Biogeochemical Approaches to Paleodietary Analysis (Advances in Archaeological and Museum Science)

Biogeochemical Approaches to Paleodietary Analysis

ADVANCES IN ARCHAEOLOGICAL AND MUSEUM SCIENCE

Series Editors: Martin J. Aitken, FRS, Oxford University Edward V. Sayre, Smithsonian Institution and R.E. Taylor, University of California, Riverside Volume 1

PHYTOLITH SYSTEMATICS: Emerging Issues

Edited by George Rapp, Jr. and Susan C. Mulholland Volume 2

CHRONOMETRIC DATING IN ARCHAEOLOGY

Edited by R.E. Taylor and Martin J. Aitken Volume 3

ARCHAEOLOGICAL OBSIDIAN STUDIES: Method and Theory Edited by M. Steven Shackley

Volume 4

SCIENCE AND TECHNOLOGY IN HISTORIC PRESERVATION Edited by Ray A. Williamson and Paul R. Nickens BIOGEOCHEMICAL APPROACHES TO PALEODIETARY ANALYSIS

Volume 5

Edited by Stanley H. Ambrose and M. Anne Katzenberg

A Continuation Order Plan is available for this series. A continuation order will bring delivery of each new volume immediately upon publication. Volumes are billed only upon actual shipment. For further information please contact the publisher.

Biogeochemical Approaches to Paleodietary Analysis Edited by STANLEY H. AMBROSE University of Illinois, Urbana-Champaign Urbana, Illinois, USA

and M. ANNE KATZENBERG University of Calgary Calgary, Alberta, Canada

Published in cooperation with the Society for Archaeological Sciences

KLUWER ACADEMIC PUBLISHERS NEW YORK, BOSTON, DORDRECHT, LONDON, MOSCOW

eBook ISBN: Print ISBN:

0-306-47194-9 0-306-46457-8

©2002 Kluwer Academic Publishers New York, Boston, Dordrecht, London, Moscow Print ©2000 Kluwer Academic / Plenum Publishers New York All rights reserved No part of this eBook may be reproduced or transmitted in any form or by any means, electronic, mechanical, recording, or otherwise, without written consent from the Publisher Created in the United States of America Visit Kluwer Online at: and Kluwer's eBookstore at:

http://kluweronline.com http://ebooks.kluweronline.com

Consulting Editors

Arnold Aspinall

Heather Lechtman

University of Bradford Bradford, England

Massachusetts Institute of Technology Cambridge, Massachusetts

Rainer Berger University of California Los Angeles, California

Jonathan E. Ericson University of California, Irvine Irvine, California

Christian Lahanier Louvre Research Center Paris, France

Hisao Mabuchi Tokyo National Research Institute of Cultural Properties Tokyo, Japan

Donald K. Grayson

Robert Maddin

University of Washington

Harvard University

Seattle, Washington

Cambridge Massachusetts

Garman Harbottle Brookhaven National Laboratory Brookhaven, New York

David Harris University of London

Patrick Martin Michigan Technological University Houghton, Michigan

Frederick R. Matson Pennsylvania State University

London, England

University Park, Pennsylvania

W. David Kingery

Risø National Laboratory Copenhagen, Denmark

University of Arizona Tucson, Arizona

B. Foss Leach Wellington, New Zealand

Vagn Mejdahl

Pieter Meyers Los Angeles County Museum of Art Los Angeles, California

Jacqueline S. Olin

Julie K. Stein

Smithsonian Institution

University of Washington

Washington, D.C.

Seattle, Washington

Ernst Pernicka Max Planck Institute for Nuclear Physics Heidelberg, Germany

Henrik Tauber National Museum of Denmark Copenhagen, Denmark

John R. Prescott University of Adelaide Adelaide, Australia

Michael S. Tite Oxford University

Frank Preusser

Oxford, England

Los Angeles, California

T. Douglas Price University of Wisconsin Madison, Wisconsin

Giorgio Torraca University of Rome

Rome, Italy

Françoise Schweizer Laboratory of the Museum of Art and History Geneva, Switzerland

Lambertus Van Zelst Smithsonian Institution Washington, D.C.

To

Harold W. Krueger founder of

Geochron Laboratories and Krueger Enterprises for his fundamental contributions to paleodietary analysis with stable isotopes and radiocarbon and stable isotope geochemistry of bone

This page intentionally left blank

Contributors Sylvia Abonyi • Department of Anthropology, McMaster University, Hamilton, Ontario L8S 4L9, Canada Stanley H. Ambrose • Department of Anthropology, University of Illinois, Urbana, IL 61801, USA Astrid Balzer • Institut für Anthropologie und Humangenetik, LudwigMaximilians-Universität, Richard-Wagner-Straße 10/I, D-80333 München, Germany Hervé Bocherens • Laboratoire de Biogéochimie Isotopique, Université P. et M. Curie, case courrier 120, 4 place Jussieu, F-75252 Paris, Cedex 05, France James H. Burton • Department of Anthropology, University of Wisconsin, Madison WI 53706, USA Gisela Grupe • Institut für Anthropologie und Humangenetik, LudwigMaximilians-Universität, Richard-Wagner-Straße 10/I, D-80333 München, Germany Norman Hammond • Department of Archaeology, Boston University, 675 Commonwealth Avenue, Boston, MA 02215, USA Robert E.M. Hedges • Research Laboratory for Archaeology, University of Oxford, 6 Keble Road, Oxford, OX1 3Q1, UK M. Anne Katzenberg • Department of Archaeology, University of Calgary, 2500 University Dr. N.W, Calgary, Alberta T2N 1N4, Canada Matthew J. Kohn • Department of Earth Science Lawrence Livermore National Lab, Livermore, CA 94550, USA Julia A. Lee-Thorp • Department of Archaeology, University of Cape Town, Rondebosch, Cape, 7700, South Africa ix

x

CONTRIBUTORS

Kim Oakberg • Department of Archaeology, Boston University, 675 Commonwealth Avenue, Boston, MA 02215, USA

Susan Pfeiffer • Human Biology and Nutritional Sciences, University of Guelph, Guelph, Ontario, NIG 2W1, Canada T. Douglas Price • Department of Anthropology, University of Wisconsin, Madison WI 53706, USA Mike P. Richards • Research Laboratory for Archaeology, University of Oxford, 6 Keble Road, Oxford, OXI 3QJ, UK Shelley R. Saunders • Department of Anthropology, McMaster University, Hamilton, Ontario, L8S 4L9, Canada Margaret J. Schoeninger • Department of Anthropology, University of Wisconsin, Madison WI 53706, USA Henry P. Schwarcz • School of Geography and Geology, McMaster University, Hamilton, Ontario, L8S 4M1, Canada Susanne Turban-Just • Institut für Anthropologie und Humangenetik, LudwigMaximilians-Universität, Richard-Wagner-Straße 10/I, D-80333 München, Germany Robert Tykot • Department of Anthropology, University of South Florida, 4202 East Fowler Avenue, SOC 107, Tampa, Florida 33620-8100, USA John W. Valley • Department of Geology and Geophysics, University of Wisconsin, Madison, Wisconsin 53706, USA Gert J. van Klinken • 14C Accelerator Laboratory, Max-Planck Institut für Biogeochemie, Sophienstrasse 10, D-07743 Jena, Germany Nikolaas J. van der Merwe • Department of Anthropology, Peabody Museum, Harvard University, 11 Divinity Avenue, Cambridge, MA 02138, USA Tamara L. Varney • Department of Archaeology, University of Calgary, 2500 University Dr. N.W., Calgary, Alberta T2N 1N4, Canada

Series Foreword This volume is the fifth in the Advances in Archaeological and Museum Science series by the Society for Archaeological Sciences (SAS). The purpose of this series is to provide summaries of advances in various topics in archaeometry, archaeological science, environmental archaeology, preservation technology, and museum conservation.

The SAS exists to encourage interdisciplinary collaboration between archaeologists and colleagues in the natural sciences. SAS members are drawn from many disciplinary fields. However, they all share a common belief that natural science techniques and methods constitute an essential component of archaeological field and laboratory studies.

xi

This page intentionally left blank

Preface The study of human diet brings together researchers from diverse backgrounds, ranging from modern human nutrition and biochemistry to the geochemistry of fossilized bones and teeth. Human paleodiet research, as studied through the chemical composition of bones and teeth, has been advanced significantly in the last 25 years, since the publication of early work on trace elements (Brown 1973) and on stable carbon isotopes (Vogel and van der Merwe 1977, van der Merwe and Vogel 1978). An important forum for such progress has been the series of Advanced Seminars on Paleodiet, held every three years since 1986. The contributions in this volume arose from the Fourth Advanced Seminar on Paleodiet, which was held in Banff, Alberta in September of 1994. The Advanced Seminars bring together a small international group of researchers interested in improving and expanding techniques for studying past diet through bone chemistry. These intensive seminars provide a unique opportunity for a group of scholars with common interests to spend several days together to discuss the latest advances and interpretations and to chart future directions for paleodietary research. The objectives of the Seminars are to: (1) present the results of new research to participants and assess their significance and implications; (2) assess the utility and validity of existing methods of isotopic, trace element and molecular analysis of bones, teeth, hair and residues of dietary resources; (3) evaluate the design of controlled diet and environment experiments; (4) discuss potential improvements in methods of sample preparation, characterization, analysis and identification of diagenesis; (5) identify new applications of methods; (6) set priorities and define problem areas for future research. xiii

xiv

PREFACE

Research presented during the Advanced Seminars has relevance for reconstructing prehistoric climates, microhabitats and environmental change, radiometric and chemical dating of bones and teeth, identification and elucidation of processes of diagenesis of organic and inorganic phases of skeletons, molecular genetic and phylogenetic studies of fossils, and skeletal biology and paleopathology. Many participants are actively pursuing these additional avenues of research, all of which are relevant to archaeology, paleoanthropology, and anthropology. This field of research also has wider implications

for biomedical, agronomic, ecological and nutritional research. For example, some studies demonstrate methods of tracing environment, nutrition and physiology by analysis of natural abundance levels of stable isotopes and trace elements rather than through the use of radioisotopes and toxic levels of chemical tracers (Coleman and Fry 1991; Boutton 1991; Wolfe 1984). Contributors to these Seminars are actively pursuing these additional avenues of research.

The first Advanced Seminar, held in Santa Fe, New Mexico resulted an the edited volume titled The Chemistry of Prehistoric Bone (Price 1989). The second seminar, held in Cape Town, South Africa, in 1988, resulted in a special issue of the Journal of Archaeological Science (edited by Sillen and Armelagos 1991). The third Advanced Seminar was held in Bad Homburg, Germany, in 1991, and produced the edited volume titled, Prehistoric Human Bone: Archaeology at the Molecular Level (Lambert and Grupe 1993). The Advanced Seminars have been crucial for evaluating the utility and validity of existing methods and results, discussing potential improvements in methods, and setting priorities for future research. A brief summary of the advances made in previous seminars and future directions of research on chemical and isotopic techniques of diet reconstruction follows. The first volume (Price 1989) included two important papers on the current state of understanding natural variations in stable carbon isotope ratios of ecosystems (Chisholm, van der Merwe), the first sophisticated theoretical isotopic mass balance model of the relationship between diet and bone collagen carbon and nitrogen isotopes (Schoeninger), and a review of potential and realized applications of trace element analyses of bones (Buikstra et al.) for determining trophic level with strontium abundances and Sr/Ca ratios (meat versus plants) and other classes of dietary items such as nuts and fruits by the analysis of zinc. The problem of diagenesis of the inorganic phase of bone was addressed by Price, Sillen, and Buikstra and colleagues. One contribution focused on reconstructing amounts of marine resources and maize to the diets of coastal and inland communities in Peru with carbon and nitrogen isotopes of bone collagen (Ericson et al.). The second Advanced Seminar (Sillen and Armelagos 1991) witnessed a significant increase in the sophistication and complexity of scientific research. Two papers focused on recent advances in the understanding of variations in stable carbon isotopes (Tieszen, van der Merwe) and one focused on variations in nitrogen isotopes (Ambrose) in terrestrial foodwebs. Two papers focused on

PREFACE

xv

theoretical aspects of diet reconstruction with stable isotopes (Schwarcz, Parkington), two focused on mineralogy and diagenesis of carbon isotopes in bone and tooth apatite (Lee Thorp and van der Merwe, Krueger) and two concentrated on mineralogy and diagenesis of trace elements in bone and tooth apatite (Lambert, Sillen and LeGeros). The most complex and advanced research presented at this seminar involved controlled diet experiments using isotopic analysis of individual amino acids (Hare et al.) and strontium isotope analysis for analysis of paleodiet and mobility patterns (Sealy et al.). Progress in these two areas has been slow, but the pace is now accelerating because of the availability of sophisticated instrumentation and improvements in analytical methods. One paper combined isotopic and elemental approaches to diet reconstruction in eastern North America (Buikstra and Milner). Outstanding questions about diagenesis and diet-tissue relationships were brought into sharper focus at this seminar. Sillen and colleagues (1991) summarize these issues in an important essay in American Antiquity titled “Chemistry and paleodiet: no more easy answers”, effectively outlining the agenda for the next Advanced Seminar. The third Advanced Seminar volume (Lambert and Grupe 1993), included the results of the first in-vivo experiments with rodents in which the isotopic composition of dietary macronutrients (protein, carbohydrate, lipid and fiber) was directly controlled (Ambrose and Norr, Tieszen and Fagre), providing the first effective tests of Krueger and Sullivan’s (1984) model of diet-tissue relationships. These experiments demonstrated conclusively that the majority of carbon atoms in consumer tissue proteins came from dietary protein rather than from fats and carbohydrates, while all dietary macronutrients contributed carbon to apatite carbonate. Controlled diet experiments for multiple trace elements were reported by Lambert and Weydert-Homeyer, and Hancock and colleagues evaluated baseline values and variations in trace elements in bone. Experiments and observations on the initial stages of diagenesis due to bone decomposition by bacteria and fungi were reported by Grupe and colleagues. Francalacci and colleagues reported on the reproducibility and reliability of trace element analysis of Medieval Italian human bones. Ericson reported on progress in identifying trace element diagenesis of bone and the use of barium for diet reconstruction. The potential of determining seasonality in trace element variation was discussed by Herrmann. Tuross discussed non-collagenous proteins and DNA in bone. Diet reconstruction and food class identification by chromatography and mass spectrometry of lipids in organic residues was discussed by Bethel and colleagues, and dental microwear was reviewed by Newesly. Katzenberg used the trophic level effect of nitrogen isotopes between mothers and their offspring to evaluate nursing and weaning patterns in human populations from Ontario. Two papers focused on reconstructing subsistence of human populations from South Africa (Lee Thorp et al.) and Ecuador (van der Merwe et al.) with carbon and nitrogen isotopes of collagen and carbonate. The great diversity of studies included in the third

xvi

PREFACE

Advanced Seminar volume reflects the greatly expanded awareness of the importance of diet reconstruction for understanding past human health and behavior. It also reflects the growing number of applications of stable isotope and trace element analyses of bones and teeth.

The foundations for chemical and isotopic techniques and new directions for paleodietary research were established at the second and third Advanced Seminars. Many of these new directions for research applications, such as determination of seasonality in diet, climate and microhabitat use, weaning and life history patterns, based on microanalysis of incremental growth structures, and analysis of the molecular components of proteins (amino acids) and lipids, have progressed slowly. Highly sophisticated instrumentation that permits analysis of microgram-sized samples is now available and the pace of research in this important and exciting area of inquiry has begun to accelerate since the Fourth Advanced Seminar. Diagenesis of bone

has become a lively field of inquiry on its own, with particular relevance to the field of chronometric dating, and workshops on diagenesis have been held at regular intervals (e.g., Schwarcz et al. 1989; Hedges and Van Klinken 1995;

Bocherens and Denys 1997, 1998).

The Fourth Advanced Seminar on Paleodiet Participants in the Fourth Advanced Seminar, as in preceding seminars, were invited based on their innovative contributions to paleodiet studies. Some of the participants have chosen to publish their work elsewhere and two chapters are by individuals who were unable to attend the seminar. The twelve papers in this volume include ten of the seventeen papers that were first presented in Banff plus the two additional papers. While the seminar began with theoretical papers we begin this volume with applied studies in which the various principles of stable isotope and trace element analyses are applied to questions of reconstructing past diet and subsistence. Some of the advances evident in the papers include the increasing use of bone and tooth carbonate as a source of stable carbon isotopes, an emphasis on ecological considerations as they relate to diet reconstruction and attempts to model the course of

ingested nutrients and their resulting chemical signatures. Katzenberg and colleagues report on a study of stable isotopes and trace elements in human bone and in meals reconstructed from historical sources for European settlers in Ontario, Canada. The use of paleodietary methods on historic samples allows the methods to be tested against historical records. Because some individuals are of known identity it is also possible to test for sex and age variation with greater precision than is possible in prehistoric samples. Van der Merwe, Tykot, Hammond, and Oakberg present the results of their studies of paleodiet and isotope ecology among the preclassic Maya of Belize. They report stable carbon and nitrogen isotope data for a wide range of terrestrial and marine faunal species. They also consider temporal and

PREFACE

xvii

spatial variation in the region by comparing their work to other studies done on neighboring peoples. Van Klinken, Richards, and Hedges present data

on carbon and nitrogen isotope variation among European samples. They attribute the observed variation to both climate and diet. European populations have received less study than those from the Americas since European agriculture was based on plants. Nitrogen isotope variation relates to the use of marine foods and also to the relative amount of animal protein in the diet. Collectively, these first three chapters provide examples of research on recent prehistoric and early historic remains from three diverse regions. They also exemplify the importance of an ecological approach to diet reconstruction, that is, considering the foods as well as the net result as reflected in the bone chemistry.

Bocherens reports on carbon and nitrogen stable isotopes in Pleistocene mammals with varying diets. His samples are drawn from France, the United Kingdom, Siberia and Alaska. Variation is observed in collagen stable isotope

values as well as in the difference between carbon isotope ratios in collagen versus biological apatite. Lee-Thorp also addresses the problem of preservation and recovery of biogenic carbon from fossilized bones and teeth. She compares carbon isotope ratios from bones and teeth of browsers and grazers from both fossil and modern contexts and concludes that enamel is less subject to post-mortem alteration even in fossils of over one million years of age. Schoeninger and colleagues also make use of tooth enamel. However, their interest is in reconstructing past climate using oxygen isotopes from enamel phosphate. The ability to chart climatic variation over the past in East Africa provides information on past habitats and the species available to early hominids. Papers by Bocherens, Lee-Thorp, and Schoeninger and colleagues demonstrate the types of information that are available in the mineral component of bones and teeth. Pfeiffer and Varney explore differential preservation of bones within a relatively small historic cemetery from Ontario. They find that the state of preservation of histological structure of bone does not correlate with preservation of collagen. Burton and Price address the subject of trace element analysis of bone mineral. These authors discuss the fact that trace element analysis has generally fallen out of favor even though two elements, strontium and barium, do provide information about paleodiet. They discuss the problems

with earlier expectations regarding the relationship between dietary trace elements and those same elements in bone and present data from their ecological studies in Wisconsin. Grupe, Blazer and Turban-Just approach diagenesis

from an experimental perspective and discuss the effects of microorganisms on bone diagenesis and the implications for stable isotope analysis. They injected bacteria found in soils into modern bone in order to simulate microbial collagen breakdown. Their work provides an estimate of potential shifts in and in collagen due to such degradation in archaeological bone.

xviii

PREFACE

The next two chapters address modeling of dietary components and resulting stable carbon isotope ratios in collagen and biological apatite.

Schwarcz explores the reasons for differences in

in collagen and apatite

by reviewing relevant biochemical principles of protein synthesis. He considers these principles in light of the model proposed by Krueger and Sullivan

(1984) and subsequent controlled feeding experiments conducted by Ambrose and Norr (1993) and Tieszen and Fagre (1993). Hedges and van Klinken present an equation for modeling the input of dietary components to specific body tissues. They make extensive use of the data provided by the controlled feeding experiments of Ambrose and Norr (1993) and Tieszen and Fagre (1993). Nitrogen isotopes are also considered with respect to metabolism and ecology. This chapter and that by Schwarcz illustrate some of the complexities of paleodietary interpretation with respect to stable isotope analysis. Such models lead researchers to consider the range of biochemical processes that

may affect resulting stable isotope ratios in the tissues analyzed for paleodiet studies. Ambrose focuses on stable nitrogen isotope ratios and reports the results of controlled diet and climate experiments on rats. He explores trophic level differences and heat stress and their effects on stable isotopes of nitrogen. As with the preceding chapters by Schwarcz and Hedges and van Klinken, Ambrose explores the various biochemical principles that explain why variation is expected under these circumstances.

All chapters in this volume have been reviewed by the editors, one external reviewer chosen by the editors, and two external reviewers chosen by the editorial board of the series on Advances in Archaeological and Museum Science. We are grateful to these reviewers for their careful reading of the contributions and for their suggestions for revisions. The editors were also the organizers of the Fourth Advanced Seminar.

Acknowledgments We gratefully acknowledge support from the Wenner-Gren Foundation for Anthropological Research, The L.S.B. Leakey Foundation and the Social Sciences and Humanities Research Council of Canada who provided financial support for the seminar. Our institutions, the Department of Archaeology, University of Calgary and the Department of Anthropology, University of Illinois, also made contributions. Finally, we thank the contributors for their patience and their hard work.

STANLEY H. AMBROSE URBANA, ILLINOIS M. A NNE KATZENBERG CALGARY, ALBERTA

PREFACE

xix

REFERENCES Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relationship of the carbon isotope

ratios of whole diet and dietary protein to those of bone collagen and carbonate. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin: Springer-Verlag: 1–37.

Bocherens, H. and Denys, C. 1997 and 1998 Third international Conference on Bone Diagenesis. Bulletin de la Société Géologique de France 168 and 169. Boutton, T.W. 1991 Tracer studies with 13C-enriched substrates: humans and large animals. In Coleman, D.W. and Fry, B., eds., Carbon Isotope Techniques. Academic Press, San Diego: 219–242. Brown, A.F.B. 1973 Bone Strontium Content as a Dietary Indicator in Human Skeletal Populations. Unpublished doctoral dissertation, Department of Anthropology, University of Michigan, University Microfilms. Coleman, D.C. and Fry, B. 1991 Carbon Isotope Techniques. Academic Press, San Diego. Hedges, R.E.M. and van Klinken, G.-J. 1995 Editorial (Special issue on bone diagenesis). Journal of Archaeological Science 22(2): 145. Krueger, H.W. and Sullivan, C.H. 1984 Models for carbon isotope fractionation between diet and

bone. In Turnlund, J.F. and Johnson, P.E., eds., Stable Isotopes in Nutrition. ACS Symposium Series 258. Washington DC, American Chemical Society: 205–222. Lambert, J.B. and Grupe, G. 1993 Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag. Price, T.D. 1989 The Chemistry of Prehistoric Bone. Cambridge, Cambridge University Press.

Schwarcz, H.P., Hedges, R.E.M. and Ivanovich, M. 1989 Editorial comments on the First International Workshop on Fossil Bone. Applied Geochemistry 4(3): 211–213. Sillen, A. and Armelagos, G. 1991 Introduction to this issue. Journal of Archaeological Science 18: 225–226. Sillen, A., Sealy, J.C. and van der Merwe, N.J. 1989 Chemistry and paleodietary research: No more easy answers. American Antiquity 54: 504–512. Tieszen, L.L. and Fagre, T. 1993 Effect of diet quality and composition on the isotopic composition of respiratory bone collagen, bioapatite and soft tissues. In Lambert, J.B. and Grupe, G., eds, Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, SpringerVerlag: 121–155.

van der Merwe, N.J. and Vogel, J.C. 1978 13C content of human collagen as a measure of prehistoric diet in Woodland North America. Nature 276: 815–816. Vogel, J.C. and van der Merwe, N.J. 1977 Isotopic evidence for early maize cultivation in New

York State. American Antiquity 42: 238–242. Wolfe, R.R. 1984 Tracers in Metabolic Research: Radioisotope and Stable Isotope/Mass Spectrometry Methods. A.R. Liss, NY.

This page intentionally left blank

Contents

Chapter 1 • Bone Chemistry, Food and History: A Case Study from 19th Century Upper Canada . . . . . . . . . . .

1

M. Anne Katzenberg, Shelley R. Saunders, and Sylvia Abonyi Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Chapter 2 • Diet and Animal Husbandry of the Preclassic Maya at Cuello, Belize: Isotopic and Zooarchaeological Evidence . . . . . . . . . . . . . . . . . . . .

