Ancient Plants and People: Contemporary Trends in Archaeobotany 9780816527106, 2014007785

Table of contents :
Cover
Title Page, Copyright
Contents
Acknowledgments
Preface
Introduction
1. Sample-Size Estimation and Interassemblage Quantification in Archaeobotany
2. Regional Exchanges in Southeastern Arabia during the Late Pre-Islamic Period: Phytolith Analysis of Ceramic Thin Sections from ed-Dur (UAE)
3. Examining Agriculture and Climate Change in Antiquity: Practical and Theoretical Considerations
4. Swahili Urban Food Production: Archaeobotanical Evidence from Pemba Island, Tanzania
5. Plant Food Subsistence in Context: An Example from Epipaleolithic Southwest Anatolia
6. Vegetation Proxy Data and Climate Reconstruction: Examples from West Asia
7. Significance of Prehistoric Weed Floras for the Reconstruction of Relations between Environment and Crop Husbandry Practices in the Near East
8. Historical Aspects of Early Plant Cultivation in the Uplands of Eastern North America
9. Routine Activities, Tertiary Refuse, and Labor Organization: Social Inferences from Everyday Archaeobotany
10. Of Crops and Food: A Social Perspective on Rice in the Indus Civilization
11. Anthracological Research on the Brazilian Coast: Paleoenvironment and Plant Exploitation of Sambaqui Moundbuilders
12. Rice of Asian Origin
13. A Review of the Research on the Origin of Six-Row Barley
14. Maize Cob Phytoliths as Indicators of Genetics and Environmental Conditions
Editors and Contributors
Index

Citation preview

Ancient Plants and People

Ancient Plants and People Contemporary Trends in Archaeobotany

Edited by Marco Madella, Carla Lancelotti, and Manon Savard

tucson

The University of Arizona Press www.uapress.arizona.edu © 2014 The Arizona Board of Regents All rights reserved. Published 2014 Printed in the United States of America 19 18 17 16 15 14   6 5 4 3 2 1 Jacket designed by Nicole Hayward Library of Congress Cataloging-in-Publication Data Ancient plants and people : contemporary trends in archaeobotany / edited by Marco Madella, Carla Lancelotti, and Manon Savard. pages cm Summary: “Ancient Plants and People is a timely discussion of the global perspectives on archaeobotany and the rich harvest of knowledge it yields. Contributors examine the importance of plants to human culture over time and geographic regions and what it teaches of humans, their culture, and their landscapes”—Provided by publisher. Includes bibliographical references and index. ISBN 978-0-8165-2710-6 (cloth : alkaline paper) 1. Plant remains (Archaeology) 2. Phytoliths. 3. Social archaeology. 4. Landscape archaeology. 5. Archaeology—Methodology. 6. Food crops— History—To 1500. 7. Plants and civilization—History—To 1500. I. Madella, Marco. II. Lancelotti, Carla. III. Savard, Manon, 1970– CC79.5.P5A525 2014 930.1028’2—dc23 2014007785 This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).

Contents

Acknowledgments

vii

Preface Martin K. Jones

ix

Introduction 3 Manon Savard, Marco Madella, and Carla Lancelotti I. Methodologies in Archaeobotany 1. Sample-Size Estimation and Interassemblage Quantification in Archaeobotany Gyoung-Ah Lee

9

2. Regional Exchanges in Southeastern Arabia during the Late Pre-Islamic Period: Phytolith Analysis of Ceramic Thin Sections from ed-Dur (UAE) Luc Vrydaghs, Paul De Paepe, Katrien Rutten, and Ernie Haerinck

26

3. Examining Agriculture and Climate Change in Antiquity: Practical and Theoretical Considerations Alexia Smith

47

II. Case Studies in Archaeobotany and Vegetation History 4. Swahili Urban Food Production: Archaeobotanical Evidence from Pemba Island, Tanzania Sarah C. Walshaw 5. Plant Food Subsistence in Context: An Example from Epipaleolithic Southwest Anatolia Danièle Martinoli

v

73

100

Contents

vi

6. Vegetation Proxy Data and Climate Reconstruction: Examples from West Asia Naomi F. Miller

120

7. Significance of Prehistoric Weed Floras for the Reconstruction of Relations between Environment and Crop Husbandry 135 Practices in the Near East Simone Riehl III. Social Archaeobotany 8. Historical Aspects of Early Plant Cultivation in the Uplands of Eastern North America Kristen J. Gremillion

155

9. Routine Activities, Tertiary Refuse, and Labor Organization: 174 Social Inferences from Everyday Archaeobotany Dorian Q Fuller, Chris Stevens, and Meriel McClatchie 10. Of Crops and Food: A Social Perspective on Rice in the Indus Civilization Marco Madella

218

11. Anthracological Research on the Brazilian Coast: Paleoenvironment and Plant Exploitation of Sambaqui Moundbuilders 237 Rita Scheel-Ybert and Maria Dulce Gaspar IV. Genetics in Archaeobotany 12. Rice of Asian Origin Yo-Ichiro Sato

265

13. A Review of the Research on the Origin of Six-Row Barley Ken-ichi Tanno

277

14. Maize Cob Phytoliths as Indicators of Genetics and Environmental Conditions Linda Scott Cummings

292

Editors and Contributors

301

Index

305

A c knowledgments

The editors are grateful to Paule Maranda for having edited part of the manuscript, saving us much-needed time for teaching and research tasks. Also, we would like to thank the many reviewers that read the book chapters or the full book. We were impressed by the amount of time and care they put into this task, and we all deeply benefited from the high intellectual standard of their comments. Without the help of these anonymous scholars, this book never would have seen the light of day. Finally, we would like to thank Allyson Carter for her friendly and supportive help during the preparation of the volume.

vii

Preface

It has sometimes been argued that the three-age system of stone, bronze, and iron misrepresents ordinary lives in the past. Much of prehistory, and indeed early history, might better be designated “the age of wood.” Trees, herbs, and a wide variety of plant tissues have dominated both the natural and created environments, as well as the food chains and material culture, of our own species for as long as we have tangible evidence for people-plant relations. That tangible evidence, however, has always been the limiting factor in archaeological research. The stone and metal objects around whose typologies prehistory has been arranged were chosen for their visibility and durability in the archaeological record. Plant-based materials have proved more elusive, even though appropriate sampling techniques show them to be more ubiquitous than any other single category of find, albeit in a fragmented and transformed state. The suite of papers brought together in this volume aptly conveys the growing potential of applying those sampling techniques and the many avenues of inquiry that can be followed. The editors refer in their introduction to a radical change over the last couple of decades after which archaeobotanical analyses “are now carried out routinely and the results of these studies are profoundly discussed for their significance” (3). In the context of their comments, it is interesting to reflect in this preface upon that period of change, in relation to selected publications of two decades back. A reasonable amount was coming into print during the 1980s, including such seminal works as the report on the Zagros flanks in the Old World (Braidwood et al. 1983) and Guila Naquitz in the New (Flannery 1986). Such important works immediately convey the emphasis upon agricultural origins in selected “centers of civilization” that very much characterized archaeobotanical research, particularly in the English-speaking world, at the time. Two other edited volumes of archaeobotanical research, Van Zeist and Casparie (1984) and Hillman and Harris (1989), allow us to place the emphasis upon this core endeavor within its wider context of contemporary archaeobotanical research. The first of these, Plants and Ancient Man, resulted from a symposium of the International Work Group for Palaeoethnobotany (IWGP) held the

ix

x

Preface

year before in Groningen, Holland. The IWGP had been the major forum for a very active community of Central and Eastern European archaeobotanists performing groundbreaking work, leading the field in many cases, and very often publishing their results in German. A combination of the limited consultation of German texts by English speakers and a traditionally guarded and cautious writing style led to an unfortunately low citation of many of their significant contributions within the growing English-language discourse of archaeology. I have argued elsewhere that many of the innovations of archaeobotanical methodology and approach explored by English-language authors of the 1970s and ’80s can be found in the brilliant archaeobotanical study of Feddersen Wierde by Udelgard Körber-Grohne, published in 1968. Plants and Ancient Man interestingly reflects the transition to Englishlanguage domination of the discourse, with only six out of thirty papers published in German. It also reflects a series of concerns that extend beyond the core endeavor of agricultural-origin studies, with sections on archaeobotany of historic periods, and on landscape and regional studies within the Bronze Age and Iron Age. The section on methodology deals with some perennial issues of statistical sampling and ecological analysis of weeds. However, the real methodological contribution comes earlier in the volume, in three papers, by Hillman, G. Jones, and Harris, that can truly be said to have transformed all subsequent archaeobotanical research by exploring the rich vein of interaction between ethnology and taphonomy. If we look beyond these substantial contributions to the limits of this volume, and by implication the limits of archaeobotanical research at the time, they seem quite clear; the work clearly emphasized one particular aspect of plant utilization by a certain type of society in certain parts of the world. The singular aspect was plant foods, particularly seeds and fruits; the type of society was agricultural; and the regions of the world clearly had Europe at their epicenter. While twenty-nine of the thirty contributions concerned the Old World, I think it is fair to say that a parallel series of emphases could be discerned within New World archaeobotany at the time. Published five years later, Harris and Hillman’s Foraging and Farming tackled two of those limitations head-on. It brought together papers delivered at the World Archaeological Congress, held in Southampton in 1986, and it remains a quite outstanding volume in terms of its global remit, bringing together contributions from every vegetated continent. It also pushed back that conceptual boundary within archaeobotany beyond the origins of agriculture, with seven papers on preagricultural and nonagricultural topics. As a necessary consequence of this, it also probed beyond the traditional archaeobotanical territory of seeds and fruits, though the real breakthroughs in the study of roots and tubers would come later

Preface

xi

than this publication. Ironically, with a small number of important exceptions, for all this breaking of boundaries, the most enduring elements of Foraging and Farming pertain to those traditional archaeological concerns of understanding the transition to agriculture in fairly familiar regions of the world. So that was the state of play of archaeobotany two decades ago, as measured by the barometer of a few conspicuous publications. Against that background, what can be said about the contemporary research presented in this volume? It is certainly true that the research here reported builds eloquently on the achievements of the 1980s, continuing theoretical explorations into the origins of agriculture and taking the ethnographictaphonomic element of methodology for granted. The geographical range is quite as global as in Foraging and Farming, and the probing back beyond agriculture to earlier epochs clearly evident, displaying the benefit of methodological advances that followed from that volume. There was a third conceptual boundary that seemed to me evident on rereading Plants and Ancient Man, and that was the emphasis on one particular aspect of plant utilization, food, particular from seeds and fruits. It is that third boundary that is repeatedly breached in this volume, and to great effect. The breaching of this boundary is manifest in two ways. In methodological terms, the chapters span a broad range of plant data sources, not just seeds and fruits but also charcoal, phytoliths, and genetics, each with its own distinct reflection of people-plant interactions. In conceptual terms, a recurrent feature of the chapters is a sophisticated interplay between the use of plants and their social or their environmental context. The intimate interplay between food, fuel, building and bedding materials, craft, technology, art, and decoration is glimpsed in a number of elegant contributions, for example, the exploration by Vrydaghs et al. on the use of agricultural by-products in the preparation of pottery and Walshaw’s work on the interplay with other arenas of production such as ironworking, pottery making, and animal husbandry. The wider social context of food is a central theme in the contribution from Fuller, Stevens, and McClatchie, and from Madella. A parallel interplay between the diverse human use of plants and engagement with the dynamics of vegetation and climate is a recurrent and important theme in papers by Miller, and Scheel-Ybert and Gaspar. The volume’s three final papers introduce an approach to contextualizing agricultural ecology within evolutionary ecology in a manner unimaginable twenty years ago—through the substantial advances in archaeogenetics. Those explorations, in which people-plant relations are integral and inseparable from their social, ecological, and evolutionary contexts, are to be welcomed, and reflect a real achievement of the last twenty years of archaeobotany. The field is also left with its challenges for future development. As mentioned above, Savard, Madella, and Lancelotti

xii

Preface

(introduction) celebrate the fact that archaeobotanical analyses (or archaeobotanical assessments) are now carried out routinely and the results of these studies are profoundly discussed for their significance. It has to be said that they are still rather more “routine” in some areas of archaeology than in others. As archaeobotany has developed, its application has moved from its core concerns with agricultural origins to a rich application forward in time, to late prehistoric and historic societies. The application backward in time has been much slower, even though the small number of Paleolithic studies, such as the research at Ohalo II (Weiss et al. 2004) and Kebara (Lev, Kislev, and Bar-Yosef 2005), have substantially molded our understanding of human paleoecology. In twenty years’ time, we might expect to see even richer archaeo­ botanical analyses of the diverse people-plant relationships that form the shape and context of human societies and ecologies, built from timber, thatch, wattle, hedgerow, and grassland; fed by woodland fuel and hundreds of plant foods; and with rich material cultures of wooden objects and plant-based textiles and dyes, just occasionally supplemented by objects of stone, bronze, and iron. I suspect we shall also be looking back on the work reported in this volume as marking an important step along that path. —Martin K. Jones

References Braidwood, L. S., R. J. Braidwood, B. Howe, C. Reed, and P. J. Watson. 1983. Prehistoric Archaeology along the Zagros Flanks. Oriental Institute Publications, vol. 105. Chicago: University of Chicago. Flannery, K. V. 1986. “Ecosystem Models and Information Flow in the TehuacánOaxaca Region.” In Guilá Naquitz, edited by K. V. Flannery, 19–28. New York: Academic Press. Harris, D. R., and G. C. Hillman. 1989. Foraging and Farming: The Evolution of Plant Exploitation. London: Unwin Hyman. Körber-Grohne, U. 1968. Geobotanische Untersuchungen auf der Feddersen Wierde. Wiesbaden: Text- und Tafelband. Lev, E., M. E. Kislev, and O. Bar-Yosef. 2005. “Mousterian Vegetal Food in Kebara Cave Mt. Carmel.” Journal of Archaeological Science 32 (3): 475–84. Van Zeist, W., and W. A. Casparie, eds. 1984. Plants and Ancient Man: Studies in Palaeoethnobotany, Proceedings of the Sixth Symposium of the International Work Group for Paleoethnobotany. Rotterdam: A. A. Balkema. Weiss, E., W. Wetterstrom, D. Nadel, and O. Bar-Yosef. 2004. “The Broad Spectrum Revisited: Evidence from Plant Remains.” Proceedings of the National Academy of Sciences USA 101 (26): 9551–55.

Ancient Plants and People

Introduction M a n o n S ava r d , M a r c o M a d e l l a , and Carla Lancelotti

Archaeobotany is not a new discipline, and it has come a long way since daunting lists of plant taxa were simply a final chapter attached to archaeological reports. These chapters lacked any (or had very little) interpretative approach in respect to the social, economical, and environmental significance of the identified plants. Fortunately, this has radically changed in the last few decades, and archaeobotanical analyses (or archaeobotanical assessments) are now carried out routinely and the results of these studies are profoundly discussed for their significance (to get a quick glimpse of the situation, just leaf through the references in this book). Indeed, plant remains recovered from archaeological sites are providing precious insight on past landscapes, human adaptation to climate change, the relationship between human groups and their environment, and food production, distribution, preparation, and consumption. These, in turn, are bringing to light important aspects of past human societies, widening the spectrum of information available to archaeologists to understand our history as a biological and cultural species. With the answers, however, often come more questions and the need for better knowledge. As a result of this very engaging and dialectic research environment, archaeobotanists are pushed to reflect on the methodological and theoretical aspects of their discipline. With Ancient Plants and People we would like to give a portrait of archaeobotany, putting forward the troubles and dilemmas of the discipline but also highlighting its enormous potential and offering perspectives into the future. Indeed, the book presents works of innovative application

3

4

Introduction

and new methods, as well as case studies in those areas where archaeobotanical analyses have entered recently. The book opens with Lee’s chapter discussing quantitative comparability between samples of various sizes and the populations they represent. It might be difficult, for instance, to distinguish between quantitative differences attributed to changes in cultural practices over time and those related to different sample size. To circumvent this problem, Lee presents a mathematical function to standardize the size of samples recovered from different sites and time periods so to produce directly comparable assemblages. A further step in the methodological development of the discipline is given by the work of Vrydaghs, De Paepe, Rutten, and Haerinck on phytolith analysis in pottery thin sections. Vegetal temper is a common component of old and modern pottery, and it is normally sourced from agricultural by-products (e.g., chaff) or wild plants (e.g., the use in some areas of Pakistan of the feathery fruits of the cattail— Typha sp.). The chapter by Vrydaghs et al. presents a case study from southeastern Arabia during the Late Pre-Islamic period that shows how the study of microremains and mineralogy from pottery allows the detection of pottery provenance and regional exchange networks. Smith’s chapter is concerned with a society’s ability to produce food under conditions of global warming and how climate affects population. The author explores how the use of an integrated agricultural data set can help in understanding how agricultural societies in the past managed to adapt to climate change, especially through changes in food production and agricultural practices. With Walshaw’s chapter we move from domestication to the organization of crop production and the beginning of urbanization in an African context. Her chapter also explores the relationships between introduced crops and established agricultural practices, and how these could be maintained depending on where production control is held. Swahili food production systems provide an interesting case study because of the close ties with Asian trade networks. However, the persistence of householdbased plant processing during the beginning and growth of the Swahili African cities is significant, because when considered in conjunction with emerging data from other arenas of production (e.g., ironworking, pottery making, and animal food), it can provide a more complete understanding of local production patterns and their interface with the Indian Ocean exchange network. Martinoli’s chapter aims at assessing optimal foraging theory, according to which foraging choices are oriented toward efficiency in food acquisition, as a result of evolutionary selection pressures. To do so, Martinoli discusses the macroremains assemblages from two Epipaleolithic cave sites from southwest Anatolia, comparing information from

Introduction

5

them with broadly contemporary sites. Focusing on the site of Malyan (Iran) in southwest Asia, Miller’s chapter discusses the importance of scale in the complex interactions between people, plants, and climate. The impact of climate shifts on vegetation and the inertia of vegetation to changes is a topic that still needs deeper understanding. Indeed, plant communities are not mere spectators but influence climate at the local or regional level, and they often maintain their own microclimate even in the face of major climatic shifts. Also, precisely dating climatic shifts and determining synchronicity and correlation between climate and cultural changes are often difficult to do. Miller further demonstrates that while people must have adapted to short- and long-term climate fluctuations, anthropic activities can also influence climate as the vegetation responds to human manipulation of the landscape. Looking at a selection of Near East late Bronze Age and early Iron Age sites, Riehl’s chapter shows how wild plant macroremains recovered from crop assemblages can provide information beyond crop husbandry practices and effectively contribute to environmental reconstructions. Gremillion discusses the role of ecological and evolutionary theories of agricultural origins in the debate over the independent development of food production, based on the study of indigenous weedy species in the Cumberland Plateau (eastern Kentucky). This development predated the introduction of maize and other Mesoamerican crops in North America. Her research puts together data from the archaeobotanical record and the geographic distribution, habitat, and ecology of three floodplain weeds that were eventually domesticated. In doing so, Gremillion brings to light “alternative pathways” of indigenous farming systems and fresh perspectives on domestication processes. Using several case studies from different times and areas of the world, the chapter by Fuller, Stevens, and McClatchie presents a theoretical discussion of archaeological context and archaeobotanical interpretation. While the authors suggest that archaeobotanical assemblages can give little help in understanding particular depositional contexts in terms of human activities, they demonstrate how archaeobotanical evidence can provide an insight on labor organization and on food production strategies. Food has a public and private role, and it permeates the life of all the people in a society. Food choice, production, and distribution probably represent the most complex indicators of social life. The chapter by Madella explores the social values of crops and food by focusing on rice within the social dynamics of the Harappan culture of the Indus valley. Madella investigates the use of plants, past human-plant relationships, production and distribution of foods, all aspects that have so far been neglected in the study of Harappan society. In their chapter Scheel-Ybert and Gaspar present the results of a pioneer work in Brazil, where the wood charcoal of coastal

6

Introduction

mound sites was investigated in an effort to understand the past environment and the anthropic impact of a complex preagricultural people. The research also sheds light on past arboreal vegetation and its stability over time. The results suggest that the plant communities of south and southwestern Brazil were not significantly affected by climate changes or anthropogenic activities. The final three chapters of this book are dedicated to genetic or genetic-­related studies. The works of Sato and Tanno illustrate how, by complementing “traditional” archaeobotanical analyses of charred material with DNA analyses, it is possible to enhance our knowledge of the processes of plant domestication. Sato attempts to identify the center of origin of cultivated rice and to trace its evolutionary trajectory, a much needed endeavor for one of the most cultivated crops in the world, which today as in the past sustains millions of people. Tanno work offers a historical review of the range of hypotheses on the origins of cultivated sixrow barley, along with the archaeological and genetic evidence available to date. The important conclusion of Tanno’s chapter is that although molecular phylogeny based on DNA can provide evidence for the question of “how many origins” a crop has, to know “when” and “where” the crop was domesticated, evidence from “classical” archaeobotany is necessary. Lastly, the work of Scott Cummings combines phytolith analysis, genetics, and an experimental crop field to investigate how microremains like phytoliths, which emerge from this study as a good proxy for maize genetics, have the potential to provide cultural information on population movements, exchange, and trade. A book like the present one is not intended, or able, to be exhaustive. We have shown archaeobotany in its many manifestations. Sometimes these lines of investigation are far apart, but they are all vital for understanding the plant-human relationship, which has been a fundamental aspect of our lineage for several million years. Since we started compiling and editing this book from an original idea born at the Society for American Archaeology Annual Meeting held in Montreal in 2004 and developed over the last few years, much more work has become available, since developments are rapid in our discipline. But our aim still holds: a guide—theoretical and practical—for the searcher approaching the discipline.

C h ap t e r 1

Sample-Size Estimation and Interassemblage Quantification in Archaeobotany G y o u n g -A h L e e

One of the most common questions in archaeobotany is how to attain quantitative compatibility of plant remains regardless of different sample sizes. Statistical consideration of sampling size itself, however, has been rarely discussed, except for a few publications (e.g., Van der Veen and Fieller 1982; Orton 2000). Building on these earlier studies, this study articulates sampling issues in two realms: a representativeness of sample to population; and a quantitative method that can cancel possible preservation difference between samples. Firstly, this paper attempts to design a simple mathematical method that can make the sizes of samples recovered from multiple populations comparable, based on a relative error and a fixed confidence level. The same degree of representativeness prevents the uncertainty of whether quantitative differences of plant remains between two periods results merely from different sample sizes rather than from real changes in cultural practices through time. Thus, this method can strengthen cultural interpretations of quantitative differences in plant remains. Secondly, the paper investigates quantitative measures that can be less sensitive to inherent differences in survival rates between samples. No quantitative measures can be free from some assumptions. If we understand the assumptions underlying the use of quantitative measures, however, we can construct ratios that are appropriate for inter- and intrasite comparisons (Miller 1988).

9

10

methodologies in archaeobotany

Figure 1.1.  Map of study area. The star indicates the location of sites compared in the Nam River valley. The upper left inset shows the location of the Nam River valley (dot) in South Gyeongsang Province (light gray area).

This research uses plant data from several Mumun-period sites in the Nam River valley of Korea (figure 1.1). Mumun, referring to the plain ware of the culture, is often regarded as a period of significant changes in Korea that include the advent of metallurgy, megalithic burials, rice agriculture, and ultimately social inequality over the 1,500 years of its

Sample-Size Estimation

11

course (Lee 2003). Two subphases, Early (1500–900 BP) and Middle (900–400 BP) Mumun, marked the establishment of intensive agricultural settlements, revealing numerous pit houses, burials, farming fields, and protective ditches surrounding the dwelling units (Crawford and Lee 2003). I collected 2,814 L of sediments from 376 features of the Early Mumun and 3,386 L (389 features) of the Middle Mumun sites. Sediment samples were floated either by a manual decanting method or a flotation apparatus, which uses a constant water supply and two sieves to separate floated remains from the heavy, sunken materials. I used a standard geological sieve of 0.212 mm mesh to catch floating materials, and one of 0.8 mm mesh to collect heavy materials. Floated remains are the subject of this study.

Defining a Population and a Sample in Cluster Sampling Most archaeological sampling is subject to cluster sampling. Archaeologists sample areas or volumes of spaces, and often pretend that they actually sample sites, artifacts, bone fragments, or plant remains (Banning 2000, 80). In such cases, objects of interest for analysis (carbonized seeds in this research) are not sampled directly and randomly from a space but are contained within a space sampled. Considering the nature of cluster sampling, this chapter uses the following essential terms and definitions (table 1.1). A spatial unit where sediments were individually collected should be treated as a “sample element” (or simply “element”) as an analytical unit, while quantities of plant remains become parameters of the sample element that contains them. Regardless of the number of bags or the volume of sediments or the number of seeds inside them, one feature or a specific space within a feature should be treated as one element. For example, I collected seven bags of sediments with a volume of 20 L each from Hearth 7 at the Middle Mumun Oun 1 site, but all these bags still represent one sample element of 140 L, not seven (table 1.1). A volume of sediment in each spatial unit is an attribute of each element. In most cases, an element is confusingly called a “sample,” a subset of a population. A sample can be a group of elements that are collected from the same feature, site, or period. The Early Mumun sample in this study consists of all the Early Mumun features at the Oun 1 site that were sampled individually. The sample size is the total number of elements in the sample. For instance, the Early Mumun sample size available for this study is 376, which represents the number of spatial units where many sediment bags (2,814 L in total) were collected.

12

methodologies in archaeobotany

Table 1.1. Terms in cluster sampling Terms

Definitions

Examples

(Sample) element

A unit of a sample. In this study, it is a spatial unit, i.e., a specific feature or a part of one feature where sediments are recovered separately from other spatial units.

Outdoor Hearth 2 of the Middle Mumun component; Floor unit F3 in House 104 of the Early Mumun component (see table 2)

Sample

A subset of population. In this study, it is a group of sample elements that come from the same feature type, site, or period.

Early Mumun sample

Sample size

Total number of sample elements in a sample

Size of the Early Mumun sample analyzed is 77 (see table 1.3).

Sampling fraction

The proportion of the sample to the whole population

Estimating a Sample Size Van der Veen and Fieller (1982) is one of the few statistical approaches to sampling in archaeobotany. The authors present formulas that depend on the required accuracy level for the difference between an estimate and a real population, and the chance of obtaining the required accuracy by means of z-score. Defining a population size as total counts of seeds in a specific context, they tabulate a required seed number that should be counted in each context to meet the accuracy level. These ready-made numbers, however, are valid only when the population size can be reasonably estimated. Although we may have some prior knowledge of the possible richness of plant remains from a certain feature as they suggested, this assumption cannot be automatically applied to unanalyzed samples from different sites. Plant assemblages in most contexts are simply tertiary in nature (Fuller, Stevens, and McClatchie, this volume) and thus estimating the size of plant remains based on the assumed functions of primary (or secondary) contexts is not meaningful. Another problem in the formulas of Van der Veen and Fieller is a misunderstanding of cluster sampling as random sampling. A collection of

Sample-Size Estimation

13

seeds from a random sample of spatial units (sampling element) is not exactly the same as a random sample of seeds from that site (Banning 2000, 81). In cluster sampling, the population that we want to analyze (i.e., seeds) is not the population that we actually sampled (i.e., space), but rather the parameters (observations) of the population. By confusing these two different sets, what their formulas estimate is actually the parameters of the samples (seed counts) rather than the actual sample size. The confusion causes analysts to overlook the effect of the spatial autocorrelation (Banning 2000, 81).

Suggested Formula Similar to what Van deer Veen and Fieller (1982) suggest, this study estimates the sample size required for analysis based on the statistical representativeness. Considering that archaeobotanical sampling is indeed cluster sampling, however, my sample-size estimation is based on the number of features (e.g., spatial units) that were individually selected for collecting sediments within, instead of the number of seeds within the feature. No matter how many seeds were found in each feature, the sample size (i.e., the number of elements) is not changed. Moreover, my formula has no need to assume the total number of seeds that may have existed in each context, which Van der Veen and Fieller’s (1982) formula requires. In order to gain comparable sizes of samples from different populations, particularly when the population is large relative to the sample, a relative error r of each sample should be controlled. An r expresses a standard error SE as a proportion of the mean μ (equation 1). In other words, we ask ourselves what percentage of the error we can tolerate. For example, do we want a standard error to be 10 percent of the mean of the population or less? The value of r is dependent on the accuracy level that we are willing to accept. The trade-off is that higher level of accuracy (a smaller r value) may require a large sample size, particularly when the parameters of samples have a rather large variation (standard deviation). r = SE/μ

(1)

The following formula can calculate the sample size where the μ and σ are the mean and standard deviation of the population. The t is a zscore, a number of the standard deviation associated with a particular confidence level. n = (σt)2/(rμ)2

(2)

Since we are unlikely to know the parameters (e.g., mean, standard deviation) of the population in advance, we should substitute them with

14

methodologies in archaeobotany

those known to us, which are the mean (x) and standard deviation (s) of the pilot sample. n = (st)2/(rx)2

(3)

This study used ten sample elements from each of the Early and Middle Mumun sites as pilot samples to determine the sample size (table 1.2). The size of pilot samples is determined arbitrarily here, and further study is needed to determine whether the size of pilot samples affects the estimation of sample sizes. Since each pilot element had volumes varying from 5 to 20 L, I used seed densities (number of seeds per 1 L of sediments) as parameters for estimation instead of seed counts (equation 4). dx = ∑xi/vi

(4)

Here n is the sample size (table 1.2); xi is the number of seeds in the element i; vi is the volume of the element i; V is total volumes in the sample; and dx is the seed density in the sample to which the element i belongs. As this study sets the confidence level at 90 percent, the z-score is 1.83 for a small sample of ten. This study decides to control a relative error, limiting it to 20 percent of the population mean (0.2 r value in equation 3). By doing so, the estimated numbers of both the Early and Middle Mumun samples will have the standard error as the same proportion of the mean (20 percent) at a 90 percent confidence level, and thus both samples can be statistically comparable. As a result, I calculated the required sample sizes of 141 elements for the Early Mumun, and 70 for the Middle Mumun for the accepted accuracy level. Again, this formula does not require a specific volume of sediments in each element, since it depends on the seed density as a parameter rather than actual counts of seeds, which may be affected by different sediment volumes of elements compared.

Application and Result Before analyzing all sample elements that my formula estimated, this study monitored whether the sizes of samples analyzed met the relative error that I set at the beginning. The relative error was computed by dividing the sample standard error (equation 5) by the sample mean instead of by the unknown population mean (equation 1). SE = √({∑[(xi/vi-dx)2x(vi.n/V)2]}/[n(n–1)])

(5)

I analyzed a total of seventy-six and sixty-three elements for the Early and Middle Mumun components, respectively (table 1.3). The relative

Table 1.2. Pilot sample elements from the Early and Middle Mumun Nam River components

Zone

Features sampled

Soil volume (L)

Seed no.

10 10 14 5 6 8 9 10 20 5 97

32 2 26 1 6 3 3 34 1 2 110

Seed density per L

Early Mumun Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1

Floor unit G1 in House 104 Floor unit F3 in House 104 Hearth in House 92 Pit in House 92 Vessel in House 65 Hearth in House 79 Floor unit C3 in House 118 Pit in House 107 Vessel in House 75 Vessel in House 84 Sum Mean Standard deviation Standard error Relative error

3.2 0.2 1.9 0.2 0.4 0.4 0.3 3.4 0.1 0.4 1.0 1.3 0.4 0.4

Middle Mumun Okbang1 Okbang1 Okbang1 Okbang1 Okbang4 Okbang4 Oun1 Oun1 Oun1 Oun1

Outdoor Hearth sa-19 Outdoor Hearth sa-28 Indoor working pit 657 Indoor working pit 660 Outdoor round pit 18 Indoor Hearth 53 Floor unit G4 in House 5 Outdoor Hearth 2 Outdoor Hearth 4 Outdoor Hearth 3 Sum Mean Standard deviation Standard error Relative error

10 10 10 10 5 10 10 10 20 20 115

6 29 8 13 5 1 0 32 26 7 127

0.6 2.9 0.8 1.3 1.0 0.1 0.0 3.2 1.3 0.4 1.2 1.1 0.4 0.3

Sample elements

Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1

H65 hearth 2 H65 hearth 3 H65 hearth 4 H65 hearth 5 H75 hearth H79 hearth 1 H79 hearth 2 H79 hearth H92 hearth H104 G10 hearth H104 D5 hearth H104 D7 hearth H118 hearth 1 H118 hearth 2 H65 hearth1 H92 floor A1 H92 floor B2 H92 floor B6

EARLY MUMUN

Sites

10.0 5.0 10.0 4.0 15.0 15.0 7.0 30.0 14.0 10.0 18.0 8.0 5.0 10.0 2.0 3.0 10.0 19.0

Soil vol. 10 1 49 0 6 6 1 1 26 2 22 3 63 186 1 0 79 19

Seed no. 1.0 0.2 4.9 0.0 0.4 0.4 0.1 0.0 1.9 0.2 1.2 0.4 12.6 18.6 0.5 0.0 7.9 1.0

Seed density Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1

Sites H104 floor G2 H104 floor G3 H104 floor G5 H104 floor G19 H104 floor H7 H104 floor A17 LP H104 floor E17 LP H104 floor G15 LP H104 floor E5 H104 floor G9 H104 floor H104 floor F7 H104 floor B3 H104 floor C19 H104 floor D13 H104 floor E9 H104 floor D6 H104 floor A14

Sample elements

Table 1.3. Seed densities of Early and Middle Mumun Nam River elements

10.0 10.0 10.0 9.0 10.0 10.0 10.0 10.0 4.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 9.0 8.0

Soil vol. 32 0 31 7 24 119 38 11 3 2 11 7 20 6 427 8 16 73

Seed no. 3.2 0.0 3.1 0.8 2.4 11.9 3.8 1.1 0.8 0.2 1.1 0.7 2.0 0.6 42.7* 0.8 1.8 9.1

Seed density

Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1

H92 floor C6 H92 floor E4 H92 floor F6 H92 floor E11 H92 floor B7 H92 floor C9 H92 floor D7 H104 floor G10 H104 floor D5 H104 floor E5 H104 floor D7 H104 floor B5 H104 floor C1 H104 floor C9 H104 floor D2 H104 floor E9 H104 floor E15 H104 floor F3 H104 floor F4 H104 floor G1

7.0 8.0 10.0 10.0 8.0 10.0 10.0 10.0 10.0 4.0 5.0 10.0 10.0 8.0 10.0 9.0 10.0 10.0 10.0 20.0

40 34 25 10 73 29 21 6 16 3 7 2 9 55 7 0 64 2 1 304

5.7 4.3 2.5 1.0 9.1 2.9 2.1 0.6 1.6 0.8 1.4 0.2 0.9 6.9 0.7 0.0 6.4 0.2 0.1 15.2

Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1

H104 floor C16 H118 floor C3 H118 floor D7 H78 round pit H92 N round pit2 H92 NE round pit H107 round pit H65 small vessel 82 H65 vessel T84 H75 vessel 1 H84 long-neck vessel H84 vessel 2 H84 vessel 4 H84 vessel 5 H84 vessel 7 H92 vessel 1 H92 vessel 3 H104 jar H107 vessel H107 H4 vessel

8.0 9.0 8.0 19.0 5.0 10.0 10.0 6.0 40.0 20.0 5.0 2.0 1.0 4.0 2.0 8.0 3.0 30.0 9.0 3.0

100 3 0 9 1 15 34 6 7 1 2 5 0 0 1 3 0 24 3 4 continued

12.5 0.3 0.0 0.5 0.2 1.5 3.4 1.0 0.2 0.1 0.4 2.5 0.0 0.0 0.5 0.4 0.0 0.8 0.3 1.3

Sample elements

Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1

hearth 1 hearth 3 hearth 4 hearth 2 west hearth 2 east hearth 7 hearth 8 hearth 22 hearth 23 hearth 26 hearth 28 H5 floor G1 H5 floor G2 H5 floor G3 H5 floor G4

MIDDLE MUMUN

Sites

Table 1.3. Continued

10.0 10.0 10.0 40.0 20.0 140.0 129.0 10.0 9.0 10.0 20 10.0 10.0 10.0 10.0

Soil vol. 6 1 0 579 9 37 89 2 1 2 12 3 0 1 0

Seed no. 0.6 0.1 0.0 14.5* 0.5 0.3 0.7 0.2 0.1 0.2 0.6 0.3 0.0 0.1 0.0

Seed density Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Okbang1 Okbang1 Okbang1

Sites H6 floor G7 H6 floor G8 H13 floor G6 H13 floor G12 H13 floor G16 H13 floor G24 H13 floor G26 H13 floor G33/42 H13 floor G35 round pit 4 round pit 19 round pit 28 hearth Sa-19 hearth Sa-28 hearth,Sa-11

Sample elements 10.0 10.0 2.0 5.0 3.0 1.0 2.0 4.0 2.0 10.0 4.0 4.0 10.0 10.0 30.0

Soil vol. 0 1 0 0 0 0 0 1 2 2 11 11 6 28 14

Seed no. 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.3 1.0 0.2 2.8 2.8 0.6 2.8 0.5

Seed density

H5 floor G5 H5 floor G6 H5 floor G7 H5 floor G8 H5 floor G9 H5 floor G10 H5 floor G11 H5 floor G12 H5 floor G13 H5 floor G14 H5 floor G15 H6 floor G1 H6 floor G2 H6 floor G3 H6 floor G4 H6 floor G5 H6 floor G6

10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0

0 4 0 0 0 2 3 1 1 2 0 0 0 0 1 1 1

0.0 0.4 0.0 0.0 0.0 0.2 0.3 0.1 0.1 0.2 0.0 0.0 0.0 0.0 0.1 0.1 0.1

Okbang1 Okbang1 Okbang1 Okbang1 Okbang1 Okbang1 Okbang1 Okbang1 Okbang4 Okbang4 Okbang4 Okbang4 Okbang4 Okbang4 Okbang4 Okbang4

Note: The letter H indicates a house. The asterisk (*) indicates an outlier.

Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1 Oun1

hearth 12-2 hearth Sa-26 hearth Sa-29 H657 working pit H658 working pit H660 working pit round pit 72 round pit 310 H53 hearth hearth 40 hearth 41 hearth 34 hearth 24 hearth 27 H44 working pit round pit 18

10.0 20.0 5.0 10.0 13.0 15.0 2.0 1.0 10.0 9.0 6.0 10.0 10.0 10.0 9.0 8.0

0 7 1 4 63 33 4 0 1 2 15 1 1 1 0 7

0.0 0.4 0.2 0.4 4.8 2.2 2.0 0.0 0.1 0.2 2.5 0.1 0.1 0.1 0.0 0.9

20

methodologies in archaeobotany

Table 1.4. Summary of the Early and Middle Mumun Nam River samples Early Mumun With outlier No. of elements

Without outlier

Middle Mumun With outlier

Without outlier

76

75

63

62

756

746

873

833

2,236

1,809

974

395

Mean seed densities

3.0

2.4

0.7

0.5

Standard deviation

6.0

3.8

2.0

0.9

Standard error

0.7

0.4

0.3

0.1

Relative error

0.2

0.2

0.4

0.2

Vol. of sediment L No. of seeds

error of the Early Mumun sample is 0.2 (table 1.4) with or without one outlier element from the Oun 1 site (Floor unit D 13 of House 104 in table 1.3). The mean and standard error of the Middle Mumun sample are 0.7 and 0.3, producing a relative error of 0.4 (table 1.4). If one outlier from Hearth 2 west in Oun 1 is excluded, however, the relative error is 0.2 and meets the accuracy level that this study regards as being acceptable. As a result I was able to complete the analysis with fewer elements than are estimated.

A Model of Ratio of Ratios Now that the samples can be comparable, the next step is to implement the quantitative method that is less sensitive to retrieval rates and preservation biases. A density is the most frequently used ratio in archaeobotany, a measure choosing a volume of sediments as a numerator against which another variable (e.g., seed counts, charcoal weight) can be measured (Miller 1988, 74). The basic assumption of density ratios is that larger sediment samples have more plant remains, all things being equal. However, all things are not equal; that is, the uniform deposition, preservation, and recovery rates cannot be taken for granted. A density ratio, like any other measure, is still vulnerable to preservation bias inherent in different types of contexts (Kadane 1988). One way to relieve this problem is to compare samples from contexts that likely have similar preservation conditions. Another solution is the interassemblage ratios, similar to what Orton (2000, 65) suggests for overcoming different preservation

Sample-Size Estimation

21

biases embedded within potsherd counts. He suggests an intuitive model regarding the effects of differential survival and retrieval rates of different types of vessels between the two samples. Differential survival and retrieval rates can widen the gap between the sampled and target assemblages. The sampled assemblage is what is retrieved at the present time, while the target assemblage is defined here as the deposited assemblage, that is, plants that were deposited into the soil matrix when the site was in use or shortly after it ceased its function (Lee 2012, 652). By maintaining the same procedures of flotation and sieve sizes for floated materials, the retrieval intensity was maintained to a similar degree in this study. Consequently, differential survival rates are more likely to be a problem rather than differential retrieval rates. Survival probabilities are related to taphonomic processes, which affect the patterning of archaeological distributions. The solution is to determine which measures are most likely to survive postdepositional processes without the need for unjustifiable assumptions about the sampling fraction, the proportion of the sample to the whole population (Orton 2000, 56). The schematic illustration in figure 1.2 shows how differential survival/ retrieval rates affect the comparison in proportions of different plant species between samples. For the sake of argument, millet and rice are assumed to have different survival probabilities (row D in figure 1.2). In this case, their retrieved raw numbers (the retrieved assemblage in row E) or retrieved ratios between them (F) cannot reflect which crop taxa were more abundant in the past (the original assemblage in row A). In house Y, millet grains were retrieved twice more than rice grains (E), but their original numbers (A) are the same. Another problem in such a direct comparison of retrieved raw numbers lies in the possible differential preservation rates between features. In this example, house Y originally had more millet than house X (A), but the retrieved result (E) is quite the opposite because house X has a better preservation condition than House Y (D). Although the compositions of samples do not reflect those of the original (target) population at the moment of deposition, the ratio of ratios between taxa across features remains the same through time. In other words, the comparison between the two samples from houses X and Y does reflect the comparison between the two original populations. The ratios between millet and rice at the time of deposition (B) are 2:1 in house X and 1:1 in house Y. Accordingly, the ratio of ratios from millet to rice between the two houses (C) is 2:1. In the sample, the ratios between millet and rice (F) are 4:1 in house X and 2:1 in house Y. The ratio of ratios in the retrieved samples (4:2 in row G) is the same as that of the target populations (2:1 in row C). Statistically, the samples and

22

methodologies in archaeobotany At deposition

A) original counts

millet

rice

50

25

millet

rice

100

100

2:1

B) Ratios between taxa

1:1

2:1

C) Ratios of ratios

House X

House Y

At recovery D) Preservation rates of taxa

0.8

0.4

0.2

0.1

E) Retrieved counts

40

10

20

10

F) Ratios between taxa G) Ratios of ratios

4:1

2:1

4:2

Figure 1.2.  Model of interassemblage ratios. (Redrawn from Orton 2000, 65.)

the taxa are independent factors, suggesting that the inter-assemblage comparison remains valid regardless of site-formation processes (Orton 2000, 66). It can be inferred that the abundance of millet relative to rice in house X is greater than in house Y, as the ratio of ratios (G) indicates. In this model, one inherent assumption is that the ratios of survival probabilities between taxa are invariant, that is, no interaction between types and assemblages. Although the survival probabilities (D) of millet and rice are different in the two houses (0.8 and 0.4 in house X, 0.2 and 0.1 in house Y), the ratios of probabilities between taxa remain the same in the two houses (0.8/0.4 = 0.2/0.1 = 2). Can this be justified? In order to answer this question, factors affecting survival rates should be considered in light of taphonomic transformation. Since the target population in this study is the deposited assemblage, the built-in assumption of invariant comparison is related to the factors that transformed the deposited into the sampled assemblages. This transformation can be further articulated into two transitions: from the moment of deposition to the present, where the remains survived through

Sample-Size Estimation

23

time, resulting in the fossil assemblage; and from what remained at the time of excavation to what was actually retrieved (Lee 2012, 652). The factors affecting these transitions are not in the realm of cultural transformation, but in the realm of natural and mechanical interferences as well as statistical inference. In the transition from the deposited to the fossil assemblages, differential rates of natural decay between taxa may cause a difference in survival rates. This bias, however, can be ruled out because uncarbonized remains are not considered ancient in origin in my study area owing to highly acidic soil conditions there. Mechanical interference may have affected the formation of the fossil assemblage. The so-called microartifacts, defined as being between 0.25 mm and 2 mm in size, are less likely to be moved horizontally by site-formation processes (e.g., plowing, trampling) than larger artifacts (Dunnell and Stein 1989, 38). Most plant remains are within this size range, so mechanical interference may not have substantially affected them, although plant remains could still be fragmented. The Nam River sites were buried 2 to 3 m below the ground by fluvial accumulations, thereby minimizing mechanical interference after seeds were deposited. In brief, the natural, mechanical interferences have probably not influenced seed taxa differently under a thick blanket of fluvial layer, and thus there is no interaction between taxa and assemblages. Consequently, archaeological patterning can be inferred from the recovered plant remains in my study area. The sampled population does not represent the totality of plant remains that were brought into a site but its subset that survived through time, namely, fossil assemblage (Lee 2012). Strictly speaking, therefore, what the ratio of ratios can indicate is which taxa may have deposited and survived more frequently, rather than which were more important in diet or in other use. Since remnants from repeated actions in the past probably have a better chance to be left in archaeological contexts, it is still possible to infer some patterning of plant use from the comparisons of taxa compositions in the retrieved assemblage. For example, experimental study of burning shows that the quantitative distribution of different carbonized cereal types found in houses and hearths reflects the actual precarbonization proportions (Gustafsson 2000, 70).

Discussion and Conclusion Equal representativeness to population between different samples is an essential requirement for making any cultural inferences derived from the samples. This allows for inducing quantitative differences (or nondifferences) between two populations from differences (or nondifferences)

24

methodologies in archaeobotany

between their samples. If not, differences in quantity may come merely from vagaries of different sampling fractions among samples compared rather than real cultural reasons that produced differences in quantities. The formula introduced in this chapter determines the sample size based on a fixed relative error and a confidence level of seed densities in pilot samples. The strength of this formula will not be affected by the difference in a sediment volume of each sample element, since the formula uses the seed densities, not the seed counts per se. Another benefit of this measure is that it can be applied retroactively as long as a soil volume of the sampling unit (e.g., feature) was recorded. Since these estimates of sample sizes are based on the parameter of total seed densities, the comparison of each taxon across samples may not yield the same representativeness as the seed densities of all taxa combined. For further comparison of taxon compositions, this paper suggested the interassemblage comparisons that can reduce biases of different preservation rates and contexts. The ratio of ratios between taxa across different samples, which is modified from vessel counts in Orton (2000, 52), can reduce biases of preservation rates, and thus can be useful in discerning relative abundance of plant taxa in deposited assemblages among different cultural entities.

Acknowledgments I am grateful to the editors of this volume and the anonymous reviewers. The Kyungnam University Museum supported my Nam River Archaeobotanical Project. I am also grateful to many individuals for their insightful feedback on earlier drafts of the chapter, including Ted Banning, Gary Crawford, Rory Walsh, Daphne Gallagher, Molly Casperson, Reecie Levin, Sandra Poaps, and Della Saunders.

References Banning, E. B. 2000. The Archaeologist’s Laboratory: The Analysis of Archaeological Data. Interdisciplinary Contributions to Archaeology. New York: Kluwer Academic/Plenum Publishers. Crawford, G. W., and G-A. Lee. 2003. “Agricultural Origins in Korea.” Antiquity 771:87–95. Dunnell, R. C., and J. K. Stein. 1989. “Theoretical Issues in the Interpretation of Microartifacts.” Geoarchaeology 4:31–42. Gustafsson, S. 2000. “Carbonized Cereal Grains and Weed Seeds in Prehistoric Houses: An Experimental Perspective.” Journal of Archaeological Science 27:65–70.

Sample-Size Estimation

25

Kadane, J. B. 1988. “Possible Statistical Contributions to Paleoethnobotany.” In Current Paleoethnobotany: Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, 206– 14. Chicago: University of Chicago Press. Lee, G-A. 2003. “Changes in Subsistence Systems in Southern Korea from the Chulmun to Mumun Periods: Archaeobotanical Investigation.” PhD diss., University of Toronto. ———. 2012. “Taphonomy and Sample Size Estimation in Paleoethnobotany.” Journal of Archaeological Science 39:648–55. Miller, N. F. 1988. “Ratios in Paleoethobotanical Analysis.” In Current Paleoethnobotany; Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, 72–85. Chicago: University of Chicago Press. Orton, C. 2000. Sampling in Archaeology. Cambridge: Cambridge University Press. Van der Veen, M., and N. R. J. Fieller. 1982. “Sampling Seeds.” Journal of Archaeological Science 9:287–98.

C h ap t e r 2

Regional Exchanges in Southeastern Arabia during the Late Pre-Islamic Period Phytolith Analysis of Ceramic Thin Sections from ed-Dur (UAE) Luc Vrydaghs, Paul De Paepe, K at r i e n  R u t t e n ,   a n d E r n i e H a e r i n c k

The archaeology of the Arabian Gulf documents exchanges of goods going back to prehistory. Pottery of the Ubeid period (5500–4000 BC) has been found at several places along the coast as far as the northern emirates. Later, during the Umm an-Nar and Wadi Suq periods, contacts with Dilmun (Bahrain), southwestern and southeastern Iran, and the Indus valley are attested. Contacts and exchanges continued during the Iron Age and the Achaemenid period. The best site illustrating this period is Mleiha. From the third to the middle of the first century BC, there was no harbor along the coast between Qatar and the Strait of Hormuz and goods could only reach Mleiha by means of camel caravans. In the second half of the first century BC, this picture changed with the emergence of the coastal site of ed-Dur in the emirate of Umm al-Qaiwain, United Arab Emirates (UAE). Located 120 km from the Strait of Hormuz and some 50 km to the north of the inland site of Mleiha (figure 2.1), ed-Dur was first occupied during the fifth/fourth millennium BC, then during the Umm an-Nar

26

Regional Exchanges in Southeastern Arabia

27

Figure 2.1. General map of the emirate of Umm al-Qaiwain with the location of ed-Dur. The dark gray area indicates the Iron Age and scatter occupations of the first century AD; and the light gray area indicates the main occupation of the first century AD.

period (second half of the third millennium BC) and the Iron Age, after which it was deserted. In the second half of the first century BC, the site was reoccupied, and this occupation lasted until the early second century AD. During this period, ed-Dur was the largest and most likely the only substantial coastal site on the southern side of the Arabian Gulf between Qatar and the Strait of Hormuz. From 1986 to 1994 a team from Ghent University (Belgium) carried out excavations at ed-Dur. The excavation of the central area of the site (figure 2.1) revealed a number of architectural remains (e.g., temple, funerary, and domestic architecture) and numerous artifacts (coins, glass vessels, beads, jewels, and bronze objects), including an extensive pottery assemblage of more than 13,000 diagnostic sherds dating from the last decades of the first century BC and the second century AD. Some of these items suggest exchanges with the Roman Empire, the eastern Mediterranean, Characene of southern Mesopotamia, southern Arabia, Iran, India, Sri Lanka, and Tanzania (De Paepe et al. 2003; Haerinck 1998, 2003). Signs of contacts with the site of Mleiha and other places in UAE and Oman are also evident from the archaeological record (De Paepe et al. 2003; Haerinck 2003).

28

methodologies in archaeobotany

Evidence of Regional Contacts from the Charcoal Analysis Ed-Dur is located in a mangrove lagoon along the coast. Mangrove vegetation is one of the major biomes of the region and consists of monospecific stands of Avicennia marina (Forssk.) Vierh. The floristic diversity of the Oman peninsula comprises more than 500 taxa, many of which show adaptation to high salinity and temperatures, and to low rainfall. However, owing to overexploitation (cutting and grazing), urbanization, and industrialization, this flora is a relict of a more complex biodiversity. Three physiographic units influence the vegetation distribution: the coastal plain, the “sand sea” or desert, and the mountains. The coastal plain comprises flat expanses of sand, gravel, and lava plains regularly incised by gullies (wadis). The vegetation of these areas (sometimes referred to as pseudosavannah) is scattered and includes, among other taxa, several species of Acacia L. and Capparis L., Balanites aegyptiaca (L.) Del., and Calotropis procera (Aiton) Aton f. (table 2.1) (Tengberg 2002). The “sand sea,” or Rub’ al-Khali, is characterized by sabkhas, gravel plains, and sand desert. Geologically recent, the Rub’ al-Khali results from subaerial erosion and sand deposition in an arid environment. Vegetation is sparse, with Calligonum crinitum Boiss on dune slopes, frequently associated with Zygophyllum mandavillei Hadidi and Cyperus conglomeratus Vahl, while trees are found along the outer margins of the desert (Acacia ehrenbergiana Hayne and Prosopis cineraria [L.] Druce). A number of artesian springs form oases where the vegetation includes natural (e.g., Phragmites sp. and Prosopis juliflora DC) and cultivated (e.g., date palm trees—Phoenix dactylifera L.) plants (table 2.1). After rainfall, grass may appear. Sabkhas are a typical feature of this landscape. They are formed by the evaporation of shallow waters leading to the formation of a salt crust. Three types are recognized: coastal sabkhas or sabkhas s.s., inland sabkhas, and fluvio-lacustrine sabkhas. Sabkhas s.s. are vast, formed near the marine shoreline, and very common. Inland sabkhas mark sea transgressions (Rice 1994, 76), while the fluvio-lacustrine sabkhas are associated with old drainage systems. Sporobolus spicatus Kunth and Prosopis cineraria (L.) Druce are the common plants of these areas. The Oman mountains, or Al Hajar, reach an altitude of more than 2,000 m and stretch from the Strait of Hormuz in the northwest to the Arabian Sea in the southeast. They represent the margin of the Arabian continental platform known as the Semail ophiolite complex, which is geologically distinct from the rest of the Arabian Peninsula. Plains of fluvial sediments flank the mountains, while dunes and sabkhas overlie the northern edge. Depending on the altitude, the vegetation includes Acacia gerardii Chaudhary, Monotheca buxifolia (Falc.) Dene. Ex Engler, or Juniperus excelsa M. Bieb (table 2.1).

Table 2.1. Modern vegetation taxa of the UAE classified according to their formation Charcoal Modern

Vegetation

Coast

Mangrove Plain

ed-Dur Mleiha

Aviciennia marina (Forssk.) Vierh A. tortilis Hayne A. raddiana Savi A. gerrardii Chaudhary A. ehrenbergiana Hayne Balanites aegyptiaca (L.) Del. Calotropis procera (Aiton) Aton f. Capparis decidua Pax. C. cartilaginea Decne Cordia gharaf Ehrenb. ex. Aschers Moringa peregrina C. Christensen Salvadora persica L. Stipograstis Nees. Inland Desert Xero­morphic Calligonum crinitum Boiss Open Cornucala arabica Botsch. shrubland Cyperus conglomeratus Vahl. Acacia tortilis Hayne A. ehreubergiana Hayne Oases Phragmites Adans. Prosopis juliflora DC Tamarix aphilla (L.) Karsten Typha L. Ziziphus spinachristi (L) Willd. Sabkhas Prosopis cineraria (L) Druce annual grasses Mountain Lower Wadis Acacia tortilis Hayne mountains Prosopis cineraria (L) Druce Ziziphus spinachristi (L) Willd. Ficus L Open 1,000– Acacia tortilis Hayne woodlands 1,500 m A. gerardii Chaudhary Steep slopes Euphorbia larica Boiss Periploca aphylla Deche 1,100– Dodonaea angustifolia L. f. 2,500 m Monotheca buxifolia (Falc.) A. DC. Olea europaea (L) Pseudosavannah

2,100– 3,000 m

+ +

+

+

+

+

+

+

+ +

+ + + + +

Juniperus excelsa M. Bieb.

Sources: www.nationalgeographic.com; www.worldwildlife.org. Note: The last two columns on the right detail the charcoals identified by Tengberg (2002) for ed-Dur and Mleiha.

30

methodologies in archaeobotany

Papadakis (1966) classifies the dominant Arabian Peninsula climate as hot subtropical desert with many frostless coasts. This climate is characterized by high summer temperatures, cold winters, and low annual rainfall. The precipitations occur between October and April, with their peak between mid-December and mid-February, and are a consequence of the Indian Ocean monsoons. However, cultivation is impossible without irrigation. The onset of the present-day conditions seems to date from 3000 BP (Sanlaville 1992). Charcoal analysis for the Oman peninsula highlights that, at the turn of our era, the mangrove stands of Avicennia marina (Forssk.) Vierh included at least one other plant taxon, Rhizophora sp. or Bruguiera sp. (Tengberg 2002). It also provided evidence for several wild species like Acacia sp., Chenopodiaceae, Calotropis procera (Aiton) Aton f., Prosopis cineraria (L.) Druce, Tamarix spp., and Ziziphus spina-christi (L.) Willd., along with some cultivated taxa like Phoenix datylifera L. and Dalbergia sissoo Roxb. (table 2.1). Charcoal from ed-Dur shows the presence of mangrove stands of Avicennia marina (Forssk.) Vierh but also of inland vegetation with Calotropis procera (Aiton) Aton f., Acacia sp, and Prosopis cineraria (L.) Druce. (Tengberg 2002) (table 2.1). These last taxa, together with date palm charcoals (Tengberg 2002) and phytoliths (Vrydaghs et al. 2001), might indicate nonlocal input of plants. These findings demonstrate that some kind of local or regional exchange was in place. In this respect, it is also worth mentioning the charcoal evidence for Rhizophora from Mleiha (Tengberg 2002, 155).

The Material Culture Numerous aspects of the material culture parallel the charcoal evidence for local and regional exchanges. Local exchange is documented by the Abi’el coins, known only from ed-Dur and Mleiha. Apart from a find of unknown origin in Yemen, none of the Abi’el coins have ever been found in northeastern Arabia or elsewhere. This suggests that this currency circulated only between ed-Dur and Mleiha (Haerinck 1998). A visual analysis of the pottery enables the identification of thirtyseven distinct types of wares distributed into two groups, referred to as Nonlocal Production and Local Production. Nine types of wares, representing 27 percent (percentages in the text are rounded to the nearest decimal) of the complete pottery assemblage, could have been manufactured in southeastern Arabia (UAE and Oman). Within this Local Production, two wares (Orange Vegetal Tempered Ware and Black Ware) are commonly found in ed-Dur and form the bulk of the local assemblage with 82 percent. The remaining seven ware types (Fine Orange Vegetal

Regional Exchanges in Southeastern Arabia

31

Painted, Fine Orange Vegetal Tempered, Coarse Buff Orange, Coarse Yellow-Pink Speckled, Coarse Pinkish Brown, Coarse Ochre, and Coarse Buff Slipped Ware) represent only 18 percent of the local assemblage (table 2.2). The local pottery shows mainly coarsely tempered fabrics. Evidence of pottery manufacture at ed-Dur is lacking. However, differences in finishing techniques, shapes, and decorations suggest the existence of several production centers; therefore, the Local Production group indicates regional contacts. The available literature shows that most of these wares are known only at contemporary sites in southeastern Arabia (Boucharlat and Mouton 1993; De Cardi, Kennet, and Stocks 1994; Højlund and Andersen 1994; Kennet 1998; Kervan and Hiebert 1991; Lecomte 1993; Lombart and Kervan 1993; Lombart and Salles 1984; Mouton 1992; Salles 1984), while others (e.g., the Black Ware) have also been found at coastal sites in the southern UAE, Yemen, and possibly along the Red Sea (De Paepe et al. 2003; Rutten 2007). To further investigate the provenance of this Local Production, studies on the petrography, chemical composition, and phytolith contents have been carried out (De Paepe et al. 2003). Thin-section ceramic petrography enables the identification of a wide variety of minerals, rock fragments, and other nonplastic inclusions of sand and gravel size dispersed throughout the fabric. Furthermore, it can provide information on the texture, shape, degree of sorting, and frequency of the tempering materials. Therefore, the data supplied by Table 2.2. Local wares from ed-Dur Ware

%

Fine Orange Vegetal Painted Ware

0.38

Fine Orange Vegetal Tempered Ware

9.24

Coarse Buff and Orange Ware

7.55

Coarse Yellow-Pink Speckled Ware

0.21

Coarse Pinkish Brown Ware

0.21

Coarse Ochre Ware

0.33

Black Ware Coarse Buff Slipped Ware Orange Vegetal Tempered Ware

60.54 0.12 21.38

Source: K. Rutten, unpublished data. Note: The column on the right gives the percentage of the local pottery. The local pottery represents ca. 27 percent of the complete pottery assemblage.

32

methodologies in archaeobotany

thin-section analysis allow the characterization of the pottery in order to identify fabric groups, to shed light on pottery-making techniques, and to establish potential areas of origin. In general, the examination in thin section of pottery of very fine fabrics or containing only common minerals and rock fragments is of very little use for provenance determination. Chemical composition can provide clues to “fingerprint” and to group pottery. To a great extent, the successful use of elemental analysis for provenance identification rests on reference materials (e.g., kiln wasters). In the absence of kiln finds among the ed-Dur samples, the origin of the ware types can only be speculated on on the basis of literature data. Therefore, petrographic and chemical analyses alone may not be sufficient to identify pottery origin. The use of phytolith abundance and other variables from the paste of pottery from the Palenque region (Mexico) has shown that phytolith analysis can provide valuable information to identify “local production and use versus exchanges at regional scale” (Rands 1987). The present research illustrates how phytolith analysis can contribute to the understanding of pottery provenance.

Material Selection, Laboratory Procedures, and Methods As a first stage, 148 sherds were selected for petrographical and chemical analysis. Chemical analysis was carried out on powdered whole-pottery samples using the atomic absorption spectrometry method (AAS). The results from the petrographic and chemical analysis provided the basis for the selection of eighty-one representative thin sections for phytolith analysis, of which twenty-one are relevant for the Local Production group (table 2.3). A small piece from each representative sherd was impregnated with the synthetic resin TRA (Humphries 1992; Murphy 1986). Thin sections with a standard thickness of ca. 30 µm were then prepared and examined under a polarizing microscope. The results of the phytolith analysis of ceramic thin sections reported in this chapter rely on three indexes: the Observed/Not Observed (O/ NO), the Morphotype Identification (MI), and the Distribution index (D) (more details on these indexes can be found in the appendix, section 1). The distribution index tells whether phytoliths are dispersed in the fired matrix or in the pores of the pasta. Phytoliths in the fired matrix are embedded in the fabric and wedging effects might be noted (for definition, see Stoops 2003). It is assumed that these phytoliths are a natural inclusion of the clay and can therefore contribute to the identification of its provenance area. Pores within the fired fabric may result from the oxidation of the plant material used as temper, possibly leading to the deposition of some phytoliths in the pores. While the plant temper may

Table 2.3a. Results of phytolith analysis of the LP sherds thin section Group

Sample

FRVTW

AR 5608

FOVTW

AR 5380

CBOW

AR 5387

DI: fm

+

AR 5554

Morphotype Identification

Sp Sp; Lc

AR 5559 AR 5561 CYPSW

AR 5557

CPBW

AR 5562

COW

AR 5560

BW

AR 5588 AR 5591

+

Sp

+

Cr

AR 5592 AR 5594 CBSW

AR 5694

OVTW

AR 5375

Rt; Lc; Bu? +

Sp

+

Hb

AR 5378 AR 5383 AR 5393 AR 5555 AR 5556 AR 5558

Note: FRVTW: Fine Red Vegetal Tempered Ware; FOVT: Fine Orange Vegetal Tempered Ware; CBOW: Coarse Buff Orange Ware; CYPSW: Coarse Yellow-Pink Speckled Ware; CPBW: Coarse Pinkish Brown Ware; COW: Coarse Ochre Ware; BW: Black Ware; CBSW: Coarse Buff Slipped Ware; OVTW: Orange Vegetal Tempered Ware. AR refers to the register of the Laboratory of Mineralogy, Petrology and Micropedology of Ghent University. DI = phytolith distribution index; p: pores; fm: fired matrix. Morphotype identification = Rt: rounded trapezoid; Sa: saddle; Tr: trapezoid; Cr: crenate; Sp: spherical spinulose; Lc: long cell; Bu: bulliform; Um: unknown morphotype.

34

methodologies in archaeobotany

Table 2.3b. Results of phytolith analysis of the LP sherds thin section Group

Sample

FRVTW

AR 5608

FOVTW

AR 5380

CBOW

AR 5387 AR 5554

DI: p +

Morphotype Identification Rt; Sp?; Lc? Sp

+

Sp; Lc

AR 5561

+

Tr; Lc

CYPSW

AR 5557

+

Rt; Tr; Lc; Um

CPBW

AR 5562

COW

AR 5560

+

Rt

BW

AR 5588

+

Rt; Lc; Bu?

AR 5378

+

Rt; Sa; Hb; Lc

AR 5383

+

Rt; Tr; Lc; Sp

AR 5559

AR 5591 AR 5592 AR 5594 CBSW

AR 5694

OVTW

AR 5375

AR 5393 AR 5555 AR 5556 AR 5558 Note: For abbreviations, see note to table 2.3a.

not always have an origin close to the manufacture area, this phytolith deposition is connected to the pottery manufacture and may be useful to document specific technological aspects. Phytoliths were classified according to a revised version of Runge (1999) and Madella, Alexandre, and Ball (2005). For comparative purposes, the Dhofar reference collection (Ball 2002) was used in combination with a general atlas of phytoliths from leaf and wood tissues (Vrydaghs 2003). The identification of the source plant relies on three criteria: morphology, morphotype assemblage, and circumstantial evidence (Vrydaghs 2003).

Regional Exchanges in Southeastern Arabia

35

All microscopic observations were conducted with a polarizing microscope at magnifications ranging between ×125 and ×1000.

Results Phytoliths were observed in the Orange Vegetal Tempered Ware, Black Ware, Fine Orange Vegetal Tempered Ware, Coarse Buff Orange Ware, Coarse Yellow-Pink Speckled Ware, Coarse Ochre Ware, and Coarse Buff Slipped Ware (table 2.3).

Orange Vegetal Tempered Ware (AR 5375, AR 5378, AR 5383, AR 5393, AR 5555, AR 5556, AR 5558) This ware is pinkish orange or, less frequently, red to brown in color. The majority of the shapes consist of large plate bowls. Closed vessels include small, short-necked jars and a variety of large broad-bellied storage jars with a flattened base. This type is quite common at ed-Dur and accounts for ca. 6 percent of the total pottery assemblage. In thin section, these sherds appear to be made with a mixture of charred vegetal matter, mineral grains (mainly quartz and feldspars), rock fragments (igneous and metamorphic), and skeletons of microfossils (tests of foraminifera and thin shell fragments) set in a fired-clay matrix with variable carbonate content and huge pores. The more likely provenance of the included rock fragments is the ophiolite complex of the southern Oman mountains, a very distinctive geological region. With the exception of AR 5383, globular echinate phytoliths typical of palms were noted in the fired matrix (table 2.3a). Since the date palm is a frequent component of the wadi vegetation, these occurrences suggest that the wadis were a possible source area for the clay. Phytoliths observed in the pores are mainly rounded trapeziforms (figure 2.2), but some saddle, hair base, and elongate phytoliths were also noted. The Dhofar phytolith reference collection (Ball 2002) comprises 500 species, and rounded trapeziforms are reported only for two Chloridoideae taxa, Sporobolus sp. and Urochondra setulosa (Trin.) Hubb. The reference collection also shows that bi- and trilobated trapeziforms are produced in Sporobolus sp. but not in Urochondra setulosa (Trin.) Hubb, and that a particular hair morphotype seems to be characteristic of Sporobolus sp., while a sclereid-like morphotype is typical of Urochondra setulosa (Trin.) Hubb (figure 2.3). Rounded trapeziforms, trapeziforms, and elongated cell appendages but not sclereid-like morphotypes were observed in the pores of the Orange Vegetal Tempered paste. This specific assemblage, repeatedly observed within the pores, suggests Sporobolus

36

methodologies in archaeobotany

Figure 2.2. Rounded trapezoids phytoliths observed in an Orange Vegetal Tempered Ware sherd.

sp. as the source of these phytoliths. Irrigation canals related to the oases economy were established in the Oman mountains during the Iron Age (Boucharlat 2003). Their abandonment led to the formation of fluvio-­ lacustrine sabkhas and thus conditions favoring the development of stands of Sporobolus sp. after precipitations. All the pottery analyzed for phytolith dates between the last century BC and the second century AD. Since the petrographical study suggests the Semail ophiolite complex, the identification of Sporobolus sp. does not contradict the evidence from the charcoal analysis relevant for the same period. As to the saddle type, the Dhofar reference collection reports them for only two taxa, Chloris sp. and Dactyloctenium sp., suggesting that, in some cases, the plant temper could have mixed at least two grass taxa. The phytolith analysis of the Orange Vegetal Tempered Ware indicates the wadis as the source area for the clay and the fluvio-lacustrine shabkhas at least for part of the vegetal temper. The combination of these data with pottery findings for submodern and traditional pot making in the mountains and wadi runoffs of the UAE (for instance, in Wadi Hakil near Ras al-Khaimah and the more southern area of Siji and Masafi; Jongbloed 2003; Stocks 1992) suggests that pottery manufacture might have taken place in the vicinity of the raw material sources between October and April, according to the precipitations regimes.

Figure 2.3.  Top, rounded trapezoid from Sporobolus sp.; bottom, sclereidlike phytolith from Urochondra setulosa. Dhofar phytolith reference collection compiled by T. Ball (2002). (Photos: T. Ball.)

38

methodologies in archaeobotany

Coarse Yellow-Pink Speckled Ware (AR 5557), Coarse Ochre Ware (AR 5560), and Coarse Pinkish Brown Ware (AR 5562) Petrographic and chemical analyses of these wares show that in this case too the clay originated from the Semail ophiolite complex (see appendix, part 2). The occurrence of palm phytoliths (globular echinates) within the fired matrix of the Coarse Ochre Ware (table 2.3a) supports the hypothesis that the clay originates from contexts similar to that of the Orange Vegetal Tempered Ware. On the other hand, the absence of palm phytoliths in the fired matrix of the Coarse Yellow-Pink Speckled (CYPS) and the Coarse Pinkish Brown (CPB) wares suggests that they derive from different areas. The phytolith assemblage noted in the pores of the CYPS and CPB wares is dominated by rounded trapeziforms (table 2.3b), and therefore they are also considered to originate from Sporobolus sp. Table 2.4 summarizes all the phytolith observations conducted for the pottery types with raw material from the Semail ophiolite complex. The phytolith analysis distinguishes at least two provenance areas for the clay. As with the Orange Vegetal Tempered Ware, it may be considered that the manufacture area of the Coarse Ochre Ware was situated near the clay and grass temper source. On the other hand, the absence of any phytolith within the fired matrix and the pores of the CPB ware seems to suggest a second manufacture area also located within the Semail ophiolite complex.

Black Ware (AR 5588, AR 5591, AR 5592, AR 5594) The color of the majority of the Black Ware sherds varies from black to grey. The shapes comprise a small number of plates and very large bowls and a broad variety of wide and narrow-necked jars as well as storage jars with horizontally ridged walls. The proportion of this ware within the edDur assemblage is about 16 percent, making it one of the most common wares. The Black Ware is heavily tempered with charred organic matter and sand-to-gravel-sized rock fragments (shale, mudstone, siltstone, and argillaceous stones). Mineral grains are not frequent and noted only in the paste. These nonplastic inclusions are clearly sedimentary in origin. For the ed-Dur location, this points to the alluvial fan around the northern edges of the Oman mountains as a possible provenance area. This area is characterized by a complex system of wadis offering appropriate conditions for the date palm tree. Unfortunately, a wedging effect (Stoops 2003) prevents us from clearly ascertaining the occurrence of palm phytoliths within the fired matrix.

Regional Exchanges in Southeastern Arabia

39

Table 2.4. Phytoliths record for the wares from the Semail ophiolite complex Palm

Sporobolus

ORVTW

X

X

COW

X

X

CYPSW

X

CPBW Note: The varying occurrence of palm phytoliths suggests the clay originates from at least two areas, while the phytoliths deriving from Sporobolus suggest two manufacture areas. OVTW: Orange Vegetal Tempered Ware; COW: Coarse Ochre Ware; CYPSW: Coarse Yellow-Pink Speckled Ware; CPBW: Coarse Pinkish Brown Ware.

The record of two crenate phytoliths in the paste appears more problematic. Crenate phytoliths are typical of the Pooideae grass subfamily (Brown 1984; Piperno and Pearsall 1998; Twiss 1992; Twiss, Suess, and Smith 1969). High temperatures and low rainfall prevail in the source area, but Pooideae favor cold conditions that, between the Tropics of Cancer and Capricorn, are encountered only at high altitude (Twiss 1992, 116). Also taking into consideration the geographic location of the possible provenance area, the presence of crenate phytoliths might result from transport, either by wind or run-off.

Coarse Buff Orange Ware (AR 5387, AR 5559, AR 5554, and AR 5561) Four sherds of this type of ware were studied (see appendix, part 2). Three of the sherds (AR 5387, AR 5559, and AR 5561) contain abundant, very poorly sorted sand- and gravel-sized sedimentary rock fragments in a groundmass with variable carbonate content. Mineral fragments, among which quartz predominates, are rather few and small in size, and show a substantial amount of secondary calcite. The occurrence throughout the sample of lithic clasts of magmatic origin significantly distinguishes AR 5554 from the other samples. A local origin is proposed for AR 5387, AR 5559, and AR 5561, while the origin of AR 5554 remains unclear. The phytolith content of these sherds is not uniform: AR 5559 does not present phytoliths, AR 5387 has palm phytoliths in the fired matrix, and AR 5554 has them in pores. Grass phytoliths occur in the pores of AR 5561 (table 2.3b). The phytolith analysis, therefore, does not help to locate the provenance area of the Coarse Buff Orange Ware.

40

methodologies in archaeobotany

Conclusions Phytoliths were observed in the ed-Dur ceramic thin sections. Palm phytoliths are present in the fired matrix. Grass phytoliths are noted in the voids and are mostly rounded trapeziforms and identify the grass temper, possibly Sporobolus sp. material collected between October and April. Different samples of a ware type present similar phytolith distribution and quantities. Phytoliths from pottery thin sections can contribute to the identification of the source area for all the wares originating from the ophiolite complex of the Omani mountains. The phytoliths distributed in the fired matrix support the clay source area proposed by the petrographic and chemical analysis. By combining these data with the possible identification of Sporobolus sp., it is suggested that we can distinguish at least two manufacture areas, one of which lay near the clay and temper sources. In this respect, phytolith analysis substantiates the archaeologically based suggestion of several regional production centers located within southern Arabia. It has been proposed that ed-Dur and Mleiha were the main settlements in the southern Arabian Gulf from the first century BC to the first century AD. These settlements were in very close contact with each other and interdependent: ed-Dur acted as a harbor and provided goods and food traded from the sea, while Mleiha delivered agricultural products and other commodities lacking on the coast. The present study establishes that these local exchanges were not exclusive and that there were other regional relations.

Acknowledgments The authors wish to thank the reviewers, Professor De Langhe and Dr. A. Degraeve, for their stimulating comments and review of our chapter. We also want to thank the editors, who offered us several opportunities to present our results and significantly helped to improve our contribution.

Appendix 1. Descriptive Indexes for Phytolith Analysis of Ceramic Thin Sections In ceramics, a distinction can be made between two fabrics: fired-clay matrix or paste and coating. Paste is the heated fabric. Phytoliths are part of the inorganic residue of biological origin of the paste, together with diatoms, sponge spicules, and chrysophycea. They

Regional Exchanges in Southeastern Arabia

41

are part of the basic mineral component of ceramic and are described using the standard concepts of phytolith analysis of soil thin sections (Bullock et al. 1985; Vrydaghs et al. 2007). Two indexes are basic to phytolith analysis of thin sections: Observed/ Not Observed (O/NO) and Morphotype Identification (MI). Phytolith analysis of ceramic thin sections also involves a Distribution index (D). The O/NO index refers to whether phytoliths are observed or not. Because of the wedging effect (Stoops 2003), no definitive conclusions about phytolith absence can be presented before conducting a scan at magnification ×1000. The observation of one phytolith suffices for a positive O/NO index, whether or not the phytoliths are relevant for the archaeological or paleoenvironmental interpretation. In the case of ceramic thin sections, to establish these indexes, the longest and widest transects through the sherd sections are scanned at magnifications ×250, ×400, and ×1000. The MI index classifies all the observations in identified or unidentified morphologies. The latter may be explained by the phytolith orientation in the thin section or the phytolith preservation. The identified types can be classified according to two systems. The first one considers simple geometrical shapes by using charts (see SACDBT 1962). This planar description of volumes is strictly descriptive and may confuse morphotypes, thus introducing errors in paleoenvironmental and archaeological interpretation. The second approach is a description referring, when available, to a specific reference collection. For the present study, one reference collection was available for Oman, that of Ball (2002). The D index refers to whether the phytoliths are present in voids or in the paste. It is assumed that phytolith occurrences in voids result from the decomposition of plant used as temper, while those in the paste result from the natural composition of the clay.

