8 Correspondence analysis plot of glume wheat ne sieveby-product samples showing owering onset/length 123

6.9 Correspondence analysis plot of glume wheat ne sieveby-product samples showing estimated epidermal cellendopolyploidy 124

6.10 Correspondence analysis plot of glume wheat ne sieveby-product samples showing length of the oweringperiod 126

6.11 Correspondence analysis plot of glume wheat ne sieveby-product samples showing vegetative spread 127

6.12 Correspondence analysis plot of glume wheat ne sieveby-product samples showing vegetative spread (perennials)and length of the owering period (annuals) 128

6.13 Correspondence analysis plot of glume wheat ne sieveby-product samples showing mean canopy dimension 130

6.14 Correspondence analysis plot of glume wheat ne sieveby-product samples showing the weed size index 131

6.15 Correspondence analysis plot of glume wheat ne sieveby-product samples showing leaf area per node 132

6.16 Correspondence analysis plot of glume wheat ne sieveby-product samples showing leaf area per node:thickness 133

6.17 Correspondence analysis plot of glume wheat ne sieveby-product samples showing specic leaf area 133

6.18 Correspondence analysis plot of glume wheat ne sieveby-product samples showing guard cell length 135

6.19 Correspondence analysis plot of glume wheat ne sieveby-product samples showing stomatal density 136

6.20 Correspondence analysis plot of glume wheat ne sieveby-product samples showing the seed longevity index 137

6.21 Correspondence analysis plot of glume wheat ne sieveby-product samples showing phytosociological class 138

L I S T O F F I G U R E S

x6.22 Schematic representation of major ecological trends inthe correspondence analysis of glume wheat ne sieveby-product samples 139

6.23 Map showing sites with samples included in thecorrespondence analysis and their crop growing conditions 140

L I S T O F F I G U R E S

xi

TABLES

1.1 The chronological framework followed in this book 122.1 Experimental, historical and ethnographic data on cereal

yields and archaeological estimates of crop and cultivatedarea requirements 23

2.2 Estimates of the cultivated area for a household of ve(requiring 1,500 kg cereals per annum), assumingnegligible deduction for seed corn 43

3.1 The relationship of the four major crop husbandry modelsto the three variables of permanence, intensity andseasonality 50

4.1 Simplication of crop identication categories 634.2 The archaeobotanical dataset selected, summarized by crop

type and crop processing classication 674.3 Period, context and bibliographic information for the

selected archaeobotanical samples 704.4 The standardized potential arable weed taxa and their

frequency in selected samples 774.5 The major categories of functional attributes used and

their ecological signicance 794.6 Flowering onset/length classes 845.1 Woodland taxa in the archaeobotanical samples 985.2 The classication of archaeobotanical samples by the

discriminant function extracted to distinguish weedassociations from autumn- and spring-sown crops(Germany), showing samples classied with high andlow probability 101

5.3 The classication of archaeobotanical samples, withChenopodium album L. removed, by the discriminantfunction extracted to distinguish weed associations fromautumn- and spring-sown crops (Germany), showingsamples classied with high and low probability 104

xii

5.4 The classication of archaeobotanical samples by thediscriminant function extracted to distinguish pulse eldsand gardens (Evvia), showing samples classied with highand low probability 106

5.5 The classication of archaeobotanical samples, withChenopodium album L. removed, by the discriminant functionextracted to distinguish pulse elds and gardens (Evvia),showing samples classied with high and low probability 109

5.6 Cross-tabulation of the classication of archaeobotanicalsamples by the semi-quantitative discriminant analyses 110

5.7 Combined classication of archaeobotanical samples byprocessing group 111

5.8 The relationship of husbandry regime to archaeologicalsite and geographical region 113

5.9 The relationship of husbandry regime to chronologicalperiod 114

6.1 Codes for weed taxa in the correspondence analysis 1166.2 The relationship of chronological period to growing

conditions 1416.3 Interpretation of observed ecological trends in terms of

crop husbandry practices 1436.4 Variability between regions included in the correspondence

analysis during the LBK 149

L I S T O F T A B L E S

xiii

ACKNOWLEDGEMENTS

I particularly want to acknowledge the guidance and support of GlynisJones, who supervised the Ph.D. on which this book is based. I am alsoindebted to the following people for their helpful comments on the manu-script at various stages: Mary Bogaard, Paul Bogaard, Mike Charles, SueColledge, Paul Halstead, Carol Palmer, Peter Rowley-Conwy, Alasdair Whittleand two anonymous reviewers. Many archaeobotanists provided advice andinformation and I would like to thank Corrie Bakels, Stefanie Jacomet andher colleagues in Basel, Sabine Karg, Angela Kreuz, Helmut Kroll, UrsulaMaier, Ulrike Piening, Simone Riehl, Manfred Rsch and Hans-Peter Stika.I am also grateful to Paul Halstead, Rdiger Krause, Hans-Christoph Strien,Jutta Meurers-Balke, Jens Lning and Wilhelm Lohmeyer for access to un-published information. John Hodgson provided expert advice on eld iden-tications and plant ecology. David Taylor kindly drew the maps. Ph.D.funding was provided by the Social Sciences and Humanities Research Councilof Canada, the Overseas Research Fund and the University of Shefeld.

1I N T R O D U C T I O N

INTRODUCTION

This book is concerned with the nature of early farming in central Europe in particular the methods used to grow crops. Current perceptions of cropcultivation in central Europe during the Neolithic vary widely and includemodels of transient and permanent cropping, small-scale hand tillage andlarge-scale cultivation with the ox-drawn ard, farming of oodplain alluviumand higher ground. Debate over crop husbandry reects conicting views ofthe way in which farming spread from the Near East to Mediterranean andtemperate Europe, the mobility of early farming communities, the extent ofsocial differentiation among households and the goals of crop production.The aim of this book is to address these conicting views of early crop hus-bandry by analysing the extensive archaeobotanical dataset available fromNeolithic sites (c. 55002200 bc) across central Europe, in particular theloess belt and Alpine Foreland.

The general intention of this book, therefore, is to bring a substantialarchaeobotanical record from central Europe into the mainstream of archae-ological discourse on European prehistory. The approach used is to interpretthe archaeobotanical data in terms of an explicit methodology for recon-structing crop husbandry practices, and to evaluate previously suggestedmodels of crop husbandry in light of the archaeobotanical evidence. Non-specialists may be surprised to nd that this analysis is not based on the cropspecies themselves, but rather on close attention to the arable weeds thatgrew and were harvested with certain crops. It is this weed evidence thatreects the fundamental agency of crop growing the time chosen to sowcrops, measures taken to encourage growth, the permanence of cultivationareas in the landscape, and so on. These choices, in turn, provide a richsource of evidence for the everyday life and longer-term transformations ofpast societies.

Archaeologists seeking to rene their accounts of agricultural practicebeyond the listing of domesticated species from archaeological sites soonface a real methodological problem. Explicitly constructed models based onrelevant features of plant and animal ecology are needed in order to relatebioarchaeological assemblages back to management regimes, and such models

I N T R O D U C T I O N

2

are still being developed (e.g. Charles et al. 1997; Halstead 1998; Mainland1998; Balasse and Tresset 2002).

An overriding concern with methodology and model-building is oftenassociated with the New Archaeology and related approaches of the 1960sand 1970s (Binford 1972; Higgs 1975), which focused on subsistence andtechnology and emphasized human adaptation to the environment. A problemwith this approach is that it sometimes made the diversity of human cultureseem redundant (Barker and Gamble 1985; Shanks and Tilley 1987). Bycontrast, post-processual archaeology (Hodder 1986, 1991; Shanks and Tilley1987; Barrett 1990, 1994) has rejected these preoccupations, emphasizingthe interpretation of meaning and symbolism in the form and functioning ofhouses, settlements and landscapes, in artefact styles or in mortuary practices.Recent accounts of the MesolithicNeolithic transition in Europe (e.g. Hodder1990; Thomas 1999; Barrett 1994; Whittle 1996a; Bradley 1998) havefocused on these forms of evidence, with only generalized consideration ofthe nature of agricultural practices themselves.

To the extent that post-processualism is not a call for methodologicalrigour but for a radical change in perspective, the construction of modelsrelating bioarchaeological assemblages to husbandry regimes might be seenas a relatively low priority. Such an anti-methodological stance is, however,self-defeating: any revolution in archaeological theory is meaningless if itfails to formulate alternative interpretations based on rationally decisivearchaeological evidence (Wylie 1992, 1996).

The remainder of this introductory section discusses the signicance ofcrop husbandry practices in archaeology, the role of middle range theory andthe interpretation of archaeobotanical weed evidence. Chapter 1 provides abrief archaeological summary of the Neolithic in the loess belt and AlpineForeland. Chapter 2 sets out previously suggested models of Neolithic crophusbandry practices in central Europe and the evidence on which they arebased. Chapter 3 considers the key ecological variables that distinguish themain crop husbandry models and their broader social and economic implica-tions. The methods used to select and analyse archaeobotanical data from thestudy area and relevant modern weed survey data are the focus of Chapter 4.Chapter 5 presents the results of statistical analyses comparing the selectedarchaeobotanical samples directly with modern weed oras developed underdifferent husbandry regimes on the basis of their weed ecological character-istics. The aim of Chapter 6 is to identify and interpret specic ecologicaltrends in weed species composition among the archaeobotanical samples,and hence differences in individual aspects of crop husbandry. Finally, Chap-ter 7 discusses the wider archaeological implications of the results presentedin Chapters 5 and 6 in terms of the models of crop husbandry reviewed inChapters 2 and 3. Brief chapter summaries are provided at the end of Chap-ters 1 through 6.