1 2 3 4 5 14

18 19 19

23

Nikolaas J. van der Merwe, Robert H. Tykot, Norman Hammond, and Kim Oakberg

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction: Preclassic Cuello . . . . . . . . . . . . . . . . . . . . . . . . . . . . Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cuello Isotopic Ecology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi

23 24 26

26 27

xxii

CONTENTS

Deer and Dogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Maya Diets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28 28 32 35 36

Chapter 3 • An Overview of Causes for Stable Isotopic Variations in Past European Human Populations: Environmental, Ecophysiological, and Cultural Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

Gert J. van Klinken, Mike P. Richards, and Robert E.M. Hedges Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Causes for Isotopic Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison of Animal and Human Data from Europe . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39 40 41 52 57 58

Chapter 4 • Preservation of Isotopic Signals (13C, 15N) in Pleistocene Mammals . . . . . . . . . . . . . . . . . . . . . . . . .

65

Hervé Bocherens Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biogenic Isotopic Signals Used to Assess the Preservation of Isotopic Signals in Fossil Samples from C3 Plant Food Webs . . . . . Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results on Modern Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 5 • Preservation of Biogenic Carbon Isotopic Signals in Plio-Pleistocene Bone and Tooth Mineral . . . . . . .

65

66 67 69 72 72 81 84

89

Julia A. Lee-Thorp Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemistry of Biological Apatites . . . . . . . . . . . . . . . . . . . . . . . . . . .

89 90 90

CONTENTS

xxiii

...........................................

100 103 106 110 111

Chapter 6 • Tooth Oxygen Isotope Ratios As Paleoclimate Monitors In Arid Ecosystems . . . . . . . . . . . . . . . . . . .

117

Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

References

Margaret J. Schoeninger, Matthew J. Kohn, and John W. Valley 117 118

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Animal Samples, Characteristics of Sampling Locality and Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

124 127 136 137

Chapter 7 • Quantifying Histological and Chemical Preservation in Archaeological Bone . . . . . . . . . . . . .

141

Susan Pfeiffer and Tamara L. Varney Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background to the Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

141 142 143 147 149 152 156 156 156

Chapter 8 • The Use and Abuse of Trace Elements for Paleodietary Research . . . . . . . . . . . . . . . . . . . . . . . . .

159

James H. Burton and T. Douglas Price Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Use of Non-alkaline Earth Elements . . . . . . . . . . . . . . . . . . . . . . . . . Strontium as a Measure of the Plant/Meat Ratio . . . . . . . . . . . . . . . .

159 160 161 163

xxiv

CONTENTS

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

166 167 167

Chapter 9 • Modeling Protein Diagenesis in Ancient Bone: Towards a Validation of Stable Isotope Data . . . . . . .

173

Gisela Grupe, Astrid Balzer, and Susanne Turban-Just Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

173 174 175 178 182 186 186

Chapter 10 • Some Biochemical Aspects of Carbon Isotopic Paleodiet Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . .

189

Henry P. Schwarcz

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

189 190 190 199 207 208 208

Chapter 11 • “Consider a Spherical Cow . . .”—on Modeling and Diet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

211

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Some Basic Biochemical Principles Relevant to Paleodiet . . . . . . . . . . Implications for Isotopic Paleodiet Studies . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Robert E.M. Hedges and Gert J. van Klinken Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

211 212

The Representation of Isotopic Fractionation Between Diet and

Body Component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flow Models of Feeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evaluating an Explicit Flow Model for the “Protein Routing” Question . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . More Speculative Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

214 221 226 231

CONTENTS

xxv

Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ...........................................

236 237 237

Chapter 12 • Controlled Diet and Climate Experiments on Nitrogen Isotope Ratios of Rats . . . . . . . . . . . . . . . . .

243

Stanley H. Ambrose Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

243 244 248 250 256 257 257

About the Editors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

261

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

263

This page intentionally left blank

Chapter

1

Bone Chemistry, Food and History: A Case Study from 19th Century Upper Canada M. ANNE KATZENBERG, SHELLEY R. SAUNDERS, AND SYLVIA ABONYI ABSTRACT Dietary indicators in preserved human bone were studied and compared to reconstructed diet for a sample of individuals who lived in southern Ontario in the 19th century. Stable isotopes of carbon were analyzed in preserved bone collagen and in foods. Foods were prepared following recipes in historical documents and using traditional cookware. Raw ingredients and finished baked goods and stew were analyzed. Results indicate small differences in between raw and cooked foods, on the order of around 1.2‰ or less. A comparison of the mean value of the reconstructed diet with the values in human bone collagen provides an estimate of diet-to-collagen spacing of 5.6‰ before correction for the difference between the present and past foodweb carbon isotopic composition. A number of trace elements were analyzed in food and in bone mineral. Trace element levels in foods varied considerably in some cases between raw and cooked foods. Most notably, iron was absorbed from the cooking pot and from water, particularly in foods such as barley and beef. Strontium values, which are higher in raw vegetables than in uncooked beef, homogenized in the stew such that cooked meat and vegetables, were indistinguishable. Biogeochemical Approaches to Paleodietary Analysis , edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

1

2

M. ANNE KATZENBERG ET AL.

isotope nitrogen exhibit little betweenStable the sexes andvalues amongforallboth agecarbon classesand except for infants. Infantvariation values are elevated due to the effect of breast-feeding. These findings match historic documents indicating that while food was plentiful, there was little variation in the daily fare.

INTRODUCTION Most paleodiet studies based on bone chemistry have been directed toward one of two approaches: reconstructing the diet of prehistoric peoples (reviewed by Katzenberg 1992; Pate 1994; Schoeninger and Moore 1992; Schwarcz and Schoeninger 1991), or carrying out experiments on laboratory or free-ranging animals, to determine how diet leaves a chemical signature in

human tissues that might be preserved in the archaeological record (Ambrose and Norr 1993; Tieszen and Fagre 1993; Lambert and Weydert-Homeyer 1993a, 1993b). When a large skeletal sample from a 19th century Anglican Church Cemetery became available for study, we saw the opportunity to combine these two approaches in a somewhat controlled situation involving buried human remains. The situation is somewhat controlled in the sense that these people lived fairly recently and there are historic records describing many aspects of their lives. However, like many historical records, there is less information on day to day activities, and the activities of women, than there is about politics, settlement data and farming practices. We are particularly fortunate because two women who lived in the Belleville area around the time that the cemetery was in use wrote books about their lives (Traill 1836, 1855; Moodie 1853). Their books were specifically directed to other women who were considering migrating to Ontario from Great Britain and they contain information about food, cooking, health and sickness in 19th century southern Ontario (known at the time as Upper Canada). In addition, letters written by William Hutton, an Irish immigrant in Belleville in the 1830’s, to his mother back in Ireland survive to this day and have been compiled into a book (Boyce 1972). Hutton wrote some detailed accounts of the food consumed by his family as well as accounts of the production and abundance of agricultural products in the region from the 1830’s through the 1860’s. In 1989, the parishioners of St. Thomas’ Anglican Church in Belleville, Ontario contracted an archaeological consulting firm to excavate the church cemetery so that they could build an addition onto the church. The parish agreed to allow study of the human remains for a period of one year before reburying the individuals in a new location. One of us (S.R.S.) oversaw all studies of the human remains. These studies have included determination of age at death (Dudar et al. 1993; Rogers 1991; Saunders et al. 1992; Saunders et al. 1993; Saunders et al. 1995a, b), estimation of sex (Rogers 1991; Rogers and Saunders 1994), pathology (Jimenez 1991; Jimenez 1994; Saunders et al.

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

3

1993; Saunders et al. 1995), osteometry (Saunders and Hoppa 1995; Saunders and Sawchuk 1995) and growth and development (Saunders and Hoppa 1993; Saunders et al. 1993).

We proposed to study diet and health by combining bone chemistry and histomorphometry. Diet would be determined by analysis of stable isotopes of carbon and nitrogen in bone protein and some preserved hair. In addition, trace elements would be quantitatively analyzed in preserved bone mineral. Abonyi (1993) participated in the study by reconstructing the diet from historical sources and analyzing various foods. Having analyzed human tissues for stable isotopes and trace elements, and foods for the same variables, we hoped to learn more about 19th century diet in southern Ontario, and at the same time, learn more about paleodiet

reconstruction. This paper addresses the following issues: 1) determination of the diet among 19th century pioneers in southern Ontario from historical sources, 2) chemical analysis of food including raw ingredients as well as prepared recipes, 3) analysis of stable isotopes of carbon and nitrogen and selected trace elements in human bone and hair, 4) determination of the relationships between

food values and tissue values, and between bone and hair, and 5) detection of sex and or age differences in bone chemistry. Another aspect of the study, determination of infant feeding practices using stable isotopes of nitrogen, is addressed in a paper by Herring and colleagues (1998).

MATERIALS The excavated human remains from St. Thomas’ Church cemetery, Belleville consist of 579 coffin burials dating from A.D. 1821 to A.D. 1874. Two fires at the church, one in 1876 and another in 1974, destroyed some of the grave markers. There was no cemetery plan indicating who was buried in each plot. Some coffin plates were preserved so there is a subsample of 80 individuals who are positively identified, including their age at death, sex and family affiliation. Parish records indicated that there were 1,564 individuals interred in the cemetery. The recovered sample makes up 37% of the total recorded interments. In addition to the studies done on the skeletons (cited in the introduction), Dr. Ann Herring has worked with the parish records and the Belleville census data (Herring et al. 1991; Saunders et al. 1995a, b). Her data on infant mortality is particularly interesting with respect to the skeletal data on stable nitrogen isotopes and weaning (Herring et al. 1998). The parish records include parish membership as well as vital events such as birth, marriage and death. One hundred and eighty-four causes of death are mentioned. Rib samples were obtained from 439 individuals. Preserved hair samples were obtained from 24 of these individuals. Sex and age at death

4

M. ANNE KATZENBERG ET AL.

were determined previously (see references in introduction). Information on positive identification of the subset of known individuals was also provided.

For the diet study, Abonyi obtained organically grown foods (whenever possible) from the Belleville area. Various recipes were prepared following historic cookbooks and newspaper articles. All but one of the recipes was prepared in an iron cooking pot from the 19th century, which was provided by the Bradley House Museum in Clarkson, Ontario.

METHODS

Human Bone—Extraction of the Organic Fraction and Amino Acid Analysis Bone protein was extracted following the method described by Sealy (1986). Bone chips were demineralized in a weak HCl solution, then soaked in 0.1 M NaOH to remove base-soluble humic substances. Remaining material, which is mainly collagen, but includes non-collagenous proteins, was freeze-dried. That material was analyzed for its amino acid composition following hydrolysis in HCl. All analyses were carried out on a Beckman 6,300

Amino Acid Analyser in the Protein Sequencing Facility in the Department of Medical Physiology, University of Calgary. Fifteen individuals, ranging in age from stillbirth to 32 years, were included in this aspect of the research. Because the purpose of this analysis was to determine if bone preservation was uniform among the younger individuals in the sample, the design is heavily weighted toward younger individuals. Specifically, six individuals under one year of age, four individuals aged one to two years, one aged two and one half years and four aged 25 to 35 years were included. Differential preservation of specific amino acids can result in variation in stable isotope ratios (Hare et al. 1991). The purpose of the amino acid analysis was to determine whether this is a factor in any variation relative to age at death.

Stable Isotope Analysis Stable isotope analyses of the organic fraction of bone and of food samples was carried out on a Micromass Prism Mass Spectrometer in the Stable Isotope Laboratory, Department of Physics, University of Calgary, under the direction of H.R. Krouse. Collagen samples were combusted in a Carlo Erba gas analyser which provides information on the carbon and nitrogen content of the samples and introduces or gases into the mass spectrometer for analysis of nitrogen or carbon stable isotopes, respectively.

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

5

Food Preparation Raw foods were freeze-dried and analyzed for carbon isotopes using mass spectrometry. Cooked foods were prepared following historic recipes, then were freeze-dried prior to analysis. For the trace element analysis, foods (both raw and cooked) were wet ashed using nitric acid in Teflon lined pressure vessels and digested in a CEM Microwave oven. Analysis of Sr, Zn, Fe, Ca and

Mg was performed using Atomic Absorption Spectrometry in the Department of Geology, University of Calgary.

Human Bone—Trace Element Analysis Nineteen bone samples were prepared for analysis of the trace elements strontium (Sr), rubidium (Rb), and zinc (Zn). The outer surface of each bone was removed with an aluminum oxide sanding wheel attached to a Dremel tool and the bone was soaked overnight in a weak acetic acid solution (Krueger and Sullivan 1984, Price et al. 1992). After rinsing to neutrality, the bone was dried then crushed in a mill. Bone powder was dry ashed in a muffle furnace

at 700°C for 18 hours. Bone ash was pressed into pellets for analysis by x-ray fluorescence spectrometry. Analyses were carried out in the Department of Geology, University of Calgary.

RESULTS

Amino Acid Analysis of Bone Collagen The fifteen samples analyzed for amino acid composition were compared to a human type I collagen standard. There are no significant differences in amino acid composition among these samples, which range in age from stillbirth to adult. Analytical error is All samples show the typical composition of type I collagen (Table 1.1). Based on these results there is no reason to suspect differential preservation of specific amino acids in the small bones of infants in comparison to bones of adults and older children.

Stable Isotope Analysis—Food Raw ingredients and cooked recipes were analyzed for stable isotopes of carbon (Abonyi 1993). To our knowledge, no one has determined if there are differences in the stable isotope values of raw and cooked foods, however Hastorf and DeNiro (1985) did compare plant remains that were heated to those that had not been heated. They found no significant difference in isotope values due to heating. Marino and DeNiro (1987) studied the effects of heating (boiling and roasting) on several types of plants to determine if cellulose

6

M. ANNE KATZENBERG ET AL.

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

7

isotope values are affected. Maize cobs, sunflower seeds, agave leaves and Pachyrrhizus tubers showed variation of less than We also analyzed mixed foods as they would have been consumed by the people buried in the cemetery. Results suggest that heating has a small effect on values and that, in some cases, mixing of ingredients has an additional effect. Differences are less than and are not always in the same direction (Table 1.2). For baked goods, cooking always results in slightly more negative values. When wet and dry ingredients are mixed, reactions take place in the dough. is released from baking powders and from yeast fermentation. Baking powder and baking soda both contain sodium bicarbonate, which is derived from natural sources such as trona ore and alkaline brines. These sources of carbonate are enriched in 13C and therefore the loss of this source in the baking process would result in an isotopically lighter finished product. Yeast feeds on carbohydrates and liberates During baking, and water vapor are lost. If cane sugar is used as a sweetener, as it was in these experiments, and yeast feeds on that sugar, then the lost in baking will be more enriched in the heavier isotope, resulting in an isotopically lighter baked product as compared to the dough. For the stew ingredients, cooking results in more enriched in carrots but in more depleted for barley, onions and potatoes. Other ingredients vary only slightly and the variation is within the precision of the analysis Each ingredient was taken from the finished stew for analysis so it is possible that there was some exchange in the cooking pot. However there is still an overall shift, with more

8

M. ANNE KATZENBERG ET AL.

negative relative to the raw ingredients. From these data, it can be concluded that cooking may have a small effect on carbon isotope values, but the directional change is variable and should not significantly affect dietary interpretations. Baking also produces a shift, which can be explained by the chemical changes that occur with leavening agents.

Stable Isotope Analysis, Bone Organic Fraction Four hundred and thirty-four samples were analyzed for stable isotopes of carbon. One hundred and thirty-two samples were analyzed for stable isotopes of nitrogen. The objectives of the carbon isotope analysis were: 1) to compare the historical information regarding the proportion of plants in the diet to the evidence from the skeletal remains, 2) to compare the values of foods to those of bone and determine the difference between diet and collagen, and 3) to see if there were changes in the consumption of plants over the time that the cemetery was in use. The first two questions are addressed in the following section. The third question was addressed, using a subsample of 45 individuals of known identity, in a study by Saunders and colleagues (1997) in which carbon isotope values are compared to caries incidence and historical evidence for an increase in the consumption of sugar over time. A weak, nonsignificant positive correlation was found with relative

to year of death. This is not surprising given that sugar would have been used as a minor ingredient, and given experimental evidence that carbon from carbohydrates will contribute less to carbon in collagen, than that obtained from dietary protein (Ambrose and Norr 1993; Krueger and Sullivan 1984; Tieszen and Fagre 1993).

A smaller number of individuals, 132 samples, was analyzed for stable isotopes of nitrogen. There were two objectives to the nitrogen isotope analysis: 1) to determine the amount of animal protein in the diet and to look for variation between the sexes and 2) to determine when infants were weaned from the breast. These objectives were addressed by selecting at least ten individuals, when available, from a number of age classes. Figure 1.1 illustrates the age distribution of the sample. Table 1.3 shows the results of these analyses. Individual values are provided in Herring et al. (1998). Mean for all samples is Mean values by age group for samples in which was also analyzed are presented in Table 1.3. Mean for 132 samples is The greater range of variation in nitrogen isotope values is the result of the effect of nursing on the values of infants. Analytical error is for and for

Estimate of

for 19th Century Diet in Belleville

On the basis of historical sources, the foods chosen and the recipes prepared for this study were those most commonly consumed during the period

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

9

Figure 1.1. Age distribution of the samples analyzed for stable isotopes of nitrogen.

of use of St. Thomas Anglican Church cemetery (sources include: Boyce 1972;

Canniff 1869; Moodie 1853; Proudfoot 1915; Talbot 1824; Tivy 1972; Traill 1855, 1836). Briefly, meat (primarily beef, mutton and pork), bread (whole wheat and com bread), and vegetables (potatoes, turnips, carrots) were the dietary staples. Sweetened breads and cakes were also commonly consumed. Although three meals a day were generally served, the menu for each was not particularly distinct. Most meals were simply prepared. Meat and vegetable items were often just boiled or fried. Food was generally prepared in iron cookware, either over a fire or in a cook stove. The major sources of variation in are the two plants, maize and sugar cane. Most other foods are and animals used for food were mainly feeding on plants. Maize and sugar cane (in the form of molasses) are found

10

M. ANNE KATZENBERG ET AL.

mainly in the corn bread and porridge. Canadian settlers used maple sugar, which is as well as cane sugar. Saunders et al. (1997) describe the increase in use of cane sugar relative to maple sugar in the second half of the 19th century in Upper Canada. Using diaries, census data, and emigrant guides as their primary sources, Kenyon and Kenyon (1992) compiled tables of the types and amounts of foods consumed on a daily basis during the nineteenth century in southern Ontario. These tables were adapted in this study to assess the relative contributions of meat, vegetables, and baked items in the adult diet. Based on the values of various recipes that incorporate the foods most commonly consumed, and the table of relative contribution of each to the diet, Abonyi (1993) derived an average for diet by multiplying the of each food group (meat, baked goods and vegetables) by their relative contribution to the diet (Fig. 1.2). The resulting estimated for the total diet is

Estimated Diet to Collagen Spacing In order to estimate the spacing between the of food and that of bone collagen, Abonyi (1993) subtracted the mean of bone collagen from the of the diet with a resulting value of This is very close to the value proposed in 1978 by Vogel for large mammals. Vogel and van der Merwe (1977) proposed a value of +5.1 for humans based on analyses of Archaic skeletons from North America assumed to have subsisted on an exclusively diet. The of atmospheric has decreased by around over the last two hundred years as a result of the burning of isotopically light fossil fuels (Marino and McElroy 1991). Therefore our estimated of diet, which is

Figure 1.2. Estimated of Belleville diet based on the percent contribution of baked goods (20%), meat (10%) and vegetables (70%) to the whole diet. Experimentally determined for each food group is indicated.

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

11

derived from modern foods, may be slightly more negative than that of the actual diet consumed in the mid 1800s. Thus the actual spacing may less than In a table complied by Ambrose (1993), free-ranging browsers and grazers have collagen to diet spacing ranging from 5.0 to 6.0 whereas laboratory mammals raised on controlled laboratory diets range from 0.5 for gerbils to 4.6 for mice. This variation may be due to variation in the experimental diets and particularly due to differences in of the protein and non-protein components in those diets (Ambrose and Norr 1993).

Stable Isotope Values in Bone and Hair Twenty-four hair samples were analyzed for stable isotopes of carbon and nitrogen. All of those same individuals were analyzed for

in bone colla-

gen and nine were analyzed for The difference between in bone collagen and hair (mean collagen value minus mean hair value) is for the 24 samples. The difference between collagen and hair for is That is, hair is heavier than bone collagen. In his carbon isotope studies of tissue differences in large mammals, Vogel (1978) found the difference in

between hide (skin) and bone collagen to be around White and differences between bone collagen and hair from 0.5 to for Nubian mummies. They attribute the range of variation to seasonal variation in diet. Seven of the nine individuals analyzed for both hair and bone are children under the age of five years. The hair and collagen values are closer together in the youngest individual (6 months of age) where the hair is slightly lighter than the bone collagen. Spacing between collagen and hair is slight for the one-year-old, then from two years on the spacing between collagen and hair is very similar (Fig. 1.3). Schwarcz (1994) report

Figure 1.3.

determined from bone collagen and hair for eight individuals.

12

M. ANNE KATZENBERG ET AL.

Trace Element Results—Food Raw and cooked food samples were analyzed quantitatively for iron, strontium, magnesium and zinc. Three potential sources of elemental increase from raw to cooked foods were identified. First, water was used to either clean the cooking vessel between uses or as an ingredient in some of the recipes. The water source was hard water from a deep water well, which contains calcium and magnesium bicarbonates, but little iron or zinc. Second, minerals may leach out of the walls of the cooking vessel (in this case, an iron pot) and be absorbed into the foods being cooked. Finally, minerals leached from one ingredient may be absorbed by another (this is in particular reference to the beef stew). Elemental loss may be accounted for by leaching or volatilization. Data are provided in the more detailed study by Abonyi (1993).

Magnesium Items that were baked increased significantly in magnesium concentration from raw to cooked form. The source may be the iron pot or the calciumand magnesium-rich water in which the pot was washed between uses. Vegetables such as carrots and turnips, which were components of a beef stew cooked in water in the iron pot, acquired magnesium. Barley, another ingre-

dient in the beef stew, increased in magnesium by approximately 20%. Beef and potatoes both lost magnesium (33% and 32%, respectively) in the cooking process. The cooking effects do not obscure the difference in magnesium levels between meat and vegetables. The range of concentration for baked items, however, falls between the ranges for meat and vegetables, overlapping with

the upper end of the vegetable range and the lower end of the meat range.

Zinc The effect of cooking on zinc concentration was highly variable, exhibiting patterns of gain and loss similar to magnesium. The zinc levels in beef remained three times higher than in all the baked foods and vegetables even after losses incurred by cooking. Iron Iron concentration increased significantly in all foods from raw to cooked form, roughly as a function of cooking time (Fig. 1.4a). The baked items, which require thirty minutes or less of cooking time, increased in iron by

15–50%. The beef stew ingredients, which simmered in excess of one hour, essentially doubled in iron concentration (with the exception of barley, in which the post cooking levels were 5 times higher than in raw barley). The spread between meat and the baked or vegetable items is retained after

Figure 1.4a. Iron concentration in raw and cooked foods. For braked foods, comparisons are between dough and baked product. For stew ingredients, raw foods are compared to foods cooked

together in an iron pot. Figure 1.4b. Strontium concentration in raw and cooked foods (as described in fig. 1.4a).

14

M. ANNE KATZENBERG ET AL.

cooking, such that iron in meat is six times higher than in any of the vegetables or baked items.

Strontium While differences in the concentration of strontium were evident between raw meat and vegetables, these differences were obscured by cooking. Strontium levels in carrots and turnips were indistinguishable from beef after cooking (Fig. 1.4b).

Trace Element Results—Hair Trace element results for hair samples were widely variable and appear to reflect contamination. They are not reported here.

Trace Elements—Bone Bones of 19 individuals were analyzed for strontium, rubidium and zinc. The number of samples was limited by the availability of bone after the stable isotope analyses were completed. Strontium was analyzed in order to test for trophic level, and to compare to other results obtained in the region on prehistoric peoples (Katzenberg 1984). Rubidium is not expected in human bone, so its presence acts as a measure of contamination. The use of zinc as a paleodietary indicator has been questioned recently (Ezzo 1994) and we were interested to see if there was any relationship between zinc content in food and bone. Mean strontium value is with a range of 126 to 292 ppm and a standard deviation of 41 ppm. These values are similar to those obtained from prehistoric and protohistoric Ontario Iroquois from the same region (Katzenberg 1984). This is not surprising since strontium levels in bone are largely dependent on background levels in soils and waters of a region. Of the three elements for which analyses were carried out, only strontium is thought to have potential as a dietary indicator (reviewed by Sandford 1992; Ezzo 1994; Burton and Price, this volume). Mean Rb for 19 samples is 6 ppm with a standard deviation of 0.7 ppm. Mean Zn for 19 samples is 571 ppm with a standard deviation of 220 ppm. The range for zinc is very large with a minimum value of 267 and a maximum value of 1,144. This range suggests that there is little to be learned regarding diet or physiology. Trace element results for bone samples are presented in Table 1.4.