2. Petrographic and Chemical Analysis of Some Ceramic Groups Discussed in This Chapter Coarse Yellow-Pink Speckled Ware (AR 5557). These five sherds (0.05 percent of the complete assemblage) are typified by a very hard but porous fabric with a mixed coloring of pink and yellow speckles. The temper consists of coarse red, grey, and black rounded and subrounded grits, and a small amount of fine calcium particles. Most fragments belong to large thick-walled storage jars. Thin-section analysis demonstrates that sherd AR 5557 of the Coarse Yellow-Pink Speckled Ware is tempered with abundant angular to subangular quartz and feldspar grains which are associated with a subordinate amount of pyroxene crystals and lithic clasts. Most quartz and feldspar grains are of very fine sand or silt size. The dominant feldspar is

42

methodologies in archaeobotany

albite-twinned plagioclase, whereas the pyroxene grains seem to have an augitic composition. The rock fragments mainly consist of serpentinized peridotite, radiolarite, and spilitic lava. The serpentinized ultramafic rock fragments are up to 1.8 mm across with an average grain size of about 0.8 mm. They are usually rounded and equidimensional to disk shaped. Much of the peridotitic nonplastic inclusions are strongly altered and oxidized, leaving very few primary olivine and pyroxene crystals unaffected. When the serpentinization process has gone to completion, the preexisting mafic minerals have been wholly converted into a fibrous mixture of minerals of the serpentine group with a highly distinctive mesh structure. The microscopic observations also show the presence of secondary carbonates in the sample. Rounded trapezoid, trapezoids, long cell morphotype, and unknown morphotype are observed in pores. No spherical echinate were recorded. The rock fragments and the composition of the clinopyroxene crystals embedded in sherd AR 5557 suggest derivation from the Semail (Oman) ophiolite complex. Serpentinized peridotite, radiolarite, and spilitic lavas are known to be major constituents of this complex. In contrast to the Orange Vegetal Tempered Ware, the Coarse Yellow-Pink Speckled Ware material submitted to phytolith analysis was too limited to provide convincing complementary evidence about the source area. Meanwhile, it is worth recalling that phytoliths typical for Palmae were not observed. Coarse Pinkish Brown Ware (AR 5562). This vessel was made with calcareous raw materials that included relatively few quartz and plagioclase fragments. The average size of the latter is of the order of 0.1–0.2 mm. The majority of the nonplastic inclusions correspond with lithic clasts characterized by a mineralogical composition and textural features indicative of a provenance from an area with abundant outcrops of peridotitic ultramafic rocks. The rock fragments are either sand or gravel sized and generally rounded in shape. The fired clay also carries charred vegetal substances. Microscopic investigation demonstrates that AR 5562 has been seriously affected by a process of secondary calcite deposition (calcitization). This process took place during the burial of the material in the soil. It is beyond dispute that the tempering elements of AR 5562 come at least in part from the Semail ophiolite in Oman. An origin in a workshop located somewhere in Oman is one of the possibilities. Coarse Ochre Ware (AR 5560). The Black Ware, composing 0.09 percent of the complete pottery assemblage, shows an ochre coloring on the exterior and a dark grey color on the interior of the vessels. The paste is tempered with coarse brown, red, and black subrounded grits, which are badly sorted throughout the matrix. Shapes are limited to two types of large open bowls. Material from the Coarse Ochre Ware carries abundant mineral and rock fragments that point to a peridotitic source region. In addition, it

Regional Exchanges in Southeastern Arabia

43

contains a considerable quantity of charred vegetal matter. The mineral fraction of the temper is dominated by grains of quartz and carbonate minerals of the finer size grades. Clinopyroxene, plagioclase feldspar, and opaque are minor constituents. The lithic debris, as well as some quartz and clinopyroxene crystals, ranges up to 3.5 mm in size. In contrast to AR 5557, many peridotite particles embedded in AR 5560 are only partially serpentinized and thus still enclose primary olivine, clinopyroxene, enstatite, and opaque minerals. Coarse Buff Orange Ware (AR 5387, AR 5559, AR 5554, and AR 5561). Three of these sherds (AR 5387, AR 5559 and AR 5561) contain abundant, very poorly sorted sand- and gravel-sized rock fragments of sedimentary origin taken in a groundmass with a very variable carbonate content. The rock fragments are very diverse in composition and up to 5.5 mm across. They include mudstone, shale, fine-grained sandstone, and limestone. The shape of mudstone fragments is invariably angular or elongated, whereas sandstone and limestone fragments tend to be subangular or subrounded. Some mudrock debris is very rich in carbonaceous matter, making it almost nontransparent in plane-polarized light. Mineral fragments are rather few, of small size grade, and quartz predominates by a large amount. They carry a substantial amount of secondary calcite and remains of plant material. With AR 5554, the abundant nonplastic inclusions scattered throughout the preparation are very poorly sorted and comprise both mineral and rock fragments. Mineral grains are less abundant than lithic clasts. Among the former, quartz and plagioclase feldspars are the commonest. Much lithic debris is of magmatic origin, with spilite, granite, and peridotite being the most widespread rock varieties of this category. Sedimentary rock fragments essentially consist of mudstone, sandstone, and chert. They may contain substantial amounts of carbonaceous material, very often subrounded. The maximum grain size of the lithic inclusions is about 3.5 mm. Remains of plant material have also been identified under the microscope. It is proposed that AR 5387, AR 5559, and AR 5561 originated from the UAE, while the origin of AR 5554 remains unclear.

References Ball, T., comp. 2002. “Dhofar Phytolith Reference Collection.” Compact disc. http://www.phytolithsociety.org/spr-resources.html Boucharlat, R. 2003. “Iron Age Water-Draining Galleries and the Iranian ‘Qanat.’” In Archaeology of the United Arab Emirates, edited by D. Potts, H. Al Naboodah, and P. Hellyer, 162–72. Proceedings of the First International Conference on the Archaeology of the U.A.E. London: Trident.

44

methodologies in archaeobotany

Boucharlat, R., and M. Mouton. 1993. “Mleiha (3e s. avant J. C.–1er/2e s. après J.C.).” In Materialien zur Archäologie der Seleukiden- und Partherzeit im südlichen Babylonien und im Golfgebiet, edited by U. Finkbeiner, 219–49. Tübingen: Wasmuth. Brown, D. A. 1984. “Prospects and Limits of a Phytolith Key for Grasses in the Central United States.” Journal of Archaeological Sciences 11:345–68. Bullock, P., N. Fedoroff, A. Jongerius, G. Stoops, and T. Tursina. 1985. Handbook for Soil Thin Section Description. Wolverhampton: Waine Research. De Cardi, F., D. Kennet, and R. L. Stocks. 1994. “Five Thousand Years of Settlement at Khatt, U.A.E. ” Proccedings of the Seminar for Arabian Studies 24:57–79. De Paepe, P., K. Rutten, L. Vrydaghs, and E. Haerinck. 2003. “A Petrographic, Chemical and Phytolith Analysis of Late Pre-Islamic Ceramics from ed-Dur (Umm al-Qaiwan, U.A.E.).” In Archaeology of the United Arab Emirates, edited by D. Potts, H. Al Naboodah, and P. Hellyer, 207–28. Proceedings of the First International Conference on the Archaeology of the U.A.E. London: Trident. Haerinck, E. 1998. “International Contacts in the Southern Persian Gulf in the Late 1st Century BC/1st Century AD: Numismatic Evidence from ed-Dur (Emirate of Um al-Qaiwan, U.A.E.).” Iranica Antiqua 33:273–302. ———. 2003. “Internationalisation and Business in Southeast Arabia during the Late 1st Century BC/1st Century AD: Archaeological Evidence from ed-Dur (Umm al-Qaiwan, U.A.E.).” In Archaeology of the United Arab Emirates, edited by D. Potts, H. Al Naboodah, and P. Hellyer, 195–206. Proceedings of the First International Conference on the Archaeology of the U.A.E. London: Trident. Højlund, F., and H. H. Andersen. 1994. Qala’at al- Bahrain, 1. The Northern City Wall and the Islamic Fortress. Aarhus: JASP 30/1. Humphries, D. W. 1992. The Preparation of Thin Sections of Rocks, Minerals, and Ceramics. Microscopy Handbooks, vol. 24. Oxford: Royal Microscopical Society. Jongbloed, M. 2003. “Pottery, Past and Present.” Al-Shindagah 51:1–4. Kennet, D. 1998. “Evidence for 4th/5th-Century Sasanian Occupation at Khatt, Ras al-Khaimah.” In Arabia and Its Neighbours: Essays on Prehistoric and Historic Developments Presented in Honour of Beatrice De Cardi, edited by C. S. Philipps, D. T. Potts, and S. Searight, 105–16. Turnhout: Abiel II. Kervran, M., and F. Hiebert. 1991. “Sohar pré-islamique: Note stratigraphique.” In Golf-Archäologie, edited by K. Schippman, A. Herling, and J. F. Salles, 337– 46. Internationale Archäologie, 6. Buch am Erlbach: M. L. Leidorf. Lecomte, O. 1993. “Ed-Dur, les occupations des 3e et 4e s. ap. J.C.: Contexte des trouvailles et material diagnostique.” In Materialien zur Archäologie der Seleukiden- und Partherzeit im südlichen Babylonien und im Golfgebiet, edited by U. Finkbeiner, 195–217. Tübingen: Wasmuth. Lombard, P., and M. Kervran. 1993. “Les niveaux hellénistiques du Tell de Qala’at al-Bahrain: Données prémiminaires.” In Materialien zur Archäologie

Regional Exchanges in Southeastern Arabia

45

der Seleukiden- und Partherzeit im südlichen Babylonien und im Golfgebiet, edited by U. Finkbeiner, 127–60. Tübingen: Wasmuth. Lombard, P., and J. F. Salles. 1984. La nécropole de Janussan (Bahrain). Lyon: TMO. Madella, M., A. Alexandre, and T. B. Ball. 2005. “International Code for Phytolith Nomenclature 1.0.” Annals of Botany 96:253–60. Mouton, M. 1992. “La péninsule d’Oman de la fin de l’âge du fer au début de la période sassanide (250 av. –350 ap. JC).” PhD diss., University of Paris. Murphy, C. P. 1986. Thin Section Preparation of Soils and Sediments. Berkhamsted, UK: AB Academic. Papadakis, J. 1966. Climates of the World and Their Agricultural Potentialities. Buenos Aires: published by the author. Piperno, D. R., and D. M. Pearsall. 1998. The Silica Bodies of Tropical American Grasses: Morphology, Taxonomy and Implications for Grass Systematics and Fossil Phytolith Identification. Smithsonian Contributions to Botany, vol. 85. Washington, DC: Smithsonian Institution Press. Rands, R. 1987. “Phytoliths in Archaeological Ceramics: Data from the Palenque Region, Mexico.” Phytolitarian Newsletter 4 (3): 5–6. Rice, M. 1994. The Archaeology of the Arabian Gulf. London and New York: Routledge. Runge, F. 1999. “The Opal Phytolith Inventory of Soils in Central Africa—Quantities, Shapes, Classification and Spectra.” Review of Palaeobotany and Palynology 107:23–53. Rutten, K. 2007. “The Roman Fine Wares of ed-Dur (Umm al-Qaiwain, U.A.E.) and Their Distribution in the Persian Gulf and the Indian Ocean.” Arabian Archaeology and Epigraphy 18 (1): 8–24. SACDBT (Systematic Association Committee for Descriptive Biological Terminology). 1962. Terminology of Simple Symmetrical Plane Shape. Chart 1a. Taxon 11, 145–56; 245–47. Salles, J. F. 1984. “Céramique de surface à Ed-Dour, Emirats Arabes Unis.” In Arabie Orientale, Mésopotamie et Iran Méridionale de l’Âge du Fer au Début de la Période Islamique, edited by R. Boucharlat and J. F. Salles, 241–70. Mémoire, 37. Paris: Editions Recherche sur les Civilisations. Sanlaville, P. 1992. “Changements climatiques dans la péninsule arabique durant le Pléistocène Supérieur et l’Holocène.” Paléorient 18 (1): 5–26. Stocks, R. 1992. “Wadi Hakil Survey: November 1992.” Proceedings of the Seminar for Arabian Studies 22:145–63. Stoops, G. J. 2003. Guidelines for Analysis and Description of Soil and Regolith Thin Sections. Edited by M. J. Vepraskas. Madison, WI: Soil Science Society of America. Tengberg, M. 2002. “Vegetation History and Wood Exploitation in the Oman Peninsula from the Bronze Age to the Classical Period.” In Charcoal Analysis: Methodological Approaches, Palaeoecological Results and Wood Uses: Proceedings of the Second International Meeting of Anthracology, Paris, September

46

methodologies in archaeobotany

2000, edited by S. Thiébault, 151–57. BAR International Series, 1063. Oxford: Archaeopress. Twiss, P. C. 1992. “Predicted World Distribution of C3 and C4 Grass Phytoliths.” In Phytolith Systematics: Emerging Issues, edited by G. Rapp Jr. and S. C. Mulholland, 113–28. Advances in Archaeological and Museum Science, 1. New York and London: Plenum Press. Twiss, P. C., E. Suess, and R. M. Smith. 1969. “Morphological Classification of Grass Phytoliths.” Proceedings of the Soil Science Society of America 33:109–15. Vrydaghs, L. 2003. “Studies in Opal Phytolith: Material and Identification Criteria.” PhD diss., Ghent University. Vrydaghs, L., Y. Devos, K. Fechner, and A. Degraeve. 2007. “Phytolith Analysis of Ploughed Land Thin Sections: Contribution to the Early Medieval Town Development of Brussels (Treurenberg site, Belgium).” In Plants, People and Places: Recent Studies in Phytolith Analysis, edited by M. Madella and D. Zurro, 13–27. Oxford: Oxbow Books. Vrydaghs, L., H. Doutrelepont, H. Beeckman, and E. Haerinck. 2001. “Identification of a Morphotype Association of Phoenix dactylifera L. Lignified Tissues Origin at ed-Durr (1st AD), Umm Al-Qaiwan (U.A.E.).” In Phytoliths: Applications in Earth Sciences and Human History, edited by J. D. Meunier and F. Colin, 239–50. Amsterdam: Balkema.

C h ap t e r 3

Examining Agriculture and Climate Change in Antiquity Practical and Theoretical Considerations Alexia Smith

Over the past few decades, as more data have been gathered on global warming, environmental and agricultural scientists have become increasingly concerned with the impact that climate change will have on different societies’ abilities to produce food. Numerous simulation models have estimated the effects of projected climate change on agricultural production, and it is clear that the climate shift will not be felt equally across the globe. Semiarid parts of the world are currently becoming more arid, which will add further constraints to agricultural production, monsoonal areas are experiencing more erratic rainfall, and temperate areas are witnessing enhanced plant growth due to higher levels of atmospheric carbon dioxide and longer growing seasons (Braswell et al. 1997; Peñuelas and Filella 2001; Vörösmarty et al. 2000). Different areas, therefore, experience climate change in different ways on both a macro- and microlevel. The current climatic flux results from human-induced greenhouse gas emissions and, based on long-term paleoclimate studies, Ruddiman (2003) argues that the rise of anthropogenic greenhouse gases likely began around 8,000 years ago following the intensification of land clearance for cultivation in Eurasia; this effect increased around 5,000

47

48

methodologies in archaeobotany

years ago owing largely to wetland rice production in Southeast Asia. While these changes had a slow, yet large cumulative effect upon climate change, greenhouse gas emissions reached unprecedented levels following the Industrial Revolution. As a result, the rate of human-induced climate flux currently differs from that of the past. Nevertheless, it is clear that agriculture is affected by changes in climate, and there is no reason to doubt that it would have been affected in antiquity. The question of how climate affects populations has been considered for centuries. In the 1300s Ibn Khaldu¯n, a prominent Tunisian statesman and scholar, wrote a history of the world, the Muqaddimah, in which he outlined his view of the nature of civilization and described the dependence of both urban and nonurban societies on their physical environment (Ibn Khaldu¯n 1989). Borrowing his ideas from Greek philosophers and geographers (Issawi 1952), Ibn Khaldu¯n further argued that geographical variation in climate led to differences in character, appearance, and custom, with civilization being concentrated in the midlatitudes where the “heat” is temperate and not too excessive. Such arguments would not be readily accepted within current anthropological thought, but the Muqaddimah provides an eloquent insight into views on the role of climate from a preindustrial Moorish perspective and demonstrates that thinking on the relationship between people and climate has deep historical roots. Climate is an omnipresent variable that every organism on the planet experiences, so it is a worthwhile endeavor to better understand its role, particularly in relation to food production. The issue should be approached from a wide range of disciplines, and archaeology has much to offer. The aim of this chapter is to discuss how climate change and agricultural production in antiquity may be considered together, using the Bronze and Iron Ages of southwest Asia as an example. For several reasons, the Levant, spanning southeastern Turkey and Syria down to the Sinai, together with eastern Syria, provides an interesting context within which to examine paleoagriculture in relation to climatic change. First, the steep NW–SE precipitation gradient makes the area very sensitive to climatic variation. Second, numerous archaeobotanical and zooarchaeological studies have been conducted there. Third, a wide range of paleoclimatic data from the region suggest that around 2250–2200 BC, the region began to experience much drier conditions, and finally, at this time the area also witnessed a large-scale urban demise and a shift toward greater emphasis upon pastoralism. Current limitations to furthering our knowledge of the impact of climate change on agriculture in antiquity are discussed. Practical considerations of data collection and management are presented first, followed by a discussion of problems of establishing contemporaneity between

Examining Agriculture and Climate Change

49

paleoclimatic and archaeological records. Data interpretation is considered last; a model describing potential responses to climatic change is presented, and the theoretical concerns with this model are discussed.

Regional Approaches to Agriculture: Practical Considerations The Oxford English Dictionary defines agriculture as “the science and art of cultivating the soil; including the allied pursuits of gathering in the crops and rearing live stock; tillage, husbandry, farming (in the widest sense).” While this definition is a modern one, and the specifics of the integration of plants and animals would have varied across space and through time, it is fair to assume that, by the Bronze Age in southwest Asia, cultivation and animal herding overlapped to some degree with the symbiotic benefits being realized by farmers. Such overlap could include the feeding of fodder, crop processing by-products, or stubble to animals as well as the use of burned or fresh manure as a fertilizer. Consequently, in order to better understand the dynamics of agricultural production, it is essential that plant and animal data sets be considered together. Owing to the amount of time it takes to train as a specialist, the need for separate subdisciplines such as archaeobotany and zooarchaeology is clear, but this does not change the fact that such a separation is an academic distinction and that, alone, neither of the subdisciplines can adequately reflect the complexities of food production in their entirety. With this issue in mind, a Near Eastern Agricultural Database (NEAD) has been constructed using Microsoft Access. Relational databases greatly enhance the efficiency of storing and examining large and complex data sets, and allow for a wide range of questions to be examined. The structure of the database and the range of information stored on both plants and animals are shown in figure 3.1. Since the importance of archaeological context in the interpretation of plant remains is now well understood, contextual information lies at the heart of the database, allowing for easy comparison of like with like (Al-Azem 1992; Hansen 1991; Hillman 1984, 1991; Jones 1984, 1998a; Jones and Halstead 1995; McCorriston 1998; Miller and Smart 1984; Van der Veen 1992; Wasylikowa 1981). There are many advantages to creating databases of published ecological data, since an understanding of regional patterns provides a framework within which to examine site-based data.1 Over the past several decades, archaeobotany has rapidly matured as a discipline, and while the number of zooarchaeological studies in southwest Asia exceeds that of archaeobotanical studies, more plant data are being published each year. Many

Chronology Reference ID Phase code* Dating method Chronological period Lower date Upper date Revised dates Uncalibrated C14 Standard deviation Lab code AMS (yes/no) Notes 1

1

Context Table Site code Reference ID Context code* Phase code Context type Plant sample () Plant recovery type Sediment volume Sediment mass Flot volume Flot mass Portion of sample sorted Animal sample () Bone recovery method Notes 1

8 8

Animal sample Animal sample ID* Context code Animal taxon code Bone type NISP MNI MNE MAU Weigh Presence only () Possible contaminant Level of identification Notes

Plant sample Plant sample ID* Context code Plant taxon code Level of identification Number per sample Ubiquity Presence only () Abundance Preservation type Possible contaminant Notes 1

1

8

Plant Taxa Table Taxon name

Seed dimensions Measurement ID* Plant sample ID Length Breadth Thickness Diameter

1

8

8

Animal Taxa Table Taxon name Taxon code* Family Common name Habitat Notes

Bone dimensions Measurement ID* Animal sample ID Sex Age Symmetry Domesticated () Length Fragment () End

Family Authority Common name Habitat Ethnobotanical uses Notes

1 Taxon code*

8 8

8

8

Figure 3.1.  Schematic diagram of the Near Eastern Agricultural Database showing the relationships between tables. Asterisks indicate the primary key for each table that is used to provide a unique link between tables.

References 1 Reference ID* 1 Primary author Date of publication Seasons covered Report type Preliminary () Specialist Notes

Site Table Site name* Site code Alternate names Region, Country Latitude, Longitude Altitude Site size Directors Notes

8

Examining Agriculture and Climate Change

51

important questions have been answered at the site level, including the definition of use areas and the nature of consumption across social strata. Such research forms the building blocks of broader considerations, but in order to adequately evaluate changes in agricultural practices in relation to climate, more data are required than a single site can possibly offer. By adopting a regional approach and examining published data from multiple sites, the number of samples available for consideration greatly increases, and identifying patterns of crop use and pasture exploitation becomes possible. Work by Colledge, Conolly, and Shennan (2004) examining Neolithic data from southwest Asia and southeastern Europe, for example, has highlighted interesting regional patterns in cultivation and has furthered knowledge on the origins and spread of agriculture. Riehl (this volume) has also identified interesting patterns in wild flora that reflect variations in local habitats. Databases of published data are not without their problems, however, and several of the more pressing issues are outlined below. The examples relate specifically to plant data, although animal data would be affected by broadly similar concerns.

Using Published Material as Data Firsthand experience with material through all stages of excavation, flotation, and analysis undoubtedly provides optimal conditions for a researcher, but this is not always possible, practical, or economically feasible. The ability to use published archaeobotanical and zooarchaeological data easily is an important issue, as it is with other archaeological reports, since once a site is excavated, these reports oftentimes provide the most accessible means of preserving the information acquired. Published reports stand as a lasting record of the remains, so if the data are not usable, solutions need to be developed. In 1977, in reaction to a synthesis of European archaeobotanical data by Hubbard (1976), Dennell (1977, 363–64) stated that “the use of published material is exceptionally hazardous, since the provenance of each sample, and the amount of mixing before, during and after excavation are rarely known.” Methods of analysis and data presentation have evolved considerably since 1977, and researchers routinely present contextual information along with their data, although the thoroughness of presentations can vary depending upon the nature of the report. Contexts can be described by a single word, such as “pit,” “hearth,” or “silo,” or by lengthy descriptions of the sediment types. It is true that unless stated by an author, the “amount of mixing before, during and after excavation” cannot be known, and, in order to use the published data, it must be assumed that since the

52

methodologies in archaeobotany

researcher considered the sample worthy of study, the context is secure. This may not always be the case.

Comparing Samples Collected and Analyzed by Different People Regional considerations necessitate the integration of data gathered and presented by different people. If we were interested in broad changes in cultivation over large geographic areas, a coordinated method of sample collection would clearly be ideal. Such an approach, while possible and desirable in the long term, would present huge logistical difficulties and immense financial and labor costs, so amassing available data provides a useful beginning. The difficulties involved in synthesizing data collected by different people cannot be dismissed, however. Potential variation could arise from different research agendas, sampling techniques, recovery methods, and levels of identification resulting from differential strengths of available reference material and the expertise and experience of the investigator, as well as the main goal of the published report. Sampling in any discipline is primarily concerned with obtaining sufficient data to understand the whole by examining only a part. Simple random sampling, while perhaps the best approach from a statistical standpoint, is not ideally suited to the excavations of tells, where certain areas are targeted to minimize costs and maximize work completed during the excavation season. Within a given area, archaeobotanical samples can be collected systematically from each locus or context, or collected only when an excavator sees charcoal or an ashy layer. The latter approach, termed targeted sampling, could lead to bias within the data, even though flotation samples are rarely collected with a specific knowledge of their botanical contents. A significant proportion of the samples in NEAD was collected when charcoal was visible to the excavator and not as part of a well-defined sampling strategy. This is a challenging problem, one that can persist even when extensive sampling is conducted and only the richest samples are chosen for examination under the microscope. Databases such as NEAD allow for a large number of samples from any given context type to be considered together. Recovery rates also differ depending upon the technique used to extract plant material from sediments (Wagner 1998; Watson 1976). The majority of archaeobotanical samples published for southwest Asia were collected using some form of flotation, although a number (mostly those collected before the advent of flotation) were handpicked. In order to employ such data, it is necessary to divide the data set and compare samples with similar recovery methods.

Examining Agriculture and Climate Change

53

Furthermore, when using published data, it must be assumed that the identifications are correct. Many domesticates, particularly cereals, can be tricky to identify (Jones 1998b; Kroll 1992; Nesbitt and Samuel 1996). Little can be done to isolate misidentifications, unless a species is rarely found in an area and its presence in the record is not explained or supported. In instances where no support is provided for species identifications that have morphological overlaps with others, such as the grains of Triticum aestivum L. (bread wheat) and T. durum Desf. (durum wheat), specimens can be entered at a lower level of classification (e.g., T. aestivum/durum), thereby losing some degree of accuracy but minimizing the potential for error. Regarding the level of identification, little can be done to control for differential preservation, differential access to reference material, or expertise of the investigator. In order to synthesize data it must be assumed that researchers identify remains to a level they are comfortable with (e.g., to the genus or species level). More experienced investigators are likely able to identify a wider range of species. It should be stressed, however, that experience may lead to more conservative identifications due to a greater awareness of subtle differences between various species. By eliminating rare taxa that appear in less than 10 percent of the samples, the influence of differential abilities will be offset to some degree, although not removed entirely. As Jones (1991) has shown, this also has the benefit of reducing noise when using multivariate statistics to analyze the data.

Geographical and Temporal Biases For the most part, geographical and temporal biases in Bronze and Iron Age archaeobotanical data from southwest Asia reflect the history of excavation and flotation in southwest Asia. First, the distribution of published archaeobotanical data does not adequately reflect the distribution of sites observed in settlement surveys. The majority of studies relate to long-term settlement; rural settlements represent a minority (e.g., Lines 1995), and no botanical samples have been collected from short-term occupations or from pastoral camps. The data, therefore, reflect a predominantly urban perspective of agriculture. Recent research is highlighting the importance of investigating pastoral camps, but as noted by Hole (1991) in reference to Syria, they may be difficult to locate undisturbed. Also, many (although not all) of the sites excavated are located in river valleys, which restricts interpretations to this type of agronomic environment. The 200–250 mm isohyet is considered to be the cutoff for rainfed agriculture (Barrow 1993), so excavation of sites in areas below the modern-day limit could yield interesting data on ancient plant use and

54

methodologies in archaeobotany

animal herding strategies. Excavations in areas as dry as this, such as Tell Sheikh Hamad in the Syrian Jazira (Kühne 1991), are rare, however, and interpretations cannot be extended to more arid parts of the landscape until further research is conducted there. Finally, a temporal bias exists in the data, with far more Bronze Age sites being represented than Iron Age sites.

Methodological Concerns It is currently difficult to fully integrate plant and animal data. As with plant remains, well-established methods exist for interpreting zooarchaeological data (e.g., Grayson 1984; Meadow and Zeder 1978; Sherratt 1983), but to date, few methods are available for considering them together; problems include differences in preservation between bones and plant remains as well as methods of weighing the relative importance of different species to the overall agricultural economy. The difficulties of interpreting quantitative data for plant and animal remains have been widely discussed as separate topics (e.g., Grayson 1984; Popper 1988), and these difficulties would be compounded when evaluating quantitative data together. Much work needs to be done before reliable methods can be established, but the construction of an agricultural data set that incorporates both archaeobotanical and zooarchaeological data from multiple sites within a region provides a beginning, and the use of exploratory multivariate statistics, such as correspondence analysis of presenceonly data, should prove fruitful in detecting patterns within the data.2 Despite the limitations of synthesizing published data, much can still be learned about the regional agricultural economy, and both socioeconomic and environmental questions can be examined. Since the entry of the zooarchaeological data is ongoing, it is too early to discuss regional patterns of plant and animal use from a statistical perspective, but several qualitative comments can be made. Work by Zeder (1991, 1995) and McCorriston (1995, 1998) has highlighted the importance of sheep in the third millennium BC, and texts from Ebla (Archi 1990; Pinnock 1990) underscore the importance of wool to the regional economy. The distribution of dye plant finds, such as Carthamus tintorius L. (safflower), fits this understanding of textile manufacture very well. Safflower produces vermilion to orange-yellow flowers, usually between July and August, and the dried petals can be used to dye wool pink to red depending upon the concentration of dye (Davis 1975). Throughout the entire region, Carthamus sp. or C. tinctorius remains are only found in northern Syria spanning from eastern Syria as far west as Umm el-Marra in the Jabbul Plain (Colledge 2003; Miller 2004; Riehl 2001; Van Zeist 2003, 1999/2000;

Examining Agriculture and Climate Change

55

Van Zeist and Bakker-Heeres 1985; Van Zeist et al. 2003). No finds have been published from the Levant, and the majority of the remains date to the first three-quarters of the third millennium BC, when the textile trade was flourishing. The concentration of finds in northern Syria likely reflects a specialized local economy. Several later finds date to the middle Bronze Age, but none have been reported for the EBIV or the late Bronze and Iron Ages. It is interesting that these dye plants are concentrated in an area where texts and zooarchaeological remains suggest an increase in sheep production and where the remains of burned dung suggest a heavy reliance on pasture, as indicated, for example, by large concentrations of Trigonella astroites Fisch. et Mey., and other pasture plants (e.g., Matilla Séquer and Nivera Núñez 1993/94; McCorriston and Weisberg 2002; Van Zeist 1999/2000; Van Zeist and Bakker-Heeres 1985). When considering plant and animal data together, burned dung fuel perhaps provides the most immediately useful method of integration, since identified taxa could yield information on regional patterns of fodder and pasture (Charles 1998; Miller 1984, 1996; Miller and Smart 1984). As Riehl has demonstrated (this volume), regional differences in wild flora do exist, and these differences can be used to examine the types of pasture exploited around a site.

Dating and the Paleoclimatic Record With an understanding of regional patterns in crop use, wild flora, and animal husbandry, agricultural change through time can be discerned more easily. Clearly, many social or political factors can lead to modifications in agricultural production over time, but as recent experience has shown, the impact of climate flux cannot be discounted. It is essential to establish contemporaneity between paleoclimatic and archaeological records before any causal link can be thoroughly assessed; currently, paleoclimatic data and archaeological evidence from southwest Asia can be compared at a temporal resolution of fifty to a hundred years, so many subtle changes may be obscured. This resolution results from probabilistic errors associated with radiocarbon dates of paleoclimatic proxies as well as from a paucity of radiocarbon dates from many Bronze and Iron Age sites and uncertainty regarding relative dates. The age of archaeobotanical and zooarchaeological samples is usually based on associated ceramic assemblages, and consequently some samples fall within a wide age bracket. Such a wide age range may be appropriate for samples collected from middens that have accumulated over long periods of time (see McCorriston and Weisberg 2002), but short-term context types such as hearths or silos should be directly radiocarbon dated. While radiocarbon

56

methodologies in archaeobotany

dating large numbers of samples would prove expensive, this is the most accurate method of demonstrating contemporaneity between paleoclimatic and archaeological data. Regarding paleoclimate, a large amount of proxy data is available from southwest Asia including palynological, anthracological, isotopic and mineralogical, and geological data (e.g., Bar-Matthews et al. 2003, 1999; Baruch 1986; Baruch and Bottema 1999; Goodfriend 1991; Stiller et al. 1983/84; Wick, Lemcke, and Sturm 2003). These records have predominantly been dated using 14C dating, although speleothem records from Soreq Cave in Israel were dated using the 230Th/U method (Kaufman et al. 1998; Kolodny et al. 2003), and the most recent records from Lake Van were dated using annual varve counts with an estimated counting error of ±0.6 percent (Landmann et al. 1996, 109). The temporal resolution (i.e., number and spacing of samples analyzed), as well as dating accuracy (which varies with the material dated and technique used), varies for each record, but uncertainty is drastically minimized when independent lines of evidence suggest a similar change. The bulk of the evidence suggests that a climatic shift occurred in southwest Asia toward the end of the third millennium BC. Evidence from Lake Van suggest a shift to drier conditions around 2210 (±13) BC based on sedimentation rates, or 2240 (±13) BC based on δ18O and Mg/Ca concentrations (Wick, Lemcke, and Sturm 2003). Issues of interpreting paleoclimatic data also need to be considered. Using pollen records from southwest Asia, Bottema (1997) has argued that there is no evidence for climate change at the end of the third millennium BC. As Miller demonstrates in this volume, palynological data are not always a useful proxy for climate change, particularly for the Bronze and Iron Ages when there is evidence for land clearance and deforestation (Miller 1990, 1997, 2004). A number of studies have demonstrated that urban societies greatly affected their surroundings in the past, either intentionally or unintentionally, and extensive use of wood for fuel or construction would clearly have affected forest composition and, thus, pollen assemblages (e.g., Redman et al. 2004). Intensive grazing would also have affected pasture flora. Fluxes in pollen assemblages represent, therefore, the end product of (species-dependent) lagged responses to climate change as well as human-induced modifications to the landscape. Separating the effects of paleoclimatic change from human-induced paleoenvironmental change using pollen data is an immensely difficult task. Since other paleoclimatic proxies, such as oxygen isotope analysis, have a much weaker anthropogenic signature, they can be considered a more reliable record of climate change. The ability to establish contemporaneity between paleoclimatic and archaeological data is an important concern. Recent attempts to use

Examining Agriculture and Climate Change

57

wiggle-match dating to reduce probabilistic error terms hold much promise for improving the chronological resolution of archaeological remains and certain paleoclimatic proxies, so the dating accuracy will likely improve in the near future (Gilboa and Sharon 2003; Manning et al. 2003; Speranza, Van der Plicht, and Van Geel 2000). These issues aside, the extent to which changes through time can be examined using Bronze and Iron Age data from southwest Asia is currently hampered by a paucity of archaeobotanical studies of remains dating to the middle and late Bronze Ages and the Iron Age.3 Thus, more studies of plant use from each time period across the region are needed.

Considerations of Climate and Agriculture: Theoretical Concerns Several scholars, notably Weiss (Weiss 1997; Weiss and Bradley 2001), have argued that climate change led to social collapse at the end of the third millennium BC in southwest Asia, but such claims are hotly contested (e.g., Zettler 2003). Establishing a link between climate change and societal collapse is no simple matter. Aside from establishing contemporaneity of data, the use of climate as a causal factor has often been rejected as overly deterministic and has met with resistance for the simple reason that many civilizations have successfully “weathered” climatic change (Abate 1994; Butzer 1997). But while some societies have weathered change, climate should not be disregarded altogether, particularly when considering food production. The reliance of a sedentary population upon cultivated plants and herded animals results in a greater dependence upon a more limited area of land for food production, which, owing to the social and economic investments in any given place, increases the likelihood of adapting to adverse climate change before considering it necessary to move. Focusing the debate on whether climate change can lead to the collapse of civilizations diverts attention from the more important question of how different societies react or adapt to climatic change and what aspects of their social structure or geographic location strengthen or weaken their ability to adapt and create a buffer. Through the use of various agricultural practices such as irrigation, terracing, raised-field agriculture, multiple crop varieties, and crop storage, societies across the globe have proven that innovation and modification of the landscape can be used to increase crop yields and effectively buffer against short-term changes in weather patterns (see, e.g., Barker and Jones 1984; Denevan, Mathewson, and Whitten 1985; Farrington and Park 1978). These buffers cannot be considered limitless, though, and they become increasingly limited as the levels of the agricultural constraints increase, for example

58

methodologies in archaeobotany

as they do near the boundary of rain-fed agriculture. Consequently, understanding the factors that affect these buffers is key to understanding how climate affects the decisions made by individuals within a society. It is fairly straightforward to examine how crop growth responds to different climatic conditions, but examining how people react is considerably more complicated and controversial. Boserup ([1965] 1993) argued that the intensification of agriculture in recent years has often been met with resistance by workers and that initial migration of farm laborers to other occupations can lead to temporary social upheaval. Could a similar process have occurred in the past? What mechanisms were in place to distribute food or help alleviate food shortages or labor shortages? With reference to data from the last few centuries, Amartya Sen (1999), a Nobel Peace Prize winner in economics, has argued that no famine has ever occurred within a functioning democracy, so where there is textual evidence for famine in the past, we can ask, what social mechanisms allowed this to happen? More to the point, how can these questions be examined using archaeological data? Where available, texts provide interesting insights into ancient agriculture and issues of environmental or climate change (Archi 1990; Civil 1994; Wiggerman 2000; Zettler 2003). Textual evidence dating to the beginning of the second millennium BC, for example, indicates that the inhabitants of Tell Bi’a along the Balikh River complained to Old Babylonian authorities at Mari about declining irrigation water supplies resulting from upstream extraction, suggesting competition for available water along the river (Villard 1987; Wilkinson 1998). Texts should not be used uncritically, however, and numerous lines of evidence, including well-dated archaeobotanical, zooarchaeological, and paleoclimatic data, as well as knowledge of settlement patterns, are essential. Cropping decisions are not entered into lightly, and farmers’ decisions result from detailed considerations of the likelihood of success of a crop and the amount of maintenance required to enhance success (such as irrigation, pest management, weeding, etc.) against its ultimate value to a society. Within Syria alone, the Ministry of Agriculture and Agrarian Reform divides the land into five agricultural settlement zones based on annual rainfall (Saliba 1997). In areas receiving more than 600 mm per annum, rain-fed crops can be planted annually with success. As the amount of precipitation declines moving south and inland, cropping intensity decreases to one or two cropping seasons every three years, and emphasis shifts from wheat and summer crop production to barley or permanent grazing (FAO 1982). Large parts of the country are considered marginal for cultivation, and in the drier areas animal husbandry forms the most feasible option. A sustained shift in climate that moved the boundary of marginal agriculture would certainly be noticed, and,

Examining Agriculture and Climate Change

59

depending upon the severity, would cause people to respond. The need to respond would likely be felt by farmers first and, depending upon the severity of the flux, may or may not be felt by the nonagrarian part of the population. Furthermore, at any given time, the need for adaptation would vary over space, with those closer to the boundary of rain-fed agriculture being affected differently from those in wetter zones. In this sense, agriculture acts as a sensitive interface between people, their environment, and the prevailing climate, a point that is underscored by the integral role of agriculture in the rise and maintenance of urban life. Generally, in areas where rainfall is low and cannot support rain-fed agriculture, and where adequate water supplies for irrigation are not easily accessible, permanent settlements tend to be sparse unless there is a strong economic or strategic incentive for inhabiting a place (e.g., Evenari, Shanan, and Tadmor 1982; Wilkinson 1997, 2000). Consequently, if rainfall belts are shifting significantly over time, settlement intensity near the lower boundary will change. Within the Levant as a whole, orographic controls affect moisture substantially, but under moister climes the limit of rain-fed agriculture generally expands to the south and southeast, thereby increasing the area of land that could be cultivated under rain-fed conditions. Over time, in the absence of conflict, such an area would likely become settled. Conversely, under drier conditions the boundary would move to the north and northeast, narrowing the area suitable for rain-fed agriculture. In such instances, several alternatives would be possible for people living at the previously established boundary for marginal agriculture: (1) Communities could continue with their established methods of food production. Under prolonged conditions of decreased rainfall, yields would likely diminish and, unless prohibited from doing so for social or political reasons, the local population would begin to migrate to more humid lands once yields fell below desired surplus levels; migration would intensify once yields were insufficient to feed the local population. (2) Farmers could modify the crops grown by diversifying the economic base and including more drought-tolerant crops, or they could “extensify” agricultural production by expanding the area of land utilized. Such “extensification” could include a heavier reliance on pastoralism. (3) Communities could remain in the area and continue growing the same crops by intensifying water conservation, water harvesting strategies, or irrigation. This might be expected in instances where people could not migrate or where settlement was economically or strategically important, as demonstrated by the Nabatean inhabitants of Avdat in the Negev desert who practiced rainwater harvesting in the more arid areas along an important trade route (Evenari, Shanan, and Tadmor 1982). Extensive settlement of dry areas in northern Mesopotamia underscores the importance of trade and potential food imports in

60

methodologies in archaeobotany

sustaining a community (Wilkinson 2000). Such trade would rely heavily upon political stability and economic ties with surrounding regions. A combination of these alternatives is also possible whereby some people remain and adapt their methods of food production and others move to different areas. It is likely that owing to differential access to and control of resources, elites would have different alternatives from other segments of society. Much of the thought relating agriculture to societal collapse in antiquity is rooted in the ideas of the Reverend Thomas Malthus (1970), the well-known author of An Essay on the Principle of Population, first published in 1798, or Ester Boserup (1993), whose publication of The Conditions of Agricultural Growth in 1965 challenged Malthus’s claims and revolutionized ideas of population dynamics within contemporary agricultural development. To some degree their ideas, particularly those of Malthus, have become received wisdom and are rarely acknowledged explicitly. After observing widespread class-based poverty in eighteenthcentury England, Malthus (1970, 71) postulated that “population, when unchecked, increases in a geometrical ratio. Subsistence increases only in an arithmetical ratio.” Because of the inelasticity of food supply, according to Malthus, once population levels increased beyond the carrying capacity of the land, the surplus population would be eliminated, either by direct starvation or by positive checks such as misery and vice or moral restraint. He also presented the possibility of migration on a local level. Boserup, on the other hand, argued that as populations increase, land utilization intensifies and agricultural methods change. She employs the transition from forest fallow, to bush and short fallow, to annual cropping, and finally multicropping within a tropical environment to support her argument (Boserup 1993). While archaeological evidence for a parallel in southwest Asia may be lacking, her basic premise that different technologies and methods of production can increase crop yields is cogent and challenges Malthus’s idea of the inelasticity of food supply. Boserup (1993, 41) does hint at upper limits of production, however, when she states, “It has no doubt happened in many cases that a population, faced with a critically increasing density[,] was without knowledge of any types of fertilization techniques. They might then shorten the period of the fallow without any other change in method. This constellation would typically lead to a decline of crop yields and sometimes to an exhaustion of land resources.” Both Malthus and Boserup directly discuss the ability of agriculture to feed ever-increasing populations, but the same arguments can be applied to considerations of agriculture and climate change, particularly when climate change has a negative impact on agricultural production and disrupts the supply balance. Boserup and Malthus are frequently considered

Examining Agriculture and Climate Change

61

to be diametrically opposed to one another, but when considering longterm changes in agriculture it is possible to incorporate aspects of both models. Methods of food production can continue unchanged as long as yields satisfy consumption or production needs. If yields decline relative to demand, adoption of an alternate technology is feasible where it is viable in terms of available labor, resources, and technology. Consequently, yields can be enhanced, but only within the limits of existing or economically viable technology; an upper yield limit for each method of production would still exist, and yields could only be increased by innovation of a more intensive technique or by “extensifying” the area under production. Under this view of Malthus and Boserup, emphasis on an inherent carrying capacity of the land is reduced and the elasticity of food production within given limits is enhanced. The shift in perspective is important, because it is precisely this elasticity that defines the capacity of different societies to buffer or adapt to climate change. Such a model is useful at a basic level, but it should be stressed that it also has serious limitations due to its restricted focus on production and consumption without any definition of geographic boundary. Clearly there is a desperate need to incorporate economic and political considerations into such models and account for trade networks and exchange with “outside” groups. The formal inclusion of such factors will allow agricultural economies and their relationship to climate change to be examined in much greater detail.

Conclusions As recent research has shown, climate change undoubtedly affects agricultural production. This chapter has discussed ways of considering the effects of climate change on agricultural production in antiquity. The current limitations to achieving this goal include practical issues of data collection and data analysis, as well as more theoretical limitations relating to data interpretation. Production and consumptions models, such as those based on the ideas of Malthus and Boserup, are useful at a basic level, but they fail to account for other mechanisms that modify the supply and demand of food, such as trade, political tension, social status, and migration. These factors are widely recognized as important, and there is an urgent need to formally incorporate them into considerations of food production in the past. There is also a need to consider multiple lines of evidence, including settlement data, economic data, and textual accounts where and when available. Understanding how climate change affected food production in antiquity is an important question that has bearing on a wide range of social

62

methodologies in archaeobotany

issues, including the maintenance of urban life. Before our understanding can be furthered, we need to enhance both the temporal resolution and dating accuracy of both paleoclimatic and archaeological records so that contemporaneity between the two can be established with a greater degree of accuracy. Temporal resolution can be enhanced by more intensive sampling of all time periods, and dating accuracy may be enhanced through the use of wiggle-match dating based on a series of radiocarbon dates. The construction of large databases of published data provides a useful starting point to evaluating regional patterns of crop production, grazing, and animal herding, and allows for more sophisticated models of agriculture to be developed. Incorporating contextual information that details the origin of each sample is important because archaeobotanists typically consider the archaeological context of a sample alongside models of plant use, deposition, and preservation when interpreting the composition of archaeobotanical samples. The examination of a large data set from similar context types across a region using correspondence analysis would allow the validity of these models to be examined more critically and would allow for interpretations of plant data to be based on a firmer foundation. Taking context into account, preliminary examination of archaeobotanical data highlights variations in crop production and weed flora that reflect both environmental differences and localized economies. In order to better understand agriculture as a whole, there is a continuing need to develop methods that allow archaeobotanical and zooarchaeological data to be integrated. While a number of limitations to examining climate change and agriculture in antiquity currently exist, identifying these problems, on practical, methodological, and theoretical levels, is an important step toward developing solutions.

Acknowledgments This chapter was originally prepared for the 2004 Society for American Archaeology Annual Meeting. I am grateful to Marco Madella and Manon Savard for organizing the session, and I am further grateful to Marco Madella, Manon Savard, and Carla Lancelotti for bringing this volume to publication. I offer thanks to Simone Riehl, Daniel S. Adler, and two anonymous reviewers for their careful reading of an earlier draft of this chapter and for their helpful comments. I thank Naomi Miller for bringing William Ruddiman’s paper to my attention. I am also indebted to Sue Colledge, James Conolly, and Stephen Shennan, who generously shared

Examining Agriculture and Climate Change

63

the design of their database with me prior to its publication in Colledge, Conolly, and Shennan (2004, S37). Their database forms part of a larger project entitled “The Origin and Spread of Neolithic Plant Economies in the Near East and Europe.” The development of the regional study discussed here was supported by the National Science Foundation under Grant No. 0210651.

Notes 1. Since this chapter was written in 2004, several online databases of southwest Asian archaeobotanical data have become publicly available. Examples include the Archaeobotanical Database of Eastern Mediterranean and Near Eastern sites (http://www.ademnes.de) maintained by Simone Riehl, and an archaeobotany database maintained by Helmut Kroll (http://www.archaeobotany. de). The Alexandria Archive Institute (http://alexandriaarchive.org) also serves as an important repository for open-access archaeological data. 2. In the past years a number of developments have been made with respect to the integration of archaeobotanical and zooarchaeological data. In 2009 Current Anthropology published a special issue dedicated to this topic (Smith and Miller 2009). Papers were coauthored by zooarchaeologists and archaeobotanists and used data from southwest Asia to explore the benefits and difficulties of considering plant and animal data together. VanDerwarker and Peres (2010) approached similar issues with a more global focus. 3. While many studies have been conducted, the geographical distribution of sites that have yielded archaeobotanical data is limited. Understanding the variation in agronomy across space needs to be better understood before well-founded temporal comparisons can be made at the regional level.

References Abate, T. 1994. “Climate and the Collapse of Civilization.” BioScience 44 (8): 516–19. Al-Azem, A. N. A-M. 1992. “Crop Storage in Ancient Syria: A Functional Analysis Using Ethnographic Modelling.” PhD diss., University College London. Archi, A. 1990. “Agricultural Production in the Ebla Region.” Annales Archéologiques Arabes Syriennes 40:50–55. Barker, G. W. W., and G. D. B. Jones 1984. “The UNESCO Libyan Valleys Survey VI: Investigations of a Romano-Libyan Farm: Part I.” Libyan Studies 15:1–43. Bar-Matthews, M., A. Ayalon, M. Gilmour, A. Matthews, and C. J. Hawkesworth. 2003. “Sea–Land Oxygen Isotopic Relationships from Planktonic Foraminifera

64

methodologies in archaeobotany

and Speleothems in the Eastern Mediterranean Region and Their Implication for Paleorainfall during Interglacial Intervals.” Geochimica et Cosmochimica Acta 67 (17): 3181–99. Bar-Matthews, M., A. Ayalon, A. Kaufman, and G. J. Wasserburg. 1999. “The Eastern Mediterranean Paleoclimate as a Reflection of Regional Events: Soreq Cave, Israel.” Earth and Planetary Science Letters 166:85–95. Barrow, C. 1993. Water Resources and Agricultural Development in the Tropics. Essex: Longman Scientific and Technical. Baruch, U. 1986. “The Late Holocene Vegetational History of Lake Kinneret (Sea of Galilee), Israel.” Paléorient 12 (2): 37–48. Baruch, U., and S. Bottema. 1999. “A New Pollen Diagram from Lake Hula.” In Ancient Lakes: Their Cultural and Biological Diversity, edited by H. Kawanabe, G. W. Coulter, and A. C. Roosevelt, 75–86. Ghent: Kenobi Productions. Boserup, E. (1965) 1993. The Conditions of Agricultural Growth: The Economics of Agrarian Change under Population Pressure. London: Earthscan Publications. Bottema, S. 1997. “Third Millennium Climate in the Near East Based upon Pollen Evidence.” In Third Millennium BC Climate Change and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 489–515. Berlin: Springer-Verlag. Braswell, B. H., D. S. Schimel, E. Linder, and B. Moore III. 1997. “The Response of Global Terrestrial Ecosystems to Interannual Temperature Variability.” Science 178:870–72. Butzer, K. W. 1997. “Sociopolitical Discontinuity in the Near East c. 2200 BCE: Scenarios from Palestine and Egypt.” In Third Millennium BC Climate Change and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 245– 96. Berlin: Springer-Verlag. Charles, M. 1998. “Fodder from Dung: The Recognition and Interpretation of Dung-Derived Plant Material from Archaeological Sites.” Environmental Archaeology 1:111–22. Civil, M. 1994. The Farmer’s Instructions: A Sumerian Agricultural Manual. Aula Orientalis Supplement, 5. Barcelona: Editorial AUSA. Colledge, S. 2003. “Plants and People.” In Excavations at Tell Brak, vol. 4, Exploring an Upper Mesopotamian Regional Centre, 1994–1996, edited by R. Matthews, 389–416. Cambridge: McDonald Institute for Archaeological Research and the British School of Archaeology in Iraq. Colledge, S., J. Conolly, and S. Shennan. 2004. “Archaeobotanical Evidence for the Spread of Farming in the Eastern Mediterranean.” Current Anthropology 45:S35–S58. Davis, P. H. 1975. Flora of Turkey and the East Aegean Islands. Edinburgh: Edinburgh University Press. Denevan, W. M., K. Mathewson, and R. Whitten. 1985. “Mounding, Mucking and Mangling: Resent Research on the Raised Fields in the Guayas Basin,

Examining Agriculture and Climate Change

65

Ecuador.” In Prehistoric Intensive Agriculture in the Tropics, Part 1, edited by I. S. Farrington, 181–83. BAR International Series, 232. Oxford: Archaeopress. Dennell, R. W. 1977. “On the Problems of Studying Prehistoric Climate and Crop Agriculture.” Proceedings of the Prehistoric Society 43:361–69. Evenari, M., L. Shanan, and N. Tadmor. 1982. The Negev: The Challenge of the Desert. Cambridge, MA: Harvard University Press. FAO (Food and Agriculture Organization). 1982. “Regional Study of Rainfed Agriculture and Agro-Climatic Inventory of Eleven Countries in the Near East Region.” World Soil Resources Report. Rome: Food and Agriculture Organization. Farrington, I. S., and C. C. Park. 1978. “Hydraulic Engineering and Irrigation Agriculture in the Moche Valley, Peru.” Journal of Archaeological Science 5:255–68. Gilboa, A., and I. Sharon. 2003. “An Archaeological Contribution to the Early Iron Age Chronological Debate: Alternative Chronologies for Phoenicia and Their Effects on the Levant, Cyprus, and Greece.” Bulletin of the American Schools of Oriental Research 332:7–80. Goodfriend, G. A. 1991. “Holocene Trends in 18O in Land Snail Shells from the Negev Desert and Their Implications for Changes in Rainfall Source Areas.” Quaternary Research 35:417–26. Grayson, D. K. 1984. Quantitative Zooarchaeology: Topics in the Analysis of Archaeological Faunas. Orlando: Academic Press. Hansen, J. M. 1991. “Beyond Subsistence: Behavioural Reconstruction from Palaeoethnobotany.” Archaeological Review from Cambridge 10 (1): 53–59. Hillman, G. C. 1984. “Interpretation of Archaeological Plant Remains: The Application of Ethnographic Models from Turkey.” In Plants and Ancient Man: Studies in Palaeoethnobotany, Proceedings of the Sixth Symposium of the International Work Group for Paleoethnobotany, edited by W. van Zeist and W. A. Casparie, 1–41. Rotterdam: A. A. Balkema. ———. 1991. “Phytosociology and Ancient Weed Floras: Taking Account of Taphonomy and Changes in Cultivation Methods.” In Modelling Ecological Change: Perspectives from Neoecology, Palaeoecology and Environmental Ecology, edited by D. R. Harris and K. D. Thomas, 27–40. London: Institute of Archaeology, University College London. Hole, F. 1991. “Middle Khabur Settlement and Agriculture in the Ninevite 5 Period.” Bulletin of the Canadian Society for Mesopotamian Studies 21:17–29. Hubbard, R. N. L. B. 1976. “Crops and Climate in Prehistoric Europe.” World Archaeology 8:159–68. Ibn Khaldu¯n. 1989. The Muqaddimah: An Introduction to History. Translated by F. Rosenthal. Edited and abridged by N. J. Dawood. Princeton: Princeton University Press. Issawi, C. 1952. “Arab Geography and the Circumnavigation of Africa.” Osiris 10:117–28.

66

methodologies in archaeobotany

Jones, G. E. M. 1984. “Interpretation of Archaeological Plant Remains: Ethnographic Models from Greece.” In Plants and Ancient Man: Studies in Palaeoethnobotany, Proceedings of the Sixth Symposium of the International Work Group for Paleoethnobotany, edited by W. van Zeist and W. A. Casparie, 43–61. Rotterdam: A. A. Balkema. ———. 1991. “Numerical Analysis in Archaeobotany.” In Progress in Old World Palaeoethnobotany: A Retrospective View on the Occasion of 20 Years of the International Work Group for Palaeoethnobotany, edited by W. van Zeist, K. Wasylikowa, and K. E. Behre, 63–80. Rotterdam: A. A. Balkema. ———. 1998a. “Distinguishing Food from Fodder in the Archaeobotanical Record.” Environmental Archaeology 1:95–98. ———. 1998b. “Wheat Grain Identification—Why Bother?” Environmental Archaeology 2:29–34. Jones, G. E. M., and P. Halstead. 1995. “Maslins, Mixtures and Monocrops: On the Interpretation of Archaeobotanical Crop Samples of Heterogeneous Composition.” Journal of Archaeological Science 22:103–14. Kaufman, A., G. J. Wasserburg, D. Porcelli, M. Bar-Matthews, A. Ayalon, and L. Halicz. 1998. “U–Th Isotope Systematics from the Soreq Cave, Israel and Climatic Correlations.” Earth and Planetary Science Letters 156:141–55. Kolodny, Y., M. Bar-Matthews, A. Ayalon, and K. D. Mckeegan. 2003. “A High Spatial Resolution ∂18O Profile of a Speleothem Using an Ion-Microprobe.” Chemical Geology 197:21–28. Kroll, H. 1992. “Einkorn from Feudvar, Vojvodina, II: What Is the Difference between Emmer-like Two-seeded Einkorn and Emmer?” Review of Palaeobotany and Palynology 73:181–85. Kühne, H. 1991. “Die rezente Umwelt von Tall Šeh Hamad und Daten zur Umweltrekonstruktion der assyrischen Stadt Dur-Katlimmu—Die Problemstellung.” In Die rezente Umwelt von Tall Šeh Hamad und Daten zur Umweltrekonstruktion der assyrischen Stadt Dur-Katlimmu, edited by H. Kühne, 21–33. Berlin: Dietrich Reimer. Landmann, G., A. Reimer, G. Lemcke, and S. Kempe. 1996. “Dating Late Glacial Abrupt Climate Changes in the 14,570 yr Long Continuous Varve Record of Lake Van, Turkey.” Palaeogeography, Palaeoclimatology, Palaeoecology 122:107–18. Lines, L. 1995. “Bronze Age Orchard Cultivation and Urbanization in the Jordan River Valley.” PhD diss., Arizona State University, Tempe. Malthus, T. (1798) 1970. An Essay on the Principle of Population. London: Penguin Books. Manning, S. W., B. Kromer, P. I. Kuniholm, and M. W. Newton. 2003. “Confirmation of Near-Absolute Dating of East Mediterranean Bronze-Iron Dendrochronology.” Antiquity 77, Project Gallery. http://antiquity.ac.uk/ProjGall /Manning/manning.html.

Examining Agriculture and Climate Change

67

Matilla Séquer, G., and D. Nivera Núñez. 1993/94. “Estudio paleoetnobotánico de Tell Qara Quzaq-I.” In Tell Qara Quzaq-I: Campañas I–III (1989–1991), edited by G. D. O. Lete, 151–82. Barcelona: Editorial AUSA. McCorriston, J. 1995. “Preliminary Archaeobotanical Analysis in the Middle Habur Valley, Syria, and Studies of Socioeconomic Change in the Early Third Millennium BC.” Canadian Society for Mesopotamian Studies Bulletin 29:33–46. ———. 1998. “Landscape and Human-Environment Interaction in the Middle Habur Drainage from the Neolithic Period to the Bronze Age.” In Espace Naturel, Espace Habité en Syrie du Nord (10e–2e millénaires av. J-C.), Actes du colloque tenu à l’Université Laval (Québec) du 5 au 7 mai 1997, edited by M. Fortin and O. Aurenche, 43–53. Québec: Canadian Society for Mesopotamian Studies; Lyon: Maison de l’Orient Méditerranéen. McCorriston, J., and S. Weisberg. 2002. “Spatial and Temporal Variation in Mesopotamian Agricultural Practices in the Khabur Basin, Syrian Jazira.” Journal of Archaeological Science 29:485–98. Meadow, R. H., and M. A. Zeder. 1978. Approaches to Faunal Analysis in the Middle East. Cambridge, MA: Peabody Museum of Archaeology and Ethnology, Harvard University. Miller, N. F. 1984. “The Interpretation of Some Carbonized Cereal Remains as Remnants of Dung Cake Fuel.” Bulletin on Sumerian Agriculture 1:45–47. ———. 1990. “Clearing Land for Farmland and Fuel: Archaeobotanical Studies of the Ancient Near East.” In Economy and Settlement in the Near East: Analyses of Ancient Sites and Materials, edited by N. F. Miller, 70–78. Philadelphia: MASCA, The University Museum of Archaeology and Anthropology, University of Pennsylvania. ———. 1996. “Seed Eaters of the Ancient Near East: Human or Herbivore?” Current Anthropology 37 (3): 521–28. ———. 1997. “Farming and Herding along the Euphrates: Environmental Constraint and Cultural Choice (Fourth to Second Millennia BC).” In Subsistence and Settlement in a Marginal Environment: Tell es-Swehat, 1989–1995, Preliminary Report, edited by R. L. Zettler, 123–81. Philadelphia: MASCA, University of Pennsylvania. ———. 2004. “Long-Term Vegetation Changes in the Near East.” In The Archaeology of Global Change: The Impact of Humans on Their Environment, edited by C. L. Redman, S. R. James, P. R. Fish, and J. D. Rogers, 130–40. Washington, DC: Smithsonian Books. Miller, N. F., and T. L. Smart. 1984. “Intentional Burning of Dung as Fuel: A Mechanism for the Incorporation of Charred Seeds into the Archeological Record.” Journal of Ethnobiology 4 (1): 15–28. Nesbitt, M., and D. Samuel. 1996. “From Staple Crop to Extinction? The Archaeology and History of the Hulled Wheats.” In Hulled Wheats, Proceedings of the

68

methodologies in archaeobotany

First International Workshop on Hulled Wheats, 21–22 July 1995, Castelvecchio Pascoli, Tuscany, Italy, edited by S. Padulosi, K. Hammer, and J. Heller, 41–100. Rome: International Plant Genetic Resources Institute. Peñuelas, J., and I. Filella. 2001. “Responses to a Warming World.” Science 294:793–95. Pinnock, F. 1990. “Patterns of Trade at Ebla in the Third Millennium B.C.” Annales Archéologiques Arabes Syriennes 40:39–49. Popper, V. S. 1988. “Selecting Quantitative Measurements in Paleoethnobotany.” In Current Paleoethnobotany: Analytical Methods of Cultural Interpretation of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, 53–71. Chicago: University of Chicago Press. Redman, C. L., S. R. James, P. R. Fish, and J. D. Rogers, eds. 2004. The Archaeology of Global Change: The Impact of Humans on Their Environment. Washington, DC: Smithsonian Books. Riehl, S. 2001. “Vorbericht der archäobotanischen Bestandsaufnahme in Emar.” Baghdader Mitteilungen (Berlin) 32:157–72, pl. 1–2. Ruddiman, W. F. 2003. “The Anthropogenic Greenhouse Era Began Thousands of Years Ago.” Climatic Change 61:261–93. Saliba, G. I. 1997. “Managing Resources, Its Socioeconomic and Environmental Consequences.” Journal of Developing Societies 13:158–67. Sen, A. 1999. Development as Freedom. New York: Alfred A. Knopf. Sherratt, A. 1983. “The Secondary Exploitation of Animals in the Old World.” World Archaeology 15 (1): 90–104. Smith, A., and N. F. Miller 2009. “Integrating Plant and Animal Data: Delving Deeper into Subsistence: Introduction to the Special Section.” Current Anthropology 50:883–34. Speranza, A., J. van der Plicht, and B. van Geel. 2000. “Improving the Time Control of the Subboreal/Subatlantic Transition in a Czech Peat Sequence by 14C Wiggle-Matching.” Quaternary Science Reviews 19:1589–604. Stiller, M., A. Ehrlich, U. Pollingher, U. Baruch, and A. Kaufman. 1983/84. “The Late Holocene Sediments of Lake Kinneret (Israel)—Multidisciplinary Study of a 5 Meter Core: Geological Survey of Israel.” Current Research 1983/84:83–88. Van der Veen, M. 1992. Crop Husbandry Regimes: An Archaeobotanical Study of Farming in Northern England, 1000 BC–AD 500. Sheffield Archaeological Monographs, 3. Sheffield: Department of Archaeology and Prehistory, University of Sheffield. VanDerwarker, A. M., and T. M. Peres, eds. 2010. Integrating Zooarchaeology and Paleoethnobotany: A Consideration of Issues, Methods, and Cases. New York: Springer. Van Zeist, W. 1999/2000. “Third to First Millennium BC Plant Cultivation on the Khabur, North-eastern Syria.” Palaeohistoria 41/42:111–25.

Examining Agriculture and Climate Change

69

———. 2003. “The Plant Husbandry of Tell al-Raqa’i.” In Reports on Archaeobotanical Studies in the Old World, edited by W. van Zeist, 7–31. Groningen: no publisher. Van Zeist, W., and J. A. H. Bakker-Heeres. 1985. “Archaeobotanical Studies in the Levant 4: Bronze Age Sites on the North Syrian Euphrates.” Palaeohistoria 27:247–316. Van Zeist, W., W. Waterbolk-Van Rooijen, R. M. Palfenier-Vegter, and G. Jan De Roller. 2003. “Plant Cultivation at Tell Hamman et-Turkman.” In Reports on Archaeobotanical Studies in the Old World, edited by W. van Zeist, 61–114. Groningen: no publisher. Villard, P. 1987. “Un conflit d’autorités à propos des eaux du Balih.” MARI 5:591–96. Vörösmarty, C. J., P. Green, J. Salisbury, and R. B. Lammers. 2000. “Global Water Resources: Vunerability from Climate Change and Populations Growth.” Science 289:284–88. Wagner, G. A. 1998. “Comparability among Recovery Techniques.” In Current Paleoethnobotany: Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, 17–35. Chicago: University of Chicago Press. Wasylikowa, K. 1981. “The Role of Fossil Weeds for the Study of Former Agriculture.” Zeitschrift für Archäologie 15:11–32. Watson, P. J. 1976. “In Pursuit of Prehistoric Subsistence: A Comparative Account of Some Contemporary Flotation Techniques.” Mid-Continental Journal of Archaeology 1 (1): 77–100. Weiss, H. 1997. “Late Third Millennium Abrupt Climate Change and Social Collapse in West Asia and Egypt.” In Third Millennium BC Climate Change and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 711–23. Berlin: Springer-Verlag. Weiss, H., and R. S. Bradley. 2001. “What Drives Societal Collapse?” Science 291:609–10. Wick, L., G. Lemcke, and M. Sturm. 2003. “Evidence of Lateglacial and Holocene Climate Change and Human Impact on Eastern Anatolia: High Resolution Pollen, Charcoal, Isotopic and Geochemical Records from the Laminated Sediments of Lake Van, Turkey.” The Holocene 13 (5): 665–75. Wiggerman, F. A. M. 2000. “Agriculture in the Northern Balikh Valley: The Case of Middle Assyrian Tell Sabi Abyad.” In Rainfall and Agriculture in Northern Mesopotamia: Proceedings of the Third Mos Symposium (Leiden 1999), edited by R. M. Jas, 171–231. Istanbul and Leiden: Nederlands Historisch-Archaeologisch Instituut and Nederlands Instituut vor het Najije Oosten. Wilkinson, T. J. 1997. “Environmental Fluctuations, Agricultural Production and Collapse: A View from Bronze Age Upper Mesopotamia.” In Third Millennium

70

methodologies in archaeobotany

BC Climate Change and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 67–106. Berlin: Springer-Verlag. ———. 1998. “Water and Human Settlement in the Balikh Valley, Syria: Investigations from 1992–1995.” Journal of Field Archaeology 25:63–87. ———. 2000. “Settlement and Land Use in the Zone of Uncertainty in Upper Mesopotamia.” In Rainfall and Agriculture in Northern Mesopotamia: Proceedings of the Third Mos Symposium (Leiden 1999), edited by R. M. Jas, 3–35. Istanbul and Leiden: Nederlands Historisch-Archaeologisch Instituut and Nederlands Instituut vor het Najije Oosten. Zeder, M. A. 1991. Feeding Cities: Specialized Animal Economy in the Ancient Near East. Smithsonian Series in Anthropological Inquiry. Washington, DC: Smithsonian Institution Press. ———. 1995. “Archaeobiology of the Khabur Basin.” Bulletin of the Canadian Society for Mesopotamian Studies 29:21–32. Zettler, R. L. 2003. “Reconstructing the World of Ancient Mesopotamia: Divided Beginnings and Holistic History.” Journal of the Economic and Social History of the Orient 46:3–45.

C h ap t e r 4

Swahili Urban Food Production Archaeobotanical Evidence from Pemba Island, Tanzania S a r a h C . W a l s h aw

Along all pathways to urbanism, food production patterns shift. Communities become more densely settled along different trajectories, but it remains a universal truth that a settlement cannot grow beyond the number of people it can feed (Boserup 1965). Archaeologists have long thought that centralized control of a food surplus was necessary to support non-food-producing individuals within a city, such as bureaucrats, religious leaders, craft specialists, and members of the military (Adams 1966, 1981; Childe 1946, 18; Redman 1978, 216; Trigger 2003, 313). Africa has become an important source of alternative models for urbanism and urbanization, showing the success of cities with diverse social organizations (Fletcher 1998; R. McIntosh 2000; S. McIntosh and R. McIntosh 1993). Moreover, African food production systems followed different trajectories than those documented elsewhere (e.g., Harlan, De Wet, and Stemler 1976; Marshall 1998; Marshall and Hildebrand 2002). Therefore, models of centralized food production developed for hierarchical systems may not apply to those African cities in which rank was based on age or kinship, and where African domesticates likely dominated. Here, I present preliminary archaeobotanical data from two Swahili communities on Pemba Island, Tanzania, undergoing changes in food production during urbanization. Swahili cities arose along the eastern

73

74

case studies in archaeobotany

coast of Africa by AD 1000 during the transition from Bantu agropastoral village lifeways to larger communities. Swahili urbanization represents a process in which leaders emerged in religious circles and through trade relations, and offers an opportunity to study the changes in food production that accompanied emergent social differences in wealth and status. Archaeological investigations on northern Pemba Island in 2002 and 2004 by LaViolette, Fleisher, and Mapunda (2003, 2004) have provided a unique opportunity to investigate Swahili plant food production systems operating between the seventh and the sixteenth centuries AD. In particular, excavations at two spatially close but temporally separate towns, Tumbe (seventh through ninth centuries AD) and Chwaka (eleventh through sixteenth centuries AD), provide an interesting view of changes in plant food production through time in these urban settings. Taken together, the architectural, subsistence, and material culture data from these excavations will provide a more complete picture of the nature of urban life on Pemba, including exciting new data from earthand-thatch households. I collected archaeobotanical samples at Tumbe and Chwaka during the 2002 and 2004 excavations (Walshaw 2003), and have also conducted ethnobotanical collections and interviews to increase our knowledge of the flora of Pemba and the agricultural activities taking place there. This is the first comprehensive study of Swahili plant food remains, and results obtained from twenty-two contexts to date show significant trends concerning Swahili subsistence during urbanization.

Pathways to Urbanism The trait-based approach to urbanism was instrumental to early discussions of the ancient city and its recognition in the archaeological record (Childe 1950; Northam 1975; Sjoberg 1960; Trigger 1972), but it was tailored to the urban and state forms then known from Mesopotamia and Mesoamerica (LaViolette and Fleisher 2005; Smith 2003). Africanist scholars have critiqued this list-based characterization because it fails to recognize many of Africa’s indigenous urban systems (see LaViolette and Fleisher 2005 for a detailed review). Many African cities successfully arose without the hierarchical organization, public monuments, and writing systems traditionally associated with urbanism (Fletcher 1998; R. McIntosh 2000; S. McIntosh and R. McIntosh 1993). Examples of such cities are Jenne-jeno (S. McIntosh and R. McIntosh 1980, 1984, 1993), Benin (Connah 1975), and Great Zimbabwe (Garlake 1973; Huffman 1972; Pikirayi 2001).

Swahili Urban Food Production

75

The momentum in anthropological studies of urban life has shifted toward a more functionalist approach, one that defines a city in terms of its “performance of specialised functions in relation to a broader hinterland” (S. McIntosh and R. McIntosh 1993, 625, paraphrased from Trigger 1972, 577). More recently, scholarship has focused on viewing cities as they were experienced by all their inhabitants, not merely those who possessed the power or wealth to influence urban planning and administration (Smith 2003). Subsistence data are particularly well suited to address these contemporary concerns in urban archaeology, because food remains are common components of household waste deposits, and because specialized food economies were often so critical in trade economies and site provisioning (Hastorf 1993; Zeder 1991). Moreover, food production systems are powerful manifestations of the urban phenomena that create them, as well as the social relations that maintain them, and as such provide a powerful lens through which to view the multiple layers of urban society (Smith 2003; Zeder 2003).

Urban Food Production Specialized and centrally managed food production systems have been documented for many cities and states around the world (Johnson and Earle 2000; Wittfogel 1957; Zeder 1991). In Mesopotamia, cuneiform tablets provide a vast amount of information regarding the administration of grain production and redistribution in that region (Damerow 1996). Figure 4.1a is a schematic diagram of the hypothetical connections among different administrative levels involved in food production within centralized hierarchical systems. In this model, farmers provide grain to the city as tax revenue, and individual laborers are amassed in local production centers to process harvested grains for storage and redistribution. Grains, or grain products, are stored in a central location until packaged (or baked into bread, brewed into beer, etc.) for redistribution as rations, trade goods, or tribute. Note that, in this model, the administration is the hub of all redistributive activities. All goods and services rendered through the exchange of grain are channeled through the administration for redistribution elsewhere. Indeed, even the workers who grind or bake in production centers are likely paid a wage of food products created by their fellow workers. Another important element of the centralized production model is the reduced opportunity for individual expression in the production process; the mechanisms of production demand uniformity and consistency, not idiosyncrasy. This schema is simplified for heuristic purposes, and does not purport to represent the precise mechanisms of,

76

case studies in archaeobotany

Figure 4.1. (A) Model of food product transfer system in centralized food production systems. (B) Model of food product transfer system in householdbased production systems.

for example, an Uruk food production system; rather, it serves to illustrate how multiple tiers of hierarchy sever the connections between the farmer and his/her surplus products. Urban food production systems in sub-Saharan Africa deserve investigation as possible sources of alternative production pathways. CoqueryVidrovitch (1971, 1997) argued that trade was the original “African mode of production” because agricultural surplus was unlikely due to technological constraints and political economic considerations. In this model, trade production operates under low conditions of administrative intervention, only that which pertains to the maintenance and regulation of the city’s marketplace. While Coquery-Vodrovitch’s model precludes the possibility of an agricultural surplus, I suggest instead that under productive environmental conditions and with the use of iron tools, Eastern African farming households could have produced sufficient surplus to permit trade of agricultural products, as indicated by ethnohistoric documents concerning the Swahili (Horton and Middleton 2000, 17). Therefore, at the level of the household or smallhold noted elsewhere for its productive potential and managerial efficiency (e.g., Netting 1993), the “African mode of production” becomes an economy in which the total surplus produced by individual farms supports a market-based trade economy. Such a system need not be centrally controlled to function effectively for all participants, although undoubtedly individual families could have

Swahili Urban Food Production

77

gained inordinate wealth and possibly status. This household production system is schematized in figure 4.1b. Individual households are responsible for the growth, harvest, and processing of crops, with families trading surplus food products in market contexts with a minimum of administrative intervention, perhaps only that necessary for market function. Furthermore, households trade for products or services that benefit them directly, such as pottery, tools, cloth, food, or medicine. This contrasts sharply with the classic Asiatic model of centralized, direct control over collective production (e.g., Wittfogel 1957) depicted in figure 4.1a, where the farmer is likely not provided the same trade opportunities as the administration.

Archaeological Signatures of an “African Mode of Production” Archaeological signatures of the processes depicted in figure 4.1 could be obtained by examining structures, cultural materials, and plant remains in both domestic and communal archaeological contexts (see table 4.1). The centralized food production systems characteristic of hierarchical administrations are best recognized by the presence of centralized food production facilities for communal labor. Alternatively, in a household mode of production, food production activities take place predominantly in or around the house, with household members or relatives serving as the primary labor source. Similarly, in a centralized system, storage facilities would have been large and under the control of the administration, whereas storage facilities in household-based production systems would be located in or adjacent to individual households. Plant-waste products generated by processing before storage would most likely be concentrated in specialized production centers in the case of centralized production, but in the case of household-based production, processing waste would be dispersed among households and middens throughout the community. Moreover, the communal labor involved in centralized production would likely result in cleaner assemblages of stored grain, while household stores could afford to contain more chaff because household labor would be available for daily processing before meal preparation (Stevens 2003). Finally, goods associated with the exchange of food products would also have different patterns of distribution depending upon the level of centralized control. Within a centralized food production system, only the administrative element of society could exchange food products for foreign goods. Conversely, according to the household mode of production, each household that could mobilize surplus food products for market exchange could possess foreign trade items.

78

case studies in archaeobotany

Table 4.1. Archaeological signatures expected for centralized vs. household-based plant food production patterns Centralized

Household-based

Production centers

Few, large specialized facilities

None; production in or near domestic contexts

Labor

Communal

Individuals or families

Storage facilities

Few; large; associated with administrative architecture

Small; numerous; associated with households

Food production waste

Near production centers, in Located in or near most some houses domestic contexts (i.e., middens)

Evidence of reciprocal products (trade)

Largely in administrative or elite contexts, therefore patchily distributed across site

Widespread; in most domestic contexts, but may vary in quantity across site

Can the archaeobotanical signatures of a household production strategy be detected in ancient cities? Sub-Saharan African communities are likely candidates, but their plant food production systems are understudied (but see Boardman 1999, 2000; Phillipson 2000). Having considered possible forms of urban food production, I now turn to a discussion of Swahili urbanism, and review the current state of knowledge concerning Swahili food production practices. Following this, I present the archaeobotanical data available thus far from excavations in northern Pemba, and contextualize these within patterns of Swahili subsistence and trade.

Swahili Urbanization and Food Production The Swahili coast of Eastern Africa occupies a narrow strip of land extending from Somalia to Mozambique, including the offshore islands, the Comoros Islands, and northwestern Madagascar. The archaeological record of this area by the mid-first millennium AD shows patterns of regional integration that persisted for a thousand years: settled communities engaged in farming, fishing, and pastoralism, used similar architectural styles, produced comparable local ceramic types, and had access to similar imported goods (Horton 1987; Wright 1993; Kusimba 1997). By AD 1000, large Swahili cities (e.g., Kilwa, Manda, Shanga) arose along Eastern Africa’s coast, and these functioned as markets, nodes of craft production, and religious centers supporting Islamic conversion (Chittick

Swahili Urban Food Production

79

1974, 1984; Horton 1996; Pouwels 2000). These communities are referred to as “stonetowns,” based on the presence of stone-built structures such as mosques, large multistory dwellings, and, more rarely, town walls. While some have identified these settlements as city-states (Abungu 1998; Kusimba 1997; Sinclair and Håkansson 2000), an alternative view suggests that Swahili communities were managed by kin groups with limited regional influence (Horton and Middleton 2000). Historically, Swahili cities were considered to be entirely non-African phenomena, created and controlled by the Arabic centers with which the Swahili traded (Hull 1976, 120; Kusimba 1999, 51). This theory was originally posited because the rise of large cities along the East African coast coincided with an increase in externally produced luxury goods in Swahili stonetowns (Chittick 1971, 1974). Recent investigations into Swahili language, culture, and society have revealed their distinctly African origins, however (Allen 1993; Nurse and Spear 1985). This has been corroborated by archaeological investigations (e.g., Horton 1996) that show Swahili involvement in diverse trade relationships with inland African pastoral communities and demonstrate that such ties extend back before Swahili involvement in Indian Ocean trade (Kusimba and Kusimba 2000). Thus, Swahili patterns of urbanization must be understood in the context of local precursors to urbanism, including local patterns of crop production and trade that supported participation in the Indian Ocean trade network (Fleisher 2001). This may prove particularly important in determining the institutional mechanisms employed by Swahili merchants to obtain local resources destined for elite consumption or regional trade (Adams 1966, 46). Linguistic and ethnohistoric data from Eastern Africa suggest that crop products were integral to local subsistence and regional trade economies, particularly those involved in the Indian Ocean trade network, such as Swahili communities (Freeman-Grenville 1962, 1975; Nurse and Spear 1985). In particular, Horton and Middleton (2000, 13; Middleton 2004, 71–73) identify Asian rice and spice plants (pepper, cinnamon, clove, nutmeg/mace) as imports brought to the Swahili coast through Indian Ocean trade, in return for exports such as sorghum, pearl millet, coconut products, mangrove species, and a diverse array of animal products. Pemba Island was reportedly an important supplier of agricultural products to Mombasa, a crucial entrepôt and powerful administrative center of the Swahili world (Middleton 1992, 57). Kirkman (1964, 179) notes that the Sultan of Mombasa’s “successor Hasan was finally recognised as ruler of Pemba in exchange for an annual tribute of rice to the Captain of Mombasa,” suggesting that rice surplus could be mobilized for political use. These expectations are only beginning to be tested against the archaeological record, because zooarchaeological and archaeobotanical

80

case studies in archaeobotany

analyses are relatively new to Eastern African archaeology (Ambrose 1984; Hoffman 1984; Horton and Mudida 1993, Robertshaw and Wetterstrom 1989; Wetterstrom 1991). Furthermore, archaeological studies of Swahili urbanism to date have relied on the incomplete picture afforded by focusing on stone architectural features, such as mosques, tombs, multistoried houses, and, at a few sites, walls encircling portions of the community. Only by examining all elements of urban life, including the earth-and-thatch structures that housed the Swahili majority, can we accurately construct a view of Swahili food production patterns (Fleisher and LaViolette 1999a, b; LaViolette, Fleisher, and Mapunda 2003; Walshaw 2003). Domesticated resources appear to be important during Swahili urbanization, but conclusions are tentative due to limited archaeological evidence. Faunal remains at Shanga and Pate by the tenth century AD suggest reliance upon fish, with some remains of goat, cattle, and chicken (Horton and Mudida 1993; Wilson and Omar 1997). Subsistence patterns differ only marginally by the thirteenth century AD: there is a slight increase in reliance upon cattle and chickens in general (Wright 1993), and at Shanga domesticates overtake fish in importance (Horton and Mudida 1993). Important faunal data have emerged from Fleisher’s (2003) survey and shovel test pit investigations on northern Pemba. Chwaka units sampled in that study show a significant amount of shellfish before AD 1300, followed by a decline in shellfish exploitation. Fish represent a significant part of the diet throughout Chwaka’s occupation, particularly the shallow-water taxa common off Pemba’s shore. The mammalian fauna at Chwaka is dominated by cattle, but the presence of chicken, dugong (sea cow), and, to a lesser extent, ovicaprids suggests some dietary breadth. Cereal cultivation was probably important by the tenth century AD, but plant remains have been recovered from only a few sites dated to this time period. At Kilwa, a concentration of sorghum grains (Sorghum bicolor ssp. bicolor L. Moench) was recovered from an ancient house floor, demonstrating probable cultivation and storage of sorghum there (Chittick 1974, 52). Flotation conducted on Dembeni phase (ninth through tenth centuries AD) sediments from the main Comoros island yielded primarily rice, but also several grains of millet (cf. Setaria), four pieces of coconut rind, and possible bean and citrus fruit remains (Hoffman 1984; Wright 1984). While the sorghum concentration at Kilwa shows direct evidence for cereal cultivation, and probably storage, cereal agriculture elsewhere is inferred from the presence of grinding stones at many Swahili archaeological sites. The stone terracing systems indicative of agricultural intensification found elsewhere in Eastern Africa by the fifteenth century AD, including Nyanga and Engaruka (Soper 1997), are absent from known Swahili sites.