Crop husbandry and middle range theory in archaeology

Crop husbandry refers to the methods farmers use to grow crops, including thetiming and method of tillage and sowing, weeding and watering of crops,middening or manuring and also longer-term rhythms of fallowing androtation. These practices largely determine the productivity, labour demands,reliability and long-term sustainability of crop growing. While attemptshave been made to understand husbandry regimes and their transformationas a function of single factors such as population pressure (Boserup 1965),environment (Higgs and Vita-Finzi 1972) or the spread of technologicalinnovations (Sherratt 1981), studies of farming societies around the worldattest to the complex cultural specication of such basic parameters as carryingcapacity, resource use, response to environmental change and the adoptionof technological innovations (Sahlins 1972: 49; Grigg 1982; Halstead 1995;Charles and Halstead 2001). Ethnographic and historical studies haveidentied links between crop husbandry regime and many other aspects offarming communities, including settlement pattern, land ownership, socialstratication and animal husbandry (Netting 1971; Goody 1976; Sherratt1981; Fleming 1985; Halstead 1987, 1990, 1995; Hodkinson 1988;Williamson and Bellamy 1987; Palmer 1998b; Forbes 1982, 2000a, 2000b).Crop husbandry is thus of central importance for understanding pastagricultural societies (see also Chapters 23) and, despite various shifts oftheoretical outlook, has been a consistent theme of synthetic works on laterEuropean prehistory (Childe 1929, 1957; Clark 1952; Piggott 1965; Dennell1983; Barker 1985; Hodder 1990; Whittle 1996a).

Anthropologists such as Bourdieu (1977, 1990) have demonstrated thatsocial reproduction takes place in the everyday habits of living. Archaeologistshave increasingly looked to habitual action (Gosden 1994: 188) or the dullcompulsion of routine experience (Edmonds 1999: 486) as the context inwhich social identities and institutions emerge and are reproduced over thelong term (Barrett 1994, 1999; Gosden 1994; Edmonds 1999). In agriculturalsocieties, crop husbandry can offer insights into these social processes as itrepresents a whole series of routines or tasks taking place on a series oftimescales (daily through seasonal, annual and inter-annual). Since it com-bines these varying timescales with the spatial dimension of arable land use,crop husbandry is of obvious relevance to recent emphasis on inhabitedlandscapes or taskscapes as a context for archaeological discourse (Ingold1993). An understanding of past crop husbandry regimes is also needed inorder to assess the enduring effects of farming on the landscape (Acheson1997; Halstead 2000).

Recent interest in routine practice highlights the need to broaden the rangeand resolution of inferences that archaeologists can make about the past.Different forms of archaeological data have been used to make inferencesabout past crop husbandry, including settlement distribution and artefactual/


4

representational data. For example, Andrew Sherratts theory of a SecondaryProducts Revolution in later Neolithic Europe, including the transformationof crop husbandry through the introduction of the ox-drawn ard, is basedlargely on these forms of evidence (Sherratt 1981). Potentially the mostinformative source of archaeological evidence for crop husbandry practices the seeds of arable weeds associated with crop material in archaeologicaldeposits has been under-utilized, however, or even misinterpreted, largelybecause the theory needed to link weed evidence with husbandry practiceshas been inadequate. In other words, archaeological inference has been lim-ited by weaknesses and inconsistencies in the linking arguments or middlerange theory (Binford 1977, 1981: 23; see also Raab and Goodyear 1984)needed to interpret archaeobotanical weed assemblages as evidence of crophusbandry practices.

Binford (1981: 2530) characterized good middle range theory as un-ambiguous, based on clear cause and effect rather than simple correlation,applicable to the past (i.e. based on plausible uniformitarian assumptions)and intellectually independent of general theory. Hodder (1982) has dis-cussed a similar concept of relational analogy as analogy based not on meresupercial similarity (formal analogy) but on some natural or cultural linkbetween the different aspects of the analogy (Hodder 1982: 16) that is, onsimilarity of causal mechanisms (Wylie 1985: 95). Criticism of middle rangetheory has tended to focus on the notion of its theoretical independence: ifall observation is theory-laden, the independence of middle range theory isillusory and arguments based upon it are circular (Hodder 1986: 107; Shanksand Tilley 1987: 122; Barrett 1990, 1994: 171 n. 1). Fortunately, however,not all forms of theory-ladenness are equally problematic (Kosso 1991; Wylie1986, 1992, 1993, 1995, 1996, 1998); thus, for example, plant ecologicaltheory relating to the behaviour of weeds under different crop husbandryregimes is based on a set of assumptions with no direct relation to broadertheories of human behaviour (cf. Charles and Halstead 2001). On the otherhand, middle range theory developed by Binford himself (1978: 45897)incorporates assumptions of human rationality and optimizing behaviourand so offers a useful heuristic tool rather than a set of innocent linkingarguments between the static record and dynamic past (Wylie 1989a; Halstead1998; Charles and Halstead 2001). Another focus for criticism has beenBinfords claim that middle range theory provides Rosetta Stones for thepast (Binford 1981: 25), with the implication that all aspects of past humanbehaviour are susceptible to reconstruction, provided the necessary middlerange theory is developed (Wylie 1989b). The claim is clearly false, butuseful middle range theory can be developed on the basis of physical, chem-ical and biological properties of humans, other organisms or artefacts thatare plausibly extrapolated to the past and largely independent of assumptionsabout human behaviour (Wylie 1985, 1986, 1993, 1995; Shennan 1993;Charles and Halstead 2001). Middle range theory is an indispensable tool for


archaeology (Cowgill 1993; Stark 1993; Trigger 1995; Wylie 1998; cf. Hodder1991); Kosso (1991) and Tschauner (1996) have highlighted its use in thewritings of Binfords own critics. The development of useful middle rangetheory not only broadens the scope of archaeological inference but also con-strains what we can claim about the past.

Approaches to the interpretation of archaeobotanical weedassemblages as evidence of crop husbandry practices

A given species of crop can generally tolerate a range of growing conditionsand may be grown using a variety of different husbandry practices (Behreand Jacomet 1991). Archaeobotanical crop remains, therefore, do not offerdetailed insight into crop husbandry, though carbon isotope studies of anci-ent grain have been used to detect irrigation (Araus et al. 1997) and ancientcrop DNA may eventually permit the identication of ecotypes adaptedto specic growing conditions (cf. Davies and Hillman 1988). At present,the most useful archaeological evidence of crop husbandry is provided by theseeds of arable weeds found in association with crop material in archae-ological deposits (Knrzer 1971, 1973, 1979, 1984; Willerding 1980, 1981,1983a, 1986; Hillman 1981, 1991; M. Jones 1981, 1988; Wasylikowa 1981;Greig 1988; Behre and Jacomet 1991; Kster 1991; G. Jones 1992, 2002;van der Veen 1992). Archaeobotanists have observed that ancient weedassemblages are often quite different from those of recent times and that thisis likely to reect differences in crop husbandry practices (Knrzer 1973;Willerding 1980, 1981, 1983a, 1986; Behre and Jacomet 1991; G. Jones1992; Kroll 1997). Weed species have different ecological requirements andpreferences (Holzner and Numata 1982; Ellenberg 1996: 87088); hence, theprominence of certain weed species at a particular time and place shouldreect the nature of the crop husbandry regime under which they thrived.

While the signicance of ancient weed assemblages for the reconstructionof crop husbandry is widely acknowledged, approaches to the interpretationof this evidence vary, with the result that the same data can be interpretedin radically different ways (Charles et al. 1997). Two of the main approachesthat have been used phytosociology and Ellenberg numbers are discussedbelow, before presenting a third approach that overcomes the major weak-nesses of previous methods.

Phytosociology

Phytosociology classies stands of vegetation into communities or syntaxabased mainly on the occurrence of character species, which are more or lessrestricted to a certain syntaxon (Westhoff and van der Maarel 1973). Anapproach seeking to identify modern syntaxa in archaeobotanical assemblagesis of limited usefulness in archaeology due to the historical contingency of


6

plant communities and their instability through time (Holzner 1978; M. Jones1984; Behre and Jacomet 1991; Hillman 1991; Kster 1991; G. Jones 1992).Other problems with the identication of weed communities in archaeo-botanical samples are that samples may not derive from a single eld andthat they do not fully reveal the original eld weed ora (M. Jones 1988;Kster 1991; G. Jones 1992). These problems also apply to attempts toreconstruct ancient weed associations that no longer exist (e.g. the Bromo-Lapsanetum praehistoricum association of Knrzer 1971).