DISCUSSION

and Age Fogel and colleagues (1989) first reported the behaviour of nitrogen isotope values in nursing infants and their mothers. They compared

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

15

fingernails of nursing infants and their mothers through the duration of breastfeeding and on into the time of weaning (i.e., when nursing was reduced and other foods were introduced into the infant diet). Their study demonstrated that there is a trophic level shift in nursing infants, who are feeding on mothers’ tissues in the form of breast milk. values are elevated in the nursing infant, then fall to values in line with those of mother as the infant is weaned. The same authors also provided prehistoric evidence for this effect using bone collagen from two skeletal samples. Subsequently, a number of researchers have demonstrated similar patterns in prehistoric and early historic samples (Katzenberg et al. 1993; White and Schwarcz 1994; Katzenberg and Pfeiffer 1995; Schurr 1998 and reviewed by Katzenberg et al. 1996). Herring and colleagues (1998) studied weaning in the St. Thomas’ Church sample from three different perspectives, values, parish records and skeletal age at death data. It appears that the infants from 19th century Belleville, as represented in this sample, were introduced to other foods at around five months of age, but that breast milk remained the main source of protein until around 14 months of age (Herring et al. 1998).

Age and Sex Differences in Stable Isotope Values Summary data by age and sex are presented in Table 1.5. The largest difference in values occurs between infants and very young children, and

16

M. ANNE KATZENBERG ET AL.

those past the weaning period. For that reason, the t-test was used to compare individuals from birth to two years of age to those older than two years. For 39 infants and 81 children and adults were compared. Results of the ttest are t = 2.1 and p = 0.039. For 50 infants and 82 children and adults were compared with t = 7.7 and p = 0.000. Infants have significantly higher

than do children and adults. For the results are also significant at the 95% confidence level. is more variable for infants and there is evidence for a smaller trophic level effect for stable carbon isotopes, on the order of l‰ (DeNiro and Epstein 1978). Individuals over 18 years of age were compared for sex differences in and values. Forty-six males have mean of –19.4‰ and 39 females have mean

of –19.7‰ (t = –2.18; p = 0.032). For

12 males

have a mean value of 10.2‰ while 11 females have a mean value of 10.6‰ (t = 1.77; p = 0.093). It is apparent that sex differences are very minor. Male values are slightly higher (more enriched in the heavier isotope while female values are slightly higher than those of males. These data do not suggest markedly different diets between the sexes. The carbon isotope data suggest that males may have consumed more beef however this is not sup-

ported by the nitrogen isotope data. Alcohol consumption was considered as a possible source of variation in carbon isotope values between the sexes. In an unrelated study, Krouse (personal communication) analyzed different types of whiskey to determine if they were made from wheat, corn or other grains. Rum, bourbon and some vodka are made from corn and have values ranging from –14 to –11‰. Whiskey made from wheat or rye has values around –25 to –26‰. Historical sources indicate that distilleries in Ontario during the 19th century primarily made use of wheat for making whiskey. This fact, plus the presumably small amount consumed relative to food, would argue against attributing carbon isotope differences to differential alcohol consumption.

Trace Elements Results of analyses of raw and cooked foods demonstrate that the greatest change appears to be related to iron absorption from the iron cooking pot.

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

17

Figure 1.5. Trace element concentrations of zinc, strontium and rubidium in bone.

Foods such as barley and beef, that absorb water while cooking, showed the greatest increase in iron content from raw to cooked. Calcium also increased from raw to cooked foods. Other elements gave mixed results. Strontium did not strictly behave like calcium. Both zinc and magnesium were variable and decreased in meat with cooking. Barley increased in all elements tested, indicating that uptake is from water, the cooking vessel and from other foods, such as beef. These results indicate that the differences between elemental concentrations in raw versus cooked foods are great enough to warrant using food samples as they would have been prepared and eaten in human diet reconstructions. Human bone strontium levels (Fig. 1.5) are within the range that is

expected for that region (Katzenberg 1984). The sample size is small and uneven with only three females and one subadult so it is not possible to comment on sex or age differences except to say that the highest Sr content was found in the sample from a child aged around years.

Trace Elements and Food Processing While the impact of food processing on levels is not significant enough to pose problems for paleodietary reconstruction, the same cannot be

18

M. ANNE KATZENBERG ET AL.

said for trace elements. In some cases it significantly decreases, if not entirely

eliminates, the predicted spread in elemental concentration between foods that form the basis for the application of the techniques to paleodietary reconstruction. In this study, the greatest change appears to be related to iron absorption from the iron cooking pot. Additionally, the water with which cooking vessels or foods are washed, and the water in which foods are cooked are major sources of change in elemental concentrations. Foods that absorb significant

amounts of water (e.g. barley and beef) during cooking show the greatest increases. In short, the results of this aspect of the study indicate that the differences between elemental concentrations in raw versus cooked foods are great enough to warrant collecting information on values for dietary items as they would have been eaten, rather than relying on raw food data. CONCLUSIONS

The diet of the 19th century residents of Upper Canada was determined from historical sources and was reproduced in order to carry out chemical analysis. Stable carbon isotope analysis of food and human bone demonstrates that the spacing between the food eaten and the bone collagen is around 5.6%o. The value may vary slightly from this estimate since the latter is based on a reconstructed diet and a large number of bone samples, which exhibit a small amount of variation. Nevertheless, this empirically derived result agrees well with estimates from field (Vogel 1978), and laboratory studies (reviewed in Ambrose 1993). While there are some shifts in values between raw unmixed ingredients and finished recipes, the shifts are not of great enough magnitude to result in different interpretations of paleodiet. Baked goods become slightly depleted due to the actions of baking powder and yeast. Stew ingredients vary in direction and magnitude of changes in Differences in trace element levels in raw versus cooked foods are much greater than differences in stable carbon isotopes. Absorption of iron from the cooking vessel and water and mixing of elements with mixing of ingredients obscures the values obtained from raw and unmixed foods. Any attempt to link specific foods to trace element levels in bone should be based on food as it is consumed. Of the three trace elements analyzed in bone samples, only strontium was thought to have any dietary significance. There is little variation in strontium content in the bone samples analyzed. Comparison of trace element levels in bone and food does not provide any specific information about diet. Rubidium levels are very low, indicating little postmortem contamination. Zinc levels are highly variable and do not suggest any dietary influence. Hair and bone collagen reflect similar dietary information but there is a difference in the diet-hair and diet-collagen spacing for both carbon and

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

nitrogen stable isotopes. Bone collagen is

19

heavier, on average, than hair

for

Bone collagen is lighter, on average, than hair for (Fig. 1.3). Sex and age differences in stable isotopes of nitrogen and carbon are not pronounced. There is no evidence that males and females were eating different foods and the only evidence for age differences, higher in infants, has been explained by the trophic level shift during the time the infant derives most of its protein from, breast milk. The small amount of variation in both and values supports the historical sources, which indicate that while food was plentiful, the diet was rather monotonous. The study of food and history in 19th century Ontario provides a rich

source of information that may be useful for interpreting bone chemistry data from prehistoric societies. It is our hope that it also provides a clearer view of the life of the early settlers of Belleville.

ACKNOWLEDGMENTS We thank the members of St. Thomas’ Anglican Church, Belleville for

permission to study the remains of their ancestors. Stable isotope analyses were carried out in the Stable Isotope Laboratory, Department of Physics, University of Calgary, under the direction of H.R. Krouse and with the assistance of Nanita Lozano. Trace element analyses were carried out in cooperation with the Department of Chemistry, University of Calgary, and in the Department of Geology, University of Calgary. Dr. Howard Yeager, Department of Chemistry, provided valuable advice and training in preparation of food samples for trace

element analysis. Thanks also to Patrick Michael, Department of Geology, for assistance in sample preparation and analysis. Dr. Don McKay, Department of Medical Biochemistry, University of Calgary provided amino acid analyses. The Bradley House Museum in Clarkson, Ontario provided the iron cooking pot used to prepare the foods. This research was funded by the Social Sciences and Humanities Research Council of Canada (#410-91-1408).

REFERENCES Abonyi, S. 1993 The Effects of Processing on Stable Isotope Levels and Mineral Concentration in Foods: Implications for Paleodietary Reconstruction. Master’s Thesis, Department of Archaeology, University of Calgary.

Ambrose, S.H. 1993 Isotopic analysis of paleodiets: Methodological and interpretive considerations. In Sandford, M.K., ed., Investigations of Ancient Human Tissue: Chemical Analyses in Anthropology. Langhorne, PA, Gordon and Breach Science Publishers: 59–130. Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of bone collagen and carbonate. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone; Archaeology at the Molecular Level. Berlin, Springer-Verlag: 1–37.

20

M. ANNE KATZENBERG ET AL.

Boyce, G.E. 1972 Hutton of Hastings. The Life and Letters of William Hutton, 1801–61. Belleville: Hastings County Council. Canniff, W. 1869 The Settlement of Upper Canada (reprinted 1971). Belleville, Mika Silk Screening, Ltd. DeNiro, M.J. and Epstein, S. 1978 Carbon isotopic evidence for different feeding patterns in two hyrax species occupying the same habitat. Science 201: 906–908. Dudar, J.C., Pfeiffer, S. and Saunders, S.R. 1993 Evaluation of morphological and histological adult skeletal age-at-death estimation techniques using ribs. Journal of Forensic Sciences 38: 677–685. Ezzo, J.A. 1994 Zinc as a paleodietary indicator: An issue of theoretical validity in bonechemistry analysis. American Antiquity 59: 606–621.

Fogel, M., Tuross, N. and Owsley, D.W. 1989 Nitrogen isotope tracers of human lactation in modern and archaeological populations. Annual report of the Director; Geophysical Laboratory, 1988–1989. Washington, D.C., Carnegie Institution of Washington: 111–117. Hare, P.E., Fogel, M.L., Stafford, T.W. Jr., Mitchell, A.D. and Hoering, T.C. 1991 The isotopic composition of carbon and nitrogen in individual amino acids isolated from modern and fossil

proteins. Journal of Archaeological Science 18: 227–292. Hastorf, C.A. and DeNiro, M.J. 1985 Reconstruction of prehistoric plant production and cooking practices by a new isotopic method. Nature 315: 489–491. Herring, D.A., Saunders, S.R. and Katzenberg, M.A. 1998 Investigating the weaning process in past populations. American Journal of Physical Anthropology 105: 425–439. Herring, D.A., Saunders, S.R. and Boyce, G. 1991 Bones and burial registers: infant mortality in a 19th-century cemetery from Upper Canada. Northeast Historical Archaeology 20: 54–70. Jimenez, S. 1991 Analysis of Patterns of Injury and Disease in an Historic Skeletal Sample from

Belleville, Ontario. Master’s Thesis, Department of Anthropology, McMaster University, Hamilton, Ontario. __ 1994 Occupational hazards in 19th-century Upper Canada. In Herring, D.A. and Chan, L., eds., Strength in Diversity: A Reader in Physical Anthropology. Toronto, Canadian Scholars’ Press Inc.: 345–364. Katzenberg, M.A. 1992 Advances in stable isotope analysis of prehistoric bones. In Saunders, S.R. and Katzenberg, M.A., eds., The Skeletal Biology of Past Peoples: Research Methods. New York, Wiley-Liss: 105–120. Katzenberg, M.A., Herring, D.A. and Saunders, S.R. 1996 Weaning and infant mortality: Evaluating the skeletal evidence. Yearbook of Physical Anthropology 39:177–199. Katzenberg, M.A. and Pfeiffer, S. 1995 Nitrogen isotope evidence for weaning age in a nineteenth century Canadian skeletal sample. In Grauer, A.L., ed., Bodies of Evidence. New York, John Wiley & Sons, Inc.: 221–235. Katzenberg, M.A., Saunders, S.R. and Fitzgerald, W.R. 1993 Age differences in stable carbon and nitrogen isotope ratios in a population of prehistoric maize horticulturists. American Journal of Physical Anthropology 90: 267–281. Kenyon, I. and Kenyon, S. 1992 Pork and potato, flour and tea: Descriptions of food and meals in Upper Canada, 1814–1867. Kewa 92: 2–25. Krueger, H.W. and Sullivan, C.H. 1984 Models for carbon isotope fractionation between diet and

bone. In Turnland, J.R. and Johnson, P.E., eds., Stable Isotopes in Nutrition. Washington D.C, American Chemical Society Symposium Series, No. 258: 205–220. Lambert, J.B. and Weydert, J.M. 1993 The fundamental relationship between ancient diet and the inorganic constituents of bone as derived from feeding experiments. Archaeometry 35: 279–294.

Lambert, J.B. and Weydert-Homeyer, J.M. 1993 Dietary inferences from element analysis of bone. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 217–228.

A CASE STUDY FROM 19TH CENTURY UPPER CANADA

21

Marino, B.D. and DeNiro, M.J. 1987 Isotopic analysis of archaeobotanicals to reconstruct past climates: Effects of activities associated with food preparation on carbon, hydrogen and

oxygen isotope ratios of plant cellulose. Journal of Archaeological Science 14: 537–548. Marino, B.D. and McElroy, M.B. 1991 Isotopic composition of atmospheric inferred from carbon in

plant cellulose. Nature 349:127–131.

Moodie, S. 1853 Life in the Clearings versus the Bush (reprinted 1989). Toronto: McClelland and Stewart, Inc. Pate, F.D. 1994 Bone Chemistry and Paleodiet. Journal of Archaeological Method and Theory 1: 161–209. Price, T.D., Blitz, J., Burton, J. and Ezzo, J.A. 1992 Diagenesis in prehistoric bone: Problems and

solutions. Journal of Archaeological Science 19: 513–529. Proudfoot, W. 1915 The Proudfoot Papers: Part I (1832). Collected by Miss Harriet Priddis. Transactions of the London and Middlesex Historical Society, Part VI. Rogers, T. 1991 Sex Determination and Age Estimation: Skeletal Evidence from St. Thomas’

Cemetery Belleville, Ontario. Master’s Thesis. Department of Anthropology, McMaster University. Rogers, T. and Saunders, S.R. 1994 Accuracy of sex determination using morphological traits of the human pelvis. Journal of Forensic Sciences 39: 1047–1056. Sandford, M.K. 1992 A reconsideration of trace element analysis in prehistoric bone. In S.R. Saunders and M.A. Katzenberg, eds., Skeletal Biology of Past Peoples: Research Methods. New

York, Wiley-Liss: 79–104. Saunders, S.R., DeVito, C., Herring, D.A., Southern, R. and Hoppa, R.D. 1993 Accuracy tests of tooth formation age estimations for human skeletal remains. American Journal of Physical

Anthropology 92: 173–188. Saunders, S.R., DeVito, C. and Katzenberg, M.A. 1997 Dental caries in nineteenth century Upper

Canada. American Journal of Physical Anthropology 104: 71–87. Saunders, S.R., Fitzgerald, C., Rogers, T., Dudar, C. and McKillop, H. 1992 A lest of several methods of skeletal age estimation using a documented archaeological sample. Journal of the Canadian Society of Forensic Sciences 25: 97–118. Saunders, S.R., Herring, D.A. and Boyce, G. 1995 Can skeletal samples accurately represent the living populations they come from? The St. Thomas’ Cemetery Site, Belleville, Ontario. In Grauer, A.L., ed., Bodies of Evidence. New York, John Wiley and Sons, Inc.: 69–89.

Saunders, S.R., Herring, D.A., Sawchuk, L.A. and Boyce, G. 1995 The Nineteenth-century Cemetery at St. Thomas’ Anglican Church, Belleville: Skeletal Remains, Parish Records and Censuses. In Grave Reflections: Portraying the Past through Cemetery Studies. Toronto, Canadian Scholars’ Press: 93–118.

Saunders, S.R. and Hoppa, R.D. 1993 Growth deficit in survivors and non-survivors: biological mortality bias in subadult skeletal samples. Yearbook of Physical Anthropology 36: 127– 152. ____1995 Long bone metrics and sexual dimorphism in a 19th-century historical sample. Paper presented at the Canadian Society for Forensic Sciences Meetings, Toronto, Ontario. Saunders, S.R., Hoppa, R.D. and Southern, R. 1993 Diaphyseal growth in a nineteenth century skeletal sample of subadults from St. Thomas’ Church, Belleville, Ontario. International

Journal of Osteoarchaeology 3: 265–281. Saunders, S.R. and Sawchuk, L.R. 1995 Biological indicators of labour and occupational stress: Gender comparisons in a Nineteenth Century pioneer community. Paper presented at the American Association of Physical Anthropology Meetings, Oakland, California. Schoeninger, M.J. and Moore, K. 1992 Bone stable isotope studies in archaeology. Journal of World

Prehistory 6: 247–296. Schurr, M.R. 1997 Stable nitrogen isotopes as evidence for the age of weaning at the Angel Site: A comparison of isotopic and demographic measures of weaning age. Journal of Archaeological Science 24: 919–927.

22

M. ANNE KATZENBERG ET AL.

Schwarcz, H.P. and Schoeninger, M.J. 1991 Stable isotope analyses in human nutritional ecology. Yearbook of Physical Anthropology 34: 283–322. Sealy, J.C. 1986. Stable Carbon Isotopes and Prehistoric Diets in the Southwestern Cape Province, South Africa. Oxford: BAR International Series 293. Tieszen, L.L. and Fagre, T. 1993 Effect of diet quality and composition on the isotopic composition of respiratory bone collagen, bioapatite and soft tissues. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 121–155. Talbot, E.A. 1824 Five Years’ Residence in the Canadas. London: Longman, Hurst, Rees, Orme, Brown and Green. Tivy, L. 1972 Your Loving Anna. Letters from the Ontario Frontier. Toronto, University of Toronto Press. Traill, C.P. 1836 The Backwoods of Canada (reprinted 1966). Toronto, McClelland and Stewart. ___ 1855 The Canadian Settler’s Guide (reprinted 1969). Toronto, McLelland and Stewart. White, C.D. and Schwarcz, H.P. 1994 Temporal trends in stable isotopes for Nubian mummy tissues. American Journal of Physical Anthropology 93: 165–187.

Chapter

2

Diet and Animal Husbandry of the Preclassic Maya at Cuello, Belize: Isotopic and Zooarchaeological Evidence NlKOLAAS J. VAN DER MERWE, ROBERT H. TYKOT, NORMAN HAMMOND, AND KIM OAKBERG ABSTRACT The diet of the Preclassic Maya at Cuello, Belize was studied by means of carbon and nitrogen isotope measurements on human and animal bones from the site, as well as on modern animals from the region. The average value for Preclassic human bone collagen was (n = 28) and for tooth enamel apatite it was (n = 33); the average in bone collagen was (n = 23). The archaeological faunal remains, in order of frequency, include white-tailed deer, freshwater turtle and dog, plus smaller numbers of armadillo, Brocket deer, peccary, and rodent. All of these are plant eaters (ave. n = 19), except dog (–15.6‰, n = 12) and armadillo (–16.4‰, n = 6). Archaeological plant remains include maize (estimated carbon and nitrogen isotope values –10‰, +3‰) and a variety of forest species. Marine foods are barely represented in the archaeological deposits. The archaeological and isotopic evidence together indicate that the people at Cuello made substantial use of maize, but were not dependent on it like later Maya populations. carbon made up ca. 30–35 percent of their bone and tooth Biogeochemical Approaches to Paleodielary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

23

24

NIKOLAAS J. VAN DER MERWE ET AL.

enamel apatite and 50–55 percent of their bone collagen. This apparent discrepancy is the result of their eating dog and armadillo, both with substantial

components in their diets. The dogs were the size of large terriers, were slaughtered in their prime, and were apparently allowed to scavenge rather than being fed a high-maize diet to fatten them for the pot. The deer neither raided the cornfields nor were they intentionally fed maize. Feeding maize to deer and dogs are phenomena which were observed in historic times, but evidently developed after the Preclassic.

INTRODUCTION: PRECLASSIC CUELLO Cuello was excavated by Hammond and co-workers between 1975 and 1993. It is the earliest known Preclassic Maya site, with a Preclassic occupation from ca. 1200 BC to AD 300 as well as later Classic period (AD 300–900) remains; the earliest pottery-using phase (Swasey, 1200–900 BC) has not yet been found at other Preclassic sites, but the Bladen (900–600 BC) and subsequent phases match occupations elsewhere in date and material culture. The Cuello excavations have been extensively described in the report edited by Hammond (1991). Of particular relevance here are the chapters on the ecology and subsistence economy (Ch. 4) by Miksicek and by Wing and Scudder, and on the human burials (Ch. 7), by Frank and Julie Saul. More recent publications have focused on the subsistence economy (Crane and Carr 1994) and on the human skeletal remains (Saul and Saul 1997). Human burials (166 in all) occur throughout the 1600 year Preclassic occupation at Cuello and so do trash deposits. Analysis of plant remains from the latter show that a range of forest species were exploited for economic purposes, including that of diet. Maize (Zea mays) is present from the earliest levels, and analysis of cupule sizes shows that early, very small-cobbed maize was replaced by progressively larger cobs (Miksicek 1991: Fig. 4.1), suggesting (but not guaranteeing) increased productivity over time. Maize was domesticated from wild teosinte in Central Mexico as early as 6000 BC, spread to the Belizean tropical lowlands by 3000 BC, and is correlated with extensive forest disturbance by 2500 BC (Pohl et al. 1996). Other staple crops included roots such as manioc (Manihot esculenta), malanga (Xanthosoma sp.) and probably sweet potato (Ipomoea batatas) (Hather and Hammond 1994). Although initially agriculture in the Cuello area was probably based on swiddening, by the late Middle Preclassic (600–400 BC) there is evidence for drained fields in low-lying wetlands, suggesting either a preference for such locations sufficient to stimulate the necessary investment of labor, or sufficient pressure on the swiddening system to require the adoption of more intensive means of cultivation. The historic Maya diet is protein-deficient, and it was long assumed that prehispanic meat intake was also low (Morley 1946: 25; Béhar 1968), although

ISOTOPIC AND ZOOARCHAEOLOGICAL EVIDENCE

25

the diversity of potential protein sources has been demonstrated by the systematic recovery of even fragmentary animal remains at Cuello (Wing and Scudder 1991: Tables 4.8–4.14). Analysis of these shows the presence of the same economic species throughout the Cuello sequence. Deer (Odocoileus vir-

ginianus) form about 50 percent of the faunal remains, both in MNI and bone frequency. Two species of freshwater turtle are the next most frequent (mud turtle, Kinosternon sp., and Staurotypus triporcatus) followed by dog (Canis familiaris). Occurring in smaller numbers are armadillo (Dasypus novemcinctus), peccary (Tayassu sp.), gibnut (Agouti paca), brocket deer (Mazama americana), opossum (Didelphus marsupialis) and other animals. Fish form less than two percent of the faunal remains and most of them are freshwater species. Mollusc shells are not plentiful, apart from an abundance of the swamp snail

Pomacea flagellata, which was apparently harvested at optimal size (Miksicek 1991: Fig. 4.2); other shells (river mussel, conch, and thorny oyster) were apparently brought to the site for manufacture into jewelry. Conch (Strombus gigas) comes from the Caribbean coast, some 50km from Cuello. Of the thorny oysters much prized in the Mayan world, Spondylus americanus is found along the Caribbean shore but Spondylus princeps is a deep-water species from the Pacific, obtained in long-distance trade. Most of the beads and other artifacts are made from Spondylus so heavily modified that the species cannot be determined, although S. americanus would seem the easier to acquire. Shellfish could, of course, have been transported to Cuello out of the shell from the Caribbean, but their shelf-life is extremely short in the tropics. We wished to establish the nature of the Preclassic Maya diet at Cuello, with special interest in the extent to which maize was a staple, and whether its importance increased over time as it became more productive. Given the importance of maize to Maya civilization, a measure of its consumption in the Preclassic is a high priority. We also wished to establish whether the Preclassic Maya specifically raised dogs as a food source and fed them maize for this purpose. The consumption of dogs was a common practice in later New World societies where other sources of meat were few (Schwartz 1997). Bishop Landa noted in Historic times (Tozzer 1941: 203) that a small, hairless breed of dog was kept in this manner by the Maya. The dogs of Cuello were not small: all were of large terrier size, but cutmarks show that they were butchered and that this was done after the initial growth period, at about 1 year of age (Clutton-Brock and Hammond 1994). It remained for us to determine whether they scavenged or were fed. A similar question arises with respect to deer. They were obviously not domesticated like dogs, but Bishop Landa observed that they were tameable: “They . . . raise other domestic animals, and let the deer suck on their breasts, by which means they raise them and make them so tame that they will not go in the woods, although they take and carry them through the woods and raise them there” (Tozzer 1941: 127). Such a procedure is likely to involve maize fodder for the deer, or at least cornfield browsing.

26

NIKOLAAS J. VAN DER MERWE ET AL.

PROCEDURE To resolve these questions, we measured the stable carbon and nitrogen isotope ratios in (1) Preclassic human skeletons from Cuello, using collagen as the sample material, but also measuring the carbon isotopes in the apatite of bone and tooth enamel; (2) animal bones from the archaeological deposits; and (3) bones and/or flesh of modern terrestrial, riverine, and marine animals collected in the area in 1992 (see also Tykot et al. 1996). The modern specimens were used to expand the database on the isotopic ecology of Cuello during the Preclassic, insofar as the modern environment can be considered to simulate the ancient one. In particular, the modern collection provided information about the marine system, since marine foods may mimic maize in values.