Swahili Urban Food Production

81

Swahili Archaeology on Pemba Island Pemba Island is a 1000 km2 coral sea mount located 60 km off the coast of Tanzania, separated from the mainland by the deep Pemba Channel (figure 4.2; Pritchard 1975). Named the “green island” by visiting Arabs, Pemba supports lush vegetation, particularly on its hilly western half, and its high agricultural productivity today suggests a plentiful food supply in the past (Pritchard 1975). Today, Pembans grow rice and cassava in large fields, and use gardens or field edges to grow coconut, banana, taro, maize, sorghum, pearl millet, tomatoes, sweet potatoes, and a variety of fruit trees (Koenders 1992; Williams 1949). While politics of the last 200 years have paired it with Unguja (Zanzibar Island) to the south, historically Pemba was more closely tied with

Figure 4.2.  Map of Pemba Island, showing location of archaeological (Chwaka) and present-day (Wete, Chake Chake) towns. (Adapted from Fleisher 2003, 143.)

82

case studies in archaeobotany

settlements on the mainland, including Mombasa. Both historic accounts and archaeological data provide evidence that Pembans actively participated in the Indian Ocean trade network, and that some individuals accumulated great wealth (Kirkman 1964; LaViolette 2000). Historic accounts document that Pemba supported important centers before the arrival of the Portuguese (Buchanan 1932; Pearce [1920] 1967). Early Arab writings reveal that Pemba had a partially Muslim population in the tenth century, who occupied some of the earliest Muslim towns on the coast (Chittick 1971, 112). Finally, we have evidence that Pemba supported large populations, confirmed by the concentration of both stone works (Garlake 1966; Kirkman 1964) and a large number of recently discovered subsurface archaeological sites (Fleisher 2001, 2003; LaViolette and Fleisher 1995). Finds from excavations at Tumbe and Chwaka in northern Pemba (figure 4.2) are still undergoing analysis, but trends reported to date show interesting changes in urbanism over time (Fleisher 2003; LaViolette and Fleisher 2005; LaViolette, Fleisher, and Mapunda 2003, 2004). Of particular concern to this discussion is the decentralized pattern emerging from data concerning several arenas of craft production, including those involving iron, textiles, and possibly beads (Fleisher 2003).

Tumbe (Seventh through Ninth Centuries AD) Tumbe is among the earliest occupations on Pemba Island and represents a large settlement comprising earth-and-thatch structures occupied from the seventh century through to the end of the ninth century AD (Fleisher 2003; LaViolette, Fleisher, and Mapunda 2004). Tumbe appears to have grown to over 20 ha, and up to possibly 30 ha, during its occupation. Survey data reveal that northern Pemba supported many smaller communities in addition to several substantial town sites at this time (Fleisher 2003). Tumbe deposits excavated in 2002 yielded evidence of an active maritime trade economy, including both consumed and produced goods (LaViolette, Fleisher, and Mapunda 2003). The imported ceramic assemblages are dominated by jar vessel forms (e.g., Sassanian Islamic), which have been implicated in the importation of foodstuffs, possibly date syrup (Horton and Middleton 2000, 77). Tumbe households included finds of imported pottery, concomitant with large amounts of production waste, including a large amount of iron slag and an unparalleled quantity of “bead grinders.” Archaeologists remain uncertain as to the actual function of pottery sherds with grooved lines (termed “bead grinders” because of a possible connection to shell bead production), but more of these items are found at Tumbe than at any other site on the Swahili coast to date (Fleisher 2003; LaViolette, Fleisher, and Mapunda 2004). Upper

Swahili Urban Food Production

83

Tumbe strata (Layer 1, possibly Layer 2) contain evidence for a separate, later, occupation (seventeenth through eighteenth centuries AD) attributed to the Mazrui; while these levels are easily differentiated from the earlier occupation on the basis of ceramics, it is evident that modern coconut root disturbance has caused some mixing between levels. Therefore, for the purposes of this discussion, I present data from below Layer 1 only, in order to represent the earliest occupation sequence, and caution should be exercised in the interpretation of potentially disturbed samples from Layer 2.

Chwaka (Eleventh through Sixteenth Centuries AD) The site of Chwaka demonstrates that a community arose adjacent to Tumbe several generations after the latter’s abandonment. Discussions are still ongoing regarding potential cultural continuities between the residents of Tumbe and Chwaka, and it seems likely that these two communities represent different immigration events from similar Swahili communities on the mainland or nearby islands. Chwaka’s stonetown component appears to attain its maximum size by AD 1300; between AD 1100 and 1500 the number of communities surrounding Chwaka decrease in number, suggesting that Chwaka’s growth stemmed from depopulation of nearby villages (Fleisher 2003). Local pottery assemblages contain significant amounts of decorated wares, including large bowls or platters, possibly indicating an increase in serving ware for public display (Fleisher 2003; LaViolette, Fleisher, and Mapunda 2003). Certain deposits at Chwaka contained less Indian Ocean pottery than their counterparts at Tumbe, which, coupled with a relative lack of stone architecture, led LaViolette, Fleisher, and Mapunda (2003) to infer that these Chwaka deposits may represent commoner rather than elite households. Most deposits yielded some evidence for craft production, but no areas of concentrated production were located, suggesting that households were the loci of production at Chwaka (LaViolette and Fleisher 2005; LaViolette, Fleisher, and Mapunda 2003). These preliminary findings suggest that accumulation of imports may be linked to the production of crafts at the household level. Does plant food production follow this same pattern?

Archaeobotanical Evidence from Excavations on Pemba Island During excavations on Pemba Island in 2002 and 2004 I collected archaeobotanical samples through bucket flotation. Sites were systematically sampled for plant remains by taking 10–20 L of sediment from every

84

case studies in archaeobotany

layer excavated, and by taking samples of various sizes from feature contexts. Abundant charred macroremains from Tumbe and Chwaka provide evidence of changes in subsistence patterns over time at these spatially related, but likely not culturally continuous, Swahili settlements. The analysis of the study samples is still ongoing, but some interesting trends are discernible from the results to date (see table 4.2).

Tumbe (Seventh through Ninth Centuries AD), N = 13 Contexts All samples from Tumbe were dominated by wood charcoal (species undetermined), and mean wood densities ranged between 0.10g/1 and 0.18g/1 (densities being measured as weight of material divided by volume of sediment processed for flotation). Coconut shell fragments (figure 4.3d), included in the tabulation if they were 2 mm in size or larger, were present in all but two samples and ranged between 0.01g/1 and 0.06g/1 in density. Interestingly, coconut is absent from Tumbe’s lowest levels, and is only sparsely present in the majority of samples. Several pieces of an undetermined nut (possibly a member of the family Arecaceae) were recovered, and future work is planned to aid in the identification of this element of the Pemban archaeobotanical record. African cereals, particularly pearl millet grains and spikelet bases, were common in the Tumbe samples. Sorghum (Sorghum bicolor ssp. bicolor L. Moench) was represented only by two poorly preserved grains, which is interesting considering that sorghum is the sole taxon found at Kilwa, in one of the only archaeobotanical finds previously reported from a Swahili site (Chittick 1974). Sorghum grains from Tumbe appear only in the lowest two layers; its absence from the top layers suggests that sorghum may have been abandoned with the introduction of rice. Pearl millet (Pennisetum glaucum L. R.Br.) grains, in contrast, were found in more than 50 percent of the samples (figures 4.3a, b). One deposit yielded a concentration of pearl millet grains (122 whole grains and 89 probable grains in 3 L of processed sediment) and possibly represents a domestic storage feature. This same feature contained the only grain of finger millet (Eleusine coracana ssp. coracana L. Gaertn.) found at the site (see figure 4.3c). Pearl millet spikelet bases were found in six of the thirteen samples, predominantly from layers also containing pearl millet grains. Spikelet bases form the attachment of the grain to the inflorescence and would have been separated from the grain during either pre- or poststorage processing of a pearl millet spike (Reddy 1997). The presence of charred pearl millet spikelet bases in Tumbe household deposits is best explained by incidental charring of processing waste after disposal, either from sweeping the floor debris into the fire or from depositing the chaff into a midden (but see Fuller, Stevens, and McClatchie, this volume, for a more detailed discussion of archaeobotanical taphonomy).

TUMBE 7th–9th C AD

6

5

13

Middle

Early

Tumbe total

2

Late

9

Chwaka total

0.15

0.18

0.16

0.10

0.65

0.94

0.68

4

2

0.34

3

Early

CHWAKA Late AD 1100–1500 Middle

Mean wood No. density samples (g/L)

0.02

0.02

0.01

0.03

0.14

0.14

0.16

0.11

Mean coconut density (g/L)

0.03

0.03

0.02

0.05

0.65

1.00

0.65

0.30

Mean total nut density (g/L)

Table 4.2. Archaeobotanical data from Tumbe and Chwaka

60

83

50

0

0

0

Pearl millet ubiquity (%)

60

83

50

0

0

30

Total African grain ubiquity (%)

0

30

100

100

100

100

Rice ubiquity (%)

20

17

0

0

25

100

20

30

100

100

100

67

Fabaceae Cotton ubiquity ubiquity (%) (%)

86

case studies in archaeobotany

Figure 4.3. Digital photographs and scanning electron micrographs of archaeobotanical finds from Tumbe and Chwaka: (A) pearl millet grain, (B) pearl millet spikelet base, (C) finger millet grain, (D) coconut endocarp fragment, (E) rice grain, and (F) cotton seed fragment.

In contrast with the assemblage of African grains recovered, Asian rice (Oryza sativa L.) is absent from the lowest strata and is present only in upper layers, similar to the pattern observed for coconut. Rice grains (figure 4.3e) are present in four (31 percent) of the samples, while rice spikelet bases are found in three (38 percent) of the samples. All of the rice fragments were found in only the upper levels of Layer 3 or in Layer 2, suggesting either a late introduction of rice during the occupation of Tumbe or contamination from overlying Mazrui (seventeenth through nineteenth century) levels. Direct dating of the Tumbe rice specimens is underway to resolve this chronology problem. The notable presence of cotton (Gossypium sp., figure 4.3f) at Tumbe indicates that it was probably an important textile or oil crop plant whose production patterns may be instructive in considering the ancient Swahili economy. Cotton fragments were found in all layers except the lowest (Layer 4) in Tumbe household deposits and are most abundant in the upper layers, following the distributional pattern of coconut and rice macrobotanical remains at Tumbe. To summarize, the archaeobotanical remains from Tumbe are relatively low in density, but show evidence of African grains such as pearl millet and, to a much lesser extent, sorghum and finger millet. Rice, coconut, and cotton are also present by the ninth century AD at Tumbe,

Swahili Urban Food Production

87

albeit only in the upper levels, but if their presence were to be confirmed by direct AMS radiocarbon dating, this chronology would be consistent with Hoffman’s (1984) archaeobotanical findings from a ninth through tenth century AD site on the Comoros Islands.

Chwaka (Eleventh through Sixteenth Centuries AD), N = 9 Contexts Chwaka samples are quite rich in archaeobotanical remains, in contrast to the low frequency of charred plant remains in samples from Tumbe. Unidentified wood charcoal was the most frequent element among the nine Chwaka samples, but unlike at Tumbe, coconut and other unidentified nut fragments dominate the nonwood categories of plant remains. Unidentified parenchymal tissue fragments are also found in all samples from Chwaka. There are a number of taxa that could be represented by such fragments, including yam (Dioscorea spp.), taro (Colacasia esculenta L. Schott), and cassava (Manihot esculenta Crantz), although the latter is unlikely because it was introduced to Eastern Africa much later than the time periods represented by Chwaka. Attempts to distinguish yam from taro and other candidates are planned for future investigations. Only one possible grain of pearl millet was discovered in all 151 L of the Chwaka sediments processed for analysis, and no sorghum or finger millet grains were found. In contrast, rice is ubiquitous in all time periods, with grains present in seven samples and spikelet bases found in all nine samples analyzed to date. Cotton was found in all of the Chwaka samples save one, and one feature sample (F-192) in particular yielded an impressive quantity of cotton (6.43g) found in conjunction with a huge number of rice spikelet bases (n = 352). The number of cottonseeds was estimated to be 404; this seed number estimate (SNE) was produced by dividing the total weight of fragments by the average weight per intact charred seed, providing the most conservative value. This feature was found in the lowest layer at Chwaka, and illustrates the significance of rice and cotton in the economy at the founding of the settlement. Bean (Fabaceae) fragments were present in all time periods, but in small quantities. Mung bean (Vigna radiata [L.] R. Wilczek) is present in an upper layer, but is represented by only one, albeit well-preserved, specimen.

Discussion Taxonomic Representation and Differential Deposition Interesting and significant patterns emerge from a comparison of macrobotanical trends at Tumbe and Chwaka based on analyses thus far. While

88

case studies in archaeobotany

the African grain crops of sorghum, finger millet, and pearl millet are present in archaeobotanical samples from Tumbe, they are represented by only one, possible, grain at Chwaka. Moreover, African grains are the only grain species present in samples from the lowest layer at Tumbe. Pearl millet remains were the main component of the sole grain concentration at Tumbe, which could be considered a storage feature given the very little wood or coconut present in this context. Rice, coconut, and cotton are found at Tumbe only above Layer 4, and appear in increasing frequency over time. At Chwaka, rice grains and chaff are the dominant grain representatives, and these, along with cotton, coconut, and other nut taxa, demonstrate an active plant production economy. While the earliest rice at Tumbe awaits direct dating, the presence of rice in Layer 3 indicates an introduction by approximately the ninth century AD. Chwaka samples were surprisingly uniform across time and space, yielding an abundance of Asian crops (rice, coconut) and cotton. Thus, by the earliest occupation of Chwaka (eleventh century AD), indigenous African crops have been largely abandoned in favor of Asian ones, and grain production is complemented by fruit, nut, and cotton production. While this apparent lack of African grains at Chwaka could be the result of a sampling bias or differential charring opportunities among staple grains, Chwaka middens are incredibly rich in archaeobotanical remains, including the preservation of incompletely charred specimens. Therefore, the absence of African millets in these contexts means that even if they were in fact being grown by Chwaka residents, the processing of African grain crops took place away from domestic contexts. A further bias against the preservation of African grain crops results from the Swahili culinary preference for grinding these grains into meal for cooking porridges (uji or ugali). In this dish, grains are partially ground and added to water to form a slurry of variable consistency, a preparation method that offers little opportunity for grains to become charred. Interestingly, however, these biases were apparently not operating at Tumbe, thus allowing preservation of African grains under the less favorable depositional circumstances existing there. Further sampling at Chwaka, coupled with ethnoarchaeological investigations of rice, sorghum, and pearl millet processing, are planned to provide a more complete picture of Chwaka plant food production practices.

Production Patterns There are several possible explanations for the archaeobotanical patterns of production emerging from Tumbe and Chwaka. Considering the apparent absence of centralized processing facilities, it could be argued that staple food items were imported, thus negating the need for large

Swahili Urban Food Production

89

facilities in which to process crops grown locally. However, this is unlikely for several of the taxa found in the Pemban archaeobotanical record to date. These issues are best addressed through an examination of each taxonomic group. Coconut. The low frequency of coconut fragments at Tumbe could potentially be explained by provisioning through long-distance trade of only whole, mature coconuts, either from Southeast Asia or from plantations elsewhere in the Swahili world. Immature coconuts (dafu in KiSwahili) are valued for their “milk,” and dafu is a common element of the Swahili diet today, but coconut in this form remains edible for only a few days, thus making it an unlikely candidate for long-distance transport. Furthermore, the coconut shell at this stage is highly fatty and combusts upon burning, thus largely removing this form of coconut from the archaeobotanical record. Edible coconut is most easily transported long-distance as copra, dried coconut “meat” that may contain up to 70 percent oil (Acland 1971, 54–55), but the presence of coconut shell at Tumbe and Chwaka makes it unlikely that coconut was being imported in the form of copra. Indeed, the high frequency of coconut endocarp at Chwaka and in the upper layers of Tumbe instead reflects probable production of copra locally for trade purposes as well as household use. Moreover, mature, dried coconut shell can be used as a fuel source, which may help explain the high frequency of charred coconut endocarp in Pemba’s archaeobotanical record. Indigenous African Grains. With respect to cereal crops, pearl millet spikelet bases were found in significant quantities in the possible storage feature at Tumbe, but also in most of the other contexts containing pearl millet grains. Moreover, no specialized extrahousehold plant food production facilities were located. Taken together, these data suggest that pearl millet processing was a widespread activity at Tumbe, likely engaged in by many households within the community. Whether the pearl millet plants were grown locally or imported from mainland communities, however, is not discernible from the present data. Pearl millet grains are easily transported while still on the head because the dense spike is a very efficient unit of transport and exchange, and the removal of grains from the head after transport would yield spikelet bases and other residual chaff elements (Reddy 1997). Ethnohistoric sources describe pearl millet and sorghum being grown on the mainland for trade purposes (Middleton 2004), and much of the sorghum available today in markets on Pemba is apparently shipped from the mainland, although many Pemban farmers still grow small quantities of these grains. Thus, while the Tumbe archaeobotanical assemblage shows evidence for household processing of pearl millet, it remains to be determined whether pearl millet was cultivated on Pemba or imported from elsewhere. Furthermore, if pearl millet

90

case studies in archaeobotany

was indeed being imported, questions remain as to how so many households were able to gain access to this likely staple food. Asian Rice. Like pearl millet, rice grains are found in tandem with rice spikelet bases, even in contexts thought to be associated with an elite dwelling, showing a continued pattern of widespread household grain processing. No specialized area of plant food processing was discovered during the 2002 or 2004 excavations at Chwaka, suggesting that the household-based processing observed at Tumbe also operated at Chwaka. Unlike pearl millet, however, rice does not bear grains in a compact head, and thus is most efficiently stored and transported after threshing and the removal of at least some of the bulky chaff elements. Thus, the high frequency of rice spikelets in domestic contexts at Chwaka is perhaps best explained by local rice cultivation followed by storage of rice in its husk. While imported rice is a foundation of the Swahili diet along the coast today, the landscape of Pemba currently offers some of the best conditions for rice cultivation in Tanzania (Middleton 2004, 72). Rice is commonly grown in the lowland depressions found throughout western Pemba, and rice processing patterns observed in Pemban households today provide a model for understanding early Swahili rice production. Rice is cut from the stalk by hand using a small metal knife, then transported to the farmer’s house (or occasionally a field house) for threshing, winnowing, and storage. After drying, virtually all stages of processing take place directly beside the house prior to cooking, and chaff elements are offered as chicken feed or thrown aside, later to be swept into a midden. This results in the storage of whole rice spikelets under a bed in the house, and provides an opportunity for chaff elements to be swept into a midden and burned or buried. This form of disposal increases the chance for rice chaff to be preserved in the archaeological record. I plan to further study rice processing on Pemba to expand this ethnoarchaeological model, including investigation of the role of the field house in crop production (Posnansky 1984).

Archaeobotanical Contributions to Understanding Trade and Subsistence Patterns in Northern Pemba The preliminary results from archaeobotanical investigations at Tumbe and Chwaka suggest that the plant production, for food and likely for fiber, was dispersed throughout the community and not concentrated into specialized production areas. A shift from African cultigens to Asian ones is probably indicative of several processes, including a specialization of food production and an engagement with Asian culture and materials through the maritime trade economy. Settlement survey data indicate

Swahili Urban Food Production

91

that many village communities were present in northern Pemba at this time (Fleisher 2003), and based on the archaeobotanical data presented here, it remains a possibility that Tumbe was importing pearl millet heads from farming communities in the countryside. Increased food production, through importation or expansion of cultivated lands, would have allowed an increase in population, including the seasonal accommodation of seafaring merchants and their crews and possible provisioning of their return voyage. Grain was likely the major dietary staple among travelers, and familiar foods such as rice and coconut may have been the preferred fare. If the archaeobotanical remains suggest that individual households were producing their own grain, and possibly using surplus grain to trade for market goods, then it follows that trade goods should be found in domestic contexts throughout the sites. Based on the preliminary analyses of the 2002 finds, this is indeed the case at Tumbe and, to a lesser extent, at Chwaka (Fleisher 2003; LaViolette, Fleisher, and Mapunda 2003). The frequency of imported pottery at Tumbe is consistent with that at Shanga, a large Swahili trading center, in the same time period (Fleisher 2003; Horton 1996). Evidence from material production at Tumbe shows a considerable amount of household production with no clearly centralized production spaces (Fleisher 2003), a pattern similar to that observed in food production macroremains. A high density of both “bead grinders” and iron slag suggests intense craft production at Tumbe; moreover, the production waste is widely distributed. Iron slag, in particular, is distributed across the site, providing supporting evidence for Mapunda’s (2002) view that iron production was carried out in small-scale furnaces at the level of the household. At Chwaka, however, a somewhat different pattern seems to be emerging. Settlement pattern data indicate that sites such as Chwaka grew because surrounding village populations moved into this urban center, suggesting that important religious, economic, or political opportunities were available at Chwaka (Fleisher 2003). Discrete concentrations of imported artifacts may signal the presence of more wealthy households at Chwaka, particularly in stone-built houses, but imports are present in small quantities throughout the site, as are evidence of production activities such as ironworking, similar to the pattern of rice production found in the archaeobotanical samples from Chwaka. Taken together, this pattern of continued craft production but differentiated consumption of imported pottery suggests several possibilities. First, household-produced goods may have been traded for goods other than imported pottery, such as cloth, which are less likely to be preserved in archaeological contexts. Textiles were reportedly an important trade item (Horton and Middleton 2000), but are virtually invisible archaeologically, particularly the

92

case studies in archaeobotany

rectangular cuts of cloth used as skirts and shawls by Swahili women today (kanga), which lack buttons or other fasteners. The problem of textile importation is further complicated by the presence of spindle whorls and cotton at Chwaka, presumably employed in local fiber production. Second, as population increased at Chwaka between AD 1100 and 1300, proximal arable land may have become scarcer, and unequal access to farming land, and consequently grain crop surplus, may have resulted. Future investigations aim to clarify the processes of wealth differentiation at Chwaka over time, and further archaeobotanical analysis of samples from various areas of the site will contribute to our understanding of Chwaka’s production industries.

Conclusions In conclusion, this first step toward understanding ancient Swahili food production indicates that grain crops were processed at the household level. While the introduction and adoption of rice and coconut were clearly related to the Indian Ocean trade economy, Asian influence apparently did not extend to the management of production, and African farmers appear to have maintained some elements of the food production system that operated before the adoption of rice cultivation. These preliminary results provide only tentative conclusions, but clearly demonstrate the need for continued and expanded research at this frontier of African urban food production. First, these results show that, given the appropriate depositional environments, identifiable charred plant remains can be recovered from tropical African archaeological sites. Swahili middens are extraordinary in their rapid accumulation of household debris, and several protected, burned contexts, such as Feature 192, afford a wealth of archaeobotanical data at Chwaka. As archaeobotanists continue to engage in discussions of how accurately archaeobotanical assemblages represent their original depositional circumstances (see Fuller, Stevens, and McClatchie, this volume), archaeobotanical finds and ethnoarchaeological models from diverse cultural and geographic contexts are needed to inform this debate. Second, this study highlights the need to consider archaeobotanical data concerning Swahili plant food production alongside long-held assumptions derived from ethnohistoric and linguistic data. For example, the historic data on whether Pemba produced or imported rice are conflicting, but the archaeobotanical data presented here strongly suggest that rice was cultivated on Pemba as early as the ninth century AD. Additionally, cotton cloth has been cited as a symbolically potent import item (Horton and Middleton 2000, 111–12), but impressive finds of cotton

Swahili Urban Food Production

93

seeds at Chwaka, coupled with the presence of spindle whorls, demand that we consider local cotton production alongside externally produced cotton textiles. As LaViolette (2004, 126) has recently suggested in her consideration of the roles of written and oral data in Swahili archaeology, “Perhaps the most resonant voice can be neither strictly historical or archaeological: a voice that can originate in the tension between sources that do not reinforce each other.” Applying this to the Pemban archaeobotanical record, local cotton textile products may have been undervalued and therefore underreported by early visitors to the Swahili coast. Also, the importation of rice reported for mainland Swahili communities does not seem to apply to Pemban rice cultivation, as indicated by the patterns of charred rice grains and chaff at Tumbe and Chwaka. This suggests that Swahili food production practices may have been more variable along the coast than previously described. Swahili food production systems provide an interesting case study due to their close ties with Asian trade networks, but the persistence of household-based plant processing throughout the transition to a ricebased food economy suggests that the household was the dominant unit of production on Pemba. This persistence of plant production patterns during urbanization is significant, because it can be used in conjunction with emerging data from other arenas of Swahili production, such as ironworking, pottery making, and animal food production, to construct a more complete understanding of Swahili local production patterns and their interface with the Indian Ocean exchange network. The archaeobotanical data presented here offer a perspective on Swahili food production that encourages a greater understanding of the local African precursors to urbanization. Whether the archaeobotanical patterns presented here are unique to Pemba or are observable throughout the Swahili world remains to be determined. This interim report is offered as a first step in the construction of ancient Swahili plant use in order to invite a critical evaluation of our current understanding of Swahili plant production systems.

Acknowledgments This work was made possible through permission from the Zanzibar Department of Archives, Museums, and Antiquities, with the assistance of Mr. Hamad Omar, Dr. Abdurahman Juma, and Mr. Salim Seif. I am indebted to the people of Tumbe village on Pemba Island, who worked for us and tolerated our bizarre ways during four months of fieldwork in 2002 and 2004. I am especially grateful to Hajj Mohammed Hajj for helping me collect ethnobotanical data, and to the many Pembans who

94

case studies in archaeobotany

patiently answered my questions and allowed my participation in farming and cooking activities. The generous support of the directors and crew of the Pemba Archaeological Project has been crucial in obtaining archaeobotanical samples for this study. In particular, I thank Adria LaViolette and Jeff Fleisher for invaluable logistical support and for helping shape my ideas through invaluable mentoring in Swahili studies. Scanning electron micrographs were taken using facilities at the Institute of Archaeology, University College London (UCL). I am grateful to Dorian Fuller (UCL) and Roy Gereau (Missouri Botanical Garden) for sharing their botanical expertise. Funding was provided by the National Science Foundation (Dissertation Improvement Grant #0431137; Standard Grant to LaViolette #0138319) and Washington University in St. Louis. This paper benefited enormously from comments by Adria LaViolette, Jeff Fleisher, and two anonymous reviewers; additionally, I am grateful to numerous proofreaders, including Jenna Hamlin, Jim Johnson, Dawn Kaufmann, Joshua Thorp, and Jill Walshaw. All remaining errors or omissions are entirely my own.

References Abungu, G. 1998. “City States of the East African Coast and Their Maritime Contacts.” In Transformations in Africa: Essays on Africa’s Later Past, edited by G. Connah, 204–18. London and Washington: Leicester University. Acland, J. D. 1971. East African Crops: An Introduction to the Production of Field and Plantation Crops in Kenya, Tanzania and Uganda. London: Longman (FAO). Adams, R. McC. 1966. The Evolution of Urban Society: Early Mesopotamia and Prehispanic Mexico. Chicago: Aldine. ———. 1981. Heartland of Cities: Surveys of Ancient Settlement and Land Use on the Central Floodplain of the Euphrates. Chicago: University of Chicago Press. Allen, J. deV. 1993. Swahili Origins. Athens: Ohio University Press. Ambrose, S. H. 1984. “Excavations at Deloraine, Rongai, 1978.” Azania 19:79–104. Boardman, S. 1999. “The Agricultural Foundation of the Aksumite Empire: An Interim Report.” In The Exploitation of Ancient Plant Resources in Ancient Africa, edited by M. van der Veen, 137–47. New York: Kluwer Academic/Plenum. ———. 2000. “Archaeobotany.” In Archaeology at Aksum, Ethiopia, 1993–97, edited by D. W. Phillipson, 127, 156, 217, 363–68, 412–14. 2 vols. London: British Institute in East Africa and the Society of Antiquaries of London. Boserup, E. 1965. The Conditions of Agricultural Growth: The Economics of Agrarian Change under Population Pressure. Chicago: Aldine.

Swahili Urban Food Production

95

Buchanan, L. A. C. 1932. The Ancient Monuments of Pemba. Zanzibar: Zanzibar Museum. Childe, V. G. 1946. What Happened in History. Harmondsworth, UK: Penguin. ———. 1950. “The Urban Revolution.” Town Planning Review 21:2–17. Chittik, N. 1971. “The Coast of East Africa.” In The African Iron Age, edited by P. L. Shinnie, 108–41. Oxford: Clarendon Press. ———. 1974. Kilwa: An Islamic Trading City on the East African Coast. Vol. 1, History and Archaeology. British Institute in East Africa Memoir 5. Nairobi: British Institute in East Africa. ———. 1984. Manda: Excavations at an Island Port on the Kenya Coast. British Institute in East Africa Memoir 9. Nairobi: British Institute in Eastern Africa. Connah, G. 1975. The Archaeology of Benin. Oxford: Clarendon Press. Coquery-Vidrovitch, C. 1971. Research into an African Mode of Production. Dakar: United Nations African Institute for Economic Development and Planning. ———. 1997. “Research on an African Mode of Production.” In Perspectives on Africa: A Reader in Culture, History, and Representation, edited by R. R. Grinker and C. B. Steiner, 129–41. Cambridge, MA: Blackwell. Damerow, P. 1996. “Production and Social Status as Documented in Proto-­ Cuneiform Texts.” In Food and the Status Quest: An Interdisciplinary Perspective, edited by P. Weissner and W. Schiefenhövel, 149–69. Oxford: Berghahn Books. Fleisher, J. 2001. “Archaeological Survey and Excavations in Northern Pemba Island, Tanzania, 1999–2000.” Nyame Akuma 56:36–43. ———. 2003. “Viewing Stonetowns from the Countryside: An Archaeological Approach to Swahili Regional Systems, AD 800–1500.” PhD diss., University of Virginia. Fleisher, J., and A. LaViolette. 1999a. “Elusive Wattle-and-Daub: Finding the Hidden Majority in the Archaeology of the Swahili.” Azania 34:87–108. ———. 1999b. “The Recovery of Swahili Settlements in the Absence of Stone Architecture: Two Preliminary Surveys from Pemba Island, Tanzania.” Nyame Akuma 52:64–78. Fletcher, R. 1998. “African Urbanism: Scale, Mobility and Transformations.” In Transformations in Africa: Essays on Africa’s Later Past, edited by G. Connah, 104–38. London and Washington: Leicester University. Freeman-Grenville, G. S. P. 1962. The Medieval History of the Coast of Tanganyika. London: Oxford University Press. ———. 1975. The East African Coast—Selected Documents from the First to the Earlier 19th Century. 2nd ed. London: Rex Collings. Garlake, P. S. 1966. The Early Islamic Architecture of the East African Coast. London and Nairobi: Oxford University Press. ———. 1973. Great Zimbabwe. London: Thames and Hudson.

96

case studies in archaeobotany

Harlan, J. R., J. M. J. de Wet, and A. B. L. Stemler. 1976. “Plant Domestication and Indigenous African Agriculture.” In Origins of African Plant Domestication, edited by J. R. Harlan, J. M. J. de Wet, and A. B. L. Stemler, 3–19. The Hague: Mouton. Hastorf, C. 1993. Agriculture and the Onset of Political Inequality before the Inka. Cambridge: Cambridge University Press. Hoffman, E. S. 1984. “Dembeni Phase Botanical Remains.” Azania 19:48–49. Horton, M. C. 1987. “Early Muslim Trading Settlements on the East African Coast: New Evidence from Shanga.” Antiquiaries Journal 67:290–323. ———. 1996. Shanga: A Muslim Trading Community on the East African Coast. Nairobi: British Institute in Eastern Africa. Horton, M. C., and J. Middleton. 2000. The Swahili: The Social Landscape of a Mercantile Society. Oxford: Blackwell. Horton, M. C., and N. Mudida. 1993. “Exploitation of Marine Resources: Evidence for the Origin of Swahili Communities of East Africa.” In The Archaeology of Africa: Food, Metal and Towns, edited by T. Shaw, P. J. J. Sinclair, B. Andah, and A. Okpoko, 673–93. London: Routledge. Huffman, T. N. 1972. “The Rise and Fall of Zimbabwe.” Journal of African History 13:353–66. Hull, R. W. 1976. African Cities and Towns before the European Conquest. New York: W. W. Norton. Johnson, A., and T. Earle. 2000. The Evolution of Human Societies. 2nd ed. Stanford: Stanford University Press. Kirkman, J. S. 1964. Men and Monuments on the East African Coast. New York: Frederick A. Praeger. Koenders, L. 1992. Flora of Pemba Island: A Checklist of Plant Species. Wildlife Conservation Society of Tanzania, Publication 2. Kusimba, C. M. 1997. “Swahili and the Coastal City-States.” In Encyclopedia of Precolonial Africa, edited by J. O. Vogel, 507–12. Walnut Creek, CA: Altamira. ———. 1999. The Rise and Fall of Swahili States. London: Altamira. Kusimba, C. M., and S. B. Kusimba. 2000. “Hinterlands and Cities: Archaeological Investigations of Economy and Trade in Tsavo, Southeastern Kenya.” Nyame Akuma 54:13–24. LaViolette, A. 2000. “Swahili Archaeology on Pemba Island, Tanzania: Pujini, Bandari ya Faraji, and Chwaka 1997–1998.” Nyame Akuma 53:50–61. ———. 2004. “Swahili Archaeology and History on Pemba, Tanzania: A Critique and Case Study of the Use of Written and Oral Sources in Archaeology.” In African Historical Archaeologies, edited by A. M. Reid and P. J. Lane, 125–62. New York: Kluwer Academic/Plenum. LaViolette, A., and J. Fleisher. 1995. “Reconnaissance of Sites Bearing Triangular Incised (Tana Tradition) Ware on Pemba Island, Tanzania.” Nyame Akuma 56:36–43.

Swahili Urban Food Production

97

———. 2005. “The Archaeology of Sub-Saharan African Urbanism: Cities and Their Countrysides.” In African Archaeology: A Critical Introduction, edited by A. B. Stahl, 327–52. London: Blackwell. LaViolette, A., J. Fleisher, and B. Mapunda, eds. 2003. Pemba Archaeological Project Preliminary Report: First Season, June–August 2002. Unpublished report submitted to the Department of Archives, Museums and Antiquities, Ministry of Education and Culture, Zanzibar, Tanzania. ———, eds. 2004. Pemba Archaeological Project Preliminary Report: Second Season, June–July 2004. Unpublished report submitted to the Department of Archives, Museums and Antiquities, Ministry of Education and Culture, Zanzibar, Tanzania. Mapunda, B. 2002. “Iron Metallurgy along the Tanzanian Coast.” In Southern Africa and the Swahili World, edited by F. Chami and G. Pwiti, 76–88. Dar es Salaam: University Press. Marshall, F. 1998. “Early Food Production in Africa.” Review of Archaeology 19 (2): 47–58. Marshall, F., and E. Hildebrand. 2002. “Cattle before Crops: The Beginnings of Food Production in Africa.” Journal of World Prehistory 16 (2): 99–143. McIntosh, R. J. 2000. “Clustered Cities of the Middle Niger: Alternative Routes to Authority in Prehistory.” In Africa’s Urban Past, edited by D. M. Anderson and R. Rathbone, 19–35. Oxford: James Currey. McIntosh, S. K., and R. J. McIntosh. 1980. “Jenne-jeno, an Ancient African City.” Archaeology 33:8–14. ———. 1984. “The Early City in West Africa: Towards an Understanding.” African Archaeological Review 2:73–98. ———. 1993. “Cities without Citadels: Understanding Urban Origins along the Middle Niger.” In The Archaeology of Africa: Food, Metal and Towns, edited by T. Shaw, P. J. J. Sinclair, B. Andah, and A. Okpoko, 622–41. London: Routledge. Middleton, J. 1992. The World of the Swahili: An African Mercantile Civilization. New Haven: Yale University Press. ———. 2004. African Merchants of the Indian Ocean: Swahili of the East African Coast. Long Grove, IL: Waveland. Netting, R. McC. 1993. Smallholders, Householders: Farm Families and the Ecology of Intensive, Sustainable Agriculture. Stanford: Stanford University Press. Northam, R. M. 1975. Urban Geography. New York: John Wiley & Sons. Nurse, D., and T. Spear. 1985. The Swahili: Reconstructing the History and Language of an African Society, 800–1500. Philadelphia: University of Pennsylvania Press. Pearce, Major F. B. (1920) 1967. Zanzibar: The Island Metropolis of Eastern Africa. London: Frank Cass & Co.

98

case studies in archaeobotany

Phillipson, D. W. 2000. “Aksumite Urbanism.” In Africa’s Urban Past, edited by D. M. Anderson and R. Rathbone, 52–64. Oxford: James Currey. Pikirayi, I. 2001. The Zimbabwe Culture: Origins and Decline of Southern Zambezian States. Walnut Creek, CA: Altamira. Posnansky, M. 1984. “The Ethnoarchaeology of Farm Shelters at Hani, Ghana.” Anthroquest 30:11–12. Pouwels, R. L. 2000. “The East African Coast, c. 780–1900 C.E.” In The History of Islam in Africa, edited by N. Levtzion and R. L. Pouwels, 251–71. Athens: Ohio University Press. Pritchrad, J. M. 1975. A Geography of East Africa. London: Evans Brothers. Reddy, S. N. 1997. “If the Threshing Floor Could Talk: Integration of Agriculture and Pastoralism during the Late Harappan in Gujarat, India.” Journal of Anthropological Archaeology 16 (2): 162–87. Redman, C. L. 1978. The Rise of Civilization. San Francisco: W. H. Freeman. Robertshaw, P. T., and W. Wetterstrom. 1989. “Plant Remains from Gogo Falls.” Nyame Akuma 31:25–27. Sinclair, P. J. J., and T. Håkansson. 2000. “The Swahili City-State Culture.” In A Comparative Study of Thirty City-State Cultures, edited by M. H. Hansen, 463–82. Copenhagen: Royal Danish Academy of Sciences and Letters. Sjoberg, G. 1960. The Preindustrial City, Past and Present. Glencoe, IL: Free Press. Smith, M. L. 2003. “Introduction: The Social Construct of Ancient Cities.” In The Social Construction of Ancient Cities, edited by M. L. Smith, 1–36. Washington, DC: Smithsonian Books. Soper, R. 1997. “Eastern African Terraced-Irrigation Systems.” In Encyclopedia of Precolonial Africa: Archaeology, History, Languages, Cultures, and Environments, edited by J. O. Vogel, 227–31. Walnut Creek, CA: Altamira. Stevens, C. J. 2003. “An Investigation of Agricultural Consumption and Production Models for Prehistoric and Roman Britain.” Environmental Archaeology 8 (1): 61–76. Trigger, B. 1972. “Determinants of Urban Growth in Pre-Industrial Societies.” In Man, Settlement, and Urbanism, edited by P. J. Ucko, R. Tringham, and G. W. Dimbleby, 575–95. London: Duckworth. ———. 2003. Understanding Early Civilizations: A Comparative Study. Cambridge: Cambridge University Press. Walshaw, S. C. 2003. “Archaeobotanical Report.” In Pemba Archaeological Project Preliminary Report: First Season, June–August 2002, edited by A. LaViolette, J. Fleisher, and B. Mapunda, 20–21. Unpublished report submitted to the Department of Archives, Museums, and Antiquities, Zanzibar, Tanzania. Wetterstrom, W. 1991. “Fleshing Out Early Agriculture with Ancient Seeds.” In Origins and Development of Agriculture in East Africa: The Ethnosystems Approach to the Study of Early Food Production in Kenya, edited by R. E. Leakey

Swahili Urban Food Production

99

and L. J. Slikkerveer, 219–34. Studies in Technology and Social Change, 19. Ames: Iowa State University. Williams, R. O. 1949. The Useful and Ornamental Plants in Zanzibar and Pemba. Timperley, Altrincham: St. Ann’s Press. Wilson, T. H., and A. L. Omar. 1997. “Archaeological Investigations at Pate.” Azania 32:31–76. Wittfogel, K. A. 1957. Oriental Despotism: A Comparative Study of Total Power. New Haven: Yale University Press. Wright, H. T. 1984. “Early Seafarers of the Comoro Islands: The Dembeni Phase of the IXth–Xth Centuries AD.” Azania 19:13–59. ———. 1993. “Trade and Politics on the Eastern Littoral of Africa, AD 800– 1300.” In The Archaeology of Africa: Food, Metal and Towns, edited by T. Shaw, P. J. J. Sinclair, B. Andah, and A. Okpoko, 658–72. London: Routledge. Zeder, M. A. 1991. Feeding Cities: Specialized Animal Economy in the Ancient Near East. Washington, DC: Smithsonian Institution Press. ———. 2003. “Food Provisioning in Urban Societies: A View from Northern Mesopotamia.” In The Social Construction of Ancient Cities, edited by M. L. Smith, 156–83. Washington, DC: Smithsonian Books.

C h ap t e r 5

Plant Food Subsistence in Context An Example from Epipaleolithic Southwest Anatolia Danièle Martinoli

The results from the archaeobotanical investigations of two Epipaleolithic cave sites, Öküzini and Karain B, in southwest Anatolia (Martinoli 2002, 2009; Martinoli and Jacomet 2004) offer the opportunity to explore the extent to which models of optimal foraging can explain aspects of subsistence decision making. The goal of this chapter is to explore whether, knowing the composition of the plant resources used by the Epipaleolithic hunter-gatherer populations of Öküzini and Karain B, it would be possible to make useful predictions about the abundance of plant resources in the environment of the sites, using a simple diet breadth optimization model. I will first present briefly the principles of optimal foraging theory and the general predictions it enables us to make. Then, I will review the archaeobotanical record of the plants exploited for food at Öküzini and Karain B, and the data available from other Late Pleistocene huntergatherer sites from different ecological zones across southwest Asia and southeast Europe. The pattern of variation will be discussed in terms of the presence or absence of the classes of plant foods nuts, roots/bulbs or tubers, and small seeds. To use models of optimal foraging, data about the ranking of the plant resources are necessary, and these data will be

100

Plant Food Subsistence in Context

101

gathered from the ethnographic and experimental literature. The application of a diet breadth model to the archaeobotanical data from Öküzini and Karain B will permit us to make predictions about the abundance of resources among the local vegetation, hypotheses that will be tested with the help of environmental reconstruction. This investigation focuses on the general strategy of adaptation and assumes important simplifications. For the optimal foraging model, we will consider only the classes of plant resources that provide the most energy. Their presence or absence will be considered to best match the level of precision commonly encountered in archaeological data on subsistence.

Optimal Foraging Models Optimal foraging models help in understanding how foraging choices are made. The most basic assumption is that human decision making is oriented toward efficiency in food acquisition as a result of evolutionary selection pressures (Kelly 1995; Winterhalder and Smith 1981; Simms 1987). To understand the factors and environmental features affecting resource selection, optimal foraging models have been developed within the field of human behavioral ecology: it is assumed that hunter-­gatherers make choices that maximize their foraging efficiency and therefore fitness. Models of optimal foraging have yielded three basic predictions: foragers should (1) prefer more profitable resources, (2) be more selective when profitable resources are common, and (3) ignore unprofitable resources that are outside the optimal diet regardless of how common they are. The assumption of efficiency is compatible with observed huntergatherer subsistence behavior (Kelly 1995; Lee 1968). More precisely, the diet breadth model, or optimal diet model, predicts that in a fine-grained environment where resources are encountered at random, a forager chooses from an available set of resources on the basis of their efficiency rank and abundance. Efficiency rank is measured as a function of the postencounter handling costs over the energy yield of the resource, the handling cost for plant resources being composed of gathering and processing. Another type of cost that helps in understanding the proportions of resources in the diet is search time, closely linked to abundance. The diet breadth model is used to predict the order in which resources will be added to or deleted from the diet. Several important predictions stem from the basic model: (1) high-ranking resources will always be taken when they are encountered, (2) the inclusion of lowerranked resources in the diet will depend not on their abundance but on the abundance of higher-ranked items, and (3) as the abundance of higher-ranked items decreases, lower-ranked items will be included in

102

case studies in archaeobotany

the diet. Conversely, as the abundance of higher-ranked items increases, lower-ranked items will be excluded, no matter how abundant they are. The principal criticism of the diet breadth model has been that human subsistence is controlled not only by energy and resource abundance but also by cultural practices, technology, cooking and eating habits, and taboos (Bettinger 1987; Gremillion 2004; Stahl 1989). However, this kind of model is not intended to address whether a behavior is optimal, but rather whether a particular hypothesis, based on specific constraints, describes the foraging behavior.

Preliminary Remarks on Plant-Based Subsistence Reconstruction and Intersite Comparison It is difficult to trace the subsistence strategy among prehistoric huntergatherer societies, for which we usually have only incomplete archaeological remains. Primary evidence for plant food from the fragile botanical remains is often highly biased by lack of recovery or poor preservation. In Near Eastern sites, most attention has been paid to the exploitation of small-seeded grasses and wild cereals as a forerunner to agricultural practices. Regardless of the species involved, such a foraging strategy can be traced back to the Levantine Epipaleolithic (Weiss 2004) and was widespread during the Pre-Pottery Neolithic in all areas where wild cereals were abundant, including the Levant and northern Fertile Crescent (Willcox 1999). However, traces of the plant subsistence base in the Upper and Epipaleolithic are still scarce and it is likely that a greater variety of foraging patterns existed. Our knowledge about the plant foods used in Upper and Epipaleolithic sites in southwest Asia is based on a limited number of studies at sites geographically scattered over a huge area (figure 5.1). Archaeobotanical records from these periods are highly biased by differential preservation, a problem that has been fully discussed elsewhere (Colledge 2001; Hillman, Madeyska, and Hather 1989). Moreover, the recovery methods (manual/machine flotation, sieving, mesh size), the contexts, the kind of sites excavated (open-air or cave site; year-round or seasonal occupation), and the sampling strategies (large scale, quantity of processed sediment) differ and are often difficult to reconstruct (table 5.1). Differences in preservation of the plant remains can also influence their level of identification. Another problem consists in quantifying the use of the plants and their importance in prehistoric diets, based on the recovered macroremains (Popper 1988). For all these reasons, an intersite comparison is only possible at a low degree of precision.

Plant Food Subsistence in Context

103

Figure 5.1.  Map showing the location of the sites mentioned in the text.

Plant-Based Subsistence in Upper and Epipaleolithic in Southwest Asia and Southeast Europe With the excavation at the Öküzini and Karain B caves in southwest Anatolia (figure 5.1) in 1990–1999 and 1996–2001, a little light was shed on plant exploitation during the Epipaleolithic in this part of southwest Asia (Martinoli 2004). Covering a stratigraphic sequence dated at 16,500– 12,000 uncal BP in Öküzini and around 15,000 uncal BP in Karain B, the sites were inhabited periodically at the end of the Late Glacial Maximum until the start of the Younger Dryas (Otte et al. 2003). The caves were not very rich in botanical material, with the bulk of the remains originating from the upper levels. Both sites yielded a very similar plant assemblage and little temporal patterning was evidenced. The principal plant foods appear to have been nuts and fruits complemented with underground storage organs like roots, bulbs, or tubers. These included wild almonds (Amygdalus graeca or orientalis), wild pistachio (Pistacia), acorns (Quercus), wild pears (Pyrus), wild grapes (Vitis vinifera subsp. sylvestris), rosehips (Rosa), hawthorns (Crataegus), and hackberry fruits (Celtis). There were also a very small number of small pulses (Vicieae) and other small seeds (table 5.2) at Öküzini. However, like the domestic cereal grains recovered, the Vicieae could represent later contaminations,

Franchthi II

Upper Paleolithic

13,000–11,000 uncal BP

Manual flotation (1.5 mm)

Cave site

Seasonal occupation

Woodland-open forest with steppic areas

Hansen 1991

Wadi Kubbaniya

Late Paleolithic

19,000–17,000 uncal BP

Dry sieving (1.7 and 0.9 mm)

Open-air site

Seasonal occupation

Arid environment with riverbank vegetation

Hillman et al. 1989

Hansen 1991; Mason et al. 2992

Open oak woodland interspersed with open patches of herbaceous vegetation

Seasonal or permanent occupation

Cave site

Manual flotation (1.5 mm)

9500–7900 uncal BP

Upper Paleolithic/ Lower-Upper Mesolithic

Franchthi III–V

Open-air site

Machine flotation (1 mm)

11,500–10,000 uncal BP

Epipaleolithic

Abu Hureyra I

Kislev et al. 1992; Kislev and Simchoni 2002; Weiss et al. 2004

Hillman 2000

Saline habitat, Therebinth-almond lakeshore habitat steppe and riverine and Mediterranean forest open parklike forest

Seasonal occupation, maybe permanent

Open-air site

Wet sieving (2 and 0.5 mm)

19,000 uncal BP

Levantine Epipaleolithic/ Kebaran

Ohalo II

Hopf and BarYosef 1987

Cave site

Wet sieving and flotation

12,000 uncal BP

Epipaleolithic/ Early and Late Natufian

Hayonim

Table 5.1. Description of the Upper and Epipalaeolithic sites in southwest Asia and southeast Europe where plant remains have been recovered

Manual flotation (0.5 mm)

12,200–11,900 uncal BP

Bucket flotation (0.3 mm)

Open-air site

10450–9550 uncal BP

Manual water flotation

Open-air site Seasonal occupation Forest-steppe and riverine forest Martinoli 2004

Steppe and river valley

van Zeist and Bakker-Heeres 1984/86

Cave site

Permanent from phase II onward

Colledge 2001

~15000 uncal BP

Epipaleolithic/ Natufian

Epipaleolithic/ Natufian

Epipaleolithic

Wadi Hammeh 28 Karain B

Mureybit I–III

Martinoli 2004

Forest-steppe and riverine forest

Seasonal occupation

Cave site

Manual flotation (0.5 mm)

16500–12000 uncal BP

Epipaleolithic

Öküzini

Rosenberg et al. 1998

Riverine woodland and deciduous mixed oak forest

Permanent

Open-air site

Machine flotation (1 and 0.35 mm)

10,600–9900 uncal BP

terminal Epipaleolithic

Hallan Çemi

O

Hallan Çemi

Wadi Hammeh 27

O

Karain B

Mureybit I–III

O

Öküzini

Hayonim

Tell Abu Hureira I

Ohalo II

Franchthi III–V

Franchthi II

Wadi Kubbaniya

Table 5.2. Presence (more than five items) and absence of selected taxa from the major plant categories at Upper and Epipalaeolithic settlements from southwest Asia and southeast Europe

Wild cereals Hordeum sp. Triticum boeticum Boiss/Secale sp. Triticum dicoccoides Schrank

?

O OO OO O

O

barley einkorn/rye emmer

Wild grasses Aegilops sp. Alopecurus sp. Avena sp. Bromus sp. Echinochloa sp. Eremopyrum sp Hordeum sp. small Lolium perenne L./ rigidum Gaudin Panicaceae Poaceae Puccinellia sp. Stipa sp. Taeniatherum caput-medusae (L.) Nevski

O

O O O O O OO O

O

cf

O O O O

O O

cf

O

O

O

goat grass black grass oat brome grass panic grass wheat grass small barley ryegrass panicaceae O grasse alkali-grass medusahead

O OO

Pulses Cicer sp. Latyrus sp. Lens sp. Lupinus pilosus Murr. Pisum sp.

O O

O O O

O

O

O O cf

cf

O

chickpea vetching O lentil lupin pea continued

O

O

cf

O

O

O

Hallan Çemi

O

Karain B

O

Öküzini

Wadi Hammeh 27

O

Mureybit I–III

Vicia ervilia (L.) Willd.

Hayonim

Large-seaded Viciae

Tell Abu Hureira I

Ohalo II

Franchthi III–V

Franchthi II

Wadi Kubbaniya

Table 5.2. Continued

O large-seeded Viciaea O bitter vetch

Small seeds Cruciferae Gundelia tournefortii L. Small-seaded Trifolieae Small-seeded Viciae

O

O

O

O

cf

O OO

Polygonum corrigioloides Jub. & Spach Bolboschoenus maritimus (L.) Palla/Scirpus tuberosus Desf.

Cruciferae O tumble thistle/ cardoon small-seeded Trifoliaea O small-seeded Vicieae knot grass

O

O

OO

OO sea club-grass

Roots/bulbs/tubers Cyperus rotundus L.

O

Diverse parenchyma tissue Bolboschoenus maritimus (L.) Palla/Scirpus tuberosus Desf.

O

O

O

O

O

O

O

O

purple nut sedge parenchymous tissue sea club-rush

Nuts Amygdalus sp./Prunus sp. Pistacia atlantica Desf./terebinthus L.

O

O

O

cf

cf

O

O O

OO O O

cf

O

O almond/plum O pistachio continued

108

case studies in archaeobotany

Pistacia lentiscus L. Quercus sp.

O

O

Hallan Çemi

Karain B

Öküzini

Wadi Hammeh 27

Mureybit I–III

Hayonim

Tell Abu Hureira I

Ohalo II

Franchthi III–V

Franchthi II

Wadi Kubbaniya

Table 5.2. Continued

lentisc oak

Soft fruits Celtis sp. (mineralised) Crataegus sp. Ficus sp. Olea europea L. sylvestris Prunus sp. Pyrus sp. Rosa sp. Vitis vinifera L. sylvestris Ziziphus spina-christi (L.) Desf.

O

O

O

O

O

O

hawthorn fig wild olive

O O

O O

O

cf O O

O O

hackberry

OO cf O O

plum pear rosehip wild grape christ’s thorn

Note: OO = important in the assemblage; O = present; cf = uncertain identification; ? = contamination. References are the same as for table 5.1.

since they often grow as cereal weeds (Martinoli 2004). Evidence of an intensive use of the small seeds was absent, despite the appearance of grinding stones at 12,500 uncal BP (Otte et al. 2003). These were not necessarily made for small-seed processing, but could be used for various other plants, meat, or minerals (Ertug 2002). Undoubtedly, the composition of the plant assemblages reflects in part poor preservation of individual specimens, bias that cannot entirely be excluded, so that the missing small seeds could eventually represent so-called missing foods (Hillman 2000; Hillman, Madeyska, and Hather 1989). Nevertheless, the trivial role of small seeds was supported by several lines of evidence (Martinoli 2004) other than the small number and very low frequency of the fossil seeds. If intensely used, small

Plant Food Subsistence in Context

109

seeds would have been more abundant in the archaeobotanical record because (1) most small-seed species identified require roasting to become edible and would therefore have good chances of preservation, and (2) the Öküzini cave, at least, bears traces of early summer occupation (Atıcı and Stutz 2002), a season when most small-seeded plants would be available.

Regional Review of Plant-Based Subsistence Strategies In table 5.2 I have listed the presence or absence and an estimate of the importance of the main plant food categories and species from the Upper and Epipaleolithic settlements from southwest Asia and southeast Europe. The environments of the sites listed in table 5.1 were taken, when available, from the authors’ descriptions based on macrobotanical, anthracological, and in some cases also zoological data. For the Upper Paleolithic, only two sites have yielded botanical remains: Wadi Kubbaniya and Franchthi II. Wadi Kubbaniya, an Egyptian site where large-scale recovery of plant remains was undertaken, produced a small number of vegetal remains, among which the major plants consisted of tubers and rhizomes of taxa growing in riverine ecosystems, augmented with rare remains of fruits of the dom palm (Hyphaena thebaica) and club rush seeds (Scirpus maritimus/tuberosus). The presence of small seeds in apparent human feces suggested their use as food. However, loose small seeds were not recovered, but they may have been lost during excavation or may not have been preserved (Hillman 2000; Hillman, Madeyska, and Hather 1989). The environment was arid but supported riverine vegetation. In zone II at Franchthi in Greece, the edible species vetch (Vicia), lentil (Lens), almond (Amygdalus communis), pear (Pyrus amygdaliformis), and wild pistachio (Pistacia cf. lentiscus) were the most numerous remains (Hansen 1991). Oat seeds (Avena) were probably also used for food, whereas the wild barley grains (Hordeum vulgare subsp. spontaneum) were considered to be contaminants. The vegetation was suggested to consist of woodland or open forest with steppic areas. For the Epipaleolithic, the archaeobotanical data become slightly more numerous. In Franchthi zones III and IV, the most abundant species were pistachio (Pistacia cf. lentiscus), almond (Amygdalus communis), pear (Pyrus amygdaliformis), and oats (Avena). Lentils (Lens), wild barley (Hordeum vulgare spontaneum), and a variety of legumes were also consistently present throughout most of the zones (Hansen 1991). New investigations found, in some samples at least, considerable quantities of parenchyma, presumably from “roots and tubers” (Mason, Hather, and Hillman 2002). The vegetation at this period had developed into open

110

case studies in archaeobotany

oak woodland interspersed with patches of herbaceous vegetation (Hansen 1991). Ohalo II in Israel provided a large collection of charred plant remains preserved in wet conditions, and 142 taxa were identified (Kislev, Nadel, and Carmi 1992; Weiss et al. 2004). The principal plant foods appear to have been grass seeds and wild cereals, interpreted as staples, augmented with a variety of nuts (Mount Tabor oak acorns/Quercus ithaburensis, almonds/Amygdalus, pistachios/Pistacia atlantica) and fruits (wild olives/ Olea europaea, hawthorns/Crataegus, wild fig/Ficus, and wild grapes/ Vitis vinifera subsp. sylvestris). Small quantities of pulses were also present. The surroundings mainly included an oak steppe forest with annual grasses among the trees and the small and saline nearby Lake Kinneret (Kislev and Simchoni 2002). Another site with a large plant assemblage was Abu Hureyra I in Syria, where more than seventy plant taxa were identified thanks to a large-scale flotation strategy (Hillman 2000). Wild cereals (rye/Secale, einkorn/Triticum boeticum/urartu, and emmer/Triticum dicoccoides), feather-grass/ Stipa, sea club rush/Scirpus maritimus/tuberosus, and knotgrass/Polygonum corrigioloides probably served as staples. Other small-seeded grasses (small barley/Hordeum murinum, etc.), pulses (lentils/Lens sp. and large seeded vetches/Vicieae), small-seeded legumes (Trifolieae), nuts (wild pistachios/Pistacia), fruits (hackberries/Celtis), and the carbohydraterich seeds of goosefoots (Chenopodiaceae) have also been interpreted as food plants. The vegetation surrounding the site was dominated by oak-terebinth-Rosaceae park woodland and steppe, shaping a mosaic of woodland and open, grass-dominated areas, and a riverine forest (Hillman 1996, 2000). The Israeli site of Hayonim Cave in western Galilee yielded a limited amount of plant remains: wild cereals (barley/Hordeum spontaneum), wild pulses (possibly pea/cf. Pisum sp., and lupin/Lupinus pilosus), and nuts (almond/Amygdalus communis) (Hopf and Bar-Yosef 1987). According to Hillman’s (1996) model of vegetation, the landscape would have been dominated by steppe and a terebinth-almond woodland-steppe. At Mureybit I–III, wild einkorn wheat (Triticum boeticum), wild barley (Hordeum), lentils (Lens), and peas (Pisum) served as human food, as well as seeds from a range of other wild small-seeded taxa (Van Zeist and Bakker-Heeres 1984/86). Wild fruit trees and shrubs exploited by the inhabitants of the site included pistachios (Pistacia), figs (Ficus), and capers (Capparis). The steppe and river valley forest constituting the natural vegetation of Wadi Hammeh 27, on the eastern margin of the Jordan rift valley, yielded only a few plant remains despite large-scale sampling, among them wild cereals and various grasses (barley/Hordeum and feather grass/Stipa), lentils (Lens), and pistachios (Pistacia) (Colledge 2001). The site was located in terebinth-almond woodland-steppe (Hillman 1996).

Plant Food Subsistence in Context

111

In Hallan Çemi, the charred plant assemblages showed a paucity of grasses and a near absence of wild cereals, but other wild plants like nuts (pistachios/Pistacia, almonds/Amygdalus) and wild pulses (lentils/Lens, bitter vetches/Vicia ervilia) were of great importance to the inhabitants of the site, as well as some other small seeds (e.g., Gundelia tournefortii) (Rosenberg et al. 1998). Surrounding vegetation comprised a deciduous mixed oak forest and riverine woodland.

Synthesis The overview of the plant subsistence base at the different sites shows some patterning (table 5.2). The majority of the assemblages yielded small-seeded plants in some proportion: wild cereals, wild grasses, legumes, or other small seeds. Various species among these categories were abundant among the remains and were therefore interpreted as staples, for example, at Ohalo II and Abu Hureyra. Exceptions to this pattern of small-seed use are represented by Wadi Kubbaniya (where, however, small seeds have been reported from human feces), Öküzini, and Karain B. In the Levantine and Euphrates sites (Ohalo II, Hayonim, Wadi Hammeh, Abu Hureyra, and Mureybit), the grasses, wild cereals, and other small seeds, like knotgrass or sea club-rush, were present and most abundant. On the other hand, at Hallan Çemi and Franchthi, legume seeds were present and seemed to play an important role. Nuts and fruits were regularly present in the southwest Asian assemblages; however, they seem to have a main role in the Öküzini assemblage only, maybe also in Hallan Çemi. Sites like Hayonim, Wadi Hammeh, and Mureybit are poor in nuts and fruits, which can probably be explained from their dry steppic environments. Roots, bulbs, or tubers were recovered at only four sites (Wadi Kubbaniya, Franchthi III–IV, Öküzini, and Karain B). It must be noted that it is only since the work of Hather (1991, 1993) that the identification of parenchymous tissues from roots, bulbs, or tubers is relevant. The absence of this kind of plant remains at most sites possibly results from an absence of recognition. Nevertheless, parenchymous root and tubers foods were likely of importance at Franchthi, Öküzini, and Karain B, especially when viewed in relation to small-seed foods, which on these sites were not or very rarely recovered. The same pattern has been noticed for European hunter-gatherer sites, where small seeds are rare, and nuts and parenchymous tissue more frequent (Mason et al. 2002). The main question arising from this review is, why are there comparatively so few small seeds at Öküzini and Karain B? To answer this question using optimal foraging theory, we need data on the cost of acquiring the main classes of plant resources.

112

case studies in archaeobotany

Use of Optimal Foraging Theory as an Explanatory Model Ranking of the Main Classes of Food Plants To apply models of optimal foraging to archaeological situations, it is necessary to acquire data on the cost of exploiting resources utilized by the autochthonous populations, which means we need to acquire estimates of handling time and a ranking of the resources. The benefits of the resources are generally expressed in calories. There are few ethnographic published data on the costs of acquiring plant resources; therefore, some authors have devised ethnographic experiments (Gremillion 2004; Simms 1987). These experiments showed that data for groups of plant types with similar morphological characteristics could be seen as being similar. Ethnographic examples of root food are numerous in hunter-gatherer as well as agrarian societies, past and present (see, e.g., Ertug 2000, 2004; Kelly 1995; Kubiak-Martens 2002; Pokotylo and Froese 1983). These studies have shown that the preference for roots over seed foods is usually a product of the lower energy costs involved in processing them (Cane 1989). Lee and Devore (1968, 7), ethnographers intimately familiar with the practicalities of subsisting on wild resources, wrote, “Our view is that vegetable food in the form of nuts, berries and roots were always available to man and were easily exploited by even the simplest of technologies.” Nuts and roots or tubers are generally considered plant foods with high return rates, but they can be quite variable depending mostly on processing requirements (Kelly 1995; Smith and Martin 2001; Talalay et al. 1984). Roots and tubers from wetland plants are particularly interesting, because they usually grow in large stands and many of them are edible (Hillman, Madeyska, and Hather 1989; Smith and Martin 2001). The exploitation of small seeds, on the other hand, needs labor-expensive processing. They vary widely in return rates, but in general are the lowest-ranked resources. The resulting return rates are generally quite low in comparison to those obtainable from different nuts and acorns, for example (Gremillion 2004).

Predictions Using the Optimal Foraging Model It seems that the main plant resources recovered at Öküzini and Karain B belonged to plant categories that usually require little time and labor to procure and process (Martinoli 2004). The plant foods originated mostly from shrubs, trees, or underground plant storage organs, were generally easy to gather, and did not need intensive processing. According to the diet breadth model, the use of mainly high-ranked plant foods reflects

Plant Food Subsistence in Context

113

their easy availability in the environment. It does not, however, imply the absence of grasses and pulses or other small seeds, but estimates that nuts, roots/bulbs, and tubers were privileged as resources because they had better return rates and were sufficiently abundant for the human population present. The exploitation in the fall of nuts like the wild almond, a slightly lower-ranked resource because it needs detoxification, can tentatively be explained by risk management behavior. It is a storable and energy-rich resource valued in anticipation of the lean winter season. Moreover, almond wood was one of the main fuel resources throughout the sequence in Öküzini and Karain B (Martinoli 2009; Thiébault 2002). The gathering of a plant resource for multiple uses can lower its exploitation cost. Two important points emerged from encounter rates developed by computer simulations (Simms 1987): the search time for plant collecting is a relatively less important factor than handling time; and with increasing handling time (i.e., low-ranked resources), abundance becomes less important in altering the procurement rate. That means that the abundance does not change the classification of the classes of plant resources, but rather accentuates the differences between the return rates. High-ranked resources tend therefore to strongly increase their return rate when they become abundant, whereas low-ranked resources, even if abundant, increase their return rates only slightly.

Test of the Predictions of Optimal Foraging Theory Late Glacial Environmental Reconstruction in Southwest Anatolia and the Abundance of Classes of Plant Resources There is no way to confidently estimate the abundance of each plant type in the local environment of the Öküzini and Karain B caves. However, the coincident use of pollen, charcoal, and archaeozoological data, together with a good knowledge of the local topography and potential vegetation, enables an ecological modeling of the vegetation in past times (e.g., Hillman 1996; Martinoli 2009). The mountains in southwest Anatolia served during the Late Glacial Maximum as a refuge for arboreal species, which, accompanied by an increase in humidity, colonized the area from 14,000 uncal BP onward (Kuzucuoglu and Roberts 1998). A sediment core taken in the plain in front of the Öküzini cave with a phase dated at 14,000 uncal BP showed that deciduous oak (Quercus), ash (Fraxinus), wild pistachio (Pistacia), and juniper (Juniperus) were already present, although the nonarboreal species dominated the pollen record (Kuzucuoglu et al. 2002). However, insect-pollinated species like almond (Amygdalus) and pear trees (Pyrus)

114

case studies in archaeobotany

are usually missing in pollen records, so the overall rate of arboreal pollen is underestimated (Woldring and Bottema 2001/2). Riverside associations were also present at 14,000 uncal BP with pollen from aquatic and semiaquatic plants (Kuzucuoglu et al. 2002). This correlates well with the existence of an ancient lake in the plain in front of the caves (Pawlikowski 2002). Anthracological analyses made at the Öküzini cave showed the presence of a forest steppe from 14,500 BP uncal to 12,200 BP uncal in which the almond tree played an important role (Thiébault 2002). Other woody steppic species identified were wormwood (Artemisia), pine (Pinus), and wild pistachio (Pistacia). However, and from the base of the sequence, the presence of mesophilous species such as deciduous oaks (Quercus), maple (Acer), plums (Prunus), boxwood (cf. Buxus sempervirens), and whitebeam/hawthorn (Sorbus/Crataegus), or of more thermophilous species such as sclerophilous oaks (Quercus) and olive trees (cf. Olea), was noted. A riverside forest was also present, composed of tamarisk (Tamarix), ash (Fraxinus), and willow/poplar (Salix/Populus) (Thiébault 2002). The archaeozoological record showed hunting of mainly ovicaprines (sheep and goats) and fallow deer and roe deer, the latter increasing in numbers during this time, which correlated with an increasingly forested environment from 16,000 to 12,000 uncal BP (Lopez Bayon, Léotard, and Kartal 2002). All the data point to the presence of a forested area rich in nuts and fruits, although its cover and density are difficult to estimate precisely. Öküzini and Karain B were situated at the junction between a plain and a mountain range, a topography that offers a high variety of ecological niches. The environment was therefore varied, with the more mesic tree species probably growing in protected valleys, the thermophilous species on the exposed foothills, the steppic bushes on dry, rocky mountain flanks, and the riverine trees and semiaquatic and aquatic species in the wet areas and marshes in the plain. Archaeological survey of the province of Antalya showed that the area supported small mobile groups (Kayan et al. 1987). It is therefore quite possible that in the forest and wetland high-ranked plant resources like the nuts and underground organs were sufficiently abundant for the subsistence of the inhabitants of Öküzini and Karain B, without resorting to the exploitation of low-ranked small seeds. Moreover, both settlements are interpreted as seasonal camps, and therefore the inhabitants had the possibility of shifting to another place when high-ranked resources (both animal and vegetal) decreased, rather than resort to low-ranked resources. A detailed examination of the local environments of all the sites mentioned here is beyond the scope of this chapter. However, it seems that, wherever present, even in small amounts, high-ranked resources like nuts

Plant Food Subsistence in Context

115

were constantly used. On the other hand, populations of sites situated in open steppic woodland vegetation have always exploited low-ranked resources like small seeds (grasses, wild cereals, legumes, and others), even when wetlands rich in root food (i.e., higher-ranked resources) were present nearby (e.g., Hallan Çemi, Ohalo). In these cases, factors like storability, nutritional content, dependability of a resource, technological skills, and so on, were probably as important as the energy provided. This is in contradiction with the predictions of optimal foraging theory, and it would be challenging to find ways to test these hypotheses.

Conclusions The use of optimal foraging theory as an interpretative model at Öküzini and Karain B to help in understanding why there are nearly no small seeds at these sites seems productive. The theory shows that the importance of nut and root food in relation to small-seed food reflects not merely a preservation pattern; rather, the explanation is likely to be found in the high return rates of nuts and roots and their easy availability in the area, together with a low human population. However, it must be kept in mind that there is no diet composed solely of nuts, roots, or seeds; this is a simplification made for the purpose of this study. It is always a variety of resources that are consumed. These first results encourage the pursuit of a more detailed investigation exploring the relation between subsistence and environment in more complexity. One way to do this would be to estimate more precisely the efficiency ranking of the plants recovered archaeologically with the help of experiments. However, as long as we are unable to specifically identify the plant remains (in particular the parenchymous tissues) and to understand how they were processed, we will be unable to estimate precisely their return rates. A better understanding of the connection between subsistence and environment will also be possible only with cautious modeling of the local environment of a site, a process that benefits greatly from multidisciplinary approaches. To conclude, much more experimentation and regional investigations are needed to better document the subtle relationship between plant food exploitation and environment, and hence the foraging behavior of prehistoric societies.

Acknowledgments I would like to thank G. Hillman and M. Nesbitt for initiating me into Near Eastern archaeobotany and handing over to me the plant material

116

case studies in archaeobotany

at Öküzini and Karain B, as well as the excavation directors I. Yalcınkaya and M. Otte and their team for their collaborative work. I wish to thank S. Jacomet, M. Nesbitt, and M. Savard for their critical review of the manuscript. This research has been made possible by the support of S. Jacomet and colleagues from the IPAS, and the financial support of the Swiss National Science Foundation (project number 1214-64974.01 and 101312-101585/1).

References Atıcı, L., and A. Stutz. 2002. “Analysis of the Ungulate Fauna from Öküzini: A Preliminary Reconstruction of Site Use, Seasonality, and Mortality Pattern.” In Öküzini: Final Paleolithic Evolution in Southwest Anatolia, edited by I. Yalçınkaya, M. Otte, J. Kozlowski, and O. Bar-Yosef, 101–8. Liège: Eraul. Bettinger, R. L. 1987. “Archaeological Approaches to Hunter-Gatherers.” Annual Review of Anthropology 16:121–42. Cane, S. 1989. “Australian Aboriginal Seed Grinding and Its Archaeological Record: A Case Study from the Western Desert.” In Foraging and Farming: The Evolution of Plant Exploitation, edited by D. R. Harris and G. C. Hillman, 99–119. London: Unwin Hyman. Colledge, S. 2001. Plant Exploitation on Epipalaeolithic and Early Neolithic Sites in the Levant. BAR International Series, vol. 986. Oxford: Archaeopress. Ertug, F. 2000. “An Ethnobotanical Study in Central Anatolia (Turkey).” Economic Botany 54 (2): 155–82. ———. 2002. “Pounders and Grinders in a Modern Central Anatolian Village.” In Moudre et broyer: L’interprétation fonctionnelle de l’outillage de mouture et de broyage dans la préhistoire et l’antiquité, edited by H. Procopiu and R. Treuil, 211–25. Paris: Édition du CTHS. ———. 2004. “Wild Edible Plants of the Bodrum Area (Mugla, Turkey).” Turkish Journal of Botany 28:161–74. Gremillion, K. J. 2004. “Seed Processing and the Origin of Food Production in Eastern North America.” American Antiquity 69 (2): 215–33. Hansen, J. 1991. The Palaeoethnobotany of Franchthi Cave. Excavations at Franchthi Cave, Greece, fascicle 7. Bloomington: Indiana University Press. Hather, J. G. 1991. “The Identification of Charred Archaeological Remains of Vegetative Parenchymous Tissue.” Journal of Archaeological Science 18:661–75. ———. 1993. An Archaeobotanical Guide to Root and Tuber Identification 1, Europe and South West Asia. Oxford: Oxbow Monograph. Hillman, G. C. 1996. “Late Pleistocene Changes in Wild Plant-Foods available to Hunter-Gatherers of the Northern Fertile Crescent: Possible Preludes to

Plant Food Subsistence in Context

117

Cereal Cultivation.” In The Origins and Spread of Agriculture and Pastoralism in Eurasia, edited by D. R. Harris, 159–203. London: UCL Press. ———. 2000. “Abu Hureyra I: The Epipalaeolithic.” In Village on the Euphrates: From Foraging to Farming at Abu Hureyra, edited by A. M. T. Moore, G. C. Hillman, and A. J. Legge, 327–98. Oxford: Oxford University Press. Hillman, G. C., E. Madeyska, and J. G. Hather. 1989. “Wild Plant Foods and Diet at Late Palaeolithic Wadi Kubbaniya: The Evidence from Charred Remains.” In The Prehistory of Wadi Kubbaniya: Stratigraphy, Palaeoeconomy and Environment, edited by F. Wendorf, R. Schild, and A. Close, 159–242. Dallas: Southern Methodist University Press. Hopf, M., and O. Bar-Yosef. 1987. “Plant Remains from Hayonim Cave, Western Galilee.” Paléorient 13 (1): 117–20. Kayan, I., A. Minzoni-Déroche, and I. Yalçınkaya. 1987. “Prospection préhistorique dans la région d’Antalya: Notices préliminaires.” Anatolia Antiqua (Varia Anatolica) 1:9–13. Kelly, R. L. 1995. The Foraging Spectrum: Diversity in Hunter-Gatherer Lifeways. Washington, DC: Smithsonian Institution Press. Kislev, M., D. Nadel, and I. Carmi. 1992. “Epipalaeolithic (19,000 BP) Cereal and Fruit Diet at Ohalo II, Sea of Galilee, Israel.” Review of Palaeobotany and Palynology 73:161–66. Kislev, M. E., and O. Simchoni. 2002. “Reconstructing the Palaeoecology of Ohalo II, an Early Epipalaeolithic Site in Israel.” In Hunter-Gatherer Archaeobotany: Perspectives from the Northern Temperate Zone, edited by S. L. R. Mason and J. G. Hather, 174–79. London: Institute of Archaeology, University College London. Kubiak-Martens, L. 2002. “New Evidence for the Use of Root Foods in Preagrarian Subsistence Recovered from the Late Mesolithic Site at Halsskov, Denmark.” Vegetation History and Archaeobotany 11:23–31. Kuzucuoglu, C., A. Emery-Barbier, M. Fontugne, and S. Kunesh. 2002. “The Öküzini Marshes: A New Upper Pleistocene Record on the Anatolian Mediterranean Coast.” In La grotte d’Öküzini: Évolution du Paléolithique Final du sud-ouest de l’Anatolie, edited by I. Yalcınkaya, M. Otte, J. K. Kozlowski, and O. Bar-Yosef, 79–82. Liège: Eraul. Kuzucuoglu, C., and N. Roberts. 1998. “Evolution de l’environnement en Anatolie de 20000 à 6000 BP.” Paléorient 2 (2): 7–24. Lee, R. B. 1968. “What Hunters Do for a Living, or, How to Make Out on Scarce Resources.” In Man the Hunter, edited by R. B. Lee and I. Devore, 30–48. Chicago: Aldine. Lee, R. B., and I. Devore. 1968. Man the Hunter. Chicago: Aldine. Lopez Bayon, I., J. M. Léotard, and M. Kartal. 2002. “Séquence stratigraphique de la grotte d’Öküzini: Remplissage naturel et remplissage anthropique.” In La

118

case studies in archaeobotany

Grotte d’Öküzini: Évolution du Paléolithique Final du Sud-Ouest de l’Anatolie, edited by I. Yalçınkaya, M. Otte, J. Kozlowski, and O. Bar-Yosef, 25–39. Liège: Eraul. Martinoli, D. 2002. “Les macrorestes botaniques de la grotte d’Öküzini.” In La grotte d’Öküzini: Évolution du Paléolithique Final du sud-ouest de l’Anatolie, edited by I. Yalçınkaya, M. Otte, J. Kozlowski, and O. Bar-Yosef, 91–94. Liège: Eraul. ———. 2004. “Plant-Food Use, Temporal Changes and Site Seasonality at Epipalaeolithic Öküzini and Karain B Caves, Southwest Anatolia, Turkey.” Paléorient 30:61–80. ———. 2009. “Reconstruction of Local Woodland Vegetation and Use of Firewood at Two Epipalaeolithic Cave Sites in Southwest Anatolia (Turkey).” In Foragers to Farmers: Papers in Honour of Gordon C. Hillman, edited by A. Fairbairn and E. Weiss, 161–70. Oxford: Oxbow Books. Martinoli, D., and S. Jacomet. 2004. “Identifying Endocarp Remains and Exporing Their Use at Epipalaeolithic Öküzini in Southwest Anatolia, Turkey.” Vegetation History and Archaeobotany 13:45–54. Mason, S., J. G. Hather, and G. C. Hillman. 2002. “The Archaeobotany of European Hunter-Gatherers: Some Preliminary Investigations.” In Hunter-Gatherer Archaeobotany: Perspectives from the Northern Temperate Zone, edited by S. L. R. Mason and J. G. Hather, 188–96. London: Institute of Archaeology, University College London. Otte, M., I. Lopez Bayon, P. Noiret, O. Bar-Yosef, I. Yalçinkaya, M. Kartal, J. M. Léotard, and P. Pettit. 2003. “Sedimentary Deposition Rates and Carbon-14: The Epi-paleolithic Sequence of Öküzini Cave (Southwest Turkey).” Journal of Archaeological Science 30:325–41. Pawlikowski, M. 2002. “Ancient Lake near Öküzini and Karain Caves (Southern Turkey).” In Öküzini: Final Paleolithic Evolution in Southwest Anatolia, edited by I. Yalçınkaya, M. Otte, J. Kozlowski, and O. Bar-Yosef, 75–77. Liège: Eraul. Pokotylo, D. L., and P. D. Froese. 1983. “Archaeological Evidence for Prehistoric Root Gathering on the Southern Interior Plateau of British Columbia: A Case Study from Upper Hat Creek Valley.” Canadian Journal of Archaeology 7 (2): 127–57. Popper, V. S. 1988. “Selecting Quantitative Measurements in Palaeoethnobotany.” In Current Paleoethnobotany: Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, 53–71. Chicago: University of Chicago Press. Rosenberg, M., M. R. Nesbitt, R. W. Redding, and B. L. Peasnall. 1998. “Hallan Çemi, Pig Husbandry, and Post-Pleistocene Adaptations along the TaurusZagros Arc (Turkey).” Paléorient 24 (1): 25–41. Simms, S. R. 1987. Behavioral Ecology and Hunter-Gatherer Foraging: An Example from the Great Basin. BAR International Series, 381. Oxford: Archaeopress.