An alternative use of phytosociology has been to interpret the occurrenceof character species in archaeobotanical assemblages as indicators of habitatconditions rather than to identify syntaxa per se. The more general groupingsof species (e.g. at the highest syntaxonomic level of class) can be applied toarchaeobotanical weed assemblages with some condence (Behre and Jacomet1991; Kster 1991; G. Jones 1992) and the occurrence of character speciesbelonging to these general groupings can be used as indicators of the habitatconditions under which the group as a whole occurs. Even with this alternativeuse, however, there remains the underlying problem that plant communitiesare linked to eld observations of growing conditions without distinguishingbetween ecological requirements and tolerances; for example, species in acommunity growing in moist conditions might be assumed to indicatemoisture generally, even though some or all of them merely tolerate a certainlevel of moisture but have a specic set of requirements for fertility or lightetc. In other words, eld observations linking phytosociological communitieswith growing conditions do not reveal which aspects of the environmentcause certain species to grow in certain locations (Charles et al. 1997).

Phytosociology has been widely used in archaeobotany to infer habitat con-ditions and crop husbandry practices (van Zeist 1974; Wasylikowa 1978, 1981;Willerding 1979, 1983a; Jacomet et al. 1989: 12844; Behre and Jacomet1991; Karg 1995; Rsch 1998b; G. Jones 2002). Some archaeobotanists haveidentied a greater prevalence of character species of the class Chenopodietea(root/row-crop or garden weeds and ruderals) in archaeobotanical assemb-lages of cereals and pulses compared with modern phytosociological studiesof winter cereals (Knrzer 1971; Willerding 1979, 1981, 1983a; Behre andJacomet 1991; G. Jones 1992). For example, G. Jones (1992) has noted thatthe weed assemblage associated with charred crop stores from late BronzeAge Assiros Toumba in Greek Macedonia is particularly rich in characterspecies of the Chenopodietea compared with modern winter cereals and pulses.G. Jones (1992) argues that this could reect the use of garden-like methodsof crop husbandry such as manuring, hand-weeding or hoeing and wateringof crops but notes that three other explanations are also possible. First,character species of this group tend to be spring-germinating and so tendto characterize spring-sown crops, suggesting perhaps that archaeobotanicalcereals/pulses associated with Chenopodietea are spring-sown (Groenman-van Waateringe 1979; Gluza 1983; Behre 1990) or that Chenopodietea-rich


assemblages are derived specically from (spring-sown) millet cultivation(Wasylikowa 1978; Kroll 1979, 1997). Second, the Chenopodietea groupalso includes many species that grow as ruderals (that is, in non-arable dis-turbed habitats), and so a further possibility is that the assemblages rich inChenopodietea are contaminated by material of ruderal (non-arable) origin.A third explanation for the occurrence of Chenopodietea species in archaeo-botanical assemblages, rst proposed by Willerding (1980, 1981, 1983a, 1985,1986: 335, 1988a, 1988b), is that Chenopodietea species reect an openstand of autumn-sown crops allowing root/row-crop weeds to germinate inthe gaps and compete with established plants.

Ellenberg numbers

Ellenberg (1950, 1979; Ellenberg et al. 1992) developed a series of scalesfor major environmental variables (light, temperature, continentality, soilmoisture, soil pH, soil nitrogen content, etc.) and scored a large number ofcentral European plant species on each of these scales. Ellenberg numbers, orindicator values (Zeigerwerte), have been widely used in archaeobotanical inter-pretation in order to infer fertility, moisture level, shadiness, etc. (Wasylikowa1978, 1981; Willerding 1980, 1983a; Jacomet et al. 1989: 14553; van derVeen 1992: 1089). Ellenberg numbers were subjectively determined foruse in central Europe, though they have been shown to correspond very wellto more objective measures of species behaviour in Britain (Thompson et al.1993). A more serious problem is that they are based on eld observations ofspecies behaviour that, as noted above in connection with phytosociology,do not distinguish between species ecological tolerances and requirementsand so cannot disentangle which ecological factor(s) determine the occurrenceof species in certain locations (Charles et al. 1997).

FIBS in archaeobotany

A new approach to the ecological interpretation of archaeobotanical data isknown as the Functional Interpretation of Botanical Surveys. FIBS providesa means of relating the behaviour of individual plant species to specicecological variables, thus overcoming the limitations of previous approachesbased on eld observations (Charles et al. 1997; G. Jones 2002). This approachwas developed at the Unit of Comparative Plant Ecology, University ofShefeld for investigating the impact of ecological processes on speciesdistribution within a wide range of habitats (Hodgson 1989, 1990, 1991;Hodgson and Grime 1990; Hodgson et al. 1999). FIBS is based on themeasurement of functional attributes morphological and behavioural traitsthat measure species potential in relation to major variables such as fertility,disturbance and moisture. In a vegetation survey of contrasting habitats, forexample, the importance of specic ecological variables can be assessed by


8

comparing functional attribute values of species associated with the differenthabitats. Species sharing the same habitat also tend to share ecologicalcharacteristics and thus belong to a distinct functional type (Grime 1979;Grime et al. 1988).

FIBS has been applied to a series of modern weed survey studies oftraditional crop husbandry regimes in Europe and the Near East. While theweed oras associated with these different husbandry regimes (e.g. irrigationversus dry farming, intensive versus extensive cultivation, different rotationregimes, etc.) can be distinguished from each other on a oristic basis alone(G. Jones et al. 1995, 1999; Palmer 1998a; Bogaard et al. 2001; Charles andHopp 2003), the modern weed oras may overlap only partially or not atall with archaeobotanical weed assemblages. A method that links speciescharacteristics rather than species per se with particular traditional crophusbandry practices, therefore, is essential to the reconstruction of ancienthusbandry regimes. Using FIBS, it has been demonstrated that the mod-ern husbandry regimes can be distinguished on the basis of the functionalattribute values of weed species associated with different husbandry practices(Charles et al. 1997, 2002, 2003; Bogaard et al. 1999, 2000, 2001; G. Joneset al. 2000a). Moreover, the use of functional attributes makes it possibleto disentangle the effect of multiple ecological factors (e.g. fertility anddisturbance, both of which may contribute to cultivation intensity G. Jones et al. 2000a).

The two main advantages of FIBS in archaeobotany, therefore, are that (1)it provides a means of comparing modern weed oras developed under knownhusbandry conditions with ancient weed assemblages, and (2) it allows dis-tinct ecological factors to be monitored independently (Charles et al. 1997;G. Jones 2002). Thus, if functional attribute data are assembled for weed spe-cies in an archaeobotanical assemblage, FIBS makes it possible to construct arelational analogy (incorporating causal mechanisms Hodder 1982: 1127;Wylie 1985; cf. Binford 1981: 2530) between the archaeobotanical weedassemblage and modern weed oras developed under particular husbandryregimes. Critically, because the terms of the comparison the functionalattributes are inherently meaningful (functional), there is also potentialto reconstruct ancient husbandry regimes for which no close modern analogueexists.

In any ecological approach to archaeobotanical weed assemblages, theuniformitarian assumption that the ecology of weed species has remainedstable through time is problematic (Behre and Jacomet 1991; G. Jones 1992,2002). The use of multiple weed species reduces the potential for erroneousconclusions due to major changes in the behaviour of individual species(G. Jones 1992, 2002; Charles and Halstead 2001). Since functional attri-butes can be measured rapidly for any species in an archaeobotanical weedassemblage, FIBS promotes the use of suites of associated species to inferpast growing conditions.


As a form of good middle range theory, therefore, FIBS satises threeof Binfords criteria set out above: the relation of functional attributes tocrop husbandry practices is one of cause and effect, plausible uniformitarianassumptions can be made based on suites of associated weed species, andassumptions about plant ecology bear no direct relation to assumptionsabout human behaviour. FIBS also goes a long way towards satisfying thefourth criterion: while the relation of functional attributes to crop husbandrypractices is not entirely unambiguous since different husbandry measuresmay have similar ecological effects, the use of functional attributes permitsthis ambiguity to be identied and assessed.

S T U D Y A R E A A N D I T S A R C H A E O L O G I C A L B A C K G R O U N D

10

1

THE STUDY AREA AND ITSARCHAEOLOGICAL

BACKGROUND

Denition of the study area

Two broad regions of central Europe have been the subject of debate overthe nature of Neolithic crop husbandry and are particularly well investigatedfrom an archaeobotanical point of view. The rst region is the loess belt, aseries of upland basins where silt-like, wind-deposited material (loess) accu-mulated to the south of the Pleistocene ice sheets, forming a discontinuousband across Europe. Excavation has taken place at hundreds of Neolithicsettlements in the loess belt, with sampling for plant remains at earlyNeolithic (Linearbandkeramik or LBK) sites in particular (Willerding 1980;Lning 1988; Kreuz 1990; Knrzer 1997). The second region is the AlpineForeland, where intensive bioarchaeological study has long formed part of

Figure 1.1 Map showing the study area (dashed line), which encompasses much ofthe loess belt (dark shading) and the Alpine Foreland

11


excavation at lakeshore settlements dating from the later Neolithic onwards(Heer 1866; Neuweiler 1905; Schlichtherle 1983, 1997a; Jacomet et al. 1989;Brombacher and Jacomet 1997; Maier 2001). Together, the loess belt andAlpine Foreland form a very broad study area bounded by the coastal plainsof Europe to the north and west and the Alps to the south (Figure 1.1). Tothe east, the study area includes the loess belt of southern Poland, Slovakiaand Hungary. LBK settlement did extend further east along with loess soils,into Romania (Transylvania, north-east Wallachia), south into Croatia and,to the north and east of the Carpathians, into the Ukraine and Moldova, butavailable archaeobotanical data from these regions are as yet limited (Dergachevet al. 1991; Wasylikowa et al. 1991; Tezak-Gregl 1993; Crciumaru 1996;Pashkevich 1997; Larina 1999). Similarly, extension of the study area toinclude LBK settlement north of the loess belt, in the North EuropeanPlain, was considered unproductive because of the restricted archaeobotanicaldataset available (Bogucki 1982: 97; Heuner 1989; Nalepka et al. 1998).