Bone from the Cuello archaeological deposits was poorly preserved and of a chalky consistency. Collagen was extracted by breaking the bone into granules of about 2 mm and dissolving them slowly in dilute hydrochloric acid. After neutralization with distilled/deionized water, the samples were treated with NaOH to remove humic acids and defatted with a 2:1:0.8 mixture of

methanol:chloroform:water. The collagen pseudomorphs were freeze-dried; combusted in closed quartz tubes with copper oxide, copper, and silver; and the carbon dioxide and nitrogen gases were isolated in a vacuum line. This is a conservative procedure which loses some potential collagen samples, but invariably yields C/N ratios near 3.2. For apatite separation, cleaned bone and tooth enamel were ground in a Spex mill; the organic phase was dissolved in

Clorox; possible calcite contaminants were dissolved in dilute acetic acid; and carbon dioxide was liberated in phosphoric acid at 90°C. The weight percent of carbon produced was typically 0.4–0.7%. The carbon and nitrogen isotope ratios were measured in a VG PRISM 2 isotope ratio mass spectrometer in the Department of Earth and Planetary Sciences at Harvard, utilizing the automatic sample manifold, a Carlo Erba analyzer, and a series of NBS and in-house standards. Results are reported in parts per mil (‰) relative to the PDB standard for carbon and to the AIR standard for nitrogen. INTERPRETATION Where in the free atmosphere has a value of –7‰, and plants are anticipated to have values of about –26.5‰ and –12.5‰ respectively (van der Merwe 1989); archaeological maize, however, typically

averages –9.5‰ (Schwarcz et al. 1985). The isotopic values of modern maize and plant foods in Mesoamerica (Wright 1994: 203–206), after correction for the Industrial Effect, average –9.6‰ and –26.4‰ respectively. Since herbivore collagen is typically enriched by +5‰ relative to the diet (van der Merwe 1989), animals from this region with a pure

plant diet should

ISOTOPIC AND ZOOARCHAEOLOGICAL EVIDENCE

27

have values around –21.5‰; the archaeological deer and peccary from Cuello average –21.0‰. Similarly, a pure plant diet should be represented by collagen values around –4.5‰. Bone apatite is typically enriched by +12‰ relative to the diet (Krueger and Sullivan 1984; Lee-Thorp et al. 1989; see also Ambrose and Norr 1993), suggesting endpoints of –14.5‰ and +2.5‰. The approximate contribution of foods to collagen and apatite is estimated by interpolation, with the understanding that these percentages may be modified by future research on diet-tissue fractionation and local endpoint values.

CUELLO ISOTOPIC ECOLOGY The modern environment at Cuello has been extensively altered by human action: the area today has extensive stands of sugar cane, open grasslands for cattle grazing, and two types of savanna. The latter include some grasses, which may be immigrants. The river banks are densely covered in mangrove and tropical forest, and remnant tropical forest patches occur nearby.

The Preclassic landscape is generally interpreted as having been tropical forest, based on the animal and plant species from the archaeological deposits. This is largely borne out by the isotope ratios of deer and peccary 0.7‰; which are consistent with a forest environment (Table 1.1; Fig. 1.1). The same is true for mud turtles 2.0‰), which evidently also had a plant diet. There is substantial evidence, however, that a plant component was present at Cuello in Preclassic times. Insect-eating armadillos average in and one mud turtle has a value of –16.4‰. This may be explained as due to the habit armadillos have of making their burrows in cornfields, where they are easily caught, and the possibility that an occasional mud turtle migrated from the river to a drainage canal next to a raised cornfield in swampy wetland. Since indigenous grasses and sedges are known from open grasslands in the lowland tropics (see e.g., Tieszen and Boutton 1989: 176), their presence in Belize cannot be discounted and deserves further investigation.

The marine isotopic environment was assessed by analyzing specimens from Caye Caulker, some 20 km offshore. The results are not detailed here, as they largely confirm those of a study of coral reef fauna from the Bahamas by Keegan and DeNiro (1988). Fish bone collagen from Caye Caulker averages – in and in while shellfish like conch, whelk, winkle, and flat tree oyster average and (n = 11). These isotopic values are very different from those encountered in studies of marine foodwebs outside the Caribbean (Fig. 2.2). They are primarily due to extensive stands of flowering marine grasses like Thalassia testudinum (turtle grass), which grow in the shallows between the mainland and the offshore reefs; they have exceptional isotope ratios, on the order of –6‰

28

NIKOLAAS J. VAN DER MERWE ET AL.

for

and 1‰ for (Keegan and DeNiro, 1988). Fortunately, as the isotopic values for the collagen and apatite of Cuello human bone show, marine foods were not a significant component of their diet, as would be anticipated for a site located some 50km inland. The carbon isotope ecology of the marine system is controlled by ocean bicarbonates and has presumably not changed since the Preclassic. The terrestrial system is controlled by atmospheric carbon dioxide, which has changed as a result of the Industrial Effect (van der Merwe 1989). This is noticeable in the collagen values of modern animals hunted in the forest, which are about 1.6‰ more negative than those from the Preclassic (Table 2.2). The number of specimens are too small to serve as a proxy measurement for atmospheric change, but they provide a reminder that modern values cannot be used unchanged in a study of archaeological diets. The canopy effect (van der Merwe and Medina 1991) will also result in depleted carbon isotope

ratios for animals living in heavily forested environments.

DEER AND DOGS The collagen values of deer bone from the Preclassic at Cuello are consistent with those of plant eaters (Table 2.1; Fig. 2.1). They did not eat maize and were evidently not tamed or loose-herded. The dogs (Table 2.1; Fig. 2.1), however, often had a substantial component in their diet. Their values average (n = 12) and their values average (n = 12). The large variation in these values, compared to those of their human owners, suggest that they were not fed a stable household diet, but that they also scavenged, foraged and hunted for food on their own. Using the endpoints described above, we estimate their collagen includes about 35–40 percent carbon. Elsewhere in Belize, similar values for both deer and dogs have been reported for Preclassic Colha, but some dogs and deer from several Classic and Postclassic sites have considerably more enriched carbon isotope ratios (White et al. 1993; 1997). Dogs from Classic Maya sites in Honduras and Guatemala typically have very enriched values indicating heavy dependence on plant foods (Gerry 1993; 1997; Gerry and Krueger 1997).

HUMANS

The

values of the Preclassic humans at Cuello (Table 2.1) average

(n = 28) in collagen, in bone apatite (n = 16), and in tooth enamel apatite (n = 33); the values in collagen average (n = 23). The discrepancy in the number of specimens is due to the fact that more teeth were available than post-cranial material, while some of the specimens contained insufficient collagen to measure the nitrogen isotope ratios. Additional bone apatite analyses are in progress.

ISOTOPIC AND ZOOARCHAEOLOGICAL EVIDENCE

29

The Preclassic human diet at Cuello can be interpreted in the light of isotope ratios for human bone and potential foods (Table 2.1; Fig. 2.1). Using the endpoints discussed above, we estimate the contribution to human collagen carbon was ca. 50–55 percent and that the contribution to human bone and tooth enamel apatite was ca. 30–35 percent. Controlled diet experiments suggest that the apatite value can be taken as the average content

of the diet, while the carbon in collagen overemphasizes the -based protein component (Ambrose and Norr 1993). This leads to the conclusion that maize provided something less than 30 percent of the Preclassic human diet at Cuello, while frequent portions of dog meat served to enlarge the

30

NIKOLAAS J. VAN DER MERWE ET AL

Figure 2.1. Carbon and nitrogen isotope values of human and animal bone collagen from the Preclassic at Cuello.

collagen carbon content to 50–55 percent. Armadillo meat played a smaller role towards the same end. Had marine foods been consumed in any quantity, this would have been evident from enriched nitrogen as well as carbon isotope values. We may conclude then that the historically observed practice of

Figure 2.2. Carbon and nitrogen isotope values of marine fauna from Belize, Ecuador (van der Merwe et al. 1993), and the southwestern Cape coast of South Africa (Sealy and van der Merwe 1986).

ISOTOPIC AND ZOOARCHAEOLOGICAL EVIDENCE

31

dog-eating on a regular basis may be extended back in time to the Preclassic, even in Belize where alternative meat sources were relatively more abundant than in other regions. It is not possible to solve the dietary equation for Preclassic Cuello more accurately at present. In this preliminary report, we have treated all of the human teeth as one group, although various molars, premolars, and some canines and incisors were analyzed. Since these teeth form at different juvenile ages, isotopic variation resulting from the introduction of solid foods and ultimately from weaning is expected (e.g., Wright and Schwarcz 1998). The average carbon isotope enrichment of the tooth enamel by about 1‰ relative to the bone apatite data set is likely a trophic effect resulting from the preweaning diet represented in early forming teeth; the larger standard deviation for the tooth enamel data set is a result of our including teeth representing both pre- and post-weaning diets. It is perhaps more instructive to compare the Preclassic collagen values with those of modern residents. The hair of the camp cook at the Cuello excavations had a value of –16.4‰ and a value of 9.8‰ (Table 2.3).

32

NIKOLAAS J. VAN DER MERWE ET AL.

To normalize these values for comparison with Preclassic collagen it is necessary to add to the value 1.5‰ for the Industrial Effect and 1‰ for the spacing between hair and collagen, yielding –13.9‰; the value of 9.8‰ remains constant. These values are quite close to the average for Preclassic

Cuello, and suggest that sources constitute about 40–50% of her diet. Béhar (1968) estimated that maize constituted about 70% of the modern Maya diet, an average value subject to regional variation and which may be further modified by late 20th century economic circumstances. Our modern resident eats about three maize tortillas with every meal and a fair amount of beef, which has been raised on a mixture of and grasses. Her dog, Blackie, has and values of –14.3‰ and 8.2‰, as normalized for the Preclassic. These values are likewise nearly identical to those of Preclassic dogs. Blackie eats the leftovers of his owner’s meals and gets somewhat smaller portions of tortilla and meat in the process. If the modern resident and her dog are representative of the general population, then these observations help to put the Preclassic diet in perspective. Finally, a detailed comparison of burials from several Preclassic phases at Cuello shows no significant chronological trends in their isotope ratios (Tykot et al. 1996). Adult males and females, however, do differ by up to 1.0‰ in values, a gender-based difference observed elsewhere for the Preclassic (White 1997; Wright 1997). Males had a dietary component as much as 10 percent higher than females, perhaps acquired in the form of maize beer. For juveniles, the specimen numbers are too small for firm conclusions. Several adult individuals from a mass burial at Cuello have distinctly enriched isotopic values relative to the single burials presented here, suggesting that they were not native to Cuello (Tykot et al. 1996).

MAYA DIETS Five other isotope studies of Maya skeletal populations in Belize are

available for comparison with Cuello (Table 2.4; Fig. 2.3). These include a time series from Preclassic to Historic at Lamanai (White and Schwarcz 1989), an Early through Terminal Classic sequence at Pacbitun (White et al. 1993),

ISOTOPIC AND ZOOARCHAEOLOGICAL EVIDENCE

33

Figure 2.3. Carbon and nitrogen isotope values of human bone collagen at Maya sites in Belize. The value for the modern sample has been corrected for collagen-hair spacing and the Industrial

Effect. Boxes represent isotopic means

one standard deviation.

34

NIKOLAAS J. VAN DER MERWE ET AL

a mostly Preclassic sample from Cahal Pech (White et al. 1996), and analyses

of the Classic populations at Mojo Cay (Norr 1991), Baking Pot and Barton Ramie (Gerry 1993; 1997; Gerry and Krueger 1997). Some differences are apparent between these sites (the carbon isotope values for the Late/Terminal Classic at Lamanai are very negative), but the general trend is that of an intensification of maize consumption from the Preclassic through the Classic to Postclassic and Historic times. Our data for Mrs. Martinez suggests that maize has become less important in the modern diet, returning to a similar level as that observed for the Preclassic. This trajectory has been observed elsewhere, Ecuador being an example (van der Merwe et al. 1993). The substantial differences which existed between Preclassic diets in Belize (White and Schwarcz 1989; Tykot et al. 1996) and in the Peten region of Guatemala (Wright 1994) can most likely be attributed to differences in local ecology, as well as in population density and in status-based access to certain food resources (Table 2.4; Fig. 2.4). Local ecological differences also may explain most of the dietary differences observed in the Classic period, for which we have data from Belize (White and Schwarcz 1989; Norr 1991; White et al. 1993; Gerry 1993), Honduras (Gerry 1993; Reed 1994), and Guatemala (Gerry 1993; Wright 1994) (Fig. 2.5). People in Guatemala and Honduras remained considerably more dependent on maize than those in Belize, with Copan residents in particular having a “corn and beans” diet. This is understandable, as Copan lies at the head of a closed mountain valley where a concentration of people would rapidly eradicate wildlife and forest in the cause of agriculture. Unlike Belize, the Copan valley of today is nearly devoid of

Figure 2.4. Carbon and nitrogen isotope values in human bone collagen from Preclassic Belize (Cuello and Lamanai) and from the Preclassic Peten (Altar de Sacrificios and Seibal).

ISOTOPIC AND ZOOARCHAEOLOGICAL EVIDENCE

35

Figure 2.5. Carbon and nitrogen isotope values in human bone collagen from Preclassic Belize; Classic Belize (Pacbitun, Baking Pot, and Barton Ramie only); the Classic Peten (Uaxactun, Holmul, Seibal, Altar de Sacrificios, Seibal, Dos Pilas, Aguateca, Itzan); and Classic Honduras (Copan).

indigenous animals and plants. The Classic Maya of the Peten evidently ate more meat than those of Copan, but they were also highly dependent on maize. The Preclassic and Classic Maya of Belize had a much more diversified diet than those of Guatemala and Honduras. They had access to a wider range of ecozones and presumably had a lower population density. The fact that geography and local ecology played the primary role in determining diet in the Maya world argues forcefully against theories about the demise of Maya civilization that are based on the collapse of maize agriculture (e.g., Santley et al. 1986; Culbert 1988; see also Wright and White 1996; White 1997; Wright 1997).

ACKNOWLEDGMENTS This research was supported by a grant from the Wenner-Gren Foundation for Anthropological Research to N.J. van der Merwe, and by Harvard University. We thank Christine White and the editors for their useful comments on an earlier draft of this paper, and Mrs. Arjelia Martinez and her dog, Blackie,

for samples of modern hair.

36

NIKOLAAS J. VAN DER MERWE ET AL

REFERENCES Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of bone collagen and carbonate. In Lambert,

J. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 1–37.

Béhar, M. 1968 Food and nutrition of the Maya before the conquest and at the present time. In Biomedical Challenges Presented to the American Indian. Washington, DC, Pan American Health Organization, Scientific Publication 165: 114–119. Clutton-Brock, J. and Hammond, N. 1994 Hot dogs: comestible canids in Preclassic Maya culture at Cuello, Belize. Journal of Archaeological Science 21: 819–826. Culbert, T.P. 1988 The collapse of Classic Maya civilization. In Yoffee, N. and Cowgill, G.L., eds., The Collapse of Ancient States and Civilizations. Tucson, University of Arizona Press: 69–101. Crane, C.J. and Carr, S.H. 1994 The integration and quantification of economic data from a late Preclassic Maya community in Belize. In Sobolik, K.D., ed., Paleonutrition: The Diet and Health of Prehistoric Americans. Occasional Paper 22, Center for Archaeological Investigations, Carbondale, University of Southern Illinois: 66–79. Gerry, J.P. 1993 Diet and Status among the Classic Maya: An Isotopic Perspective. PhD Dissertation, Harvard University. Ann Arbor, MI, University Microfilms. ___ 1997 Bone isotope ratios and their bearing on elite privilege among the Classic Maya. Geoarchaeology 12: 41–69. Gerry, J.P. and Krueger, H.W. 1997 Regional diversity in Classic Maya diets. In Whittington, S.L. and Reed, D.M., eds., Bones of the Maya: Studies of Ancient Skeletons. Washington, DC, Smithsonian Institution Press: 196–207. Hammond, N. 1991 Cuello. An Early Maya Community in Belize. Cambridge, Cambridge University Press. Hather, J. and Hammond, N. 1994 Ancient Maya subsistence diversity: root and tuber remains from Cuello, Belize. Antiquity 68: 330–335, 487–488. Keegan, W.F. and DeNiro, M.J. 1988 Stable carbon- and nitrogen- isotope ratios of bone collagen used to study coral-reef and terrestrial components of prehistoric Bahamian diet. American Antiquity 53: 320–336. Krueger, H.W. and Sullivan, C.H. 1984 Models for carbon isotope fractionation between diet and bone. In Turnlund, J.E. and Johnson, P.E., eds., Stable Isotopes in Nutrition. American Chemical Society, Symposium Series 258: 205–222. Lee-Thorp, J.A., Sealy, J.C. and van der Merwe, N.J. 1989 Stable carbon isotope ratio differences between bone collagen and bone apatite, and their relationship to diet. Journal of Archaeological Science 16: 585–599. Miksicek, C.H. 1991 The ecology and economy of Cuello: the natural and cultural landscape of Preclassic Cuello. In Hammond, N., ed., Cuello: An Early Maya Community in Belize. Cambridge, Cambridge University Press: 70–84. Morley, S.G. 1946 The Ancient Maya. Stanford, Stanford University Press. Norr, L.C. 1991 Nutritional Consequences of Prehistoric Subsistence Strategies in Lower Central America. PhD dissertation, University of Illinois at Urbana-Champaign. Ann Arbor, MI, University Microfilms. Pohl, M.D., Pope, K.O., Jones, J.G., Jacob, J.S., Piperno, D.R., deFrance, S.D., Lentz, D.L.O., Gifford, J.A., Danforth, M.E. and Josserand, J.K. 1996 Early agriculture in the Maya lowlands. Latin American Antiquity 7: 355–372. Reed, D.M. 1994 Ancient Maya diet at Copan, Honduras, as determined through the analysis of stable carbon and nitrogen isotopes. In Sobolik, K.D., ed., Paleonutrition: The Diet and Health of Prehistoric Americans. Occasional Paper 22, Center for Archaeological Investigations, Carbondale, University of Southern Illinois: 210–221.

ISOTOPIC AND ZOOARCHAEOLOGICAL EVIDENCE

37

Santley, R.S., Killion, T.W. and Lycett, M.T. 1986 On the Maya collapse. Journal of Anthropological Research 42: 123–159. Saul, J.M. and Saul, F.P. 1997 The Preclassic skeletons from Cuello. In Whittington, S.L. and Reed, D.M., eds., Bones of the Maya: Studies of Ancient Skeletons. Washington, DC, Smithsonian Institution Press: 28–50. Schwartz, M. 1997 A History of Dogs in the Early Americas. New Haven, Yale University Press.

Schwarcz, H.P., Melbye, J., Katzenberg, M.A. and Knyf, M. 1985 Stable isotopes in human skeletons of southern Ontario: reconstructing paleodiet. Journal of Archaeological Science 12: 187–206. Sealy, J.C. and van der Merwe, N.J. 1986 Isotope assessment and seasonal-mobility hypothesis in

the southwestern Cape of South Africa. Current Anthropology 27: 135–150. Tieszen, L.L. and Boutton, T.W. 1989 Stable carbon isotopes in terrestrial ecosystem research. In Rundel, P.W., Ehleringer, J.R. and Nagy, K.A., eds., Stable Isotopes in Ecological Research. New York, Springer-Verlag: 167–195.

Tozzer, A.M., ed. 1941 Landa’s Relacion de las Cosas de Yucatan. Papers of the Peabody Museum of Archaeology and Ethnology, Harvard University 18, Cambridge, MA. Tykot, R.H., van der Merwe, N.J. and Hammond, N. 1996 Stable isotope analysis of bone collagen, bone apatite, and tooth enamel in the reconstruction of human diet. A case study from Cuello, Belize. In Orna, M.V., ed., Archaeological Chemistry: Organic, Inorganic, and Biochemical Analysis. ACS Symposium Series 625, Washington, DC, American Chemical Society: 355–365. van der Merwe, N.J. 1989 Natural variation in 13

13

C concentration and its effect on environmental

12

reconstruction using C/ C ratios in animal bones. In Price, T.D., ed., The Chemistry of

Prehistoric Human Bone. Cambridge, Cambridge University Press: 105–125. van der Merwe, N.J. and Medina, E. 1991 The canopy effect, carbon isotope ratios and foodwebs

in Amazonia. Journal of Archaeological Science 18: 249–259. van der Merwe, N.J., Lee-Thorp, J.A. and Raymond, J.S. 1993 Light stable isotopes and the subsistence base of Formative Cultures at Valdivia, Ecuador. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone. Archaeology at the Molecular Level. Berlin, Springer Verlag: 63–98. White, C.D. 1997 Ancient diet at Lamanai and Pacbitun: implications for the ecological model of collapse. In Whittington, S.L. and Reed, D.M., eds., Bones of the Maya: Studies of Ancient

Skeletons. Washington, DC, Smithsonian Institution Press: 171–180. White, C.D., Healy, P.F. and Schwarcz, H.P. 1993 Intensive agriculture, social status, and Maya diet at Pacbitun, Belize. Journal of Anthropological Research 49: 347–375. White, C.D., Longstaffe, F. and Song, R.-J. 1996 Preclassic Maya diet at Cahal Pech, Belize: the isotopic evidence from human bone collagen. Paper presented at the 61st Annual Meeting of the Society for American Archaeology, New Orleans. White, C.D., Pohl, M. and Schwarcz, H. 1997 Maya husbandry of deer and dog. Paper presented at the 62nd Annual Meeting of the Society for American Archaeology, St. Louis. White, C.D. and Schwarcz, H.P. 1989 Ancient Maya diet: as inferred from isotopic and elemental analysis of human bone. Journal of Archaeological Science 16: 451–474. Whittington, S.L. and Reed, D.M. 1996 Evidence of diet and health in the skeletons of Iximché. In Robinson, E., ed., Estudios Kaqchiqueles: In Memory of William R. Swezey. CIRMA, Antigua, Guatemala. Wing, E.S. and Scudder, S.J. 1991 The ecology and economy of Cuello: the exploitation of animals.

In Hammond, N., ed., Cuello: An Early Maya Community in Belize. Cambridge, Cambridge University Press: 84–97. Wright, L.E. 1994 The Sacrifice of the Earth? Diet, Health, and Inequality in the Pasión Maya Lowlands. Ph.D. dissertation, University of Chicago. Ann Arbor, MI, University Microfilms. ___ 1997 Ecology or society? Paleodiet and the collapse of the Pasión Maya lowlands. In Whittington, S.L. and Reed, D.M., eds., Bones of the Maya: Studies of Ancient Skeletons. Washington, DC, Smithsonian Institution Press: 181–195.

38

NIKOLAAS J. VAN DER MERWE ET AL.

Wright, L.E. and Schwarcz, H.P. 1998 Stable carbon and oxygen isotopes in human tooth enamel: identifying breastfeeding and weaning in prehistory. American Journal of Physical Anthropology 106: 1–18. Wright, L.E. and White, C.D. 1996 Human biology in the Classic Maya collapse: evidence from paleopathology and paleodiet. Journal of World Prehistory 10: 147–198.

Chapter

3

An Overview of Causes for Stable Isotopic Variations in Past European Human Populations: Environmental, Ecophysiological, and Cultural Effects GERT J. VAN KLINKEN, MIKE P. RICHARDS, AND ROBERT E.M. HEDGES ABSTRACT The reconstruction of ancient human diet based on carbon and nitrogen analysis has proved to be relatively simple in the New World: utilization of and marine resources causes relatively large shifts in these stable isotopes. In Europe those food sources seem to play a relatively limited role, resulting in a relatively small range of values observed in past human populations, making dietary analysis much less straightforward. However, an apparent non-random pattern is often observed when isotopic values of regional or temporal groups are compared. In this paper we will describe the factors that contribute to those small-scale isotopic variations, and assess the importance of environmental factors as opposed to anthropogenic, cultural factors. This assessment is Biogeochemical Approaches to Paleodietary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

39

40

GERT J. VAN KLINKEN ET AL.

possible in part through the use of a large data base generated by the dating

process in our laboratory. We conclude that most variation in human values in Europe is related to regional differences in climate, and that most of the variation in seems to be caused by variable meat consumption, combined with poorly understood variation in plant nitrogen values.