Plant Food Subsistence in Context

119

Smith, C. S., and W. Martin. 2001. “Sego Lilies and Prehistoric Foragers: Return Rates, Pit Ovens, and Carbohydrates.” Journal of Archaeological Research 28:169–83. Stahl, A. B. 1989. “Plant-Food Processing: Implications for Dietary Quality.” In Foraging and Farming: The Evolution of Plant Exploitation, edited by D. R. Harris and G. C. Hillman, 171–96. London: Unwin Hyman. Talalay, L., D. R. Keller, and P. J. Munson. 1984. “Hickory Nuts, Walnuts, Butternuts, and Hazelnuts: Observations and Experiments Relevant to Their Aboriginal Exploitation in Eastern North America.” Prehistory Research Series 6 (2): 338–59. Thiébault, S. 2002. “Approche de l’environnement végétal du site d’Öküzini (Turquie) au Tardiglaciaire par l’analyse anthracologique.” In La grotte d’Öküzini: Évolution du Paléolithique Final du sud-ouest de l’Anatolie, edited by I. Yalçınkaya, M. Otte, J. Kozlowski, and O. Bar-Yosef, 95–99. Liège: Eraul. Van Zeist, W., and J. A. H. Bakker-Heeres. 1984/86. “Archeobotanical Studies in the Levant. 3. Late-Palaeolithic Mureybit.” Palaeohistoria 26:171–99. Weiss, E., W. Wetterstrom, D. Nadel, and O. Bar-Yosef. 2004. “The Broad Spectrum Revisited: Evidence from Plant Remains.” Proceedings of the National Academy of Science 101 (26): 9551–55. Willcox, G. 1999. “Agrarian Change and the Beginnings of Cultivation in the Near East: Evidence from Wild Progenitors, Experimental Cultivation and Archaeobotanical Data.” In The Prehistory of Food, edited by C. Gosden and J. Hather, 479–500. London: Routledge. Winterhalder, B., and E. A. Smith. 1981. Hunter-Gatherer Foraging Strategies: Prehistoric Archeology and Ecology. Chicago: University of Chicago Press. Woldring, H., and S. Bottema. 2001/2. “The Vegetation History of East-Central Anatolia in Relation to Archaeology: The Eski Acıgöl Pollen Evidence Compared with the Near Eastern Environment.” Palaeohistoria 43/44:1–34.

C h ap t e r 6

Vegetation Proxy Data and Climate Reconstruction Examples from West Asia N a o m i F. M i l l e r

“Climate” may be defined as a thirty-year average of the weather. Its influence on culture and culture change is important, but difficult to assess. In West Asia, climate change has most recently been implicated in both agricultural origins (Hillman et al. 2001; Moore and Hillman 1992) and a late third-millennium collapse of civilization (Weiss 1997, 2000; Weiss et al. 1993). To evaluate its significance, we need dates precise enough to determine that the proposed climate shift happened before the proposed cultural change. We also need to determine whether the climate shift was big enough to affect established, traditional cultural responses to normal annual and interannual variability in a positive or negative way. A variety of complementary techniques have been applied to these questions: dendroclimatology, oxygen isotope analysis, sedimentology, palynology, and archaeobotany. Tree rings give precise dates and they reflect growing conditions well. Unfortunately, those data for West Asia are not yet good enough to be useful (Kuniholm 1990). Paleoclimatologists have many other ways of reconstructing climate with proxy data from natural sediments, though each technique has its own uncertainties. Some lines of evidence are relatively direct indicators of climate, such as oxygen isotope ratios.1 Other materials give more direct evidence of vegetation and plant cover, which, at a subcontinental scale, reflects climate. For example, rapid sediment deposition in the Persian Gulf is evidence of erosion upstream, which implies

120

Vegetation Proxy Data and Climate Reconstruction

121

bare ground in the Tigris-Euphrates watershed, which, in turn, has been used to infer a late third-millennium drought (Aqrawi 2001; Cullen et al. 2000; see also deMenocal 2001). Pollen gives evidence of vegetation, but has its own interpretive issues that limit its direct application to climate reconstruction. There are two obvious ways a pollen assemblage is not a direct representation of vegetation cover and an even less precise indicator of climate: pollen production and dispersal vary among plant taxa. For example, wind-pollinated plants produce far more pollen than insectpollinated plants, and some pollen is local to a lake, some washes in, and some blows in from a long distance (see Bottema 1997). Plant macroremains from archaeological sites complement pollen, because archaeological visibility is based on use and usefulness to people rather than on pollen production (Miller 1997). On a broad scale of time and space, vegetation data are very good proxies for tracking climate changes. For example, pollen cores from Lake Zeribar in the central Zagros register the retreat of cold, dry steppe and the spread of pistachio and oak forest at the end of the last Glacial period (Van Zeist and Bottema 1977). But if you want to limit the time frame, specify a location, or consider individual species, ambiguities are unavoidable. One problem is that different plant taxa respond differently to climate shifts, sometimes depending on what is (or is not) already growing in a place.2 In the Zagros, both pistachio and oak increase at the end of the Pleistocene, but at different rates. At what point does the pistachio-oak forest become the oak-pistachio forest? Whole categories of plants may respond at different rates to changed conditions. For example, grasses established themselves before trees. Archaeologists should remember that climate change does not alter whole ecosystems, but rather affects individual organisms and groups of organisms at different rates. When one turns from geological to archaeological time, it becomes much more difficult to correlate the dates of the different lines of evidence. And the time scale at which climate variability occurs is important, because people respond to actual conditions in particular places, not theoretical or average ones. Even when some variables cannot be measured, analyses should not gloss over considerations such as the time scale and synchronicity of the presumed climate shift, its spatial scale, and its magnitude. Time, space, and degree all have bearing on the magnitude of presumed cultural shift. And reasonable archaeologists can and do disagree about what constitutes a major break in the archaeological record.

The Younger Dryas Sometimes, botanical data are good proxies for climate. The Younger Dryas is characterized as a cold, dry period that occurred after the climate

122

case studies in archaeobotany

had begun to ameliorate at the end of the last Glacial period. In West Asia, its impact on vegetation has been documented from the Mediterranean to the Zagros (Baruch and Bottema 1999; Stevens et al. 2001), and appears to have occurred before human activities had significantly altered the vegetation (figure 6.1). Lemcke and Sturm (1997) see the Younger Dryas in the ∂18O/16O ratios of the Lake Van cores. At Lake Zeribar, in the central Zagros of Iran, Stevens, Wright, and Ito (2001, 753) see the Younger Dryas in enriched ∂18O comparable to Lake Van. It may not have taken long for climate to become a dependent variable, however: a paleoclimatologist, William F. Ruddiman (2003), has identified a nonnatural increase in the greenhouse gases methane and carbon dioxide as early as 7000 (cal) BC that he attributes to rice agriculture in South Asia3 (for more in-depth discussion of rice cultivation in Asia, see chapters by Madella and by Sato in this volume). Neil Roberts (2002) proposes that human activities—burning in this case—could have slowed forest advance in West Asia at the end of the Pleistocene. And Yasuda, Kitagawa, and Nakagawa (2000) have identified forest clearance associated with early agriculture in a pollen core from Lake Ghab, Syria. So the question of causality seems to be getting more complicated.

Third-Millennium Drought Ancient texts suggest that nomad invasions precipitated the collapse of the Akkadian empire of Mesopotamia at the end of the third millennium BC. Deteriorating climate across Eurasia could explain why nomads were moving and how agriculturally weakened polities of the rainfall agriculture zone might have been the first to succumb. Changing stream flows and weather patterns affecting the irrigation agriculture areas would have disrupted the economy and society of lower Mesopotamia (Weiss 1997, 2000; Weiss et al. 1993). In the past decade, evidence has been emerging that might support the argument that parts of West Asia experienced a sudden, severe, widespread drought at about 2200 BC (calendar years). On the other hand, based on critical reading of the epigraphic evidence and archaeological arguments for the ceramic dating, Richard Zettler (2003) questions the severity of the proposed collapse. Drought or not, he concludes that the evidence does not support arguments for radical settlement shifts in northern Mesopotamia at that time. The scientific evidence for drought at 2200 BC includes isotope studies at Lake Van (Lemcke and Sturm 1997) and arguably at Lake Zeribar, where Stevens, Wright, and Ito (2001) see a Late Holocene dry spell between 4000 and 3500 BP (approximately 3350–3150 cal BC to 1775– 1875 cal BC; extrapolated from Stuiver et al. 1998); they comment that

Vegetation Proxy Data and Climate Reconstruction

123

BLACK SEA

L. Van L. Urmiya

CASPIAN SEA

L. Zeribar

s gro Za

is

r Tig

Kurban Höyük Hacınebi Sweyhat Abu L. Ghab Hureyra

un

Mo

ph

n tai

Eu

s

ra tes

Malyan 0

400 km

Figure 6.1.  Map showing the location of places mentioned in the text.

the “timing of this event appears to be coeval with the abandonment of farming sites in northern Mesopotamia during the Akkadian empire . . . although the large errors associated with our dates preclude a more robust correlation” (753). As mentioned above, increased sedimentation rates in the Persian Gulf date to this time (see also Kay and Johnson 1981 for synthetic discussion of Tigris-Euphrates streamflow). For the eastern Mediterranean, Rosen (1995) discusses the social implications of drought in the region with an innovative study of phytoliths. Yet, Bottema and Cappers (2000) see no pollen evidence for a third-millennium drought in the Lake Ghab core from northern Syria, and Bottema (1997) reaches a similar conclusion in his discussion of the Zeribar core. Rather, they see vegetation responding to human activities. Archaeobotanical evidence could point to third-millennium deforestation in some places, if you accept the premise that the proportion of charred seeds to charred wood is an indicator of dung fuel to wood fuel (Miller 1997). But drought is not the only conceivable cause of deforestation.

124

case studies in archaeobotany

The third millennium is the time of first city-states of Mesopotamia. Irrigation technology, attendant population increase, and the development of fuel-intensive technologies such as bronze metallurgy are hallmarks of the period. Ax-wielding people can cause deforestation; so, if people switch from wood fuel to dung, is it because the climate dried up or because they cut down the trees? The climate of the rainfall agriculture zone of northern Mesopotamia follows “the general Mediterranean pattern, characterised by cool to mild, rainy winters and hot and dry summers” (Zohary 1973, 27). A farmer’s first line of defense against drought would be to grow more droughttolerant crops, like barley. As you go from the wetter north to the drier south, barley does become more important relative to wheat (figure 6.2). The Kurban Höyük sequence spans the period of the postulated drought of 2200 BC. If anything, an increased emphasis on barley occurred in the mid-third millennium, before the great drought (Miller 1997).4 Deforestation as indicated by an inferred increase in the use of dung fuel is also dated to the mid-third millennium (Miller 1997, figure 7.3). It is perhaps no accident that the mid-third millennium is also the time of maximum population and settlement in the region (Miller 1997, 129; Wilkinson 1990). In a nutshell, there are indeed widespread changes in vegetation, but they are not uniform in their causes or effects. Both the timing and the severity of the proposed drought are also at issue (Charles and Bogaard 2001, 325–26; Miller 2004). Two alternative hypotheses may explain apparent vegetation changes at about 2200 BC. First: over a vast expanse of the Eurasian continent there was a period of drought severe enough to cause serious deforestation, loss of even herbaceous plant cover, and consequent erosion; the ramifications for people were movement of nomads to greener pastures and subsequent stress on urban civilization in Mesopotamia. Second: most of the changes in the environmental record can be explained by human activity. The truth probably lies somewhere in between. That is, there is some evidence for drought, but the resilience of local people and economies in northern Mesopotamia seems to have allowed them to adjust, at least at a scale that is archaeologically visible.

Malyan The site of Malyan in the southern Zagros of Iran provides a comparison with northern Mesopotamia. At about 3000 BC Malyan was the largest site in the Kur River basin. The settlement covered about 45 ha, and survey estimates for the aggregate area of rural settlement in the valley are about 30 ha. Archaeological survey and excavation at Malyan show there

Vegetation Proxy Data and Climate Reconstruction Late 4th

Early 3rd

Mid 3rd

125

Late 3rd/Early 2nd

Kurban

Hacinebi

Sweyhat

Barley Wheat

Figure 6.2.  Proportions of wheat and barley in the rainfall agriculture zone of the Euphrates River. The total weight of identified cereal grains is low. Kurban Höyük: late 4th millennium—2.81 g; early 3rd—9.11 g; mid-3rd—6.03 g; late 3rd/early 2nd—0.18 g. Hacinebi: late 4th—2.77 g. Sweyhat: late 3rd/early 2nd—4.56 g. (Source: Miller 1997, fig. 7.6.)

was almost no settlement for about 400 years in the middle of the third millennium BC (Sumner 1989; 2003, 54–55). Then, at about 2200 BC, Malyan once again became the center of a much larger and more complex settlement system—with an estimated area of 130 ha for the city and 278 ha for the valley as a whole (Sumner 1990). The ceramics of the earlier and later periods are quite different from each other, but one small excavation unit has produced a few sherds consistent with a transitional assemblage that suggests there was some cultural continuity (Miller and Sumner 2004). Beyond the Kur River basin, a paucity of settlement sites in the middle of the third millennium BC extends across the southern and central Zagros generally. Archaeologically speaking, does absence of settlement mean absence of people here? The presence of nomadic pastoralists is the most reasonable explanation for this pattern (Miroschedji 2003, 24). From historic times to the present, Qashqa’i nomads have passed through the valley on their way to summer pasture, though we do not know for certain how far back that pattern of transhumance goes. For the third millennium, the closest site

126

case studies in archaeobotany

of any significance is at Jalyan, about 200 km to the south of Malyan. It is one of few archaeological sites thought to have been left by nomads, but it was a cemetery, not a settlement (Miroschedji 1974). Thus, independent archaeological evidence points to a nomadic population occupying the region continuously during the third millennium. Archaeobotanical evidence can be used to address specific questions about landscape and climate as well as about the impact of the nomadic population on the vegetation in the region. The climate regime of the Zagros Mountains of western Iran is similar in some respects to that of the mountains of the eastern Mediterranean, with a pattern of wet winters and dry summers, but it is more continental, with cooler winters (Zohary 1973, 37). As rainfall decreases (toward the south and at lower elevations), pistachio-almond forest replaces oak. The Kur River basin straddles the border between the oak forest to the north and west and the pistachio-almond forest to the south and east (figure 6.3; see Zohary 1973, map). The valley bottom, on which Malyan is located, is now largely cultivated. Remnant pistachio and almond grow on the edges at the north end of the valley. Two of the other woods, common in the archaeological samples but not in the present forest, are maple and juniper. Maple is a component of both forests. Juniper today occurs rarely in scattered locations in this part of Iran; it is also absent or rare in the pollen record of the central Zagros. The most likely species, Juniperus excelsa Bieb. (=J. polycarpos C. Koch), is probably a cold- and droughtresistant type (Sabeti 1966; Zohary 1973, 351) that is an “excellent fuel and said to yield good charcoal” (Townsend and Guest 1966, 93). The distribution across time in the major forest wood types shows some patterning (table 6.1, figure 6.4). The data come from counts of about 1,500 identified pieces of the handpicked charcoal (Miller 1982). In general, the counts and weights of handpicked and flotation charcoal are correlated (Miller 1985). A plausible assumption essential to this analysis is that the main cost of wood fuel is transport, so all things being equal, trees growing closer to a settlement will be cut first for fuel. A significant reduction in the proportion of juniper appears to have occurred between the Banesh and Kaftari settlement periods. The 400-year gap would have been long enough for any climate change to have an archaeologically observable effect on the vegetation. So, could the decline of drought-resistant juniper signify severe drought at that time? Judging from the increase in oak, which is a tree of the moister northern woodlands, probably not. Conversely, could the decline in juniper indicate a mid-third-millennium moisture increase so great that juniper could no longer compete with more moisture-loving plants? This, too, seems unlikely, because poplar/willow and hackberry, which grow in fairly moist places along streams in the Kur River basin, also decline in the later

Vegetation Proxy Data and Climate Reconstruction

127

Malyan

Persian Gulf 0

100

200 km

Oak Pistachio-almond IRAN

Central steppe Tropical savanna & desert Halophytic vegetation

Figure 6.3. Vegetation zones of southwestern Iran. (Source: Zohary 1973.)

levels (table 6.1). In any case, no one has proposed climate change for the mid-third millennium. If climate was not a factor in the juniper decline, perhaps people had something to do with it. Although absence of settlement characterizes the mid-third millennium in the region, archaeological traces of nomadic pastoralists are notoriously hard to find. Maybe what we are seeing here is the otherwise invisible impact of nomads on the landscape: juniper is slow growing, is unlikely to grow back from a stump, and is a good fuel; its fruits and shoots might be eaten by sheep and goats,5 and it might have been preferentially cut. In contrast, the nut-producing trees (almond, pistachio, and oak) might have been protected by the herders, and in any case can grow from stumps if nibbled. But juniper, already under stress from fuel cutting, could not recover from grazing pressure.

128

case studies in archaeobotany Middle Banesh

Late Banesh

Earlier Kaftari

Later Kaftari

Juniper

Oak

Almond

Pistachio

Maple 3400 BC

2900 BC

2600 BC

[gap]

2200 BC

1600 BC

Figure 6.4.  Major forest woods at Malyan.

Assuming the present-day woodland vegetation is an indicator of climatically established past distribution, the increase in oak during the Banesh period and contemporaneous decline of juniper, pistachio, and almond would represent an expanding radius of fuel procurement. Today, the boundary between the pistachio-almond and oak vegetation lies just at the northern edge of the valley. One could make the argument that, initially, juniper (along with pistachio and almond) grew on the valley bottom, but even the small urban system of the Banesh period put some pressure on wood resources.

Vegetation Proxy Data and Climate Reconstruction

129

Table 6.1. Identified charcoal from Malyan based on samples that can be assigned to phase Middle Banesh

Late Banesh

Earlier Kaftari

Later Kaftari

Number identified (1,491)

490

140

731

130

Number of samples (62)

 24

  9

 19

 10

Juniper (Juniperus sp.)

0.34

0.22

0.06

Oak (Quercus sp.)

0.01

0.24

0.22

0.35

Maple (Acer sp.)

0.03

0.04

0.16

0.03

   0

Almond (Prunus sp.)

0.16

0.06

0.15

0.21

Pistachio (Pistacia sp.)

0.11

0.01

0.23

0.04

Poplar/willow (Populus/Salix)

0.27

0.06

0.01

0.02

Hackberry (cf. Celtis sp.)

0.05

0.33

0.05

0.04

Other

0.03

0.04

0.10

0.31

Source: Miller 1982. Note: Banesh samples assigned to stratigraphically and ceramically defined subphases; temporal assignment of Kaftari samples based on relative stratigraphy (William M. Sumner, pers. comm., 5 March 2004). Measurements in the main body of the table are percentages expressed over 1 rather than 100.

The proportions of almond do not show much change over the long term; one imagines it to be a protected tree, or perhaps it can withstand trimming and browsing. The edible nuts have an intense almond flavor with a strongly bitter aftertaste. I have seen wild almond (probably Prunus scoparia [Spach] C.K. Schneid.) grow from unnoticeable to large shrubs within twenty-five years (between 1978 and 2003), just north of Shiraz; from the hillslopes down to the highway, there used to be a completely degraded landscape with no plant growing taller than about a half meter; wild almond is now thriving there, probably because roadside development has made it inaccessible to grazers. The pistachio distribution is interesting, too. It looks like both the earlier and later urban systems stressed the trees, but that pistachio was able to regenerate during the 400-year gap in settlement. I have no neat explanation for the apparent changes in the amounts of maple; perhaps it replaced juniper during the 400-year settlement gap and, having no use as food, was cleared during the more populous Kaftari period. As is true for all of these examples, one point is well worth remembering: whether influenced by climate or human activities, each species within an ecosystem has its own trajectory.

130

case studies in archaeobotany

Conclusions The interactions of people, plants, and climate are complex. Even though vegetation responds to changing climate conditions, individuals and taxa do so at different rates. Vegetation responds to human manipulation of the landscape, too; and one should also bear in mind that vegetation and human activities can influence climate. Several specific conclusions that apply to climate and landscape in the ancient Near East can be drawn from this assessment of vegetation data as a proxy for climate: First, from an archaeological perspective, the most important result of the Malyan study is that the decline in juniper that seems to have occurred during the 400-year settlement gap in the Kur River basin was not caused by climate change. Rather, this evidence supports the idea that pastoral nomads were a significant presence in the mid-third-millennium Zagros. In particular, archaeobotanical data suggest that those populations had a major impact on the landscape. More broadly, climate has short- and long-term fluctuations to which people must adapt. Farmers in West Asia have always dealt with uncertainty and year-to-year variability in the weather. In both the short and the long term, crop choice, irrigation, and mobility are how they deal with it. By the third millennium BC, people in the Near East were clearly having an impact on the vegetation. Finally, from a paleoclimatological perspective, the comparison of the cultural sequence in southern Iran and northern Mesopotamia should raise at least some questions about the geographical extent of the great drought of 2200 BC. Even if there were some drying trend in northern Mesopotamia at that time, archaeologists and paleoclimatologists should probably maintain a healthy skepticism of each other’s methods and explanatory schemes.

Acknowledgments I would like to thank Lindsay Shafer, Museum Applied Science Center for Archaeology, University of Pennsylvania Museum of Archaeology and Anthropology, who prepared the illustrations.

Notes 1. The 18O isotope is heavier than the more common oxygen, 16O. Lighter isotopes evaporate first and fall in precipitation last, but condensation processes are affected by temperature. The 18O-depleted sediments are associated with colder temperatures or winter rainfall (Stevens, Wright, and Ito 2001). Lemcke and

Vegetation Proxy Data and Climate Reconstruction

131

Sturm (1997) use oxygen isotope ratios as proxies for humidity in their analysis of cores from Lake Van. 2. Lewin (1985) discusses a study by Kenneth Cole of packrat middens that suggested that the observed lag between climate change and vegetation is a result not just of the inherent immobility of plants but also of the fact that the existing vegetation may limit the spread of plants more suited in principle to the new climate—a phenomenon called “vegetational inertia.” See also Von Holle, Delcourt, and Simberloff (2003). 3. He suggests an uncalibrated radiocarbon date of about 8000 BP (Ruddiman 2003, 273). 4. Charles and Bogaard (2001) reach a similar conclusion based on the continuing use of the relatively moisture-demanding free-threshing wheat throughout the sequence at Tell Brak, in northeastern Syria. 5. Juniper is probably not preferred if other plants are available. In Texas, where juniper control is a problem, a rancher reported that “not all goats will eat juniper . . . the first step in using goats as a cedar [i.e., Juniperus] control tool is to identify which animals in the herd have a taste for the juniper berry” (Briskin 2003). In a Texas rangeland study, Cory (1927) found that “the fruit of both species [J. monosperma and J. utahensis] is palatable to goats.” In England, sheep eat the seedlings in winter, not summer (Fitter and Jennings 1975). The Kur River Basin is on the Qashqa’i migration route, closer to the summer pastures (Beck 1991, map 3), but William M. Sumner has seen winter camps about 14 km northeast of Malyan (W. M. Sumner, pers. comm., 18 March 2004). If the animals eat the fruit, that could limit the ability of juniper to reproduce. It is likely that the cones (juniper “berries”) would have been ripe during the spring migration to the summer pastures.

References Aqrawi, A. A. M. 2001. “Stratigraphic Signatures of Climatic Change during the Holocene Evolution of the Tigris-Euphrates Delta, Lower Mesopotamia.” Global and Planetary Change 228:267–83. Baruch, U., and S. Bottema. 1999. “A New Pollen Diagram from Lake Hula: Vege­tational, Climatic, and Anthropogenic Implications.” In Ancient Lakes: Their Cultural and Biological Diversity, edited by H. Kawanabe, G. W. Coulter, and A. C. Roosevelt, 75–86. Ghent: Kenobi Productions. Beck, L. 1991. Nomad: A Year in the Life of a Qashqa’i Tribesman in Iran. Berkeley: University of California Press. Bottema, S. 1997. “Third Millennium Climate in the Near East Based upon Pollen Evidence.” In Third Millennium BC Climate Change and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 489–515. Berlin: Springer-Verlag.

132

case studies in archaeobotany

Bottema, S., and R. T. J. Cappers. 2000. “Palynological and Archaeobotanical Evidence from Bronze Age Northern Mesopotamia.” In Rainfall and Agriculture in Northern Mesopotamia, edited by R. M. Jas, 38–70. Istanbul: Nederlands Historisch-Archaeologisch Instituut te Istanbul. Briskin, J. 2003. “Goats Used to Control Juniper and Restore Rangeland.” http:// www.countryworldnews.com/news-archives/CTX/2003/ct0918rangeland.php. Charles, M., and A. Bogaard. 2001. “Third-Millennium BC Charred Plant Remains from Tell Brak.” In Excavations at Tell Brak, vol. 2, Nagar in the Third Millennium BC, by D. Oates, J. Oates, and H. McDonald, 301–26. London: British School of Archaeology in Iraq. Cory, V. L. 1927. Activities of Livestock on the Range. Texas Agricultural Experimental Station Bulletin 367. Texas A&M University, College Station, TX. Cullen, H. M., P. B. deMenocal, S. Hemming, G. Hemming, F. H. Brown, T. Guilderson, and F. Sirocko. 2000. “Climate Change and the Collapse of the Akkadian Empire: Evidence from the Deep Sea.” Geology 28:379–82. deMenocal, P. B. 2001. “Cultural Responses to Climate Change during the Late Holocene.” Science 292:667–73. Fitter, A. H., and Jennings, R. D. 1975. “The Effects of Sheep Grazing on the Growth and Survival of Seedling Juniper (Juniperus communis L.).” Journal of Applied Ecology 12:637–42. Hillman, G., R. Hedges, A. Moore, S. Colledge, and P. Pettitt. 2001. “New Evidence of Lateglacial Cereal Cultivation at Abu Hureyra on the Euphrates.” The Holocene 11:383–93. Kay, P. A., and D. L. Johnson. 1981. “Estimation of Tigris-Euphrates Streamflow from Regional Paleoenvironmental Proxy Data.” Climatic Change 3:251–63. Kuniholm, P. I. 1990. “Archaeological Evidence and Non-Evidence for Climatic Change.” Philosophical Transactions of the Royal Society of London A 330:645–55. Lemcke, G., and M. Sturm. 1997. “∂18O and Trace Element Measurements as Proxy for the Reconstruction of Climate Changes at Lake Van (Turkey): Preliminary Results.” In Third Millennium BC Climate Change and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 653–78. Berlin: Springer-Verlag. Lewin, R. 1985. “Plant Communities Resist Climatic Change.” Science 228: 165–66. Miller, N. F. 1982. “Economy and Environment of Malyan, a Third-Millennium BC Urban Center in Southern Iran.” PhD diss., University of Michigan, Ann Arbor. ———. 1985. “Paleoethnobotanical Evidence for Deforestation in Ancient Iran: A Case Study of Urban Malyan.” Journal of Ethnobiology 5:1–21. ———. 1997. “Farming and Herding along the Euphrates: Environmental Constraint and Cultural Choice (Fourth to Second Millennia BC).” MASCA Research Papers in Science and Archaeology 14:123–32.

Vegetation Proxy Data and Climate Reconstruction

133

———. 2004. “Long-Term Vegetation Changes in the Near East.” In The Archaeology of Global Change: The Impact of Humans on Their Environment, edited by C. L. Redman, S. R. James, P. R. Fish, and J. D. Rogers, 130–40. Washington, DC: Smithsonian Institution Press. Miller, N. F., and W. M. Sumner. 2004. “The Banesh-Kaftari Interface: The View from Operation H5, Malyan.” Iran 42:77–89. Miroschedji, P. 1974. “Tépé Jalyan, une nécropole du IIIe millénaire av. J.-C. au Fars Oriental (Iran).” Arts Asiatiques 30:19–64. ———. 2003. “Susa and the Highlands: Major Trends in the History of Elamite Civilization.” In Yeki Bud, Yeki Nabud: Essays on the Archaeology of Iran in Honor of William M. Sumner, edited by N. F. Miller and K. Abdi, 17–38. Los Angeles: Cotsen Institute of Archaeology, University of California. Moore, A. M. T., and G. C. Hillman. 1992. “The Pleistocene to Holocene Transition and Human Economy in Southwest Asia: The Impact of the Younger Dryas.” American Antiquity 57:482–94. Roberts, N. 2002. “Did Prehistoric Landscape Management Retard the PostGlacial Spread of Woodland in Southwest Asia?” Antiquity 76:1002–10. Rosen, A. M. 1995. “The Social Response to Environmental Change in Early Bronze Age Canaan.” Journal of Anthropological Archaeology 14:26–44. Ruddiman, W. F. 2003. “The Anthropogenic Greenhouse Era Began Thousands of Years Ago.” Climatic Change 61:261–93. Sabeti, H. 1966. Native and Exotic Trees and Shrubs of Iran [in Persian]. Tehran: University of Tehran. Stevens, L. R., H. E. Wright Jr., and E. Ito. 2001. “Proposed Changes in Seasonality of Climate during the Lateglacial and Holocene at Lake Zeribar, Iran.” The Holocene 11:747–55. Stuiver, M., P. J. Reimer, E. Bard, J. W. Beck, G. S. Burr, K. A. Hughen, B. Kromer, G. Mccormac, J. van der Plicht, and M. Spurk. 1998. “INTCAL98 Radiocarbon Age Calibration, 24,000–0 cal BP.” Radiocarbon 40:1041–83. Sumner, W. M. 1989. “Anshan in the Kaftari Phase: Patterns of Settlement and Land Use.” In Archaeologia Iranica et Orientalis: Miscellanea in Honorem Louis Vanden Berghe, edited by L. De Meyer and E. Haerinck, 135–61. Ghent: Peeters Press. ———. 1990. “An Archaeological Estimate of Population Trends since 6000 BC in the Kur River Basin, Fars Province, Iran.” In South Asian Archaeology 1987, edited by M. Taddei, 1–16. Rome: IsMEO. ———. 2003. Early Urban Life in the Land of Anshan: Excavations at Tal-e Malyan in the Highlands of Iran. Malyan Excavation Reports, vol. 3. Philadelphia: University of Pennsylvania Museum of Archaeology and Anthropology. Townsend, C. C., and E. Guest, eds. 1966. Flora of Iraq. Vol. 2. Baghdad: Ministry of Agriculture. Van Zeist, W., and S. Bottema. 1977. “Palynological Investigations in Western Iran.” Palaeohistoria 19:19–85.

134

case studies in archaeobotany

Von Holle, B., H. R. Delcourt, and D. Simberloff. 2003. “The Importance of Biological Inertia in Plant Community Resistance to Invasion.” Journal of Vege­ tation Science 14:425–32. Weiss, H. 1997. “Late Third Millennium Abrupt Climate Change and Social Collapse in West Asia and Egypt.” In Third Millennium BC Climate Change and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 711–23. Berlin: Springer-Verlag. ———. 2000. “Beyond the Younger Dryas: Collapse as Adaptation to Abrupt Climate Change in Ancient Western Asia and the Eastern Mediterranean.” In Environmental Disaster and the Archaeology of Human Response, edited by G. Bawden and R. M. Reycraft, 75–99. Albuquerque: Maxwell Musuem of Anthropology. Weiss, H., M. A. Courty, W. Wetterstrom, F. L. Guichard Sr., R. Meadow, and A. Curnow. 1993. “The Genesis and Collapse of Third Millennium North Mesopotamian Civilization.” Science 261:995–1004. Wilkinson, T. J. 1990. Town and Country in Southeastern Anatolia. Vol. 1, Settlement and Land Use in the Lower Karababa Basin. Oriental Institute Publications, vol. 109. Chicago: The Oriental Institute. Yasuda, Y., H. Kitagawa, and T. Nakagawa. 2000. “The Earliest Record of Major Anthropogenic Deforestation in the Ghab Valley, Northwest Syria: A Palynological Study.” Quaternary International 73/74:127–36. Zettler, R. L. 2003. “Reconstructing the World of Ancient Mesopotamia: Divided Beginnings and Holistic History.” Journal of the Economic and Social History of the Orient 46:3–45. Zohary, M. 1973. Geobotanical Foundations of the Middle East. Stuttgart: Gustav Fischer.

C h ap t e r 7

Significance of Prehistoric Weed Floras for the Reconstruction of Relations between Environment and Crop Husbandry Practices in the Near East Simone Riehl

One of the greatest challenges in the consideration of ancient plant remains from archaeological sites is the assessment of their meaning for economy and ecology. It was often argued that carbonized wild plant floras from cultural layers in archaeological sites are not suited for environmental reconstruction, because the taxa mainly derive from crop processing and therefore belong to the prehistoric crop weed flora (Jones 1992; Jones et al. 1999; Jones, Bogaard, and Charles 2000; Van Zeist 1993, 1999). Therefore, ecological consideration of wild plant taxa has mainly focused on the reconstruction of crop husbandry systems (Bogaard et al. 1998, 1999; Bogaard, Jones, and Charles 2001; Charles, Jones, and Hodgson 1997). While there is no doubt that weeds usually represent the largest proportion of archaeobotanical assemblages, and that they reflect a specific ecological range that is directly associated with crop husbandry practices, it is important to note that there are several other sources of origin of plant material in archaeological sites. These include the intensively

135

136

case studies in archaeobotany

investigated dung remains that have been recognized and identified by archaeobotanists since the 1980s and that contain a broad spectrum of different wild plant taxa and thus a wide range of different ecotypes (Miller 1984; Charles 1998). It is impossible to directly determine the relationship of one plant taxon to another, but a careful consideration of specific associations among plant taxa from many sites may illuminate the way archaeological sites and their plant remains are associated with each other and related to other factors. In this chapter, the working hypothesis is that wild plant floras contain information beyond that of crop husbandry practices. Impressive cultural shifts in the Near East and the eastern Mediterranean are evident from textual and archaeological sources for the whole Bronze and Iron Ages (ca. 3500–600 BC). This time span encompasses the late Bronze Age expansion of the Hittite Empire from central Anatolia to the periphery in western Anatolia and northern Syria, as well as its decline. Warfare was a very important element, but adoption of religious and cultural ideas from other political systems also took place. Sociopolitical dynamics may have also affected the economic systems, eventually even crop plant production, but the recognition of such factors in the archaeological or archaeobotanical record is problematic. As Miller points out (this volume), there are various methodological problems in using the presence of plant taxa as climate proxies. In her contribution on the linkage between climate change and the late thirdmillennium collapse of civilization in the Near East, she reveals that culture change is sometimes attributed to climate change without integration of the results from other paleoenvironmental studies and without careful consideration of the variables of climate change, such as time, space, and degree, demonstrating the need for more interdisciplinary research. Comprehensive comparison of archaeological sites from different geographic and ecological regions is necessary to address the question of whether wild plant floras from cultural layers in archaeological sites do reflect any ecological pattern independent from their function as weeds. Usually pollen and wood remains are used for environmental reconstructions exclusively, ignoring the high potential of wild plant taxa from prehistoric carpological assemblages in providing environmental information. The tentative results presented here (based on a small selection of sites and the limited time scale of late Bronze and early Iron Age) are part of a larger project on the evolution and development of prehistoric wild plant floras in the Near East. They aim to compare late Bronze and early Iron Age sites on the basis of their remains of wild plant taxa to contribute to the reconstruction of prehistoric environmental conditions.

Significance of Prehistoric Weed Floras

137

Evidence for the high potential of wild plant taxa from archaeological excavations as proxy records for environmental conditions and change is presented.

Methods and Methodological Problems Environmental change can be caused either by people or by climate shift. And, as Miller stresses (this volume), clearance of woodland might be reflected in the archaeobotanical record as a switch from wood fuel to dung or, as often argued in older works with far-reaching conclusions, as a change in preference for more drought-tolerant crops in combination with grain size (Hopf 1978, 1983).1 Usually it remains unclear whether environmental change occurred because the climate drifted toward an extreme or because human impact was incisive; one of the very few exceptions is the Proto-Elamite (3400–2600 BC) and Elamite (2400–1800 BC) site of Malyan (Miller, this volume). In addition, culture change by culture contact, as often visible in later prehistoric and historic societies, may also result in economic changes and alteration in food production, and may thus obscure the interpretation of changing plant spectra. It seems logical that the chances of assessing the role of specific taxa or plant groups from the archaeological record increase with the breadth and depth of data. It is necessary to consider and compare as many archaeological records as possible (a broad spectrum of diverse plant remains and taxa) over multiple chronological sequences from an area large enough to include different landscapes, climatic regimes, and cultural spheres. Given the aim of addressing such expansive questions, the urgent need and the recent popularity of developing large databases becomes intelligible (see also Alexia Smith’s contribution on the NEAD database, this volume). The association of wild plant remains and crops in a site can be interpreted in various ways, but particularly for more recent periods, wild plants are often exclusively interpreted as crop weeds. The plant spectra from some Pre-Pottery Neolithic (PPN) sites (e.g., the Epipaleolithic/ PPNA Tell site of Mureybit [Van Zeist and Bakker-Heeres 1984] or middle PPNB Tell Ghoraifé [Van Zeist and Bakker-Heeres 1982]) already contain a broad spectrum of wild plants, including typical weeds such as Chenopodium L. spp., Fumaria L. spp., Centaurea L. spp., Silene L. spp., and so on. However, nobody would interpret, for example, the find of Echinochloa crus-galli (L.) P. Beauv. at Epipaleolithic/PPNA Mureybit (10500–9600 uncal BP) as an indicator of irrigated fields, because it is known today as a typical weed of moist places, and particularly not if we

138

case studies in archaeobotany

also take into consideration that the status of domestication for emmer wheat is somewhat questionable in this site (Nesbitt 2002). The various possible interpretations for the find of a specific taxon in an archaeological site make it necessary to include the whole set of information we have, that is, the find situation of one taxon in relation to archaeological features (context), the numerical relation of one taxon to other taxa from the same sample (sample composition), as well as modern ecology of the considered taxon. Despite the knowledge of strongly differing seed production in different plant species, archaeobotanists usually take the number of seeds, often in combination with the frequency of a specific taxon at a site, as an indicator of its usefulness or dominance among the archaeobotanically represented taxa. This includes statements on the “importance” of specific crops, as well as on the harmfulness of specific weeds. Such statements, sometimes more or less justified, are made in almost every archaeobotanical report based on a large amount of samples. The approach used here is the comparison of patterns of taxa composition from different sites under consideration of diverse geographically relevant factors (bioclimatic range of the sites). Correspondence analysis (CA) was used to work out clusters of similarities and differences in the plant spectra from late Bronze and Iron Age sites in the region (including the periphery) of the New Hittite Kingdom (figure 7.1).2 CA is widely used in community ecology and archaeology (Baxter 1994; Birks 1991; Gauch 1982; Ter Braak 1987). It is sometimes applied to assess the composition of archaeobotanical assemblages (Colledge 1998; Jones 1991; Lange 1990; Riehl 1999), and in combination with a graphic software program, patterns and trends in plant assemblages through time can be plotted. The plots show the position of each sample relative to all other samples and to each taxon. Additionally, they present the relationship of each taxon to all the other taxa. Colledge (1998) investigated covariation between the taxonomic composition of samples at PPN sites in Syria. She found patterns, which she calls “vegetational fingerprints,” that are best explained as ecological groups. In particular, she demonstrated for Mureybit that “the dominant ecological groups for each period of occupation may be indicative of the most prevalent vegetation type at that time” (Colledge 1998, 128). Before applying multivariate statistics to different geographical areas of plant production (Aegean and Near Eastern sites), one has to be aware of a regional patterning caused by different crop spectra in different sites, and thus different economic systems and preferences in plant production.3 The patterning is visible not so much in the crop taxa themselves, because most of the main crops were cultivated in both areas during the Bronze and Iron Ages (hulled and naked wheats, barley, lentil, grape, etc.)

Significance of Prehistoric Weed Floras

139

Figure 7.1.  Map of the research area and the sites referred to in the text: (1) Tiryns, (2) Kastanas, (3) Assiros Toumba, (4) Troy, (5) Beycesultan, (6) Kuşakli, (7) ‘Ain Dara, (8) Tilbeshar, (9) Tell Shiuk Fawqani, (10) Umm el-Marra, (11) Emar, (12) Tell Munbaqa, (13) Tell Hadidi, (14) Tell Bderi, (15) Tell Schech Hamad.

(Riehl and Nesbitt 2003). But regional patterning is visible in the proportions of specific crops. The continuing dominance of hulled wheats in the Aegean during the Bronze and Iron Ages is not visible in northern Syrian sites. In contrast, free-threshing wheat reaches much higher proportions in the Near East, and the two-row variety of barley was also intensively cultivated at some Syrian sites. Bearing this in mind, one has to consider that statistical patterning of the whole archaeobotanical data set of a site will be attributed to differences in the presence or actual counts of crop taxa. Therefore, analyses of the data set were conducted several times with slightly differing data sets (with the whole data set, including the crop taxa; a smaller data set limited to the wild plant taxa; and an even more reduced data set that excluded weeds probably associated with the dominant crops). Another methodological problem is a missing linkage of botanical samples to the fine chronology of the site or an incorrect chronological designation of archaeobotanical samples, which happens often but is hard for the archaeobotanist to demonstrate. An additional and general problem for the analysis of data from a broad geographical area is that many sites have little dating information, and some sites have no 14C dates at all. A further factor exacerbating the difficulties of evaluating assemblages within this chronological framework is the underrepresentation of

Late Bronze Age Late Assyrian period (900–700 BC) Middle Assyrian period (1275–1075 BC) Late Bronze Age II Iron Age I Iron Age II Late Bronze Age Late Helladic IIIB Late Helladic IIIC Iron Age? Late Bronze Age (1250–1180 BC) Iron Age (1150–1000 BC) Late Bronze Age

22 15 15 18 6 20 13 29 82 11 23 12

TMun TSH-LA TSH-MA TSF_BA2 TSF_IA1 TSF_IA2 Tb_LBA T-LHIIIB T-LHIIIC Tro_VIIa Tro_VIIb2 UeM_II

Tell Munbaqa Tell Schech Hamad Tell Schech Hamad Tell Shiukh Fawqani Tell Shiukh Fawqani Tell Shiukh Fawqani Tilbeshar Tiryns Tiryns Troia Troia Umm el-Marra

Note: See figure 7.1 for the location of these sites. a Work in progress; see http://perso.wanadoo.fr/g.willcox/.

Iron Age Late Bronze Age I Late Bronze Age II Late Bronze Age Late Bronze Age Younger late Bronze Age Earlier Iron Age (“wood building phase”) Earlier Iron Age (“loam building phase”) Hittite (14th/13th C BC) Hittite (“gate,” layer 2) (14th/13th C BC) Hittite (“temple,” layer 2) (14th/13th C BC) Mittani period Late Bronze Age IB (1550–1400 BC)

35 36 18 7 5 41 46 83 unknown unknown unknown 3 46

ADa AT-LBAI AT-LBAII BS-LBA Em_LBA K-4 (16-14) K-6 K-7 (8-4) Kus-II Kus-II_G Kus-II_T TB-M Had-LB

‘Ain Dara Assiros Toumba Assiros Toumba Beycesultan Emar Kastanas Kastanas Kastanas Kusakli Kusakli Kusakli Tell Bderi Tell Hadidi

Period

No. of samples

Phase code

Site name

Table 7.1. List of the sites and phases included in the analysis

Syria Syria Syria Syria Syria Syria Turkey Greece Greece Turkey Turkey Syria

Syria Greece Greece Turkey Syria Greece Greece Greece Turkey Turkey Turkey Syria Syria

Country Crawford 1999 Jones 1983 Jones 1983 Helbaek 1961 Riehl, unpub. data Kroll 1983 Kroll 1983 Kroll 1983 Pasternak 1998 Pasternak 1998 Pasternak 1998 Van Zeist 1999/2000 Van Zeist and Bakker-Heeres 1985 Küster 1989 Van Zeist 1999/2000 Van Zeist 1999/2000 Willcoxa Willcoxa Willcoxa Willcoxa Kroll 1982 Kroll 1982 Riehl 1999 Riehl 1999 Miller 2000

Author

Significance of Prehistoric Weed Floras

141

Iron Age sites in the study area, which makes comparisons with the late Bronze Age difficult. Other methodological problems concern differences in the breadth of the plant spectra from individual sites, such as the missing of plant taxa caused by limited sampling (few samples, small sample volumes, low diversity in sampled contexts). Very often this information, which is important for the decision whether or not to include a site in the analysis, is missing. Data selection for this study was conducted according to the following criteria (table 7.1). An ideal number of thirty samples per settlement phase was considered critical because it is assumed that the representation of a site within its plant remains increases with the number of samples. Exceptions were made, because an exclusion of sites with fewer samples would have resulted in too small a data set. The minimal number of taxa was set at fifteen, to maximize the number of usable sites.

First Results Some introductory remarks on correspondence diagrams will be followed by the preliminary results for a few selected sites from the research area.4 The simplified two-dimensional figures represent the original multidimensional quality of the data clustering. The ordination technique arranges the samples along the axes according to the spectra and number of taxa (variables), so “the position of each sample is shown relative to all other samples and to all the species, and of each species relative to all other species and to all the samples in the analysis” (Colledge 1998, 127, quoting Lange 1990). In CA the first principal (horizontal) axis accounts for the largest proportion of variation (highest eigenvalues) among the data. It is assumed that the produced axes represent the underlying causes of taxa variation among samples. As such, the diagrams are used to build hypotheses on the causes. Hypotheses are tested via classification of data points according to extrinsic variables, such as sample chronology and origin or taxa ecology. Interpretation of the data points in the diagram is mainly conducted according to following aspects: •• The 0,0 coordinates express the fact that samples near this intersection are similar in composition and taxa are usually frequent. •• Distance from the 0,0 coordinates shows the degree of divergence of the data points, that is, difference of sample composition from the general sample mean, or in case of taxa, its restrictedness to single samples.

142

case studies in archaeobotany

•• Vicinity of samples in the diagram indicates the degree of similarity, and vicinity of taxa indicates the degree of their association in the same samples, whereas vicinity of samples and taxa cannot be directly related to each other. Further evaluation of the original data and additional analysis (e.g., attribute plots) have to be conducted to answer such questions. •• To simplify, distance is the visualization of difference, and vicinity of data points stands in most cases for similarity or correlation. CA of all fifteen sites was conducted for the whole set of variables (crop and wild plant taxa counts), but also for the crops and for the wild plant taxa separately in order to see the role of crops and wild plant taxa in data patterning. Gathered fruits (as primarily indicating human preferences in the composition of diet) and obvious weed taxa (and thus data that would cluster like the crop data) were eliminated from the data set of the wild plants.5 The graph for the complete data set shows a clear patterning coincidental with that of the geographical position of the sites, but also gives some clues about husbandry practices (figure 7.2). The Euphrates sites Emar, Tell Shiukh Fawqani, Tell Hadidi, and Tell Munbaqa separate clearly from other Syrian sites at the Khabur (Tell Schech Hamad, Tell Bderi) and the Mediterranean coast (‘Ain Dara), as well as from the Aegean and the Anatolian sites along the first principal axis. Following the accepted opinion that these patterns are based on differing crop husbandry systems or even more simply on the occurrence and proportion of each crop taxon, we can see, by taking a closer look at the crops from each site, that the Euphrates sites separate from the rest of the samples along the first axis due to a high proportion of two-row barley (figure 7.3). A more detailed view of the samples in the left part of the diagram shows that the separation is not very clear, but gives the impression that samples cluster along the second axis according to their proportions of wheats (top of the diagram) and six-row barley (center of the diagram). The data analysis based exclusively on the crops gives a slightly different clustering than that based on the whole data set, but neither fully explains the strong geographical patterns visible in figure 7.2. Analysis of the data set consisting of wild plants (without weeds) yields clustering that is even clearer than in the previous diagrams: Aegean and Syrian sites separate along the first axis, although the position of the site ‘Ain Dara together with Aegean sites cannot be explained at that point. Looking at the ecology of the wild taxa, which was obtained from Davis (1965–88), the strong presence of taxa able to grow in saline places clustering with the Euphrates sites and the high proportion of taxa usually growing in moist but freshwater habitats in the bulk of the Aegean

Figure 7.2. Correspondence (CA) plot of the complete data set, showing a clear patterning coincidental with that of the geographical position of the sites.

Figure 7.3. Correspondence (CA) plot of the sites, showing their geographic clustering according to dominant crops represented in the sites.

144

case studies in archaeobotany

Figure 7.4. Correspondence (CA) plot of the sites and wild plant taxa, showing their geographic and ecological clustering.

sites becomes visible (figure 7.4). The clustering of the sites is easy to explain if one considers the modern mean annual precipitation (figure 7.5). Similarly, the diagram position of ‘Ain Dara, which enjoys a higher mean annual precipitation than other Syrian sites owing to its proximity to the Mediterranean coast, becomes clear. The wild plant spectrum of ‘Ain Dara is more similar to those of Aegean sites than to other Syrian sites, whereas no such strong patterns are visible in the analysis of the crop spectra alone. This also corresponds well to the potential natural vegetation as reconstructed by Hillman (in Moore, Hillman, and Legge 2000, figure 3.7). While most of the Syrian sites are located according to the potential vegetation under modern climatic conditions in a woodland steppe or moist and medium-dry steppe, with close vicinity of riverine woodland, ‘Ain Dara and the Aegean sites are situated in the Mediterranean woodland zone. Generally, the clustering of the wild plants seems to be directly related to the environmental conditions; none of the crops is exclusive to one area, so crops cannot explain the clustering totally.

Significance of Prehistoric Weed Floras

145

Figure 7.5. Correspondence (CA) plot of the sites and wild plant taxa, showing their clustering under the classification of sites according to mean annual precipitation and ecological classification of wild plant taxa.

In detail it also becomes visible from the analysis that salinity of the soils, eventually caused by small-scale irrigation of specific crops (such as pulses or grapes), should have been a problem for people living in the late Bronze and early Iron Age sites along the northern Euphrates, whereas there is only weak indication of moist and saline conditions in the Khabur basin (five records of Carex divisa Hudson for the site Tell Schech Hamad). This is surprising, as mean annual precipitation is below the amount necessary for rain-fed agriculture, and irrigation had surely been practiced. A probable cultivation in the flood plain may have avoided the problem of salinization by inundation and subsequent rinsing of salt accumulation. The high proportions of barley raise another question. Since two-row barley is today mainly cultivated in areas with marginal rainfall, one may ask with regard to the Euphrates, where there is potential for irrigation, why the inhabitants of these sites did not grow six-row barley in larger amounts? Were the salinization problems of large-scale irrigation already known, and therefore avoided where possible?

146

case studies in archaeobotany

Another noticeable environmental factor at Tell Schech Hamad is chronologically significant. Namely, during the middle Assyrian period the wild taxa indicate the presence of moist and damp habitats there (probably caused by irrigation practice), whereas during late Assyrian time no such indicators are present (figure 7.5).

Conclusions The general conclusion from the analysis of this selected data set of Bronze and Iron Age sites from a geographical area spanning from the Aegean to northern Syria is that a consideration of wild plants provides information about environmental conditions at the sites beyond that of crop husbandry. Furthermore, when nonweedy wild plants are considered separately, new patterns become visible, as was also demonstrated by Colledge (1998) for PPN sites in Syria. It is interesting to see this also confirmed for more recent sites, where one would expect technological progress in plant production to have had a great impact on the vegetation, and the wild plant taxa to be almost exclusively weeds. CA supports the hypothesis that crop and particularly wild plant spectra of archaeobotanical assemblages of Aegean and Near Eastern sites do not simply reflect crop husbandry practice. Indeed, they are also influenced by the geographical position of the sites, which determines the vegetation and climatic variables (e.g., mean annual precipitation). A striking detail of the analysis is that, supposing a direct relationship between irrigation and salinization of the soil, habitats characterized by saline soils were strongly represented in the northern Euphrates region, but only scantly represented in the Khabur region (Tell Schech Hamad and Tell Bderi; the latter site was not presented in this study). Particularly bearing the southern limit of rain-fed agriculture in northern Syria in mind (Wilkinson 2003), the obviously different distribution of saline soils in the area needs further explanation. The weakness of chronological patterning is also noteworthy (see figure 7.2), but may also be related to a lack of Iron Age sites. In the analyses the chronological factor is superimposed by the geographical factor, so Late Bronze Age and Iron Age crop spectra and proportions in one site are more similar than in one period comparing several sites. However, the analysis of the wild plant taxa demonstrated that, at Tell Schech Hamad, for example, the differences between different periods are based on the composition of the wild plants, in contrast to a more or less stable composition of the wild plant taxa in time at the Aegean sites. To test the hypothesis that there are areas with more obvious changes from one period to another, we must await the analysis of the whole data set (including 250 sites).

Significance of Prehistoric Weed Floras

147

In the CA of the single periods, the patterns already visible for all the data become clearer. The analyses separate the geographical regions of the Euphrates sites (which have high proportions of taxa typical of saline places and open habitats) from the Khabur site, Tell Schech Hamad, and from the Aegean sites, which have high proportions of plants growing near or in freshwater habitats (see also figure 7.4, but for Bronze and Iron Age data). Looking at the Iron Age data, the Aegean sites obviously show a higher proportion of woodland taxa compared to coastal Troy, whereas saline habitats are mainly represented by the Syrian sites near the Euphrates River valley. To speculate further on these results, the hypothesis should be tested whether the occasionally mentioned aridification from the late Bronze to the early Iron Age (Lemcke and Sturm 1997) was indeed an important environmental factor in the Near East, and if so, whether it is possible to further specify the environmental conditions for the settlements at the Khabur and the Euphrates.

Future Perspectives The use of prehistoric wild plant floras from archaeological sediments for environmental reconstructions is the research perspective of this chapter. Although there are still several methodological problems to solve, including that of the precise dating of the archaeobotanical assemblages, correspondence analysis has a high potential to reveal direct relations between environmental factors and prehistoric wild plant taxa. A very clear example is the clustering of the archaeological sites according to their precipitation regimes, regardless of whether irrigation was practiced or not. If wild plant remains (excluding typical weeds) directly mirrored crop husbandry practices, the flora from irrigated fields (which would consist of intruding species from the riverside flora) should place the related data points in the diagrams into the regime of higher precipitation rates. The actual patterns would not show as close a correspondence to precipitation regimes as they do. There are surely factors other than crop husbandry practices responsible for clustering. Careful dating of archaeobotanical samples will help us to refine our results on early environments. There is strong demand for research on wild and domesticated plant assemblages of late Bronze and Iron Age Near Eastern sites to address questions on economic or environmental impact factors in a politically turbulent cultural region.

148

case studies in archaeobotany

Acknowledgments This research was made possible by financial support from the Ministry for Science and Arts (Baden-Württemberg, Germany) and the German Research Council. I would like to thank the editors for giving me the opportunity to present my work in their book. The text benefited from critical comments by Naomi Miller (University of Pennsylvania Museum of Archaeology and Anthropology), Alexia Smith (Boston University), and an anonymous reviewer. Special thanks to Irina Kuzyakova (Department of Bioinformatics, University Hohenheim, Germany) for discussing statistical questions with me. Also many thanks to Naomi Miller and James Greig for correcting the English.

Notes 1. Hopf assumes free-threshing wheat to have been more important than barley at Jericho until the end of the early Bronze Age. With the transition from early to late Bronze Age, six-row barley becomes more frequent than the wheat species at Jericho. This is interpreted by Hopf to be caused by deforestation, because of a decrease in the groundwater level between 2600 and 2300 BC. The more salt-tolerant and drought-resistant barley may have grown better than the more sensitive wheat (Hopf 1983, 579). Hopf further claims that the economic basis of agriculture was destroyed by erosion in the course of the early Bronze Age, supported by a decrease in seed size of the crops. 2. Correspondence Analysis (CA) is an ordination technique for data showing unimodal environmental gradients and is therefore widely used for pattern searching in complex variable-by-sample data. The analysis was conducted with the software package Canoco 4.5 and CanoDraw by Ter Braak and Šmilauer (1998). 3. A side effect of CA on a supraregional level is that sites with archaeobotanical data not representative of a specific area for taphonomic reasons would be displayed as “outliers” in the diagram. In other words, the patterning of sites according to their geography can be used as a measure of the representativeness of their data for all the sites from that specific region. 4. Since the results presented here are part of a larger research project at the University of Tübingen still in progress, they are very preliminary. The actual database, still in development, will include about 250 sites, which will be included in the final data analysis. 5. The filtering of plant taxa was also conducted by a simple analysis of the correlation coefficient. Wild plant taxa in relation to each crop plant with a coefficient of p > 0.95 were considered to have been very likely deposited together with the crop. In some cases it was very probable that they also grew together in the

Significance of Prehistoric Weed Floras

149

field. Others were just the result of accumulation. Each taxon was separately considered for its most probable origin, based on the available floristic information from Davis (1965–88), the correlation coefficient, and additional archaeological information. The wild plant species and genera considered in the analysis are Aeluropus littoralis (Gouan) Parl., Aizoon hispanicum L., Alisma L. sp., Alkanna Tausch sp., Andrachne L. sp., Androsace L. sp., Arnebia Forsskal sp., Artemisia L. sp., Asphodelus L. sp., Astragalus L. sp., Bellevalia Lapeyr. sp., Berula erecta Hudson, Bunias erucago L., Carex divulsa Stokes, Carex divisa Hudson, Carex remota L., Chara L. sp., Chenopodium urbicum L., Chenopodium ficifolium Sm., Chenopodium murale L., Cladium mariscus (L.) Pohl, Cyperus longus L., Echinaria capitata (L.) Desf., Eleocharis R.Br. sp., Euphorbia falcata L., Euphorbia helioscopia L., Fimbristylis Vahl sp., Geranium L. sp., Glycyrrhiza L. sp., Gypsophila L. sp., Helianthemum Miller sp., Hordeum geniculatum All., Hordeum murinum L., Hyoscyamus L. sp., Hypericum L. sp., Isoëtes duriei Bory, Juncus L. sp., Lepidium L. sp., Malcolmia R.Br. sp., Myrtus communis L., Nepeta L. sp., Neslia Desv. sp., Onobrychis Adanson sp., Onopordum acanthium L., Ornithogalum L. sp., Plantago L. sp., Polypogon maritimus Willd., Polygonum aviculare L., Polygonum corregioloides L., Polygonum venantianum L., Prosopis L. sp., Raphanus raphanistrum L., Ranunculus sardous Crantz, Reseda L. sp., Rhinanthus L. sp., Rosa L. sp., Rumex L. sp., Salsola L. sp., Sambucus ebulus L., Schoenus nigricans L., Scilla L. sp., Scirpus maritimus L., Scleranthus L. sp., Scorpiurus L. sp., Scrophularia L. sp., Solanum nigrum L., Spergula L. sp., Spergularia marina (L.) Griseb., Stachys L. sp., Stellaria aquatica L., Stellaria media (L.) Vill., Suaeda L. sp., Teucrium chamaedrys L., Thalictrum lucidum L., Thymelaea Miller sp., Thymus L. sp., Typha latifolia L., Verbascum L. sp., Verbena officinalis L., Veronica hederifolia L., Ziziphora L. sp.  The reduction in taxa resulted in a decrease in the number of records, so that some sites had to be excluded from the analysis of the data set without crops and weeds because they no longer fulfilled the requirements (seventy records per site was an absolute minimum to be included in the analysis). The sites excluded were Tell Bderi, Tilbeshar, Tiryns LH IIIB, and Beycesultan.

References Baxter, M. J. 1994. Exploratory Multivariate Analysis in Archaeology. Edinburgh: Edinburgh University Press. Birks, H. J. B. 1991. “Some Selected Published Applications of Canonical Correspondence Analysis, Redundancy Analysis, and Detrended Canonical Correspondence Analysis Using Canoco.” Unpublished manuscript. Bogaard, A., J. G. Hodgson, P. J. Wilson, and S. R. Band. 1998. “An Index of Weed Size for Assessing the Soil Productivity of Ancient Crop Fields.” Vegetation History and Archaeobotany 7:17–22.

150

case studies in archaeobotany

Bogaard, A., G. Jones, and M. Charles. 2001. “On the Archaeobotanical Inference of Crop Sowing Time Using the FIBS Method.” Journal of Archaeological Science 28:1171–83. Bogaard, A., C. Palmer, G. Jones, M. Charles, and J. G. Hodgson. 1999. “A FIBS Approach to the Use of Weed Ecology for the Archaeobotanical Recognition of Crop Rotation Regimes.” Journal of Archaeological Science 26:1211–24. Charles, M. 1998. “Fodder from Dung: The Recognition and Interpretation of Dung-Derived Plant Material from Archaeological Sites.” Environmental Archaeology 1:111–22. Charles, M., G. Jones, and J. G. Hodgson. 1997. “FIBS in Archaeobotany: Functional Interpretation of Weed Floras in Relation to Husbandry Practices.” Journal of Archaeological Science 24:1151–61. Colledge, S. 1998. “Identifying Pre-Domestication Cultivation Using Multivariate Analysis.” In The Origins of Agriculture and the Domestication of Crop Plants in the Near East, edited by A. B. Damania, J. Valkoun, G. Willcox, and C. O. Qualset, 121–31. Aleppo: ICARDA. Crawford, P. L. 1999. “Botanical Remains.” In The Iron Age Settlement at ‘Ain Dara, Syria: Survey and Surroundings, edited by E. C. Stone and P. E. Zimansky, 113–21. BAR International Series, vol. 786. Oxford: British Archaeological Reports. Davis, P. H. 1965–88. Flora of Turkey and the East Aegean Islands. Edinburgh: Edinburgh University Press. Gauch, H. G. 1982. Multivariate Statistics in Community Ecology. Cambridge: Cambridge University Press. Helbaek, H. 1961. “Late Bronze Age and Byzantine Crops at Beycesultan in Anatolia.” Anatolian Studies 11:77–97. Hopf, M. 1978. “Plant Remains, Strata V-I.” In Early Arad: The Chalcolithic Settlement and Early Bronze City, edited by R. Amiran, 64–82. Jerusalem: Israel Exploration Society. ———. 1983. “Jericho Plant Remains.” In Excavations at Jericho V: The Pottery Phases of the Tell and Other Finds, edited by K. M. Kenyon and T. A. Holland, 576–621. London: British School of Archaeology in Jerusalem. Jones, G. E. M. 1983. “The Use of Ethnographic and Ecological Models in the Interpretation of Archaeological Plant Remains: Case Studies from Greece.” PhD diss., University of Cambridge. ———. 1991. “Numerical Analysis in Archaeobotany.” In Palaeoethnobotany, edited by W. van Zeist, K. Wasylikowa, and K. E. Behre, 63–80. Rotterdam: A. A. Balkema. ———. 1992. “Weed Phytosociology and Crop Husbandry: Identifying a Contrast between Ancient and Modern Practice.” Review of Palaeobotany and Palynology 73:133–43. Jones, G., A. Bogaard, and M. Charles. 2000. “Distinguishing the Effects of Agricultural Practices Relating to Fertility and Disturbance: A Functional

Significance of Prehistoric Weed Floras

151

Ecological Approach in Archaeobotany.” Journal of Archaeological Science 27:1073–84. Jones, G., A. Bogaard, P. Halstead, M. Charles, and H. Smith. 1999. “Identifying the Intensity of Crop Husbandry Practices on the Basis of Weed Floras.” Annual of the British School at Athens 94:167–89. Kroll, H. 1982. “Kulturpflanzen von Tiryns.” Archäologischer Anzeiger 1982: 467–85. ———. 1983. Kastanas: Ausgrabungen in einem Siedlungshiigel der Bronze- und Eisenzeit Makedoniens 1975–1979: Die Pflanzenfunde. Berlin: Volker Spiess. Küster, H. 1989. “Bronzezeitliche Pflanzenreste aus Tall Munbaqa.” Mitteilungen der Deutschen Orientgesellschaft 121:85–91. Lange, A. G. 1990. De Horden near Wijk bij Duurstede: Plant Remains from a Native Settlement at the Roman Frontier: A Numerical Approach. Amersfoort: ROB. Lemcke, J., and M. Sturm. 1997. “∂ 18O and Trace Element Measurements as Proxy for the Reconstruction of Climate Changes at Lake Van (Turkey): Preliminary Results.” In Third Millennium BC Climate Change and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 653–78. Berlin: Springer-Verlag. Miller, N. F. 1984. “The Use of Dung as Fuel: An Ethnographic Example and an Archaeological Application.” Paléorient 10:71–79. ———. 2000. “Plant Remains.” In Excavation and Survey in the Jabbul Plain, Western Syria: The Umm el-Marra Project 1996–1997, edited by G. M. Schwartz, H. H. Curvers, F. A. Gerritsen, J. A. MacCormack, N. F. Miller, and J. A. Weber, 438–47. American Journal of Archaeology 104 (3): 419–62. Moore, A. M. T., G. C. Hillman, and A. J. Legge. 2000. Village on the Euphrates: From Foraging to Farming at Abu Hureyra. New York: Oxford University Press. Nesbitt, M. 2002. “When and Where did Domesticated Cereals First Occur in Southwest Asia?” In The Dawn of Farming in the Near East, edited by R. T. J. Cappers and S. Bottema, 113–32. Studies in Early Near Eastern Production, Subsistence and Environment, 6. Berlin: Ex oriente. Pasternak, R. 1998. “Übersicht über die Ergebnisse der archäobotanischen Arbeiten in Kuşaklı 1994–1997 und ein Interpretationsansatz zu den Befunden.” Mitteilungen der Deutschen Orient-Gesellschaft zu Berlin 130:160–70. Riehl, S. 1999. Bronze Age Environment and Economy in the Troad: The Archaeobotany of Kumtepe and Troy. Tübingen: Mo Vince. Riehl, S., and M. Nesbitt. 2003. “Crops and Cultivation in the Iron Age Near East: Change or Continuity?” In Identifying Changes: The Transition from Bronze to Iron Ages in Anatolia and Its Neighbouring Regions: Proceedings of the International Workshop, Istanbul, Nov. 8–9, 2002, edited by B. Fischer, 301–12. Istanbul: Türk Eskiçağ Bilimleri Enstitüsü. Ter Braak, C. J. F. 1987. “The Analysis of Vegetation-Environment Relationships by Canonical Correspondence Analysis.” Vegetatio 69:69–77.

152

case studies in archaeobotany

Ter Braak, C. J. F., and P. Šmilauer. 1998. CANOCO Reference Manual and User’s Guide to Canoco for Windows: Software for Canonical Community Ordination (version 4). Ithaca: Microcomputer Power; Wageningen, Neth.: Centre for Biometry. Van Zeist, W. 1993. “Archaeobotanical Evidence of the Bronze Age Field Weed Flora of Northern Syria.” In Dissertationes Botanicae; The Zoller Festschrift: Contributions to the Philosophy and History of the Natural Sciences, Evolution and Systematics, Ecology and Morphology, Geobotany, Pollen Analysis, and Archaeobotany, edited by C. Brombacher, S. Jacomet, and J. N. Haas, 499– 511. Berlin: J. Cramer in der Gebrueder Borntraeger Verlagsbuchhandlung, E. Schweizerbart’sche Verlagsbuchhandlung. ———. 1999. “Evidence for Agricultural Change in the Balikh Basin, Northern Syria.” In The Prehistory of Food: Appetites for Change, edited by C. Gosden and J. G. Hather, 351–73. London: Unwin. ———. 1999/2000. “Third to First Millennium BC Plant Cultivation on the Khabur, North-Eastern Syria.” Palaeohistoria 41/42:111–25. Van Zeist, W., and J. A. H. Bakker-Heeres. 1982. “Archaeobotanical Studies in the Levant. 1. Neolithic Sites in the Damascus Basin: Aswad, Ghoraifé, Ramad.” Palaeohistoria 24:165–256. ———. 1984. “Archaeobotanical Studies in the Levant. 3. Late-Palaeolithic Mureybit.” Palaeohistoria 26:171–99. ———. 1985. “Archaeobotanical Studies in the Levant. 4. Bronze Age Sites on the North Syrian Euphrates.” Palaeohistoria 27:247–316. Wilkinson, T. J. 2003. Archaeological Landscapes of the Near East. Tucson: University of Arizona Press.

C h ap t e r 8

Historical Aspects of Early Plant Cultivation in the Uplands of Eastern North America Kristen J. Gremillion

For much of the twentieth century, discussions of agricultural origins in North America stuck to basic historical issues (Gremillion 1993b): when did corn and other so-called tropical cultigens bring to the region the potential for the development of advanced societies with monumental architecture? How far down the road of progress could a culture proceed on a hunter-gatherer base? In the Southeast, Caldwell (1958) took the time to consider in some depth the implications of “Primary Forest Efficiency” as an alternative adaptation that might have supported complex systems of meaning and exchange such as Hopewell. For the most part, however, serious consideration of the ecological dynamics of early food production was not to become the center of debate until issues of human adaptation to the natural environment rose to prominence in the broader discipline of anthropology. With this change, diffusion and migration lost much of their credibility as explanatory devices for major transitions in human history. Process took center stage as archaeologists labored to relate the archaeological record, culture change, and adaptation. In eastern North America, emerging interest in ecological and evolutionary theories of agricultural origins went hand-in-hand with the increasingly convincing case for the independent development of food production based on indigenous weedy species—a farming tradition that

155

156

social archaeobotany

predated the introduction of maize and other Mesoamerican crops (Smith 1987b). Morphological changes in seeds documented the domestication process in sumpweed (Iva annua L.), sunflower (Helianthus annuus L.), and, eventually, indigenous cultivated forms of goosefoot (Chenopodium berlandieri Moq.) and Cucurbita gourd (Cucurbita pepo L. ssp. ovifera [L.] D. S. Decker) (Asch and Asch 1977, 1985; Crites 1987, 1993; Fritz and Smith 1988; King 1985; Smith 1989, 1992a, 1997; Smith and Cowan 1993; Smith, Cowan, and Hoffman 1992; Yarnell 1972, 1978). Early finds of Cucurbita gourd rind and seeds lost their status as indicators of diffusion, and the millennia-long history of association of this species with human populations came to be widely accepted (Crawford 1982; Decker 1988; Smith and Cowan 1993; Smith, Cowan, and Hoffman 1992). A growing archaeobotanical database established that the cultivation of several starchy and oily seeds native to the midcontinent was widespread in the Midwest by Middle Woodland times (Smith 1989, 1992d). Bruce Smith’s gradualistic “floodplain weed theory” (Smith 1992c) rose to prominence as an explanation of the initial cultivation of crops in riverine habitats. This model synthesized ecological knowledge of plant evolution in disturbed habitats with the record of midwestern climate and settlement change to show convincingly how farming might have developed naturally from unintended human impacts on floodplain environments. The emphasis on independent domestication and gradualistic, ecologically sophisticated theories of subsistence change was a much-needed antidote to the sterile diffusion-migration scenarios that satisfied many archaeologists of previous generations. However, there are signs of renewed interest in the historical aspects of agricultural origins in regions such as the eastern United States, where the autochthonous nature of plant domestication has gained widespread acceptance. The key issue now is not the influence of Mesoamerican agricultural knowledge, but rather the “alternative pathways” followed by indigenous farming systems as they adapted to the variable environmental conditions encountered across space and over time in the region (Fritz 1990; Harris 1977). The adaptive responses we infer from the archaeological record arise by way of general evolutionary, economic, and social processes, but they are elicited by particular historical circumstances. One such particular is the differential geographic distribution of wild populations of the weedy annuals that became domesticates. Many of these plants are adapted to dynamic floodplain habitats yet outgrew this niche under human management. How this occurred, and how this transfer of plants and knowledge about them affected the human groups that adopted them, is the topic of this chapter. The development of food production on the Cumberland Plateau of eastern Kentucky is especially relevant to this issue. At least some of the

Historical Aspects of Early Plant Cultivation

157

so-called floodplain weeds (goosefoot, sumpweed, and Cucurbita gourd) seem not to be members of the local flora today, or to occur only rarely or as small, isolated populations (Cowan 1985b; Smith and Cowan 2003). How were they introduced to this rugged landscape from their midcontinental home, by what cultural mechanisms, and in what form? Were they ever present as free-living populations in this landscape of high relief and narrow stream valleys, or were they introduced as crop plants? Were the seeds introduced through trade or intermarriage, and were they accompanied by relevant information about cultivation practices? Was the cultivation of plants a true innovation to the people of eastern Kentucky who acquired them? What role did gradual coevolution of human populations and weedy annuals play in the emergence of food production in this region? These questions have remained largely unaddressed in favor of the issue of why food production became a successful adaptation. However, we cannot afford to ignore the history of crop diffusion. Whether and how the floodplain weeds were introduced to the Cumberland Plateau may well have explanatory relevance not only to why food production arose when it did but also to understanding the path along which it developed thereafter. The fact that the floodplain weeds were not native to the region introduces the possibility of considerable dependence of these plants on human management for local persistence. In addition, the fact that these plants were introduced as crops brings into sharper focus the question of what benefits were gained by adding small-scale farming to a foraging economy. Adoption of innovations carries risks as well as opportunities, both of which must be factored in when assessing the costs and benefits of cultivating crops. Perceived risk and uncertainty tend to inhibit or slow the diffusion of innovations such as new crops among contemporary peasant farmers (Ghadim and Pannell 1999; Marra, Pannell, and Ghadim 2002; Sarkar 1998). It seems likely that the same factors affected the pace of subsistence change as cultivation of food crops became established among the foragers of the Cumberland Plateau. However, I will argue that seed crops are more likely to have reduced the total variance in harvest yields than to have increased it. This may account for their adoption despite the high costs associated with processing these small grains. The goal of this chapter is to assess the role of crop diffusion in the origins of food production on the Cumberland Plateau. First I provide some background on the domestication of native plants in eastern North America, and the record of early plant cultivation in eastern Kentucky. Next I discuss evidence for the geographic distribution, habitat, and ecology of the three floodplain weeds that were domesticated in the region: goosefoot, sumpweed, and native gourd. I discuss archaeobotanical and floristic evidence for their presence in eastern Kentucky prior to their establishment as crops. Finally I take up the implications of patterns and

158

social archaeobotany

paths of the introduction of floodplain weeds for the explanation of the forager-farmer transition in this region.

Background: Origins of Farming in the Uplands of Eastern Kentucky The Cumberland Plateau first attracted professional archaeological attention because of the many rockshelters whose dry, sandy sediments offered ideal conditions for the preservation of organic remains, including textiles, plant remains, and paleofeces (Funkhouser and Webb 1929, 1930; Webb and Funkhouser 1936) (figure 8.1). The collection recovered from the Newt Kash shelter (15MF1) in the Licking River drainage (Jones 1936) played a critical role in providing empirical support for the premaize “Eastern Agricultural Complex” that had been hypothesized by Ralph Linton (Linton 1924). The research effort led by C. Wesley Cowan in the 1970s used data from several sites to investigate the development of early plant cultivation along the North Fork of the Red River in an area known as the Red River Gorge. This work was the first to place the rich botanical assemblages of the region’s rockshelters in ecological context using a variety of data sources including pollen, wood charcoal, faunal remains, and sediments (Cowan 1985a, b; Cowan et al. 1981). The centerpiece of Cowan’s project was Cloudsplitter rockshelter (15MF36). Specimens of domesticated or cultivated forms of indigenous weedy annual plants including chenopod (Smith and Cowan 1987), sumpweed, maygrass (Phalaris caroliniana), and knotweed (Polygonum) were abundant in Early Woodland contexts, in contrast to their scarcity in the Late Archaic. Work at the Cold Oak shelter in the Big Sinking Creek drainage about 20 km south of the Red River Gorge has provided further support for many of the patterns documented at Cloudsplitter (Gremillion 1993c, 1995, 1998). Increased quantities of seeds of cultivated plants and the appearance of storage features distinguish post-3000 BP deposits from earlier ones at both sites. At Cold Oak, the increase in cultivated plant seeds is accompanied by a general expansion of the anthropogenic component of the macrobotanical assemblage, which is dominated by weedy herbs in contrast to the woody, closed-canopy taxa represented in pre-3000 BP deposits. The temporal association of abundant crop seeds with storage features at Cold Oak and Cloudsplitter supports Cowan’s hypothesis that the seasonal utility of seed crops as reserves for times of scarcity was an important factor affecting their economic role (Cowan 1985a, b). Other empirical support for the dietary importance of food storage is provided by indicators of seasonal nutritional stress in desiccated human feces from a Late Woodland rockshelter occupation in the Red River Gorge

Historical Aspects of Early Plant Cultivation

159

Figure 8.1.  Map of the study area showing location of sites discussed in the text.