The chronological framework used in this book follows the scheme pro-posed by Lning (1996: Figure 1) for central Europe. Table 1.1 summarizesthe relevant periods and culture-historical groupings.

The earlymiddle Neolithic (c. 55004400 bc)

The earlymiddle Neolithic archaeology considered here comprises theLinearbandkeramik or LBK (c. 55005000 bc) and subsequent LBK-relatedcultures (i.e. Rssen, Lengyel, Stichbandkeramik or SBK, Oberlauterbach, etc.)of the middle Neolithic (c. 50004400 bc) in the loess belt what Bogucki(1988) calls the Primary Neolithic of central Europe, a period of over onethousand years. Earlymiddle Neolithic sites are at palimpsests of post-holes and trenches of longhouses and associated pits; erosion and disturbancehave destroyed oor surfaces and occupation layers, leaving only negativefeatures lled with archaeological deposits (Whittle 1996a: 160). LBK long-houses tend to have a tripartite organization, with front, central and backsections divided by transverse rows of posts, though two- and perhaps one-section longhouses also occur and structures vary considerably in overall length,from c. 10 to 40 m (Modderman 1988; Coudart 1998: 19, 278, 534).Middle Neolithic longhouses often lack the internal tripartite division andtend to be trapezoidal in shape rather than rectangular (Coudart 1998: 51,54, 56).

The concept of the longhouse as a farmstead situated within its ownyard (the Hofplatz model) was developed in the course of extensive rescueexcavations of LBK sites in the lower Rhine basin, including the completeexcavation of a 1.3 km stretch of the Merzbach valley in the AldenhovenPlateau, which suggested that each longhouse was surrounded by a charac-teristic set of pits (Lning 1982b, 1988, 1997; Stehli 1989). It was suggestedfurther that the sequential replacement of longhouses over time resulted in a


12

Tab

le 1

.1T

he c

hron

olog

ical

fra

mew

ork

foll

owed

in

this

boo

k, w

ith

cult

ure-

hist

oric

al g

roup

ings

for

eac

h re

gion

lis

ted

in a

ppro

xim

ate

chro

nolo

gica

l or

der

(LB

K =

Lin

earb

andk

eram

ik,

SBK

= S

tich

band

kera

mik

, T

RB

= T

rich

terb

eche

rkul

tur)

Maj

or p

erio

ds14

C c

al.

BC

N F

ranc

e, B

elgi

um

Low

er R

hine

M

euse

bas

in

Sout

hern

Ger

man

y,Sw

itze

rlan

d

Cen

tral

Ger

man

y,B

ohem

ia,

Sout

hern

Pol

and

Aus

tria

, H

unga

ry,

Slov

akia

, M

orav

ia

Sour

ces :

The

ove

rall

str

uctu

re a

nd 1

4 C d

atin

g fo

llow

s L

ning

(19

96:

Figu

re 1

). Su

pple

men

tary

sou

rces

are

Kee

fer

(199

3: 1

71),

Koo

ijm

ans

(199

3:Fi

gure

9),

Whi

ttle

(19

96a:

Fig

ure

6.3)

, P

reu

(19

98:

supp

lem

ent

1).

Ear

ly N

eoli

thic

5500

500

0

Rub

an

rce

nt

LBK

Mid

dle

Neo

lith

ic50

00

4400

Cer

ny/B

licq

uy

Rs

sen

Hin

kels

tein

,G

rog

arta

ch,

Rs

sen/

Leng

yel/

SBK

/O

berl

aute

rbac

h,B

isch

heim

SBK

Leng

yel

Lat

er N

eoli

thic

4400

350

0

Cha

sse

n

Mic

hels

berg

Aic

hbh

l,M

iche

lsbe

rg,

Schu

ssen

ried

,A

lthe

im,

Pfy

n-C

orta

illo

d

Late

Len

gyel

,T

RB

Late

Len

gyel

3500

280

0

Sein

eO

ise

Mar

ne

Hor

gen/

Cha

m/

War

tber

g

Ber

nbur

g, B

aden

,G

lobu

lar

Am

phor

a

Bad

en

2800

220

0

Cor

ded

War

e

Bel

l B

eake

r

13


lateral drift of structures (Lning 1982b). This view is supported by seriationof nds (especially ceramics) from pits associated with individual longhouses,allowing the sequential replacement of longhouses to be traced throughmany generations (Lning 1988, 1997). As new longhouses were constructed,there appears to have been a tendency to avoid overlap with earlier structures(Pavlu 2000: 243).

Since earlymiddle Neolithic sites represent palimpsests of drifting long-houses through time, what appear to be dense concentrations of longhousesmay represent the replacement of a single structure over time or include veryfew contemporary longhouses separated by considerable distances, from 1020 m to 100 m or more (Hamond 1981; Milisauskas 1986: 34; Whittle1996a: 151). While many LBK settlements appear to consist of one to a fewlonghouses at any one time, large sites with a number of contemporary long-houses are also known: at Langweiler 8 in the Merzbach valley (AldenhovenPlateau), for example, 11 contemporary longhouses covering c. 7 ha are attestedin one phase (Lning 1988, 1997). Population estimates for earlymiddleNeolithic sites range from less than ten (a single longhouse) to severalhundred or more (Modderman 1970: 2057; Milisauskas 1986: 21920;Milisauskas and Kruk 1989a; Coudart 1998: 91). Most, if not all, settlementswould not be demographically viable (cf. Wobst 1974), and recent strontiumisotope work on human bone from LBK sites in the Rhineland suggestsmovement of women in particular between communities in different regions,perhaps as a result of intermarriage and patrilocality (Price et al. 2001;Bentley et al. 2002, 2003).

Lning (1997) has argued that even large sites such as Langweiler 8were not true villages, with an emphasis on communal, supra-householdorganization, but loose groupings of farmsteads (Streusiedlungen). The newlyexcavated LBK settlement at Vaihingen in south-west Germany, however,was enclosed in one phase by a ditch containing burials and appears toreect a more cohesive, village-like community (Krause 2000). In someregions at least, sites appear to become more nucleated in the middle Neolithic(e.g. fewer, larger sites in the lower RhineMeuse basin), and greater cohesionmay also be indicated by the construction of enclosures at some sites, imply-ing communal cooperation (Lning 1982b, 2000: 16; Starling 1985, 1988;Pavuk 1991; Hodder 1990: 1229). The length of earlymiddle Neolithicsite occupations varies but is often of the order of several centuries; somelarge settlements were occupied for more than four hundred years (Lning1997, 2000: 15).

Earlymiddle Neolithic sites tend to occur in clusters, often strung outalong small- to medium-sized river valleys (Hamond 1981; Bogucki 1988:74; Lning 1997). In some cases these clusters have been shown to include asingle large site plus a number of smaller ones (Lning 1997). These clusters,in turn, occur within broader concentrations of sites apparent on a continentalscale, referred to as settlement cells (Siedlungskammern) and often circumscribed


14

by topographical features such as hills surrounding basins (Hamond 1981;Bogucki 1988: 723).

It is well known that early Neolithic sites tend to occur in areas of loess,though there are many exceptions (e.g. Paris basin, lower Oder, Kuyavia),and middle Neolithic settlement expanded into the moraine landscapes ofthe North European Plain and the Alpine Foreland (Lning 2000: 17). Theassociation with loess has usually been interpreted as a preference for thehigh fertility of loess soils, though other factors may have been of greaterimportance, such as location in at areas near the conjunction of river valleysand watersheds (Bogucki 1988: 77; Bogucki and Grygiel 1993). This posi-tion enabled access to oodplains providing seasonal grazing for livestockand/or fertile alluvium for cultivation (Kruk 1973, 1980: 267, 504, 634, 1988; Bakels 1978: 139; Wasylikowa 1989).

Mixed deciduous woodland is generally considered to have been thedominant vegetation in the Neolithic (Kster 1995a: 6970, 745; Jacometand Kreuz 1999: 23140; Lning 2000: 257). Some authors, however,have suggested that the natural vegetation of central Europe is open park-land or wood pasture due to the impact of large native herbivores (Geiser1992; May 1993; Vera 2000). Zoller and Haas (1995) argue in favour ofmixed deciduous woodland but emphasize that it would exist in a mosaic ofregeneration states at any one time. The location of earlymiddle Neolithicsites along river valleys like the later emergence of lakeshore settlement inthe Alpine Foreland (pp. 1819) may reect a preference for relatively openvegetation (Zoller and Haas 1995).