INTRODUCTION Palaeodietary research faces a rather different situation in Europe than in the Americas, tropical Asia and Africa. In the latter cases, the variations found in stable isotopic ratios of carbon and nitrogen are often large, due to a much wider spread in the isotopic ratios of the food sources. For example, hunter-gatherer communities in North America show values that reflect food webs based on plants with photosynthesis, while the most important crop grown in ancient American agricultural communities is maize, a plant with a very different carbon isotopic signature (Bender et al. 1981; Van der Merwe

and Vogel 1978; White and Schwarcz 1989). Also, coastal communities seemed to have utilized large amounts of marine resources (e.g., deep-water fish/shellfish (Aufderheide et al. 1994; Chisholm et al. 1982; Walker and DeNiro 1986); or reef- and sea grass-meadow organisms (Keegan and DeNiro 1988; Van Klinken 1991), resulting in shifts in both carbon and nitrogen isotopic ratios of or more when compared to values. So both and marine signals are a common occurrence outside Europe. In Europe the situation is radically different. Isotopic variations at the bottom end of the food web, brought about by differences in plant metabolism versus and CAM), and origin of the fixed (causing a systematic discrepancy between terrestrial and marine ecosystems) are generally absent. This leads to relatively small variations in 13C/12C ratios at the beginning of the food chain. Human consumption (millet) is very limited: only one clear case is known (Magdalenska Gora, Slovenia; Murray and Schoeninger 1988). Extensive consumption of sea food is not widespread either: in the very large majority of European human samples that arrive in our dating laboratory (and others) no marine isotopic ratios are found, even from sites relatively close to the sea. Only in Mesolithic times was there appreciable local utilization of seafood. Skeletal remains from some sites in coastal regions of Denmark (Noe-Nygaard 1988), Portugal (Lubell et al. 1994), and Northern England (Clutton-Brock and Noe-Nygaard 1990) show clear marine stable isotopic signals. Because of the general absence of and marine signals, the stable isotopic variations as measured in bone collagen samples across Europe are relatively small in scale. Despite the fact that the overall variability across Europe is small, we quite often find that between-group (interpopulation) variation exhibits a definite pattern or structure, which may either take the shape of small-scale

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

41

changes with time, or as regional patterning. The possible causes for the patterns are most likely environmental factors (such as variations in the of

atmospheric and climatic variations) and behavioral factors (food selection resulting from cultural choices, i.e. selection of food items from the local food chain). This paper sets out to give a preliminary overview of the potential sources of variability in human isotopic values in a temperate, European context, based on current knowledge of relevant environmental and biological processes. Some trends in 13C and 15N found in animal and human data from Europe are discussed, although not in detail, as that will be done elsewhere (Richards et al. in prep).

CAUSES FOR ISOTOPIC VARIABILITY Variations in the Variations in the

and

of Atmospheric

of atmospheric

and

come in two forms: global,

temporal variations and spatial variations. Temporal variations of a global nature have been observed in oceanic, well-mixed (Mook 1986). They include the fossil fuel effect, which is caused by the admixture of anthropogenic (fossil fuel) Use of fossil fuels has led to a steady change in atmospheric 13 12 C/ C ratio during the last 150 years, from—6.5‰ in 1744 to around –8.0‰ in 1990 (Keeling et al. 1989; Marino and McElroy 1991). This is an important factor to take into account when comparing modern food items with ancient human bone values, where normally about 1.5‰ has to be added to convert modern to pre-industrial values. On a longer time scale, ice core data (Leuenberger et al. 1992) suggests that pre-Holocene was about 0.32‰ more negative than the last (pre-industrial) millennium. The possibility of temporal variation in atmospheric values of has to be kept in mind when even older (for instance pre-Pleistocene) material is studied, and compared to data from other time periods. Spatial variations include differences between marine and continental air due to the influence of the continental biosphere (Keeling et al. 1989, 1980). To make matters more complicated, these differences vary with latitude and season. An extreme example of spatial variations is what is called the canopy effect (Van der Merwe and Medina 1989, 1991; Vogel et al. 1990). The re-use of plant-fractionated, respired in dense vegetation, which very locally changes the of atmospheric to values between –21 and –26‰

(Keeling 1961), can cause a systematic bias between plant and animal species living in forests floor versus forest canopy and open environments. For example, Rodière et al. (1996) describe canopy effects in roe deer in France

and Ambrose and DeNiro (1986, 1989; Ambrose 1993) describe it in several species in Kenya. Canopy effects are likely to influence the carbon isotope ratios of forest-dwelling Mesolithic hunter-gatherer communities that consume

42

GERT J. VAN KLINKEN ET AL.

roe deer, boar, aurochs and forest bison; Neolithic agricultural communities that consume foods from more open landscapes may have less negative values. Air is well-mixed, and acts as a very large reservoir: both factors buffer against much variability in the stable isotope ratios of (Mariotti 1983).

Climate and Plant Photosynthesis Isotopic variability in plants is the primary determinant of the carbon isotopic signature in the food eaten by humans. Climatic differences cause regional patterning across Europe in the of plants (Van Klinken et al. 1994). The climatic effect is brought about by the influences of temperature and/or relative humidity on the photosynthetic process in plants. This influence acts in two ways: i) on the diffusion of gases into and from the plant leaves through the stomata; ii) on the biochemical (enzymatic) step, the actual fixation of carbon, which is affected by temperature and partial pressure of However, models of photosynthetic fractionation (Francey and Farquhar 1982; O’Leary 1981; Vogel 1980) do not allow quantitative explanation of field isotopic observations in terms of the climatic parameters such as the ones mentioned above. The of plants in part depends on the ratio of internal (interstitial) to external concentrations which is influenced by the rate of carbon fixation and the rate of stomatal conductance. It is on both these parameters that environmental factors of temperature, relative humidity or water stress act. For instance, high stomatal conductance, low water use efficiency, and low rates of photosynthesis (Leavitt and Danzer 1992; Van de Water et al. 1994) lead to high values of and thus more negative plant values. Conversely, during periods of high temperature and low humidity the stomata are closed in order to conserve water (low stomatal conductance). Interstitial concentrations decrease (low and proportionately more is fixed, resulting in less negative plant values. As a result, in the Holocene a clear pattern of climate-induced variation in values of plant material and in bone collagen exists across Europe (Van Klinken et al. 1994). Figures 3.1 and 3.2 show that wood and collagen values are closely correlated, demonstrating that the trend is passed on in the food chain. The trend of enrichment from northwestern to southern Europe correlates very strongly with the climatic pattern across Europe; the climatic isotopic variability is on the order of 2 to 4‰. There are also indications that climate-related shifts in wood of 1 to 3‰ have occurred in time: such shifts have been observed at the Pleistocene-Holocene boundary (Becker et al. 1991; Van de Water et al. 1994), and at later (Epstein and Krishnamurthy 1990) and earlier times (Aucour et al. 1993; Leavitt and Danzer 1992). Other temporal variations in plants can be seasonal (Leavitt and Long 1991; Loader et al. 1995).

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

Figure 3.1. Regional variation in

43

values due to climatic influences as observed in Holocene

charcoal samples. Note the north-south climatic trend, combined with an Atlantic-continental trend (west to east).

Variability within plants and between individual plants from the same location, can be quite extensive (in the order of several ‰), often as a result of differences in growing conditions (light intensity, humidity), that can vary over a very short distance (Saurer et al. 1995). Systematic differences seem to occur between life forms such as trees, shrubs, cushion plants, etc. (Tieszen 1991; Tieszen and Fagre 1993a; Valentini et al. 1995, 1992), and between different botanical groups (for example, deciduous versus coniferous trees; Leavitt and Newberry 1992; Ramesh et al. 1986; Stuiver and Braziunas 1987).

Variability in Plants, Nitrogen Uptake and Environmental Factors Plants take up inorganic nitrogen, either as atmospheric

through

symbiotic nitrogen fixation in legumes such as pulses, some shrubs and trees,

Figure 3.2. Average carbon isotopic ratios for all human, herbivore, carnivore, and omnivore bone samples from the European Holocene in the data base: A: Uncorrected ratios B: climatecorrected ratios Only countries with more than 10 samples are included. For a description of the climatic correction procedure see text.

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

or as nitrate (non-nitrogen fixers).

45

values in nitrogen fixers seem to be

closer to atmospheric values, i.e., 0‰, than non-nitrogen fixers, although not in an unequivocal way (Binkley et al. 1985; Domenach et al. 1989; Shearer and Kohl 1986; Virginia and Delwiche 1982; Yoneyama et al. 1990). For instance, experimental comparison (Meints et al. 1975) of maize with soybean (a nitrogen fixer) revealed a consistent difference of about 4‰ (where soybean had values of –3‰), even when considerable amounts of N fertilizer were applied. It seems that non- fixer values are determined by soil properties (the form in which nitrogen is present) and especially microbial activity in the soil, and normally reflect soil nitrogen (nitrate) values (Garten 1993). There seem to be four processes that play a role: nitrogen immobilization by soil microorganisms that make nitrogen unavailable for plants (Davidson et al. 1991,1992; Garten 1993); and volatilization of ammonia, nitrification, and denitrification which latter three are all isotopically fractionating (Delwiche and Steyn 1970; Mariotti et al. 1981, 1982). Shearer et al. (1974) developed a model predict-

ing that the of soil nitrate and ammonium depend upon the ratio of the rates of nitrogen immobilization versus the rate of nitrification, but even then the natural isotopic variability in soil nitrate and ammonia is such that this model is not predictive in actual case studies. The interpretation of the factors mentioned above in the context of palaeodietary research is not straightforward: a lack of nitrogen isotope data

in relevant plant species makes the situation even more complicated. The observed variability in plants, even within ecosystems, is so extensive and so unpredictable that modelling of the behaviour of natural nitrogen abundances in plants is fraught with difficulties: “because there are no simple, universal “laws” governing the site-specific details of the N cycle, there will be no simple, universal “laws” of (in plants)” (Handley and Raven 1992; Handley and Scrimgeour 1996). A relatively uncomplicated case is the use of nitrogen fixers as food items. There is abundant archaeological evidence for the importance of pulses in the

Mediterranean in Neolithic and later periods. Because pulses have values of around 0‰, the use of pulses as food stuffs should be detectable. Recently, we have measured human values in some Greek sites that indicate that a few individuals in the cemetery were almost completely dependent on pulse protein (Van Klinken and Triantaphyllou, in prep.). A special case is the observation of very depleted nitrogen ratios in forest trees and soils. Values of less than –6‰ have been reported, connected with low availability of nitrogen in such soils (Garten 1993). This is in agreement with observations of gradual isotopic enrichment of nitrogen when cultivation is started (Mariotti 1982). These depleted nitrogen ratios in plants are transferred to forest animals: roe deer show values nearly as depleted as the tree leaves they eat (Rodière et al. 1996). Subarctic vegetation shows similarly low

values in some life forms (Michelsen et al. 1996a, 1996b; Nadelhoffer et al. in press; Schultze et al. 1994); in this case the mechanism seems to be

46

GERT J. VAN KLINKEN ET AL.

related to the ectomycorrhizal uptake of nitrogen in these plants. Both ecosystems with their relatively simple variations seem ideal to investigate plant—herbivore interactions. Heaton (1987) compared plants from different climatic zones in South Africa and Namibia, and found that climate-related changes in in plants caused by increased aridity were relatively small, but that increases in salinity however can increase plant values by up to 10‰ (Heaton 1987). He and others found a clear correlation between aridity and values in animals and humans (Heaton et al. 1986; Sealy et al. 1987). Low rainfall causes shifts in for instance to values of up to +14.6‰ in African elephants, and seem to be linked with nitrogen metabolism, especially with the increased urea excretion due to increased water stress: the study of African mammals by Ambrose shows the higher the urea excretion, the higher the (Ambrose 1991). In Europe, Fizet et al. (1995) describe an increase in nitrogen values of 2–3‰ in reindeer, bison and horse from a 40–45,000BP layer in the Late Pleistocene French site of Marillac. They ascribe the increase to drier, colder conditions. This is probably the only published account of a climatic nitrogen effect in Europe. We ourselves have found no indications in our data base of systematic discrepancies between Northern and Southern European nitrogen data for herbivores, at least not to the extent that regional differences exceed more local, microhabitat driven variations. This is not to say that no climatic nitrogen effects can be expected under European climatic circumstances, only that they do not seem to play a similarly dominant role as in arid environments in Africa.

Food Webs, and Carbon and Nitrogen Isotopes: “Trophic Level Effects” A number of systematic changes in carbon and nitrogen isotopic ratios have been observed when organic matter is passed upwards in the food chain.

A shift in both carbon (around +0.5 to 1.0‰) and nitrogen (+3–4‰) is observed during the transfer of biomass in an ecosystem between lower and higher trophic levels (e.g., from herbivores such as deer to carnivores such as wolf). This has often been referred to in the palaeodietary literature, both for carbon and nitrogen, as the trophic level effect (Schoeninger 1985). We will discuss carbon and nitrogen together, in an effort to analyze the similarities and differences between the two.

Carbon “Trophic Level Effects” In carbon, “an uncertainty exists which arises because we have neither sufficient empirical data nor an accepted theoretical model to explain the quantitative relationships between dietary carbon isotope signals and those

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

47

assimilated by consumers” (Tieszen and Fagre 1993b). Attempt to develop models are published elsewhere in this volume (Hedges and Van Klinken, Schwarcz); here we will try to summarize the available data for Europe. A difference of 0.5 to 1‰ between carnivores and herbivores has been observed in modern and ancient food webs (Bocherens et al. 1994a, 1991; Fizet

et al. 1995; Lee-Thorp et al. 1989; Schoeninger 1985; Schoeninger and DeNiro 1984). We will discuss our data below. The small shift might have its origin in preferential uptake of tissue with a different biochemical composition) that has a that differs from the mean of the organism that is eaten. In animals, values have been reported by Huebner (1985) for muscle (–23‰), brain tissue (–21‰), collagen (–20‰), pancreas (–25‰), hair (–20‰), subcutaneous fat (–28‰), and milk (–26‰). Differences in values between plant tissues are much less pronounced than in animals, most plant tissues are close to the average value of carbohydrates (–25 to – 27‰). For that reason “trophic level” shifts should be less pronounced in herbivores than in carnivores. It is difficult to measure a plant-herbivore shift, because we need to take the collagen—diet spacing (for which slightly different values are quoted in the literature), and the variability in plant into account. In our opinion this makes the use of the term “carbon trophic level effect” in herbivores of little value, and consequently it would be clearer to use the term “carnivore effect” for the carbon trophic level effect in carnivores.

Nitrogen “Trophic Level Effects” The nitrogen isotope values of an organism at each level depend on the value of its nitrogen source, and on the metabolic effects within the organism. Nitrogen metabolic effects are more pronounced than is the case in carbon, and accumulate during transfer in the food web, thus causing a substantial range in mean values of organisms. Nitrogen trophic level effects, expressed as the difference between diet and bone collagen nitrogen, or are caused by biochemical kinetic isotope fractionation effects plus fractionation that happens in urea metabolism and excretion in the kidneys. Kinetic fractionations were observed by Macko et al. (1986) during amino acid transamination by the enzyme glutamate-oxaloacetate transaminase, an enzyme shared by the majority of organisms and utilized to transfer nitrogen from one amino acid to the other. As a result, the nonessential amino acids (i.e., synthesized by the body, not originated intact from food) show enrichments in when food and body tissue are compared (Gaebler et al. 1966). Also, physiological effects in urea production and kidney function will play a role, especially when the organism relies on water retention mechanisms in arid circumstances (Ambrose 1991). Stepwise enrichment in nitrogen was first observed in marine (coastal) ecosystems (Minagawa and Wada 1984; Miyake and Wada 1967). The

latter quote an average

of

per trophic level, occurring

48

GERT J. VAN KLINKEN ET AL.

independently of habitat, form of nitrogen excretion and growth rate. Similar amounts of enrichment (about +3–4‰) were also reported in terrestrial ecosystems (Bocherens et al. 1994a, 1994b, 1991; Fizet et al. 1995; Schoeninger et al. 1983; Vogel et al. 1990), and in controlled diet studies with rodents and pigs (Gaebler et al. 1966; DeNiro and Epstein 1981; Hare et al. 1991). We will discuss relevant data from our data base below. There are exceptions to this general trend of an evenly-spaced stepwise enrichment: African ruminants often show much larger (Ambrose 1986; Ambrose 1991; Heaton et al. 1986; Sealy et al. 1987) due to water stress. Waterconserving species such as gazelle, hyrax and dikdik have significantly higher values than water-dependent species such as waterbuck and wildebeest (Ambrose 1991). Differences in due to water stress are an environmentally induced variability, and is a factor to consider only in the more arid zones of Europe such as the Mediterranean, although we found no indications for this in our data base (see below). It can be argued that the gut flora provides an additional trophic level in ruminants (Steinhour et al. 1982). This should result in a larger in ruminants than in non-ruminant herbivores. As many domesticated animals are ruminants this is a factor that has to be taken into account. Available data does not seem to show systematic differences between the categories ruminants and non-ruminants, although systematic species differences exist. How do carbon and nitrogen trophic level effects compare? We have argued that the cause of shifts in carbon might be diverse, and that these shifts are probably variable in size. Especially because of this ambiguity we feel that preferably the term trophic level effect should be avoided for carbon. Given that nitrogen trophic level effects are much more pronounced and universally fairly similar in size (possibly outside of arid areas), it should be possible to calculate food values by subtracting the trophic level effect from ancient bone values.

Trophic Level Effects in Humans What are the reported carbon and nitrogen effects in humans? Not much data has been published regarding herbivore-human carbon shifts. We will discuss our data in a later section. More is known about nitrogen. In a study of modern humans where diet components (protein, lipid and carbohydrate) were measured against the corresponding body components, a shift of between 4.2 and 4.4‰ was observed for nitrogen in both plasma protein and hair (Schoeller et al. 1986). This is just outside the usual 3–4‰ range. Salmon fishers from coastal British Columbia are enriched by 3‰ compared to their diet (Chisholm et al. 1983). Ancient Mexicans have constant values, as reported by DeNiro and Epstein (1981): 8–10‰, and While and Schwarcz (1989): In the latter

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

49

study, ancient Maya diet shows a of 4.5‰ (humans-herbivores). The values for mixed-diet humans in Schoeninger et al. (1983) seem always somewhat too positive for their supposed food: European agriculturalists are about 8–10‰. Also, Bocherens et al. (1991) and Lubell et al. (1994) give similar values (around +9‰) for medieval French and Neolithic humans from Portugal, respectively. The values (+9.3 and 11.6‰) of two human (Neanderthal) samples (Fizet et al. 1995) are very similar to those of

associated carnivores but are only slightly higher than those of Neolithic humans.

Anthropogenic Changes The last cause for isotopic change is the influence that humans have had, and still have, on their environment, one which also has consequences for the isotopic composition of food items. Many natural ecosystems exhibit low values in plants and organic soil components as a result of a tightly closed cycle of N (Högberg 1990). There are indications that increased human

working of the soil disturbs this close cycle and causes an often dramatic increase (5–8‰) in values of soil nitrogen and plants (Mariotti 1982). The application of increasing amounts of natural fertilizers also dramatically increases values in plants (Högberg 1990; Meints et al. 1975). The “manuring effect” will be transported up the food chain, and influence values in human consumers. It seems highly likely that when improved farming techniques increased the working of land, shifts in soil and plant values resulted. Thus there is a need to check for anthropogenic effects in the archaeological food chain, which can be done by measuring associated plant and animal remains.

Freshwater Fish In Europe, freshwater fish bones have been found in some inland sites. or marine resources, as there is no appreciably-sized set of data available. There are indications (Fry 1991; Hesslein et al. 1991; Hobson 1990; Katzenberg 1989; Minagawa and Wada 1984; Schoeninger and DeNiro 1984) that values are sometimes significantly more negative, although quite variable, and that values are higher than those of terrestrial animals but comparable to marine fish. Consumption of both freshwater and marine fish increases values in the consumer. Much of the high values are linked with the trophic level of the consumed fish. Fish that are used for human consumption are often carnivores, and will have high values (Fry 1991). Fry also noted a difference of 2–3‰ between lake and river fish of comparable trophic level. Katzenberg (1989), Hobson (1990), and Hesslein (1991) found the following in Canada: much variability in values of carnivorous fish (–23 to –30‰) Testing for freshwater fish consumption is not so straightforward as for

50

GERT J. VAN KLINKEN ET AL.

and values ranging from 10 to over 14‰ in lake fish, and lower values for river fish (7 to 9‰).

Does Collagen

Represent Total Diet or Mainly Protein?

The next aspect is the focus of the most recent debate in palaeodietary reconstructions: the main assumption in (palaeo-) diet reconstructions with stable isotopes is that mixing of two isotopically distinct food sources is essentially linear, i.e., that collagen adequately reflects the whole body composition. Lately, it has become increasingly clear that the isotopic ratio of collagen, a protein, reflects more the protein component in the diet than the overall diet. This can lead to a significant bias such as over-representation of the protein component in diet reconstructions based on collagen measurements (Ambrose

and Norr 1993; Tieszen and Fagre 1993b). This would be the case in e.g., a coastal human population where marine seafood, which is mainly protein, supplements an otherwise land-derived diet (Keegan and DeNiro 1988; Van Klinken 1991), which will almost always have a major non-protein component. In this case collagen values will overestimate the marine component in the diet. This problem has been the focus of experimental work (Ambrose and Norr 1993; Tieszen and Fagre 1993b), and modelling efforts (Hedges and Van Klinken this volume; Schwarcz this volume; Krueger and Sullivan 1984; Lee-Thorp et al. 1989). It now seems an established fact that hydroxy-

lapatite carbonate more reflects the general catabolic (fuel) component of metabolism: it reflects the of the respiratory shifted due to the +9‰ fractionation that occurs during carbonate precipitation from the bicarbonate pool in the blood (Mook 1989), while collagen carbon is more representative for the protein input (Ambrose and Norr 1993; Tieszen Fagre 1993). The spacing is smaller in carnivores than in herbivores (Lee-Thorp et al. 1989), which would make it possible to measure the degree of “carnivory” in humans. Few values for human spacings have been reported. From Fig. 6 of Krueger and Sullivan (1984) we can extract values of around 7‰ for based diets, 5‰ for diets; marine diets and African pastoralists are about 2‰. Lee-Thorp’s (1989) mean value of coastal hunter-gatherers from the southwestern Cape is These values for marine-based, and East African pastoralist diets are considerably lower than those of carnivores (with a of around 4), which seems to imply that these humans were more carnivorous than carnivores, which is impossible. The explanation must be found by looking into which selection of food types has been made: usage as “fuel” of depleted lipids or enriched proteins, as Lee-Thorp suggests. We refer to Hedges and Van Klinken (this volume) for further discussion of the theoretical aspects of spacing, and will include measurements of apatite-collagen spacings in future measurements.

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

51

Nitrogen is mainly taken in by humans in the form of proteins, so there is no possibility for ambiguity here: nitrogen isotopic ratios must reflect those of the protein in the diet.

Plant Versus Animal Protein, and

Measurements

One of the most useful aspects of nitrogen measurements (together with apatite-collagen spacings) in the European archaeological context is the potential of measuring meat consumption. The higher the nitrogen value, the more the ingested protein originates from higher trophic levels (i.e., meat, milk and dairy products). However, the proportion of protein is not the same in plant food as in meat (Schwarcz et al. 1985). Plants are low to very low in protein, while animal flesh is high. Maize has 11% protein based on 1.75% N (wt. %) and a nitrogen:protein conversion factor of 6.25, while meat has 90% protein (14.4% N). This means that a relatively small weight proportion of animal protein in a diet has a large effect on the consumer tissue The archaeologically most meaningful measure would be the proportion of animal protein in the total diet (weight consumed of total weight). What is a more straightforward indicator of meat intake based on measurements, is the proportion of animal protein of the total protein intake. When considering this, it is important to keep the following in mind:

1. Total intake is in fact the digested part of the food intake., as the part of the food that is excreted is neither used for energy generation nor for anabolic activity. 2. The protein content of plant foods is most of the time so different from that of animal-derived foods that large amounts of plants will have to be consumed with relatively little effect on consumer values. For an archaeologically meaningful interpretation of isotopic measurements this “weighting problem” needs to be considered. 3. It is impossible to differentiate between high and low level meat consumption if the main protein input is only from one source (for instance meat), assuming that the anabolic minimum N requirement is met. An example will illustrate this. Assuming that no plant protein is consumed, it is not possible to distinguish between someone who eats 100g and another person who eats 500g animal protein. The point at which people are de facto single-source protein consumers needs to be established: is it when plant protein drops below 5% total protein, or is it at higher levels? All these factors make comparisons with other archaeological indicators of paleodiet more complicated. Nevertheless, nitrogen trophic level effects,

together with measurement of the

spacing, seems to be the best

way to quantify proportions of animal protein in the diet.

52

GERT J. VAN KLINKEN ET AL.

COMPARISON OF ANIMAL AND HUMAN DATA FROM EUROPE Assessment of the Current State of Research The discussion above lacks basic data: the purpose of our inventory is mainly to raise issues that need to be addressed in the future, and to try to develop a framework that relates these issues to each other, than to supply this lacking data. Because of that, the question of whether aspects of isotopic variation discussed above can be unequivocally identified in the archaeological record in Europe cannot yet be answered. We can, however, state that some

form of patterning (as opposed to random variation) can often be observed. In many cases we observe patterns without knowing the precise causes, conceivably because they are the result of more than one factor: e.g., a climatic and a cultural effect.