(Cowan 1978) and the association of crop seeds with seasonal foods in paleofeces from the Hooton Hollow and New Kash shelters dating to ca. 3000 BP (Gremillion 1996). Most recently, investigations at the Courthouse Rock shelter (15P0322) have produced additional documentation of the production and storage of seed crops in rockshelters between 3200 and 2500 BP (Gremillion 1999). Analysis of materials from test excavations at Courthouse Rock indicate the use of a wide range of indigenous crops, including Cucurbita gourd, maygrass, sunflower (Helianthus annuus), sumpweed, bottle gourd (Lagenaria siceraria), and goosefoot. The expansion of investigations in recent years supports the explanation proposed initially by Cowan: that seed crops were a form of insurance that reduced the risk of food shortage during the winter months. Hickory nuts, acorns, chestnuts, and walnuts were all staple sources of carbohydrates and fats, but their yields are variable from year to year (Gardner 1997). Reliance on nuts is also risky because they must be harvested quickly once they have fallen; there is a narrow window of opportunity for collection before competing wildlife depletes the crop. Unlike trees, annual plants could be managed effectively because they could be relocated and their yields manipulated (Cowan 1985a). This explanation

160

social archaeobotany

for the success of annual seed crops in the region is consistent with the relatively low profitability of these plants compared to most of the local wild flora and fauna (Gremillion 2004). Their utility is greatly enhanced by their storability, ease of manipulation, and ability to offset the risk associated with variable nut yields.

Floodplain Weeds in Eastern Kentucky: Native or Introduced? Cowan was convinced that the crops grown by the people of the Red River drainage were not locally native. Based on his research at Cloudsplitter and their absence from the modern flora of the valley, he proposed that “a wholesale introduction of the plants and the methods to cultivate and process their edible parts seems to have taken place” (Cowan 1985a, 371). More recent research has supported this assertion for the most part but has also enriched it with details that have explanatory relevance. In the following section I discuss the ecological niche of the three floodplain-­adapted plants that came into widespread use as domesticates in eastern North America as well as evidence for their prehistoric geographical distribution.

The Ecological Niche of Floodplain Weeds Goosefoot, sumpweed, and gourd are all closely adapted to frequently inundated floodplains. Sumpweed and gourd rely on floodwaters for dispersal, and all occupy an ecological niche that includes colonization of open patches on which woody plants cannot become established. All three can form large stands on floodplains, where they can be harvested with a relatively small amount of labor. Because their persistence relies on ongoing disturbance of floodplain soils, where they have a competitive advantage, they are not effective colonizers of most upland landforms. However, they do grow outside the floodplain habitat, in some cases having expanded into human-disturbed habitats such as roadsides and waste areas (Smith 1987a, 1992b; Smith, Cowan, and Hoffman 1992).

The Archaeological Record of Wild Populations in Eastern North America Of the three taxa considered here, Cucurbita gourd appears to have the longest history of association with human populations (table 8.1). Remains of Cucurbita have been recovered from Early Archaic (ca. 7000 BP)

Historical Aspects of Early Plant Cultivation

161

Table 8.1. Locations and approximate dates of earliest archaeological occurrences of selected crop plants of eastern North America Taxon

Location

Date (RCYBP)

Reference

West-central IL (Koster)

8500 (earliest) 4000 (harvested)

Asch and Asch 1985

3450 + 50

Smith and Cowan 1987

Chenopodium berlandieri   Wild type

 Cultigen type Eastern KY (Cloudsplitter) Iva annua   Wild type

West-central IL (Koster)

7320 (earliest) Asch and 4900 (harvested)* Asch 1985

 Cultigen type West-central IL 4500 + 500 (Napoleon Hollow)

Smith and Cowan 2003

Cucurbita pepo ssp. ovifera   Wild type

West-central IL (Koster)

 Cultigen type MO (Phillips Spring)

ca. 7000

Asch and Asch 1985

4440 + 75

Smith and Cowan 2003

Note: “Harvested” indicates that quantities or context are highly indicative of intensive collecting by humans. *Most recent date from this deposit.

archaeological contexts in west-central Illinois (Asch and Asch 1985). However, specimens exhibiting the relatively large seeds and thick rinds of domesticated forms postdate 5000 RCYBP (cal 3000 BC). The earliest archaeological examples of sumpweed come from Middle Archaic deposits at Koster in west-central Illinois (Asch and Asch 1985). The large quantity of seeds from Horizons 7–6 at Koster indicate intensive exploitation of wild stands. Coring of noncultural deposits has revealed that wild populations of sumpweed were present in the Illinois valley by 12,000 RCYBP. However, direct dating of seeds from the Napoleon Hollow site has pushed back the date of earliest domestication of sumpweed to not later than 4500 RCYBP (cal 3197 BC). Chenopodium berlandieri is present in domesticated form as early as 3450 RCYBP (cal 1785 BC) in eastern Kentucky (Smith and Cowan 1987). However, seeds representing the genus Chenopodium are common on prehistoric sites dating back to the Early Archaic in eastern North America, although they do not appear

162

social archaeobotany

in substantial quantities until after 3000 RCYBP (Smith 1985a, b; Yarnell and Black 1985). In west-central Illinois, Chenopodium berlandieri makes its first archaeological appearance in the Early Archaic in the form of a few seeds lacking morphological signs of domestication (Asch and Asch 1985). Quantities suggestive of the harvesting of wild stands are present there at ca. 4000 RCYBP.

Natural Populations of Floodplain Weeds on the Cumberland Plateau The earliest examples of all three floodplain weeds come from the Illinois River valley. Sumpweed and goosefoot were both harvested there prior to their establishment as garden crops. This pattern is due in part to the high volume of archaeobotanical samples drawn from that region over the past twenty-five years. However, it is also fully consistent with the known habitat and geographical distribution of these plants today. However, in the uplands of eastern Kentucky, there is very little evidence pointing to a subsistence role for any of these plants before the appearance of cultigentype morphology around 3500 BP. In the section to follow, I review data supporting this view from three areas of research: the archaeobotanical record of seed utilization, the paleoecology of the Cumberland Plateau (specifically the upper Kentucky River drainage), and the contemporary floristics of the region.

The Archaeobotanical Record Unlike in Illinois, preagricultural deposits in eastern Kentucky have provided no evidence of the harvesting of any of the floodplain weeds, with the possible exception of the native gourd. Gourd remains resembling fruits and seeds of free-living Cucurbita pepo ssp. ovifera have been directly dated to 5130 +60 at Cloudsplitter and 5080 +60 BP at the Mounded Talus shelter in the Big Sinking Creek drainage (Gremillion and Mickelson 1996). That a small, hard-shelled gourd was used on the plateau at this early date cannot be contested. Cowan argues that oilrich seeds were first harvested from local wild populations, which were brought into cultivation, showing signs of domestication by 3000 BP (Cowan 1997). The morphology of the earliest gourd seeds and rind from eastern Kentucky is consistent with the exploitation of wild populations. It is also possible that these gourds were managed in some way without undergoing selection for larger seeds or fruits. The vast majority of Chenopodium berlandieri seeds from eastern Kentucky sites exhibit the thin seed coats and rectanguloid cross-section

Historical Aspects of Early Plant Cultivation

163

characteristic of domesticated forms (Smith 1985b). However, there is more variability in size and shape than in the case of sumpweed. Two Chenopodium seeds from the Cold Oak shelter associated with Terminal Archaic dates are identical to seeds from modern wild populations. They also resemble the weed morph of Salts Cave (Gremillion 1993a). These could be from naturally occurring wild populations, but they are associated with cultigen-type seeds and do not predate them. Thus the possibility cannot be ruled out that these wild-type seeds are variants within a cultivated crop, or represent weed varieties associated with it. Earlier finds of Chenopodium at the Mounded Talus shelter dating to the Middle Archaic are C. missouriense, not C. berlandieri. Judging on the basis of the size of seeds and achenes, there are no archaeological examples of sumpweed from the Cumberland Plateau that are not from domesticated populations. Thus there is no record of its presence prior to its establishment as a garden crop.

Paleoecological Evidence for Persistently Disturbed Habitats The three floodplain weeds considered here form persistent populations only where environmental conditions—such as annual flooding or other frequent or continual disturbance—limit competing vegetation (Smith 1987a, 1992b; Smith and Cowan 1993; Smith, Cowan, and Hoffman 1992). Floodwaters are also an important means of seed dispersal. For these reasons, wild populations of these plants have difficulty becoming established in forest gaps even under conditions of anthropogenic disturbance, simply because they do not compete well with more ecologically generalized weedy colonizers. Plants adapted to dynamic river-margin habitats typically lack effective adaptations for dispersal between forest openings (Marks 1983). Weedy annuals in general are dominant in early successional habitats only during the first year following a disturbance; in eastern North America, biennial and then perennial species take over thereafter, followed by seedlings of woody plants. Once a canopy has re-formed, weedy colonizers are generally excluded (Bazzaz 1996, 140). Several lines of evidence indicate the general lack of suitable habitat for floodplain weeds in the valleys of the Red River and Big Sinking Creek, both of which drain into the Kentucky River. The rockshelters that have produced the bulk of evidence for early plant domestication are located in sandstone cliffs adjacent to steep forested slopes. The Big Sinking floodplain is limited to occasional benches or small terraces. The Red River is a larger stream and has developed substantial terraces in its lower reaches, but the floodplain itself is narrow enough to be in large part shaded by vegetation. Floodwaters do cover the terraces in some years, but they recede rapidly and do not seem to interfere with forest development. The

164

social archaeobotany

forest canopy extends nearly to the water’s edge. The North Fork of the Red in its upper portion is known as the Red River Gorge for the steep cliffs that directly border the channel. This environment contrasts with the typical habitat of gourds, sumpweed, and goosefoot in the larger midwestern river valleys, whose soils are flooded annually. Paleoecological data support the limited nature of forest opening on the Cumberland Plateau before the watershed date of 3000 BP, after which cultivated plants seem to have become established as major components of human diets. The pollen record from Cliff Palace Pond, located about 25 km south of the Big Sinking drainage, shows a sharp increase in ragweed, cheno-ams, and other successional herbs only after 3000 RCYBP (Delcourt et al. 1998). Charcoal size and density in the pond deposits, along with increasing pollen rain from oaks and pines, indicate that fire played a major role in this shift. Anthropogenic fires used to clear land for farming are a likely cause for these changes. A similar shift is represented in seed remains from other rockshelters in the area. The ratio of seeds of early successional plants to those of canopy and understory trees and shrubs is relatively low prior to the appearance of domesticated plants in the archaeological record (figure 8.2). There were therefore few opportunities for colonizing annuals of any kind to develop large stands. Forest gaps were probably small, ephemeral, and widely spaced. Low-quality patches of this kind are likely to be passed up by human foragers in favor of richer ones that offer a better rate of energy gain.

Modern Floristics of the Cumberland Plateau The contemporary landscape of the Cumberland Plateau, in particular the Upper Kentucky drainage, does not include substantial populations of any of the three floodplain weeds. Sumpweed is present but not abundant and is considered to be a relatively recent introduction from west of the Mississippi (Hargan 1991; Medley 1993). Chenopodium berlandieri has not been reported at all from the eastern part of the state, with one exception from an Ohio River valley locality (USDA and NRCS 2002). In some areas, goosefoot at least has widened its niche to become established in anthropogenic habitats outside of floodplains (Asch and Asch 1985), but this is not the case in eastern Kentucky. A similar pattern characterizes Cucurbita pepo ssp. ovifera, which has been reported growing in the wild in Kentucky only in counties immediately adjacent to the Ohio River. The possibility that these examples are escapes from cultivation cannot be ruled out. It is also possible that all three floodplain weeds were more widely distributed in prehistory. This is clearly the case in the context of plant

Historical Aspects of Early Plant Cultivation

165

Figure 8.2. Ratios of weed seeds to tree seeds for four archaeobotanical assemblages from eastern Kentucky rockshelters. Notes on data, by site: 15Le50 (Early Woodland), weed seed count excludes 2,653 seeds of Amaranthus sp. from a single feature; 15Po322, results from Unit B only; 15Le77, counts from Middle Archaic contexts only.

domestication. However, it seems unlikely that they were able to stray far from the Ohio River valley without human assistance. Had they been more widespread, their sparse distribution on the Cumberland Plateau today would seem puzzling.

Floodplain Weeds as Introduced Crops: Implications for an Explanation The Mosaic Character of Agricultural Origins Based on their archaeological occurrence, there is no reason to think that the various grain crops of eastern North America (including the floodplain weeds) were introduced or adopted as a package. Despite the rapid increase in their archaeological visibility after 3000 BP, goosefoot, gourd, and sumpweed make their first appearance at different times. Gourd is the earliest, dating back to the late Middle Archaic, when human disturbance of the landscape was still very limited in scope. However, it maintained an association with humans for several millennia before undergoing morphological changes as a result of domestication. Cowan

166

social archaeobotany

(1997) attributes its initial use to the oil-rich seeds, with the thick rind needed for containers postdating 3000 BP and nonbitter, fleshy varieties still later. Before these changes took place, its status is uncertain, and it may have been growing naturally along the river valleys of the Cumberland Plateau. However, its presence in rockshelters far removed from the substantial floodplains that are its natural habitat suggests some sort of human management. Goosefoot occurs on the plateau primarily as a domesticated plant, with very few archaeological seeds whose morphology is compatible with wild populations. In any case, these examples do not predate the cultigen varieties. Finally, sumpweed provides a clear-cut case of introduction to the region exclusively as a crop plant. There are some locally native weedy annuals that were also probably cultivated prehistorically. These include amaranth, giant ragweed, and several species of Polygonum that are common and locally abundant in eastern Kentucky today (Hargan 1991; USDA and NRCS 2002). However, they do not predate the floodplain weeds in the archaeological record, as might be expected if disturbance of local vegetation led to domestication as it did in the larger river valleys of the Midwest. In fact, it is possible that local weeds were harvested in quantity only after the introduction of crops and the consequent opening of the forest. This is what Cowan suggests for giant ragweed, which is a particularly aggressive colonizer of disturbed habitats.

The Hilly Flanks as a Marginal Zone? The fact that floodplain weeds were outside their native habitat in eastern Kentucky has ecological significance relevant to the path taken by food production in the region. Once introduced, sumpweed and goosefoot probably required human maintenance in order to persist because they lacked a local refugium to which they could retreat and found successful wild populations. They would be quickly out-competed in abandoned gardens or fields. They had to be either maintained as crops or quickly lost. That the former prevailed reflects the botanical knowledge of local farmers. It also may explain divergent patterns in the archaeobotanical records of eastern Kentucky and west-central Illinois. In the latter region, there are considerable gaps in time between the earliest evidence of harvested wild populations of sumpweed and goosefoot and their establishment as economically important crops, which is not apparent until the Middle Woodland (2000 BP or so). In contrast, in eastern Kentucky evidence of harvesting and domestication are largely coeval. Storage economies based on cultivated plants were established by 3000 BP and persisted

Historical Aspects of Early Plant Cultivation

167

until at least 1300 BP. The removal of these plants from their natural habitat necessitated a degree of dependence on human management that was not replicated in regions where populations could more readily revert to a wild state. This difference may account for the early establishment of food production in Kentucky compared to Illinois (although preservation biases are relevant as well). In a sense, the “hilly flanks” of eastern Kentucky can be considered a “marginal zone” with respect to the quality of habitat available to floodplain weeds without human assistance. Rather than being harvested initially from substantial wild stands and evolving into domesticates, as they seem to have done in their native habitat, these plants may have been placed on the fast track to staple food status in areas such as the Cumberland Plateau where they had to be cultivated annually to ensure their availability.

Cultural Transmission and the Rate of Subsistence Change The role of cultural transmission in agricultural origins also has implications for the pace of subsistence change on the Cumberland Plateau. Empirical studies and theoretical analyses both indicate that the adoption of technological innovations entails distinctive costs and benefits (Ghadim and Pannell 1999; Marra, Pannell, and Ghadim 2002; Sarkar 1998). Studies of the diffusion of innovations among contemporary peasant farmers converge in the frequent finding that the risks associated with innovation are important factors inhibiting the adoption of new techniques. Risk of failure to produce is perceived as being high because knowledge is limited and outcomes are uncertain. Adoption is more likely if it can be accomplished incrementally, by adding techniques or items piecemeal, or by experimenting first with a small input of resources. Thus it is to be expected that the greater the quantity of relevant information that accompanies an innovation, the greater the likelihood of adoption. If the costs of risk and uncertainty are addressed, innovation through social learning has the potential to initiate rapid changes in cultural practice. Acquisition of materials and information from a large social network can actually cut some of the costs associated with experiential learning, which may involve repeated trial and error. In comparison to genetic inheritance, cultural transmission greatly enhances the speed with which behavioral traits can spread through and between populations. In the case of food production on the Cumberland Plateau, there is reason to believe that local ecological conditions were favorable to innovation. First, the early appearance of Cucurbita gourd in the region suggests some familiarity with the tending of useful plants, even if not

168

social archaeobotany

in the context of major dietary dependence. This knowledge would have reduced the time costs of learning effective gardening techniques. Second, the role of seed crops as insurance against seasonal food shortages implies that their adoption promised to diminish the risk of starvation or malnutrition during the winter months. This would have offset, at least to some degree, the risk entailed by taking on an untried technology. Risk was further limited by the characteristics of the crop plants themselves, which can be grown, maintained, and harvested with a relatively small input of labor. Seeds are costly to process for consumption, but processing for storage takes little or no time. Winnowing and grinding could be deferred until winter, when competing activities were few and time less valuable than during the peak harvest season of late summer and fall.

Summary and Conclusions Both cultural and ecological events and processes can be said to have led up to the establishment of food production on the Cumberland Plateau. It remains important to address the ecological conditions that made food production an adaptive strategy for human groups in certain environments. Resource density and distribution have significant impacts on economic costs and benefits; as the structure of the landscape changes in time and over space, people adapt by choosing which resources to pursue, and when. This follows from the assumption of human behavioral ecology that human decisions about diet and subsistence tend, over time and on average, to enhance fitness. This approach to explanation is generalizing in nature; it views subsistence behavior in terms of ultimate (evolutionary) causality. However, it is important not to allow the generalizing goal to overshadow or devalue historical particulars as components of explanation. The introduction of goods and information—such as seeds and cultivation techniques—is an example of the kind of historical fact that must be included in a well-rounded explanation of subsistence change. Diffusion affects agricultural origins on a number of levels—not only by providing raw material in the form of crops but by influencing the costs and benefits that play a central role in evolutionary accounts of change. The particular history of individual crop plants reminds us that food production is not a monolithic entity, but a set of behaviors and traditions that can be articulated in diverse ways. Displacement of crop plants from their native habitats may change the type and scale of management that is optimal. Novelty adds start-up costs, including the time to learn effective techniques; poor initial returns from farming slow its integration into

Historical Aspects of Early Plant Cultivation

169

subsistence. However, incremental exposure to crops and technologies can limit perceived risk and accelerate the pace of change. Diffusion of innovations enlarges the pool of information available to individuals and groups beyond that available through experiential learning. This ability seems to be a key feature of human adaptation generally. It should not be surprising that a complex mix of social learning, historical particulars, and ecological dynamics underlies the shift to farming, not just on the Cumberland Plateau but also worldwide.

Acknowledgments This chapter was originally prepared for the 2004 Society for American Archaeology Annual Meeting.

References Asch, D., and N. Asch. 1977. “Chenopod as Cultigen: A Reevaluation of Some Prehistoric Collections from Eastern North America.” Midcontinental Journal of Archaeology 2:3–45. ———. 1985. “Prehistoric Plant Cultivation in West-Central Illinois.” In Prehistoric Food Production in North America, edited by R. I. Ford, 149–204. Anthropological Papers, vol. 75. Ann Arbor: Museum of Anthropology, University of Michigan. Bazzaz, F. A. 1996. Plants in Changing Environments: Linking Physiological, Population, and Community Ecology. Cambridge: Cambridge University Press. Caldwell, J. R. 1958. Trend and Tradition in the Prehistory of the Eastern United States. Memoirs of the American Anthropological Association, no. 88. Menasha, WI: American Anthropological Association. Cowan, C. W. 1978. “Seasonal Nutritional Stress in a Late Woodland Population: Suggestions from Some Eastern Kentucky Coprolites.” Tennessee Anthropologist 3:117–28. ———. 1985a. “From Foraging to Incipient Food Production: Subsistence Change and Continuity on the Cumberland Plateau of Eastern Kentucky.” PhD diss., University of Michigan. ———. 1985b. “Understanding the Evolution of Plant Husbandry in Eastern North America: Lessons from Botany, Ethnography, and Archaeology.” In Prehistoric Food Production in North America, edited by R. I. Ford, 205–43. Anthropological Papers, vol. 75. Ann Arbor: Museum of Anthropology, University of Michigan. ———. 1997. “Evolutionary Changes Associated with the Domestication of Cucurbita pepo: Evidence from Eastern Kentucky.” In People, Plants, and

170

social archaeobotany

Landscapes, edited by K. J. Gremillion, 63–85. Tuscaloosa: University of Alabama Press. Cowan, C. W., E. E. Jackson, K. Moore, A. Nickelhoff, and T. Smart. 1981. “The Cloudsplitter Rockshelter, Menifee County, Kentucky: A Preliminary Report.” Southeastern Archaeological Conference Bulletin 24:60–75. Crawford, G. W. 1982. “Late Archaic Plant Remains from West-Central Kentucky: A Summary.” Midcontinental Journal of Archaeology 7:205–24. Crites, G. D. 1987. “Middle and Late Holocene Paleoethnobotany of the Hayes Site (40ML139): Evidence from Unit 990N918E.” Midcontinental Journal of Archaeology 12:3–15. ———. 1993. “Domesticated Sunflower in Fifth Millennium BP Temporal Context: New Evidence from Middle Tennessee.” American Antiquity 58:146–48. Decker, D. S. 1988. “Origin(s), Evolution, and Systematics of Cucurbita pepo (Cucurbitaceae).” Economic Botany 42:4–15. Delcourt, P. A., H. Delcourt, C. R. Ison, W. Sharp, and K. J. Gremillion. 1998. “Prehistoric Human Use of Fire, the Eastern Agricultural Complex, and Appalachian Oak-Chestnut Forests: Paleoecology of Cliff Palace Pond, Kentucky.” American Antiquity 63:263–78. Fritz, G. J. 1990. “Multiple Pathways to Farming in Precontact Eastern North America.” Journal of World Prehistory 4:387–435. Fritz, G. J., and B. D. Smith. 1988. “Old Collections and New Technology: Documenting the Domestication of Chenopodium in Eastern North America.” Midcontinental Journal of Archaeology 13:3–27. Funkhouser, W. D., and W. S. Webb. 1929. The So-Called “Ash Caves” in Lee County, Kentucky. Reports in Archaeology and Anthropology, 1 (2). Lexington: University of Kentucky. ———. 1930. Rock Shelters of Wolfe and Powell Counties, Kentucky. Reports in Archaeology and Anthropology, 1 (4). Lexington: University of Kentucky. Gardner, P. S. 1997. “The Ecological Structure and Behavioral Implications of Mast Exploitation Strategies.” In People, Plants, and Landscapes, edited by K. J. Gremillion, 161–78. Tuscaloosa: University of Alabama Press. Ghadim, A. K. A., and D. J. Pannell. 1999. “A Conceptual Framework of Adoption of an Agricultural Innovation.” Agricultural Economics 21:145–54. Gremillion, K. J. 1993a. “Crop and Weed in Prehistoric Eastern North America: The Chenopodium Example.” American Antiquity 58:496–509. ———. 1993b. “Paleoethnobotany.” In The Development of Southeastern Archaeology, edited by J. Johnson, 132–59. Tuscaloosa: University of Alabama Press. ———. 1993c. “Plant Husbandry at the Archaic/Woodland Transition: Evidence from the Cold Oak Shelter, Kentucky.” Midcontinental Journal of Archaeology 18:161–89. ———. 1995. Archaeological and Paleoethnobotanical Investigations at the Cold Oak Shelter, Kentucky. Report submitted to National Geographic Society, Grant Number 5226-94 by The Ohio State University.

Historical Aspects of Early Plant Cultivation

171

———. 1996. “Early Agricultural Diet in Eastern North America: Evidence from Two Kentucky Rockshelters.” American Antiquity 61:520–36. ———. 1998. “Changing Roles of Wild and Cultivated Plant Resources among Early Farmers of Eastern Kentucky.” Southeastern Archaeology 17:140–57. ———. 1999. National Register Evaluation of the Courthouse Rock Shelter (15P0322), Powell County, Kentucky. Report submitted to USDA Forest Service, Winchester, Kentucky, by The Ohio State University. ———. 2004. “Seed Processing and the Origins of Food Production in Eastern North America.” American Antiquity 69:215–34. Gremillion, K. J., and K. R. Mickelson. 1996. National Register Evaluation of the Mounded Talus Shelter (15LE77), Lee County, Kentucky. Report submitted to USDA Forest Service, Winchester, Kentucky, by The Ohio State University. Hargan, P. D. 1991. Weeds of Kentucky and Adjacent States. Lexington: University of Kentucky Press. Harris, D. R. 1977. “Alternative Pathways toward Agriculture.” In Origins of Agriculture, edited by C. H. Reed, 179–243. The Hague: Mouton. Jones, V. 1936. “The Vegetal Remains of Newt Kash Hollow Shelter.” In Rock Shelters in Menifee County, Kentucky, edited by W. D. Funkhouser, 147–65. Reports in Archaeology and Anthropology, 3 (4). Lexington: University of Kentucky. King, F. B. 1985. “Early Cultivated Cucurbits in Eastern North America.” In Prehistoric Food Production in North America, edited by R. I. Ford, 73–98. Anthropological Papers, vol. 75. Ann Arbor: Museum of Anthropology, University of Michigan. Linton, R. 1924. “The Significance of Certain Traits in North American Maize Culture.” American Anthropologist 26:345–59. Marks, P. L. 1983. “On the Origin of the Field Plants of the Northeastern United States.” American Naturalist 122:210–28. Marra, M., D. J. Pannell, and A. K. A. Ghadim. 2002. “The Economics of Risk, Uncertainty and Learning in the Adoption of New Agricultural Technologies: Where Are We on the Learning Curve?” Agricultural Systems 75:215–34. Medley, M. E. 1993. “An Annotated Catalogue of the Known or Reported Vascular Flora of Kentucky.” PhD diss., University of Louisville. Sarkar, J. 1998. “Technological Diffusion: Alternative Theories and Historical Evidence.” Journal of Economic Surveys 12:131–76. Smith, B. D. 1985a. “Chenopodium berlandieri ssp. jonesanum: Evidence for a Hopewellian Domesticate from Ash Cave, Ohio.” Southeastern Archaeology 4:107–33. ———. 1985b. “The Role of Chenopodium as a Domesticate in the Pre-Maize Garden Systems of the Eastern United States.” Southeastern Archaeology 4:51–72. ———. 1987a. “The Economic Potential of Chenopodium berlandieri in Prehistoric Eastern North America.” Journal of Ethnobiology 7:29–54.

172

social archaeobotany

———. 1987b. “Independent Domestication of Indigenous Seed-Bearing Plants in Eastern North America.” In Emergent Horticultural Economies of the Eastern Woodlands, edited by W. F. Keegan, 3–47. Carbondale: Center for Archaeological Investigations, Southern Illinois University. ———. 1989. “Origins of Agriculture in Eastern North America.” Science 246:1566–71. ———. 1992a. “Agricultural Origins in Eastern North America.” In Agricultural Origins in World Perspective, edited by C. W. Cowan and P. J. Watson, 101–20. Washington, DC: Smithsonian Institution Press. ———. 1992b. “The Economic Potential of Iva annua in Eastern North America.” In Rivers of Change: Essays on Early Agriculture in Eastern North America, edited by B. D. Smith, 185–200. Washington, DC: Smithsonian Institution Press. ———. 1992c. “The Floodplain Weed Theory of Plant Domestication in Eastern North America.” In Rivers of Change: Essays on Early Agriculture in Eastern North America, edited by B. D. Smith, 19–34. Washington, DC: Smithsonian Institution Press. ———. 1992d. “Hopewellian Farmers of Eastern North America.” In Rivers of Change: Essays on Early Agriculture in Eastern North America, edited by B. D. Smith, 201–48. Washington, DC: Smithsonian Institution Press. ———. 1997. “The Initial Domestication of Cucurbita pepo in the Americas 10,000 Years Ago.” Science 276:932–34. Smith, B. D., and C. W. Cowan. 1987. “Domesticated Chenopodium in Prehistoric Eastern North America: New Accelerator Dates from Eastern Kentucky.” American Antiquity 52:355–57. ———. 1993. “New Perspectives on a Wild Gourd in Eastern North America.” Journal of Ethnobiology 13:17–54. ———. 2003. “Domesticated Crop Plants and the Evolution of Food Producing Economies in the Eastern United States.” In People and Plants in Ancient Eastern North America, edited by P. E. Minnis, 105–25. Washington, DC: Smithsonian Books. Smith, B. D., C. W. Cowan, and M. P. Hoffman. 1992. “Is It an Indigene or a Foreigner?” In Rivers of Change: Essays on Early Agriculture in Eastern North America, edited by B. D. Smith, 67–102. Washington, DC: Smithsonian Institution Press. USDA (United States Department of Agriculture) and NRCS (Natural Resources Conservation Service). 2002. “The PLANTS Database, Version 3.5. National Plant Data Center, Baton Rouge, LA 70874-4490 USA.” http://plants.usda.gov. Webb, W. S., and W. D. Funkhouser. 1936. Rock Shelters in Menifee County, Kentucky. Reports in Anthropology and Archaeology, 3 (4). Lexington: University of Kentucky. Yarnell, R. A. 1972. “Iva annua var. macrocarpa: Extinct American Cultigen?” American Anthropologist 74:335–41.

Historical Aspects of Early Plant Cultivation

173

———. 1978. “Domestication of Sunflower and Sumpweed in Eastern North America.” In The Nature and Status of Ethnobotany, edited by R. I. Ford, 289– 300. Anthropological Papers, vol. 67. Ann Arbor: Museum of Anthropology, University of Michigan. Yarnell, R. A., and M. J. Black. 1985. “Temporal Trends Indicated by a Survey of Archaic and Woodland Plant Food Remains from Southeastern North America.” Southeastern Archaeology 4:93–106.

C h ap t e r 9

Routine Activities, Tertiary Refuse, and Labor Organization Social Inferences from Everyday Archaeobotany Dorian Q Fuller, Chris Stevens, a n d M e r i e l M c C l at c h i e

Two sets of forces have recurrently structured human history in the long term. On the one hand are the constraints imposed by environment and climate: biotic productivity, water availability, and the predictability of annual cycles. On the other are the constraints of social history, those cultural traditions that shape how a society is organized and how human energy is expended to meet individual and group needs. At the interface between these two sets of forces are human subsistence practices. As climate structures the temporal and spatial availability of both wild and domestic plant foods, so traditions that mold the scheduling of subsistence activities adjust and overcome such constraints. Hence we find that humans organize subsistence strategies around the pooling and exchanging of labor (both animal and human) and through the storing and redistribution of key resources, such as the harvested crop, the seed grain, the land, and tools. One of archaeology’s major contributions to social science is in developing an understanding of the long-term dynamics of how the scheduling of labor, including that involved in food production,

174

Routine Activities, Tertiary Refuse, Labor Organization 175 has changed, and how it has modified environments and been modified by new technologies and other social transformations. Through the application of botanical knowledge and laboratory techniques to the study of archaeological questions, archaeobotany has the potential to provide insights into questions of social organization and change in the past. In examining past social organization, one of the most fundamental issues we need to address is how labor was organized and scheduled within that society. In this chapter we will outline the reasoning and analytical implications of seeing most charred plant assemblages, with an emphasis on food or crop remains, as the product of incidental loss or waste disposal from recurrent, routine activities. Viewed as such, archaeobotanical assemblages have the potential to contribute to an understanding of how labor was deployed and scheduled on a site-by-site basis. Taken at a broader level they can then reveal varying patterns of how labor was organized in past societies, a point central to any conceptualization of greater societal change. In order to make this case, we will deal with the processes involved in the formation of archaeobotanical assemblages, focusing on possible routes to charring and how these can be distinguished. We consider the relationship between the archaeobotanical assemblage and the archaeological context from which it is recovered, and conclude that interpretations considering such a relationship to be socially significant are often misguided. Instead we argue that it is rather the internal composition—by which we mean all those seeds/chaff from a single flotation sample—of the assemblage itself that is most informative. Through a consideration of this composition, we address the vexed issue of the role of dung burning in the creation of Old World seed assemblages. We then consider a range of case studies, from Pakistan, England, South Asia, and Ireland, each briefly illustrating the utility of archaeobotanical evidence for inferring aspects of social organization, especially with regard to the deployment and scheduling of labor in food production and processing over the annual cycle (figure 9.1).

Breaking the Tyranny of Context: The Power of Content Archaeologists like to talk about context, and the importance of context for interpretation, but there are two very different concepts meant by context. On the one hand there are “contextual archaeologies,” for example, when finds or sites and their interpretations are understood in relation to the known social order of their time and place. This is essentially the “contextual” archaeology of Hodder (1991). Or one might understand context in terms of its ecological and geomorphic landscape setting, as in

176

social archaeobotany

Figure 9.1.  Map showing regions discussed in the text.

the “contextual archaeology” of Butzer (1982). Archaeological context, by contrast, is generally used to refer to the specific deposit from which archaeological finds are derived, that is, the continuous unit of sedimentary matrix together with a range of artifacts of other evidence. This depositional context is the basic unit of most field excavation recording systems and also the starting point for many analyses of the significance of finds in terms of past human activities. Indeed, “context” in this sense has been advocated as the key starting point for relating archaeobotanical remains to human activities (e.g., Dennell 1972, 1974, 1976a; Hastorf 1999; Miller 1991, 153; Pearsall 1988; 2000, 241). In Europe this has been the case since the application of flotation, which produced significantly increased archaeobotanical evidence more widely across sites. There is unfortunately a tendency to assume a direct relationship between activities that created an archaeobotanical assemblage and the excavated context from which it was recovered. The earliest examples of such beliefs were advocated by Robin Dennell (1972, 1974) in the early 1970s and have since been echoed on both sides of the Atlantic (e.g., Hastorf 1988, 1999; Hillman 1981). It should be mentioned that Gordon Hillman (1984a) was keen to draw attention to the different role context played in his own and in Dennell’s approach. While Dennell would interpret archaeobotanical remains by “external” reference to the

Routine Activities, Tertiary Refuse, Labor Organization 177 past function of the context in which they were found, Hillman (1981, 1984a) advocated that the function of context itself could be understood by the examination of the “internal” composition of the charred assemblage. It is not the aim of this chapter to further this debate. Rather it is to question why we should assume any relationship between context and archaeobotanical assemblage at all. The assumption that archaeologists can find a direct reflection of activities, such as cereal dehusking, on the basis of where remains are found was widespread in archaeology before the rise of formation process studies (La Motta and Schiffer 1999; Schiffer 1987). Such misconceptions seem to persist in relation to archaeobotanical data. Thus we find that postholes from granaries are sampled to find out what crops they held, floor layers to examine the location of processing activities, and field ditches to see what crops grew within them. Although we may dismiss such thinking as entrenched in the naïveté of field excavators, these issues are perhaps worthy of some archaeobotanical introspection. For while comparisons with modern crop husbandry have played a major role in the interpretation of charred plant remains (Hillman 1973, 1981, 1984a, 1984b, 1985; G. Jones 1984, 1987), the reasons as to how and why plant remains become charred and subsequently deposited within archaeological contexts have received comparably little attention.

Categorizing Charred Assemblages To state the obvious, charred remains only become charred and preserved through virtue of having come into contact with fire. It is then surely impossible to interpret the relationship between context and charred plant remains without considering how both relate to the fire responsible for the assemblage’s preservation. It is then perhaps surprising to find that this relationship is underplayed or sometimes even ignored in many of those studies that advocate the importance of archaeological context in the study of charred remains (e.g., Dennell 1972, 1974, 1976a; Hastorf 1988; Pearsall 1988). It is therefore essential that archaeobotanists are clear about the form of preservation (charred) and their assumptions or inferences about how the remains came to be charred and then deposited archaeologically. This amounts to a “behavioral context” (sensu La Motta and Schiffer 2001) in which plant processing, burning, and archaeobotanical evidence are linked. The “critical variables” (sensu La Motta and Schiffer 2001) that structure these activities are to be found in the relationships of plant structure, edible versus inedible parts, and the impact of charring on plant parts, which are determined by the characteristics of plants and thus transcend cultural variation.

178

social archaeobotany

One author who has addressed this issue on countless occasions is Richard Hubbard (Hubbard 1975, 1976a, 1976b; Hubbard and Clapham 1992; see also Wilkinson and Stevens 2003, 151–52). In their paper on taphonomy, Hubbard and Clapham divide assemblages into three groups according to the relationship between context and assemblage. The first group (class A) is where the remains have been burned in the context from which they were recovered. In this case, the context itself should display signs of burning. While this constitutes a case of primary deposition (sensu La Motta and Schiffer 1999), it is important to be clear that the material that entered the fire is not necessarily primary refuse. The second group (class B) represents assemblages that come from one discrete burning event, but have then been moved to the context from which they were excavated (secondary deposition). In this case the context itself shows no signs of burning, and in some cases the source of the material may be directly evident on the site. For example, an assemblage recovered from a pit next to a kiln may contain waste from the firing of that kiln. Or, to give a well-cited example, large spreads of burned material at the site of Assiros, in Greece, could be traced to the burning of second-floor storerooms located above the surface from which the remains were excavated (G. Jones et al. 1986). It is quite possible, however, that the relationship between context and assemblage may be unfathomable. To interpret charred assemblages from class B contexts, we must be aware that they are the product of at least three distinct groups of activities. The first are those activities that created the assemblage before it became charred; for example, the collecting of firewood, as well as the growing, harvesting, and processing of crops. The second are those that involve the burning of the assemblage, for example, the lighting of the fire, the discard of waste into a fire, or the heating and accidental burning of cereals while drying. Finally, there are those activities involved in the deposition of the waste, for example, the throwing of material onto a midden or the digging of a pit to bury midden material. Where the location of the fire in which the assemblage became charred is archaeologically invisible, we must consider the nature of the relationship between the context and each group of activities (e.g., Kreuz 1990). We often do not know, however, how far the botanical material resulting from various activities has been transported before it reached the fire—it is possible that the waste from the fire was further transported before it became incorporated into an archaeological feature. The final group (class C) was considered by Hubbard and Clapham (1992) to be the most ubiquitous of all. Class C assemblages differ from those of class B in that, rather than coming from a single event, they are formed from many different charring events and so, by inference, many

Routine Activities, Tertiary Refuse, Labor Organization 179 different activities (see Hubbard 1976a, 1976b). Hubbard and Clapham (1992) see these as the most difficult of all to interpret, a view that, as shall be seen, is not necessarily shared by the authors of this chapter. A curious point that arises from these categories is the validity of random sampling, so keenly advocated by many authors (Lennstrom and Hastorf 1992; Van der Veen 1984; for further consideration of sampling issues, see Lee, this volume). The first premise behind random sampling must be that most assemblages relate to different activities and events, in other words, they are of the class A and B varieties. The remains of plant-processing activities would therefore have to be disposed of and charred where the activities were carried out, or discarded into fires and then deposited without being mixed with residues of other activities. If this is the case, then patterns in the distribution of these activities should be detected by random sampling if enough samples are taken (Orton 2000, 14–39). Experience demonstrates, however, that class A and B assemblages occur only rarely in discrete locations throughout a site (see Miksicek 1987). In fact, such assemblages are only likely to be recovered through judgmental sampling of these specific contexts as the excavator identifies them (see M. Jones 1991). Contrary to what proponents of random sampling propose, this sampling strategy will almost certainly fail to recover class A and class B assemblages. As we hope will be seen, experience often reveals most assemblages to be of the class C variety. From this perspective, random sampling is likely to produce, on the whole, similar homogenized assemblages, a point demonstrated all too clearly in Van der Veen’s (1984) initial study. Similarly, a test of a blanket sampling approach demonstrated most contexts are very much the same regardless of context type, with only a few that stand out as atypical (Lennstrom and Hastorf 1995). Such homogenization often means that variation between samples is either related to different phases of occupation (see Pearsall 1988) or to postdepositional factors, such as proximity to the active soil horizon (Miksicek 1986, 1987).

Toward a Prevalent Taphonomy Instead of using context as a means of studying charred plant remains, we propose that it is the content of the assemblages themselves that is most informative about past human activities. The nature of most archaeobotanical evidence is such that its final resting place is usually only tenuously, if at all, connected to the activities that produced it. Just as archaeozoologists routinely look at the representation of body parts in relation to studies of differential utility, ethnographic butchery, and survivability (e.g., Lyman 1994), archaeobotanists require an understanding of

180

social archaeobotany

how the composition of closed assemblages—the proportions of different plant parts and species—represents past behavioral patterns. What we require is a methodological uniformitarianism (see Bailey 1981) for the creation of charred plant assemblages deposited on human occupation sites. The presence of fires on human occupation sites is a universal phenomenon within all societies. The ash and charcoal that are produced by such fires are equally routinely disposed of. Although disposal practices may be structured (e.g., Moore 1986, 109–10), fire waste can be expected to be deposited near settlements and, as such, is prone to redeposition and transport through wind, rainwash, walking, animal trampling, and sweeping, to finally accumulate within rubbish deposits and pits, against walls, and in ditches. Owing to the small size and lightweight nature of charred plant remains, we can expect them to linger within the soils and sediments that are found upon human habitation sites. We can further expect that, in time, charred material will form a significant proportion of the general background noise of refuse, even if disposal practices lead to its concentration within certain areas or deposits. The preservation of this material will obviously be affected not only by deposition but also by postdepositional factors. These may include destruction through trampling and weathering, with increased survival and preservation where such deposits are within soils and sediments that are subject to little bioturbation and have been rapidly buried. Such situations may be found within the basal deposits of deep pits; where mud-brick houses have collapsed and been leveled; or within fine, or rapidly silting, waterlain deposits. Thus, central to any archaeobotanical study of quantitative composition of charred assemblages is the question of how plants came to enter fires and how fire waste came to be disposed of. Recurrent influx to fire, along with outflux to archaeological “fill,” is most likely to result from frequently conducted, routine activities (figure 9.2). This is undoubtedly true on the grounds of pure statistical probability. Disregarding for the moment the events leading to the deposition of charred material, charred plant components resulting from routine activities that are conducted day-in, day-out are 365 times more likely to be represented than the once-in-a-year or occasional event, for example, the burning of old thatch, the cereal-processing accident, the rare medical ritual, or a life-passage rite (Fuller 2002, 264; Stevens 2003b). It is often the case that wood charcoal makes up the bulk of archaeobotanical assemblages. Wood as fuel is intentionally burned, in quantity, and thus wood charcoal is routinely produced in quantity. Seeds are generally a smaller proportion of the assemblage, but one of the remarkable features that every archaeobotanist will have experienced is the uniformity of samples from across given sites, cultures, and phases. The recurrence of the

Routine Activities, Tertiary Refuse, Labor Organization 181

Figure 9.2. A simplified diagram of the major pathway toward preservation of charred seed assemblages in Old World agricultural societies.

same species and parts of species (e.g., grains, seeds, nut shells, glume chaff, etc.), whether of wild foods, crops, or probable weeds, can often be detected in a narrow range of proportions. What such a pattern implies is that, within any region, during a defined period of time, the same type of material is not only routinely being burned but also, in all probability, is being produced by a limited number of well-defined activities and such activities are being repeated on a continuous basis. It is then our suggestion that the most important source for seeds and chaff are routine, “daily” activities of final crop processing for food preparation, in which crops, and their contaminants, are taken from storage and processed toward consumption (Stevens 2003b). The waste and incidental loss of crop that result from these activities are then disposed of in the fire, whether as intentional fuel or rubbish disposal. The charred products from each of these different daily burning events will then become combined and in quantitative terms averaged, as material is first amassed in the fire and then mixed in subsequent disposal and reworking of rubbish and sediment. What we propose, therefore, is that rather than being related to one particular event or activity, charred assemblages signify something that is essentially tertiary, or at best secondary, refuse (in the terms of Schiffer 1972), reflecting average and recurrent patterns of activity. Crop processing provides “critical variables” (sensu La Motta and Schiffer 2001, 25) that allow us to compare poststorage processing between archaeological sites and ancient social contexts. We shall now show how these activities are still interpretable from the examination of the internal composition of these essentially class C charred assemblages.

182

social archaeobotany

Assemblage Content and Recurrent Activities European archaeobotanists have been developing an understanding of the relationships between assemblage composition (content) and past human activities for more than three decades. As the application of sieving and flotation began to be more widely deployed, greatly increasing the volume of archaeobotanical evidence, professional archaeobotanists became a reality of archaeological research (see Fuller 2002, 257 figs.; Weber 2003). It was during this period that German archaeobotanists, Körber-Gröhne (1967) and Knörzer (1971), commented on the recurrent nature of archaeobotanical assemblages. In general, charred assemblages clearly represented a very limited range of the floristic diversity of the European flora. While many wild species were present, almost all were known from modern associations as weeds, with the exception of wild edible plants. Yet by far the most well represented remains were those of cereals, represented through both their grain and chaff. So with the exception of the remains of wild plant foods, most of this material could be seen as derived from arable plant communities rather than the environment at large. This had two significant implications for archaeobotanical analysis and interpretation. Firstly, the wild species present, if largely arable weeds, provided an important data set for inferring arable ecology and human practices of cultivation that created and maintained agricultural environments, such as tillage, manuring, and irrigation. Indeed, this has been a continuing focus of archaeobotanical analysis for the past three decades (e.g., Behre and Jacomet 1991; Bogaard 2002; Hillman 1981, 1991; G. Jones et al. 1995; G. Jones et al. 2000; M. Jones 1978, 1981, 1985; Knörzer 1971; Körber-Gröhne 1981; Küster 1991; Van der Veen 1992). The second implication is that the composition of assemblages reflected the outcome of filtering processes imposed by activities employed in processing crops to obtain clean grain. In order to understand the taphonomic effects of processing activities upon charred assemblages, Gordon Hillman undertook ethnoarchaeological work in Turkey in the early 1970s (see Hillman 2004, 77–78). As outlined by Hillman (1973), what ethnoarchaeology would provide was a link between assemblage composition and human activities, regardless of whether human activities could be inferred from archaeological context. Indeed, Hillman even went so far as to suggest that the function of the context could be inferred from the charred assemblage itself (Hillman 1981). This represents an early example of the kind of behavioral archaeology promoted by Schiffer (e.g., Schiffer 1976; La Motta and Schiffer 2001). This pioneering work (Hillman 1981, 1984a, 1984b, 1985) was followed by similar research by Glynis Jones (1984, 1987),

Routine Activities, Tertiary Refuse, Labor Organization 183 who contributed the important insight of the significance of weed seed size and weight (aerodynamic properties) and more robust statistical assessment of such models. Since that time, valuable studies have been carried out on a wider range of Old World crops, including millets (D’Andrea et al. 1999; Lundström-Baudais et al. 2002; Reddy 1997, 2003; Young and Thompson 1999) and pulses (Butler et al. 1999), with further investigations of wheat (Triticum spp.) and barley (Hordeum vulgare L.) (Peña-Chocarro 1999; Viklund 1998). While Hillman (1981), like Dennell (1974), emphasized the application of crop-processing assemblage signatures in the inference of activities, such an approach is misguided for reasons already outlined. Rather, the linkage of activities to recurrent patterns across assemblages and contexts provides a basis for meaningful social interpretation from archaeobotanical evidence. The residues of crop processing then need to come into contact with fire, and experiments have shown that this is likely to have biased samples by destroying certain botanical components preferentially according to their physical nature (Boardman and Jones 1990; Viklund 1998; Wilson 1984). This bias favors the preservation of more robust items, such as cereal grains/grass caryopses and pulses, with poorer (or no) preservation of fragile associated elements (e.g., chaff, pulse pods, etc.). Taken together, these studies indicate that traditional forms of nonmechanized crop processing are constructed from a limited set of actions that structure assemblages in predictable ways. Crop processing can be divided into two basic sets of activities, those that break apart the crop plant and those that separate out the various freed components. The first includes threshing to break apart cereal ears, or to separate pods of some pulses. Another, later stage for hulled crops is the dehusking to remove hulls and glumes still attached to the grain, and for “pod-threshing” pulses to remove the pod. Separation involves the removal of nonfood items, such as chaff, stems, capsules, seed heads, and weed seeds. A universal approach to separation is to rely on the physical properties of various components. Thus, winnowing separates elements according to weight and aerodynamics, while sieving separates according to size. Waste from various stages will therefore have characteristics of weight or size in common (G. Jones 1987; Stevens 2003b). As the harvested crop progresses through the crop-processing sequence, larger proportions of edible grain are retained relative to waste (mainly chaff and weed seeds); in addition, remaining weed seeds will be closer in size and weight to cereal grains, that is, large seeds will increase relative to small seeds (figure 9.3). When we consider how crop processing structures charred assemblages, alongside the implications discussed earlier in this chapter concerning the bias of charred evidence toward routine activities (figure 9.2),

184

social archaeobotany

Figure 9.3. A diagrammatic summary of the effect of crop processing on the composition of grain, chaff, and weed assemblages. The main crop-processing activities are numbered in order: 1. threshing, 2. raking, 3. winnowing, 4. coarse sieving, 5. first medium-coarse sieving, 6. pounding, 7. second winnowing/medium-coarse sieving, 8. fine sieving, 9. sorting.

an important implication logically arises. Following Hillman (1981, 1984a), the processing sequence can be divided into two groups of activities: those that precede storage and those that are conducted as crops are taken from stores. Those occurring prior to storage are usually conducted in bulk at harvest time, once or twice a year (depending on the local climate/latitude). These activities are less likely to contribute to the archaeobotanical evidence. Firstly, they are often conducted in the field or on specially prepared threshing floors, understandably located away from fires (see also Reddy 1999, 69; 2003). Secondly, because of the limited number of times such activities occur within a single year, the waste from them may only be present for a few weeks. These processing activities stand in marked contrast to those activities carried out when a crop is removed from storage. Such stages are repeated regularly throughout the year, usually within settlements and so in proximity to fires. Hence, there is a high probability for charred waste to enter the record in an iterative, numerically significant fashion.

Routine Activities, Tertiary Refuse, Labor Organization 185

Figure 9.4. The effects of storage strategies on daily processing activities and recurrent assemblage formation. Three alternative storage strategies are indicated, each of which requires different degrees of labor mobilization during the harvest period. This relates, therefore, to how many crop-processing stages, shown in the top row, are achieved prior to storage.

Thus, storage plays a crucial role in dividing crop processing into the routine (poststorage stages), which will be present (predominately) in the archaeological evidence, and the seasonal (prestorage, harvest processing), which will be either absent or rarely present (figure 9.4). Importantly, this dividing line of storage reflects a decision, arguably a strategy, on the part of past humans (agents) to carry out a certain amount of processing before storage and to leave the remainder to be conducted from day to day. Such strategic decisions can be related to demands on labor. The organization and scheduling of the agricultural work force is therefore planned according to how much labor is available for processing at the time of harvest, when other activities also demand attention. This is, of course, further dependent on how much time there is to conduct such

186

social archaeobotany

operations before crops need to be stored to stay ahead of the weather. If the time and labor are available, processing can be taken further, and the work required throughout the remainder of the year until the next harvest is reduced. Alternatively, fewer helping hands would promote storage in a less-processed state and the routine full processing of smaller amounts throughout the year (Stevens 2003b). In addition, the state at which crops are stored will determine how much storage space is needed, with grains taking less space than spikelets and considerably less than sheaths. Grain still in its spikelet is, however, more resistant to the attacks of pests (Hillman 1981; Reddy 2003, 77) and might therefore be strategically chosen as a way to store. An example of how charred assemblages reflect storage patterns can be drawn from Bronze Age Assiros in Greece, where the collapse of burned granaries could be inferred from archaeological features (G. Jones et al. 1986). In this case, the structural, archaeological context demonstrates that we are dealing with grain stores, indicating storage as semiclean spikelets of wheat. Had this fortunate chance of preservation not occurred, however, we could nevertheless infer the form of storage from the chaff, grain, and weed assemblages that recurred in domestic contexts and fill layers on the site. These samples would differ from the catastrophically charred stores only in a lower proportion of grain, as further processing should have removed some grain into the human food chain. Even had the Assiros material been redeposited elsewhere on the site, the composition would still indicate storage as spikelets.

Dung: Don’t Be Distracted Another potential source for charred seeds that is often discussed is burned dung. While the use of dung as fuel has been considered a likely source for the Near East (Charles 1998; Miller 1984; 1991, 154; 1996; Miller and Smart 1984), we expect it to be generally absent from sites in northwestern Europe owing to climatic constraints on drying dung and the ready availability of wood fuels. Is it realistic in well-wooded environments to suppose that prehistoric populations dried dung, which during most seasons would have required fire to do so, as a fuel? Accepting that this is extremely unlikely, the derivation of charred remains directly from crop-processing waste, either used as fuel or incidentally disposed of in fire, seems likely. This source through recurrence provides the only logical explanation for common assemblages of northwest Europe. Is it logical to assume, as proponents of a dung source for charred seeds do, that the explanation for the preservation of charred seeds in archaeological sites is different in the Near East (and perhaps southeast Europe) from

Routine Activities, Tertiary Refuse, Labor Organization 187 that in northwest Europe? In New World contexts like prehistoric eastern North America, we confront a similar quandary: in the absence of large domestic fauna herds and abundant forest resources, how do we explain the recurrent presence of charred seeds? We contend that the simplest, and most widely applicable, explanation is that charred seeds are derived from burned waste generated during the regular processing and preparation of food. This is not to deny that other sources of charred seeds, such as the burning of dung, might contribute to some assemblages, some of the time, but rather to emphasize that certain recurrent and crosscultural practices, such as plant-food processing, are likely to be the more quantitatively significant. We believe that the importance of routine crop-processing waste for archaeobotany can be demonstrated through some simple interregional comparisons. For the periods of the Pre-Pottery Neolithic Near East, domestic animals are absent or recent enough adoptions as to be fairly minor contributors to the charred seed record, and the range of wild species that may be weeds is by and large the same as that of the later Neolithic period, when domestic animals are an important part of the economy (table 9.1). These lists are very similar and suggest that the addition of domestic animals did little to change the composition of the archaeological wild seed roster. This suggests continuity in what these wild seeds represent. We concur with Hillman, Legge, and Rowley-Conwy (1997) that late Pleistocene foragers are unlikely to have invested time to gather wild ungulate dung during an era when wood sources were abundant (contra Miller 1996, 1997). The general similarity in the weed floras represented within charred assemblages from both pastoral and prepastoral sites suggests recurrent processes, and the routine use of livestock dung is not a possibility in the case of the former. If dung were the most significant source of seed remains, then surely the adoption of livestock would be expected to register a more significant impact upon the archaeobotanical record. If we compare this to the wet environments in Europe, represented in table 9.2 by weeds from the British and Irish Bronze Age through Anglo Saxon sites, we see a general similarity. Differences can all be explained in biogeographic terms as more temperate species and genera replace related taxa of more arid environments. What all of these seed taxa rosters represent are predominantly fast-turnover, disturbance-adapted species—namely, weeds. Certainly dung is a fuel source, especially in semiarid environments such as the Near East. Our point, however, is to make the case that the evidence from dung may well be swamped out numerically by the routine, or at the worst just add noise to evidence of arable weed-chaff assemblages. In a targeted ethnoarchaeological study on crop-processing waste in India, where dung is a major fuel source, Reddy (1999; 2003,

Table 9.1. Recurrent weeds in Near Eastern Neolithic archaeobotany Potamogetonaceae # Potamogeton sp.

# Triticum boeoticum Boiss. *

Liliaceae # Asparagus sp. * # Asphodelus sp. # Bellevalia sp. * # Liliaceae sp.

Ranunculaceae # Adonis sp. *

Juncaceae # Juncus sp. Cyperaceae # Carex divisa Huds. type Cyperus sp. * # Eleocharis sp. * # Scirpus sp. * Poaceae Aegilops crassa Boiss. * Aegilops sp. * Avena sp. * # Bromus sterilis L. # Bromus sp. * # Cynodon sp. dactylon (L.) Pers. type * # Echinochloa crusgalli (L.) Beauv. # Eremopyrum sp. * # Hordeum sp. (wild) * # Lolium sp. * Phalaris sp. * Phragmites sp. * # Secale sp. * # Setaria sp. *(incl. viridis) # Stipa sp. * # Taeniatherum sp.

Fumariaceae # Fumaria sp. Papaveraceae # Glaucium sp. * Caryophyllaceae # Gypsophila sp. * # Saponaria type * # Silene sp. * Stellaria sp. *

Lupinus sp. * # Medicago sp. * Melilotus sp. * Prosopis sp. * # Trifolium sp. * # Trigonella astroites Fisch. & C. A. Mey type # Trigonella sp. * # Vicia sp. # Vicia/Lathyrus sp. * Lythraceae # Alkanna sp. Brassicaceae # Alyssum sp. * # Capsella type * # Lepidium type

Chenopodiaceae # Atriplex sp. * # Chenopodium album L. * Malvaceae # Chenopodium sp. * # Malva sp. * # Suaeda sp. * Thymeleaceae Amaranthaceae # Thymelaea * # Amaranthus sp. Primulaceae Aizooaceae Anagallis sp. * Aizoon sp. * # Androsace maxima L.* Portulaceae # Portulaca sp. * Solanaceae Hyoscyamus sp. * Polygonaceae # Solanum sp. # Polygonum sp. * # Rumex sp. * Convolvulaceae # Convolvulus sp. * Geraniaceae # Cuscuta sp. Erodium sp. * Boraginaceae Fabaceae # Arnebia decumbens # Astralgus sp. * (Vent.) Coss. & Hippocrepis sp. * Kral * # Lathyrus cicera L. # Arnebia linearifolia # Lathyrus nissolia (L.) A. DC. * Döll

Routine Activities, Tertiary Refuse, Labor Organization 189 Table 9.1. Continued # Arnebia/ Lithospermum sp. * # Buglossoides arvensis (L.) I. M. Johnst. # Echium sp. * # Heliotropium sp.* # Lithospermum arvense L. * # Lithospermum tenuifolium L. f. * Rubiaceae # Crucianella sp. * Galium sp.(small) verum/palustre/ mullugo *

Galium aparine/ tricornutum * Galium spurium L. * # Galium sp. *

Apiaceae Foeniculum type * Torilis arvensis/ japonica *

Plantaginaceae Plantago sp. *

Dipsacaceae Cephalaria sp. *

Scrophulariaceae # Verbascum sp.

Linaceae # Linum sp. *

Lamiaceae # Ajuga sp. * # Micromeria sp. *

Valerianaceae # Valerianella sp.

# Stachys type # Teucrium sp. * # Ziziphora sp. *

Asteraceae # Centaurea sp. * # Helianthemum sp. *

Sources: Helbaek 1969; Van Zeist and Bakker-Heeres 1984; Van Zeist and Buitenhuis 1983; Van Zeist and De Roller 1995; Van Zeist and Waterbolk-Van Rooijem 1985; Willcox and Fornite 1999. Note: Symbol # indicates weeds likely on earlier Neolithic sites prior to the adoption of livestock in mid-Pre-Pottery Neolithic B; symbol * indicates weeds likely on later Neolithic sites after the adoption of livestock from mid-Pre-Pottery Neolithic B through the ceramic Neolithic.

158–60) found that crop-processing waste outnumbered seed input from dung when the charcoal from modern hearths was collected. The use of dung as fuel does not negate the need for households to carry out routine poststorage crop processing, which is therefore likely to contribute to the archaeobotanical record (Samuel 2001). Thus, we should be justified in approaching charred assemblages with some of the same expectations derived from crop-processing models, and the expectations of routine waste as a reflection of recurrent labor deployment, wherever we study agricultural societies that must process cereal crops. Having made this point, it is nevertheless wise to consider each site and set of assemblages on their own terms to assess the relative amount of noise added by dung burning, as Charles (1998) has. Another way to do so is through multiple lines of evidence, such as using phytolith assemblages from the same archaeological fills, since these can suggest the input of dung burning or other sources (Madella 2003). We briefly present such an example from a historical site in Pakistan.

Table 9.2. Recurrent weed species in British and Irish archaeobotany Juncaceae Juncus sp. Cyperaceae Carex sp. Cladium mariscus L. Eleocharis sp. Poaceae Anisantha sterilis (L.) Nevski (1) Arrhenatherum elatius (L.) P.Beauv. ex J.Presl & C.Presl Avena sp. Brizia media L./ Glyceria maxima (Hartm.) Holmb./ Milium effusum L. Bromus sp. Dactylis glomerata L. (1) Danthonia decumbens (L.) DC (including Sieglingia decumbens) Deschampsia sp. (1) Festuca/Lolium sp. Hordeum sp. (wild) Lolium sp. Phleum sp. Poa sp. Ranunculaceae Adonis annua L. (1) Ranunculus sp. Fumariaceae Fumaria sp. Papavaraceae Papaver sp.

Caryophyllaceae Agrostemma githago L. Cerastium sp. Dianthus deltoids L./ armeria L. (1) Lychnis flos-cuculi L. (1) Scleranthus sp. Silene sp. Spergula arvensis L. Stellaria sp. Chenopodiaceae Atriplex sp. Chenopodium album L. Chenopodium ficifolium Sm. Chenopodium polyspermum L. Chenopodium urbicum L. Chenopodium sp.

Medicago sp. Trifolium sp. Vicia hirsura (L.) Gray/tetrasperma Vicia tetrasperma (L.) Schreb. Vicia sp. Rosaceae Aphanes arvensis L. (1) Potentilla sp. Urticaceae Urtica sp. Brassicaceae Brassica sp. Lepidium sp. (1) Raphanus raphanistrum L. Sinapis alba L./ arvensis L. Malvaceae Malva sp.

Portulaceae Montia Fontana L.

Primulaceae Anagallis sp. (1)

Polygonaceae Fallopia convolvulus (L.) Á.Löve Polygonum sp. Rumex sp.

Solanaceae Hyoscyamus niger L.

Euphorbiaceae Euphorbia sp. Violaceae Viola arvensis Murray/ tricolor L. (1) Fabaceae Lathyrus sp. Lotus sp. (1)

Boraginaceae Lithospermum arvense L. Myosotis arvensis (L.) Hill (1) Rubiaceae Galium aparine L./ tricornutum Dandy Galium sp. Sherardia arvensis L.

Routine Activities, Tertiary Refuse, Labor Organization 191 Table 9.2. Continued Plantaginaceae Plantago sp. Schrophulariaceae Euphrasia/Odontites verna Odontites verna (Bell.) Dumortier Rhinanthus sp. (1) Veronica sp. Lamiaceae Clinopodium acinos (L.) Kuntze/Mentha sp. (1)

Galeopsis sp. Lamium sp. (1) Mentha sp. Prunella vulgaris L. (1) Teucrium cf. chamaedrys L. (1) Apiaceae Aethusa cynapium L. (1) Daucus carota L. Torilis sp.

Valerianaceae Valerianella sp. Asteraceae Anthemis cotula L. Anthemis sp. Centaurea sp. Cirsium/Carduus sp. Lapsana communis L. Tripleurospermum inodorum (L.) Sch. Bip.

Notes: Data are from the following sites and periods. For England: Barton Court Farm (NEO, IA, RB), Coburg Rd (MBA, LBA), Dairy Lane (MBA, RB), Fengate (NEO), Glanfeinion (MBA), Gravelly Guy (NEO, EBA), Hinxton (EBA, RB), Mount Farm (NEO, EBA, LBA, IA, RB, AS), Plan Fogerddan (NEO), Scotch Corner (IA, RB), St. Giles (IA), Sutton Common (IA), Tewkesbury (IA), Tewkesbury (RB), Watsons Lane (IA, RB), Yarnton (IA, RB, AS). For Ireland: Ballyglass (LBA), Ballypriorbeg (LBA), Ballyveelish (MBA, LBA), Ballyvelly I (LBA), Cappamore (LBA), Chancellorsland A and C (MBA), Cloghers (LBA), Crossreagh East (MBA), Curraghatoor (LBA), Dun Aonghasa (LBA), Dundalk (MBA), Haughey’s Fort (LBA), Kilmahuddrick (LBA), Lough Gur (LBA), Mannin Bay 2 (LBA), The Heath (MBA). Period abbreviations: NEO = Neolithic, EBA = Early Bronze Age, MBA = Middle Bronze Age, LBA = Late Bronze Age, IA = Iron Age, RB = Romano-British, AS = Anglo Saxon. Those species marked with (1) occur only once in the data set examined, while other species occur more than once.

Routine Crop Processing and Occasional Dung Burning: The Case of Hund, Northwest Pakistan The site of Hund, in Peshawar District (northwest Pakistan), is a substantial archaeological mound. At the end of the University of Peshwar’s excavation season in 1996–1997, Fuller was able to join the excavations to carry out a limited program of archaeobotanical sampling, including bulk flotation samples from layers, pit fills, and ovens, as well as sediment samples for phytoliths from each of the same contexts. This site was formerly a significant regional city. While historical references indicate that this was the major river crossing of the Indus, along the major

192

social archaeobotany

trade route from Kabul to India and where Alexander the Great crossed the Indus, the occupation of the site excavated began slightly later, in the second century BC. Stratigraphy and building phases continued through the sixteenth century AD, with a sequence readily datable by coin finds (Ali 1999). Analyses of plant macroremains and phytoliths (Cooke 2002; Fuller, unpublished data; Hassan 2000) provide a complementary picture of plant inputs to the site, with probable dung contributions being minor except for a few contexts that stand out from the norm. Rare contexts with dung fragments also have charred culm nodes, different weed taxa, and a phytolith assemblage higher in straw/culm morphotypes in comparison to most contexts dominated by spikelet phytoliths. This is illustrated in figure 9.5, where the typical dendritic phytolith–dominated assemblage can be contrasted with the occasional smooth rod–dominated assemblage. The high presence of spikelet phytoliths suggests that it is these parts that were most often produced and charred, even though charred macroremains of wheat and barley chaff tend to be underrepresented in comparison to grains. This may indicate the storage of wheat and barley ears or hulled barley spikelets, since the rachis segments of free-threshing wheat and barley are removed early in processing, whereas the harvested straw was removed prior to storage and utilized as fodder. It is the occasional context that sticks out from the typical macroremains assemblage. While most assemblages are dominated by wheat and barley, with large-seeded weed species, a few contexts have lower seed densities, the presence of charred culm nodes (from straw), and fragments of charred dung. This evidence warrants additional observations with regard to detecting dung-derived seed assemblages. It has been suggested (e.g., Miller 1984; 1991, 154) that the wood charcoal to seed ratio may reflect the likelihood or degree of burned dung contributing to an assemblage, with less wood and more seeds reflecting more dung. In Hund flotation samples, however, those samples with burned dung fragments, as well as culm nodes, actually have higher wood charcoal to seed ratios than those samples that represent routine processing, in which dung is absent and likely to be only a minor contributor, if at all. In the samples with the highest seed to wood ratio (by either volume or weight), no dung is present and the seed assemblage is dominated by fully formed grains of the well-known cereal weed Lolium temulentum L. The samples that are high in seeds and low in charcoal, which Miller’s approach would categorize as more likely dung derived, also contain quantities of grain remains and, in contrast to other samples, contained no charred dung or culm nodes. Overall, two conclusions can be drawn from the example of Hund. First, while dung contributes to a few unusual contexts, most of the

Figure 9.5. Relative frequency of phytolith morphotype categories in two selected samples from Hund. The graph at the top is typical of most samples from fills, whether inside or outside structures, while the pattern shown at the bottom was found in very few “special” contexts, including an oven fill and some pit fills.

194

social archaeobotany

archaeobotanical seed remains are best interpreted as the results of dayto-day cereal processing for human consumption. The quantitative implication from Hund is that when all of the samples are tallied together, the overriding picture is one reflecting recurrent daily cereal processing, in this case involving the threshing of naked wheat ears (reflected in chaff phytoliths), fine sieving, and final stages of handpicking large weeds. The other remarkable feature relating to these data is that the evidence is consistent from the start of the site, at ca. 200 BC, to the end of the sequence, ca. AD 1600. Although a few of the rare accidental plant inclusions, from fruits and oilseeds, may have changed over time, the predominant domestic crop-processing pattern for staple foods remained consistent. This is despite the major political upheavals and religious changes (from Buddhist to Hindu rulers to Muslim military chiefs) to which this site was central, according to historical sources. In fact, relatively little changed in the organization of agricultural activities. This seems a testament to the conservative nature of domestic labor organization and food production, thereby providing a useful way to compare sites, regions, and societies over the long term and across world regions.

Community Traditions in Labor Mobilization: The Case of Iron Age Britain Community traditions in the organization of crop processing can be well documented in the Iron Age and Roman era of southern Britain. In this region numerous settlement sites have been subjected to flotation and archaeobotanical analysis. Examples from these sites show that individual sites display fairly consistent patterns in terms of weed, crop grain, and chaff assemblages, but that not all sites are uniform. Thus, while some sites consistently show one pattern, in terms of the proportion of key elements, other sites show another, implying that different groups of sites have different systems for the organization of labor (Stevens 2003b). Two patterns of archaeobotanical data with implied differences in labor organization can be illustrated by comparing three Iron Age hillfort sites with three smaller Iron Age settlements in the Thames Valley (for another case study where patterns in archaeobotanical data might be related to labor organization, see Walshaw in this volume). The archaeobotanical data from these sites are plotted in figure 9.6. This plot shows the proportion of weed seeds to grains, against the proportion of large to small weed seeds. In general, smaller weed seeds decline comparatively to grain and large weed seeds as we progress through the processing sequence (figure 9.3). Thus the lower right-hand side of the diagram contains samples that represent waste from only the last processing stages (figure 9.6). Hence

Routine Activities, Tertiary Refuse, Labor Organization 195

Figure 9.6. Data from six sites in Iron Age Britain, plotted in terms of the crop-processing stage indicators of weed seed proportions, as outlined in figure 9.3. (Data sources: Danebury [M. Jones 1984], Balksbury [De Moulins 1995], Asheldam Camp [Murphy 1991], Groundwell West [Stevens and Wilkinson 2001], Sherbourne House Lechlade, Gloucestershire [Stevens 2003b], Mingies Ditch [M. Jones 1993].)

they are indicative of crops stored probably as semiclean spikelets or grain (figure 9.4). Samples falling in the top left-hand corner are then representative not just of waste from the earlier processing stages but also, because they are also frequently rich in glumes (Stevens 2003b), of waste from the entire sequence. They are then representative of waste from crops that were stored relatively unclean, perhaps as partially threshed ears or even sheaves (figure 9.4). What is striking is the separation of the two distributions, with a modal tendency. All of the samples from the Thames Valley sites clearly indicate more processing stages routinely contributing to archaeobotanical evidence, while those from the hillfort sites indicate fewer stages. The implication of this is a division between these two groups of sites in terms of how processing was organized and how much labor was normally mobilized. At the three hillfort sites, with archaeological features indicating their importance as communal centers (Cunliffe 1984; Wainwright and Davies 1995) and a fortified site, crops were stored more fully processed

196

social archaeobotany

and had therefore arrived on the site having undergone a greater amount of processing and hence labor input during the harvest period, probably due to the mobilization of a larger workforce. By contrast, at the Thames Valley sites, more typical of small, dispersed Iron Age settlements, the crop appears to have been stored as unclean spikelets/grains or even sheaths, with little additional labor being employed during the harvest period. This then led to the bulk of all of the processing stages, and thus a greater range of weed species (and chaff), being represented in the archaeobotanical samples. This pattern has implications for the diversity of modes of labor organization in Iron Age Britain. Indeed, one might expect differences within each community in their ability to call upon labor in the increasingly complex and hierarchical social organization of this period. Whether through some centralized pull of labor or through different scales of household organization, these groups of contemporary sites differ (Fuller and Stevens 2009). That this difference is highly embedded within social formations, rather than being either transient or regional, is suggested by the fact that both patterns are found in the Thames Valley and endure for the sites’ entire occupations, extending for periods of up to a millennium (Stevens 2003b). One of the most intriguing implications of these data is that there are at least two alternative models for how coexisting communities organized labor and dealt with harvest and crop processing. These approaches reflect different scales of labor mobilization, with more people or more time relative to the quantity of crop in those settlements storing as semiclean spikelets, versus those with less time and expendable labor storing as sheaths/ears. Furthermore, these approaches were consistent within communities and showed a remarkable consistency over time. Indeed, the conservatism of crop-processing/labor-mobilization patterns of particular communities is remarkable, and appears not to have been transformed by the political change that is Romanization (Stevens 2003b). As in the case of Hund, the case of Iron Age and Roman Britain suggests that sometimes changes in who is running a society may have little impact on how communities organize their most basic social routines.

Putting Labor in the Landscape: Ashmounds and Villages of Neolithic South India A routine versus seasonal processing perspective on archaeobotanical remains can provide important insights into the wider social systems of an archaeological culture. As an example, we will take the South Indian Neolithic, where agricultural village sites with evidence of the cultivation and routine processing of native small millets and pulses occur contemporary

Routine Activities, Tertiary Refuse, Labor Organization 197

Figure 9.7.  Schematic representation of inferred seasonal movement pattern of transhumant pastoralists during the South Indian Neolithic.

with isolated “ashmound” sites (figure 9.7) formed by cyclical accumulations and conflagrations of cattle dung (Fuller 2001, 2003; Fuller, Korisettar, and Venkatasubbaiah 2001; Fuller et al. 2004; Korisettar et al. 2001; Korisettar, Venkatasubbaiah, and Fuller 2001). These two kinds of archaeological sites are very different, but by considering them in terms of routine crop processing versus seasonal mass processing and redistribution, they can be understood as an integrated system in the scheduling of food production, labor, pastoral herd movement, and ritual gatherings. The agricultural settlement sites were often located on the flat summits of granite tors (inselbergs) that break the flat plains of the Mysore Plateau. These hilltop sites often include deeply stratified archaeological deposits and have yielded much artifactual evidence, as well as structural features in the form of round huts. Bone refuse from animal consumption is frequent, as is charcoal, including recurrent assemblages of crops and some weeds (figure 9.8). The crops represent a Neolithic package domesticated within the region, consisting of two small millet taxa, browntop millet and bristly foxtail (Brachiaria ramosa L. and Setaria verticillata [L.] P. Beauv.), and two grain legumes, the mungbean (Vigna radiata [L.] R. Wilczek) and horsegram (Macrotyloma uniflorum [Lam.] Verdc.) (Fuller, Korisettar, and Venkatasubbaiah 2001; Fuller et al. 2004). These are regarded as major foodstuffs, both because they are known crops today (some admittedly very rare) and because, as charred evidence, they are likely to reflect routine processing waste rather than dung. In addition to the general reasons already given above, dung can be excluded on the grounds of the extremely restricted taxa diversity (Fuller 2003).

198

social archaeobotany

Figure 9.8. Typical flotation sample from Sanganakallu, a hilltop village site in South India, represented as relative frequencies of major categories.

Two grasses recur in all samples, with very few other grass remains. This is despite being located in an ecologically zone rich with savannah grasses, including some 120 species, many of which are ecologically predicted to be much more common than the rare millets that have been selected as crops. In addition, charred dung fragments are absent. The small quantity of weed seeds points toward later processing stages, while the fact that the wild taxa encountered were consistently similar in size to millet caryopses suggests that these weeds accompanied the millets. A small percentage of the millet grains had fragments of husk adhering to them, indicating that these are grains that had been incompletely dehusked or were not yet dehusked, the kind of waste that develops around the outside of a mortar when dehusking. It is these accidentally lost, still-hulled grains, as well as some lost dehusked grains, that are most likely to be swept up with other waste, like husks and weeds seeds, and thrown onto the fire. The lack of husks is simply a product of the greatly biased destruction of husks when burned. Local dehusking is suggested by the numerous mortarlike depressions in the granite boulders that surround these sites. These millet waste groups have then become mixed with pulse waste, perhaps from the accidental loss of pulses in dry roasting. On the whole, archaeobotanical assemblages are highly consistent through individual southern Neolithic sites, and across sites (some twelve were studied by Fuller), and indicate the importance of summer- (monsoon-) grown crops, which were then processed throughout much of the year at the hilltop sites. By contrast, the evidence of ashmounds bespeaks seasonal (or sporadic), shorter-term encampments. The role of pastoralism at these sites is clear from evidence for penning, dung accumulation, and animal bones (Allchin 1963; Paddayya 1998), but plant food consumption is implied

Routine Activities, Tertiary Refuse, Labor Organization 199 by the presence of a number of quern stones. Abundant charred remains of crops are, however, lacking. Although the lack of routine processing on these sites is confounded with the lack of densely stratified occupation layers, the ashmounds provide a clear contrast with the hilltop villages, where routine millet dehusking and pulse roasting were conducted. The artifacts, plant assemblages, and nature of archaeological deposits between settlements and ashmounds are highly dissimilar (Korisettar, Venkatasubbaiah, and Fuller 2001b). Nevertheless, because of their contemporaneity, as well as shared ceramic and lithic repertoires, these sites should be linked into one social system, and thus the ashmounds can be interpreted as seasonal encampments and festival centers to which staple plant foods are brought in small quantity from village sites in the regional settlement system, but not as sites of crop storage with routine poststorage processing.