The visibility and often dispersed distribution of earlymiddle Neolithiclonghouses has fostered interest in the household as the fundamental unit ofdecision-making (Bogucki 1988: 21415; Lning 1988: 86; Halstead 1989a;Bogucki and Grygiel 1993; Lning 2000: 180) by analogy with ethnographicand historical studies of small-scale agrarian societies (Sahlins 1972; Nettinget al. 1984). This perspective implies that agricultural practices reect theaspirations and motivations of individual households (cf. Bogucki 1988:215). The origins of social hierarchy have been sought in the relationshipsbetween households and their differential success (Halstead 1989a, 1989b;Bogucki 1993, 1999: 21018). There is disagreement, however, over thedegree of social differentiation among households in the earlymiddleNeolithic (see Chapter 3).

The crop spectrum of the early Neolithic consists of emmer and einkornwheat, common pea, lentil and ax, with the addition of opium poppy (inthe western LBK) and barley (among other rare cereals and pulses) at somesites, though some or all of these additional species may have grown asweeds of the major crops. (Willerding 1980; Jacomet and Kreuz 1999:2947). It has been suggested that emmer and einkorn were grown togetheras a mixed or maslin crop based on the ubiquitous mixture of these speciesin archaeobotanical samples (Willerding 1980, 1983b; Bakels 1991b; Knrzer

15


1997, 1998), though the lack of clear in situ storage contexts makes thisdifcult to demonstrate conclusively ( Jacomet and Kreuz 1999: 295).Recent work on the morphology of glume wheat remains at the LBK siteof Vaihingen (Bogaard unpublished data) suggests that a third type ofglume wheat, recently described in Neolithic assemblages from Greece andresembling modern Triticum timopheevi Zhuk (G. Jones et al. 2000b), wasalso grown. The LBK spectrum of cereals and pulses is narrow comparedto that known from the Neolithic in the southern Balkans and Greece(Halstead 1989a), which notably includes pulses such as bitter vetch, chickpea and grass pea that are mostly Mediterranean crops in Europe today(Zohary and Hopf 2000: 108, 116, 119). Free-threshing wheat and nakedbarley emerged as common crops in the study area during the middleNeolithic (Bakels 1991a, 1997a).

Cattle dominate most earlymiddle Neolithic animal bone assemblagesfrom the study area, while the relative importance of pig and sheep/goatappears to vary (Lning 2000: 110). The prominence of cattle throughoutthe Neolithic may reect their suitability for browsing in woodland as wellas their greater reliability compared with smaller stock as indirect banksfor surplus crops (Halstead 1989a, 1992b; cf. Bogucki 1988: 91). Availablemortality data for cattle and sheep/goat assemblages from earlymiddleNeolithic sites are generally too limited to infer herd management strategiesreliably (Halstead 1989a; Glass 1991: 69; Arbogast 1994: 91), but theyappear to reect predominant juvenile mortality and hence meat use (Arbogast1994: 93; Benecke 1994a: 95, 1994b: 1223). Benecke (1994a: 96) hasargued that a high proportion of adult females among cattle and sheep/goatremains at the middle Neolithic (Rssen) site of Knzing-Unternberg inLower Bavaria indicates a combined meat/milk strategy. Ceramic sieves fromLBK sites, interpreted as cheese strainers for separating curds and whey(Bogucki 1982, 1984, 1986), have been used to suggest that cattle wereexploited for their dairy products, perhaps as part of a more generalizedmilkmeatblood use strategy (cf. Glass 1991: 75).

In contrast to later Neolithic settlement in the North European Plain andthe Alpine Foreland, the LBK has often been treated as a textbook case ofmigration (Clark 1952: 958; Piggott 1965: 502; Ammerman and Cavalli-Sforza 1971, 1984: 61, 634; Vencl 1986; Bogucki 1987, 1996; Price et al.1995). Continuity with the Mesolithic in some aspects, however, especiallyin lithic assemblages (Tillmann 1993), as well as heterogeneity within theLBK (Lning 2000: 110) and possible evidence for Mesolithic agriculture(Erny-Rodmann et al. 1997), has been used to argue that the LBK represents acomplex pattern of indigenous adoption, with limited migration or no move-ment at all from the homeland of LBK material culture in the HungarianPlain (Dennell 1983: 176; Modderman 1988; Whittle 1996a: 3634, 1996b,1997; Kind 1998; Gronenborn 1999; Bogucki 2000; Jochim 2000; Zvelebil2000a, 2000b; Price et al. 2001; Bentley et al. 2002, 2003).


16

The later Neolithic (c. 44002200 bc)

The transition from the middle to the later Neolithic represents the end ofthe Bandkeramik tradition of longhouses in central Europe. SubsequentNeolithic settlement extended well beyond the loess belt including theAlpine Foreland, which forms part of the study area (Figure 1.1) andcontinued a trend towards increasing regionalization of material culture.Later Neolithic sites vary considerably in location (on and off loess; lakeshoresand interuves as well as valley margins), size (from large settlements sur-rounded by palisades and ditches to dispersed farmsteads) and duration (fromlong-lived settlements lasting several centuries to dendrochronologically datedlakeshore villages of less than twenty years occupation).

The later part of the Neolithic (c. 44002200 bc) has been characterizedas a period of profound changes. Sherratt (1981, 1997) proposed a secondaryproducts revolution based on the intensive use of renewable resources fromdomesticated animals (traction, milk and wool/hair) in the fourth and thirdmillennia bc. The temporal and geographic coherence of this horizon hasbeen questioned, as has the extent and nature of its impact on societiesacross Europe (Chapman 1982; Rowley-Conwy 1987, 2000a; Glass 1991:77; Halstead 1995; Lning 2000: 12). Recently, however, Bogucki (1993,1999: 22730) has advocated an animal traction revolution, arguing thatox-drawn ard (scratch plough) cultivation and wheeled transport freedlater Neolithic households from their inherent labour limitations and thatdifferential access to traction promoted economic differentiation betweenhouseholds (see Chapter 2).

Another potential cause of changes in settlement and society in the laterNeolithic is the fusion of indigenous hunter-gatherer and existing agricul-tural (i.e. LBK-related) communities, particularly in the North EuropeanPlain and the Alpine Foreland (Kruk 1988; Bogucki 1987, 1988: 1079,1996, 2000; Sherratt 1990, 1995). Contrasting Neolithic traditions sup-posedly founded by immigrant farmers versus indigenous farmers havebeen linked directly to contrasting crop husbandry regimes (Bogucki 1996),as discussed further in Chapter 2.

Within the loess belt, regional survey of Neolithic sites in southern Poland(Kruk 1973, 1980: 289, 547, 64, 1988) documented a tendency for laterNeolithic TRB (Trichterbecherkultur or Funnel-necked Beaker culture) sitesto be located in the interuve (or watershed) zone, away from the marginsof river valleys. Another regional study of Neolithic settlement, in the ElbeSaale area (Starling 1985, 1988), detected a similar trend for larger sites (hilltopenclosures or Hohensiedlungen) in the interuves but also greater continuity ofsettlement on valley margins. Kruk (1973, 1980: 289, 547, 64, 1988)interprets the focus on interuves, characterized by poorer, non-loess soils, asevidence for shifting cultivation in the TRB, while other authors infer a greateremphasis on animal husbandry, including animal-drawn ard cultivation

17


(Bogucki 1988: 1767; Howell 1989). Subsequent settlement evidence ofthe Baden culture in southern Poland also extends into the interuves, andan emphasis on stockbreeding and plough cultivation has been inferred(Sherratt 1981, 1997; Kruk 1988; Milisauskas and Kruk 1989a; Lning2000: 189). Bronocice, located on a loess ridge above a tributary of theVistula in south-east Poland, has produced settlement evidence for the TRBand Baden periods in the eastern loess belt, comprising extensive spreads ofpits and enclosure ditches but no detailed evidence of settlement or houselayout (Milisauskas and Kruk 1993).

Archaeological evidence of later Neolithic settlement from the westernpart of the loess belt is variable; here again, settlement remains are oftenconned to pits and ditches, without any clear evidence of house or settle-ment layout. The proliferation of monumental earthworks and hilltop enclos-ures in various regions suggests increasing concern with communal defence(Hodder 1990: 15861). Sites of the Michelsberg culture extending fromthe lower Rhine to the Swabian Alb and from eastern France to Bohemiaand Moravia consist of substantial earthworks, in some cases with evidenceof settlement in the enclosed area (Keefer 1993: 149).

Well-preserved settlements of the Aichbhl and Schussenried cultures ofsouth-west Germany (the latter with close links in ceramic tradition toMichelsberg) have been excavated on loess (e.g. Hochdorf ) as well as underwaterlogged conditions off loess (e.g. Ehrenstein), revealing closely spaced,post-built houses, of smaller dimensions than the earlier longhouses, eachcontaining a hearth and baking oven (Keefer 1993: 12845). In lowerBavaria, settlements of the Altheim culture, some with enclosures, have asimilar layout (Ottaway 1999: 250). Settlements of related ceramic traditionare also known further south, in lakeshore sites of the Alpine Foreland (seep. 18).