Below we will describe some individual cases where we think that may be possible to pinpoint the cause for isotopic variations, and will try to give a general idea of human variations that can be found in Europe. The characterization will not be comprehensive, given the fact that we have only recently started analyzing and comparing isotopic data across Europe, but we hope that a start can be made with matching actually observed variations and the causes behind them.

Results from our Data Base Currently we have a data base of values from approximately 800 samples of bone collagen extracted from human and faunal bones that were submitted for radiocarbon dating at both the University of Oxford and Groningen University dating laboratories. Approximately 250 of these samples also have associated values. Diagenetic contamination of bone by humic substances can change the C/N ratio and the value of bone collagen. We

only included a sample if the collagen C/N ratio was between 2.7 and 3.3, to help ensure that we were measuring intact collagen (DeNiro 1985). Contamination can also shift the to more negative values (up to –24‰). However, we did not exclude more negative values, mainly because otherwise that would have eliminated indications of canopy effects and fresh-water fish consumption. 13

C/12C Ratios: Climatic Correction

Before we could begin to compare the values from different regions of Europe we needed to climate-correct the values, because natural climatic

differences result in different values for different regions, in plants (Fig. 3.1) and in bone collagen (Van Klinken et al. 1994). Climate correction

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

53

eliminates environmental variability, and therefore remaining differences should be entirely due to human food choice. The regression line for wood, taken from Van Klinken et al. (1994) was used to calculate expected values, which were then used to climate-correct our bone values. The relationship we find in wood can be expressed as:

where y = the expected value (in ‰), and x is the average July daily temperature in °C. Similar results were obtained by Stuiver and Braziunas (1987). We used modern weather data from Müller (1982) to obtain the modern average July temperature at the weather station nearest to the archaeological site where each of our samples originated. We calculated the difference between an arbitrary value of –26‰ for wood and the expected value from the climate data. This difference was then added to the measured bone values, resulting in a climate-corrected value Table 3.1 provides an example of the utility of climate correction. It compares human bone values from two very climatically different areas: Scotland and Spain. Before climate-correction the of the Scottish samples were quite negative, while the Spanish samples were more positive. One could incorrectly conclude that the Spanish values are due to a greater input of marine food than the Scottish sample. After correcting for climate variation we can see that the two samples are actually quite similar, and both reflect a terrestrial diet.

54

GERT J. VAN KLINKEN ET AL.

Figure 3.2 shows the average uncorrected

and climate corrected values for all mammalian species analyzed for certain countries plotted on a map of Europe. The values in Germany, Netherlands, Belgium, France, Austria, and Russia have been altered the least by the climate correction, and all have similar average values. The values in Ireland, Scotland, England, Sweden, Spain and Malta (and to a lesser extent Greece) have been brought closer to the average values of the above countries. A possible explanation for the variation observed between countries is that it may not always be valid to group together values in a political, rather than a geographic unit. As an example, Italy is very diverse geographically, and the reported average contains measurements from sites in the Alps as well as the south.

Values of Animals and Humans Table 3.2 shows the and values of herbivores, omnivores, carnivores and humans. The (climate-corrected) trophic level effect between herbivores and carnivores is 0.90‰. Human values are closer to carnivore and omnivore values than to herbivore values. The human values are on average 0.66‰ more positive than the herbivore values, a good estimate for a carnivore effect in humans (see section on trophic level effects, below). The average human value is which would indicate that Holocene humans in Europe had a diet that consisted of terres-

trial foods, which is as might be expected. By looking at the humans separate from the total bone data set, we notice potential human food selection (Fig. 3.3): we can see a non-climatic pattern, which is much less uniform than in the total bone data set (Fig. 3.2b). Italy has a much more negative value than the Czech Republic Spain = –19.3‰) and Greece (–18.9‰; but the of 9.0‰ does not indicate marine food), while the northern European countries are closer to a value of –20‰. What the actual causes are for this pattern in the human

samples is not clear; to better understand these variations it is best to consider, where possible, the values with the values.

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

55

Figure 3.3. Average carbon isotopic ratios for human bone collagen samples, corrected for climatic trends. Only countries with more than 10 samples are included. For a description of the climate correction procedure see text.

Values Table 3.3 shows the average values for different animal species taken from our data base. Unfortunately the number of carnivores and omnivores is considerably smaller than the herbivores and humans. The average values are similar to values published elsewhere (Fizet et al. 1995; Schoeninger 1985). Because of the connection of variability in human values with differences in dietary choice, it seems less meaningful to generalize about average human data in the same manner as for the animal categories. Nevertheless a few observations can be made. The difference between human

and herbivore (for instance, average values are the greatest. The human values are surprisingly high, more similar to the carnivore values than to the omnivore At face value this would mean that all humans that were measured were high level meat eaters, and certainly were not vegetarians. More realistically, human consumption patterns are in between herbivores and carnivores. Several explanations can be offered for high values among humans.

56

GERT J. VAN KLINKEN ET AL.

In the first part of our paper, we proposed three possible explanations for high human values. Firstly, if the trophic level model used to explain nitrogen values is valid it appears that humans fall into the higher end of the observed values, and are therefore behaving like carnivores. Second,

human agricultural practices, particularly manuring, may alter the value of domesticated plants. Modern plants grown in manured fields can have values as high as 8‰. However, this does not account for high values we have observed for humans from pre-agricultural societies, or for the lower values observed in agricultural societies. Third, it is also possible that inland freshwater fish was used in some areas, which may have higher values. Fourth, high human values may be partly explained by the fact that even among agriculturalists most protein may come from animals because plants have low proportions of protein. Furthermore, analyzing our 15N data in a similar manner as we did with the 13C data does not reveal similar climaterelated patterning across Europe. We have found indications of temporal changes in humans older than 10.000 BP (Van Klinken, in prep), but within the Holocene the 13N variations seem to be more linked with non-climatic effects associated with interactions between soil nutrients and vegetation than with climate. Finally, it is possible that humans have high values for physiological reasons, although strict vegetarians (vegans) are quite comparable to herbivores (O’Connell 1996). We mentioned above that published measurements of human trophic level effects seem to be larger than in carnivores, which is another indication that humans are systematically different in their

nitrogen isotopic values, and that a physiological cause is conceivable. It is not the purpose of this paper to described human isotopic variability in Europe in detail (this will be done elsewhere: Richards, in prep), but to

ENVIRONMENTAL, ECOPHYSIOLOGICA1, AND CULTURAL EFECTS

57

illustrate the diversity of

in Europe, we have presented average human values from a number of archaeological sites in Table 3.4. The variation observed between sites from different countries is about 4‰, and may be due to local differences in plant and animal values, as well as some of the cultural effects discussed above. However, differences between sites from the same area, such as the Bronze and Iron Age sites in England, are unlikely to be due

to “natural” variation, and are most likely related to dietary differences (such as different proportions of plant and animal protein). Before we can decide the best models to explain the human values from specific sites we need to take a number of factors into account. We need to look at the values from the range of faunal remains present at the sites, especially at differences between wild species and domesticated animals. Measuring the of plant remains would also be very useful for looking at the question of the influence of manuring or other cultivation effects on their values, and for their trophic level relationship with the humans. A comparison between human bone carbonate and collagen values, combined with the data will provide two measures for the amount of meat consumption in human populations. We are beginning to start this line of research, and hope to soon publish some results from selected sites.

CONCLUSIONS Isotopic variation in European human remains is caused by a combination of environmental, biological and cultural factors. For carbon, the main influence is the dominance of photosynthesis in all vegetation types. The

factors that can cause deviations from this general trend are the climatic effect,

58

GERT J. VAN KLINKEN ET AL.

canopy effect, consumption of river or marine fish, and occasional consumption of millet in the south. Climatic effects can be adequately corrected; and the canopy effect can be demonstrated by measuring local herbivores. When

two or more of these factors occur together, discrimination might be difficult (for example, canopy effect and fresh water fish consumption give the same result). For nitrogen, the main determinants are the nitrogen isotopic signature of plants, and variation in the degree of animal versus plant protein consumption (the degree of “carnivory”). The former is notoriously difficult to analyze, and is the main “unknown” in the analysis of most local food webs; the latter is complicated by differences in between herbivores. The factors that can cause extra complication are the consumption of marine or fresh water fish, and potentially anthropogenic effects on ecosystems. Most variation in human values in Europe is related to regional differences in climate; most of the variation in seems to be caused by variable consumption of animal protein, combined with poorly understood variation in plant nitrogen values. Finally, we also conclude that the causes of slightly high values in humans are not yet well-understood.

REFERENCES Ambrose, S.H. 1986 Stable carbon and nitrogen isotope analysis of human and animal diet in Africa. Journal of Human Evolution 15: 707–731. ___1991 Effects of diet, climate and physiology on nitrogen isotope abundances in terrestrial food webs. Journal of Archaeological. Science 18: 293–317. ___1993 Isotopic analysis: Methodological and interpretive considerations. In Sandford, M.K., ed., Investigations of Ancient Human Tissue: Chemical Analysis in Anthropology. Langhorne, PA.: Gordon and Breach Scientific: 59–130. Ambrose, S.H. and DeNiro, M.J. 1986 The isotopic ecology of East African mammals. Oecologia 69: 395–406. ___1989 Climate and habitat reconstruction using stable carbon and nitrogen isotope ratios of collagen in prehistoric herbivore teeth from Kenya. Quaternary Research 31: 407–422. Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of bone collagen and carbonate. In Lambert,

J.B. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 1–37. Aucour, A.-M., Hillaire-Marcel, C. and Bonnefille, R. 1993 A 30,000 year record of 13C and 18O changes in organic matter from an equatorial peat bog. Geophysical. Monographs 78: 343–351.

Aufderheide, A.C., Kelley, M.A., Rivera, M., Gray, L., Tieszen, L.L., Iversen, E., Krouse, H.R. and Carevic, A. 1994 Contributions of chemical dietary reconstruction to the assessment of adaptation by ancient highland immigrants (Alto Ramirez) to coastal conditions at Pisagua, North Chile. Journal of Archaeological Science 21: 515–524. Becker, B., Kromer, B. and Trimborn, P. 1991 A stable-isotope tree-ring timescale of the Late Glacial/Holocene boundary. Nature 353: 647–649. Bender, M.M., Baerreis, D.A. and Steventon, R.L. 1981 Further light on carbon isotopes and Hopewell agriculture. American Antiquity 46: 346–353.

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

59

Binkley, D., Sollins, P. and McGill, W.B. 1985 Natural abundance of nitrogen-15 as a tool for tracing

alder-fixed nitrogen. Soil Science Society of America Journal 49: 444–447. Bocherens, H., Fizet, M. and Mariotti, A. 1994a Diet, physiology and ecology of fossil mammals as inferred from stable carbon and nitrogen isotope biogeochemistry: implications for

Pleistocene bears. Palaeogeography, Palaeodimatology, Palaeoecology 107: 213–225. Bocherens, H., Fizet, M., Mariotti, A., Gangloff, R.A. and Burns, J.A. 1994b Contribution of isotopic biogeochemistry (13C, 15 N, 18O) to the paleoecology of mammoths (Mammuthus primigenius). Historical Biology 7: 187–202. Bocherens, H., Fizet, M., Mariotti, A., Olive, C., Bellon, G. and Billiou, D. 1991 Application de la biochimie isotopique (13C, 15N) à la détermination du regime alimentaire des populations humaines et animates durant les périodes antique et médédvale. Archives des Sciences de

Genève 44: 329–340. Chisholm, B.S., Nelson, D.E. and Schwarcz, H.P. 1982 Stable-carbon isotope ratios as a measure of marine versus terrestrial protein in ancient diets. Science 216: 1131–1132. __1983 Dietary information from and measurements on bone collagen. PACT 8:

391–397. Clutton-Brock, J. and Noe-Nygaard, N. 1990 New osteological and C-isotope evidence on Mesolithic dogs: Comparison to hunters and fishers at Star Carr, Seamer Carr and Kongemose. Journal of Archaeological Science 17: 643–653. Davidson, E., Hart, S.C., Shanks, C.A. and Firestone, M.K. 1991 Measuring gross nitrogen mineralization, immobilization, and nitrification by 15N isotopic pool dilution in intact soil cores. Journal of Soil Science 42: 335–349.

Davidson, E.A., Hart, S.C. and Firestone, M.K. 1992 Internal cycling of nitrate in soils of a mature coniferous forest. Ecology 73: 1148–1156. Delwiche, C.C. and Steyn, P.L. 1970 Nitrogen isotope fractionation in soils and microbial reactions. Environmental Science and Technology 4: 929–935. DeNiro, M.J. 1985 Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to palaeodietary reconstruction. Nature 317: 806–809. DeNiro, M.J. and Epstein, S. 1981 Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica et Cosmochimica Acta 45: 341–351. Domenach, A.M., Kurdali, F. and Bardin, R. 1989 Estimation of symbiotic dinitrogen fixation in alder forest by the method based on natural 15N-abundance. Plant and Soil 118: 51–59.

Epstein, S. and Krishnamurthy, R.V. 1990 Environmental information in the isotopic record in trees. Philosophical Transactions of the Royal Society London A 330: 427–439. Fizet, M., Mariotti, A. and Bocherens, H. 1995 Effect of diet, physiology and climate on carbon and nitrogen stable isotopes of collagen in a Late Pleistocene anthropic palaeoecosystem: Marillac, Charente, France. Journal of Archaeological Science 22: 67–79. Francey, R.J. and Farquhar, G.D. 1982 An explanation of 13C/12C variations in tree rings. Nature 297: 28–31. Fry, B. 1991 Stable isotope diagrams of freshwater food webs. Ecology 72: 2293–2297. Gaebler, O.H., Vitti, T.G. and Vukmirovich, R. 1966 Isotope effects in metabolism of 14N and 15N from unlabeled dietary proteins. Canadian Journal of Biochemistry 44: 1249–1257.

Garten, C.T., Jr. 1993 Variation in foliar 15N abundance and the availability of soil nitrogen on Walker Branch watershed. Ecology 74: 2098–2113.

Handley, L.L. and Raven, J.A. 1992 The use of natural abundance of nitrogen isotopes in plant physiology and ecology. Plant Cell Environment 15: 965–985. Handley, L.L. and Scrimgeour, C.M. in press Terrestrial plant ecology and 15N abundance: the present limits to interpretation for uncultivated systems with original data from a Scottish

Old Field. Advances in Ecological Research. Hare, P.E., Fogel, M.L., Stafford, T.W., Jr., Mitchell, A.D. and Hoering, T.C. 1991 The isotopic composition of carbon and nitrogen in individual amino acids isolated from modern and fossil proteins. Journal of Archaeological Science 18: 277–292.

60

GERT J. VAN KLINKEN ET AL.

Heaton, T.H.E. 1987 The

ratio of plants in South Africa and Namibia: relationship to

climate and coastal/saline environments. Oecologia 74: 236–246. Heaton, T.H.E., Vogel, J.C., Von la Chevallerie, G. and Collett, G. 1986 Climatic influence on the isotopic composition of bone nitrogen. Nature 322: 822–823. Hedges, R.E.M. and Van Klinken, G.J. 2000 “Consider a Spherical Cow. . . .”—On Modelling and Diet. In Ambrose, S. and Katzenberg, M.A., eds., Biogeochemical Approaches To Paleodietary Analysis. Kluwer Academic/Plenum Publishers, New York. Hesslein, R.H., Capet, M.J., Fox, D.E. and Hallard, K.A. 1991 Stable isotopes of sulphur, carbon, and nitrogen as indicators of trophic level and fish migration in the lower Mackenzie River Basin, Canada. Canadian Journal of Fisheries and Aquatic Science 48: 2258–2265. Hobson, K.A. 1990 Stable isotope analysis of marbled murrelets: evidence for freshwater feeding and determination of trophic level. Condor 92: 897–903. Högberg, P. 1990 Forests losing large quantities of nitrogen have elevated 15N/14N ratios. Oecologia 84: 229–231. Huebner, H. 1985 Geographische variationen der isotopen zusammensetzung leichter elemente im tierischen, insbesondere menschlichen organismus. Mitteilungen Zentralinstitut fuer Isotopen- und Strahlenforschung, Leipzig 106: 1–91.

Katzenberg, M.A. 1989 Stable isotope analysis of archaeological faunal remains from Southern Ontario. Journal of Archaeological Science 16: 319–329. Keegan, W.F. and DeNiro, M.J. 1988 Stable carbon- and nitrogen-isotope ratios of bone collagen used to study coral-reef and terrestrial components of prehistoric Bahamian diet. American Antiquity 53: 320–336. Keeling, C.D. 1961 The concentration and isotopic abundances of atmospheric carbon dioxide in rural and marine air. Geochimica et Cosmochimica Acta 24: 277–298. Keeling, C.D., Bacastow, R.B., Carter, A.F., Piper, S.C., Whorf, T.P., Heimann, M., Mook, W.G. and Roeloffzen, H. 1989 A three-dimensional model of atmospheric transport based on observed winds: 1. Analysis of observational data. Geophysical Monographs 55: 165–236. Keeling, C.D., Bacastow, R.B. and Tans, P. 1980 Predicted shift in the 13C/12C ratio of atmospheric carbon dioxide. Geophysical Research Letters 7: 505–508. Krueger, H.W. and Sullivan, C.H. 1984 Models for carbon isotope fractionation between diet and

bone. In Turnlund, J.S. and Johnson, P.E., eds., Stable Isotopes in Human Nutrition. American Chemical Society Symposium Series 258. U.S.A., American Chemical Society: 205–220. Leavitt, S.W. and Danzer, S.R. 1992 variations in plants over the past 50,000 years. Radiocarbon 34: 783–791. Leavitt, S.W. and Long, A. 1991 Seasonal stable-carbon isotope variability in tree rings: possible palaeoenvironmental signals. Chemical Geology (Isotope Geoscience Section) 87: 59–70. Leavitt, S.W. and Newberry T. 1992 Systematics of stable-carbon isotopic differences between gymnosperm and angiosperm trees. Plant Physiology (Life Science Advances) 11: 257–262. Lee-Thorp, J.A., Sealy, J.C. and Van der Merwe, N.J. 1989 Stable carbon isotope ratio differences between bone collagen and bone apatite, and their relationship to diet. Journal of Archaeological Science 16: 585–599. Leuenberger, M., Siegenthaler, U. and Langway, C.C. 1992 Carbon isotope composition of atmospheric from an Antarctic ice core. Nature 357: 488–490. Loader, N.J., Switsur, V.R. and Field, E.M. 1995 High-resolution stable isotope analysis of tree rings: implications of “microdendroclimatology” for palaeoenvironmental research. The Holocene 5: 457–460.

Lubell, D., Jackes, M., Schwarcz, H., Knyf, M. and Meiklejohn, C. 1994 The Mesolithic-Neolithic transition in Portugal: isotopic and dental evidence of diet. Journal of Archaeological Science 21: 201–216. Macko, S.A., Estep, M.L.F., Engel, M.H. and Hare, RE. 1986 Kinetic fractionation of stable nitrogen isotopes during amino acid transamination. Geochimica et Cosmochimica Acta 50: 2143–2146.

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

61

Marino, B.D. and McElroy, M.B. 1991 Isotopic composition of atmospheric inferred from carbon in plant cellulose. Nature 349: 127–131. Mariotti, A. 1982 Apports de la géochimie isotopique à la connaissance du cycle de l’azote. Geoscience. Paris, Université Pierre et Marie Curie, Mémoires des Sciences de Terre, No 82–13. ___1983 Atmospheric nitrogen is a reliable standard for natural 15N measurements. Nature 303:

685–687. Mariotti, A., Germon, J.C., Hubert, P., Kaiser, P., Letolle, R., Tardieux, A. and Tardieux, P. 1981 Experimental determination of nitrogen kinetic isotope fractionation: some principles; illustration for the denitrification and nitrification processes. Plant and Soil 62: 413–430. Mariotti, A., Mariotti, F., Champigny, M.L., Amarger, N. and Moyse, A. 1982 Nitrogen isotope fractionation associated with nitrate reductase activity and uptake of by pearl millet. Plant Physiology 69: 880–884. Meints, V.W., Boone, L.V and Kurtz, L.T. 1975 Natural 15N abundance in soil, leaves, and grain as influenced by long term additions of fertilizer N at several rates. Journal of Environmen-

tal Quality 4: 486–490. Müller, M.J. 1982 Selected Climatic Data for a Global Set of Standard Stations for Vegetation Science. The Hague, Junk. Michelsen, A., Jonasson, S., Sleep, D., Havström, M. and Callaghan, T.V. 1996a Shoot biomass,

nitrogen and chlorophyll responses of two dwarf shrubs to in situ shading, nutrient application and warming simulating climatic change. Oecologia 105: 1–12. Michelsen, A., Schmidt, I.K., Jonasson, S., Quarmby, C. and Sleep, D. 1996b Leaf 15N abundance of subarctic plants provides field evidence that ericoid, ectomycorhizal, and non- and arbus-

cular mycorhizal species access different sources of soil nitrogen. Oecologia 105: 53–63. Minagawa, M. and Wada, E. 1984 Stepwise enrichment of 15N along food chains: Further evidence and the relation between and animal age. Geochimica et Cosmochimica Acta 48: 1135–1140. Miyake, Y. and Wada, E. 1967 The abundance ratio of 15N/14N in marine environments. Records of Oceanographic Works, Japan 9 (new series): 32–53.

Mook, W.G. 1986 13C in atmospheric

Netherlands Journal of Sea Research 20: 211–223.

____ 1989 Principles of Isotope Hydrology, introductory course on isotope hydrology. Amsterdam, Vrije Universiteit. Murray, D.L. and Schoeninger, M.J. 1988 Diet, status and complex social structure in iron age Central Europe: some contributions of bone chemistry. In Gibson, D.B. and Geselowitz, M.N., eds., Tribe and Polity in Late Prehistoric Europe. New York, Plenum: 155–176. Nadelhoffer, K., Shaver, G., Fry, B., Giblin, A., Johnson, L. and Mackane, R. in press 15N natural abundances and N use by tundra plants. Oecologia. Noe-Nygaard, N. 1988 values of dog bones reveal the nature of changes in man’s food

resources at the Mesolithic-Neolithic transition. Isotope Geoscience 73: 87–96. O’Connell, T.C. 1996 The isotopic relationship between diet and body proteins: implications for the study of diet in archaeology. Ph.D. dissertation, University of Oxford. O’Leary, M.H. 1981 Carbon isotope fractionation in plants. Phytochemistry 20: 553–567. Ramesh, R., Bhattacharya, S.K. and Gopalan, K. 1986 Stable isotope systematics in tree cellulose

as palaeoenvironmental indicators—a review. Journal of the Geological Society of India 27: 154–167. Rodière, E., Bocherens, H., Angibault, J.-M, and Mariotti, A. 1996 Particularités isotopiques de

l’azote chez le chevreuil (Capriolus capriolus L.): implications pour les reconstitutions paléoenvironnementales. Comptes Rendus de l’Académie des Sciences de Paris 323: 179–185. Saurer, M., Siegenthaler, U. and Schweingruber, F. 1995 The climate-carbon isotope relationship in tree rings and the significance of site conditions. Tellus 47B: 320–330. Schoeller, D.A., Minagawa, M., Slater, R. and Kaplan, I.R. 1986 Stable isotopes of carbon, nitro-

gen and hydrogen in the contemporary North American human food web. Ecology of Food and Nutrition 18: 159–170.

62

GERT J. VAN KLINKEN ET AL.

Schoeninger, M.J. 1985 Trophic level effects on 15N/14N and 13C/12C ratios in bone collagen and

strontium levels in bone mineral. Journal of Human Evolution 14: 515–525. Schoeninger, M.J. and DeNiro, M.J. 1984 Nitrogen and carbon isotopic composition of bone collagen from marine and terrestrial animals. Geochimica et Cosmochimica Acta 48: 625–639. Schoeninger, M.J., DeNiro, M.J. and Tauber, H. 1983 Stable nitrogen isotope ratios of bone colla-

gen reflect marine and terrestrial components of prehistoric human diet. Science 220: 1381–1383. Schultze, E.-D., Chapin, F.S. and Gebauer, G. 1994 Nitrogen nutrition and isotope differences among life forms at the northern treeline of Alaska. Oecologia 100: 406–412. Schwarcz, H.P, Melbye, J., Katzenberg, M.A. and Knyf, M. 1985 Stable isotopes in human skeletons of southern Ontario: reconstructing paleodiet. Journal of Archaeological Science 12: 187–206. Sealy, J.C., Van der Merwe, N.J., Lee-Thorp, J.A. and Lanham, J.L. 1987 Nitrogen isotopic ecology in southern Africa: implications for environmental and dietary tracing. Geochimica et Cosmochimica Acta 51: 2707–2717. Shearer, G., Duffy, J., Kohl, D.H. and Commoner, B. 1974 A steady-state model of isotopic fractionation accompanying nitrogen transformations in soil. Soil Science Society of America Proceedings 38: 315–322.