Linking Productive Labor: Postharvest Processing and Pottery Production As the South Indian and British cases illustrate, the organization of labor for crop processing can be usefully considered in a regional and seasonal landscape context. Agricultural labor can be scheduled for particular times of the year and at particular places in the settlement system. So it is with other forms of productivity, such as raw material procurement and activities of artifact fabrication. When all these areas of activity can be linked, a more holistic understanding of ancient social organization should become possible. In the case of pottery production, it may be possible to more directly link agricultural labor when agricultural products, and by-products such as chaff, are incorporated into ceramics. An illustrative case from Bronze Age Ireland will be explored (concentrating on middle and late Bronze Age sites owing to the very small quantity of early Bronze Age assemblages), with some brief comments on how the linkage in the monsoonal tropics may necessarily be different, conditioned by different climatic seasonality. New research into arable agricultural systems of Bronze Age Ireland has highlighted the need for consideration of a range of archaeobotanical data in the reconstruction of cereal economies. Studies of arable agricultural systems throughout the world have regularly utilized data from seed and chaff impressions on ceramic vessels in the reconstruction of past economies (e.g., Costantini 1983; Helbaek 1952, 1959; Jessen and Helbaek 1944; Klee and Zach 1999; Stemler 1990; Vishnu-Mittre 1969). Cereal components can become incorporated into ceramic vessels during manufacture and may be preserved through charring, or may

200

social archaeobotany

be destroyed during the firing of a pot, leaving morphologically identifiable impressions of the material remaining in the fabric of vessels. It has regularly been proposed for Ireland and Britain that the incorporation of cereals into the fabric of prehistoric ceramic vessels is a result of the presence of crops in manufacturing areas, whereby components are inadvertently incorporated (Cleary 1987, 35; Godwin 1975, 405; Jessen and Helbaek 1944, 10). The actual identification of chaff as a temper in Irish vessels has proved to be a contentious issue (Cleary 2000, 125–27; Ó Ríordáin 1954, 327; Sheridan 1993, 49). The intentional inclusion of cereals may, however, have occurred due to technical requirements of potters, for example, in the use of chaff as a tempering agent (Gibson 2002, 35; 2003, vi; Gibson and Woods 1997, 33), and may also have occurred for symbolic, social, or stylistic reasons (Darvill 2004, 204n2; Gibson 2003, vi; Schiffer and Skibo 1987, 596). Cereal components may even have represented a valuable commodity for ceramic manufacturers as tempering agents and also for use in fires. However, the utilization of mineral tempers and certain organic tempers, such as bone and shell, have often received far more attention than cereals and grasses. Previous studies have proposed that the frequency of various cereal types recorded from ceramic vessels represents the relative economic importance of each cereal type (Costantini 1983; Godwin 1975, 405; Helbaek 1952; Jessen and Helbaek 1944, 10; Possehl 1999, 459). Others have argued that a range of processes and behavioral patterns affected the ways in which cereals were incorporated into ceramic vessels, and that the predominance of certain cereal types at various times is unlikely to be related to their economic importance. Evidence from cereal impressions may reveal little about local agricultural systems, as ceramic vessels may have been deposited at a considerable distance from their location of manufacture (Dennell 1976b, 13). Hubbard (1975, 200) has suggested that the types of crops predominant in cereal impressions may reflect particular activities that incorporate various cereals, rather than their overall economic status, while M. Jones (1980) has noted that the absence of cereal impressions on particular pottery styles does not mean that crops were economically unimportant to the manufacturers and consumers of those vessels. Until recently, seed and other plant impressions on ceramic vessels constituted the main macroremains evidence for arable agriculture in Bronze Age Ireland (Jessen and Helbaek 1944; Monk 1986, 32–33). Jessen and Helbaek’s study (1944) of cereal impressions on more than twenty Bronze Age vessels indicated that barley—naked barley in particular—was by far the predominant cereal type recorded, with occasional evidence for hulled barley. Although we might now ascribe vessels to different subperiods of the Bronze Age than those ascribed by Jessen

Routine Activities, Tertiary Refuse, Labor Organization 201 and Helbaek (many of the vessels are now thought to date to the earlier Bronze Age), the general pattern that Jessen and Helbaek encountered remained unchanged throughout the Bronze Age period. A small number of later publications noted the rare occurrence of wheat, naked where identified as to variety, as well as more evidence for barley (Hartnett 1957, 259; Ó Ríordáin and Waddell 1993, 113 and 126). It seemed clear from these studies that wheat played a very minor role in Bronze Age agriculture, while barley, particularly naked, was the focus of arable activity at this time (figure 9.9). The validity of Jessen and Helbaek’s study in reconstructing arable economies of the period has occasionally been questioned (Monk 1986, 32–33), since the ceramic vessels originated from mainly funerary contexts in the north and east of Ireland only, rather than from settlement areas throughout the island. It was also questioned whether the preference for barley as a tempering agent necessarily represented the economic predominance of this cereal. Such suspicions, until now, remained speculative. The recent collation (by McClatchie) of around twenty published and unpublished archaeobotanical assemblages of charred macroremains from Bronze Age sites in Ireland, many of which were associated with settlements that were distributed over many parts of the island, has provided a very different picture with regard to the economic status of various cereal types (figure 9.9). Wheat seems to have been much more significant (particularly during the middle Bronze Age) than previously considered, and hulled barley also played a prominent role, the latter being recorded on 43 percent of middle Bronze Age sites and 38 percent of late Bronze Age sites. It is clear that this new study does not correlate well with Jessen and Helbaek’s findings from analysis of cereal impressions. The exclusive use of evidence from cereal impressions in ceramic vessels does not, therefore, seem appropriate in determining the economic roles of various cereal types in arable agricultural economies of this period. The strong association of naked barley with ceramic vessels may be better viewed as representing a relationship between activities associated with the processing of naked barley and ceramic vessel production. It may be significant that naked barley, and in one incidence free-threshing wheat, are predominant in the cereals identified from seed impressions, while hulled barley and glume wheats are less well represented. It is possible that crops requiring a greater amount of processing to extract grains, such as hulled barley and glume wheats, were not fully processed before storage, being stored in spikelet or sheaf form. Indeed, the only incidence of hulled barley in the seed impression record is of a floret rather than a seed alone (Jessen and Helbaek 1944, 21). Processing crops to spikelet or sheaf stage would lessen their chance of being included as a tempering agent or inadvertently being incorporated into ceramic vessels, as

202

social archaeobotany

Figure 9.9. Ceramic impressions versus macroremains. The chart at the top shows the relative frequency of major cereal types recorded as impressions on ceramic vessels from Bronze Age Ireland, as well as charred macroremains recovered from archaeological deposits dating to the same period. The chart on the bottom provides more detailed information on the types of cereals recorded from middle and late Bronze Age macroremains assemblages. (Data sources for seed impressions: Hartnett [1957], Jessen and Helbaek [1944], Ó Ríordáin and Waddell [1993]. Data sources for macroremains: Brewer [2002, 2003], Church [unpublished data], Collins [unpublished data], Doyle [2001], Johnston [2007], McClatchie [2014], Monk [1987a, 1987b], Tierney and Hannon [2003], Weir [1996].)

their mass would have made them unsuitable. Free-threshing wheats and naked barley may have been more fully processed at an earlier stage, separating grains from lighter chaff, which would therefore have constituted more suitable material for tempering or unintentional incorporation. The processing of naked barley and wheat to an advanced stage may have occurred when a large number of people were mobilized, for

Routine Activities, Tertiary Refuse, Labor Organization 203

Figure 9.10. The inferred basic seasonal cycle for Bronze Age Ireland.

example, at harvest time (figure 9.10). If the cereals were spring sown, harvesting would have occurred during the autumn period. This also coincides with a time when the production of ceramic vessels would have been advantageous, coming at the end of the driest season, thus facilitating the preparation and firing of ceramic vessels, which need to first be slowly dried to the “leatherhard” state, as well as the preparation of fuel (Arnold 1985, 61–77). Harvest time in temperate areas like Ireland may have roughly coincided with ceramic production, the latter being to some extent a seasonal activity, with a concentrated production of vessels being scheduled to occur at a time around harvest. It has previously been suggested that pottery production was a seasonal endeavor, since cereals were available for incorporation into fabrics at harvest time (Howard 1981, 25). Although this approach fails to recognize that cereals can be stored, and therefore utilized, over relatively long periods, the ready availability of cereals, particularly the recently processed chaff of naked cereals, at harvest time represents another reason why pottery production would have been more favorable at this time of year. While ceramic production could undoubtedly have been a year-round activity (Gibson and Woods 1997, 46–48), particularly when dedicated drying facilities were constructed, it does seem more beneficial

204

social archaeobotany

to produce vessels around harvest time when environmental factors are advantageous for drying unfired vessels. At this time, large numbers of people are mobilized for harvest and processing, and some proportion of them could also be involved in the production and distribution of vessels. The potential scheduling conflict (Arnold 1985, 99; Kramer 1985, 80) might suggest that those people involved in potting were focused on this activity rather than agricultural labor, implying some degree of specialization, at least seasonally. A ready supply of cereal components would become available for use as temper (or, if inadvertently incorporated, would have been present in vessel manufacturing areas), but significantly only of those species that are more readily fully processed into clean grain and light chaff components, such as naked barley. Hulled cereals—including hulled barley and emmer wheat (Triticum dicoccum Schrank)—are instead likely to have been stored as semiprocessed spikelets or even as sheaths. Plant impressions in ceramics, when considered both as temper and as crop-processing waste, provide important insights into the links between labor deployed in food production and another productive labor—potting. During the seasonal period of harvesting and mass processing prior to storage, we expect that most societies will produce abundant straw and naked cereal chaff, while an even greater amount of labor would be required to fully clean hulled cereals, such as glume wheats. For this reason, we might expect glume wheats to be proportionately underrepresented in pottery impressions. Yet, a review of the Neolithic and early Bronze Age impressions from England, identified by Helbaek (1952), provides a stark contrast to the evidence recovered from these periods since the application of regular flotation. In the ceramics, emmer wheat spikelet forks and whole spikelets regularly occur, suggesting that if this is the waste of postharvest mass processing, hulled cereals were being dehusked. By contrast, most Neolithic seed assemblages, although generally poor, produce cereal grains without chaff (Robinson 2000), totally unlike the macroremains evidence for later periods (e.g., the Iron Age and Roman period discussed above) and contemporary ceramic impressions (Stevens 2007). However, these contrasts between on-site flotation samples and ceramic impressions can be seen as indicative of seasonal versus routine practices of cereal processing, with the seasonal constraints on potting coinciding with those of the agricultural season. By contrast, under different seasonal constraints, the relationship between charred remains and pottery impressions may be markedly changed (figure 9.11). In India, where the summer brings monsoon rains that provide the basis for much cultivation, potting is impossible at that time because of the rains and humid atmosphere. The mass, prestorage

Routine Activities, Tertiary Refuse, Labor Organization 205

Figure 9.11. The inferred basic seasonality for monsoonal India, such as the Ganges Neolithic.

processing of monsoon crops, both rice and millets, takes place at the end of the monsoon. Traditional potting in India occurs in the winter and spring months—the dry season (Arnold 1985; Kramer 1985). During these months, daily processing takes place. It should therefore come as no surprise that in monsoonal rice growing regions, such as the Ganges valley, vegetable-tempered ceramics, found at Neolithic sites like Senuwar, Mahagara, and Koldihwa, are dominated by rice husks, with the occasional whole rice spikelet (observations by Fuller; see also Saraswat 2004; Sharma et al. 1980, pl. MGR XVI; Vishnu-Mittre 1969), precisely the dehusking waste we would expect from daily processing during the dry season when potting is possible. This suggests that in wet-summer tropical regions, chaff temper in ceramics is more likely to derive from routine daily waste, whereas within temperate regions with wet winters and drier summers, the late summer or autumn harvest processing is more likely to be linked to ceramic tempering. Such expected patterns, at least, provide a basis for developing a comparative analysis of the relationships between ceramic production and agricultural activities across periods and cultures.

206

social archaeobotany

Labor Scheduling as an Agenda for Social Archaeobotany Archaeobotany can become an essential component of a contextual archaeology only once we accept that its interpretation does not rely on archaeological context. Standard charred archaeobotanical samples represent redeposition from fire contexts, where plant remains may already represent secondary refuse. Archaeobotanical assemblages are unlikely to be readily interpretable from, or contribute to, the understanding of particular depositional contexts in terms of human activities. This means that only very rarely will flotation samples contribute to studies of spatial patterning and activity areas on archaeological sites, except perhaps at a very coarse scale on the very largest sites. Despite their tertiary nature, however, the recurrent patterns of archaeobotanical assemblages across sites reflect recurrent practices in the past—the routine. As such, they provide an important window on traditions of daily household labor. These patterns of daily labor can then be compared between sites, between cultural phases and regional cultural traditions, to build up a larger comparative perspective on the evolution of systems of human labor organization. In this way archaeobotanical evidence contributes to the contextualization of sites in terms of labor organization and food-production strategies. Productive activities related to food production or procurement and storage represent important scheduling decisions. It is obviously the case that the availability of many foods is seasonal. With the exception of modern, industrialized supermarket economies, seasonal patterns in food consumption are ethnographically and historically universal (De Garine 1994). The seasonality of labor needs in relation to agricultural production is an important arena that creates need for assistance between human groups, such as between households, and provides a recurrent situation in which relationships of social debt and reciprocity develop (see, e.g., Dietler and Herbich 2001; Peletz 1992; Stone, Netting, and Stone 1990). The size of households or other social groups that can be organized for harvest-period processing versus the daily labor requirement of food preparation in the household are reflected in archaeobotanical assemblages. Similarly, other productive activities, such as craft production, must be scheduled, either to avoid conflict with labor needed elsewhere or to take advantage of shared resources and weather conditions. With the rise of increasingly complex societies, more craft production might be supported by redistributed surplus sequestered from the labor of others, and increasingly craftsman are freed from immediate ties to the production and labor schedule of domestic food production. Archaeobotanical evidence, when analyzed through crop-processing models in relation to the social context of seasonal scheduling and scale of labor

Routine Activities, Tertiary Refuse, Labor Organization 207 groups, has the potential to contribute to comparative studies of social structure and social evolution.

Acknowledgments Fieldwork at Hund was made possible by Professor Ihsan Ali and the hospitality of the University of Peshawar Archaeology Department. We would like to thank Anies Hassan and Miriam Cooke for their hard work in analyzing some of the data from the site of Hund. Research carried out in South India was done in collaboration with Professor Ravi Korisettar (Karnatak University, Dharwad) and Dr. P. C. Venkatasubbaiah (Andhra University, Kuppam). Fieldwork and research visits by Dorian Fuller to Mahagara and Koldihwa in North India were made possible by Professor J. N. Pal and Dr. S. C. Gupta of Allahabad University, and archaeobotanical analyses were carried out by Emma Harvey. Current research on Comparative Pathways to Agriculture (ComPAg) by D. Q F. and C. J. S. is supported by European Research Council Grant 323842. Meriel McClatchie would like to thank the following archaeobotanists and archaeologists for access to unpublished data from Irish Bronze Age sites: Claire Breen, Thaddeus Breen, Abigail Brewer, Seamus Caulfield, Mike Church, Eamon Cody, Brenda Collins, Clare Cotter, Geraldine Crowley, Ian Doyle, Martin Doody, Lar Dunne, James Eogan, Audrey Gahan, Margaret Gowen, Alan Hayden, Penny Johnston, Valerie Keeley, Jacinta Kiely, Paul Logue, Rob Lynch, Jim Mallory, Catherine McLoughlin, Cormac McSparron, Mick Monk, Charles Mount, Richard O’Brien, Edmond O’Donovan, John O’Neill, Hilary Opie, Gill Plunkett, Martin Reid, Barry Raftery, Emmett Stafford, Ian Suddaby, John Tierney, Gerry Walsh, Paul Walsh, and David Weir. McClatchie would also like to thank staff at the National Museum of Ireland for access to Bronze Age ceramic vessels.

References Ali, I. 1999. “Excavations at Hund on the Banks of the Indus: The Last Capital of Gandhara.” In The Indus River: Biodiversity, Resources, Humankind, edited by A. Meadows and P. S. Meadows, 269. Karachi: Oxford University Press. Allchin, F. R. 1963. Neolithic Cattle Keepers of South India: A Study of the Deccan Ashmounds. Cambridge: Cambridge University Press. Arnold, D. E. 1985. Ceramic Theory and Cultural Process. Cambridge: Cambridge University Press. Bailey, G. 1981. “Concepts, Time-Scales and Explanations in Economic Prehistory.” In Economic Archaeology, edited by A. Sheridan and G. Bailey, 97–117. BAR International Series, vol. 96. Oxford: Archaeopress.

208

social archaeobotany

Behre, K. E., and S. Jacomet. 1991. “The Ecological Interpretation of Archaeobotanical Data.” In Progress in Old World Palaeoethnobotany: A Retrospective View on the Occasion of 20 Years of the International Work Group for Palaeoethnobotany, edited by W. van Zeist, K. Wasylikowa, and K. E. Behre, 81–108. Rotterdam: A. A. Balkema. Boardman, S., and G. E. M. Jones. 1990. “Experiments on the Effects of Charring on Cereal Plant Components.” Journal of Archaeological Science 17 (1): 1–12. Bogaard, A. 2002. “Questioning the Relevance of Shifting Cultivation to Neolithic Farming in the Loess Belt of Europe: Evidence from the Hambach Forest Experiment.” Vegetation History and Archaeobotany 11 (1–2): 155–68. Brewer, A. 2002. “Appendix 3: Plant Remains from Cappamore.” In Archaeological Excavation Report, Cappamore, Co. Limerick, Excavation Licence no. 00E0595, edited by J. Kiely. Unpublished report. ———. 2003. “Plant Remains for Cloghers, 00E0065.” In Archaeological Excavation Report, Cappamore, Co. Limerick, Excavation Licence no. 00E0595, edited by J. Kiely. Unpublished report. Butler, A., Z. Tesfay, C. D’Andrea, and D. Lyons. 1999. “The Ethnobotany of Lathyrus sativus L. in the Highlands of Ethiopia.” In The Exploitation of Plant Resources in Ancient Africa, edited by M. van der Veen, 123–36. New York: Kluwer Academic/Plenum. Butzer, K. W. 1982. Archaeology as Human Ecology. Cambridge: Cambridge University Press. Charles, M. 1998. “Fodder from Dung: The Recognition and Interpretation of Dung-Derived Plant Material from Archaeological Sites.” Environmental Archaeology 1:111–22. Cleary, R. M. 1987. “Appendix VI: The Pottery from Ballyveelish 2 and 3.” In Archaeological Excavations on the Cork-Dublin Gas Pipeline (1981–82), edited by R. M. Cleary, M. F. Hurley, and E. A. Twohig, 31–35. Cork: Department of Archaeology, University College, Cork. ———. 2000. “The Potter’s Craft in Prehistoric Ireland with Specific Reference to Lough Gur, Co. Limerick.” In New Agendas in Irish Prehistory: Papers in Commemoration of Liz Anderson, edited by A. Desmond, G. Johnson, M. McCarthy, J. Sheehan, and E. Shee-Twohig, 119–34. Wicklow: Wordwell. Cooke, M. 2002. “Investigating Changing Agricultural Production and Patterns of Subsistence at Hund, in the Vale of Peshawar, Pakistan.” BSc diss., Institute of Archaeology, University College London. Costantini, L. 1983. “The Beginning of Agriculture in the Kachi Plain: The Evidence of Mehrgarh.” In South Asian Archaeology 1981, edited by B. Allchin, 29–33. Cambridge: Cambridge University Press. Cunliffe, B. 1984. Danebury: An Iron Age Hillfort in Hampshire. Vol. 2, The Excavations 1969–1978: The Finds. Council for British Archaeology Research, Research Report no. 52. London: Council for British Archaeology.

Routine Activities, Tertiary Refuse, Labor Organization 209 D’Andrea, C., D. Lyons, M. Haile, and A. Butler. 1999. “Ethnoarchaeological Approaches to the Study of Prehistoric Agriculture in the Highlands of Ethiopia.” In The Exploitation of Plant Resources in Ancient Africa, edited by M. van der Veen, 101–22. New York: Kluwer Academic/Plenum. Darvill, T. 2004. “Soft-Rock and Organic Tempering in British Neolithic Pottery.” In Monuments and Material Culture: Papers in Honour of an Avebury Archaeologist: Isobel Smith, edited by R. Cleal and J. Pollard, 193–206. Salisbury: Hobnob. De Garine, I. 1994. “The Diet and Nutrition of Human Populations.” In Companion Encyclopedia of Anthropology: Humanity, Culture and Social Life, edited by T. Ingold, 226–64. London: Routledge. De Moulins, D. 1995. “Charred Plant Remains.” In Balksbury Camp, Hampshire, Excavations 1973 and 1981, edited by G. J. Wainwright and S. M. Davies, 87–92. English Heritage Archaeological Report 4. London: English Heritage. Dennell, R. W. 1972. “The Interpretation of Plant Remains: Bulgaria.” In Papers in Economic Prehistory, edited by E. S. Higgs, 149–59. Cambridge: Cambridge University Press. ———. 1974. “Botanical Evidence for Prehistoric Crop Processing Activities.” Journal of Archaeological Science 1:275–84. ———. 1976a. “The Economic Importance of Plant Resources Represented on Archaeological Sites.” Journal of Archaeological Science 3:229–47. ———. 1976b. “Prehistoric Crop Cultivation in Southern England: A Reconsideration.” Antiquaries Journal 56:11–23. Dietler, M., and I. Herbich. 2001. “Feasts and Labor Mobilization: Dissecting a Fundamental Economic Practice.” In Feasts: Archaeological and Ethnographic Perspectives on Food, Politics, and Power, edited by M. Dietler and B. Hayden, 240–64. Washington, DC: Smithsonian Institution Press. Doyle, I. 2001. “A Prehistoric Ring-Barrow in Kilmahuddrick, Co. Dublin.” Archaeology Ireland 1 (4): 16–19. Fuller, D. Q. 2001. “Ashmounds and Hilltop Villages: The Search for Early Agriculture in Southern India.” Archaeology International 4:43–46. ———. 2002. “Fifty Years of Archaeobotanical Studies in India: Laying a Solid Foundation.” In Indian Archaeology in Retrospect, vol. 3, Archaeology and Interactive Disciplines, edited by S. Settar and R. Korisettar, 247–364. Delhi: Manohar. ———. 2003. “Indus and Non-Indus Agricultural Traditions: Local Developments and Crop Adoptions on the Indian Peninsula.” In Indus Ethnobiology: New Perspectives from the Field, edited by S. A. Weber and W. R. Belcher, 343–96. Lanham: Lexington Books. Fuller, D. Q, R. Korisettar, and P. C. Venkatasubbaiah. 2001. “Southern Neolithic Cultivation Systems: A Reconstruction Based on Archaeobotanical Evidence.” South Asian Studies 17:171–87.

210

social archaeobotany

Fuller, D. Q, R. Korisettar, P. C. Venkatasubbaiah, and M. K. Jones. 2004. “Early Plant Domestications in Southern India: Some Preliminary Archaeobotanical Results.” Vegetation History and Archaeobotany 13:115–29. Fuller, D. Q, and C. J. Stevens. 2009. “Agriculture and the Development of Complex Societies.” In From Foragers to Farmers: Papers in Honour of Gordon C. Hillman, edited by A. Fairbairn and E. Weiss, 37–57. Oxford: Oxbow Books. Gibson, A. 2002. “Aspects of Manufacture and Ceramic Technology.” In Prehistoric Britain: The Ceramic Basis, edited by A. Woodward and J. D. Hill, 34–37. Oxford: Oxbow. ———. 2003. “Prehistoric Pottery: People, Pattern and Purpose: Some Observations, Questions and Speculations.” In Prehistoric Pottery: People, Pattern and Purpose, edited by A. Gibson, v–xi. BAR International Series, vol. 1156. Oxford: Archaeopress. Gibson, A., and A. Woods. 1997. Prehistoric Pottery for the Archaeologist. 2nd ed. London: Leicester University Press. Godwin, H. 1975. The History of the British Flora. 2nd ed. Cambridge: Cambridge University Press. Hartnett, P. J. 1957. “Excavation of a Passage Grave at Fourknocks, Co. Meath.” Proceedings of the Royal Irish Academy, Section C, 58:197–277. Hassan, A. 2000. “An Initial Investigation into Aspects of Archaeobotany: Applying Phytolith Analysis from the Site of Hund, Pakistan.” MSc diss., Institute of Archaeology, University College London. Hastorf, C. 1988. “The Use of Palaeoethnobotanical Data in the Prehistoric Studies of Crop Production, Processing, and Consumption.” In Current Palaeo­ ethnobotany: Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, 119–44. Chicago: University of Chicago Press. ———. 1999. “Recent Research in Paleoethnobotany.” Journal of Archaeological Research (1): 55–103. Helbaek, H. 1952. “Early Crops in Southern England.” Proceedings of the Prehistoric Society 18:194–233. ———. 1959. “The Domestication of Food Plants in the Old World.” Science 130:365–72. ———. 1969. “Plant-Collecting, Dry-Farming and Irrigation Agriculture in Prehistoric Deh Luran.” In Prehistory and Human Ecology of the Deh Luran Plain: An Early Village Sequence from Khuzistan, Iran, edited by F. Hole, K. V. Flannery, and J. A. Neely, 383–426. Memoirs of the Museum of Anthropology, 1. Ann Arbor: University of Michigan. Hillman, G. C. 1973. “Crop Husbandry and Food Production: Modern Models for the Interpretation of Plant Remains.” Anatolian Studies 23:241–44. ———. 1981. “Reconstructing Crop Husbandry Practices from Charred Remains of Crops.” In Farming Practice in British Prehistory, edited by R. Mercer, 123–62. Edinburgh: Edinburgh University Press.

Routine Activities, Tertiary Refuse, Labor Organization 211 ———. 1984a. “Interpretation of Archaeological Plant Remains: The Application of Ethnographic Models from Turkey.” In Plants and Ancient Man: Studies in Palaeoethnobotany, Proceedings of the Sixth Symposium of the International Work Group for Paleoethnobotany, edited by W. van Zeist and W. A. Casparie, 1–41. Rotterdam: A. A. Balkema. ———. 1984b. “Traditional Husbandry and Processing of Archaic Cereals in Recent Times: The Operations, Products and Equipment Which Might Feature in Sumerian Texts, Part I: The Glume Wheats.” Bulletin on Sumerian Agriculture 1:114–52. ———. 1985. “Traditional Husbandry and Processing of Archaic Cereals in Recent Times, Part II: The Free-Threshing Cereals.” Bulletin on Sumerian Agriculture 2:1–31. ———. 1991. “Phytosociology and Ancient Weed Floras: Taking Account of Taphonomy and Changes in Cultivation Methods.” In Modeling Ecological Change— Perspectives from Neoecology, Palaeoecology and Environmental Archaeology, edited by D. R. Harris and K. D. Thomas, 27–40. London: UCL Press. ———. 2004. “Investigating the Start of Cultivation in Western Eurasia: Studies of Plant Remains from Abu Hureyra on the Euphrates.” In The Widening Harvest: The Neolithic Transition in Europe: Looking Back, Looking Forward, edited by A. J. Ammerman and P. Biagi, 75–97. Boston: Archaeological Institute of America. Hillman, G. C., A. J. Legge, and P. A. Rowley-Conwy. 1997. “On the Charred Seeds from Epipalaeolithic Abu Hureyra.” Current Anthropology 38 (4): 651–55. Hodder, I. 1991. Reading the Past. 2nd ed. Cambridge: Cambridge University Press. Howard, H. 1981. “In the Wake of Distribution: Towards an Integrated Approach to Ceramic Studies in Prehistoric Britain.” In Production and Distribution: A Ceramic Viewpoint, edited by H. Howard and E. Morris, 1–30. BAR International Series, vol. 120. Oxford: Archaeopress. Hubbard, R. N. L. B. 1975. “Assessing the Botanical Component of Human Palaeo-­Economies.” Bulletin of the Institute of Archaeology 12:197–205. ———. 1976a. “Crops and Climate in Prehistoric Europe.” World Archaeology 8:159–68. ———. 1976b. “On the Strength of the Evidence for Prehistoric Crop Processing Activities.” Journal of Archaeological Science 3:257–65. Hubbard, R. N. L. B., and A. Clapham. 1992. “Quantifying Macroscopic Plant Remains.” Review of Palaeobotany and Palynology 73:132–73. Jessen, K., and H. Helbaek. 1944. Cereals in Great Britain and Ireland in Prehistoric and Early Historic Times. Biologiske Skrifter 3, no. 2. Copenhagen: Ejnar Minksgaard. Johnston, P. 2007. “Analysis of Carbonised Plant Remains.” In The Bronze Age Landscapes of the Pipeline to the West: An Integrated Archaeological and Environmental Assessment, edited by E. Grogan, 70–79. Wicklow: Wordwell.

212

social archaeobotany

Jones, G. E. M. 1984. “Interpretation of Archaeological Plant Remains: Ethnographic Models from Greece.” In Plants and Ancient Man: Studies in Palaeoethnobotany, Proceedings of the Sixth Symposium of the International Work Group for Paleoethnobotany, edited by W. van Zeist and W. A. Casparie, 42–61. Rotterdam: A. A. Balkema. ———. 1987. “A Statistical Approach to the Archaeological Identification of Crop Processing.” Journal of Archaeological Science 14:311–23. Jones, G. E. M., A. Bogaard, M. Charles, and J. G. Hodgson. 2000. “Distinguishing the Effects of Agricultural Practices Relating to Fertility and Disturbance: A Functional Ecological Approach in Archaeobotany.” Journal of Archaeological Science 27:1073–84. Jones, G. E. M., M. Charles, S. Colledge, and P. Halstead. 1995. “Towards the Archaeobotanical Recognition of Winter-Cereal Irrigation: An Investigation of Modern Weed Ecology in Northern Spain.” In Res Archaeobotanicae, edited by H. Kröll and R. Pasternak, 49–68. Kiel: Institut für Ur- und Fruhgeschichte der Christian-Albrecht-Universität. Jones, G. E. M., K. Wardle, P. Halstead, and D. Wardle. 1986. “Crop Storage at Assiros.” Scientific American 254:96–103. Jones, M. K. 1978. “The Plant Remains.” In The Excavation of an Iron Age Settlement, Bronze Age Ring-Ditches and Roman Features at Ashville Trading Estate, Abingdon (Oxfordshire), 1974–76, edited by M. Parrington, 93–110. Council for British Archaeology, Research Report 28. London: Council for British Archaeology. ———. 1980. “Carbonised Cereals from Grooved Ware Contexts.” Proceedings of the Prehistoric Society 46:61–63. ———. 1981. “The Development of Crop Husbandry.” In The Environment of Man, the Iron Age to the Anglo-Saxon Period, edited by M. K. Jones and G. Dimbleby, 95–127. BAR, British Series, vol. 87. Oxford: Archaeopress. ———. 1984. “The Plant Remains.” In Danebury: An Iron Age Hillfort in Hampshire, vol. 2, The Excavations 1969–1978: The Finds, edited by B. Cunliffe, 483–95. Council for British Archaeology, Research Report 52. London: Council for British Archaeology. ———. 1985. “Archaeobotany beyond Subsistence Reconstruction.” In Beyond Domestication in Prehistoric Europe, edited by G. W. Barker and C. Gamble, 107–28. New York: Academic Press. ———. 1991. “Sampling in Palaeoethnobotany.” In Progress in Old World Palaeoethnobotany: A Retrospective View on the Occasion of 20 Years of the International Work Group for Palaeoethnobotany, edited by W. van Zeist, K. Wasylikowa, and K. E. Behre, 53–62. Rotterdam: A. A. Balkema. ———. 1993. “The Carbonised Plant Remains.” In The Prehistoric and Iron Age Enclosed Settlement at Mingies Ditch; Hardwick-with-Yelford, Oxon. Thames Valley Landscapes: The Windrush Valley, vol. 2, Oxford, edited by T. G. Allen

Routine Activities, Tertiary Refuse, Labor Organization 213 and M. A. Robinson, 121–23. Oxford: Oxford Archaeology Unit/Oxford University Committee for Archaeology. Klee, M., and B. Zach. 1999. “The Exploitation of Wild and Domesticated Food Plants at Settlement Mounds in North-East Nigeria (1800 cal BC to Today).” In The Exploitation of Plant Resources in Ancient Africa, edited by M. van der Veen, 81–88. New York: Kluwer/Plenum. Knörzer, K. H. 1971. “Urgeschichtliche Unkrauter im Rheinland: Ein Beitrag zur Entstehung der Segetalgesellschaften.” Vegetation 23:89–111. Körber-Gröhne, U. 1967. Geobotanische Untersuchungen auf der Feddersen Wierde. Wiesbaden: Steiner. ———. 1981. “Distinguishing Prehistoric Grains of Triticum and Secale on the Basis of Their Surface Patterns Using Scanning Electron Microscopy.” Journal of Archaeological Science 8:197–204. Korisettar, R., P. P. Joglekar, D. Q Fuller, and P. C. Venkatasubbaiah. 2001. “Archaeological Re-investigation and Archaeozoology of Seven Southern Neolithic Sites in Karnataka and Andhra Pradesh.” Man and Environment 26:47–66. Korisettar, R., P. C. Venkatasubbaiah, and D. Q Fuller. 2001. “Brahmagiri and Beyond: The Archaeology of the Southern Neolithic.” In Indian Archaeology in Retrospect, vol. 1, Prehistory, edited by S. Settar and R. Korisettar, 151–238. New Delhi: Manohar. Kramer, C. 1985. “Ceramic Ethnoarchaeology.” Annual Review of Anthropology 14:77–102. Kreuz, A. 1990. “Searching for ‘Single-Activity Refuse’ in Linearbandkeramik Settlements: An Archaeological Approach.” In Experimentation and Reconstruction in Environmental Archaeology, edited by D. E. Robinson, 63–74. Oxford: Oxbow Books. Küster, H. 1991. “Phytosociology and Archaeobotany.” In Modeling Ecological Change—Perspectives from Neoecology, Palaeoecology and Environmental Archaeology, edited by D. R. Harris and K. D. Thomas, 17–26. London: UCL Press. La Motta, V. M., and M. B. Schiffer. 1999. “Formation Processes of House Floor Assemblages.” In The Archaeology of Household Activities, edited by P. M. Allison, 19–29. London: Routledge. ———. 2001. “Behavioral Archaeology: Toward a New Synthesis.” In Archaeological Theory Today, edited by I. Hodder, 14–64. Oxford: Blackwell. Lennstrom, H. A., and C. A. Hastorf. 1992. “Testing Old Wives’ Tales in Palaeoethnobotany: A Comparison of Bulk and Scatter Sampling Schemes from Pancán, Peru.” Journal of Archaeological Science 19:202–29. ———. 1995. “Interpretation in Context: Sampling and Analysis in Paleoethnobotany.” American Antiquity 6 (4): 701–21. Lundström-Baudais, K., A. M. Rachoud-Schneider, D. Baudais, and B. Poissonnier. 2002. “Le broyage dans la chaîne de transformation du millet (Panicum miliaceum): Outils, gestes et écofacts.” In Moudre et Broyer, vol. 1, Méthodes,

214

social archaeobotany

edited by H. Procopiou and R. Treuil, 155–80. Paris: Comité des Travaux Historiques et Scientifiques. Lyman, R. L. 1994. Vertebrate Taphonomy. Cambridge: Cambridge University Press. Madella, M. 2003. “Investigating Agriculture and Environment in South Asia: Present and Future Contributions from Opal Phytoliths.” In Indus Ethnobiology: New Perspectives from the Field, edited by S. Weber and W. R. Belcher, 199–249. Lanham: Lexington Books. McClatchie, M. 2014. “Food Production in the Bronze Age: Analysis of Plant Macro-remains from Haughey’s Fort, Co. Armagh.” Emania. Forthcoming. Miksicek, C. H. 1986. “Plant Remains from the Tanque Verde Wash Site.” In Archaeological Investigations at the Tanque Verde Wash Site, edited by M. Elson, 371–94. Tucson: Institute for American Research. ———. 1987. “Formation Processes of the Archaeobotanical Record.” In Advances in Archaeological Method and Theory, vol. 10, edited by M. B. Schiffer, 211–47. Albuquerque: University of New Mexico Press. Miller, N. F. 1984. “The Use of Dung as Fuel: An Ethnographic Example and an Archaeological Application.” Paléorient 10 (2): 71–79. ———. 1991. “The Near East.” In Progress in Old World Palaeoethnobotany: A Retrospective View on the Occasion of 20 Years of the International Work Group for Palaeoethnobotany, edited by W. van Zeist, K. Wasylikowa, and K. E. Behre, 133–60. Rotterdam: A. A. Balkema. ———. 1996. “Seed Eaters of the Ancient Near East: Human or Herbivore?” Current Anthropology 37 (3): 521–28. ———. 1997. “Reply to Hillman et al.” Current Anthropology 38 (4): 655–59. Miller, N. F., and T. L. Smart. 1984. “Intentional Burning of Dung as Fuel: A Mechanism for the Incorporation of Charred Seeds into the Archaeological Record.” Journal of Ethnobiology 4:15–28. Monk, M. A. 1986. “Evidence from Macroscopic Plant Remains for Crop Husbandry in Prehistoric and Early Historic Ireland: A Review.” Journal of Irish Archaeology 3:31–36. ———. 1987a. “Appendix 1: The Charred Plant Remains.” In Archaeological Excavations on the Cork–Dublin Gas Pipeline (1981–82), edited by R. M. Cleary, M. F. Hurley, and E. A. Twohig. Cork: Department of Archaeology, University College, Cork. ———. 1987b. “Appendix V: The Charred Plant Remains from Ballyveelish.” In Archaeological Excavations on the Cork–Dublin Gas Pipeline (1981–82), edited by R. M. Cleary, M. F. Hurley, and E. A. Twohig, 30–31. Cork: Department of Archaeology, University College, Cork. Moore, H. L. 1986. Space, Text and Gender: An Anthropological Study of the Marakwet of Kenya. Cambridge: Cambridge University Press. Murphy, P. 1991. “Cereals and Cropweeds.” Essex Archaeology and History 22:31–35.

Routine Activities, Tertiary Refuse, Labor Organization 215 Orton, C. 2000. Sampling in Archaeology. Cambridge: Cambridge University Press. Ó Ríordáin, B., and J. Waddell. 1993. The Funerary Bowls and Vases of the Irish Bronze Age. Galway: Galway University Press. Ó Ríordáin, S. P. 1954. “Lough Gur Excavations: Neolithic and Bronze Age Houses on Knockadoon.” Proceedings of the Royal Irish Academy, Section C, 5:297–459. Paddayya, K. 1998. “Evidence of Neolithic Cattle-Penning at Budihal, Gulbarga District, Karnataka.” South Asian Studies 14:141–53. Pearsall, D. 1988. “Interpreting the Meaning of Macro Remains Abundance: The Impact of Source and Contest.” In Current Paleoethnobotany: Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, 97–118. Chicago: University of Chicago Press. ———. 2000. Paleoethnobotany: A Handbook of Procedures. 2nd ed. New York: Academic Press. Peletz, M. G. 1992. A Share of the Harvest: Kinship, Property, and Social History among the Malays of Rembu. Berkeley: University of California Press. Peña-Chocarro, L. 1999. Prehistoric Agriculture in Southern Spain during the Neolithic and Bronze Age: The Application of Ethnographic Models. BAR International Series, vol. 818. Oxford: Archaeopress. Possehl, G. L. 1999. Indus Age: The Beginnings. Philadelphia: University of Pennsylvania Press. Reddy, S. N. 1997. “If the Threshing Floor Could Talk: Integration of Agriculture and Pastoralism during the Late Harappan in Gujarat, India.” Journal of Anthropological Archaeology 16:162–87. ———. 1999. “Fueling the Hearths in India: The Role of Dung in Paleoethnobotanical Interpretation.” Paléorient 24 (2): 61–70. ———. 2003. Discerning Palates of the Past: An Ethnoarchaeological Study of Crop Cultivation and Plant Usage in India. Ethnoarchaeological Series, 5. Ann Arbor: International Monographs in Prehistory. Robinson, M. A. 2000. “Further Considerations of Neolithic Charred Cereals, Fruits and Nuts.” In Plants in Neolithic Britain and Beyond, edited by A. S. Fairbairn, 85–90. Neolithic Studies Group Seminar Papers, vol. 5. Oxford: Oxbow Books. Samuel, D. 2001. “Archaeobotanical Evidence and Analysis.” In Peuplement rural et aménagements hydroagricoles dans la Moyenne Vallée de l’Euphrate fin VII–XIX siècle, edited by S. Berthier, 343–481. Damas: Institut Français de Damas. Saraswat, K. S. 2004. “Plant Economy of Early Farming Communities.” In Early Farming Communities of the Kaimur, edited by B. P. Singh, 416–535. Jaipur: Publication Scheme. Schiffer, M. B. 1972. “Archaeological Context and Systematic Context.” American Antiquity 37:156–65. ———. 1976. Behavioural Archaeology. New York: Academic Press.

216

social archaeobotany

———. 1987. Formation Processes of the Archaeological Record. Albuquerque: University of New Mexico Press. Schiffer, M. B., and J. M. Skibo. 1987. “Theory and Experiment in the Study of Technological Change.” Current Anthropology 28 (5): 595–622. Sharma, G. R., V. D. Misra, D. Mandal, B. B. Misra, and J. N. Pal. 1980. Beginnings of Agriculture (Epi-Palaeolithic to Neolithic): Excavations at ChopaniMando, Mahadaha, and Mahagara. Allahabad: Abinash Prakashan. Sheridan, A. 1993. “The Manufacture, Production and Use of Irish Bowls and Vases.” In The Funerary Bowls and Vases of the Irish Bronze Age, edited by B. Ó. Ríordáin and J. Waddell, 45–75. Galway: Galway University Press. Stemler, A. B. 1990. “A Scanning Electron Microscopic Analysis of Plant Impressions in Pottery from Sites of Kadero, El Zakiab, Um Direiwa and El Kadada.” Archéologie du Nil Moyen 4:87–106. Stevens, C. J. 2003a. “The Arable Economy.” Transactions of the Bristol and Gloucestershire Archaeological Society 121:76–81. ———. 2003b. “An Investigation of Agricultural Consumption and Production Models for Prehistoric and Roman Britain.” Environmental Archaeology 8:61–76. ———. 2007. “Reconsidering the Evidence: Towards an Understanding of the Social Contexts of Subsistence Production in Neolithic Britain.” In Archaeobotanical Perspectives on the Origin and Spread of Agriculture in Southwest Asia and Europe, edited by S. Colledge and J. Conolly, 375–90. London: UCL Press. Stevens, C. J., and K. N. Wilkinson. 2001. “Economy and Environment.” In An Iron Age Site at Groundwell West, Blunsdon St. Andrew, Wiltshire: Excavations in 1996, edited by G. Walker, B. Langton, and N. Oakey. Cirencester: Cotswold Archaeological Trust Ltd. Stone, G. D., R. Mcc. Netting, and M. P. Stone. 1990. “Seasonality, Labor Scheduling, and Agricultural Intensification in the Nigerian Savanna.” American Anthropologist 92:7–23. Tierney, J., and M. Hannon. 2003. “Appendix II: Charred Plant Remains.” Proceedings of the Royal Irish Academy, Section C, 103:152–57. Van der Veen, M. 1984. “Sampling for Seeds.” In Plants and Ancient Man: Studies in Palaeoethnobotany, Proceedings of the Sixth Symposium of the International Work Group for Paleoethnobotany, edited by W. van Zeist and W. A. Casparie, 193–99. Rotterdam: A. A. Balkema. ———. 1992. Crop Husbandry Regimes: An Archaeobotanical Study of Farming in Northern England 1000 BC–AD 500. Sheffield Archaeological Monographs, vol. 3. Sheffield: J. R. Collis. Van Zeist, W., and J. A. Bakker-Heeres. 1984. “Archaeobotanical Studies in the Levant 3. Late Palaeolithic Mureybet.” Palaeohistoria 26:171–99. Van Zeist, W., and H. Buitenhuis. 1983. “A Palaeobotanical Study of Neolithic Erbaba, Turkey.” Anatolica 10:47–89.

Routine Activities, Tertiary Refuse, Labor Organization 217 Van Zeist, W., and W. A. Casparie, eds. 1984. Plants and Ancient Man: Studies in Palaeoethnobotany, Proceedings of the Sixth Symposium of the International Work Group for Paleoethnobotany. Rotterdam: A. A. Balkema. Van Zeist, W., and G. J. de Roller. 1995. “Plant Remains from Asiliki Höyük, a Pre-Pottery Neolithic Site in Central Anatolia.” Vegetation History and Archaeobotany 4:179–85. Van Zeist, W., and W. Waterbolk-Van Rooijem. 1985. “The Palaeobotany of Tell Bouqras, Eastern Syria.” Paléorient 11 (2): 131–47. Viklund, K. 1998. Cereals, Weeds and Crop Processing in Iron Age Sweden. Archaeology and Environment, 14. Umeaa: Department of Archaeology, University of Umeaa. Vishnu-Mittre. 1969. “Remains of Rice and Millet.” In Excavations at Ahar (Tambavati), edited by H. D. Sankalia, S. B. Deo, and Z. D. Ansari, 229–36. Pune: Deccan College Postgraduate and Research Institute. Wainwright, G. J., and S. M. Davies. 1995. Balksbury Camp, Hampshire, Excavations 1973 and 1981. English Heritage Archaeological Report, vol. 4. London: English Heritage. Weber, S. A. 2003. “Archaeobotany at Harappa: Indications for Change.” In Indus Ethnobiology: New Perspectives from the Field, edited by S. A. Weber and W. R. Belcher, 75–198. Lanham, MD: Lexington Books. Weir, D. 1996. “Charred Seed Remains from Mannin Bay 2 (MN2).” Journal of Irish Archaeology 7:79–80. Wilkinson, K., and C. J. Stevens 2003. Environmental Archaeology: Approaches, Techniques and Applications. Stroud: Tempus Publishing. Willcox, G., and S. Fornite. 1999. “Impressions of Wild Cereal Chaff in Pisé from the 10th Millennium uncal BP at Jerf el Ahmar and Mureybet: Northern Syria.” Vegetation History and Archaeobotany 8:21–24. Wilson, D. G. 1984. “The Carbonisation of Weed Seeds and Their Representation in Macrofossil Assemblages.” In Plants and Ancient Man: Studies in Palaeoethnobotany, Proceedings of the Sixth Symposium of the International Work Group for Paleoethnobotany, edited by W. van Zeist and W. A. Casparie, 201–6. Rotterdam: A. A. Balkema. Young, R., and G. Thompson 1999. “Missing Plant Foods? Where Is the Archaeobotanical Evidence for Sorghum and Finger Millet in East Africa?” In The Exploitation of Plant Resources in Ancient Africa, edited by M. van der Veen, 63–72. New York: Kluwer Academic/Plenum Publishers.

C h ap t e r 1 0

Of Crops and Food A Social Perspective on Rice in the Indus Civilization Marco Madella

Introduction Both the cultural ecology and agriculture of the Indus Civilization have attracted significant attention (for an overview, see Fuller and Madella 2002; Madella and Fuller 2006). However, there has not been much effort to explore the role of plants and animals as food in Indus society (see, e.g., Fuller 2005). Food choice implies the selection of “ingredients” and their consumption, and encompasses what is eaten, why, where, and how. Food choice has an important role in the social, economic, and symbolic aspects of life because it conveys information on preferences, identities, and culture. Food permeates the life of all the people in a society, with both a public and a private role. The range of factors that can intervene in food choice is remarkably varied. The three most important are the life course of an individual, the influences to which an individual is exposed, and the personal choices an individual makes (Sobal et al. 2006). A food choice process model is illustrated in figure 10.1. Among the influences that can control food choice, ideals are the norms people have acquired through acculturation and socialization, while social factors are the relationships to which people are exposed. These factors affect the cultural judgments on which

218

Of Crops and Food

219

Figure 10.1. The food-choice process model suggested for the Indus Civilization (based on Sobal et al. 2006).

foods are acceptable and preferable, as well as how and when they are eaten (Sobal et al. 2006).

The Indus Civilization The Indus Civilization of the greater Indus Valley (figure 10.2), in what are today Pakistan and India, was the apex of a long cultural trajectory that started in the western highlands of the valley and then shifted onto the Indus plains. In this respect the Indus Civilization has similarities with two other ancient civilizations, the Egyptian and the Mesopotamian, where major rivers flow through desert or semidesert environments and the agriculture is very much linked to the floods of the river. However, a peculiarity of the Indus River is its very variable course due to the continuous aggrading of the flood plain that leaves the river “suspended” in respect to the surrounding land, making course changes during the flood very common (Lambrick 1967, 1986). Several chronologies have been proposed for the Indus Civilization, but those of Kenoyer (1998) and Possehl (2002) are probably the most thorough and will be followed in the present chapter (see table 10.1 for

220

social archaeobotany

Figure 10.2.  Map of the greater Indus valley with the major sites of the Mature Harappan stage (Harappan phase). The northern part of the GhaggarHakra area is shaded in gray.

a comparison). The early food-producing communities (7000–4300 cal BC) are to be found mainly on the hills and piedmonts of Baluchistan, North-Western Frontier Province, and Sindh. Neolithic sites may also be present in the plains, but they are deeply buried beneath alluvial deposits or later sites. These villages represent both early cultivators (e.g., Mehrgarh in Baluchistan) (Costantini 1984) and pastoralists. With the consolidation of these early communities (4300–3200 cal BC), there is a continuous growth in settled life and a possible geographical expansion. More sites are found in the plains, for example, in the area of Fort Derawar in Cholistan along the ancient inland delta of the now-dried Ghaggar-Hakra River (see figure 10.2; Mughal 1982, 1997). There is a clear cultural continuity with the previous period but also technological developments with, for instance, the introduction of the potter’s wheel (Possehl 2002). The Early Harappan (3200–2600 cal BC, or in Kenoyer’s scheme, 3300–2800 cal BC [Ravi phase] and 2800–2600 cal BC

c. 2600–2450

c. 2200–1900

1900?–1700

1700?–1300

Harappan Phase

Harappan Phase

Harappan Phase

Harappan / Late Harappan Transition

Late Harappan (Cemetery H)

3A

3B

3C

4

5

c. 2450–2200

c. 2800–2600

Early Harappan/Kot Diji

2

c. 3300–2800

Years B.C.

Early Harappan/Ravi

Phase

1 A/B

Period

Chronology of the Indus Age according to Kenoyer (1998) and based on the stratigraphy at Harappa

Posturban Harappan (several non-contemporaneous phases) Early Iron Age of N. India and Pakistan

1000–500

Mature Harappan (five contemporaneous phases)

7

6

5

4

Early Harappan / Mature Harappan Transition

1900–1000

2500–1900

2600–2500

3

Early Harappan (four contemporaneous phases)

2

Developed village farming communities and pastoral societies (two non-contemporaneous phases)

4300–3800 and 3800–3200 3200–2600

1

Beginning of village farming communities and pastoral societies (two noncontemporaneous phases)

Stage

7000–5000 and 5000–4300

Years B.C.

Chronology of the Indus Age according to Possehl (2002) and based on the Greater Indus Valley area

Table 10.1. Chronology of the Indus Civilization according to Kenoyer (1998) and Possehl (2002).

222

social archaeobotany

[Kot Diji phase]) is still characterized by cultural continuity with the previous period associated with growth and expansion of the farming communities into new territories. Notwithstanding the evidence for incipient urbanism, during this stage there seems to be little indication of social segmentation. The transition between the Early and Mature Harappan (2600–2500 cal BC, or period 3A in Kenoyer’s chronology) is not well understood, and more needs to be done to shed light on a pivotal moment of the Indus Civilization. Undoubtedly, however, it is during this transitional period that the cultural innovations and the unifying attributes that will characterize the Mature Harappan start to appear (Allchin and Allchin 1997; Mughal 1991, 1994). The Mature Harappan phase (2500–1900 cal BC and Kenoyer’s period 3B and 3C) sees the establishment of the Indus Civilization over an enormous territory, spanning north-south from the Gulf of Mumbai (Bombay) to the Himalaya and east-west from Rajasthan, Haryana, and Gujarat to Baluchistan and Afghanistan. This stage was characterized by an elaborated and clearly stratified society, the presence of complex architecture, a sophisticated material culture, and an overarching ideology (Kenoyer 1991, 1998; Possehl 2002). It is during the Mature Harappan that we see the setting of urbanization with the rapid growth of cities like Harappa, Mohenjo-Daro, Ganweriwala, and Dholavira (see figure 10.2).

Indus Civilization Agriculture and Evidence for Rice Wheat (Triticum sp.) and barley (Hordeum vulgare L. sensu lato) were the basis of the Indus agriculture, at least in the core area of the greater Indus valley (Costantini and Biasini 1985; Fairservis 1971; Franke-Vogt 1995; Fuller and Madella 2002; Kajale 1991; Madella 1995, 1997; Meadow 1989, 1996, 1998; Saraswat 1992; Vishnu-Mittre and Savithri 1982; Willcox 1992). Wheat and barley were part of a winter (rabi) crop package that also included legumes (pulses): peas (Pisum sativum L., incl. P. arvense L.), grasspea (Lathyrus sativus L.), lentils (Lens culinaris Med.), and chickpea (Cicer arietinum L.) (Fuller and Madella 2002). Wild rice (Oryza rufipogon Griff. and O. nivara Sharma et Shastry) does not occur in the Indus valley, and the cultivation and importance of rice in the Indus society have been the subject of intense debate (Weber 1999; for an overview, see Fuller 2001b; Fuller and Madella 2002; Madella 2003). For a list of sites with evidence of cultivated rice in northern South Asia, see table 10.2 and figure 10.3. The earliest documented evidence for the use of rice in the Indian subcontinent is from Lahuradewa (District Sant Kabirnar, Uttar Pradesh) (Tewari et al. 2003; Tewari et al. 2006; Tewari et al. 2008). Rice is found

Of Crops and Food

223

Figure 10.3. Distribution and ages of the archaeological sites with rice remains in the northern part of South Asia.

from period IA with calibrated dates from wood charcoal of 8000– 9000 BC, while an AMS on rice grain gave a date of 6400 BC (Tewari et al. 2008). A contentious issue is whether the Lahuradewa rice is effectively domesticated or whether it represents the harvesting of wild rice from the nearby lake (see Fuller, Stevens, and McClatchie, this volume). It is possible that both forms are present in the assemblage. Rice has been reported from several other sites on the subcontinent; however, most of these findings have chronological problems and might not be as early as previously thought. The Vindhyan Neolithic site of Chopani Mando (Allahabad, Uttar Pradesh) (Chakrabarti 1988) was considered to be of the seventh millennium BC and with evidence for rice. A single radiocarbon age from the level with pottery is around 3300 cal BC. Renewed research on the site shows that the charred plant remains are very scarce, with no crops and few charcoals, and the site might have been occupied only seasonally (Fuller, pers. comm.). The handmade pottery Vindhyan Neolithic of Koldihwa (Allahabad, Uttar Pradesh) (Chakrabarti 1988) has three radiocarbon ages: 7505–7033 cal BC, 6190–5764 cal BC, and 5432–5051 cal BC. However, these dates appear to be quite early. The ceramic evidence, based on comparison to other settlements, could date the site to 2500 cal BC. Recent radiocarbon ages on grains (including rice) seem to frame this site to the second and first millennium BC (Fuller, pers. comm.). Other Neolithic sites with evidence of rice are Barahuna and Manigara (Eastern Neolithic) and Mahagara (Vindhyan Neolithic) (Kharakwal

224

social archaeobotany

Table 10.2. Harappan and adjacent archaeological sites with rice remains (data from Fuller and Madella 2002). Site

Geographical area

Period BC

Type of remain

HARAPPAN CIVILIZATION Pirak I

Baluchistan

1950–1550

Impressions (dubious)

Harappa

Northern and Eastern Provinces

2250–2000

Seeds and phytoliths

Daulatpur

Northern and Eastern Provinces

2000–1700

Seeds

Mithatal

Northern and Eastern Provinces

2000–1400

Seeds

Hulas

Upper Ganga

1800–1200

Seeds

Lal Quila

Upper Ganga

1800–1200

Seeds

Atranjikhera

Upper Ganga

1800–1200

Seeds

Lothal

Gujarat

2000–1700

Impressions

Rangpur IIA

Gujarat

2600–2200

Impressions

Rangpur III

Gujarat

2000–1700

Impressions

ADJACENT OR IN CONTACT WITH HARAPPAN CIVILIZATION Semthan I

Kashmir (Neolithic)

1500–600

Seeds

Ghaleghay I

Swat Valley (Chalcolithic)

3000–2500

Impressions

Ghaleghay II

Swat Valley (Chalcolithic)

2500–2000

Impressions

Ghaleghay III

Swat Valley (Chalcolithic)

1900–1700

Impressions

Bir-KotGhwandai

Swat Valley (Chalcolithic)

1700–1400

Seeds

Loebanhr 3

Swat Valley (Chalcolithic)

1700–1400

Seeds

Ahar

Maharashtra (Banas)

2600–1500

Seeds

Dangwada

Maharashtra (Banas)

2000–1500

Seeds

Inamgaon

Maharashtra (Malwa)

1700–1500

Seeds

Daimabad IV

Maharashtra (Jorwe)

1500–1100

Seeds

Navdatoli

Maharashtra (Jorwe)

1500–1100

Seeds

Of Crops and Food

225

et al. 2004). Mahagara is dated at ca. 1700 cal BC (Kumar 2001), with rice in the earliest levels accompanied by millets and pulses but not barley and wheat (Fuller, pers. comm.). Further findings of rice are from the sites of Damdana (ca. 2500 cal BC) (Fujiwara et al. 1992; Fuller 2001a), Kunal (Haryana, 2850–2600 cal BC) (Saraswat and Srivastava 1996), and from the pre-Harappan (ca. 2300–2000 cal. BC) levels at Balu (Haryana) (Saraswat and Srivastava 1996). Macrobotanical remains and impressions are reported from Harappan sites in Gujarat (Ghosh and Lal 1963; Weber 1991), Rajasthan (Kajale 1996; Vishnu-Mittre 1969), Uttar Pradesh (Saraswat 1992, 1993), Punjab (Weber 1997), Haryana (Willcox 1992), and Baluchistan (Costantini 1979; Costantini and Biasini 1985). Other sites from the Northern Neolithic, which overlaps with the Indus chronology, are Gufkral (2800–1500 cal BC) and Semthan (Kharakwal et al. 2004). Phytoliths from glumes (double-peaked cells and serrated elements) of Oryza sativa L. have been recovered from deposits in Harappa dating after ca. 2200 cal BC (Madella 2003). This confirmed the finding of Fujiwara et al. (1992) of leaf bulliform cells from the Mature Harappan levels of the AB mound (“the citadel”) as well as the finding of charred seeds by Weber (2003). Possible rice phytoliths were also identified in some of the pottery sherds from the platforms, the granary area, and mound F, but, owing to the small sample size, identification remains unsafe (Fujiwara et al. 1992). As most of the Indus area of cultural influence does not fall in the zone of distribution of wild rice, the presence of Oryza sativa L. phytoliths from all parts of the plant (leaves and chaff) should be interpreted as evidence of local cultivation (see Harvey and Fuller 2005; Madella 2003). On the other hand, it is intriguing that the phytolith evidence is not abundant, something that we should have expected if the final rice processing was routinely (daily) carried out in the household (Harvey and Fuller 2005; Fuller, Stevens, and McClatchie, this volume). On the basis of the above discussion, it is apparent that rice use is established by ca. 6400 BC in the middle Ganges valley, and later evidence makes it clear that rice was a crop in South Asia dating from ca. 2500 BC. Genetic studies suggest that indica and japonica populations diverged 200,000 to 400,000 years ago (Sweeney and McCouch 2007 and references; for an in-depth discussion, see also Sato, this volume), and that the establishment of current indica varieties required hybridization of native wild rices from South Asia with rice strains from East Asia (japonica). As proposed by Fuller and Weisskopf (2009–11), a possible explanation for the genetic evidence is that rice farmers in South Asia received, through trade from the east (e.g., China), rice that already had domestication genes and, probably appreciating the characteristics

226

social archaeobotany

of the eastern rice, decided to hybridize this rice with their local (possibly inferior) varieties. Fuller and Weisskopf (2009–11) suggest that this process took place in Pakistan and northwest India at around 1900 BC, because at this time several crops of Chinese origin (common millet, hemp, peaches, apricot) and harvesting tools similar to those seen earlier in China started to appear. The evidence from the weeds recovered from South Asian sites shows that dry rice was the early rice of the Ganges floodplains (Fuller and Qin 2009), cultivated by taking advantage of the monsoon and floods.

Indus Society, Crops, and Food An important aspect of complex societies is the control and redistribution of goods, of which food is an important part. Indeed, the archaeological study of social complexity has been in general based on the assumption that the existence of stratified societies and the presence of an elite is indicated by the occurrence of luxury or prestige goods, symbolizing wealth and power, along with sophisticated and complex architecture and burials (e.g., Turkon 2004). Specific material goods can also play a role in shaping social relations (Hodder 1982, 1986), and this role can be controlled by the elite to establish and preserve their role and status (e.g., DeMarrais, Castillo, and Earle 1996; Douglas and Isherwood 1996, 90; Helms 1979; McGuire 1992). During the Mature Harappan, new products that were the result of technically complex expertise appear in the Indus cultural area. The range of products is quite extensive, from metallurgy to bead-making or faience crafts, and some of these new goods are unequivocally urban (Allchin and Allchin 1997; Bhan, Vidale, and Kenoyer 2002; Kenoyer 1998; Vidale and Miller 2000). These products were traded within the Indus sphere, and trade suggests that, for certain goods, political influence must have been exercised through segregation and control of production and distribution. However, it is essential to underline that even if there is plenty of evidence for a segmented society, with certain segments having preferential access to goods and value items, the overall structure of the Indus Civilization appears to be truncated. Indeed, there is no clear sign of a ruling class, and, for instance, burials do not present any status differentiation (neither in style nor in grave goods). This could signify little vertical social differentiation (heterarchy; see, e.g., Crumley 1995; Potter and King 1995) as well as the use of different sets of values to express status, and it is in this direction that I would like to direct my inquiry. If the Indus society constructed bonds and relationships through alternative and less obvious ways, we need to start opening out our investigation to other avenues that might help in better understanding the

Of Crops and Food

227

peculiarities of these people. I think that food is one such avenue that has not been properly explored. Food can be seen as one of the many “things” that are made by people, and it carries both functional and social messages that cannot be separated. In the area of production and control, food and its ingredients play a fundamental role in substantiating social status and identity through a connection between what is eaten and why it is eaten. To convert products into food is, indeed, to create an intense semiotic device (Appadurai 1981, 1986). Undeniably, food has a social and a political role and plays a focal function in activities related to identity and power (Dietler 1996, 2001; Gumerman 1997). Food can be seen as a medium of contact between human beings, and “many important ideas concerning sharing, redistribution and power are expressed in the idiom of food” (Appadurai 1981, 495). The production and exchange of the ingredients that will make up the food eventually link the political economy with everyday domestic life (Dietler 1996, 2001; Dietler and Hayden 2001). In fact, it has been suggested by Sherrat (1999) that a characteristic of cultural development is a growing complexity in a society’s diet and cuisine, in parallel with the development of more complex social and technological structures. Indeed, preferential access to food or control of the raw ingredients can bestow social status in the same manner as access to and control of other valued goods (Cobb 1996; Fox 1996) and raw materials. Certain types of food can undeniably acquire special significance through cultural constructions maneuvered by a part of the society. This process would establish or reinforce, in the social perspective, the status (high/ low) symbolism of that particular food as well as the status of the people eating it (Dietler 1996, 98–99; 2001; Hayden 2001). Other forms of food control can also be implemented to establish, maintain, and reinforce social positions. In certain cases, it is not the control of rare or complex foods but the control of production, accumulation, and redistribution of staple foods that can endorse the elite. Production control can be exercised as a control on the land and the means of cultivation (e.g., water availability or labor organization) as well as through forms of taxation (Cobb 1996; Dietler 1996; Fox 1996). Accumulation, as a display of wealth, will reinforce status, and redistribution—in periods of need or on a regular basis—would act to strengthen the bonds between elite and lower classes and to increase the sense of obligation toward the elite (Sahlins 1972). It is possible to hypothesize that a Mature Harappan example of the process of production-with-redistribution of staple food would have been the Ghaggar-Hakra area (Kenoyer 1998; Fuller and Madella 2002). This area could have acted as a breadbasket for the production of staple grains (wheat and barley). The cereal surplus was probably then channeled toward the major centers and subsequently further

228

social archaeobotany

redistributed in the settlements of Sindh and Punjab. This could have happened through an exchange of goods or by straightforward acquisition. In fact, when the area of the Ghaggar-Hakra became unproductive owing to a rearrangement of the local hydrology, the consequences were felt on a much bigger scale, one of the plausible explanations for the de­ urbanization that took place during the Late Harappan (2000–1700 BC) (for an extensive discussion of this theme, see Fuller and Madella 2002; Madella and Fuller 2006).

The Role of Rice in the Indus Society Models of crop adoption basically fall into two groups: food stress and food choice (see, e.g., Fuller 2003, 374–76). Food stress or push models speculate that new crops can be adopted when changes in food supply occur (Minnis 1985). These changes can be related to population pressure, shortages in staple food, and environmental changes. Food choice or pull models see the adoption of new crops for social reasons, to have access to higher surpluses, or for producing special foods that have a function in social display and the legitimization of roles (e.g., Fuller 2001a; Hayden 1996; Sherratt 1999). The specific social role of crops or food is a subject that has not been much discussed in Indus archaeology. Usually, changes in the cultivated plants and acquisition of new crops were discussed only on the basis of food stress theories (e.g., Possehl 1997). A departure from this approach can be seen during the last decade (see, e.g., Fuller 2001a; Fuller and Madella 2002; Weber 1999). Evidence from the greater Indus valley shows that the addition of summer crops happened at many sites during and after the Mature Harappan (2500–1900 cal BC). The reasons for such a diversification in the crop package might be related to buffering an overreliance on winter staple cereals or to a more rational use of the labor force, avoiding bottlenecks when a large proportion of the labor force needed to be mobilized for the harvest (Fuller 2001a; Fuller and Madella 2002). A further explanation for the acquisition of some of these new crops can be found in the use of food elements as social symbols (e.g., ideals, social values, context), importantly, in the negotiation of status relationships (Wiessner 1996). Why is rice seldom found in the Indus sites of the core area? And why does rice appear only from the late Mature and Late Harappan periods? Rice is not a crop indigenous to the core Indus area. In the last decades several sites encompassing the Mature Harappan have been excavated, and macrobotanical remains were part of the recovered archaeological

Of Crops and Food

229

record (Fuller and Madella 2002, and references; Kharakwal et al. 2004). Many sites have now also been investigated for phytoliths (Eksambekar and Kajale 2007; Kajale and Eksambekar 2007; Madella 2003, and references). Rice, when present, has consistently been recovered in small quantities, in the form of charred material, imprints, or phytoliths. This evidence reinforces the view of rice as a rare crop and not one providing an alternative staple or buffer crop. Rice, for instance, never becomes abundant during the 400 years of the Mature Harappan period (2600/2500–1900 cal BC) at the important site of Harappa, from which dozens of samples have been analyzed for phytoliths and charred remains (Madella 2003, and unpublished data; Weber 2003). I suggest that there are two possible scenarios for the “secondary” role of rice in the Indus society. As discussed above, by the Mature Harappan (2600–1900 cal BC), the Indus Civilization shows a certain control related to goods production and distribution, and accessibility to the different levels of the society. This structure of control might have acted in preventing the entrance of a different type of food. The social reorganization as well as the possible partial breakdown of control occurring during the Late Harappan (1900–1300/1000 cal BC) rearranged the power balance, opening the possibility for peasants to acquire crops from other cultural spheres. Rice, being a summer crop (kharif), could have come along with the acquisition of millets (Fuller and Madella 2002; Ma­della and Fuller 2006), but somehow at a much slower pace. A second hypothesis is that during the last part of the Mature Harappan rice was acquired as a further symbol of class identity. Indeed, I doubt that rice cultivation was the result of a trend toward diversification or labor rearrangement. Other­wise, we would expect rice to become more abundant over time in the Indus agricultural package as well as as a cuisine ingredient. Also, if rice had become a staple food, its processing would be more and more evident at the household level (e.g., more chaff phytoliths), but this never happens. Rice remains a “rare” ingredient and, according to the evidence we have so far, specific to major settlements. I would argue, therefore, that in the case of this second hypothesis, the importance of rice rested exactly in its constrained availability and its associated social significance, possibly specifically related to the creation or reinforcement of class identity. In that it was an exceptional ingredient, only a certain part of the social group was able to acquire it. A specific social group might also have had a role in controlling production and distribution of the crop, as was the case for other valued commodities. This control need not have been expressed as enforcement or coercion, however, and it could have taken the form of belief, for instance, through a system of purity/impurity in which ideology would

230

social archaeobotany

prevent a segment of the society from accessing rice. Also, there is no reason to see this class identity in terms of the presence of “rulers.” Indeed, status differentiation could have been a vertical differentiation within the same kinship (leaders of kin groups) or a horizontal differentiation between kinships (religious leaders). The suggestion that earlier indica varieties might have been of inferior quality in respect to the japonica strains, and that hybridization around 1900 BC improved them, seems to reinforce the status of rice as a secondary (possibly because less productive) but sought-after product of the Harappan package. Indeed, after hybridization had brought out better varieties, and with the breakdown of the Harappan social structure after 1900 BC, rice could have become assimilated into a diversifying agriculture, possibly also losing its special status.

Conclusions Since the beginning of South Asian archaeobotanical studies, there has been a rather generalized attitude of discussing crop acquisition and agricultural practices as the result of climate influence. Past societies have often been seen as passive subjects of climatic vagaries. In certain cases it is true that climate or other nonsocietal factors might have been the mechanism that shaped change. However, social factors related to crop (and therefore food) choice have on the whole attracted little interest. Rice, because of its wild species distribution, requirements for cultivation, and specific postharvest processing, is the ideal crop to explore social dynamics intervening in crop acquisition and food consumption. The present discussion has tried to address rice acquisition and consumption through a social perspective. It has by no means the pretense of clarifying or solving all the aspects of rice acquirement, cultivation, and consumption, and more work is indeed needed, especially in the field, to expand the present database on this crop. However, it is hoped that this foray will at least open the arena for a wider discussion on the agricultural strategies and food choices of the Indus Civilization and that it will foster debate on the social role of plants in South Asian prehistory.

Acknowledgments I would like to express my thanks to Heather M. L. Miller, Dorian Q Fuller, and Carla Lancelotti for the useful comments on a preliminary draft of this chapter.

Of Crops and Food

231

References Allchin, B., and F. R. Allchin. 1997. Origins of a Civilization: The Prehistory and Early Archaeology of South Asia. New Delhi: Penguin Books India. Appadurai, A. 1981. “Gastro-Politics in Hindu South Asia.” American Ethnologist 8 (3): 494–511. ———. 1986. “Introduction: Commodities and the Politics of Value.” In The Social Life of Things: Commodities in Cultural Perspective, edited by A. Appadurai, 3–63. Cambridge: Cambridge University Press. Bhan, K. K., M. Vidale, and J. M. Kenoyer. 2002. “Some Important Aspects of the Harappan Technological Tradition.” In Indian Archaeology in Retrospect, vol. 2, Protohistory, edited by S. Settar and R. Korisettar, 223–71. New Delhi: Manohar. Chakrabarti, D. K. 1988. Theoretical Issues in Indian Archaeology. New Delhi: Munshiram. Cobb, C. R. 1996. “Specialization, Exchange, and Power in Small-Scale Societies and Chiefdoms.” Research in Economic Anthropology 17:251–94. Costantini, L. 1979. “Plant Remains at Pirak.” In Fouilles de Pirak, edited by J.-F. Jarrige and M. Santoni, 1:326–33. Paris: Diffusion de Boccard. ———. 1984. “The Beginnings of Agriculture in the Kachi Plain: The Evidence from Mehrgarh.” In South Asian Archaeology 1981, edited by B. Allchin, 29– 33. Cambridge: Cambridge University Press. Costantini, L., and L. Costantini Biasini. 1985. “Agriculture in Baluchistan between the 7th and 3rd Millennium B.C.” Newsletter of Baluchistan Studies 2:16–37. Crumley, C. L. 1995. “Heterarchy and the Analysis of Complex Societies.” In Heterarchy and the Analysis of Complex Societies, edited by R. M. Ehrenreich, C. L. Crumley, and J. E. Levy, 1–5. Archaeological Papers of the American Anthropological Association, no. 6. Arlington, VA: American Anthropological Association. DeMarrais, E., L. J. Castillo, and T. Earle. 1996. “Ideology, Materialization, and Power Strategies.” Current Anthropology 37 (1): 15–31. Dietler, M. 1996. “Feasts and Commensal Politics in the Political Economy: Food, Power, and Status in Prehistoric Europe.” In Food and the Status Quest: An Interdisciplinary Perspective, edited by P. Weissner and W. Schiefenhövel, 87–125. Providence: Berghahn Books. ———. 2001. “Theorizing the Feast: Rituals of Consumption, Commensal Politics and Power in African Contexts.” In Feasts: Archaeological and Ethnographic Perspectives on Food, Politics, and Power, edited by M. Dietler and B. Hayden, 65–114. Washington, DC: Smithsonian Institution Press. Dietler, M., and B. Hayden. 2001. “Digesting the Feast: Good to Eat, Good to Drink, Good to Think.” In Feasts: Archaeological and Ethnographic Perspectives

232

social archaeobotany

on Food, Politics, and Power, edited by M. Dietler and B. Hayden, 1–22. Washington, DC: Smithsonian Institution Press. Douglas, M., and B. Isherwood. 1996. The World of Goods. Oxford: Routledge. Eksambekar, S. and M. D. Kajale. 2007. “Microstratigraphy of an Early Historic Refuse Pit: A Phytolithological Approach.” In Plants, People and Places: Recent Studies in Phytolith Analysis, edited by M. Madella and D. Zurro, 110–17. Oxford: Oxbow. Fairservis, W. A., Jr. 1971. The Roots of Ancient India: The Archaeology of Early Indian Civilization. London: George Allen and Unwin. Fox, J. G. 1996. “Playing with Power: Ballcourts and Political Ritual in Southern Mesoamerica.” Current Anthropology 37 (3): 483–509. Franke-Vogt, U. 1995. “Cultural Ecology of the Greater Indus Valley and Beyond.” Archaeological Review (Karachi) 4:13–64. Fujiwara, H., M. R. Mughal, A. Sasaki, and T. Matano. 1992. “Rice and Ragi at Harappa: Preliminary Results by Plant Opal Analysis.” Pakistan Archaeology 27:129–42. Fuller, D. Q. 2001a. “Harappan Seeds and Agriculture: Some Considerations.” Antiquity 75:410–14. ———. 2001b. “Fifty Years of Archaeobotanical Studies in India: Laying a Solid Foundation.” In Indian Archaeology in Retrospect, vol. 3, Archaeology and Interactive Disciplines, edited by S. Settar and R. Korisettar, 247–363. New Delhi: Manohar. ———. 2003. “Indus and Non-Indus Agricultural Traditions: Local Developments and Crop Adoptions in the Indian Peninsula.” In Indus Ethnobiology: New Perspectives from the Field, edited by S. A. Weber and W. R. Belcher, 343–96. Lanham, MD: Lexington Books. ———. 2005. “Ceramics, Seeds and Culinary Change in Prehistoric India.” Antiquity 79 (306): 761–77. Fuller, D. Q, and M. Madella. 2002. “Issues in Harappan Archaeobotany: Retrospect and Prospect.” In Indian Archaeology in Retrospect, vol. 2, Protohistory, edited by S. Settar and R. Korisettar, 317–90. New Delhi: Manohar. Fuller, D. Q, and L. Qin. 2009. “Water Management and Labour in the Origins and Dispersal of Asian Rice.” World Archaeology 41 (1): 88–111. Fuller, D. Q, and A. Weisskopf. 2009–11. “The Early Rice Project: From Domestication to Global Warming.” Archaeology International 13/14. doi:http:// dx.doi.org/10.5334/ai.1314. Ghosh, S. S., and K. Lal. 1963. “Plant Remains from Rangpur and Other Explorations in Gujarat.” Ancient India 18–19:161–75. Gumerman, G. 1997. “Food and Complex Societies.” Journal of Archaeological Method and Theory 42:105–39. Harvey, H. L., and D. Q Fuller. 2005. “Investigating Crop Processing Using Phytolith Analysis: The Example of Rice and Millets.” Journal of Archaeological Science 32 (5): 739–52.

Of Crops and Food

233

Hayden, B. 1996. “Feasting in Prehistoric and Traditional Societies.” In Food and the Status Quest: An Interdisciplinary Perspective, edited by P. Wiessner and W. Schiefenhövel, 127–48. Providence: Berghahn Books. ———. 2001. “Fabulous Feasts: A Prolegomenon to the Importance of Feasting.” In Feasts: Archaeological and Ethnographic Perspectives on Food, Politics, and Power, edited by M. Dietler and B. Hayden, 23–64. Washington, DC: Smithsonian Institution Press. Helms, M. W. 1979. Ancient Panama: Chiefs in Search of Power. Austin: University of Texas Press. Hodder, I. 1982. Symbols in Action: Ethnoarchaeological Studies of Material Culture. Cambridge: Cambridge University Press. ———. 1986. Reading the Past: Current Approaches to Interpretation in Archaeology. Cambridge: Cambridge University Press. Kajale, M. D. 1991. “Current Status of Indian Palaeoethnobotany: Introduced and Indigenous Food Plants with a Discussion of the Historical and Evolutionary Development of Indian Agriculture and Agricultural Systems in General.” In New Light on Early Farming: Recent Developments in Palaeoethnobotany, edited by J. M. Renfrew, 155–89. Edinburgh: Edinburgh University Press. ———. 1996. “Palaeobotanical Investigations at Balathal: Preliminary Results.” Man and Environment 21 (1): 98–102. Kajale, M. D., and S. Eksambekar. 2007. “Phytolith Analytical Study on a Late Chalcolithic–Early Historical Archaeostratigraphical Sequence from Balathal, South Rajasthan, India.” In Plants, People and Places: Recent Studies in Phytolith Analysis, edited by M. Madella and D. Zurro, 118–33. Oxford: Oxbow. Kenoyer, J. M. 1991. “The Indus Valley Tradition of Pakistan and Western India.” Journal of World Prehistory 5 (4): 331–85. ———. 1998. Ancient Cities of the Indus Valley Civilization. Karachi: Oxford University Press. Kharakwal, J. S., A. Yano, Y. Yasuda, V. S. Shinde, and T. Osada. 2004. “Cord Impressed Ware and Rice Cultivation in South Asia, China and Japan: Possibilities of Inter-Links.” Quaternary International 123–25:105–15. Kumar, A. 2001. “Origin, Development and Growth of the Mahagara Neolithic Culture.” Pragdhara 11:119–24. Lambrick, H. T. 1967. “The Indus Flood-Plain and the ‘Indus’ Civilization.” Geographical Journal 133 (4): 483–95. ———. 1986. Sindh, General Introduction. Vol. 1. Hyderabad: Sindh Publishing House. Madella, M. 1995. “A Preliminary Study of Phytolith Analysis, Agriculture and Use of Plants at Kot Diji (Sindh–Pakistan).” Ancient Sindh 2:93–108. ———. 1997. “Phytolith Analysis from the Indus Valley Site of Kot Diji, Sind, Pakistan.” In Archaeological Sciences 1995, edited by A. Sinclair, E. Slater, and J. Gowlett, 294–302. Oxbow Monograph 64. Oxford: Oxbow.