Settlement evidence from the nal phases of the Neolithic (c. 35002200bc, e.g. Horgen, Cham, Baden, Globular Amphora, Corded Ware and BellBeaker cultures) is limited in the loess belt (Rieckhoff 1990: 4862; Keefer1993: 161; Ottaway 1999: 2518; Lning 2000: 1920). Waterloggedsettlements of the Alpine Foreland provide the best evidence of houseand settlement layout for this vast period (see below). Together with theGlobular Amphora culture, the Corded Ware and Bell Beaker complexeshave been associated with pastoral nomadism due to the predominance ofburial sites and lack of settlement evidence (Kruk 1973, 1980: 5861,1988), but there is no positive evidence for reliance on herding (Milisauskasand Kruk 1989a, 1989b; Keefer 1993: 16970). More recently, the conceptof deliberate homogenization of material culture has to some extent replacedmigrationist interpretations (Shennan 1986; Hodder 1990: 175; Rieckhoff1990: 4857).

Bioarchaeological evidence from later Neolithic sites in the loess beltindicates essentially the same range of crops as that attested in the middle


18

Neolithic (Lning 2000: 667). Cattle bone assemblages from Michelsbergand Chassen sites in the western loess belt suggest a general increase in themaintenance of older animals compared with the earlier Neolithic, possiblya reection of dairying and/or use for traction (Arbogast 1994: 96). Nitrogenisotope analysis of cattle teeth from the Chassen site of Bercy (early fourthmillennium bc) is consistent with early weaning to increase the amount ofmilk available for human consumption (Balasse and Tresset 2002). To theeast, the TRB-Baden site of Bronocice has yielded possible evidence forcattle traction and wool production (Milisauskas and Kruk 1989a, 1991).

South of the loess, in the Alpine Foreland, the sequence of lakeshore settle-ments preserved by waterlogging begins c. 4300 bc and continues in someareas through to the Corded Ware phase (c. 2400 bc) (Schlichtherle 1995;Whittle 1996a: 21619; Ptrequin et al. 1998; Gross-Klee 1997; Schibler andJacomet 1999). Rather than an edge effect of settlement concentrated ondry soils, lakeshore settlement appears to reect a real preference for wetareas (lake and marsh edges) in the later Neolithic, along with houses thatare smaller than the earlier longhouses and coherent villages, often surroundedby fences or palisades, rather than loose groupings. High resolution dating bydendrochronology shows that the occupation of these villages was relativelybrief, ranging from less than twenty years up to eighty years.

The good preservation of lakeshore settlements in the Alpine Forelandappears to demonstrate independent household production at some sites.Such evidence is particularly clear at Hornstaad-Hrnle IA, an early lakeshoresettlement on Lake Constance (dendrochronologically dated to 3915 bc),where each house had its own crop stores (charred in the destruction ofthe settlement by re) and a standard tool kit, including wooden handards (Frchenstcke), polished stone axes, int points and shing equipment(Dieckmann 1991; Dieckmann et al. 1997). At other sites, crop stores werelocated away from houses in separate structures outside the village proper, inpart perhaps as protection from re (Ptrequin and Ptrequin 1995). Thesmall size of houses in Goldberg III sites (Alleshausen-Tschenwiesen andAlleshausen-Grundwiesen) in the Federsee region, dated to the end of theNeolithic, appears to preclude household crop storage (Schlichtherle 1995,1997b). Furthermore, botanical analyses at Alleshausen-Grundwiesen appearto indicate site specialization in ax production (Maier and Schlichtherle1993; Schlichtherle 1997b). These developments at the end of the Neolithichave been interpreted as evidence that independent household productionwas being eroded (Schlichtherle 1995, 1997b).

The crop spectrum of Neolithic lakeshore settlements resembles that ofthe later Neolithic in the loess belt, but with notable emphasis on free-threshing wheat prior to the Horgen period in contrast to the focus onemmer and einkorn wheat in the loess belt from the early Neolithic onwards( Jacomet and Kreuz 1999: 302). Mortality curves for cattle assemblagesfrom the Alpine Foreland suggest dairying at sites of the Pfyn-Cortaillod

19


cultures (Higham 1967; Becker 1981; Jacomet and Schibler 1985; Halstead1989a; Gross et al. 1990; Hster-Plogmann and Schibler 1997), though thelack of large-scale pasture and the labour intensity of fodder collection wouldhave limited the scale of animal husbandry (Hster-Plogmann et al. 1999).

The lakeshore settlement at Weier (Pfyn period) has provided the earliestdirect evidence of stalling in the study area (Rasmussen 1989; Robinson andRasmussen 1989; Overgaard Nielsen et al. 2000). Such management practiceswould have increased the availability of milk for human consumption byencouraging the let down of milk (Halstead 1998). Modelling of the humandiet suggests, however, that crops remained the chief food source (Grosset al. 1990; Schibler and Brombacher 1995). Evidence for a crisis in foodproduction c. 37003600 bc (Pfyn period) at various lakeshore settlements(e.g. at Lake Zurich and Lake Biel) suggests that declining crop yields weresupplemented not by intensication of animal husbandry but by higherlevels of hunting and foraging (Schibler et al. 1997a, 1997b; Hster-Plogmannet al. 1999).

In the subsequent Horgen levels of lakeshore settlements at Lake Zurich,age/sex data for cattle show evidence of use as work animals, coinciding withthe rst evidence of wheeled vehicles and yokes in the region (Hster-Plogmann and Schibler 1997; Schibler and Jacomet 1999); as discussedfurther in Chapter 2, this evidence has been associated with more extensivearable cultivation and greater availability of land for grazing. It appears thatcows and bulls were used for traction in the Horgen period, whereas osteo-logical data from Corded Ware (and Early Bronze Age) contexts at LakeZurich suggest increased use of oxen (Hster-Plogmann and Schibler 1997).Sheep mortality data from Corded Ware contexts at Lake Zurich are con-sistent with milk production (Hster-Plogmann and Schibler 1997).

Summary

The study area consists of two adjacent regions with extensive Neolithic(c. 55002200 bc) archaeobotanical datasets: the westerncentral portionsof the loess belt (extending from the Low Countries in the west to Poland,Slovakia and Hungary in the east) and the Alpine Foreland.

Residential household groups are visible throughout the period underconsideration, from the longhouses of the earlymiddle Neolithic (whichvary considerably in size and form) through to smaller, one- or two-roomstructures of the later Neolithic. Exceptional preservation at the laterNeolithic lakeshore settlement of Hornstaad-Hrnle IA provides clearevidence for household-level production and consumption.

In general, Neolithic settlements were too small to be demographicallyviable, and the importance of interaction between settlements is borneout by regional similarities in material culture and long-distance tradein materials such as Spondylus ornaments.


20

The early Neolithic crop spectrum, dominated by emmer and einkornwheat, is narrow compared with that known from Neolithic sites insouth-east Europe. Crop spectra of the middle and later Neolithic werebroadened by the addition of other cereal species. Cattle generally dom-inate animal bone assemblages throughout the Neolithic in the studyarea, with variability in the relative importance of pig versus sheep/goat.

According to the secondary products revolution model, the general ex-pansion of later Neolithic settlement beyond the loess was conditionedby use of animal traction and the ard (scratch plough), as well as by theemergence of milking and wool production.

21

M O D E L S O F C R O P H U S B A N D R Y

2

MODELS OF CROP HUSBANDRYIN NEOLITHIC CENTRAL

EUROPE

Introduction

The aim of this chapter is to summarize and discuss crop husbandry modelspreviously applied to the Neolithic in the study area. Crop husbandry regimesare often characterized as intensive or extensive in the archaeological liter-ature, but the denition of these terms varies (Halstead 1992a). In thisbook, intensive husbandry refers to regimes involving high inputs of labourper unit area, resulting in high area yields; extensive regimes involve smallerinputs of labour per unit area, resulting in smaller area yields (Slicher vanBath 1963: 2403; Upton 1976: 196; Grigg 1984: 49, 174).

Shifting cultivation

Shifting cultivation (also known as slash-and-burn, swidden, long-fallow orforest-fallow) involves the clearance of primary or secondary woodland, usu-ally by burning, and cropping of the newly cleared soil for one to ve years.While new plots are cleared and cultivated, old plots are left to regeneratefor twenty years or more. Burning both reduces the need for tillage andweeding by damaging the viability of seeds or rhizomes in the soil (Ellenberg1996: 770) and mobilizes nutrients from organic material, resulting in highcrop yields over the short term (Sigaut 1975: 1829, 99). In tropical regions,with high rainfall and rapid leaching of soil nutrients, shifting cultivation iswidely attested (Grigg 1974: 5774; Bayliss-Smith 1982: 2536; Steensberg1993: 1698). There is also historical evidence for shifting cultivation in partsof Europe and North America (Manninen 1932; Mead 1953; Montelius 1953;Grigg 1974: 623; Sigaut 1975: 1829; Steensberg 1955, 1993: 1516,98153; Larsson 1995; Lning 2000: 524).