Shearer, G. and Kohl, D.H. 1986

fixation in field settings: estimations based on natural

15

N

abundance. Australian Journal of Plant Physiology 13: 699–756. Steinhour, W.D., Stokes, M.R., Clark, J.H., Rogers, J.A., Davis, C.L. and Nelson, D.R. 1982 Estimation of the proportion of non-ammonia-nitrogen reaching the lower gut of the ruminant derived from bacterial and protozoal nitrogen. British Journal of Nutrition 48: 417–431. Stuiver, M. and Braziunas, T.F. 1987 Tree cellulose 13C/12C isotope ratios and climatic change.

Nature 328: 58–60. Tieszen, L.L. 1991 Natural variations in the carbon isotope values of plants: implications for archaeology, ecology, and paleoecology. Journal of Archaeological Science 18: 227–248. Tieszen, L.L. and Fagre, T. 1993a Carbon isotopic variability in modern and archaeological maize. Journal of Archaeological Science 20: 25–40. Tieszen, L.L. and Fagre, T. 1993b Effect of diet quality and composition on the isotopic composition of respiratory bone collagen, bioapatite, and soft tissues. In Lambert, J.B.

and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 121–155. Valentini, R., Scarascia Mugnozza, G.E., De Agnelis, P. and Matteucci, G. 1995 Coupling water and carbon metabolism of natural vegetation at integrated time and space scales. Agricultural and Forest Meteorology 73: 297–306. Valentini, R., Scarascia Mugnozza, G.E. and Ehleringer, J.R. 1992 Hydrogen and carbon isotope ratios of selected species of a Mediterranean macchia ecosystem. Functional Ecology 6: 627–631. Van de Water, PK., Leavitt, S.W. and Betancourt, J.L. 1994 Trends in stomatal density and 13C/12C ratios of Pinus flexilis needles during last glacial-interglacial cycle. Science 264: 239–243. Van der Merwe, N.J. and Medina, E. 1989 Photosynthesis and 13C/12C ratios in Amazonian rain

forests. Geochimica et Cosmochimica Acta 53: 1091–1094. ___1991 The canopy effect, carbon isotope ratios and foodwebs in Amazonia. Journal of

Archaeological Science 18: 249–259. Van der Merwe, N.J. and Vogel, J.C. 1978 13C Content of human collagen as a measure of prehistoric diet in woodland North America. Nature 276: 815–816. Van Klinken, G.J. 1991 Dating and Dietary Reconstruction by Isotopic Analysis of Amino Acids in Fossil Bone Collagen—with Special Reference to the Caribbean. Ph.D. dissertation, University of Groningen, The Netherlands. Van Klinken, G.J., Van der Plicht, J. and Hedges, R.E.M. 1994 Bone 13C/12C ratios reflect (palaeo)

climatic variations. Geophysical Research Letters 21: 445–448.

ENVIRONMENTAL, ECOPHYSIOLOGICAL, AND CULTURAL EFECTS

63

Virginia, R.A. and Delwiche, C.C. 1982 Natural 15N abundance of presumed -fixing and non -fixing plants from selected ecosystems. Oecologia 54: 317–325. Vogel, J.C. 1980 Fractionation of the Carbon Isotopes During Photosynthesis. New York, Springer. Vogel, J.C., Eglinton, B. and Auret, J.M. 1990 Isotopic fingerprints in elephant bone and ivory. Nature 346: 747–749. Walker, PL. and DeNiro, M.J. 1986 Stable nitrogen and carbon isotope ratios in bone collagen as indices of prehistoric dietary dependence on marine and terrestrial resources in Southern California. American Journal of Physical Anthropology 71: 51–61. White, C.D. and Schwarcz, H.P. 1989 Ancient Maya diet: As inferred from isotopic and elemental analysis of human bone. Journal of Archaeological Science 16: 451–474.

Yoneyama, T., Murakami, T., Boonkerd, N., Wadisirisuk, P., Siripin, S. and Kouno, K. 1990 Natural 15N-abundance in shrub and tree legumes, Casuarina, and non-fixing plants in Thailand. Plant and Soil 128: 287–292.

This page intentionally left blank

Chapter

4

Preservation of Isotopic Signals in Pleistocene Mammals HERVÉ BOCHERENS ABSTRACT The preservation of biogenic isotopic signals in fossil bones and teeth is critical in order to interpret paleodiets. Some patterns of variation of these biogenic isotopic signals are characteristic of modern mammals, and their recognition in fossil samples provides a clue for the preservation of biogenic paleodietary signals. In plant food webs, the values of collagen are slightly enriched in carnivores (~–20‰) relative to herbivores (~–21‰). In carbonate hydroxylapatite, the values are higher in herbivores (~–1l‰) than in carnivores (~–14‰). Thus, the spacing values between collagen and carbonate hydroxylapatite are different in herbivores (~ +8–9‰) and carnivores (~ +4–5‰). The values in collagen show a large enrichment (~ +4‰) at each trophic step. Moreover, there is an enrichment between dentine and bone collagen in species with definite tooth growth, such as carnivores and cervids, but not in species with continuous tooth growth, such as horses. This is probably due to the effect of suckling milk in the young individual, when part of the dentine is formed. These isotopic differences between tissues within individuals between species and trophic levels can be used to check the preservation of the isotopic signatures in Pleistocene samples of cold temperate and arctic environments. Biogcochemical Approaches to Paleodietary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

65

66

HERVÉ BOCHERENS

Examples are presented using samples from Western Europe (France, United

Kingdom), Siberia and Alaska.

INTRODUCTION Before using the carbon and nitrogen isotopic abundances in bones and teeth of Pleistocene mammals to reconstruct paleodiets, it is essential to assess the preservation of biogenic isotopic signals. These signals are recorded in the organic and inorganic phases of vertebrate phosphatic tissues, and both phases are susceptible to diagenetic alteration during the fossilization processes. Several methods are available to identify alteration of collagen (the dominant protein in bones and teeth), including analysis of atomic C/N ratios and amino acid composition (DeNiro 1985; Ambrose 1990; Tuross et al. 1988; DeNiro and Weiner 1988; Bocherens et al. 1991a; Fizet et al. 1995). However, minor alterations of the amino acid composition of a collagen-like organic extract may lead to important shifts of the isotopic abundances, especially in the case of nitrogen (Grupe et al. 1993). For the carbonate incorporated into the hydroxylapatite crystals, independent monitors of alteration using elemental chemistry or crystallinity are under development, but none yet provide a consistent signal of isotopic alteration (Lee-Thorp and Van der Merwe 1991; Bryant et al. 1994). As a result, alteration of these isotopic signals can be assessed by examining the extent to which expected biological carbon and nitrogen isotope patterns are disrupted in fossils. This paper is mainly focused on bones and teeth from cold and temperate environments. Indeed, Pleistocene samples from these areas seem to have retained large proportions of original organic matter during tens of thousands of years, due to favorable taphonomic conditions, such as low temperatures and trapping in caves (Gillespie et al. 1984; Hare and von Endt 1990; Long et al. 1989). Until recently, this organic matter has been mainly used for radiocarbon dating and few applications of stable isotope analysis have been performed (Bocherens et al. 1991a, 1991b, 1994, 1995a, 1995b, 1996; Koch, 1991; Fizet et al. 1995). Continental environments in temperate and cold high latitudes have much less variability in food web stable isotopic composition than in tropical areas. The two different photosynthetic pathways (i.e., the photosynthetit pathway used by all trees and shrubs and the photosynthetic pathway used by savanna grasses) have very different carbon isotope ratios (Bender 1968; Smith and Epstein 1971; Deines 1980; O’Leary 1988). Both types of plants occur in tropical areas, and the relatively strict selection of these plants by browsers versus grazers, provides a way to predict carbon isotopic values in herbivorous mammals from these areas (Vogel 1978; Ambrose and DeNiro 1986; Van der Merwe 1986; Sillen 1988; Koch et al. 1991). In a natural state, only plants are present in cold and temperate environments, since plants have a greater efficiency of uptake under limited amounts of water

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

67

and high summer temperatures (Berry 1975). Ecological conditions at low latitudes seem also to generate more variation in nitrogen isotopic abundances than in high latitudes (Ambrose and DeNiro 1986; Heaton et al. 1986; Sealy et al. 1987). Unfortunately, the climatic conditions in tropical areas preclude preservation of organic matter in bones in tropical areas for long periods; thus most of the isotopic data are retrieved from the inorganic (carbonate in

hydroxylapatite) fraction of these pre-Holocene fossil bones and teeth (LeeThorp and Van der Merwe 1993).

BIOGENIC ISOTOPIC SIGNALS USED TO ASSESS THE PRESERVATION OF ISOTOPIC SIGNALS IN FOSSIL SAMPLES FROM C3 PLANT FOOD WEBS

Carbon-13 In a wide range of animals the values of the whole body are enriched by about l‰ relative to the diet (DeNiro and Epstein 1978). In large mammals, collagen is enriched by around 5‰ relative to the diet and presumably +4‰ relative to the whole body. This enrichment is similar in herbivores and carnivores (Vogel 1978; Vogel et al. 1990). There is a slight enrichment in between herbivore and carnivore collagen, up to 2‰, at each trophic step (Van der Merwe 1986; Lee-Thorp et al. 1989). However, it has not been observed in all terrestrial ecosystems studied to date. This difference in collagen values between herbivores and carnivores seems to be observed only in ecosystems where all the plants have a photosynthetic pathway (Van der Merwe 1986). Variations in bone collagen in foodwebs are illustrated in Figure 4.1. The difference between the values of enamel carbonate hydroxylapatite between herbivores and carnivores has been attributed to the isotopic difference between lipids and carbohydrates of their respective food items (Krueger and Sullivan 1984), although more recent investigations support an interpretation of a homogenization of dietary carbon from all macronutrient fractions, including protein (Ambrose and Norr 1993; Tieszen and Fagre 1993). The high content of isotopically depleted lipids in the carnivore diet would tend to shift the carbonate hydroxylapatite values to more negative values. In carbonate hydroxylapatite, the values are enriched by around +10‰ in carnivores and +14‰ in herbivores relative to the average body (Bocherens and Mariotti 1992). The first value is similar to the results of feeding experiments on mice and rats (Ambrose and Norr 1993; Tieszen and Fagre 1993). Large herbivores are usually ruminants or at least species with microbial symbionts in their digestive tracts. It is noteworthy that methanogenic microbes in herbivore guts produce highly -enriched when they produce -depleted methane (Metges et al. 1990). This will lead to higher carbonate values than in carnivores and other nonmethanogenic-symbiont species (Hedges, this volume). Thus, there are

68

HERVÉ BOCHERENS

Figure 4.1.

values of bone collagen in

plant ecosystems.

differences in the spacing value between collagen and carbonate hydroxylapatite in herbivores (~ +8–9‰) and carnivores (~ +4–5‰) (Bocherens and Mariotti 1992). This pattern looks slightly different than in sub-Saharan Africa, where the spacing value for herbivores is lower, around 7‰ (Lee-Thorp 1989; Koch et al. 1990). Nitrogen-15 A clear enrichment in has been shown to exist between herbivore and carnivore collagen. This enrichment in is reported to range in terrestrial ecosystems from 2.8‰ (Schwarcz 1991) to 5.7‰ (Ambrose and DeNiro 1986) with an average value between 3 and 4‰ (Schoeninger and DeNiro 1984; Schoeninger 1985; Sealy et al. 1987). The nursing of young mammals makes them consume a diet one trophic level higher than their adult diet, and thus enriched in (Fogel et al. 1989). This fact, connected to the fact that dentine starts to accumulate before the end of nursing in some mammal species (Hillson 1986), leads to an enrichment between dentine and bone collagen in species with definite tooth growth, such as carnivores and deer, but not in species with continuous tooth growth, such as horses (Bocherens et al. 1994). Some environmental parameters, such as aridity and dietary stress, also seem to increase nitrogen isotopic abundances at each trophic level (Ambrose and DeNiro 1986; Heaton et al. 1986; Sealy et al. 1987; Ambrose 1991).

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

69

Variations in bone collagen nitrogen isotope ratios in C3 foodwebs are illustrated in Figure 4.2. In this paper, I attempt to refine the predictable isotopic differences between collagen and carbonate that can be found in modern faunas from temperate and cold areas, using samples from Europe, Siberia and northwestern North America. Some of the results presented here have been published previously (Bocherens et al. 1991a, 1991b, 1994, 1995a, 1995b, 1996; Bocherens and Mariotti 1992; Fizet et al. 1995) but additional new data are reviewed as well in order to present a new synthesis. This should provide a framework that can be used to assess the quality of preservation of the isotopic signatures in Pleistocene mammal bones and teeth from these areas.

MATERIALS AND METHODS

Materials Modern samples of herbivorous and carnivorous species come from

Europe, Siberia and northwestern North America (Alaska and British

Figure 4.2.

values of mammal bone collagen in

plant ecosystems.

70

HERVÉ BOCHERENS

Columbia). They have been killed recently or have been obtained from museum collections (Table 4.1). The latter may be as much as several decades old. Herbivorous species have been sampled. Bears have been treated separately from other carnivores due to their omnivorous rather than strictly

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

71

carnivorous diet (Herrero 1978; Clevenger et al. 1992), and because of the special interest in the ursid group in Pleistocene localities (Bocherens et al. 1990, 1991b, 1994, 1995a).

Pleistocene samples come from several localities: Two localities yielded an abundance of herbivores and carnivores, Marillac, France, 40,000 to 45,000 BP, with 12 well-defined stratigraphic layers (Bocherens et al. 1991a; Fizet et al. 1995) and Kent’s Cavern, Great Britain, around 35,000 to 60,000BP (years before present), without a good stratigraphic record (Bocherens et al. 1995b). Four additional French localities provided mostly bears, and some complementary herbivore and carnivore samples (Bocherens et al. 1994): Aldéne Cave, ~30,000 to 450,000 BP, comprising three chronologically distinct units but without a good stratigraphic record within each unit (Bocherens et al. 1991b), Mialet Cave, ~30,000 to 60,000 BP, without a good stratigraphic record), Azé Cave, ~200,000 BP, and Escale Cave, ~600,000 BP. Samples from herbivorous and carnivorous species from three localities near Fairbanks, Alaska, Late Wisconsin, around 40,000–10,000 BP, have been considered as well (Bocherens et al. 1995a). Pleistocene samples of extant herbivores (bison, horse) and carnivores (wolf, lion) come from Siberia, around 50,000–10,000 BP (Bocherens et al. 1996).

Methods The extraction of collagen from recent bones was performed using the protocol described in Bocherens et al. (1991a). The biochemical purity of collagen was routinely checked by measurement of the quantity of carbon and nitrogen in the extracted organic matter. Isotopic values have been measured

on and either by a modified Dumas combustion (Bocherens et al. 1991a), or by combustion in a CHN elemental analyzer connected to a mass spectrometer (Bocherens et al. 1995a). Both techniques have been carefully

intercalibrated and give identical isotopic results. For isotopic analysis of the inorganic phase, the samples have been pretreated according to Bocherens et al. (1991a). The powdered bone of recent or enamel of fossil animals was soaked in 2–3% NaOCl for up to 3 days at 20°C to oxidize organic residues, rinsed with distilled water, then treated with 1M acetic acid-Ca acetate buffer (pH = 4.75) for 20hrs at 20°C to remove exogenous carbonate without losing too much of the bone mineral. Carbon dioxide was produced from the treated powders by reaction of ~40 mg with in 100% at 25°C for 3 days or at 50°C for 5 hours. Carbon dioxide was purified by cryogenic distillation on a vacuum line, and

carbon isotope composition was measured on a VG Sira 9 gas source mass spectrometer. Isotopic abundances are normalized to international and

internal laboratory calcite standards that are analyzed concurrently with the apatite samples.

72

HERVÉ BOCHERENS

RESULTS ON MODERN SAMPLES values of collagen range from –23.4 to –18.8‰ (Table 4.1). In herbivores, the values of collagen cover the whole range, whereas the range is narrower for bears (–21.5 to –18.9‰) and for carnivores (–21.6 to 18.9‰). The average values are very similar for each trophic category for herbivores, for bears, for carnivores), with a slight average enrichment in bears and carnivores relative to herbivores. values of carbonate hydroxylapatite range from –14.8 to –10.3‰ in herbivores. They range from –17.4 to –15.4‰ in bears, and from –16.1‰ to –14.5 in carnivores (Table 4.1). The average values clearly differ between herbivores bears and carnivores The C values of carbonate hydroxylapatite of bears are slightly more negative than those of other carnivores. The difference between the values of carbonate hydroxylapatite and collagen of the same individual, range from 5.8 to 9.2 in herbivores, from 2.7 to 5.7 in bears, and from 4.3 to 5.5‰ in carnivores (Table 4.1). There is no overlap of values between herbivores on one hand, and bears and carnivores on the other. The average value is slightly larger in carnivores than in bears, but with a large overlap. The values of collagen range from 1.8 to 6.1%o in herbivores and from 6.7 to 12.0‰ in carnivores, with no overlap (Table 4.1). The values of bears range from 3.8 to 8.7‰, overlapping the values of herbivores and carnivores. The average value of collagen is clearly higher in carnivores than in bears and in herbivores The C values of bone and dentine collagen extracted from a mandible and a lower first molar from a Siberian wolf are identical (Table 4.2). On the contrary, the value of bone collagen is clearly lower (8.2‰) than the value of dentine collagen (9.8‰).

DISCUSSION Isotopic Framework for Modern Temperate and Cold Areas Due to the way the modern bones have been sampled, only general patterns can be described from the measured isotopic values. The samples come

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

73

from different continents with different ecological settings. Even in pure plant environments, small variations of values are observed. The principal ones are a “canopy-effect” (Schlesser and Jayasekera 1985; Van der Merwe and Medina 1991), an altitudinal effect (Körner et al. 1988; Leavitt and Long 1992), and a gradient of values within Europe, correlated with rainfall,

temperature and other climatic factors (Van Klinken et al. 1994, this volume). Water stress and salinity are additional factors that lead plants to exhibit less negative values (Farquhar et al. 1982; Guy et al 1986). All these effects have an amplitude of a few permil at most. An additional interfering factor is the fact that the carbon isotope ratio of atmospheric has changed during the last century, shifting from a value around –6.5‰ before 1850, which marks the beginning of the industrial age and the release of -depleted fossil fuel to values around –7.8‰ today (Marino and McElroy 1991). This shift in values of atmospheric has been observed in collagen values of moose teeth that lived in the years 1948–1973 (Bada et al. 1990). Thus, all samples coming from osteology collections for which there are no records of date of death may carry an isotopic signature related to atmospheric several decades old, and thus -enriched relative to modern atmospheric A significant decline occurs from pre- to post10,000 BP in both wood and, with a smaller magnitude, in all plant matter, which may change the baseline values of foodwebs for the fossil specimens (Leavitt and Danzer 1991). Changes of the value of atmospheric have been recorded in ice core bubbles, but the range of variations is only around l‰ for the last 40,000 years (Leuenberger et al. 1992). However, these sources of foodweb variation should not influence the patterns of between apatite and collagen. Local variations of nitrogen isotopic abundances are also documented (Nadelhoffer and Fry 1994). For example, plants growing on acidic soils are -depleted, their values being as low as –5‰ (Mariotti et al. 1980), and even lower values have been measured in a deciduous forest south of Paris (Rodiére et al. 1996). Conversely high latitude arid saline soils, for example Saskatchewan, Canada, have relatively high contents (Karamanos et al. 1981). These major sources of environmental variation must be considered in assessing global average values. Variation within trophic levels within ecosystems should be less. Moreover, more accurate estimates of trophic level differences should be made using such sample sets. Diachronic sample sets may include individuals that experienced different paleoenvironments and are thus less useful for evaluating trophic level effects.

Among modern herbivores (Table 4.1), three species exhibit

values

higher than –20‰: reindeer (19.8 to –19.4‰), chamois (–19.9‰) and the horse from Northern Greece (–18.8‰). High values of reindeer have been

interpreted as a consequence of the consumption of lichen during winter (Fizet et al. 1995). The high value of the chamois is probably due to the consumption of plants from high altitudes, which are -enriched relative to plants from low ones (Körner et al. 1988). The high value of the Greek

74

HERVÉ BOCHERENS

horse is consistent with the gradient observed by Van Klinken et al. (1994) in Europe. The value of the moose is particularly low –23.4‰). This could be due to the canopy effect if this was a forest-dwelling cervid; similar values have been reported by Bada et al. (1990) for moose from Isle Royale National Park (USA). Thus, despite the far from ideal type of sampling, some isotopic differences seem to appear between different groups of herbivores living in pure environments. The values of carnivores and bears do not show any clear tendency. It is however noteworthy that the highest collagen value for a bear has been measured on a specimen which has been killed before 1880, which is when the value of atmospheric was around 1.5‰ higher than today. The values of carbonate hydroxylapatite of the herbivores are less negative than those of carnivores and bears, with almost no overlap. The values of carbonate hydroxylapatite of the bears are slightly more negative than those of the other carnivores. As a result, difference between values of carbonate hydroxylapatite and those of collagen is larger in herbivores than in carnivores and in bears The value is on average 3.2‰ lower in carnivores than in herbivores. The variation in values is smaller for carnivores than for bears and for herbivores, but only five samples of carnivores have been analyzed. In herbivores, the value can be very different for two different specimens from the same

species (cow: 7.6 and 9.2; reindeer: 6.9 and 8.9). In bears, the range of the values is quite large, from 2.7 to 5.7‰, which may be linked to the omnivorous diet of these animals, with a large variety of food items. However, the values for bears are less than those of carnivores, although the significance of this is uncertain given the small sample size. This raises questions about the hypothesis that this difference in values is due to the trophic level effect rather than methanogenesis rate (Metges et al. 1990; Ambrose 1998). The values of herbivores are clearly lower than those of carnivores. The difference between the average values of both trophic groups is 4.6‰. This enrichment is similar to those reported in terrestrial ecosystems: 2.8‰ (Schwarcz 1991) to 5.7‰ (Ambrose and DeNiro 1986) with an average value between 3 and 4‰ (Schoeninger and DeNiro 1984; Schoeninger 1985, Sealy et al. 1987). The omnivorous bears have values overlapping the ranges of herbivores and carnivores, reflecting their omnivorous diet (Clevenger et al. 1992). In herbivores and carnivores, the ranges of

values are quite large,

but the significance of these variations is uncertain considering the small sample size and diverse geographic origins. The wolf jaw sample has a higher value in dentine than in bone collagen. This is due to the influence of -enriched collagen formed before weaning in dentine, relative to the collagen incorporated into bone, probably reflecting the adult diet, which has lower values than milk (Bocherens et al. 1994). Such a difference is not observed

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

75

in species such as horse, where teeth continue to grow a long time after

weaning, thus forming their dentine collagen from the same pool of nitrogen as bone (Bocherens et al. 1994). In summary, the following differences are observed in modern mammals from cold and temperate areas (Figs. 4.3 and 4.4): (1)

an enrichment of around l‰ in in collagen between herbivores and carnivores; (2) a clear difference in carbonate hydroxylapatite values between herbivores and carnivores, and thus a difference between values of herbivores and carnivores In this respect, bears behave mostly like other carnivores. Thus digestive physiology may be more important for values than the type of food actually consumed. (3)

a clear enrichment of around 4‰ in between herbivore and carnivore bone collagen. (4) an enrichment of around l–2‰ in between bone and dentine collagen in species where teeth grow partly before weaning and stop their growth.

More subtle isotopic differences may also be present, such as differences between micro-environments, or between species due to the selection of special sources of food. For example, a modern population of roe-deer Capreolus capreolus dwelling in a forest 35 km south-west of Paris have negative values, ranging from –2.8 to –0.4‰, probably because of local acidic soil conditions and the very territorial habits of this deer species (Rodiére et al. 1996). Such differences between habitats should be confirmed by further studies in modern undisturbed ecosystems. This would provide a very useful framework for understanding sources of variation in Pleistocene species. The approach proposed to check the preservation of isotopic signatures in Pleistocene samples from cold and temperate areas is to look for these specific signatures whenever possible, by selecting the appropriate specimens in the studied localities, before trying to interpret isotopic variations in fossil samples. The next section will provide examples of this approach in several cases published for Eurasia and Alaska.

Preservation of the Isotopic Signals in Selected Sites The difference of collagen

values between herbivores and

carnivores (Fig. 4.5) can be checked within the samples from Marillac

(Bocherens et al. 1991b; Fizet et al. 1995), Kent’s Cavern (Bocherens et al.

76

HERVÉ BOCHERENS

Figure 4.3. and values of bone collagen of herbivores, carnivores and bears in arctic and temperate ecosystems.

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

77

Figure 4.4. values of bone collagen of herbivores, carnivores and bears in arctic and temperate ecosystems.