234

social archaeobotany

———. 2003. “Investigating Agriculture and Environment in South Asia: Present and Future Contributions from Opal Phytoliths.” In Indus Ethnobiology: New Perspectives from the Field, edited by S. A. Weber and W. R. Belcher, 199–249. Lanham, MD: Lexington Books. Madella, M., and D. Q Fuller. 2006. “Palaeoecology and the Harappan Civil­ isation of South Asia: A Reconsideration.” Quaternary Science Review 25: 1283–301. McGuire, R. H. 1992. A Marxist Archaeology. New York: Academic Press. Meadow, R. 1989. “Continuity and Change in the Agriculture of the Greater Indus Valley: The Palaeoethnobotanical and Zooarchaeological Evidence.” In Old Problems and New Perspectives in the Archaeology of South Asia, edited by J. M. Kenoyer, 61–74. Wisconsin Archaeological Reports, vol. 2. Madison: University of Wisconsin. ———. 1996. “The Origins and Spread of Agriculture and Pastoralism in Northwestern South Asia.” In The Origins and Spread of Agriculture and Pastoralism in Eurasia, edited by D. R. Harris, 390–412. London: UCL Press. ———. 1998. “Pre- and Proto-Historic Agricultural and Pastoral Transformations in Northwestern South Asia.” In “The Transition to Agriculture in the Old World,” edited by O. Bar-Yosef, special issue, Review of Archaeology 19 (2): 12–21. Minnis, P. E. 1985. Social Adaptation to Food Stress: A Prehistoric Southwestern Example. Prehistoric Archeology and Ecology Series. Chicago: University of Chicago Press. Mughal, R. 1982. “Recent Archaeological Research in the Cholistan Desert.” In Harappan Civilization: A Contemporary Perspective, edited by G. L. Possehl, 85–96. Warminster: Aris & Phillips. ———. 1991. “The Cultural Patterns of Ancient Pakistan and Neighbouring Regions c. 7000–1500 BC.” Pakistan Archaeology 26:218–37. ———. 1994. “Ancient Cities of the Indus.” Lahore Museum Bulletin 7 (1–2): 53–59. ———. 1997. Ancient Cholistan: Archaeology and Architecture. Lahore: Ferozsons. ———. 1997. “Climate and the Eclipse of the Ancient Cities of the Indus.” In Third Millennium BC Climate Changes and Old World Collapse, edited by H. N. Dalfes, G. Kukla, and H. Weiss, 193–243. Heidelberg: Springer-Verlag. Possehl, G. L. 1997. “The Transformation of the Indus Civilization.” Journal of World Prehistory 11 (4): 425–72. ———. 2002. The Indus Civilization: A Contemporary Perspective. Walnut Creek, CA: Altamira Press. Potter, D. R., and E. M., King. 1995. “A Heterarchical Approach to Lowland Maya Socioeconomics.” In Heterarchy and the Analysis of Complex Societies, edited by R. M. Ehrenreich, C. L. Crumley, and J. E. Levy, 17-32. Archaeological

Of Crops and Food

235

Papers of the American Anthropological Association, no. 6. Arlington, VA: American Anthropological Association. Sahlins, M. 1972. Stone Age Economics. London: Tavistock. Saraswat, K. S. 1992. “Archaeobotanical Remains in Ancient Cultural and SocioEconomical Dynamics of the Indian Subcontinent.” Palaeobotanist 40:514–45. ———. 1993. “Plant Economy of Late Harappan at Hulas.” Puratattva 23:1–12. Saraswat, K. S., and C. Srivastava. 1996. “Palaeobotanical and Pollen Analytical Investigations.” Indian Archaeology 1991–92—A Review: 139–40. Sherratt, A. 1999. “Cash-Crops before Cash: Organic Consumables and Trade.” In The Prehistory of Food: Appetites for Change, edited by C. Gosden and J. Hather, 13–34. London: Routledge. Sobal, J., C. A. Bisogni, C. M. Devine, and M. Jastran. 2006. “A Conceptual Model of the Food Choice Process over the Life Course.” In The Psychology of Food Choice, edited by R. Shepherd and M. Raats, 1–18. Wallingford: CABI. Sweeney, M. T., and S. R. McCouch. 2007. “The Complex History of the Domestication of Rice.” Annals of Botany 100:951–57. Tewari, R., R. K. Srivastava, K. K. Singh, K. S. Saraswat, and I. B. Singh. 2003. “Preliminary Report of the Excavation at Lahuradewa, District Sant Kabirnagar, U.P.: 2001–2002.” Pragdhara 13:37–68. Tewari, R., R. K. Srivastava, K. K. Singh, K. S. Saraswat, I. B. Singh, M. S. Chauhan, A. K. Pokharia, A. Saxena, V. Prasad, and M. Sharma. 2006. “Second Preliminary Report of the Excavations at Lahuradewa District Sant Kabir Naga, U.P.: 2002–2003–2004 & 2005–06.” Pragdhara 16:35–68. Tewari, R., R. K. Srivastava, K. S. Saraswat, I. B. Singh, and K. K. Singh. 2008. “Early Farming at Lahuradewa.” Pragdhara 18:347–73. Turkon, P. 2004. “Food and Status in the Prehispanic Malpaso Valley, Zacatecas, Mexico.” Journal of Anthropological Archaeology 23 (2): 225–51. Vidale, M., and H. M-L. Miller. 2000. “On the Development of Indus Technical Virtuosity and Its Relation to Social Structure.” In South Asian Archaeology 1997, edited by M. Taddei and G. De Marco, 115–32. Rome: Istituto Italiano per l’Africa e l’Oriente (IsIAO); Naples: Istituto Universitario Orientale. Vishnu-Mittre. 1969. “Remains of Rice and Millet.” In Excavations at Ahar (Tambavati), edited by H. D. Sankalia, S. B. Deo, and Z. D. Ansari, 229–35. Pune: Deccan College Postgraduate and Research Institute. Vishnu-Mittre, and R. Savithri. 1982. “Food Economy of the Harappans.” In Harappan Civilization: A Contemporary Perspective, edited by G. L. Possehl, 205–21. New Delhi: Oxford and IBH. Weber, S. A. 1991. Plants and Harappan Subsistence: An Example of Stability and Change from Rojdi. New Delhi: Oxford and IBH. ———. 1997. “Harappa Archaeobotany: A Model for Subsistence.” In South Asian Archaeology 1995, edited by R. Allchin and B. Allchin, 115–17. Oxford and New Delhi: IBH.

236

social archaeobotany

———. 1999. “Seeds of Urbanism: Palaeoethnobotany and the Indus Civilization.” Antiquity 73 (282): 813–26. ———. 2003. “Archaeobotany at Harappa: Indications for Change.” In Indus Ethnobiology: New Perspectives from the Field, edited by S. A. Weber and W. R. Belcher, 175–98. Lanham, MD: Lexington Books. Wiessner, P. 1996. “Introduction: Food, Status, Culture, and Nature.” In Food and the Status Quest: An Interdisciplinary Perspective, edited by P. Weissner and W Schiefenhövel, 1–18. Providence: Berghahn Books. Willcox, G. 1992. “Some Differences between Crops of Near Eastern Origin and Those from the Tropics.” In South Asian Archaeology 1989, edited by C. Jarrige, 291–99. Monographs in World Archaeology, vol. 14. Madison, WI: Prehistory Press.

C h ap t e r 1 1

Anthracological Research on the Brazilian Coast Paleoenvironment and Plant Exploitation of Sambaqui Moundbuilders R i ta S c h e e l -Y b e r t a n d M a r i a D u l c e G a spa r

Archaeobotanical and zooarchaeological studies are still relatively rare in Brazil. They have only recently assumed the role of independent archaeological disciplines, rather than auxiliary techniques for archaeology (in the case of archaeobotany, seldom employed). Archaeologists with a solid formation in biological sciences who are using this knowledge to approach archaeological problems are mainly responsible for this change. On the other hand, the increasing refinement of theory, methods, and interpretation in Brazilian archaeology has also led to a greater demand for archaeobotany. Analyses of plant remains are no longer presented in the form of “laundry lists” attached to archaeological reports. Plant remnants are actively searched for, special techniques are employed, and taxonomic relations have become significant to the extent that they are interpreted simultaneously both according to their landscape and paleoecological significance and as products of the relation of prehistoric groups and their environment.

237

238

social archaeobotany

Until recently, charcoal remains, which are generally abundant in archaeological sediments, were collected exclusively for dating purposes. However, they can provide invaluable information on the paleoenvironment, firewood economy, diet, and so on (Miller, this volume; Scheel-Ybert 2000, 2001a). This chapter presents an example of recent archaeobotanical research in Brazil based on charcoal analysis. In addition to landscape reconstruction, the study tries to understand the possible outcomes of the environmental stability and relative social stability of the coastal moundbuilders society.

Regional Setting Eight sites were studied, all located on the south-southeastern coast of Brazil (figure 11.1). Forte, Salinas Peroano, Meio, and Boca da Barra are situated in Cabo Frio (22°53'S, 42°03'W). The first site is located between the Itajuru Channel and the Atlantic Ocean, on the west coast of this channel, which connects the Araruama Lagoon to the sea; the others are situated on low inland crystalline rocks hills on the east side of the channel. Ponta da Cabeça is situated on the Arraial do Cabo peninsula (22°57'S, 42°14'W), on an igneous rocks hill near the Massambaba beach. Beirada and Pontinha are situated at Saquarema (22°55'S, 42°33'W), at the rear of the Pleistocene beach ridge between the Saquarema Lagoon and the ocean. In most of Rio de Janeiro State, the climate in the coastal zone is tropical wet, hot, and rainy in summer with a mild dry season in winter (Barbiére 1984). In Saquarema, the mean annual temperature is 24– 26ºC and mean precipitation is 1,000 mm/yr (Sá 1992). In Cabo Frio and Arraial do Cabo, the particularly dry climate is a variant of a semiarid hot climate due to the upwelling of cold waters along the coast, which reduces local precipitation. Mean annual temperature is 25ºC and precipitation rarely exceeds 800 mm/yr (Barbiére 1984). Plant associations in the southeast of Rio de Janeiro State, as on a great part of the Brazilian coast, vary according to physiographic conditions and distance from the ocean. The land-sea interface, especially the edges of rivers and lagoons, presents mangrove and saltwater marshes. The restinga ecosystem, a mosaic of vegetation types with diverse physiognomies, occupies the coastal sandy beach ridges. It varies from sparse open plant communities, such as herbaceous and scrub formations (open restinga) to dense evergreen forest (restinga forest), and is characterized by a high percentage of plants from the Myrtaceae family. Rocky outcrops in the Cabo Frio region are occupied by xeromorphic forest, which is a drier variety of the Atlantic Rain Forest. Farther from

Anthracological Research on the Brazilian Coast

239

Figure 11.1. Geographical location of the studied sites: 1–7, southeastern Rio de Janeiro State; 8, southern Santa Catarina State. Stars indicate the locations of sambaquis. (1) Forte sambaqui, (2) Boca da Barra sambaqui, (3) Meio sambaqui, (4) Salinas Peroano sambaqui, (5) Ponta da Cabeça sambaqui, (6) Beirada sambaqui, (7) Pontinha sambaqui, (8) Jabuticabeira II sambaqui. (Adapted from Scheel-Ybert 2001b.)

the ocean, low mountains support forests similar in composition to those of the Atlantic Rain Forest, although trees may be shorter than at higher altitudes. The site of Jabuticabeira II (28°36'S, 48°57'W) is situated in Jaguaruna, Santa Catarina State, on the side of a paleodune, at about 1 km

240

social archaeobotany

from the southwestern margins of the Garopaba do Sul lagoon and about 6 km from the sea (figure 11.1). In this region the climate is temperate sub-hot, with mean winter temperatures over 15ºC and no dry season. Mean annual temperature is 20ºC and mean precipitation is 1,400 mm/yr (Nimer 1989). Although the natural vegetation is almost absent from this region today, the site lies in the phytosociological domain of the restinga ecosystem. Restinga forest probably dominated the paleodunes, and there may have been flooded restinga forest in the plains. The Atlantic Forest is located farther inland, as is the subtropical forest typical of the high plateaus over 800 m in altitude. The Atlantic Forest in Santa Catarina State is in fact the extratropical continuance of the dense ombrophilous forest that extends along the Atlantic Coast between the states of Rio Grande do Norte (5°S) and Rio Grande do Sul (29°S). In the southern region, the mild winter temperatures along the coast and the abundant rains, well distributed throughout the year, ensure the development of an evergreen forest with high biodiversity.

Archaeological Framework The Brazilian coast was colonized by a society of fisher-gatherer moundbuilders during the Holocene, from at least 7100 to 1000 BP (Gaspar 1996). Shellmounds are frequent in various parts of the world, as maritime adaptations became relatively common after the Pleistocene. Each culture presented its own specificities, whether environmental or cultural, and it is consequently difficult to compare the societies of different shellmound constructors. Some sites may be mere shell middens, others are habitation sites containing shellfish and other food remains in association with different vestiges of human activity, and yet others are habitation sites where shells have been used as construction material (Widmer 1989, cited in Claassen 1991). Brazilian shellmounds have long been considered the product of sedentary people, used for such different activities as living, burial, and waste discard (Gaspar 2000). Current interpretations see them as exclusively funerary structures, resulting from repeated ritual events (De Blasis et al. 1998). Around 5000 BP the Brazilian shore was intensely occupied. There is isolated evidence, which needs confirmation, indicating that the process of colonization of the coast started by 9000 BP (De Blasis 2001). Remains of more ancient sites are very difficult to find because they have probably been destroyed by rising sea levels at the beginning of the Holocene. Indeed, the sea level rose along the entire Brazilian coast during

Anthracological Research on the Brazilian Coast

241

the Late Pleistocene and Early Holocene, to reach a maximum of about 4 meters higher than the present zero around 5000 BP (Angulo and Lessa 1997; Martin, Suguio, and Flexor 1979). The mounds, called sambaquis, were intended as landscape markers and in some locations attained monumental size (figure 11.2). Some of the sites reach 30 m in height and more than 400 m in width (Fish et al. 2000; Gaspar 2000), but they were probably much bigger before destruction by modern exploitation. They are usually located on the margins of large water bodies in sectors where the coastline is indented and in the vicinity of a wide range of ecological habitats (sea, lagoons, rivers, restingas, mangroves, forest, etc.). They are considered to have had close relations to the restinga environment, as well as to coastal lagoons (Scheel-Ybert 2000). The archaeological sediments are constituted mainly of mollusk shells and fish bones, frequently alternating with sandy layers rich in organic matter and faunal remains. Composition of the sediments varies from one site to another, but the artifacts show similar characteristics in all the sites (Gaspar 1992). The association of artifacts, burials, and hearths is recurrent. The interpretative scheme previously used assumed that the different stratigraphic layers that make up sambaquis were the remains of successive episodes of occupation by itinerant gatherers, and that mollusks were the subsistence staple (Heredia et al. 1989; Kneip 1980). Only quite recently did researchers begin to acknowledge that fish was the most important dietary intake (De Masi 1999; Figuti 1993; Klökler 2000). Sambaquis archaeology is currently undergoing a major theoretical shift. Social complexity (De Blasis et al. 1998; Lima 1997) and differentiated social status (Gaspar 2000) have entered the discussion arena. A debate has arisen about whether or not zoomorphic sculptures (figure 11.3) indicate the presence of specialist producers (Lima 1997). The grandiosity of the sites is considered to be the result of social labor (Gaspar 2000). It is interesting to note that there is no indication of warfare or other forms

Figure 11.2. Santa Marta sambaqui, Santa Catarina State. (Photo M. D. Gaspar.)

242

social archaeobotany

Figure 11.3. Naturalist zoomorphic sculptures from various sites of Santa Catarina State. (a) Otter-like zoolith, from Jabuticabeira; (b) birdlike zoolith, from Pântano do Sul sambaqui, Ilha de Santa Catarina (green diorite, 13 cm); (c) fish representation, from Laguna de Imaruí (brown diabase, ca. 30 cm); (d) tapir representation, from Jaguaruna; (e) armadillo representation, from the north of Santa Catarina Island (green diorite). (All pieces from the Colégio Catarinense Museum, Florianópolis. Photo: P. Piles.)

of physical conflict in the sambaqui society (Lessa and Medeiros 2001; Neves and Wesolowski 2002). Spatial analysis and studies of several site profiles suggest that this society was sedentary (Gaspar 1992, 1995/96). Sedentariness is also suggested by the long occupation of most sites, without any confirmed period of site abandonment. The sites studied on the southeastern coast, for instance, present occupation periods that vary from about 500 to over 3,000 years, according to conservative estimates (Scheel-Ybert 2000). The Jabuticabeira II site, on the southern coast, was built up over at least 1,000 years (Eastoe et al. 2002; Fish et al. 2000). Although there is little systematic research on demography, there are some indicators of a demographic increase. The spatial organization of sambaquis, frequently grouped, indicates territorial stability (De Blasis et al. 1998). Despite a primary focus on marine resources subsistence, the practice of human management or incipient cultivation (horticulture) of tubers and fruit trees has been suggested for sites in southeast Brazil

Anthracological Research on the Brazilian Coast

243

(Scheel-Ybert 2001a) and for some southern sites (Wesolowski 2000), thereby proposing a more complex subsistence pattern among these populations than was previously envisaged. These activities are the fruit of cultural choices that indicate a deep knowledge of plant life cycles.

Materials and Methods For most of the sites discussed here, archaeological excavations were conducted during the 1980s, when botanical remains were not routinely recovered. For this reason, charcoal fragments were subsequently collected from vertical profiles covering the full height of the sambaqui deposits from the old trenches left by previous archaeological work. Excavation of the sites followed natural layers. However, since archaeological layers are usually thick, internal artificial levels were also defined. Therefore, each sample consists of sediment from a layer of 10 cm in thickness within trenches of 1 m2. The sediment was dry-sieved in the field using a 4 mm mesh. Residual charcoal was later recovered in the laboratory with a flotation device (Ybert, Scheel, and Gaspar 1997). At the Ponta da Cabeça sambaqui, charcoal fragments were collected during the archaeological excavation, and plant remains were sorted by hand. Charcoal samples were examined under a reflected light brightfield/ darkfield microscope. Transversal, tangential longitudinal and tangential radial sections were broken manually. Systematic determination was obtained by comparing the anatomical structures observed with that of extant charred samples and with descriptions and photographs from the literature. Since the wood anatomy of tropical plants is very poorly known, and especially so in the studied region, we have assembled a reference collection of charred woods containing over 2,000 identified samples. Charcoal identification was facilitated by the elaboration of a program for computer-aided identification especially conceived for charcoal analysis, coupled to a data bank of anatomical features from extant and ancient charcoal (Scheel-Ybert, Scheel, and Ybert 2002). Wood anatomy analysis allows identification at the genus level and, sometimes, at the species level. All charcoal fragments over 4 mm were analyzed. In the tropics, smaller fragments are normally impossible to identify because generally they do not present a sufficiently large array of characters. Conventionally, results are presented by ubiquity, by weight, or by count of charcoal fragments. While some authors argue that only charcoal weight is reliable (Carcaillet and Thinon 1996), others assume that counting better expresses the plants’ biomass (Thiébault 1997). Still others maintain that both methods are comparable, even if counting is preferable for practical reasons (Chabal 1997). Under tropical conditions,

244

social archaeobotany

we consider that charcoal pieces must be counted for two main reasons: (1) tropical woods present a very wide range of density values (from 0.1 to 1.3 in dry wood, according to Briane and Doat 1985) that may distort biomass estimations based on weight; and (2) charcoal fragments, especially from shellmounds, are frequently impregnated with carbonates that increase the sample weight in a way that cannot be estimated. Advocates of ubiquity argue that counting or weighing fragments is overly influenced by biases in sampling, preservation, fragmentation, and volume of sampled sediment (just to name a few processes), and therefore it does not accurately reflect past vegetation (Popper 1988; Willcox 1974). Nevertheless, studies based on the analysis of saturation curves and Gini-Lorenz concentration curves1 (Scheel-Ybert 2002), as well as the application of multivariate analysis comparing charcoal records to modern vegetation (Scheel-Ybert 2000), have demonstrated that the relative percentage values of each identified taxon in the assemblage show a representative image of the proportions of these taxa in the surrounding vegetation, provided that a minimum of 100 to 200 charcoal fragments per sample are examined. Multivariate analysis was here applied to both charcoal assemblages and to the phytosociological data of the modern vegetation resulting from specialized literature (Scheel-Ybert 2000). Correspondence analyses (CA) was performed with “Statoscope” program (version 1.6).

Results Complete anthracological diagrams are presented for four sites, which are considered to be representative of each one of the studied localities (figures 11.4–11.6). Complete results have been published previously (Scheel-Ybert 2000). Summary diagrams, where the different taxa have been grouped by plant formation, are presented for all sites (figure 11.7). Transversal sections of some of the identified charcoal fragments are presented in figure 11.8.

Discussion A great floristic diversity characterizes all the charcoal diagrams. Each sample has, in general, more than thirty taxa (figures 11.4–11.6), indicating that domestic firewood was not collected on a selective basis. Frequent traces of decay and insect larvae attack before charring (figure 11.9) are interpreted as signs of deadwood gathering (Scheel-Ybert 2001a). This mode of provision is an additional demonstration of random firewood collection. The random gathering of deadwood, as well as the

Anthracological Research on the Brazilian Coast

245

high floristic diversity, supports the hypothesis that the archaeological charcoal remains are associated with a long temporal activity and represent the vegetation of a relatively large area surrounding the sites. These are essential premises for reliable paleoecological interpretation based on archaeological charcoal (Chabal 1992). The results of this study show that sambaqui moundbuilders occupied essentially the restinga ecosystem, establishing their sites in the vicinity of diverse plant communities, particularly mangroves and coastal forests (the xeromorphic, or dry forest, in the Cabo Frio region, and the Atlantic Rain Forest elsewhere), which were also exploited. All diagrams show strong dominance of Myrtaceae, especially in southeastern sites. Although common in different communities of Brazilian vegetation, high percentages of taxa from this family are typical of the restinga formations, particularly in Rio de Janeiro State (Araujo and Henriques 1984). There is significant contribution of forest taxa in the Salinas Peroano and Boca da Barra sambaquis, both situated in the phytosociological domain of the dry forest, as well as in Jabuticabeira II. Open restinga taxa predominate in the Beirada and Pontinha sambaquis, both situated in the phytosociological domain of the open restinga. A mixed environment is observed in the Forte and Ponta da Cabeça sambaquis, which are more clearly situated on ecotones. Anthracological spectra reflect the local vegetation very well. Indeed, the predominance of one type of vegetation or another depends on the geographical location of each site. Only a few microcharcoal fragments were studied for the Meio sambaqui. They were retrieved at the base of the site during excavation work and are dated at about 5200 BP. The presence of two Myrtaceae species and Condalia sp. suggest that restinga-type vegetation existed near the site at that time (Scheel-Ybert 2000). The charcoal assemblage at each site did not change significantly during the several centuries of occupation. The small oscillations in the relative percentages of the vegetation types in the Beirada sambaqui, and in some levels of Pontinha, Salinas Peroano, and Jabuticabeira II (figure 11.7) are not significant, because of the extremely small number of charcoal pieces in these samples. However, even if coastal vegetation was not affected by either climatic or anthropogenic activities, climatic oscillations did occur. They affected the mangrove vegetation of the southeastern coast, as described below. At Arraial do Cabo (Ponta da Cabeça), mangrove elements increase significantly from about 2100 BP onward (figure 11.7). This may be caused by an increase in mangrove vegetation in the environment, possibly explained by climatic reasons, or may be related to the population increase recorded during this period (Tenório, Barbosa, and Portela 1992),

Figure 11.4. Charcoal percentage diagram of the Forte sambaqui. Total number of analyzed charcoal pieces (Nt) = 8,097; number of taxa (Nsp) = 102. (Adapted from Scheel-Ybert 2000.)

Figure 11.5. Charcoal percentage diagrams of the Salinas Peroano (top, Nt = 2,052; Nsp = 59) and Pontinha (bottom, Nt = 1,621; Nsp = 54) sambaquis. (Adapted from Scheel-Ybert 2000.)

Figure 11.6. Charcoal percentage diagram of the Jabuticabeira II sambaqui; Nt = 1,904; Nsp = 51 (Adapted from Scheel-Ybert 2001b.)

Figure 11.7.  Summary charcoal diagrams of the Forte, Salinas Peroano, Boca da Barra, Ponta da Cabeça, Beirada, Pontinha (Rio de Janeiro State), and Jabuticabeira II (Santa Catarina State) sambaquis. Radiocarbon ages are expressed as 14C yr BP normalized to a δ13C of 25‰ PDB. Calibrated ages (yr cal BP) are presented in parentheses. (Adapted from Scheel-Ybert 2000.)

252

social archaeobotany

Figure 11.8. Transversal plane SEM views of some of the charcoal fragments identified in Forte sambaqui. (a) Securinega sp. (Euphorbiaceae), 200–210 μm; (b) Rheedia sp. (Guttiferae), 240–50 cm; (c) Lauraceae, 70–80 cm; (d) Myrtaceae type 1, 240–50 cm. Scale bar = 100 µm.

which probably led to widening the site catchment area for firewood. According to this last hypothesis, mangrove vegetation may or may not have varied in the environment, but its presence was not recorded before the extension of the site catchment area. At Cabo Frio (Forte, on the east side of the channel; Salinas Peroano and Boca da Barra, on the west side) the reduction in mangrove vegetation (figure 11.7) was associated with a drier climate when lower precipitation probably produced a rise in salinity in the neighboring lagoon. Indeed, changes in lagoon sedimentation observed in several cores suggest regional climatic variations during the Holocene. Our results have been compared to the curve of the variation in composition of isotopic

Figure 11.9. Charcoal fragments of Avicennia aff. schaueriana (Verbenaceae). Left, normal structures; right, signs of decay before charring. Scale bar = 100 µm.

254

social archaeobotany

carbonates in sediments of the Araruama Lagoon (Tasayco-Ortega 1996). Salinity in the lagoon was low until about 5000 BP. This can be correlated with the high percentages of mangrove elements at the base of the charcoal diagrams. After this point, salinity was high until about 2300 BP. During this period, mangrove was rare on both sides of the channel. A net reduction of salinity is recorded between 2300 and 2000 BP. This can be correlated with the reestablishment of mangrove elements at the Forte sambaqui after 2300 BP, which is certainly related to a humid episode. We are unable to estimate its duration, but the fact that it is not recorded on the east side of the channel suggests it was brief. The low salinity levels recorded before 5000 BP and between 2300 and 2000 BP correspond to periods when the mangrove was well developed. The period of high salinity between 5000 and 2300 BP, related to a drier climate, corresponds to a decrease in mangrove vegetation in all sites. The stability of the mainland vegetation seems to contradict the climatic variations recorded by mangrove elements. However, this is probably a consequence of the edaphic character of the coastal vegetation, especially the restinga ecosystem (related to sandy soils and to the geomorphological nature of sandy beach ridges), which results in a greater resilience to climatic change (Scheel-Ybert 2001b). Moreover, it must be taken into account that the absence of significant variation in the land plant associations around sambaquis, here considered as “vegetation stability,” cannot be associated with a stationary or immutable landscape. In addition to the attested climatic variations and to the activities of human populations who inhabited the region for thousands of years, significant sea level variations changed the coastal outline and the local lagoon conformation (Angulo and Lessa 1997; Martin, Suguio, and Flexor 1979). What probably happened both in southeastern as well as in southern Brazil was a retreat of the coastline and a reorganization of the coastal lagoons’ and bays’ size and distribution, as a consequence of a lower sea level. In this process, the restinga vegetation advanced inland, changing its distribution in the landscape, but maintaining its structure and floristic composition. Although the paleoenvironmental history of Brazil is still poorly known, various studies have demonstrated that significant climate changes affected this area during the last millennia (e.g., Ledru et al. 1995; Ybert et al. 2003). Climate changes led to vegetational modifications in all the studied regions. However, paleoenvironmental studies in coastal areas are still rare. In Rio Grande do Sul State, palynological evidence from the coastal area has shown climatic oscillations, but these were essentially inferred from the alternation between marine, marshland, and continental elements in the lagoon cores, and no significant variations in the continental vegetation have been recorded during the Late Holocene

Anthracological Research on the Brazilian Coast

255

(Cordeiro and Lorscheitter 1994; Neves and Lorscheitter 1991). In São Paulo State, the palynological study of a peat core from a flooded forest showed no variations in the coastal vegetation that could be attributed to climate changes between ca. 4300 and 1000 BP (Ybert et al. 2003). This seems to confirm that coastal formations, strongly dependent on edaphic conditioning, are much more resilient to change from climatic variability. Most of the ancient taxa identified correspond to plant associations currently existing in the region. For evaluation purposes, anthracological results from the southeastern sites were compared to phytosociological data of similar plant associations from a variety of Brazilian localities using correspondence analysis (CA). The analysis was conducted at the genus level, to allow comparison with the charcoal assemblages (ScheelYbert 2000). Despite floristic similarities between restinga and Atlantic forests (Araujo and Lacerda 1987; Rizzini 1979), CA that takes into consideration various modern phytosociological studies produces a clear separation of these formations, even at the genus level (figure 11.10). Different factor analyses demonstrate that the archaeological samples are more similar to each other than to present vegetation types, when taken separately, even though all fossil samples are related to the tropical Atlantic and restinga forest samples, and not to subtropical plant formations (figure 11.10). The mixing of taxa from different plant associations in the archaeological samples explains their resemblance to each other, while their similarity to modern tropical vegetation types corroborates the robustness of the anthracological method (Scheel-Ybert 2000). More detailed analysis has shown a stronger resemblance of some sites to some specific vegetation types, indicating that charcoal assemblages mainly represent local vegetation. The analysis of isolated sites provides only local information, and a regional reconstruction depends on the study of numerous sites distributed over a relatively large area. In spite of all the processes that affect a charcoal assemblage from the moment of wood gathering until its analysis (physical transformations of charred wood, postdepositional processes, charcoal fragmentation, etc.), our results show that the anthracological taxonomic assemblage resembles present-day plant associations fairly well (Scheel-Ybert 2000).

Conclusions Some archaeologists have attributed supposed economic modifications and, eventually, the disappearance of sambaqui populations to environmental changes (Dias 1987; Uchôa 1981/82). On the one hand, this assertion is related to an outdated scheme of interpretation that associated

Figure 11.10.  Factor correspondence analyses. (A) Phytosociological studies only. (B) Correlation between archaeological samples and phytosociological studies. Each symbol represents a group of superposed samples. (C) Archaeological samples only. Sambaqui Meio modern sample is supplementary. R = restinga; F = Atlantic Forest; letters after the numbers stand for geographical location. (Adapted from Scheel-Ybert 2000.)

Anthracological Research on the Brazilian Coast

257

these people with itinerant bands of mollusk gatherers. On the other hand, it is essentially based on the widespread idea that coastal environments are extremely sensitive to climate changes and that coastal vegetation responds quickly to such oscillations. However, new findings from the archaeology and archaeobotany of sambaqui sites contradict these views. Sambaqui people were sedentary and had territorial stability. Some kind of social stability persisted for a very long period while new territories were being incorporated. It was a peaceful society in a process of increasing social complexity, living in a relatively stable environment, whose sociocultural system was maintained for at least 6,000 years. These populations occupied the coast for a long period of time until the expansion of horticulturalists and ceramic-producing groups, around the beginning of the Christian era. We can speculate that the social and environmental stability gave rise to an extremely well-structured society that was unable to create an efficient answer—for example, war—to defend its territory when it was invaded by the Tupi from the Amazon region and the Gê from Central Brazil, both of which were very warlike groups. Environmental stability, having important outcomes for prehistoric populations, even if not determinant for their lifeways, was probably a decisive factor in the expansion of sambaquis, in their sedentariness, and in preserving their sociocultural system. On the other hand, such stability is surprising in a territory that has been intensely occupied for thousands of years, where sambaqui builders intentionally modified the landscape by the construction of impressive mounds. Yet, even if plant consumption was important, these populations were specialized in the exploitation of aquatic resources, which causes very little damage to the terrestrial environment. Also, the practice of human management or horticulture should not imply deforestation or milieu overexploitation, as suggested by Wesolowski (2000) and Scheel-Ybert (2001a). Exploitation of deadwood for firewood is very conservationist in environmental terms. We have no idea, up to now, of the size of individual communities, and in consequence it is difficult to estimate the impact of woodcutting for timber (house construction, wood implements, canoes, etc.). Nevertheless, present data seem to point to specific woodland management strategies that ensured the sustainable exploitation of natural resources and common land. In addition to conducting the first paleoenvironmental reconstruction of this area (the first one associated with Brazilian archaeological sites), archaeobotanical research in coastal Brazil is providing new clues to understanding wood resources exploitation, social organization, and diet in the moundbuilders society. Also, this work is contributing to the important development of methodological and taphonomical aspects of anthracology in tropical regions, as well as advancing archaeobotanical

258

social archaeobotany

research, since both the paleoenvironmental and paleoethnobotanical approaches are based on the same data set.

Acknowledgments The authors are supported by funds from the CNPq (National Council of Technological and Scientific Development) and FAPERJ (Carlos Chagas Filho Foundation for Research Support of the State of Rio de Janeiro).

Note 1. Gini-Lorenz concentration indexes provide a measure of plant equitability (an aspect of plant diversity). The higher the concentration index, the lower its evenness in a plant formation. It can be used to evaluate sampling reliability in modern and fossil assemblages, because small samples generally present an abnormally even distribution, since the vegetation structure is not well represented (Scheel-Ybert 2002).

References Angulo, R. J., and G. Lessa. 1997. “The Brazilian Sea Level Curves: A Critical Review with Emphasis on the Curves from the Paranaguá and Cananéia Regions.” Marine Geology 140:141–66. Araujo, D. S. D., and R. P. B. Henriques. 1984. “Análise florística das restingas do estado do Rio de Janeiro.” In Restingas: Origem, Estrutura, Processos, edited by L. D. Lacerda, D. S. D. Araujo, R. Cerequeira, and B. Turcq, 159–94. Niterói: CEUFF. Araujo, D. S D., and L. D. Lacerda. 1987. “A natureza das restingas.” Ciência Hoje 6 (33): 26–32. Barbiére, E. B. 1984. “Cabo Frio e Iguaba Grande, dois microclimas distintos a um curto intervalo especial.” In Restingas: Origem, Estrutura, Processos, edited by L. D. Lacerda, D. S. D. Araujo, R. Cerequeira, and B. Turcq, 3–14. Niterói: CEUFF. Briane, D., and J. Doat. 1985. Guide technique de la charbonisation: La fabrication du charbon de bois. Aix-en-Provence: EDISUD, Centre Technique Forestier Tropical. Carcaillet, C., and M. Thinon. 1996. “Pedoanthacological Contribution to the Study of the Evolution of the Upper Treeline in the Maurienne Valley (North French Alps): Methodology and Preliminary Data.” Review of Palaeobotany and Palynology 91:399–416.

Anthracological Research on the Brazilian Coast

259

Chabal, L. 1992. “La représentativité paléo-écologique des charbons de bois archéologiques issus du bois de feu.” Bulletin de la Société Botanique de France 139 (2/3/4): 213–36. ———. 1997. “Forêts et sociétés en Languedoc (Néolithique Final, Antiquité Tardive): L’anthracologie, méthode et paléoécologie.” Documents d’Archéologie Française 63:1–188. Claasen, C. 1991. “Normative Thinking and Shell-Bearing Sites.” In Archaeological Method and Theory, edited by M. B. Schiffer, 249–98. Tucson: University of Arizona Press. Cordeiro, S. H., and M. L. Lorscheitter. 1994. “Palynology of Lagoa dos Patos Sediments, Rio Grande do Sul, Brazil.” Journal of Paleolimnology 10:35–42. De Blasis, P. 2001. “Os mais antigos soberanos da costa.” In XI Congresso da Sociedade de Arqueologia Brasileira, Resumos, 43–44. Rio de Janeiro: SAB. De Blasis, P. A. D., S. K. Fish, M. D. Gaspar, and P. R. Fish. 1998. “Some References for the Discussion of Complexity among the Sambaqui Moundbuilders from the Southern Shores of Brazil.” Revista de Arqueologia Americana 15:75–105. De Masi, M. A. N. 1999. “Prehistoric Hunter-Gatherer Mobility on the Southern Brazilian Coast: Santa Catarina Island.” PhD diss., Stanford University. Dias, O. 1987. “Pré-história e arqueologia da região sudeste do Brasil.” Boletim do Instituto de Arqueologia Brasileira, sér. Catálogos 3:155–64. Eastoe, C. J., S. Fish, P. Fish, M. D. Gaspar, and A. Long. 2002. “Reservoir Corrections for Marine Samples from the South Atlantic Coast, Santa Catarina State, Brazil.” Radiocarbon 44 (1): 145–48. Figuti, L. 1993. “O homem pré-histórico, o molusco e os sambaquis: Considerações sobre a subsistência dos povos sambaquieiros.” Revista do Museu de Arqueologia e Etnologia, São Paulo 3:67–80. Fish, S. K., P. A. D. De Blasis, M. D. Gaspar, and P. R. Fish. 2000. “Eventos incrementais na construção de sambaquis, litoral sul do estado de Santa Catarina.” Revista do Museu de Arqueologia e Etnologia, São Paulo 10:69–87. Gaspar, M. D. 1992. “Aspectos da organização social de um grupo de pescadores, coletores e caçadores que ocupou o litoral do estado do Rio de Janeiro.” In Paleontologia e Paleoepidemiologia: Estudos Multidisciplinares, edited by A. J. G. Araújo and L. F. Ferreira, 95–109. Rio de Janeiro: Ensp, Fundação Oswaldo Cruz. ———. 1995/96. “Território de exploração e tipo de ocupação dos pescadores, coletores e caçadores que ocuparam o litoral do estado do Rio de Janeiro.” CLIO, sér. Arqueologia 1 (11): 153–74. ———. 1996. “Análises das datações radiocarbônicas dos sítios de pescadores, coletores e caçadores.” Boletim do Museu Paraense Emilio Goeldi, sér. Ciências da Terra 8:81–91. ———. 2000. Sambaqui: Arqueologia do litoral Brasileiro. Rio de Janeiro: Jorge Zahar Editor.

260

social archaeobotany

Heredia, O. R., M. C. Tenório, M. D. Gaspar, and A. M. G. Buarque. 1989. “Environment Exploitation by Prehistorical Population of Brazil.” In Coastlines of Brazil, edited by C. Neves, 230–39. New York: American Society of Civil Engineers. Klökler, D. M. 2000. “Construindo ou deixando um sambaqui? Análise de sedimentos de um sambaqui do litoral meridional Brasileiro: Processos formativos, região de Laguna.” MSc diss., Museu de Arqueologia e Etnologia Universidade de São Paulo. Kneip, L. M. 1980. “A seqüência cultural do sambaqui do Forte, Cabo Frio, Rio de Janeiro.” Pesquisas, sér. Antropologia 31:87–100. Ledru, M. P., P. I. S. Braga, F. Soubiès, M. Fournier, L. Martin, K. Suguio, and B. Turcq. 1995. “The Last 50,000 Years in the Neotropics (Southern Brazil): Evolution of Vegetation and Climate.” Palaeogeography, Palaeoclimatology, Palaeoecology 123:239–57. Lessa, A., and J. C. Medeiros. 2001. “Reflexões preliminares sobre a questão da violência em populações construtoras de sambaquis: Análise dos Sítios Cabeçuda (SC) e Arapuan (RJ).” Revista do Museu de Arqueologia e Etnologia, São Paulo 11:77–93. Lima, T. A. 1997. “The Shellmound-Builders: Emergent Complexity along the South/Southeast Coast of Brazil.” Paper presented at the 62nd Annual Meeting of the Society for American Archaeology, Nashville. Martin, L., K. Suguio, and J. M. Flexor. 1979. “Le quaternaire marin du littoral Brésilien entre Cananéia (SP) et Barra de Guaratiba (RJ).” In Proceedings of the International Symposium on Coastal Evolution in the Quaternary, 296– 331. São Paulo. Neves, P. C. P., and M. L. Lorscheitter. 1991. “Upper Quaternary Palaeoenvironments in the Northern Coastal Plain of Rio Grande do Sul, Brazil.” Quaternary of South America and Antarctic Peninsula 9:39–67. Neves, W. A., and V. Wesolowski 2002. “Economy, Nutrition and Disease in Prehistoric Coastal Brazil: A Case Study from the State of Santa Catarina.” In The Backbone of History: Health and Nutrition in the Western Hemisphere, edited by R. H. Steckel and J. C. Rose, 376–402. Cambridge: Cambridge University Press. Nimer, E. 1989. Climatologia do Brasil. 2nd ed. Rio de Janeiro: IBGE. Popper, V. S. 1988. “Selecting Quantitative Measurements in Paleoethnobotany.” In Current Paleoethnobotany: Analytical Methods and Cultural Interpretation of Archaeological Plant Remains, edited by C. A. Hastorf and V. S. Popper, 53–71. Chicago: University of Chicago Press. Rizzini, C. T. 1979. Tratado de fitogeografia do Brasil. Vol. 2. São Paulo: Hucitec. Sá, C. F. C. 1992. “A vegetação da restinga de Ipitangas, reserva ecológica estadual de Jacarepiá, Saquarema (RJ): Fisionomia e listagem de angiospermas.” Arquivos do Jardim Botânico do Rio de Janeiro 31:87–102.

Anthracological Research on the Brazilian Coast

261

Scheel-Ybert, R. 2000. “Vegetation Stability in the South-Eastern Brazilian Coastal Area from 5500 to 1400 yr BP Deduced from Charcoal Analysis.” Review of Palaeobotany and Palynology 110:111–38. ———. 2001a. “Man and Vegetation in the Southeastern Brazil during the Upper Holocene.” Journal of Archaeological Science 28 (5): 471–80. ———. 2001b. “Vegetation Stability in the Brazilian Littoral during the Late Holocene: Anthracological Evidence.” Revista Pesquisas em Geociências 28 (2): 315–23. ———. 2002. “Evaluation of Sample Reliability in Extant and Fossil Assemblages.” In Charcoal Analysis: Methodological Approaches, Palaeoecological Results and Wood Uses, edited by S. Thiébault, 9–16. BAR International Series, vol. 1063. Oxford: Archaeopress. Scheel-Ybert, R., M. Scheel, and J. P. Ybert. 2002. Atlas Brasil: Databank for Charcoal Analysis and Computerized Key to Charcoal Determination (in Portuguese, English and French). Version 2.2. Compact Disc. Tasayco-Ortega, L. A. 1996. “Variations paléohydrologiques et paléoclimatiques d’une région d’upwelling au cours de l’Holocène: Enregistrement dans les lagunes côtières de Cabo Frio (état de Rio de Janeiro, Brésil).” PhD diss., University Pierre et Marie Curie. Tenório, M. C., M. Barbosa, and T. Portela. 1992. “Pesquisas arqueológicas no sítio Ponta de Cabeça, Arraial do Cabo, Rio de Janeiro.” Anais da IV Reunião da Sociedade de Arqueologia Brasileira 2:279–91. Thiébault, S. 1997. “Early-Holocene Vegetation and the Human Impact in Central Provence (Var, France): Charcoal Analysis of the Baume de Fontbrégoua.” The Holocene 7 (3): 343–49. Uchôa, D. P. 1981/82. “Ocupação do litoral sul-sudeste Brasileiro por grupos coletor-pescadores Holocênicos.” Arquivos do Museu de História Natural, UFMG 6–7:133–43. Wesolowski, V. 2000. “A prática da horticultura entre os construtores de sambaquis e acampamentos litorâneos da região da Baía de São Francisco, Santa Catarina: Uma abordagem bio-antropológica.” MSc diss., Universidade de São Paulo. Widmer, R. J., 1989. “Archaeological Research Strategies in the Investigation of Shell-Bearing Sites, a Florida Perspective.” Paper presented at the 74th Annual Meeting of the Society for American Archaeology, Atlanta. Willcox, G. H. 1974. “A History of Deforestation as Indicated by Charcoal Analysis of Four Sites in Eastern Anatolia.” Journal of the British Institute of Archaeology at Ankara 24:117–33. Ybert, J. P., W. M. Bissa, E. L. M. Catharino, and M. Kutner. 2003. “Environmental and Sea-Level Variations on the South-Eastern Brazilian Coast during the Late Holocene with Comments on Prehistoric Human Occupation.” Palaeogeography, Palaeoclimatology, Palaeoecology 189:11–24.

262

social archaeobotany

Ybert, J. P., R. Scheel, and M. D. Gaspar. 1997. “Descrição de alguns instrumentos simples utilizados para a coleta e concentração de elementos fósseis de pequenas dimensões de origem arqueológica ou pedológica.” Revista do Museu de Arqueologia e Etnologia, São Paulo 7:181–89.

C h ap t e r 1 2

Rice of Asian Origin Y o -I c h i r o S at o

Asian common rice has been supporting more than two billion people all over the world. It is certainly an important crop with a long history of cultivation, yet until recently there has been little agreement about its origin and evolution. In the last decade, molecular genetic analyses have strongly indicated different parentage for the two major varietal groups of cultivars, indica and japonica. Recent archaeological studies, particularly those in China, have suggested cultivation of incipient japonica in the middle and lower basins of the Yangtze River. Here I present a new hypothesis on the phylogeny and homeland of cultivated rice.

Phylogenetic Relationships of Wild and Cultivated Rice DNA Analyses of indica and japonica Asian cultivated rice has been recognized for over a century as pertaining to a single species (O. sativa). The ancestral species has been considered to be an AA genome species (O. rufipogon Griff.) or, previously, an Asian form of O. perennis Moech. Two major clusters of O. sativa have been recognized at subspecies level: indica and japonica (Kato et al. 1930). How do indica and japonica groups differ? Are they different types originating from the same species? From the isozyme and DNA analyses performed in recent years, it is reasonable to say that indica and japonica belong to two different races. Isozyme analyses in the 1980s revealed that these two races could be distinguished from the associations of alleles at isozyme loci (Glaszmann 1987; R. Sano and Morishima 1992). In the 1990s isozyme analyses

265

266

genetics in archaeobotany

were followed by DNA analyses. Polymorphism at noncoding regions, in which the DNA sequences do not act as a gene, was found to be useful for distinguishing between indica and japonica. In phylogenetic analyses, restriction fragment length polymorphism (RFLP) in nuclear DNAs (Kawase et al. 2008; Z. Wang and Tanksley 1989; Zhang et al. 1992, etc.) apparently showed that indica and japonica belonged to different clades. Differences in the length of spacers in rDNA codons were also observed in cultivates of indica and japonica (Y. Sano and R. Sano 1990). The number of ribosomal DNA (rDNA) loci could also be counted by fluorescent in situ hybridization (FISH). Strains of japonica tended to have two NORs, while many strains of indica cluster showed four NORs (Fukui, Oomido, and Khush 1993). The degree of differentiation of DNA highlighted by these analyses suggests that indica and japonica may have very different origins. DNA is stored mainly in the nucleus, but also exists outside the nucleus, in chloroplasts and mitochondria (organelle DNA). In most higher plants, organelle DNA is inherited only by the maternal line, and through its analysis, it is possible to detect maternal lineage. A sixty-nine base-pair deletion of chloroplast DNAs (cpDNAs), called the D-type, is frequent (more than 90% occurrences) in indica, and it is rare (less than 8% occurrences) in japonica (Chen et al. 1993). The plastid subtype ID region (PS-ID) (Nakamura et al. 1997) in the cpDNA is also distinct in the two groups: many japonica cultivars had 6C7A or 7C6A types, while indica cultivars display other types (figure 12.1). The results of these analyses clearly show that indica and japonica represent two different maternal lineages. Genealogical studies by the late H. I. Oka and his group suggested that rice originated from one common source (Oka 1958, 1983, 1988; Oka and Chang 1962; Oka and Morishima 1982). Oka’s theory that “wild rice has no indica-japonica differentiation” is doubtful, because recent studies show that indica-japonica differentiation is present in both the chloroplast and nuclear DNAs of wild rice species. Until recently, the explanation was that rice could be divided into wild and cultivated groups and that cultivated rice could be further divided into indica and japonica types. The theory proposed here is that ancestral wild rice of modern-day cultivated strands originally differentiated into indica and japonica. Only after this separation did indica and japonica become domesticated. It is therefore inaccurate to say that indica and japonica are the same type of “rice.” It is more plausible that they are distinct crops derived independently from different predomestication races or species.

Evidence for the indica-japonica Differentiation in Wild Rice Chen et al. (1993) discovered that the D-type cpDNA was frequent in O. nivara (an annual wild rice), while being rare in the perennial species

Rice of Asian Origin

267

Figure 12.1.  Sequence of the PS-ID region of Oryza species. (Source: Nakamura and Sato, unpublished data.)

O. rufipogon. This evidence strongly supports the opinion that indica and japonica were derived from O. nivara and O. rufipogon, respectively, at least as far as cpDNA is concerned. The D-type cpDNA originated in the O. rufipogon complex, and was transferred to indica cultivars (Chen et al. 1993). Analyses of RFLPs in the nuclear DNA of wild rice species also reveal two distinct groups that correspond to the indica and japonica cultivated varieties, suggesting that the indica/japonica differentiation occurred before the domestication of rice (Y. Fukuta, unpublished data). A phylogenetic tree based on the length polymorphisms of a family of short inserted elements (pSINE-r1) showed that indica cultivars are very similar to O. nivara, while japonica cultivars are very similar to O. rufipogon (Cheng et al. 2003; figure 12.2). Strains of vegetatively propagated O. rufipogon, from the Mekong delta (Vietnam), had the same combination of alleles that characterizes japonica (Yamanaka et al. 2002).

Geographic Origin and Distribution of Wild and Cultivated Rice Previous Hypotheses Earlier theories on the origin of rice were based on dispersal models (Harlan 1975). In many of these works, China and India were regarded as the center of origin for domesticated rice (see, e.g., De Candolle [1883] 1959; Randhawa 1980; Zhou 1948). However, these early works were mostly based on limited evidence and few archaeological excavations. It then became accepted that the Asian cultivated rice (O. sativa L.) originated in the Himalayan foothills (Morinaga 1955) or the Assam-Yunnan

268

genetics in archaeobotany

Figure 12.2. Classification of the AA genome of wild-rice strains based on pSINE-v1 polymorphism. I: O. s. indica; N: O. nivara; R: O. rufipogon; J: O. s. japonica.

area (Chang 1976; Watabe 1977). This area stretches from the Assam district of India, upper Burma, northern Thailand, Laos, and Vietnam to the southwestern provinces of China. Indeed, Watabe (1977) suggested, from observing the shape of spikelets used as temper in old bricks from South and Southeast Asia, that cultivated rice was domesticated in the “Assam-Yunnan area” (figure 12.3). A similar hypothesis was also proposed by Chang (1976).

Remains in the Middle Yangtze River Basin According to the archaeological evidence, rice cultivation in southwest China (including the Yunnan province) started ca. 4,500 years ago. This is more than 2,500 years later than rice cultivation in the middle and lower basin of the Yangtze River (Luo 1995; T. Wang 1986). Therefore, the more western Assam-Yunnan area should be excluded as the homeland of cultivated rice. On the other hand, from the evidence of phytoliths from archaeological sediments, rice might have represented a gathered source

Rice of Asian Origin

269

Figure 12.3. Distribution of archaeological sites with evidence for rice cultivation.

of food from as early as 16,000–11,000 years ago. A few sites dating from the earlier period (10,000 years BP or older) were discovered in caves in limestone mountains on the south margin of the middle Yangtze River basin. At Xianrendong, for example, a site in Jiangxi Province, many phytoliths derived from rice husks were discovered in layers as old as 11,000 to 14,000 years BP (Zhao et al. 1995). Unfortunately, the morphology of husk phytoliths gives no information about the domestic or wild status of the rice. Four rice grains were recovered in 1996 from the site of Yuchanyan (12,000 BP) in southern Hunan Province (Sato, unpublished data). Judging from the slender shape and the traces of natural shattering, these grains appear to be from wild rice. The oldest firmly dated remains of cultivated rice in the lower basin of the Yangtze River were obtained from the site of Hemudu in Zhejiang Province (ca. 7,000 years old, based on radiocarbon dating) (Luo 1995; T. Wang 1986). A great number of rice grains were excavated from the fourth layer, together with fragments of wood, suggesting that the area was part of a storage house. Sato (2002) observed five spikelets of wild rice among eighty-one samples from this site. These five spikelets had traces of a developed awn with a long needle as well as the shattering characteristics, both significant features of wild rice. Some other spikelets, however, had features intermediate between the domesticated and

270

genetics in archaeobotany

nondomesticated forms. This evidence may indicate the presence of primitive cultivars during the Hemudu period and a process of incipient domestication. In the middle and lower basins of the Yangtze River, more than 100 sites with rice cultivation have been excavated so far (Yan 1982, 2002). The age of these sites is reported to be not older than 9000 BP. However, their dating is based on pottery typology and not on radiocarbon dating. Among these sites, the Chengtoushan site in Hunan Province deserves special attention. The archaeological record from Chengtoushan has shown that a complex agricultural system was established here, and that the settlement was a central city of the Yangtze River Civilization. Rice grains were also part of the archaeological record, and some were identified as japonica on the basis of their DNA sequence in the PS-ID region (6C7A) (Yano et al. 2000, table 1). In India and Pakistan, many rice sites have been excavated all over the country, especially in the Ganga basin, but none of these is older than 2200 BC (Randhawa 1980; see also Madella, this volume). In Thailand and Vietnam, some sites are known, but they are not older than 3000 BC (Agrawal and Kharakwal 2002). According to the data available so far, rice cultivation in the lowland Asian tropics seems to have began not earlier than 3000 BC. In summary, the archaeological studies indicate that the first rice cultivars appeared along the middle-lower Yangtze valley around 7000– 9000 BC.

Varieties of Rice Cultivars in the Yangtze Valley What type of rice originated in the middle-lower Yangtze valley? Various hypotheses have been suggested, and many researchers think that both indica and japonica originated in this area. However, the main evidence on which this hypothesis is based is that the carbonized rice grains excavated from the Yangtze valley archaeological sites (e.g., Hemudu) are round and slender. You (1979), on the basis of the variability of shapes of the excavated grains, suggested that the rice in ancient China (Hemudu period, ca. 7000 years BP) was a mixture of indica and japonica. Originally it was thought that indica grains were slender and japonica round (Morinaga 1968), but recent studies show that grain shape does not reliably distinguish indica and japonica. This difference seems to reflect popular views with no substantial basis (Sato 1991). Abundant rice grains have been excavated from archaeological sites. The grains are often blackish and seem to be charred. In fact, some of them are burned, while others were still fresh when buried. Several researchers have attempted the extraction of DNA from these “fresh” and

Rice of Asian Origin

271

unburned rice grains. DNA in the remains was usually fragmented because of the long burial, but up to several hundred base-pairs DNA could be amplified using polymerase chain reaction (PCR) (Yamanaka et al. 2002; Yano et al. 2000). Detection of the PS-ID type (Nakamura et al. 1997) has been used to identify the varietal group of ancient rice cultivars, as mentioned above. To date, the PS-ID types of twenty-eight grains excavated in China, Korea, Japan, and Thailand have been identified by direct sequencing of PCR products. Among these grains, twenty-seven are from China, Korea, and Japan (dated from 1900 to 6000 years BP) and displayed either 6C7A or 7C6A sequences, which are both japonica-specific. Six grains out of seven analyzed from the Chengtoushan site had 6C7A cpDNA, suggesting that the cultivar in this settlement was predominantly japonica (Yano et al. 2000). Only one grain from Thailand (Lop buri Province) has been successfully analyzed, and it was identified as a 7C7A type, which is typical of indica (Sato, unpublished data). The grain was retrieved from a brick of the Narayana Rachanives Palace, built in AD 1622 (RRI 1992). These data strongly suggest that eastern Asia was the homeland of japonica.

Ecological Environment in the Area of Rice Origin During the early Holocene, the middle-lower Yangtze River valley was covered with a dense laurilignosa (laurel) forest, and the river created wide marshlands in the lowlands. Between the marshland and forest, there was a plain zone that was rich in water during the rainy season and dry in the dry season. This type of environment is assumed to be ideal for the origin of rice. It is speculated that wild rice, perhaps O. rufipogon, was abundant in this region. When rice cultivation expanded, the climate was shifting from warm to cold, and the northern limit of the Yangtze valley wild rice contracted (to 40°N latitude). The wild O. rufipogon is, in effect, japonica with a perennial trait (e.g., the production of long tillers as lateral shoots that send up new vegetative parts intended for the wet season).

Origin of japonica in the Yangtze Valley As noted above, indica and japonica have different origins, with no evidence for their origin in the same region. This is the view presented in Sato (1996, figure 4). Evidence for this theory is provided by cpDNA differences in chlorophyll from indica and japonica DNA derived from carbonized rice grains. When studies were carried out on carbonized rice grains uncovered in several sites near the Yangtze River, we found that

272

genetics in archaeobotany

each grain carries its own DNA trait derived from japonica. Although the result has to be further substantiated owing to the small number of grains and excavation sites used, it is suggested here that japonica was cultivated at ca. 4000–5000 BC, strengthening our theory that japonica originated in the Yangtze valley. Chinese civilization is mainly characterized as a Yellow River Civilization, with the Yangtze valley regarded as undeveloped, but there have been suggestions of an unknown Yangtze valley civilization that may have existed much earlier. Many of the oldest rice cultivation sites have been found in the middle and lower Yangtze valley in the past ten years (i.e., between coastal Zhejiang and southeast Jiangsu, and the Tai Lake region). The size and quantity of ancient artifacts from the capital of Zhejiang, various Liangzhu locations, western Hangzhou, and western Hunan suggest an ancient state. The Liangzhu site is dated to 3000 BC at the earliest, while the Hunan sites date to 4000 BC. Rice cultivation spread through the Yangtze valley at 4000– 3000 BC, and it could be said that the Yangtze valley civilization was based on japonica.

Origin of indica We must also reconsider the origin of indica. Southeast and South Asia may be places where the indica varieties originated. However, rice cultivation in this area does not pre-date 2000 BC in the southeast regions, and the onset of rice cultivation may have been much later compared to China. This is possibly due to differences in the ecosystems and the effect of climate change. In a temperate climate the problems resulting from a cooling of the environment and preserving gathered food might have pushed the people to begin rice cultivation and hence to start agriculture at an early date. While agriculture in the West emerged in the Fertile Crescent at the time of the so-called Neolithic revolution (Childe 1925), studies on the origins of rice and other crops indicate a range of independent agricultural activities in eastern Asia. A recent study of chloroplast DNA from wild rice species suggests that O. officinalis, carrying a CC genome, is the most likely maternal ancestor of the indica varieties (I. Nakamura, unpublished data). Masood et al. (2004) determined all the sequences of cpDNA from a strain of O. nivara, and concluded that the sequences differed from indica cultivars. However, the donor of nuclear DNA would be O. nivara or a closely related species having an AA genome. A possible explanation for the origin of indica is that cytoplasm was donated by O. officinalis (and with it the organelles DNA), while nuclear genome was donated by O. nivara or a related species. Further studies are needed to untangle the origin of rice.

Rice of Asian Origin

273

Conclusion Asian cultivated rice is now thought to have originated by diffusion. Cultivars of japonica were probably the first developed in the middle and lower basins of the Yangtze River 11,000 to 14,000 years ago, from a type of rufipogon with a nondeleted cpDNA. Domestication was perhaps started when strains of O. rufipogon acquired relatively high seed productivity as a natural response to stress conditions created by a cooler and drier climate. Cultivars of japonica were diversifying all over China by about 3,000 years ago, at the time of the unification of the China states. Some strains may have reached the tropics across the Yunnan mountains. If cultivated japonica strains hybridized with indigenous annual types of wild rice to produce indica cultivars, then this may explain the later emergence of indica in the marshland of the tropical plain. The high genetic diversity in rice of the “Oriental fertile crescent” is explained by the introduction of indica and japonica cultivars from the tropics and eastern China, respectively, and subsequent hybridization between them (Second 1981). The resulting genetic diversity has been preserved there because of the complex conditions of climate, geography, and cultural diversity. A sustainable way of agriculture may play an important role in the preservation of such genetic diversity. The genetic diversity of rice in western Asia represents a secondary center of domestication (Vavilov 1926). In future work, analysis of plant and animal remains (bioarchaeology), in particular archaeological DNA analysis, will be increasingly important for unraveling the complex history of rice.

References Agrawal, D. P., and J. S. Kharakwal. 2002. South Asian Prehistory. Delhi: Aryan Books International. Chang, T. T. 1976. “The Origin, Evolution, Cultivation, Dissemination, and Diversification of Asian and African Rices.” Euphytica 25:431–41. Chen, W. B., I. Nakamura, Y.-I. Sato, and H. Nakai. 1993. “Distribution of Deletion Type in cpDNA of Cultivated and Wild Rice.” Japanese Journal of Genetics 68:597–603. Cheng, C., R. Motohashi, S. Tsuchimoto, Y. Fukuta, H. Ohtsubo, and E. Ohtsubo. 2003. “Polyphyletic Origin of Cultivated Rice: Based on the Interspersion Patterns of SINEs.” Molecular Biology and Evolution 20:67–75. Childe, G. V. 1925. The Dawn of European Civilization. London: Kegan Paul. De Candolle, A. (1883) 1959. Origin of Cultivated Plants. New York: Noble Offset Printers.

274

genetics in archaeobotany

Fukui, K., N. Oomido, and G. S. Khush. 1993. “Variability in rDNA Loci in the Genus Oryza Detected through Fluorescence in Situ Hybridization.” Theoretical and Applied Genetics 87:893–99. Glaszmann, J. C. 1987. “Isozymes and Classification of Asian Rice Varieties.” Theoretical and Applied Genetics 74:21–30. Harlan, J. R. 1975. Crops and Man. Madison, WI: American Society of Agronomy. Kawase, M., N. Kishimoto, T. Tanaka, T. Yoshimura, S. Yoshimura, K. Saito, A. Saito, M. Yano, N. Takeda, N. Nagamine, and M. Nakaghara. 2008. “Intraspecific Variation and Genetic Differentiation Based on Restriction Fragment Length Polymorphism in Asian Cultivated Rice, Oryza sativa L.” In Rice Genetics II, Proceedings of the 2nd International Rice Genetics Symposium, 14-18 May 1990, edited by S. J. Banta, 467–74. Manila: International Rice Research Institute. Kato, S. 1930. “On the Affinity of the Cultivated Rice Varieties of Rice Plants, Oryza sativa L.” Journal of the Faculty of Agriculture, Kyushu University 2:241–75. Luo, E. H. 1995. A Summary of Excavated Samples in the Neolithic Age in China. Kyoto: Kyoto University. Masood, M. S., T. Nishikawa, S.-i. Fukuoka, P. K. Njenga, T. Tsudzuki, K.-i. Ka­dowaki. 2004. “The Complete Nucleotide Sequence of Wild Rice (Oryza nivara) Chloroplast Genome: First Genome-Wide Comparative Sequence Analysis of Wild and Cultivated Rice.” Gene 2:133–39. Morinaga, T. 1968. “Origin and Geographical Distribution of Japanese Rice.” Japan Agricultural Research Quarterly 3:1–5. Nakamura, I., N. Kameya, K. Kato, S. I. Yamanaka, H. Jomori, and S. I. Sato. 1997. “Six Different Plastid Sub-types Were Found in the O. sativa-O. rufipogon Complex.” Rice Genetics Newsletter 15:80–82. Oka, H. I. 1958. “Interval Variation and Classification of Cultivated Rice.” Indian Journal of Genetics and Plant Breeding 18:79–89. ———. 1983. “Life-History Characteristics and Colonizing Success in Plants.” American Zoology 23:99–109. ———. 1988. Origin of Cultivated Rice. Tokyo: JSSP; Amsterdam: Elsevier. Oka, H. I., and W. T. Chang. 1962. “Rice Varieties Intermediate between Wild and Cultivated Forms and the Origin of the japonica Type.” Botanical Bulletin of Academia Sinica 3:109–31. Oka, H. I., and H. Morishima. 1982. “Ecological Genetics and the Evolution of Weeds.” In Biology and Ecology of Weeds, edited by W. Holzner and N. Numata, 73–89. The Hague: W. Junk. Randhawa, M. S. 1980. A History of Agriculture in India. Vol. 1. New Delhi: Indian Council for Agricultural Research. RRI (Rice Research Institute). 1992. Rice and Thai People [in Thai]. Bangkok: Department of Agriculture.

Rice of Asian Origin

275

Sano, R., and H. Morishima. 1992. “Indica-japonica Differentiation of Rice Cultivars Viewed from Variations in Key Characters and Isozymes, with Special Reference to Landraces from the Himalayan Hilly Areas.” Theoretical and Applied Genetics 84:266–74. Sano, Y., and R. Sano. 1990. “Variation of the Intergenic Spacer Region of the Ribosomal DNA in Cultivated and Wild Rice Species.” Genome 33:209–18. Sato, Y.-I. 1991. “Variation in Spikelet of the indica and japonica Rice Cultivars in Asian Origin.” [In Japanese with English summary.] Japanese Journal of Breeding 41:121–34. ———. 1996. DNA ga kataru inasaku bunmei: Kigen to tenkai. Tokyo: Nihon Hoso Shuppan Kyokai. ———. 2002. “Origins of Rice Cultivation in the Yangtze River Basin.” In The Origins of Pottery and Agriculture, edited by I. Yasuda, 143–50. New Delhi: Lustre Press and Roli Books. Sato, Y.-I., R. Ishikawa, and H. Morishima. 1990. “Nonrandom Association of Genes and Characters in indica and japonica in Rice.” Heredity 65:75–79. Second, G. 1981. “Origin of the Genetic Diversity of Cultivated Rice (Oryza spp.): Study of the Polymorphism Scored at 40 Isozyme Loci.” Japanese Journal of Genetics 57:25–57. Vavilov, N. I. 1926. “Studies on the Origin of Cultivated Plants.” Bulletin of Applied Botany 16:139–248. Wang, T. T. 1986. “Discussion on Origin and Spread of Agriculture in China.” [In Chinese.] Agricultural Archaeology 2:25–32. Wang, Z. Y., and S. D. Tanksley. 1989. “Restriction Fragment Length Polymorphism in Oryza sativa L.” Genome 32:1113–18. Watabe, T. 1977. Rice Road [in Japanese]. Tokyo: Nippon Housou Kyoukai. Yamanaka, S., Y. Fukuta, R. Ishikawa, I. Nakamura, T. Sato, and Y.-I. Sato. 2002. “Phylogenetic Origin of Waxy Rice Cultivars in Laos Based on Recent Observations for ‘Glutinous Rice Zones’ and dCAPS Markers of Waxy Gene.” Tropics 11:109–20. Yan, W. 1982. “The Origin of Rice Agriculture in China.” [In Chinese.] Agricultural Archaeology 1:2. ———. 2002. “The Origins of Rice Agriculture, Pottery and Cities.” In The Origins of Pottery and Agriculture, edited by Y. Yasuda, 151–56. Singapore: Lustre Press. Yano, M., Y. Katayose, M. Ashikari, U. Yamanouchi, L. Monna, T. Fuse, T. Baba, K. Yamamoto, Y. Umehara, Y. Nagamura, and T. Sasaki. 2000. “Hd1, a Major Photoperiod Sensitivity Quantitative Trait Locus in Rice, Is Closely Related to the Arabidopsis Flowering Time Gene CONSTANS.” Plant Cell 12:2473–84. You, X.-L. 1979. “On the Origin, Differentiation and Spread of the Cultivated Rice in China—An Investigation with the Carbonized Rice Found at the Hemudu Excavation.” Acta Agronomica Sinica 5:1–10.

276

genetics in archaeobotany

Zhang, Q. F., M. A. Saghai-Maroof, T. Y. Lu, and B. Z. Shen. 1992. “Genetic Diversity and Differentiation of indica and japonica Rice Detected by RFLP Analysis.” Theoretical and Applied Genetics 83:495–99. Zhao, Z., D. Pearsall, and Q. Jiang. 1995. “Analysis of the Phytoliths from Xian Ren Dong and Wang Dong.” In Origin of Rice Agriculture: The Preliminary Report of the Sino-American Jiangxi (PRC) Project, edited by R. S. MacNeish and J. G. Libby, 47–52. Publications in Anthropology, no.13, El Paso Centennial Museum. El Paso: University of Texas.

C h ap t e r 1 3

A Review of the Research on the Origin of Six-Row Barley Ken-ichi Tanno

Barley is one of the earliest cereals in the world, along with einkorn and emmer wheats. Its origin is thought to be in western Asia and dates back to the tenth millennium uncalibrated BP. Barley has been continuously cultivated for 10,000 years and ranks as the fourth cereal produced in the world today. This continuity is peculiar to barley and contrasts with einkorn and emmer wheats, which gradually diminished in importance and are fairly rare today. Barley did not undergo polyploidization as did bread wheat; therefore, we can say that the evolution of barley after domestication is based on hybridization and accumulation of spontaneous mutations. The geographical distribution of barley covers the temperate zones (eastward to Japan and Australia, and westward to Argentina and Chile), and barley has adapted to each local area and forms landraces. Because of the continual exchange of seeds and resulting introgression, it is difficult to know which type is truly original. For a long time, a simple question has puzzled barley researchers: is the progenitor of cultivated barley two-row or six-row (figure 13.1)? For the two-row cultivated barley, there is no doubt that it was domesticated from two-row wild barley, subsp. spontaneum. However, the origin of six-row cultivated barley is less clear, because six-row brittle barley such as f. agriocrithon or f. laguncliforme could be the candidate for the progenitor, as well as two-row cultivated barley (and, less possibly, subsp. spontaneum). Harlan (1995) concluded that the domestication of barley “originated in a limited area from a two-row ancestor, that three recessive mutants were of key importance and that, as a corollary and contrary to earlier

277

Figure 13.1. Two- (a) and six-row (b) barleys. A spikelet unit in two- (c) and six-row (d) barleys. Bottom view for six-row barley (e). Two-rowed barley has two sterile lateral flowers (c). In six-row barley, all three flowers are fertile and have well-developed awns on lateral spikelets (d). Because the spikelets are set alternately, six-kernel rows are distinctive in six-row barley (e).

Review of the Research on the Origin of Six-Row Barley 279 views, six-row ‘wild’ barleys (such as agriocrithon types) reflect introgression from cultivars.” Two-row cultivated barley is surely the most probable progenitor of six-row barley, but do we have enough evidence for a definite conclusion? After long debate on the row type of the progenitor, there is much roundabout evidence that implies that six-row cultivated barley is derived from two-row cultivated barley by mutation. However, we seem to have little direct evidence for this conclusion at present, and debate continues concerning the taxonomic status of six-row brittle barley f. agriocrithon.

Long-Term Debate The first researcher who raised the question regarding the row type of the progenitor of cultivated barley was De Candolle in 1883. In his important book Origin of Cultivated Plants, he suggested that both tworow and six-row cultivated barleys were derived from the wild two-row form subsp. spontaneum. There is a close resemblance between subsp. spontaneum and two-row cultivated barley except for some typical wild characteristics, such as the brittle rachis in subsp. spontaneum. But he also mentioned the possibility of an extinct wild six-row barley as the progenitor, because all the barley remains from ancient sites at that time were six-row. Kornicke and Werner (1885) also wrote that subsp. spontaneum is more likely to be the progenitor for two- and six-row cultivated barleys, but von Tschermak (1914) disputed the two-row hypothesis and proposed a six-row progenitor, saying that both intermediate and six-row forms would only appear if the six-row were the progenitor. Schiemann (1932) also had an opinion concerning the six-row progenitor hypothesis based on her detailed morphological observation with some genetic analysis. Her hypothesis was a bit extreme: all two- and six-row cultivated barley and two-row wild subsp. spontaneum were derived from an imaginary six-row wild barley. In 1938 Åberg found two barley kernels in a wheat sample from Taofu, China, collected by H. Smith, and obtained six-row brittle barley after growing the kernels (Åberg 1938). He insisted that this six-row brittle barley was the progenitor of six-row cultivated barley and called it Hordeum agriocrithon (Åberg 1940), and this conclusion was widely accepted at that time. Freisleben (1940), based on his exploration in the Hindu Kush, also advocated a “di-phyletic hypothesis” where six-row cultivated barley was derived from f. agriocrithon in the eastern part of Central Asia and then spread, invading the area of subsp. spontaneum; then two-row cultivated barley arose (table 13.1). Takahashi (1955) supported this hypothesis, because his genetic analysis of two different tough-rachis genes

280

genetics in archaeobotany

Table 13.1. Two-, six-, and di-phyletic progenitor hypothesis Two-row hypothesis

Diphyletic hypothesis

De Candolle (1883) Kornicke and Werner (1885)

Six-row hypothesis De Candolle (1883)

Freisleben (1940)

Takahashi (1955)

Tschermak (1914) Schiemann (1932) Åberg (1938) Åberg (1940) Nevski (1941) Parodi (1947) Covas (1950) Kamm (1954)

Helbaek (1959) Zohary (1959) Zohary (1960) Zohary (1964) Harlan and Zohary (1966) Helbaek (1960) Harlan (1968) Bothmer et al. (1989) Bothmer et al. (1995) Harlan (1995) Zohary and Hopf (2000) Tanno et al. (2002) Tanno and Takeda (2004)

Shao and Li (1987) Zhang et al. (1992) Zhang et al. (1994)

Yin et al. (2003)

showed clear geographical distribution (western and eastern types), which corresponded to the di-phyletic hypothesis. However, around 1950 evidence to support the two-row hypothesis began to increase. Helbaek (1953) found a two-row domestic type from the early agricultural site of Jarmo in Iraq. Since the six-row type was not found there, Helbaek (1959) concluded that the oldest cultivated barley was two-row, and that the six-row barley derived from two-row cultivated barley by mutation after agriculture was established in irrigated land in

Review of the Research on the Origin of Six-Row Barley 281 southern Mesopotamia. Six-row mutants had been obtained from tworow barley through radiation (Notzel 1952; Nybom 1954; Stubbe and Bandlow 1947), and Helbaek (1960) also got the six-row mutant from subsp. spontaneum by using X-rays. But there was no mutant that had changed from a six- to a two-row spikelet (thus far no two-row barley has been obtained by mutation from six-row barley). Zohary (1959, 1960) also supported the two-row progenitor hypothesis based on an experimental field observation using Israeli six-row brittle barley and its progeny. He insisted that the six-row brittle barleys were derived from natural hybrids, based on the fact that the progeny of sixrow brittle barley showed segregation in several morphological characteristics, suggesting high heterogeneity (Zohary 1964). Harlan and Zohary (1966) clarified the natural geographical distribution of subsp. spontaneum, and Harlan (1968) reviewed the origin of cultivated barley according to the archaeological evidence at the time. He pointed out that the appearance of two-row cultivated barleys in archaeological sites was older than that of the six-row forms, supporting the two-row progenitor hypothesis. The present consensus concerning the two-row progenitor hypothesis was established at that time. Since then, much archaeological data has accumulated (Zohary 1986; Zohary and Hopf 2000), and it appears that two-row cultivated barley is older than the six-row type. Before the appearance of two-row cultivated barley, hunter-gatherers had used subsp. spontaneum. At the present time, the earliest six-row barley is from Tell Abu Hureyra, Syria (8115 ±80 uncal BP—estimated from six-row barley seeds; Moore, Hillman, and Legge 2000); whereas the oldest two-row cultivated barley, based on a rachis fragment, is from Tell Aswad, Syria (9300 ±60 uncal. BP) (the date was taken from emmer grains of the same layer, and the grain sample was originally collected by Van Zeist; G. Willcox, pers. comm.). After much debate, the two-row progenitor hypothesis is widely accepted today. Six-row cultivated barley was derived from two-row (cultivated) barley by mutation, and two-row cultivated barley was domesticated from subsp. spontaneum. These conclusions are based on the following points: 1. The six-row spikelet characteristic is controlled by a single recessive gene at the vrs1 locus. Mutation of this gene always results in the appearance of six-row mutants from two-row original barley, and the reverse is unknown. 2. Six-row brittle barley is inferior to two-row wild barley in its ability to survive in nature. The six-row brittle barley does not have distinctive geographical distribution as seen in subsp. spontaneum, and has been found in artificially disturbed locations such as cultivated

282

genetics in archaeobotany

fields and roadsides. Consequently, it cannot be regarded as genuine wild barley. 3. Archaeobotanical evidence indicates that two-row cultivated barley appeared earlier than six-row cultivated barley. Before the appearance of the two-row cultivated barley, hunter-gatherers had used subsp. spontaneum. There is no evidence of the existence of six-row brittle barley from archaeological sites. Points 1 and 2 state that six-row brittle barley is unlikely to be a progenitor of cultivated barley, and they reject the possibility of the six-row progenitor hypothesis by a process of elimination. Only point 3 is direct evidence to support the two-row progenitor hypothesis, but the problem is that there is currently insufficient archaeological evidence. Even in western Asia, one of the best excavated regions of the world, more sites are needed for conclusive evidence concerning the domestication and initial evolution of barley, and little excavation has been done in Central Asia, where the origin of six-row brittle barley, f. agriocrithon, is suspected.