While shifting cultivation is generally characterized as an extensive hus-bandry regime, with low labour inputs per unit area (Boserup 1965: 24, 29),clearance work may be considerable (Lning 2000: 524) and high yieldshave generally been assumed. In his description of pioneer farming in southernOntario, Canada, Schott (1936: 169) reports area yields of c. 15003400 kg/ha


22

for the rst wheat crop sown on newly cleared forest soil, with little tillageand no weeding or manuring (Table 2.1). More often, historical yields arereported as seed-yield ratios for example, 2050:1 or even 100:1 (Soininen1959) (Table 2.1). As various authors (Sigaut 1975: 11920; Rowley-Conwy1981; Halstead 1990) have pointed out, however, these seed-yield ratiosmust be interpreted in light of the sowing techniques used and amount ofseed sown. Historical descriptions of shifting cultivation often specify dib-bling (i.e. dropping a few seeds into individual holes), which uses much lessseed corn than broadcasting and tends to produce much higher seed-yieldratios. Where area yield gures are not available, therefore, it is unclear towhat extent high seed-yield ratios translate into high area yields, or whetherhigh seed-yield ratios were caused primarily by the efciency of dibblingrather than the quality of growing conditions per se. The results reportedfrom the experiments at Draved (Steensberg 1979) and Butser (Reynolds1977) do not suggest spectacular area yields compared with intensive per-manent cultivation (Table 2.1). On the other hand, Rsch et al. (2002)report high area yields of up to 2500 kg/ha and 4000 kg/ha in the rstcultivation season after clearance and burning from two sites near Stuttgart(Wackershofen, Forchtenberg) where experimental shifting cultivation wasconducted (Table 2.1).

Archaeobotanical weed evidence has played a limited role in debate overthe importance of shifting cultivation in Neolithic Europe (cf. Engelmark1989; Dennell 1992), in part due to the lack of modern comparative data onthe arable weed oras that develop under a shifting cultivation regime. Fewobservations on the weed oras growing with crops in shifting elds arefound in historical descriptions of shifting cultivation in Europe and NorthAmerica. Though weed growth may be limited by burning and/or by luxuri-ant crop growth in the rst cultivation season, there are indications thatweed growth increases in the second and third cultivation seasons (Sigaut1975: 1829, 99; Engelmark 1995). Historical accounts provide very littleinformation on the actual composition of these weed oras, and most relateto shifting cultivation on poor soils in coniferous woodland areas; weed orasin deciduous woodland may be quite different (Engelmark 1995).

Shifting cultivation in the earlymiddle Neolithic

Childe (1929: 456) invoked shifting cultivation to explain the dispersalof LBK farming communities across central Europe. Despite the emergenceof an alternative model of permanent elds cropped on a regular basis(Modderman 1971; Kruk 1973; Lning 1980, 2000: 4950, 1879; Sherratt1980, 1981; Rowley-Conwy 1981; Dennell 1983: 172; Barker 1985: 1413;Bogucki 1988: 7982), the shifting cultivation model has continued toinuence discussion of earlymiddle Neolithic cultivation in the loess belt(Sangmeister 1983; Ammerman and Cavalli-Sforza 1984: 43, 114; Wasylikowa

23


Tab

le 2

.1E

xper

imen

tal,

hist

oric

al a

nd e

thno

grap

hic

data

on

cere

al y

ield

s an

d ar

chae

olog

ical

est

imat

es o

f cr

op a

nd c

ulti

vate

d ar

ea r

equi

rem

ents

Sour

ce

Exp

erim

enta

l da

ta

shi

ftin

g cu

ltiv

atio

n:D

rave

d ex

peri

men

tB

utse

r ex

peri

men

t

Wac

kers

hofe

n an

dFo

rcht

enbe

rgex

peri

men

ts

Exp

erim

enta

l da

ta

con

tinu

ous

cere

als:

At

Wob

urn

1877

192

0

Are

a yi

eld

(kg/

ha)

no

retu

rn16

00 y

ear

1, 1

400

year

2,

900

year

3up

to

2500

and

4000

(ye

ar 1

),m

inim

al-z

ero

yiel

d in

lat

erye

ars

480

163

0(m

ean

820)

840

275

0(m

ean

1640

)11

41

950

(mea

n 93

0)11

003

560

(mea

n 21

50)

Cro

p hu

sban

dry

Whe

at,

nom

anur

eW

heat

, m

anur

ed

Bar

ley,

no

man

ure

Bar

ley,

man

ured

Ref

eren

ce

Stee

nsbe

rg (

1979

)R

eyno

lds

(197

7)

Rs

ch (

2000

a),

Rs

ch e

t al

.(2

002)

Rus

sell

and

Voe

lcke

r(1

936)

*

Kg/

pers

on(%

die

t)Se

ed-y

ield

rati

o

>20

:1

ha/h

ouse

hold

No.

peo

ple/

hous

ehol

d


24

Tab

le 2

.1co

ntin

ued

Sour

ce

Exp

erim

enta

l da

ta

con

tinu

ous

barl

ey:

At

Rot

ham

sted

1852

196

2

Exp

erim

enta

l da

ta

whe

at/b

are

fall

ow:

At

Rot

ham

sted

1852

196

4

Exp

erim

enta

l da

ta

row

-sow

n ce

real

s:G

row

n fo

r15

yea

rs a

tLi

ttle

But

ser

His

tori

cal

data

s

hift

ing

cult

ivat

ion:

Fore

st f

arm

ing

in C

anad

aSh

ifti

ng c

ulti

vati

onin

Fin

land

Ref

eren

ce

Rot

ham

sted

Exp

erim

enta

lSt

atio

n (1

970)

Rot

ham

sted

Exp

erim

enta

lSt

atio

n (1

970)

Rey

nold

s (1

992)

Scho

tt (

1936

)

Soin

inen

(19

59)

Cro

p hu

sban

dry

No

man

ure

Man

ured

No

man

ure

Man

ured

No

man

ure,

3x

hoei

ng d

urin

gcr

op g

row

th

As

abov

e, e

mm

erro

tate

d w

ith

broa

d be

ans

Are

a yi

eld

(kg/

ha)

450

150

0(m

ean

928)

1880

351

0(m

ean

2960

)

800

238

0(m

ean

1410

)18

703

550

(mea

n 28

40)

Em

mer

: 36

03

030

(mea

n 16

50);

Spel

t: 6

702

610

(mea

n 14

90)

780

311

0(m

ean

2080

)

1500

340

0in

rs

t ye

ar

Seed

-yie

ldra

tio

30:1

30:1

205

0:1,

up t

o10

0:1

Kg/

pers

on(%

die

t)N

o. p

eopl

e/ho

useh

old

ha/h

ouse

hold

25


His

tori

cal

data

e

xten

sive

cer

eal

cult

ivat

ion:

Med

ieva

l E

urop

e(8

001

150 ad

)H

isto

rica

l da

ta f

rom

cent

ral

Eur

ope

(med

ieva

l pe

riod

,18

th c

entu

ry)

Stat

isti

sch-

Top

ogra

phis

ches

Bur

eau

for

1850

190

5

Eth

nogr

aphi

c da

ta

int

ensi

ve c

erea

l cu

ltiv

atio

n:A

stur

ias,

NW

Spa

in

Oth

er e

stim

ates

for

pre

hist

oric

cer

eal

cult

ivat

ion:

Not

es:

*Run

of

bad

year

s in

192

0s a

t W

obur

n om

itte

d.

Yie

ld d

ata

repo

rted

as

ten-

year

ave

rage

s.

250

200

for

mal

ead

ult

(55

65

% o

f di

et)

65

80%

die

t20

0 (8

0% d

iet)

150

(65%

die

t)

100

(52%

die

t)

100,

150

or

200

(50

100

%di

et)

300

Hen

ning

(19

94)

Lni

ng(1

979/

1980

),T

egtm

eier

(199

3: 5

)

Gre

gg (

1988

)

P.

Hal

stea

d,e

ld n

otes

Hal

stea

d (1

981a

,19

87)

Bak

els

(198

2)M

ilis

ausk

as a

ndK

ruk

(198

9b)

Jaco

met

et

al.

(198

9: 9

01

)G

ross

et

al.

(199

0)B

illa

mbo

z et

al.

(199

2)

Hal

stea

d (1

995)

450

550

900

800

(ran

ge65

01

050)

(800

)17

001

900

800

100

0

800

500

600

650

600,

800

1000

2.5

3:1

3:1

4:1

5:1

5:1

Man

urin

g, h

and

wee

ding

6 4

8

61

05 5

6

5 5 5

c. 2.

5, o

r3.

75w

ith

bare

fall

ow1.

63

.2

1.5

32 1.

52

1 12

23


26

et al. 1985; Beranova 1987, 1989; Godlowska et al. 1987; Kruk 1988;Milisauskas and Kruk 1989a; Wasylikowa 1989; Rsch 1990a; Rulf 1991;Whittle 1996a: 1602, 1996b, 1997; Gerht et al. 2002). Though Childe(1929) originally linked shifting cultivation with the spread of migrantfarmers across Europe, it has come to be associated with the indigenous,Mesolithic (and hence mobile) identity of Europes rst farmers; the latterassociation appears to underlie Whittles (1996a, 1996b) recent characteriza-tion of LBK communities as both more indigenous and more mobile thanpreviously thought.

Arguments advanced in support of earlymiddle Neolithic shifting cul-tivation have included the lack of tell formation (Childe 1929: 456, 1957:1056) and apparent evidence for discontinuity in settlement occupation(Soudsky and Pavlu 1972), the assumption that relatively good soils wouldbe rapidly exhausted (Childe 1929: 456, 1957: 1056) whereas relativelypoor soils could not be improved (Kruk 1973, 1980: 547, 1988) andpollen evidence for changes in woodland composition linked with clearanceand burning (Wasylikowa et al. 1985; Wasylikowa 1989; Godlowska et al.1987; Rsch 1990a). All of these arguments are open to question. First, theabsence of tells is easily explained by the lack of mudbrick architecture(Sherratt 1981). Second, the tendency of longhouses to drift horizontally,avoiding overlap with earlier structures, can explain apparent discontinuityin settlement occupation (Modderman 1970: 20811, 1971). Third, assump-tions of soil exhaustion appear unwarranted given experimental evidencefrom Britain and Germany for the long-term stability of crop yields overdecades of continuous cultivation on relatively good soils (Lning 1980,2000: 174; Rowley-Conwy 1981; Reynolds 1992); at the same time, manuring,watering and weeding of cultivation plots can greatly enhance crop-growingconditions (G. Jones et al. 1999). Fourth, in addition to the need for adequatedating and appropriate calculation of pollen diagrams (Rowley-Conwy 1981;Kalis and Meurers-Balke 1998), apparent clearance and burning episodes donot necessarily reect past arable land use. These changes could, for example,relate to the management of separate woodland or grazing areas (Rowley-Conwy 1981; Brombacher and Jacomet 1997; Kalis and Meurers-Balke 1997,1998).

Critics of the slash-and-burn model for the earlymiddle Neolithic havealso emphasized differences between the environmental context of tropicalswidden cultivators (e.g. thin, rapidly leached soils and high rainfall) andthat of early farmers in central Europe (Modderman 1971; Jarman and Bay-Petersen 1976). A further contrast can be drawn between the loess belt andareas of northern Europe where historical shifting cultivation was associatedwith marginal soils and limited availability of good arable land (Sherratt1980; Rowley-Conwy 1981). Moreover, shifting cultivation was often a formof outeld cultivation in marginal areas of northern Europe, practised along-side a more intensive form of ineld cultivation (Rowley-Conwy 1981).

27


In the absence of modern comparative data on the sort of weed ora thatdevelops under a shifting cultivation regime, the same archaeobotanical evid-ence has been used to support different conclusions. Knrzer (1971) inter-preted the repeated association between a narrow range of weed species (theso-called Bromo-Lapsanetum praehistoricum weed community) and charredcrop material on LBK-Rssen sites in the lower Rhine basin as evidence forpermanent elds cultivated each year using the same methods. Bakels (1978:69), on the other hand, has argued that repetition of the same weed assemblagecould reect a shifting cultivation regime in which areas chosen for clearance,methods of clearance and sowing, etc. were consistent. A particular focus ofconicting interpretations is Nipplewort (Lapsana communis L.), which con-stitutes one element of the Bromo-Lapsanetum assemblages. It has beensuggested that Lapsana communis indicates long-fallow cultivation by virtueof its shade tolerance and hence its ability to grow in heavily shaded shiftingplots in woodland (Beranova 1987; cf. Whittle 1997). Other authors, however,have interpreted this species as an indicator of permanent cultivation plotsshaded by surrounding hedges (Knrzer 1967, 1971, 1988; Groenman-vanWaateringe 1971) (see also p. 39). While a permanent eld model hastended to nd favour in recent archaeobotanical studies of earlymiddleNeolithic crop husbandry in central Europe (Kreuz 1990; Stika 1996), therehas also been acknowledgement of the difculty of excluding shifting cul-tivation on archaeobotanical grounds (Brombacher and Jacomet 1997).

A different approach to the inference of cultivation on newly clearedforest soil is to treat the absence of weed seeds in archaeobotanical cropsamples as indicative. This is based on the observation that weeds may besuppressed in the rst cultivation season following woodland clearance(p. 22). Bakels (1991b) has suggested that weed-poor crop samples tend toderive from LBK sites established in new areas (i.e. without previous cul-tivation) whereas samples from sites in established settlement areas tend tocontain more weed seeds, reecting the continuous cultivation of plots. Anobvious problem with this line of reasoning is that crop material may befree of weed seeds for a variety of other reasons (e.g. crop processing, hand-weeding of crops, preservation, etc.). Furthermore, as Bakels (1991b) makesclear, initial woodland clearance would be necessary under any cultivationregime.

Shifting cultivation in the later Neolithic

It has recently been claimed that shifting cultivation formed the principal crophusbandry regime of later Neolithic lakeshore communities in the AlpineForeland (Bocquet et al. 1987; Rsch 1987, 1989, 1990a, 1990b, 1996, 2000a;Ptrequin 1996; Bailly et al. 1997; Ptrequin et al. 1998; Rsch et al. 2002;see also Schlichtherle 1989, 1992, 1995, 1997a; Whittle 1996a: 21622),though the actual weed assemblages accompanying charred crop stores from


28

lakeshore sites have been interpreted as evidence of xed-plot cultivation( Jacomet et al. 1989: 234; Brombacher and Jacomet 1997; Maier 1999, 2001:78109) (see also pp. 4041). The main arguments in favour of shiftingcultivation are based on more indirect forms of evidence. Rsch interpretspollen and microscopic charcoal sequences from the Lake Constance area asevidence of cyclical changes in woodland composition and burning, respect-ively, and argues that shifting cultivation was dominant through to theend of the Neolithic (Rsch 1990b), or at least during the earlier part of thelater Neolithic (Rsch 1996, 2000a). He reasons that shifting cultivationin the later Neolithic was necessitated by the deterioration of soils as a resultof xed cultivation without manuring or fallowing in the earlymiddleNeolithic and argues further that shifting cultivation contributed to poorsoil conditions in the Bronze Age (Rsch 2000a). Ptrequin (1996; seealso Bailly et al. 1997; Ptrequin et al. 1998) infers shifting cultivationprior to c. 3000 bc for lakeshore settlement in the French Jura, mainlyon the basis of the age and species composition of house timbers from siteson the shores of the Clairvaux and Chalain lakes. Both approaches are opento criticism since the link between the evidence cited and arable land useis tenuous; the pollen and timber evidence may instead reect woodlandmanagement practices related to animal husbandry, for example (Rowley-Conwy 1981; Kalis and Meurers-Balke 1998; Lning 2000: 502; Rschet al. 2002).

Shifting cultivation has also been identied as the major form of laterNeolithic crop husbandry in loess areas such as southern Poland (TRB culture Kruk 1973, 1980: 547, 1988) and the Paris basin (SeineOiseMarneculture Howell 1983). This is based on the association of settlement inthis period with interuves (see Chapter 1) and the assumption that handcultivation of xed plots on the drier upland loess would be impractical. Inaddition, Howell (1983) cites the occurrence of charcoal layers as evidence ofslash-and-burn cultivation in north-west France, though initial clearance byburning could relate to other forms of land use.

Extensive ard cultivation

Cultivation with the animal-drawn ard (scratch-plough) requires less humanlabour per unit area than cultivation by hand (Halstead 1995; Lning 2000:181). Ethnographic evidence indicates that ard cultivation also results inless-thorough tillage, unless it is accompanied by hoeing (G. Jones et al.1999; cf. Halstead 1995). Area yields are low compared with hand cultiva-tion regimes (Gallant 1991: 51; Halstead 1995) (Table 2.1). Particularlywhen specialized plough oxen are used, however, the total area undercultivation is considerably larger than that worked under a hand cultivationregime, allowing the production of surplus on a large scale (Goody 1976;Halstead 1995).

29


Extensive ard cultivation in the earlymiddle Neolithic

According to Lning (1979/80, 1980, 2000: 1601, 163, 181), Lning andStehli (1989) and Tegtmeier (1993: 5), LBK cereal production must havetaken place on a relatively large scale, with the help of an ox-drawn ard, inorder to provide the staple food source. Lning (1979/1980) has calculatedthat a family of six would need to cultivate 2.5 ha of cereals. This is based onan annual requirement of 250 kg of cereals per person and cereal yields of900 kg/ha, of which 300 kg is reserved for seed corn (Table 2.1). If eldswere left fallow every third year as in a medieval three-eld system, the totalcultivation area per family would be 3.75 ha (a bare, or cultivated, fallow isapparently assumed here cf. Lning 1980). Lning (1979/1980) claims thata household would need an ard to cultivate this area: given a conservativework rate for ard cultivation of 5001000 square metres per day, at leastthirty work days would be needed to cultivate 3.75 ha (cropped plus barefallow area) once, and in fact two to three cross-ploughings with the ardwould be likely. The cultivation area could easily be accommodated withinthe 10 ha estimated to have been available for each LBK longhouse in theMerzbach valley of the Aldenhoven Plat

bogaard, amy - neolithic farming in central europe, an archaeobotanical study of crop husbandry...

Documents

neolithic central europe

neolithic society

neolithic centraleurope

neolithic periodeurope

models of crop husbandry

nature of crop husbandry

permanence of cultivation

major crop husbandry