Further Discussion of Six-row Brittle Barley, f. agriocrithon The origin of f. agriocrithon has long been debated. Two contrary opinions were offered, both based on field study. Shao and Li (1987) pointed out that f. agriocrithon in Tibet matures a half-month earlier than sixrow cultivated barley. Because of this reproductive isolation, they decided that f. agriocrithon and six-row cultivated barley in Tibet were separate from each other and that f. agriocrithon is the wild progenitor for the six-row cultivated barley of this region. On the other hand, Bothmer, a taxonomist of the genus Hordeum, found a series of variations and high weediness in barley fields in Tibet, and concluded that f. agriocrithon is not a real wild barley but rather of hybrid origin (Bothmer, Yen, and Yang 1989; Bothmer et al. 1995). This kind of incongruity is also evident in biochemical and DNA marker analysis. Zhang et al. (1994) compared their results from esterase isozyme and rDNA analysis in East Asian barleys to previous reports from the Western world. They found some alleles distributed only in Tibet but never seen in the West, while some alleles found in the Western world cannot be seen in Tibetan barley. Based on this specific distribution, they supported Freisleben’s di-phyletic hypothesis. Yin, Ma, and Ding (2003) pointed out that the genetic diversity of Hordein (storage proteins of grain endosperm) is higher in the Tibetan population than in the Israeli and Jordanian populations. They insisted that Tibet is one of the centers for the origin of cultivated barley and that the immediate progenitor for

Review of the Research on the Origin of Six-Row Barley 283 Tibetan six-row cultivated barley is f. agriocrithon. On the other hand, Konishi (2001), supporting the two-row progenitor hypothesis, speculated that hybridization between Tibetan six-row cultivated barley and Afghan subsp. spontaneum produced f. agriocrithon in Tibet. His study is based on morphological and physiological analysis and esterase isozyme variations. Esterase loci (Est1, Est2, Est4) showed particular haplotypes (Est1-Est2-Est4: Ca-Un-Nz, Pr-Fr-At, and Pr-Fr-Su) in Tibetan f. agriocrithon, and the same haplotypes were also found in six-rowed cultivated barley in this area but not found in subsp. spontaneum. Tanno and Takeda (2004) estimated the nucleotide diversity of f. agriocrithon using a DNA marker closely linked (0.1cM) to the row-type gene (the vrs1 locus). The results indicated a considerably low diversity value in f. agriocrithon. This suggests that f. agriocrithon does not have genuine wild status but is rather hybrid in origin (details for this study appear later in this chapter).

Genetic Study of Six-Row Barley When six-row barley is crossed with two-row barley, the hybrid (F1 plant) has two-row spikelets; the next generation (F2 plants) segregates twoand six-row spikelets at a ratio of 3:1. This is what we call the Mendelian rule: the row type is regarded as under the control of a single major gene where the two-row characteristic is dominant and the six-row is recessive. Ubish (1916) was the first researcher to indicate that the row type characteristic adheres to the Mendelian rule. Thus, classical genetics began to focus on the row-type gene early in the last century. But it took time to apply this work to population genetics. In the 1980s biochemical marker analysis methods, including isozyme and allozyme, become available for population studies of barley, and its diversity and variation were argued. The advantage of using biochemical markers is that they are not affected by environmental or artificial selection. In the 1990s the DNA marker gradually replaced the biochemical marker, because it too is unaffected by environmental and artificial selection, and also produces more critical and clear-cut results, sometimes covering a whole genome. Using biochemical and DNA analysis, many studies showed distinct variation between two- and six-row cultivated barleys. But these results were often based on samples from limited areas and appear to reflect the different use of two- and six-row barleys. For example, in Europe, two-row barley is used for brewing while six-row barley is used for animal feed, and they have been separately cultivated. As a result, the two-row barleys are apt to be similar within their population, and the six-row barleys similar within theirs.

284

genetics in archaeobotany

However, when the analysis is based on worldwide samples, the difference between the two- and six-row cultivated barleys is more ambiguous. For example, RFLP (Saghai-Maroof, Zhang, and Biyashev 1995) studies based on a large number of worldwide accessions did not show a big difference between two- and six-row cultivated barleys in either allele composition or diversity. This similarity found in worldwide accessions is probably due to frequent hybridization between two- and six-row cultivated barleys. This type of introgression is extremely influential for changing allele composition. Since cultivated barley can easily hybridize with subsp. spontaneum, both two- and six-row cultivated barleys would have accumulated a number of variations from wild barley. The best way to determine the accurate origin for the two- and six-row cultivated barleys is to carry out the cloning of a related gene (i.e., row-type gene), because this process removes the influence of introgression.

An Experiment to Determine the Origin of Six-Row Barley To determine the origin of six-row barley, it would be preferable to carry out the cloning of the row-type gene (the vrs1 locus) and to count the number of mutations that cause six-row morphology. But at present cloning of the row-type gene has not been reported. (After this manuscript was submitted, the six-row gene was successfully cloned by Komatsuda et al. [2007]. The six-row gene had indeed multiple origins derived from independent mutations from two-rowed barleys.) An operation to clone a morphological gene entails high cost and takes time, but fortunately genome mapping studies have advanced considerably, and I have analyzed a linkage marker that is tightly linked to the row-type gene. Here I introduce a study by myself and my colleagues of modern barley accessions from Eurasia that focuses on whether six-row barley was a mutant origin derived from two-row barley by using a linkage marker tightly linked to the row-type gene (the vrs1 locus). This work was published by Tanno et al. (1999), Tanno et al. (2002), and Tanno and Takeda (2004). The closest known DNA marker to the vrs1 locus was cMWG699. This marker was produced by Graner et al. (1991) as a restriction fragment length polymorphism (RFLP) marker derived from cDNA. In their study, the cMWG699 did not show segregation with the row-type gene in ninety-two individuals of the F2 population of “Igri” × “Franka.” Komatsuda et al. (1999) modified this marker to be able to use it with the PCR method, and got a 0.1cM map distance from a cross-population of “Kanto nakate Gold” × “Azumamugi.” This distance means that even if hybridization took place between six-row and two-row barleys, the possibility of the

Review of the Research on the Origin of Six-Row Barley 285

Figure 13.2. (a) Recombination at meiosis. During recombination, homologous chromosomes break and reattach on different chromosomes to secure diversity in the progeny. For details about recombination, see Kinebuchi et al. (2004). (b) Linkage between the row-type gene and a gene locus called cMWG699 on chromosome 2H in barley. Because of their close linkage, there is little chance for a recombination to occur between the two loci (especially during the 10,000 years after domestication).

appearance of a recombinant between the row-type gene and cMWG699 is about 1/1,000 (figure 13.2). Taking into consideration two things—one being the highly self-fertile nature of barley and the other the appearance of a recombinant (with 1/1,000 probability) that has to survive under competition (i.e., genetic drift in evolutionary biology)—the probability of the recombinant seems to be negligibly low. If the six-row barley has derived from a single mutant, the cMWG699 marker may give a single type of sequence. There was no report that applied a linkage marker to this kind of phylogenetic study. The results based on the 823bp sequence are shown in figure 13.3. The wild two-row barley, subsp. spontaneum, showed nine different sequences out of ten accessions at the cMWG699 marker locus, therefore we can accept that subsp. spontaneum has considerably high genetic variation. Two-row cultivated barley does not match subsp. spontaneum in variation, but it still had high variations. On the other hand, six-row cultivated barley had only three types of sequences out of thirteen accessions (at first, only two sequences, type I and II, were known [Tanno et al. 1999; Tanno et al. 2002], but later type III was found [Tanno and Takeda 2004]). Out of the three sequences, types I and III differed only by one nucleotide substitution, and since type III was distributed only in East Asia, this type seemed to be a derivation from the type I sequence. Types I and II differed by four nucleotide substitutions. This difference is significant because if the divergence time for the intron region was estimated using a mutation rate at 5–30 × 10−9 (Wolfe, Li, and Sharp 1987), the date should be 138–830 thousand years. This divergence time is too early to conclude that the two sequences diverged after the emergence

286

genetics in archaeobotany

Figure 13.3. Nucleotide substitution site at cMWG699 marker locus that is closely linked to the six-rowed gene. In this figure, e.g., the 64th nucleotide is substituted from C (cytosine) to T (thymine) in the accession OUH707. Wild (subsp. spontaneum) and two-rowed cultivated barleys accumulate nucleotide substitutions while six-rowed cultivated and six-rowed brittle barleys show only small variation.

of the six-row characteristic (the oldest six-row barley remains were from about 8000 uncal BP, as mentioned above). Therefore this result suggests that there are two origins for six-row cultivated barley, assuming that the recombination had not happened between the cMWG699 and the rowtype gene. When we look at the distribution of type I and II sequences in sixrow cultivated barley, we see that type I spreads to a wide area over the

Review of the Research on the Origin of Six-Row Barley 287 temperate Eurasian zone (Japan, China [2], Afghanistan, Iraq, Ethiopia [2], Tunisia, Italy, and Finland), while type II was limited to near the Mediterranean region (Spain, Bulgaria). The same sequences were found in two-row barley from Turkmenistan (subsp. spontaneum OUH730) for type I, and from Morocco (subsp. spontaneum OUH777) and elsewhere in North Africa (cultivated barley named “Palmela Blue”) for type II. Thus, the origin of six-row cultivated barley may be not far from these regions. But we should take into consideration that the number of accessions analyzed is not large and DNA analysis based on modern plants always contains a risk that the distribution of the modern population differs from the past distribution. The existence of f. agriocrithon has aggravated the problem of the origin of cultivated barley. Is it really a “wild” barley? How much diversity does it have? If it has a long history as a true wild barley, it should have accumulated DNA variations like subsp. spontaneum. In other words, if f.  agriocrithon had large variations similar to subsp. spontaneum, that would indicate that it was a genuine wild barley, but if it showed low variation in the cMWG699, as in six-row cultivated barleys, this would suggest that it was hybrid in origin. The results of DNA sequence analysis demonstrated that only three sequences were found in forty-two accessions, and the three sequences were the same as those of the six-row cultivated barley. The nucleotide diversity values for f. agriocrithon were π = 0.0014 (Nei’s [1987] method), θ = 0.0014 (Watterson’s [1975] method), and therefore were about 2.8 and 4.5 times lower than those of subsp. spontaneum (π = 0.0040, θ = 0.0064). This means that f. agriocrithon did not have the same level of diversity as seen in the true wild species. Consequently, f. agriocrithon does not appear to represent genuinely wild populations but more probably originated from hybridization between subsp. spontaneum and six-row cultivated barley. In the results of these studies, the DNA marker cMWG699, which is closely linked to the row-type gene, indicated little variation in six-row barley in comparison to two-row barley. This evidence agrees with the two-row progenitor hypothesis whereby six-row barley is considered to be derived from two-row barley by mutation. The weaker evidence obtained from this study was that the six-row barley might be derived from two independent mutation events. But we could not eliminate the possibility of misjudgment that comes from past recombination events between the vrs1 and the cMWG699. Especially for the type II six-row barleys, reconsideration of an independent origin or a recombination event may be needed, because it has a limited distribution where there is no archaeological evidence. This point will perhaps be solved in the future when the row-type gene is cloned. (Later analysis by Komatsuda et al. [2007] proved the multiple origins of six-row genes.) If

288

genetics in archaeobotany

six-row barley were derived from two independent mutants, it would be a quite rare phenomenon. But there are some examples for multiple origins for a single gene for agronomic characteristics. For example, the low and less amylose grain characteristic has five origins in foxtail millet based on a DNA (RFLP) analysis of a related gene (waxy gene) (Fukunaga, Kawase, and Kato 2002). Since this characteristic gives “sticky texture” to boiled or steamed grains, this type of cereal is preferred, especially in East Asia, and has been targeted for selection. In cultivated barley, the most important domestication characteristic, nonshattering (nonbrittle rachis), has two origins based on a cross allelism test of hybrids (Takahashi 1955). Alternatively, if six-row barley does not have a dual origin but the results of cMWG699 were from recombination events, this is still important information for further study of barley evolution, because it implies that current genetic studies based on DNA markers, like AFLP and RAPD methods, may no longer be effective for the study of diversification in barley.

Molecular Phylogeny and Archaeobotany Molecular phylogeny based on DNA (and amino acids as well) provides strong evidence for answering the question “how many origins?” when the cloning of related gene(s) (such as a nonshattering gene) was carried out. However, if we want to know when and where, molecular phylogeny based on modern plants may not always tell us the truth, because the distribution of modern plants is an outcome of migration according to climate changes and human impact. To resolve the problem of when and where, evidence from archaeobotany is necessary. At the present time, the number of excavated archaeological sites is not enough, and archaeobotany is less advanced than DNA-based study. Development of archaeobotany is highly necessary. Both genetics and archaeology should work in cooperation to elucidate crop origins.

Acknowledgments I would like to thank Elizabeth Willcox for help with the English. My thanks also go to all my coworkers, especially T. Komatsuda (NIAS, Japan) and K. Takeda (Okayama University, Japan), on DNA-based studies using the linkage marker of the row-type gene. I am also grateful to the Japan Society for the Promotion of Science for its grant (JSPS Postdoctoral Fellowships for Research Abroad).

Review of the Research on the Origin of Six-Row Barley 289 References Åberg, E. 1938. “Hordeum agriocrithon nova sp., a Wild Six-Rowed Barley.” Annals of the Royal Agricultural College of Sweden 6:159–212. ———. 1940. “The Taxonomy and Phylogeny of Hordeum L. Sect. Cerealia Ands.” Symbolae Botanicae Upsalienses 4:1–156. Bothmer, R. von, C. Yen, and J. Yang. 1989. “Does Wild Six-Rowed Barley, Hordeum agriocrithon, Really Exist?” FAO/IBPGR Plant Genetic Resources Newsletter 77:17–19. Bothmer, R. von, N. Jacobsen, C. Baden, R. B. Jørgensen, and I. Linde-Laursen, eds. 1995. An Ecogeological Study of the Genus Hordeum. 2nd ed. Systematic and Ecogeographic Studies on Crop Genepools, 7. Rome: International Plant Genetic Resources Institute. Covas, G. 1950. “El prototipo de las cebadas cultivadas.” Ciencia e Investigación 6:236–37. De Candolle, A. (1883) 1959. Origin of Cultivated Plants. New York: Noble Offset Printers. Freisleben, R. 1940. “Die phylogenetische Bedeutung asiatisher Gersten.” Züchter 12:257–72. Fukunaga, K., M. Kawase, and K. Kato. 2002. “Structural Variation in the Waxy Gene and Differentiation in Foxtail Millet (Setaria italica (L.) P. Beauv): Implications for Multiple Origins of Waxy Phenotype.” Molecular Genetics and Genomics 268:214–22. Graner, A., A. Jahoor, J. Schondelmaier, H. Siedler, K. Pillen, G. Fischbeck, G. Wenzel, and R. G. Herrmann. 1991. “Construction of an RFLP Map of Barley.” Theoretical and Applied Genetics 83:250–56. Harlan, J. R. 1968. “On the Origin of Barley.” In Barley: Origin, Botany, Culture, Winterhardiness, Genetics, Utilization, Pests, 9–31. Agricultural Handbook, no. 338. Washington, DC: U.S. Department of Agriculture. ———. 1995. “Barley.” In Evolution of Crop Plants, edited by J. Smartt and N. W. Simmonds, 140–47. London: Longman. Harlan, J. R., and D. Zohary. 1966. “Distribution of Wild Wheats and Barley.” Science 153:1074–80. Helbaek, H. 1953. Archaeology and Agricultural Botany. Ninth Annual Report of the University of London, Institute of Archaeology, 45–59. ———. 1959. “Domestication of Food Plants in the Old World.” Science 130:365–72. ———. 1960. “Ecological Effects of Irrigation in Ancient Mesopotamia.” Iraq 22:186–96. Kamm, A. 1954. “The Discovery of Wild Six-Rowed Barley and Wild Hordeum intermedium in Israel.” Annals of the Royal Agricultural College of Sweden 21:287–320.

290

genetics in archaeobotany

Kinebuchi, T., W. Kagawa, R. Enomoto, K. Tanaka, K. Miyagawa, T. Shibata, H. Kurumizaka, and S. Yokoyama. 2004. “Structural Basis for Octameric Ring Formation and DNA Interaction of the Human Homologous-Pairing Protein DMC1.” Molecular Cell 14:363–74. Komatsuda, T., W. Li, F. Takaiwa, and S. Oka. 1999. “High Resolution Map around the vrs1 Locus Controlling Two- and Six-Rowed Spike in Barley, Hordeum vulgare.” Genome 42:248–53. Komatsuda, T., M. Pourkheirandish, C. He, P. Azhaguvel, H. Kanamori, D. Perovic, N. Stein, A. Graner, T. Wicker, A. Tagiri, U. Lundqvist, T. Fujimura, M. Matsuoka, T. Matsumoto, and M. Yano. 2007. “Six-rowed Barley Originated from a Mutation in a Homeodomain-­leucine Zipper I-class Homeobox Gene.” PNAS 104:1424–1429. Konishi, T. 2001. “Genetic Diversity in Hordeum agriocrithon E. Åberg, SixRowed Barley with Brittle Rachis, from Tibet.” Genetic Resources and Crop Evolution 48:27–34. Kornicke, F., and H. Werner. 1885. Handbuch de Getreidebaues. Berlin: Verlag von Paul Parey. Moore, A. M. T., G. C. Hillman, and A. J. Legge. 2000. Village on the Euphrates: From Foraging to Farming at Abu Hureyra. New York: Oxford University Press. Nei, M. 1987. Molecular Evolutionary Genetics. New York: Columbia University Press. Nevski, S. A. 1941. “Materialien zur der Kenntnis der wildwachsenden Gersten im Zusammenhang mit der Frage über den Ursprung von Hordeum vulgare L. und Hordeum distischum L. (Versch einer Monographie der Gattung Hordeum L.).” [In Russian with German summary.] Acta Instituto Botanico Nomine V. L. Komarovii, Academy of Science U.S.S.R. 1 (5): 64–255. Notzel, H. 1952. “Genetische Untersuchungen an rontgeninduzierten Gerstenmutanten.” Kuhn Archaeologische 66:72–132. Nybom, N. 1954. “Mutation Types in Barley.” Acta Agricola Scandinavia 4:430–56. Parodi, L. R. 1947. “ La especie progenitora de las cebadas domésticas.” Ciencia e Investigación 3:267–69. Saghai-Maroof, M. A., Q. Zhang, and R. Biyashev. 1995. “Comparison of Restriction Fragment Length Polymorphisms in Wild and Cultivated Barley.” Genome 38:298–306. Schiemann, E. 1932. “Entstehung der Kulturpflanzen.” Handbuch der Vererbungswissenschaft 3:161–74. Shao, Q., and A. Li. 1987. “Unity of Genetic Population for Wild Barley and Cultivated Barley in Himalaya Area.” In Barley Genetics V, edited by S. Yasuda and T. Konishi, 35–41. Okayama, Japan: Sanyo Press. Stubbe, H., and G. Bandlow. 1947. “Mutationsversuche an Kulturpflanzen I Rontgenbestrahlung von Winter-und-Sommergerstein.” Zuchter 17/18:365–74. Takahashi, R. 1955. “The Origin and Evolution of Cultivated Barley.” Advances in Genetics 7:227–66.

Review of the Research on the Origin of Six-Row Barley 291 Tanno, K., F. Takaiwa, S. Oka, and T. Komatsuda. 1999. “A Nucleotide Sequence Linked to the vrs1 Locus for Studies of Differentiation in Cultivated Barley (Hordeum vulgare L.).” Hereditas 130:77–82. Tanno, K., and K. Takeda. 2004. “On the Origin of Six-Rowed Barley with Brittle Rachis, Agriocrithon (Hordeum vulgare ssp. vulgare f. agriocrithon [Åberg] Bowd.), Based on a DNA Marker Closely Linked to the vrs1 (Six-Row Gene) Locus.” Theoretical and Applied Genetics 109:145–50. Tanno, K., S. Taketa, K. Takeda, and T. Komatsuda. 2002. “A DNA Marker Closely Linked to the vrs1 Locus (Row Type Gene) Indicates Multiple Origins of Six-Rowed Cultivated Barley (Hordeum vulgare L.).” Theoretical and Applied Genetics 104:54–60. Tschermak, E. von. 1914. “Die Verwertung der Bastardierung für phylogenetische Fragen in der Getreidegruppe.” Zeitschrift für Pflanzenzücht 2:291–312. Ubish, G. V. 1916. “Beitrag zu einer Faktorenanalyse von Gerste.” Zeitsschrift für Induktive Abstammungs- und Vererbungslehre 17:120–52. Watterson, G. A. 1975. “On the Number of Segregating Sites in Genetical Models without Recombination.” Theoretical Population Biology 7:256–76. Wolfe, K. H., W. H. Li, and P. M. Sharp. 1987. “Rate of Nucleotide Substitution Varies Greatly among Plant Mitochondrial, Chloroplast and Nuclear DNAs.” Proceedings of the National Academy of Science 84:9054–58. Yin, Y. Q., D. Q. Ma, and Y. Ding. 2003. “Analysis of Genetic Diversity of Hordein in Wild Close Relatives of Barley from Tibet.” Theoretical and Applied Genetics 107:837–42. Zhang, Q., M. A. Saghai Maroof, and G. Yang. 1992. “Ribosomal DNA Polymorphism and Oriental-Occidental Genetic Differentiation in Cultivated Barley.” Theoretical and Applied Genetics 84:682–87. Zhang, Q., G. P. Yang, X. Dai, and J. Z. Sun. 1994. “A Comparative Analysis of Genetic Polymorphisms in Wild and Cultivated Barley from Tibet Using Isozyme and Ribosomal DNA Markers.” Genome 37:631–38. Zohary, D. 1959. “Is Hordeum agriocrithon Åberg the Ancestor of Six-Rowed Cultivated Barley?” Evolution 13:279–80. ———. 1960. “Studies on the Origin of Cultivated Barley.” Bulletin Research Council of Israel 9D:21–42. ———. 1964. “Spontaneous Brittle Six-Rowed Barley, Their Nature and Origin.” Barley Genetics I: Proceedings of the 1st International Barley Genetics Symposium, Wageningen, 26-31 August 1963, 27–31. Wageningen: Centre for Agricultural Publications and Documentation. ———. 1986. “The Origin and Early Spread of Agriculture in the Old World.” In The Origin and Domestication of Cultivated Plants, edited by C. Barigozzi, 3–20. Amsterdam: Elsevier. Zohary, D., and M. Hopf. 2000. Domestication of Plants in the Old World: The Origin and Spread of Cultivated Plants in West Asia, Europe, and the Nile Valley. Oxford: Clarendon Press.

C h ap t e r 1 4

Maize Cob Phytoliths as Indicators of Genetics and Environmental Conditions L i n d a S c o t t C u mm i n gs

Maize (Zea mays L.) cobs carry within them signatures of both their genetics and the environmental conditions under which they grew. Since cobs represent part of the reproductive structure of maize, their cells can be expected to be modified when humans select for certain characteristics to make maize a more attractive or productive crop, such as soft glumes. For instance, when people selected maize that had soft glumes, which are the papery elements that partially cover the kernels—and get stuck in your teeth when you eat corn on the cob—rather than hard glumes of teosinte, they selected cobs with different physical characteristics that were passed on to future generations (Doebley 1990, 1992; Fedoroff 2003, 1158–59; Jaenicke-Després et al. 2003). Hard glumes are more silicified, and also cobs are very rigid structures that have adapted by silicifying many of their cells, providing us with an excellent phytolith record of the shape or form of individual cells. Therefore, phytoliths recovered from maize cobs are expected to reflect the genetics of maize, since they are silica casts of the inside of the cells in cobs and glumes (Piperno 2006, 49; Piperno and Pearsall 1998, 158–63). Maize requires a certain amount of water to produce ears that are filled with kernels, and thus attractive to humans. From the maize plant’s perspective, the most critical time to receive moisture is during the two weeks before and the two weeks following silking or tasseling. Silking, in much of the temperate American Southwest, including southwestern and

292

Maize Cob Phytoliths as Indicators of Genetics

293

western Colorado, happens approximately mid-August (Adams, Muenchrath, and Schwindt 1999, 492). Therefore, we will consider August precipitation to be the most important to maize and, hence, human agriculturalists growing maize. In addition, the soil must be moist enough in the late spring for maize to germinate. The importance of both quantity and timing of precipitation is expected to be reflected in cells of maize cobs, since Adams, Muenchrath, and Schwindt (1999) report that sizes of maize cobs vary with variations in precipitation. Archaeologists working in areas with maize agriculture recover maize cobs in a variety of sizes. Very small cobs often contain eight or twelve or even more rows of kernels, which is similar to the number of rows on larger cobs. Cobs don’t modify their number of rows with their numeric diameter; they modify the size of the cupules and kernels according to overall cob size. Therefore, one might expect that the size of individual cells also varies in maize cobs of different sizes. Certainly, if this is true, it should not be necessary to digest maize cobs to recover and measure microscopic phytoliths only to report that one maize cob was larger than another. However, when we have either fragmentary or whole cobs, the information that we retrieve from digesting the cobs, and recovering and measuring the phytoliths, includes both genetics and environmental conditions, which allows us to distinguish between two sizes of cobs either as coming from different races or as representing a single race of maize grown under different environmental conditions. This is not possible by visual inspection of the cobs alone. Examining and measuring the phytoliths allows for the possibility that more than one race of maize was grown at a site, and phytolith analysis would therefore be able to identify the presence of multiple races, provided the sampling population was large enough to detect all potential races present and perform appropriate statistical analysis. Different environmental conditions might be as subtle as reflecting differences across a single field, one edge of which received more water through irrigation than the other. Or it might reflect the strategy of planting maize in more than one location during a single year so that one, or perhaps both, of the crops might reach maturity and provide food. Alternatively, recovering cobs of different sizes might reflect the strategy of storing maize for more than one year against the possibility of future crop failure. One other possibility must be mentioned: it is possible that some of the small ears are tiller ears and should be considered anomalous. Tiller ears are only a small portion of any potential harvest. The reference population selected for testing consists of phytoliths recovered from an experimental population of maize grown in New Mexico during the summers of 1992 and 1993 by Deb Muenchrath (1995). Both years received essentially the same amount of precipitation during the

294

genetics in archaeobotany

Table 14.1. Watering regimes for Tohono O’odham maize experimental garden Total precipitation

1992

1993

973 mm

964 mm

Mean daily air temp.

22.4ºC

23.2ºC

Cumulative growing degree days

1,462

1,515

Precipitation received during critical flowering stage

78%

20%

Yield (total 588 ears)

251

337

Mean Row Number

10

10

Row Numbers

6, 8, 10, 12, 14

6, 8, 10, 12, 14

summer: 973 mm in 1992 and 964 mm in 1993 (table 14.1). In addition, mean daily air temperature was very similar: 22.4ºC in 1992 and 23.2ºC in 1993. Even cumulative growing degree days were similar: 1,462 in 1992 and 1,515 in 1993. However, timing of moisture during the two years was very different. In 1992, 78 percent of the total precipitation was received during the critical flowering stage. In contrast, during 1993, only 20 percent of the summer rain occurred during flowering. Moisture is critical at the time of pollination, silking, and grain fill, meaning that the 1993 growing season was a poor one for maize, while the 1992 season was excellent. Identification of the significance of the timing of precipitation, compared to total growing season precipitation, “raises concerns about modelling dryland maize productivity based on the tree ring record of reconstructed precipitation,” according to Adams, Muenchrath, and Schwindt (1999, 492). Since conifer tree ring growth reflects precipitation both prior to and early in the growing season that is not available to maize, current tree ring reconstructions are not a good proxy record for growing conditions that would have produced a good maize crop. Therefore, recovering and measuring phytoliths from maize cobs provides a unique opportunity to examine growing conditions directly relevant to maize production. In fact, the Adams, Muenchrath, and Schwindt (1999, 492) study “shows that, even for 2 years with identical total growing season precipitation, the timing and amounts of individual rain events influenced total maize grain yield, especially for maize receiving no supplemental moisture.” Note that 1992, with more moisture available at the critical time of year, produced fewer cobs than 1993, which was more stressful. Cobs

Maize Cob Phytoliths as Indicators of Genetics

295

produced in 1993 were generally smaller than those from 1992, as were the kernels. In fact, measurements of cobs lead Adams, Muenchrath, and Schwindt (1999, 491) to conclude that “environmental moisture factors influence the size of structures responsible for supporting the kernels,” which means that the size of phytoliths or silica casts of the cells in these supporting structures should vary with environmental conditions.

Experimental Maize Population Eight of the experimental Tohono O’odham maize cobs grown in New Mexico were selected for analysis. Although this study started as a blind test, it was too large to complete without funding. Therefore, after examining twelve cobs, the blind test was canceled, and I selected four cobs from the 1992 and four cobs from the 1993 season that were examined to compare them statistically. The 1992 cobs representing maize watered every two weeks and every six weeks and 1993 cobs representing maize watered every two weeks and every four weeks were selected. This provides data from cobs representing the more extreme conditions for our statistical verification. Using phytoliths from these eight cobs, ANOVA statistical tests were used to compare sizes of phytoliths (table 14.2). Fifty phytoliths were measured for size and shape on both the upper and lower faces of each sample. When measuring phytoliths, we generate both size and shape Table 14.2. Experimental Tohono O’odham maize population (N = 50) Size

p-value

Shape

Area

.00025

Form factor

.215

Perimeter

.001

Compactness

.542

Convex perimeter

.001

Extent

.842

Length

.002

Convexity

.307

Breadth

.003

Aspect ratio

.373

Fiber length

.002

Elongation

.392

Width

.004

Curl

.455

Equivalent diameter

.001

Roundness

.575

Inscribed radius

.011

Solidity

.360

p-value

Note: Null Hypothesis = All cobs represent a single population. Reject Null Hypothesis at p < 0.02. All Values for Size REJECT Null Hypothesis. All Values for Shape ACCEPT Null Hypothesis.

296

genetics in archaeobotany

Table 14.3. Definitions for phytolith size categories Size Category

Definition

 Area

Measurement of interior bounded space

 Convex Area

Taut string area (this fills in indentations and irregularities)

 Perimeter

Total measurement around outside of phytolith

 Convex Perimeter Taut string perimeter (fills indentations)  Length

Longest chord distance on phytolith (largest difference between minimum and maximum coordinates at any angle of rotation or maximum Feret’s diameter)

 Breadth

Shortest chord distance on phytolith (smallest difference between minimum and maximum coordinates at any angle of rotation or minimum Feret’s diameter)

  Fiber length

Longest continuous fiber that can be drawn in the phytolith or (.5*perimeter)(area/perimeter)

 Width

The width of an ellipse of the same area and length as those measured for the phytolith (4*area)/(p*length); this is a derived parameter

Inscribed radius

Radius of the maximum circle that can be inscribed in the phytolith

Equivalent Diameter Diameter of a circle with same area as the phytolith Note: Terms are defined using text in Russ (1990, 182–203).

data. Sizes are hypothesized to reflect environmental conditions under which the maize grew. Nine size categories were measured, which are defined in table 14.3. In this preliminary study, Area appears to be the most sensitive measure and Inscribed Radius the least sensitive (table 14.2). Phytolith size variability (noted on the left side of this table), reflecting environmental conditions, is significant at the .004 to .00025 level, indicating two distinct size ranges of cells—representing maize grown in good environmental conditions and maize grown in adverse environmental conditions, and rejecting the Null Hypothesis that these sizes represent a single, unvaried population of maize. The second type of measurement that is generated reflects nine shape parameters (table 14.4). Many of these shapes are not intuitive. At this

Maize Cob Phytoliths as Indicators of Genetics

297

Table 14.4. Definitions for shape categories Shape Category

Definition

Angularity Elongation

Form factor

Ratio of area to perimeter (4 p Area/Perimeter2)

Roundness

Ratio of area to length (longest chord) (4 Area/p Length2)

Convexity

Ratio of taut string perimeter to external perimeter

Solidity

Ratio of area to convex or taut string area

somewhat

Extent

Ratio of area to bounding parallelogram

X

Compactness

Ratio of the maximum moment of inertia to that of a circle of the same area (moments of inertia; see Exner and Hougardy 1988)

X

Aspect ratio

Ratio of measured length to either breadth or width

X

Elongation

Square root of the ratio of maximum to minimum moment of inertia

X

Ratio of measured length to fiber length (objects that are straight yield a number approaching 1.0)

X

Curl

X X X

somewhat

Note: Terms are defined using text in Russ (1990, 182–203).

early stage, Extent, Roundness, and Compactness appear to be most sensitive, measuring both angularity and elongation or roundness of the phytoliths. Examining the statistics for shape parameters on the right side of table 14.2 indicates that shape varies very little within the experimental population. For all shape parameters, the value of p is greater than .2 and is noted as high as .84, indicating that the Null Hypothesis of shapes representing a single population cannot be rejected—a double negative—which is consistent with the fact that a single race of maize was planted in the experimental plots. This supports the hypothesis that

298

genetics in archaeobotany

phytolith shapes are, indeed, an excellent proxy for genetics, such as in identifying race.

Archaeological Maize Examination of maize from the archaeological record has been restricted to several cobs from Fremont occupations in northwestern Colorado, a single cob from southwestern Utah, and a single cob from West Texas. The Fremont cobs from the Kuck Shelter in northwestern Colorado represent occupation about AD 930 (table 14.5). We first examined two cobs from two separate features at the site that yielded very similar shape measurements and different size measurements, indicating that although they represented maize of the same race, these cobs represent crops grown under different environmental conditions, suggesting either storage of crops from different years or perhaps use of different locations for fields as an adaptive strategy to maximize the possibility of successful harvest. Next we measured phytoliths from a cob recovered at a nearby site. These phytoliths were similar in size to the cob that appeared to be grown in good conditions. Comparing the archaeological populations with the modern experimental population provides a large enough number of measurements for statistical analysis (table 14.6). Once again, the Null Hypothesis can be Table 14.5. Fremont maize cob phytolith measurements, in μm (N = 50) KUCK SITE, 5RB3157

5RB705

SIZE

179 Upper

46 Upper

2 Upper

Area

8.0886

12.0392

12.5

Convex area

8.529

12.2242

12.651

Perimeter

11.7498

13.6934

14.7382

Convex perimeter

10.9554

12.9322

13.96

Length

4.313

5.1032

5.679

Breadth

2.5372

2.9972

2.9652

Fiber length

4.6466

5.3066

5.813

Width

2.2126

2.7712

2.72

Equivalent diameter

3.049

3.6946

3.884

Inscribed radius

1.1778

1.4684

1.4528

Maize Cob Phytoliths as Indicators of Genetics

299

Table 14.6. Comparison of modern Tohono O’odham with Fremont, Kayenta Anasazi, and West Texas maize Shape

p-value

Fremont

Compactness

.004

Kayenta Anasazi

Compactness

.000012

West Texas

Compactness

.000004

Note: Null Hypothesis = All cobs represent a single population. Reject Null Hypothesis at p < 0.02.

rejected at p = 0.02. Our data for shape indicates that the modern Tohono O’odham and the Fremont maize are not the same genetic population, as one might expect. In addition, modern Tohono O’odham and Kayenta Anasazi are statistically distinct. Although the sample is not yet large enough, it appears that Fremont and Kayenta Anasazi maize also are statistically distinct from one another. We don’t yet have data for Hohokam or Mesa Verde Anasazi with which to compare the Fremont, Kayenta Anasazi, or West Texas archaeological maize. Our data for size appear to reflect different growing conditions for different cobs of the same race (table 14.5). Although the population size is still small, the trend emerging appears to indicate that many cobs will sort out into size ranges indicating receipt of sufficient moisture during the critical part of the growing season or failure to receive sufficient moisture during that time, much like an “on” or “off” switch, rather than a continuum. It will be interesting to observe whether size tends to vary differently in different races of maize, or whether phytolith size appears to be an independent variable reacting solely to moisture and growing conditions, irrespective of race. In addition, initial observations of p-values (table 14.6) suggest that once populations are large enough, we might be able to assign a relative measure of “genetic distance” to archaeological populations of maize by examining shapes of phytoliths. Further testing will be very helpful in support or contradiction of this statement. Note the difference between the extremely small p-values when comparing modern Tohono O’odham and Kayenta Anasazi or Tohono O’odham and West Texas maize, then note the p-value comparing Tohono O’odham and Fremont maize, which is small enough to reject the Null Hypothesis at less than 0.02, and indeed, indicates an expected error in only four cases per thousand, indicating two distinct genetic populations. At this time, however, sample size is too small for any definitive statements of “genetic distance” to be made.

300

genetics in archaeobotany

Conclusions Based on the ethnographic concept that different families, lineages, or tribes own seed that is planted, harvested, saved, and passed along within that family or lineage (Whiting 1939, 67–69), phytoliths (which appear in this study to be a good proxy for maize genetics) have the potential to provide much cultural information. This information can tell us about relationships of people (perhaps representing lineages) within a site, relationships of people and populations at one site to another, or movement of people across the landscape, settlement patterns, and trade.

References Adams, K. R., D. A. Muenchrath, and D. M. Schwindt. 1999. “Moisture Effects on the Morphology of Ears, Cobs and Kernels of a South-Western U.S. Maize (Zea mays L.) Cultivar, and Implications for the Interpretation of Archaeological Maize.” Journal of Archaeological Science 26:483–96. Doebley, J. 1990. “Molecular Evidence and the Evolution of Maize.” Economic Botany 44 (3, suppl.): 6–27. ———. 1992. “Mapping the Genes That Made Maize.” TIG 8:302–7. Exner, H. E., and H. P. Hougardy, eds. 1998. Quantitative Image Analysis of Microstructures: A Practical Guide to Techniques, Instrumentation, and Assessment of Materials. Oberursel: DGM Informationsgesellschaft. Fedoroff, N. 2003. “Prehistoric GM Corn.” Science 302:1158–59. Jaenicke-Després, V., E. S. Buckler, B. D. Smith, M. Thomas, P. Gilbert, A. Cooper, J. Doebley, and S. Päabo. 2003. “Early Allelic Selection in Maize as Revealed by Ancient DNA.” Science 302:1206–8. Muenchrath, D. A. 1995. “Productivity, Morphology, Phenology, and Physiology of a Desert-Adapted Native American Maize (Zea mays L.) Cultivar.” PhD diss., Iowa State University. Dissertation Abstracts 95-40927. Piperno, D. R. 2006. Phytoliths: A Comprehensive Guide for Archaeologists and Paleoecologists. Oxford: Altamira Press. Piperno, D. R., and D. M. Pearsall. 1998. The Origins of Agriculture in the Lowland Neotropics. New York: Academic Press. Russ, J. C. 1990. Computer-Assisted Microscopy: The Measurement and Analysis of Images. New York: Plenum Press. Whiting, A. F. 1939. Ethnobotany of the Hopi. Museum of Northern Arizona Bulletin, no. 15. Flagstaff: Northern Arizona Society of Science and Art.

E d i tors a nd Contr ibu tors Editors Marco Madella has a first degree in natural sciences (botany) from the Universitá Statale di Milano and a PhD in environmental archaeology from the University of Cambridge. He has been an affiliated lecturer at the University of Cambridge and director of studies in archaeology and anthropology at St. Edmund’s College. He currently is ICREA research professor at the Spanish National Research Council (CSIC) in Barcelona. He has been a visiting fellow at the Universidad Nacional de Mar del Plata (Argentina), Universidade de São Paulo (ESALQ—Brazil), Universitá Statales di Firenze (Italy), and the Research Institute for Humanities and Nature (Japan). Marco Madella’s research interests are in archaeobotany, social modeling, and paleoecology. Specifically, he is interested in plant paleoeconomy and the human-environment relationship both in hunter-gatherer and agricultural societies. He has been working on several international projects in South Asia, the Near East, South America, and Africa. Having completed a BA in archaeology at the University of Bologna and an MSc in archaeological sciences at the University of Milan in Italy, Carla Lancelotti obtained her PhD degree in archaeobotany at the University of Cambridge in 2010. She is interested in human-environment interactions, especially regarding the exploitation and use of nonfood plant resources. She applies archaeobotanical, geochemical, and ethnographic methods to the understanding of socioecological dynamics in prehistory. She is especially interested in South Asia but has also worked in European contexts. She is currently a postdoctoral researcher at the Universitat Pompeu Fabra in Barcelona. After completing a BSc and an MSc in geography at the Université du Québec à Montréal, and a DEA in environmental archaeology at the Université de Paris I Panthéon-Sorbonne, Manon Savard obtained a PhD in archaeology, specializing in archaeobotany, at the University of Cambridge. She is now a professor of geography and archaeology at the Université du Québec à Rimouski, where she is a founding member of a

301

302

Editors and Contributors

laboratory of archaeology and heritage. She has extensive fieldwork experience in the Near East, the Balkans, and the Mediterranean, as well as in France and Québec. Her research interests include hunter-gatherer subsistence strategies, the origins and diffusion of agriculture, past human response to environmental change, and human impact on the environment. More recently, she developed archaeological projects in Quebec, including a field school, along with an interest in public archaeology and heritage.

Contributors Paul De Paepe is a retired soil scientist formerly at the Department of Geology and Soil Science, University of Ghent, in Belgium. Dorian Q Fuller is a professor of archaeobotany at the Institute of Archaeology, University College London in London, United Kingdom, an archaeobotanist, and an archaeologist working on the origins of agriculture and the spread of agriculture and particular crops, as well as the later intensification of agriculture. Maria Dulce Gaspar is a professor in the doctoral and master’s programs in archaeology at the Department of Anthropology, National Museum, Federal University of Rio de Janeiro, in Brazil, where she works on dynamics of human settlement of the Brazilian territory. Kristen J. Gremillion is an archaeologist and palaeoethnobotanist in the Department of Anthropology at The Ohio State University in Columbus, Ohio. She studies the evolution of human diet and subsistence practices in ecological context. Ernie Haerinck is a retired professor in the Department of Near Eastern Archaeology, University of Ghent, in Belgium. He has worked on several archaeological projects in Central Asia and the Emirates. Martin K. Jones is the George Pitt-Rivers Professor of Archaeological Science in the Department of Archaeology at the University of Cambridge in the United Kingdom. He works on archaeobotany and archaeogenetics in the context of the broader archaeology of food. Gyoung-Ah Lee is a professor in the Department of Anthropology, University of Oregon, in Eugene, Oregon. Her research focuses on human-­ environmental interactions, social complexities, transitions to

Editors and Contributors

303

agriculture, ethnography of traditional farming, crop origins, and quantitative archaeology. Danièle Martinoli is Science Officer at the Swiss Biodiversity Forum, Swiss Academy of Sciences, in Bern. She was working until 2007 at the Institute for Prehistory and Archaeological Sciences (IPAS) at the University of Basel in Switzerland, where she investigated patterns of use of plant foods among hunter-gatherers and the origins and mechanisms of the spread of agriculture in southwest Asia. Meriel McClatchie is a post-doctoral research fellow at the Department of Archaeology and Palaeoecology, School of Geography, Queen’s University, Belfast, in Ireland. She undertakes research into plant use by both farming and nonfarming societies, including the social implications of these activities. Naomi F. Miller is a consulting scholar at the University of Pennsylvania Museum of Archaeology and Anthropology in Philadelphia and conducts archaeobotanical research primarily in west and central Asia. Simone Riehl is a lecturer in environmental archaeology at the Institute for Archaeological Sciences and senior researcher at the Senckenberg Center of Human Evolution and Palaeoenvironment at the University of Tübingen in Tübingen, Germany. Her interests include palaeoethnobotany, the emergence and development of agriculture, and palaeoecology and environmental archaeology of the Middle East. Katrien Rutten is a pottery specialist who worked at the Department of Near Eastern Archaeology, University of Ghent, in Belgium. Yo-Ichiro Sato is a researcher at the Research Institute for Humanities and Nature (RIHN) in Kyoto, Japan, where he conducts research on the genetics of rice. Rita Scheel-Ybert is a professor in the doctoral and master’s programs in archaeology at the National Museum of Rio de Janeiro. She is an archaeobotanist specializing in anthracology, and her main research interest is in coastal archaeology. Lynda Scott Cummings is the director of the PaleoResearch Institute in Golden, Colorado, and a palynologist-phytolith-starch analyst whose work includes palaeoenvironmental, subsistence, and palaeonutrition interpretations.

304

Editors and Contributors

Alexia Smith is an assistant professor in the Department of Anthropology at the University of Connecticut in Storrs, Connecticut. She works in agricultural development, climate change, and landscape use, with a focus on Bronze and Iron Age archaeology of the Near East. Chris Stevens is a researcher at the Institute of Archaeology, University College London, in London, United Kingdom. He is an archaeobotanist interested in the methodology of research, environmental archaeology, and domestication processes. Ken-ichi Tanno is a member of the Faculty of Agriculture, Yamaguchi University in Yamaguchi, Japan. He works on plant genetics and the archaeobotany of prehistoric West Asian archaeological sites, with special interest on breeding of winter crops and the origins of agriculture. Luc Vrydaghs is a member of the Centre de Recherches Archaeologiques et Patrimoine (ULB, Belgium) and is the founder of the Research Team in Archaeo- and Palaeo-Sciences in Brussels, Belgium. He is an expert on phytolith studies applied to archaeological contexts in Europe, (sub)tropical zones, and dry zones (Jordan, Israel, and Emirates). Sarah C. Walshaw teaches African history in the History Department at Simon Fraser University in Canada, where she also holds an adjunct professorship in archaeology. Her latest research focuses on the social relations of colonial and current agricultural practices in Zanzibar and Pemba, and she conducts archaeobotanical research at Songo Mnara in Tanzania, and Juffure in The Gambia.

Index

O isotope, 130–31n1

Antalya province, archaeological survey of, 114 antiquity: agriculture and climate change in, 47–49, 61–62; climate change and societal collapse in, 57–58; crop production in, 5, 58, 62, 79, 90; dating of, and the paleoclimatic record, 55–57; geographical and temporal biases concerning, 53–54; methodological issues concerning, 54–55; published material as data for, 51–52; regional approaches to agriculture, 49, 51; sample comparisons data, 52–53; theoretical concerns of, 57–61 Arabia: regional exchanges in southeastern, 26–27; southeastern, 30; southern, 27 Arabian Gulf, 27 Arabian Peninsula, 28, 30 archaeobotany, 3–6, 49, 120; and molecular phylogeny, 288; and social organization, 175. See also labor scheduling, as an agenda for social archaeobotany archaeology: “contextual,” 175–76; and contributions to social science, 174–75 Artemisia sp., 114 artesian springs, 28 ash (Fraxinus), 113, 114 Asia (East), 225, 285 Asia (South), 175, 225–26, 230, 268. See also Indus civilization Asia (Southwest), 51, 53, 60; climate shift in, 56; plant-based subsistence in Upper Southwest Asia, 103, 108–9; pollen records of, 56; social collapse at the end of the third millennium in, 57

18

Åberg, E., 279 Abi’el coins, 30 Abu Hureyra, 111 Acacia: Acacia sp., 28, 30; A. gerardii, 28 Acer, 114, 126 acorns (Quercus sp.), 103, 110, 112, 159. See also oak Adams, K. R., 294, 295 Aegean sites, 142, 144, 146, 147 Afghanistan, 222 Africa: eastern Africa, 79–80, 87; sub-Saharan Africa, 78 African grains, Chwaka and Tumbe production patterns of, 89–90 “African mode of production,” 76–77; archaeological signatures of, 77–78 agriculture, 76–77; “extensification” in, 59; and societal collapse, 60–61; and stone terracing systems, 80. See also antiquity: agriculture and climate change in; plant cultivation: in eastern North America ‘Ain Dara, 142; wild plant spectrum of, 144 Alexandre, A., 34 almonds (Amygdalus communis), 109, 111, 113. See also wild almonds (Amygdalus graeca or orientalis) amaranth, 166 Amygdalus. See almonds (Amygdalus communis); wild almonds (Amygdalus graeca or orientalis) Anatolia: mountains of, 113; western, 136 animals, 49, 57, 131n5, 218; domestic animals, 187

305

306

Index

Asia (West), 120–21; farmers in, 130; impact of the third-millennium drought in, 122–24; impact of the Younger Dryas on vegetation in, 121–22 Asian rice (Oryza sativa, O. nivara, O. officinalis, and O. rufipogon), 86, 265, 273; Chwaka and Tumbe production patterns, 90; early hypotheses concerning origins of, 267–68; evidence for indica-japonica differentiation in wild rice, 266–67; indica, 272; japonica, 271–72; phylogenetic relationships of wild and cultivated rice, 265–66; rice remains, 268–70; varieties of rice cultivars, 270–71. See also rice Assiros, 186 Atlantic Rain Forest, 238–39; in Santa Catarina State, 240; and similarities to restinga, 255 atomic absorption spectrometry (AAS), 32 Avdat, rainwater harvesting in, 59 Avicennia marina (Forssk.) Vierh, 28, 30 Balanites aegyptiaca (L.) Del., 28, 29 Ball, T. B., 34 Balu (Haryana), 225 Baluchistan, 220, 222, 225 banana (Musa sp.), 81 barley, 6, 58, 138, 139; discovery at Tell Aswad, 281; domestication of, 277, 279; drought-resistant, 148n1; grains (Hordeum vulgare subsp. spontaneum), 109, 183, 281, 282, 285, 287; hulled barley, 200; naked barley and ceramic vessels, 201; processing of naked barley, 202–3; resemblance between subsp. spontaneum and two-row cultivated barley, 279; and Tibet, 282–83; two-row barley, 145; two-row wild barley (subsp. spontaneum), 277, 279. See also Hordeum; six-row barley Big Sinking Creek drainage, 158, 162, 163 Boserup, E., 60

Bothmer, R., 282 Bottema, S., 123 bottle gourd (Lagenaria siceraria), 159 boxwood (cf. Buxus sempervirens), 114 Brazil/Brazilian coast: and archaeological research, 237–38, 240–43, 244–45, 252, 254–55, 257–58; carbonate sediments of the Araruama Lagoon, 254; and charcoal, 243, 244; climate change in, 254–55; and climate oscillations in the Rio Grande do Sul State, 254; dominance of Myrtaceae in, 245; exploitation of deadwood for firewood in, 257; human occupation of Brazilian shore, 240–41; incipient cultivation (horticulture) of tubers and fruit trees in, 242–43; and mangrove elements at Arraial do Cabo, 245, 252, 254; materials and methods used in the study of, 243–44; and plant associations southeast of Rio de Janeiro State, 238; and rocky outcrops in the Cabo Frio region, 238–39; sedentary nature of the Sambaqui people in, 257; site excavations in, 243; spatial analysis concerning, 242; and specific regional sites studied, 238–40, 245; stability of mainland vegetation in, 254; war in, 257. See also Atlantic Rain Forest; shellmounds (sambaquis) bristly foxtail (Setaria verticillata), 197 Bronze Age: in Asia, 48, 53, 55; in Near East, 136, 138–39, 145, 146; Late Bronze Age, 141, 14 Burman, 268 Butzer, K. W., 176 Caldwell, J. R., 155 Calotropis procera, 28, 30 Capparis, 28 Cappers, R. T. J., 123 Carex divisa (divided sedge), 145 Carthamus sp., 54; C. tintorius (safflower), 54 cassava (Manihot esculenta), 81, 87 Celtis sp. (hackberry), 103, 126

Index Centaurea spp., 137 cereal, 199–200; and ceramic vessels, 200–201; dehusking, 177; hulled, 204 charcoal, 5, 30, 52, 113, 126, 189, 197, 238; analysis, 36, 243–44; assemblages, 245, 255; flotation, 126; wood, 84, 87, 158, 180, 192, 223 charred plant assemblages, 111, 175, 180; categorization of, 177–79; content of, 179–80; and crop processing, 181; and recurrent activities, 182–86; as waste disposal from routine activities, 175, 180–81. See also crop processing Chen, W. B., 266 Chengtoushan, 270 Chenopodiaceae (goosefoots), 30 Chenopodium: Chenopodium spp., 137, 158; C. berlandieri, 156, 159, 160, 161–62, 163, 164, 165, 166; C. missouriense, 163 chestnuts, 159 chicken, 80, 90 China, 225, 270, 271, 279; crops originally from, 226; and Hunan Province, 269, 270; and Jiangxi Province, 269; and rice remains in Yangtze River basin, 268–70; southwest, 268; Tai Lake region of, 272; Yangtze valley of, 272; as Yellow River Civilization, 272; and Zhejiang Province, 269, 272 Chopani Mando (Allahabad, Uttar Pradesh), 223 Chrysophycea, 40 Chwaka, 74; archaeobotanical excavations at, 87; archaeobotanical patterns of production emerging from, 88–90; archaeological findings concerning the diet at, 80; archaeology of, 83; comparison of macrobotanical trends at Chwaka and Tumbe, 87–88; lack of African grains at, 88; pottery assemblages of, 83; settlement pattern data for, 91–92; stonetown component of, 83 Clapham, A., 178–79

307

Cliff Palace Pond, pollen record from, 164 climate, 130; definition of, 120; influence of on culture and cultural change, 120. See also climate change; Malyan climate change, 121; lag between climate change and vegetation, 131n2. See also antiquity: agriculture and climate change in; Brazil/Brazilian coast: climate change in; Malyan climate/environmental reconstruction, 120–21, 135. See also weed floras Cloudsplitter rockshelter, 158–59, 162 cluster sampling: conclusions concerning, 23–24; defining the population of, 11 coconuts (Cocos nucifera), 80, 81, 84, 86, 88; coconut shell fragments at Tumbe, 84; production patterns of at Chwaka and Tumbe, 89 Colacasia esculenta, 81, 87 Cold Oak shelter, 158, 163 Colledge, S., 51, 138, 146 Comoros Islands, 78; cereal cultivation on, 80 Conditions of Agricultural Growth, The (Boserup), 60 Conolly, J., 51 Coquery-Vidrovitch, C., 76 correspondence analysis (CA), 138, 141, 142, 146, 148nn2–3 cotton (Gossypium sp.), 86, 87, 88; cotton cloth, 92–93 Courthouse Rock shelter, 159 Cowan, C. W., 158–60, 166 Crataegus sp., 103, 114 crop husbandry, 177. See also weed floras: and relationship to environment and crop husbandry crop processing, 181, 189; crop-processing assemblage signatures, 183; importance of to archaeobotany, 187; non-mechanized crop processing, 183; postharvest processing, 199–205; processing of crops to spikelet or sheaf stage, 201–2; processing of monsoon crops, 205; processing of naked barley,

308

Index

crop processing (continued) 202–3; the processing sequence, 184; residues of, 183; role of storage in crop processing, 185–86; and the structure of charred assemblages, 183–84; two basic activities of, 183. See also Hund (northwest Pakistan), routine crop processing and dung burning in; labor mobilization (for crop processing), in Iron Age Britain Cucurbita gourd (Cucurbita pepo ssp. ovifera), 156, 159, 160–61, 162, 164, 165–66; appearance of, in eastern North America, 167–68 Cumberland Plateau: archaeobotanical record of, 162–63; development of food production in, 156–57; modern floristics of, 164–65; natural populations of floodplain weeds on, 162; paleoecological evidence for persistently disturbed habitats of, 163–64 Dalbergia sissoo, 30 date palm charcoals (Phoenix dactylifera), 30 Davis, P. H., 142, 148–49n5 De Condolle, A., 279 deforestation, 122, 123 dendroclimatology, 120 Dennell, R. W., 51, 176–77, 183 Devore, I., 112 Dhofar reference collection, 34 diatoms, 40 differential survival and retrieval rates, 21 Ding, Y., 282 Dioscorea spp., 87 Distribution (D) index, 32 dom palm (Hyphaena thebaica), 109 drought, 59, 130; third-millennium drought, 121, 122–24 dugong (sea cow), 80 dung, 55, 189, 191; as fuel, 123–24, 137, 187, 189; as a source for charred seeds, 186–87. See also Hund (northwest Pakistan), routine crop processing and dung burning in

“Eastern Agricultural Complex,” 158 Echinochloa crus-galli, 137 ed-Dur, 26, 40; absence of kiln finds in, 32; and evidence of regional contacts, 28, 30; excavations at, 27; mangrove stands of, 30; pottery of, 27, 31; vegetation of (pseudosavannah), 28; vegetation distribution on, 28 Elamite, 137 Eleusine coracana coracana, 84, 87, 88. See also millet (cf. Setaria); pearl millet Emar, 142 Engaruka, 80 England, 175; class-based poverty in eighteenth-century, 60. See also labor mobilization (for crop processing), in Iron Age Britain Epipaleolithic cave sites. See Karain B; Öküzini Essay on the Principle of Population, An (Malthus), 60 Euphrates sites, 142, 146; northern Euphrates, 145 Europe, 176; northwest Europe, 187; southeastern Europe, 51 Fabeae, 103, 108 famine, 58 Fertile Crescent, 102 Fieller, N. R. J., 12, 13 fish, 80, 241; fish bones, 241 Fleisher, J., 74, 80, 83 floatation, application of, 176 floodplain weeds, 5, 157–58, 162, 163, 164–65, 166–67; “floodplain weed theory,” 156 floodwaters: and forest development, 163–64; as a means of seed dispersal, 163 food production, and the organization of labor, 174–75 food production, in Swahili communities during urbanization (Pemba Island, Tanzania), 73–74, 92–93; and archaeobotanical evidence from excavations, 83–87; in settlement

Index communities in northern Pemba, 90–91; and Swahili archaeology on Pemba Island, 81–83; and trade and subsistence patterns in northern Pemba, 90–92; and the trait-based approach to urbanism, 74; and types of foods grown, 81; and urban food production, 75–77; urbanization and food production, 78–80. See also Chwaka; Tumbe food production systems, in Africa, 73; household-based production, 77–78; Uruk food production system, 76 food supply, inelasticity of, 60 foraging models, optimal, 101–2, 115; and criticism of diet breadth model, 102; and diet breadth model (optimal diet model), 101; and efficiency rank, 101; predictions of, 101, 112–13; vs. predictions of the basic model, 101–2; and ranking of main classes of food plants, 112; and search time, 101, 113; testing the predictions of, 113–15 Fort Derawar, 220 Franchthi II (Greece), 109 Fraxinus sp., 113, 114 Freisleben, R., 279, 282 fruit/fruit trees, 81, 88, 110, 131, 142, 242–43; citrus fruit, 80. See also individual fruits Fuller, D. Q., 191, 225, 226 Fumaria spp., 137 Ganges Valley, 225 Garner, A., 284 Ghaggar-Hakra River, 220 giant ragweed (Ambrosia trifida), 166 Gini-Lorenz concentration curves, 244 glumes, 183, 195; hard glumes, 292; Oryza sativa, 225; soft glumes, 292 goosefoot (Chenopodium berlandieri), 156, 159, 160, 161–62, 164, 165, 166; archaeobotanical record of, 162–63 Gossypium sp., 86, 87, 88 grain spikelet, 186

309

grapes (Vitis vinifera), 138, 145. See also wild grapes (Vitis vinifera subsp. sylvestris) gravel plains, 28 greenhouse gases, 47–48, 122 grinding stones, 80, 108 Gujarat, 222 Gulf of Mumbai (Bombay), 222 hackberry fruit (Celtis sp.), 103, 126 Hallan Çemi, 111 handling time, in plant collection, 113 Harappan, 225; Early Harappan, 220; Late Harappan, 228, 229; Mature Harappan, 222, 226, 227, 228, 229 Harlan, J. R., 277, 281 hawthorn (Crataegus sp.), 103, 114 Hayonim, 111 Helbaek, H., 200–201, 204, 280–81 Helianthus annuus (sunflower), 156, 159 hemp (Cannabis sativa), 226 Hemudu, 269, 270 hickory nuts (Carya tomentosa), 159 Hillman, G. C., 144, 176, 177, 183, 187; and ethnoarchaeological work in Turkey, 182 Hittite Empire, 136; New Hittite Kingdom, 138 Hodder, I., 175 Holocene, 240; Early Holocene, 241; Late Holocene, 122 Hooten Hollow shelter, 159 Hopewell, 155 Hoph, M., 148n1 Hordein, genetic diversity of, 282 Hordeum: H. agriocrithon , 279, H. vulgare subsp. spontaneum, 109, 183, 279, 281, 282, 285, 287. See also barley; six-row barley horsegram (Macrotyloma uniflorum), 197 Horton, M. C., 79 Hubbard, R. N. L. B., 51, 178–79 human behavioral ecology, 101 Hund (northwest Pakistan), routine crop processing and dung burning in, 191–92, 194, 196

310

Index

hunter-gatherers, 100–101, 112, 281, 282; use of H. vulgare subsp. spontaneum by, 282 Hyphaene thebaica, 109 Ibn Khaladūn, 48 Illinois River Valley, 162 India, 27, 204, 270; Assam district of, 268; North-Western Frontier Province, 220 Indian Ocean, trade economy of, 79, 82, 92 Indus civilization, 219–20; crops and foods of Indus society, 226–28; evidence for rice in, 222–23, 225–26, 230; food choice in, 218–19; role of rice in, 228–30; and specific sites of rice evidence, 223, 225 industrialization, 28 Ipomoea batatas, 81 Iran, 27, 122; climate of western Iran, 126; southern Iran, 130 Iraq, 280 Ireland, 175; Bronze Age Ireland agricultural systems, 199–200 Iron Age (Asia), 48, 53, 55 Iron Age (Near East), 136, 138–39, 141, 145, 146; Early Iron Age, 147 Ito, E., 122 Jalyan, 126 Japan, 271 Jericho, 148n1 Jessen, K., 200–201 Jones, G. E. M., 53, 182–83 juniper (Juniperus sp.), 113, 126; decline in drought-resistant, 126; as food for grazers, 131n5; Juniperus excelsa, 28, 126; reduction in, between the Banesh and Kaftari settlement periods, 126–27, 130 Juniperus: Juniperus sp., 113, 126–27, 130; J. excelsa, 28, 126 Karain B, 100–101, 111, 112–13, 115; plant-based subsistence in,

103, 108–9. See also plant-based subsistence strategies Kenoyer, J. M., 219, 220 Kentucky River drainage, 162, 163 Khabur region, 142, 145, 146, 147 Kilwa, 78; cereal cultivation at, 80, 84 Kitigawa, H., 122 Knörzer, K. H., 182 knotweed (Polygonum sp.), 158 Komatsuda, T., 284 Körber-Gröhne, U., 182 Korea, 271. See also Mumun-period sites (Korea) Kornicke, F., 279 Koster, 161 Kuck Shelter (Colorado), 298 Kur River basin, 124, 125, 126, 130 Kurban Höyük sequence, 124; and Bogaard, A., 131n4 labor mobilization (for crop processing), in Iron Age Britain, 194–96 labor scheduling, as an agenda for social archaeobotany, 206–7 Lagenaria siceraria, 159 Lake Ghab, 122 Lake Van 56, 122; isotope studies at, 122, 130–31n1 Lake Zeribar, 122, 123; isotope studies at, 122, 130–31n1; pollen cores from, 121 Laos, 268 Late Glacial Maximum, 103, 113 Late Woodland, 158–59 lavas: peridotite lava, 42; radiolarite lava, 42; spilitic lava, 42 LaViolette, A., 74, 83, 93 Lee, R. B., 112 Legge, A. J., 187 Lemcke, G., 122, 130–31n1 Lens sp., 109, 111 lentils (Lens sp.), 109, 111 Levantine Epipaleolithic, 102 Lewin, R., 131n2 Li, A., 282 Licking River drainage, 158

Index Ma, D. Q., 282 Macrotyloma uniflorum (horsegram), 197 Madella, M., 34 Mahagara, 223, 225 maize (Zea mays), 81, 156, 300; archaeological, 298–299; experimental populations of, 295–98; Fremont and Kayenta Anasazi, 299; measurement of, 295–97; Mesa Verde, 299; phytoliths as indicators of genetics and environmental conditions, 292–95; Tohono O’odham, 299; Verde Anasazi, 299 Malthus, T., 60–61 Malyan (Iran), 130, 137; almond and oak distribution in, 128–29; ceramics of, 125; decline of juniper forest in, 126–27; pistachio and almond forests in, 126; presence of nomadic pastoralists near, 125–26; settlements in, 124–25 Manda, 78 maple (Acer sp.), 114, 126 Mapunda, B., 74, 83, 91 material culture, 30–32 material selection, laboratory procedures, and methods, 32, 34–35 maygrass (Phalaris caroliniana), 158, 159 McCorriston, J., 54 mechanical interference, 23 Mediterranean, the, 122; coast (‘Ain Dara), 142, 144; eastern, 27, 123, 236; woodland zone of, 144 Mesoamerica, 74; crops of, 156 Mesopotamia, 27, 74; administration of grain production in, 75; city-states of, 124; collapse of Akkadian Empire of, 122, 122; northern, 123, 130 Middle Woodland, 166 Middleton, J., 79 Miller, N. F., 56, 136 millet (cf. Setaria), 80, 183, 198, 226; browntop millet (Brachiaria ramosa), 197; finger millet (Eleusine coracana spp. coracana), 84, 87, 88. See also pearl millet

311

mineral grains, 43 Mleiha, 26, 27, 30, 40 molecular phylogeny, and archaeobotany, 288 Mombasa, 82 Monotheca buxifolia, 28 monsoons, 204; Indian Ocean monsoons, 30; processing of monsoon crops, 205 Morphotype Identification (MI) index, 32, 41 Mounded Talus shelter, 163 Mozambique, 78 mudstone fragments, 43 Muenchrath, Deb, 293, 294, 295 Mumun-period sites (Korea), 10–11; Early Mumun, 11; Middle Mumun, 11, 14; Middle Mumun Oun 1 site, 11 mung beans (Vigna radiata), 87, 197 Muqaddimah (Ibn Khaldūn), 48 Mureybit, 111, 137–38 Myrtaceae, 245 Mysore Plateau, 197 Nakagawa, T., 122 Narayana Rachanives Palace, 271 Near East, the, 130, 135, 136. See also weed floras: and relationship to environment and crop husbandry Near Eastern Agricultural Database (NEAD), 49, 52, 137 New Kash shelter, 159 North America, 5; prehistoric eastern, 187. See also plant cultivation: in eastern North America Northern Neolithic, 225 nuts, 110, 112, 113; risks involved in reliance on, as a food source, 159. See also individual nuts Nyanga, 80 oak: deciduous (Quercus sp.), 113, 114, 121; increase of, during the Banesh period, 128; pollen rain from, 164. See also acorns (Quercus sp.) Observed/Not Observed (O/NO) index, 32, 41

312

Index

Ohalo II, 111 Öküzini, 100–101, 111, 112–13, 115; archaeozoological record of, 114; plant-based subsistence in, 103, 108–9; small pulses (Fabeae) and small seeds found at, 103, 108; woody steppic species found at, 114. See also plant-based subsistence strategies Olea sp., 114 olive trees (cf. Olea sp.), 114 Oman, 27, 30; Al Hajar Mountains of, 28; floristic diversity of, 28 Origin of Cultivated Plants (De Candolle), 279 Orton, C., 20–21, 24 Oryza: O. sativa L., O. nivara, O. officinalis and O. rufipogon, 86, 225, 265, 273, 267–68; O. sativa subsp. indica, 265–66, 272; O. sativa subsp. japonica, 265–66, 271–72; O. indica-japonica, 266–67. See also Asian rice (Oryza sativa, O. nivara, O. officinalis, and O. rufipogon); rice oxygen isotope analysis, 120 packrat middens, 131n2 Pakistan, 175, 270; North-Western Frontier Province, 220; Sindh, 220, 228 Palenque region, 32 palynology, 120 Papadakis, J., 30 pastoralism, 59, 78; nomadic pastoralists, 125–26 Pate, archaeological findings concerning the diet at, 80 peaches (Prunus persica), 226 pearl millet, 81, 88, 89–90; grains (Pennisetum glaucum), 84; spikelet bases, 84, 89 pears (Pyrus amygdaliformis), 109. See also wild pears (Pyrus sp.) Pennisetum glaucum, 84. See also millet (cf. Setaria); pearl millet Pemba Island. See food production, in Swahili communities during

urbanization (Pemba Island, Tanzania) Peres, T. M., 63n2 Persian Gulf, 123; sediment deposition in, 120–21 petrography, 31; thin section petrography, 31–32 Phalaris caroliniana, 158, 159 Phoenix datylifera, 30 phytoliths, 30, 225, 268; classification of, 34–35; grass phytoliths, 40; indexes for phytolith analysis, 40–43; phytolith analysis, 32, 192; results of the analysis of, 35–36, 38–40; rice phytoliths, 225 pine (Pinus sp.), 114; pollen rain from, 164 Pistacia sp. See wild pistachios (Pistacia sp.) pistachios. See wild pistachios (Pistacia sp.) plagioclase feldspars, 43 plant assemblages, 12, 138, 199; composition of, 108–9, 180. See also charred plant assemblages plant-based subsistence strategies, 102; regional review of, 109–11; in the Upper Paleolithic, 109; in Upper Southwest Asia, the Epipaleolithic period, and southeastern Europe, 103, 108–9 plant cultivation: and agricultural/ farming origins, 158–60, 165–66; and crop diffusion, 157; cultural transmission, subsistence change, and, 167–68; in eastern North America, 155–58, 168–69; and floodplain weeds, 160, 165–68; and geographic distribution of weedy annuals, 156; of North American wild populations, 160–62; and introduction of Mesoamerican crops to eastern North America, 156; and storage of crop seeds, 158–59 plants: annuals, 159–60, 163; inherent immobility of, 131n2; insectpollinated, 113, 121; wind-pollinated,

Index 121. See also plant assemblages; plantbased subsistence strategies; plant cultivation; plant taxa plant taxa: as climate proxies, 121–22, 136; filtering of, 148–49n5 Pleistocene, 122; Late Pleistocene, 241; Late Pleistocene foragers, 187 plums (Prunus sp.), 114 pollen/pollen record, 113–14, 121, 164; from aquatic and semiaquatic plants, 114; fluxes in pollen assemblages, 56; pollen cores, 121; pollen production, 121; pollen rain (from oak and pine), 164 Polygonum sp., 158, 166 poplar (Populus sp.), 114 population/population dynamics, 60–61 Populus sp., 114 Possehl, G. L., 219 pottery: Black ware, 30, 38–39; Coarse Buffed Slipped Ware, 31; Coarse Buff Orange Ware, 31, 39, 43; Coarse Ochre Ware, 31, 42–43; Coarse Pinkish Brown Ware, 31, 42; Coarse Yellow-Pink Speckled, 31, 38, 41–42; and dedicated drying facilities, 203–4; Fine Orange Vegetal Painted, 30–31; Fine Orange Vegetal Tempered, 31; “leatherhard” state of, 203; local production of, 30, 31, 32; nonlocal production of, 30; Orange Vegetal Tempered Ware, 30, 35–36, 42; production in Bronze Age Ireland, 199–205; and relationship between charred remains and pottery impressions, 204–5; of Vindhyan Neolithic of Koldihwa, 223 Pre-Pottery Neolithic (PPN), 102, 137, 187 “Primary Forest Efficiency,” 155 Prosopis cineraria, 30 Proto-Elamite, 137 Prunus: Prunus sp., 114; P. scoparia , 129 pulses, 113, 145, 183, 196, 198, 222, 225; “pod-threshing” pulses, 183; small pulses (Fabeae), 103; wild pulses, 110, 111

313

Punjab, 228 Pyrus: Pyrus sp., 103, 113; P. amygdaliformis, 109 Qashqa’i nomads, 125 Qatar, 26, 27 quantitative compatibility, 9 quartz, 43 rainfall: and cropping decisions, 58; harvesting of, 59; and permanent settlements, 59–60. See also monsoons ratio of ratios model (survival ratios), 20–23 Red River/Red River Gorge, 158, 163, 164; human Late Woodland rockshelter occupation in, 158–59 Rhizophora sp., charcoal evidence for, 30 rice, 81, 88; direct dating of Tumbe rice specimens, 86; hybridization of, 230; indica and japonica strains of, 230; rice grains and fragments from Tumbe, 86; rice phytoliths, 225; surplus of rice mobilized for political use, 79. See also Asian rice (Oryza sativa L., O. nivara, O. officinalis, and O. rufipogon); Indus civilization: evidence for rice in; Oryza Roberts, N., 122 Roman Empire, 27 Romanization, 196 roots/tubers, 112 rosehips (Rosa sp.), 103 Rosen, A. M., 123 Rowley-Conwy, P. A., 187 Ruddiman, W. F., 47, 122 Runge, F., 34 sabkhas, 28; fluvio-lacustrinein sabkhas, 28; inland sabkhas, 28 salinity, 28, 252, 254; soil salinity, 145 Salts Cave, 163 sample size: application and results of, 14, 20; estimation of, 11–12; suggested formula for, 13–14 sand desert, 28 “sand sea” (Rub’ al-Khali), 28 Sato, Y.-I., 269

314 Schiemann, E., 279 Schwindt, D. M., 294, 295 Scirpus maritimus/tuberosus, 109. See also seeds search time, in plant collection, 101, 113 sedimentology, 120 sediments: fluvial sediments, 28; natural sediments, 120; sediment samples, 11 seed number estimate (SNE), 87 seeds: club rush seeds (Scirpus maritimus/tuberosus), 109; fossil seeds, 108; oat seeds, (Avena sp.), 109; Old World seed assemblages, 175; roasting of small seeds, 109; seed dispersal by means of floodwaters, 163; small seeds, 112; trivial role of small seeds, 108–9 Semail ophiolite complex, 28, 42 Sen, Amartya, 58 Setaria, 80, 183, 198,. 226; S. verticillata, 197. See also millet (cf. Setaria); pearl millet Shanga, 78; archaeological findings concerning the diet at, 80 Shao, Q., 282 shellfish, 80, 240 shellmounds (sambaquis), 240, 241–42 Shennan, S., 51 Silene sp., 137 silking, 292–93 Sindh, 220, 228 six-row barley, 148n1, 277, 279; discovery of earliest, at Tell Abu Hureyra, 281; and DNA-marker analysis of f. agriocrithon, 282–83, 287–88; experiment to determine the origins of, 284–88; genetic study of, 283–84; origin of, 277; and research concerning two-row barley as progenitor of, 279–82; RFLP studies of, 284; six-row brittle barley (f. agriocrithon or f. laguncliforme), 277, 279, 287; six-row brittle barley (Hordeum agriocrithon), 279. See also barley; Hordeum

Index Smith, A., 137 Smith, B., 156 Smith, H., 279 social history, constraints of, 174 social science, archaeology’s contributions to, 174–75 Somalia, 78 Sorbus sp., 114 sorghum, 81, 88, 89; sorghum grains (Sorghum bicolor ssp. bicolor), 80, 84, 87 South Indian Neolithic: “ashmound sites,” 197–99; agricultural and social systems of, 196–99; major foodstuffs of, 197–98 spikelet bases, 84, 86, 87, 89, 90 sponge spicules, 40 Sporobolus sp., 28, 35, 38, 40 Sri Lanka, 27 Stevens, L. R., 122 stonetowns. See Swahili cities (stonetowns) Strait of Hormuz, 26, 27 Stuiver, M., 122–23 Sturm, M., 122, 130–31n1 sumpweed (Iva annua), 156, 157, 158, 159, 160, 163, 164, 165–66; Koster samples of, 161 sunflower (Helianthus annus), 156, 159 survival probabilities, 21 Swahili cities (stonetowns), 78–79, 83; as non-African phenomena, 79 Swahili communities/Swahili coast. See food production, in Swahili communities during urbanization (Pemba Island, Tanzania) sweet potatoes (Ipomoea batatas), 81 Syria, 49, 53, 142, 144; agricultural settlement zones in, 58–59; eastern Syria, 49, 54; northern Syria, 54, 123, 136; PPN sites in, 138, 146; the Syrian Jazira, 54 Takahashi, R., 279–80 Takeda, K., 283, 284 tamarisk 30, 114 Tamarix sp., 30, 114

Index Tanno, K., 283, 284 Tanzania, 27. See also food production, in Swahili communities during urbanization (Pemba Island, Tanzania) taro (Colacasia esculenta), 81, 87 Tell Abu Hureyra, 281 Tell Aswad, 281 Tell Bderi, 142, 146 Tell Bi’a, 58 Tell Brak, 131n4 Tell Ghoraifé, 137 Tell Hadidi, 142 Tell Munbaqa, 142 Tell Schech Hamad, 142, 145, 146, 147 Tell Sheikh Hamad, 54 Tell Shiukh Fawqani, 142 Texas, 298, 299 Thailand, 268, 270, 271 Thames Valley, Iron Age settlements in, 194–96 Tibet, as a center for the origin of cultivated barley, 282–83 Tigris-Euphrates watershed, 121, 123 tomato (Solanum lycopersicum), 81 trade: archaeobotany and understanding, in northern Pemba, 90–92; importance of, to community sustenance, 59–60; Indian Ocean, 79; as original “African mode of production,” 76–77 tree rings, 120 Triticum sp.: T. aestivum L. (bread wheat), 53; T. dicoccum, 204; T. durum (durum wheat), 53. See also wheat (Triticum spp.) Troy, 147 Tumbe, 74; archaeobotanical excavations at, 84, 86–87; archaeobotanical patterns of production from, 88–90; archaeology of, 82–83; “bead grinders” and iron slag at, 82, 91; comparison of macrobotanical trends at Chwaka and Tumbe, 87–88; later occupation of, by the Mazrui, 82–83; maritime trade economy of, 82 Turkey, 182; southeastern, 48 Turkmenistan, 287

315

Ubeid period (5500–4000 BC), pottery of, 26 Ubish, G. V., 283 Umm an Nar period, 26–27 Umm al-Qaiwain, 26 Unguja (Zanzibar Island), 81 United Arab Emirates (UAE), 26, 27, 30, 43 urbanization, 28; in Africa, 73–75 Utah, 298 VanDerwarker, A. M., 63n2 Van der Veen, M., 12, 13 vetch (Vicia sp.), 109; bitter vetch (Vicia ervilia), 111 Vietnam, 268, 270 Vicia sp., 109, V. ervilia, 111 Vitis vinifera subsp. sylvestris, 103 von Tschermak, E., 279 Wadi Kubbaniya, 109, 111 Wadi Suq period, 26 walnuts (Juglans sp.), 159 weed floras: conclusions concerning, 146–47; future perspectives concerning, 147; initial results concerning, 141–42, 144–46, 148n4; methodological problems concerning, 139, 141; methods used with, 138–39, 141, 142, 146, 147, 148nn2–3; and relationship to environment and crop husbandry, 135–35; wild plants species and genera considered in the analysis of, 148–49n5 Weiss, H., 57 Weisskopf, A., 225, 226 Werner, H., 279 wheat (Triticum spp.), 183; dominance of hulled wheat in the Aegean, 139; emmer wheat (Triticicum dicoccum), 204; free-threshing wheat, 148n1, 202; various types of, 138–39. See also Triticum sp. whitebeam (Sorbus sp.), 114 wiggle-match dating, 57 wild almonds (Amygdalus graeca or orientalis), 103

316 wild almonds (Prunus scoparia), 129 wild grapes (Vitis vinifera subsp. sylvestris), 103 wild pears (Pyrus sp.), 103, 113 wild pistachios (Pistacia sp.), 103, 109, 111, 113, 114, 121 willow (Salix sp.), 114 wood, as fuel, 137 wormwood (Artemisia sp.), 114 Wright, H. E., Jr., 122 Xianrendong, 269 yam (Dioscorea spp.), 87 Yasuda, Y., 122

Index Yin, Y. Q., 282 You, X.-L., 270 Younger Dryas, 121 Yuchanyan, 269 Zagros Mountains: central Zagros, 121, 122, 125; and climate in Iran, 126; southern Zagros, 124 Zeder, M. A., 54 Zettler, R., 122 Zhang, Q., 282 Ziziphus spina-christi, 30 Zohary, D., 281 zooarchaeology, 49 zoomorphic sculptures, 241