1995b), Mialet (Bocherens et al. 1994), Fairbanks (Bocherens et al. 1995a) and Siberia (Bocherens et al. 1996) (Figs. 4.6–4.9). In all cases, the values measured on fossil samples show a good general agreement with the modern equivalents. In Kent’s Cavern, a slight overlap occurs between herbivore and carnivore values (Fig. 4.6). This might be due to a poor stratigraphic

78

HERVÉ BOCHERENS

Figure 4.5. and values of bone collagen from Marillac mammals (~45,000 BP) compared with those of recent mammals from cold and temperate areas (average 1 s.d.). Isotopic abundances of Marillac mammals are from Fizet et al. (1995).

resolution of the sampling in this site. Indeed, collagen from three different bones from Kent’s Cavern used in this study have been radiocarbon dated and the ages range from 40,600 to 27,640 BP (Hedges et al. 1995). Kent’s Cavern values seem more negative than Marillac and modern values. This could be explained either by different values in the plants, with values lower in Great Britain than in continental Europe, as described by Van Klinken et al. (1994) for modern Europe, or by differences in the herbivore species composition. Indeed, reindeer in Marillac have less negative values than horses, probably due to the consumption of lichens (Fizet et al. 1995). Differences in atmospheric values and climate differences should also be considered.

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

79

Figure 4.6. and values of bone collagen from Kent’s Cavern mammals (~35,000 – 60,000 BP) compared with those of recent mammals from cold and temperate areas (average 1 s.d.). Isotopic abundances of Kent’s Cavern mammals are from Bocherens et al. (1995b).

The enrichment in dentine collagen relative to bone collagen in species with definite tooth growth is observed in samples from Marillac, Kent’s Cavern, Aldéne and Mialet caves (Fig. 4.10). The highest values are mea-

sured for deciduous teeth in fossil samples (Bocherens et al. 1994; Fizet et al. 1995). No enrichment is observed between dentine and bone collagen in horses (Bocherens et al. 1995b; Fizet et al. 1995).

Similar values are measured in carbonate hydroxylapatite for herbivores, carnivores and bears, respectively, in localities from 40,000 to 600,000 years old (Fig. 4.11). The spacing value between collagen and carbonate hydroxylapatite is larger in Kent’s Cavern herbivores and carnivores

80

HERVÉ BOCHERENS

Figure 4.7. and values of bone collagen from Mialet cave mammals (~35,000 – 60,000 BP) compared with those of recent mammals from cold and temperate areas (average 1 s.d.). Isotopic abundances of Mialet cave mammals are from Bocherens et al. (1994).

than in modern ones in Kent’s Cavern herbivores versus in modern herbivores; in Kent’s Cavern carnivores versus in modern carnivores). Nonetheless, the difference between the spacing value of herbivores and carnivores is preserved in Kent’s Cavern samples. It is noteworthy that cave bears show a different pattern than modern brown bear, with values lower than those of other carnivores (Bocherens et al. 1994). Since this difference occurs in all studied localities, it is more probably reflecting a biological difference rather than diagenetic alteration. This “speleoid” isotopic pattern occurs in Ursus spelaeus deningeroides and in Ursus deningeri as well (Bocherens et al. 1994), maybe indicating a distinct physiology in this bear lineage.

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

81

Figure 4.8. and values of bone collagen from Pleistocene Alaskan mammals (~ 10,000 – 40.000BP) compared with those of recent Alaskan mammals. Isotopic abundances of Pleistocene mammals are from Bocherens et al. (1995a).

CONCLUSION The following differences have been observed in modern mammals from cold and temperate areas with no plants: a slight enrichment in in collagen between herbivores and carnivores; a clear difference in carbonate hydroxylapatite values between herbivores and carnivores (including bears), and thus a difference between values of herbivores and carnivores; a clear enrichment in between herbivore and carnivore bone collagen,

82

HERVÉ BOCHERENS

Figure 4.9. and values of bone collagen from Pleistocene Siberian mammals (~10,000 – 50,000 BP) compared with those of recent Siberian mammals. Isotopic abundances of Pleistocene mammals are from Bocherens et al. (1996).

around 4‰; and finally, an enrichment in around l–2‰ between bone and dentine collagen in species where permanent teeth start to grow before weaning. Similar differences have been measured in the studied Pleistocene localities of Eurasia and Alaska, suggesting a good preservation of isotopic abundances in these localities. The approach proposed to check the preservation of isotopic signals in fossil vertebrate bones and teeth is appropriate not only for Pleistocene cold and temperate areas, but also during geological periods before the

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

83

Figure 4.10. Differences in values between bone and dentine collagen from Pleistocene and recent mammals. Isotopic abundances are from Bocherens et al. (1994, 1995b) and this paper.

ecological dominance of plants. The difference in carbon isotopic abundances between herbivores and carnivores in purely -ecosystems suggest a way to test the preservation of the isotopic signatures, especially in enamel, during the pre-Miocene geological periods, where plants are absent or rare.

84

HERVÉ BOCHERENS

Figure 4.11. Average collagen (“c”) and carbonate hydroxylapatite (“a”) values in modern and Pleistocene herbivores, carnivores and bears. Isotopic abundances for modern samples are from this paper, those for Kent’s Cavern samples are from Bocherens et al. (1995b) and those for other Pleistocene localities are from Bocherens et al. (1994).

REFERENCES Ambrose, S.H. 1990 Preparation and characterization of bone and tooth collagen for isotopic analysis. Journal of Archaeological Science 17: 431–451. 1991 Effects of diet, climate and physiology on nitrogen isotope abundances in terrestrial foodwebs. Journal of Archaeological Science 18: 293–317.

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

85

1998 Prospects for stable isotopic analysis of Later Pleistocene hominid diets in West Asia

and Europe. In Akazawa, T, Aoki, K. and Bar-Yosef, O., eds., Neanderthals and Humans in West Asia, New York, Plenum Press: 277–289. Ambrose, S.H. and DeNiro, M.J. 1986 The isotopic ecology of East African mammals. Oecologia

69: 395–406. Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relationship of the carbon isotope

ratios of whole diet and dietary protein to those of bone collagen and carbonate. In Lambert, J. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level Berlin, Springer-Verlag: 1–37. Bada, J.L., Peterson, R.O., Schimmelmann, A. and Hedges, R.E.M. 1990 Moose teeth as monitors of environmental isotopic parameters. Oecologia 82: 102–106. Bender, M.M. 1968 Mass spectrometry studies of carbon-13 variations in com and others grasses. Radiocarbon 10: 468–472. Berry, J.A. 1975 Adaptation of photosynthetic processes to stress. Science 188: 644–650. Bocherens, H., Emslie, S.D., Billiou, D. and Mariotti, A. 1995a Stable isotopes and paleodiet of the giant short-faced bear (Arctodus simus). Comptes Rendus de l’ Académie des Sciences de Paris 320: 779–784.

Bocherens, H., Fizet, M. and Mariotti, A. 1990 Mise en évidence du régime alimentaire végétarien de l’ours des cavernes (Ursus spelaeus) par la biogéochimie isotopique du collagéne fossile. Comptes Rendus de l’ Académie des Sciences 311: 1279– 1284. Bocherens, H., Fizet, M., Mariotti, A., Lange-Badré, B., Vandermeersch, B., Borel, J.P and Bellon,

G. 1991a Isotopic biogeochemistry of fossil vertebrate collagen: implications for the study of fossil food web including Neandertal Man. Journal of Human Evolution 20: 481–492. Bocherens, H., Fizet, M., Mariotti, A., Billiou, D., Bellon, G., Borel, J.R and Simone, S. 1991b

Biogéochimie isotopique

et paléoécologie des ours Pléistocénes de la grotte

d’Aldéne. Bulletin du Muséé d’Anthropologie Préhistorique de Monaco 34: 29–49. Bocherens, H., Fizet, M. and Mariotti, A. 1994 Diet, physiology and ecology of fossil mammals as inferred from stable carbon and nitrogen isotope biogeochemistry: implications for Pleistocene bears. Palaeogeography, Palaeoclimatology, Palaeoecology 107: 213– 225.

Bocherens, H., Fogel, M.L., Tuross, N. and Zeder, M. 1995b Trophic structure and climatic information from isotopic signatures in a Pleistocene cave fauna of southern England. Journal of Archaeological Science 22: 237–340. Bocherens, H. and Mariotti, A. 1992 Biogéochimie isotopique du carbone dans les os de mammiféres actuels et fossiles de zones froides et tempérées. Comptes Rendus de l’ Academic des Sciences de Paris 315: 1147–1153. Bocherens, H., Pacaud, G., Lazarev, P. and Mariotti, A. 1996 Stable isotope abundances in collagen and soft tissues from Pleistocene mammals from Yakutia. Implications for the paleobiology of the mammoth steppe. Palaeogeography, Palaeoclimatolology, Palaeoecology

126: 31–44. Bryant, J.D., Luz, B. and Froelich, P.N. 1994 Oxygen isotopic composition of fossil horse tooth phosphate as a record of continental paleoclimate. Palaeogeography, Palaeoclimatology,

Palaeoecology 107: 303–316. Clevenger, A.P., Purroy, F.J. and Pelton, M.R. 1992 Food habit of brown bears (Ursus arctos) in the Cantabrian mountains, Spain. Journal of Mammalogy 73: 415–421. Deines, P. 1980 The isotopic composition of reduced organic carbon. In Fritz, P. and Fontes, J. Ch., eds., Handbook of Environmental Isotope Geochemistry, Vol. 1: The Terrestrial Environment, A. Amsterdam, Elsevier: 329–406. DeNiro, M.J. 1985 Post-mortem preservation and alteration of in vivo bone collagen isotope ratios

in relation to palaeodietary reconstruction Nature 317: 806–809.

86

HERVÉ BOCHERENS

DeNiro, M.J. and Epstein, S. 1978 Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta 42: 495–506. DeNiro, M.J. and Weiner, S. 1988 Chemical, enzymatic and spectroscopic characterization of collagen and other organic fractions in prehistoric bone. Geochimica et Cosmochimica Acta 52: 2197–2206. Farquhar, G.D., Ball, M.C., von Laemmer, S. and Roksandic, Z. 1982 Effect of salinity and humidity on values of halophytes. Evidence for diffusional isotope fractionation determined by the ratio of intercellular atmospheric

under different environmental conditions.

Oecologia 52: 121–124. Fizet, M., Mariotti, M., Bocherens, H., Lange-Badré, B., Vandermeersch, B., Borel, J.P and Bellon, G. 1995 Effect of diet, physiology and climate on carbon and nitrogen stable isotopes of collagen in a late Pleistocene anthropic paleoecosystem (France, Charente, Marillac). Journal of Archaeological Science 22: 67–79. Fogel, M.L., Tuross, N. and Owsley, D.W. 1989 Nitrogen isotope tracers of human lactation in modern and archeological populations. Annual Report of the Director of the Geophysical Laboratory of the Carnegie Institution, Washington 1988–1989. Washington, D.C., Carnegie

Institution of Washington: 111–117. Gillespie, R., Hedges, R.E.M. and Wand, J.O. 1984 Radiocarbon dating of bone by Accelerator Mass Spectrometry . Journal of Archaeological Science 11: 165–170. Grupe, G., Dreses-Werringloer, U. and Parsche, F. 1993 Initial stages of bone decomposition: Causes and consequences, In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone, Archaeology at the Molecular Level. Berlin, Springer-Verlag: 257–274. Guy, R.D., Reid, D. and Krouse, H.R. 1986 Factor affecting ratios of inland halophytes. I. Controlled studies on growth and isotopic composition of Puccinellia nuttaliana. Canadian Journal of Botany 64: 2693–2707. Hare, PE. and von Endt, D. 1990 Variable preservation of organic matter in fossil bone. Annual

Report of the Director of the Geophysics Laboratory. Carnegie Institution, Washington. 1989–1990. Washington, D.C., Carnegie Institution of Washington: 115–118. Heaton, T.H.E., Vogel, J.C., Chevallerie, G.V.L. and Collett, G. 1986 Climatic influence on the isotopic composition of bone nitrogen. Nature 322: 822–824. Hedges, R.E.M., Lee-Thorp, J.A. and Tuross, N.C. 1995 Is tooth enamel carbonate a suitable material for radiocarbon dating? Radiocarbon 37: 285–290. Herrero, S. 1978 A comparison of some features of the evolution, ecology and behavior of black and grizzly/brown bears. Carnivore 1: 7–17. Hillson, S. 1986 Teeth. Cambridge, Cambridge University Press: 376. Karamanos, R.E., Voroney, R.P. and Rennie, D.A. 1981 Variation in natural N-15 abundance of central Saskatchewan soils. Soils Science Society of America Journal 45: 826–828. Koch, PL. 1991 The isotopic ecology of Pleistocene proboscideans. Journal of Vertebrate Paleontology 11: A40. Koch, PL, Behrensmeyer, A.K., Tuross, N. and Fogel, M.L. 1990 Isotopic fidelity during bone weathering and burial. Annual Report of the Director of the Geophysical Laboratory of the Carnegie Institution, Washington, 1989–1990. Washington, D.C., Carnegie Institution of Washington: 105–110. Koch, PL., Behrensmeyer, A.K. and Fogel, M.L. 1991 The isotopic ecology of plants and animals in Amboseli National Park, Kenya. Annual Report of the Director of the Geophysicsal Laboratory, Carnegie Institution, Washington 199–1991. Washington, D.C., Carnegie Institution of Washington: 163–171. Körner, Ch., Farquhar, G.D. and Roksandic, Z. 1988 A global survey of carbon isotope discrimination in plants from high altitude. Oecologia 74: 623–632. Leavitt, S.W. and Danzer, S.R. 1991 Implications of variations in plants over the past 55,000 years. Radiocarbon 33(2): 221. Leavitt, S.W. and Long, A. 1992 Altitudinal differences in of bristlecone pine tree rings. Naturwissenschaften 79: 178–180.

ISOTOPIC SIGNALS

IN PLEISTOCENE MAMMALS

87

Lee-Thorp, J.A. 1989 Stable carbon isotopes in deep time. The diets of fossil fauna and hominids. Ph.D. Thesis, University of Cape Town, South Africa, 174 pp. Lee-Thorp, J.A. and Van der Merwe, N.J. 1991 Aspects of the chemistry of modern and fossil biological apatites. Journal Archaeological Science 18: 343–354. 1993 Stable carbon isotope studies of Swartkrans fossils. In C.K. Brian, ed., Swartkrans, A Cave’s Chronicle of Early Man Transvaal Museum Monograph 8: 251–256. Lee-Thorp, J.A., Sealy, J.C. and Van der Merwe, N.J. 1989 Stable carbon isotope ratio differences between bone collagen and bone apatite, and their relationship to diet. Journal of Archaeological Science 16: 585–599. Leuenberger, M., Siegenthaler, U. and Langway, C.C. 1992 Carbon isotope composition of atmospheric during the last ice age from an Antarctic ice core. Nature 357: 488–490. Long, A., Wilson, A.T., Ernst, R.D., Gore, B.H. and Hare, P.E. 1989 AMS radiocarbon dating of bones at Arizona. Radiocarbon 31: 231–238. Marino, B.D. and McElroy, M.B. 1991 Isotopic composition of atmospheric inferred from carbon in plant cellulose Nature 349: 127–131. Mariotti, A., Pierre, D., Vedy, J.C., Burckert, S. and Guillemot, J. 1980 The abundance of natural nitrogen 15 in the organic matter of soils along an altitudinal gradient (Chablais, HauteSavoie, France). Catena 7: 293–300. Metges, C., Kempe K. and Schmidt, H.-L. 1990 Dependence of the carbon-isotope contents of breath carbon dioxide, milk, serum and rumen fermentation products on the value of food in dairy cows. British Journal of Nutrition 63: 187–196. Nadelhoffer, K.J. and Fry, B. 1994 Nitrogen isotope studies in forest ecosystems. In Lajtha, K. and Michener, R.H., eds., Stable Isotopes in Ecology and Environmental Science. Blackwell Scientific Publications: 22–44. O’Leary, M.H. 1988 Carbon isotopes in photosynthesis. BioScience 38: 328–336. Rodiére, É., Bocherens, H., Angibault, J.-M. and Mariotti, A. 1996 Particularités isotopiques chez

le chevreuil (Capreolus capreolus L.): Implications pour les reconstitutions paléoenvironnementales. Comptes Rendu de l’Académie des Sciences de Paris 323: 179–185. Schlesser, G.H. and Jayasekera, R. 1985 13C-variations of leaves in forests as an indication of reassimilated CO2 from the soil. Oecologia 65: 536–542. Schoeninger, M.J. 1985 Trophic level effects on and ratios in bone collagen and strontium levels in bone mineral. Journal of Human Evolution 14: 515–525. Schoeninger, M.J. and DeNiro, M.J. 1984 Nitrogen and carbon isotopic composition of bone collagen from marine and terrestrial animals. Geochimica et Cosmochimica Acta 48: 625–639. Schwarcz, H.P. 1991 Some theoretical aspects of isotope paleodiet studies. Journal of Archaeological Science 18: 261–275. Sealy, J.C., Van der Merwe, NJ., Lee-Thorp, J.A. and Lanham, J.L. 1987 Nitrogen isotopic ecology in southern Africa: implications for environmental and dietary tracing. Geochimica et Cosmochimica Acta 51: 2707–2717. Sillen, A. 1988 Elemental and isotopic analyses of mammalian fauna from southern Africa and their implications for paleodietary research. American Journal of Physical Anthropology 76: 49–60. Smith, B.N. and Epstein, S. 1971 Two categories of ratios for higher plants. Plant Physiology 47: 380–384. Tieszen, L.L. and Fagre, T. 1993 Effect of diet quality and composition on the isotopic composition of respiratory bone collagen, bioapatite, and soft tissues, In Lambert, J. and Grupe, G., eds., Prehistoric Human Bone—Archaeology at the Molecular Level. Berlin, Springer: 121–155. Tuross, N., Fogel, M.L. and Hare, P.E. 1988 Variability in the preservation of the isotopic composition of collagen from fossil bone. Geochimica et Cosmochimica Acta 52: 929– 935. Van der Merwe, N.J. 1986 Carbon isotope ecology of herbivores and carnivores, Palaeoecology of Africa and the Surrounding Islands 17: 123–131.

88

HERVÉ BOCHERENS

Van der Merwe, NJ. and Medina, E. 1991 The canopy effect, carbon isotope ratios and foodwebs in Amazonia. Journal of Archaeological Science 18: 249–259. Van Klinken, G.J., Van der Plicht, H. and Hedges, R.E.M. 1994 Bone

ratios reflect

(palaeo-)climatic variations. Geophysical Research Letters 21: 445–448. Vogel, J.C. 1978 Isotopic assessment of the dietary habits of ungulates. South African Journal of Science 74: 298–301. Vogel, J.C., Talma, A.S., Hall-Martin, A.J. and Viljoen, P.J. 1990 Carbon and nitrogen isotopes in elephants. South African Journal of Science 86: 147–150.

Chapter

5

Preservation of Biogenic Carbon Isotopic Signals in Plio-Pleistocene Bone and Tooth Mineral JULIA A. LEE-THORP ABSTRACT A key test for establishing the integrity of stable carbon isotope data from the mineral phase of fossil bones and teeth was the demonstration that only small differences existed between mean values for modern browsers and their fossil counterparts. The small difference increased with age; hence it was interpreted solely in terms of diagenesis. However, re-examination of the data suggests that the modern values used as references could themselves be anomalously depleted due to the effects on modern atmospheric of fossil fuel burning. Here carbon isotope results for animals of predictable diet at both ends of the carbon isotope “spectrum” are compared from sites with isotopically different depositional contexts and ages, in order to re-evaluate the effects of diagenesis. Where matrix carbonates are relatively enriched in the distinction between depleted browser values and enriched matrix material remains the most useful test, but where deposit carbonates are more depleted, the comparison between enriched grazers and depleted matrix values is more useful. The results for enamel and bone apatite of both browsers and grazers indicate that for about the first 100,000 years at least, isotopic differences observed between the modern and their fossil counterparts are attributable mainly to Biogeochemical Approaches to Paleodietary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

89

90

JULIA A. LEE-THORP

shifts of the modern atmosphere. Larger shifts for samples in the million year (Ma) range may more plausibly be ascribed to diagenetic effects. INTRODUCTION Use of the mineral phase of bone and teeth as sample material in stable carbon isotope dietary or climate reconstruction studies holds both advantages and disadvantages. Firstly, according to our current understanding, carbon isotope ratios in carbonate within mineral reflect those of the entire, integrated diet, while those in collagen tend to reflect rather the protein isotope values (Ambrose and Norr 1993; Tiezsen and Fagre 1993). Hence combined analysis of both phases provides a much fuller understanding about elements of the diet, which as suggested earlier (Lee-Thorp et al. 1989a), permits distinction of animal- versus cereal-based components of prehistoric diets, and potentially expands isotopic dietary assessments into mono-isotopic environments such as Europe where application was previously limited (Ambrose 1998; Bell and Lee-Thorp 1997; van Klinken et al., this volume). The remaining points hinge largely on the same process: fossilization, or its correlate, diagenesis. Mineral in bones and teeth usually survives far longer than collagen. Collagen tends to hydrolyze, denature and dissolve away over millennial time-scales, so that only in exceptional conditions do quantities sufficient for analysis survive into the Pleistocene, although Ambrose (1998) has recently reported collagen survival up to 70 Ka in cool Central Europe. Survival is usually far shorter for bones buried in warmer regions, and none has been reported in southern Africa beyond ~10Ka, other than for a cool high altitude location in the highlands of Lesotho (Vogel 1982). In contrast, the mineral survives because it alters and stabilizes, i.e., there is a price to longterm survival. Therefore, changes which occur over time in the mineral structure after death and burial are integral to the fossilization process. The extent to which these processes do or do not obliterate biogenic isotopic or trace element signals is the source of long-standing controversy; yet resolution of these questions is important if we are to satisfactorily address questions about human diet in the distant past. The purpose of this paper is to re-evaluate the effects of diagenesis on carbon isotope signals in mineral, using as tests a series of fauna with predictable diets, from different depositional environments and time ranges. Before doing so it is useful to review those aspects of calcified tissue chemistry of relevance to this issue.

CHEMISTRY OF BIOLOGICAL APATITES The minerals in skeletal mammalian tissues are biological apatites (the plural is used here deliberately because they exist as a continuum of structures

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

91

differing in kind and degree of substituents, function, crystal size, and properties). They belong to the group of hexagonal calcium phosphates, of which hydroxyapatite is perhaps the best known and fairly close in structure to biological apatites. But the latter differ from synthetic or geologically occurring apatites in a number of ways, including non-stoichiometry, high degree of isomorphic substitution and adsorption including carbonate, lattice distortions and small crystal size (Pellegrino and Biltz 1968; Biltz and Pellegrino 1977; Betts et al. 1981; LeGeros 1981). These properties are linked, and the net result is that biological apatites are poorly crystalline (the exception being enamel), with a number of important implications for form and

function in vivo, and post-mortem changes or diagenesis. Crystallinity is a term which refers to both crystal size and amounts of defects and distortions. As a result of enhanced surface area effects due to small size and reactive surfaces with net charge and adsorbed ions (Simpson 1972), and high number of substitutions that tend to increase solubility (Blumenthal et al. 1975; Betts et al 1981; LeGeros 1981; LeGeros and Tung 1983; Posner 1985), bone apatite is relatively soluble. This is in fact vital because one of the functions of bone is to act as a store for ready release of and ions into the bloodstream as required (Wheeler and Lewis 1977). At the same time it is the carbonate substituted within the apatite which holds the “dietary” carbon isotope signal. Apatite in bone and dentine is so poorly crystalline in vivo that it has been referred to as submicro-crystalline or para-crystalline (Wheeler and Lewis 1977; Nelson et al. 1982). Carbonate occurs in several locations in crystals, both as (i) adsorbed ions on crystal surfaces, and (ii) within the unit cell substituted mainly in the phosphate position and most likely to a lesser extent in the hydroxyl position (Driessens et al. 1978; Eanes 1979; Chickerur et al. 1980; Besha et al. 1990; Rey et al. 1991). Adsorbed carbonates are more labile (reactive) (Poyart et al. 1975), while the substituted carbonates are more stable, and here are referred to as “structural carbonates” in keeping with our earlier work (Lee-Thorp 1989; Lee-Thorp and van der Merwe 1991). Enamel apatite differs from bone and dentine, being more crystalline with fewer substitutions, less distortion and greater long-range order, and with crystals about an order of magnitude larger (LeGeros 1983). Other important differences exist, e.g., in higher-order structures and the kind and quantity of organic matrix. Bone and dentine apatite is formed in close association with collagen fibrils, and deposition and orientation of apatite crystallites is regulated by fibril periodicity (Boskey 1981), whereas enamel forms as rods on templates comprising organic tubules (Boyde 1967; Eisenmann 1985). During maturation the organic matrix, which consists of phosphoproteins and amelogenins (Boskey 1981, LeGeros 1983), decreases in quantity so that in mature enamel there is very little left—perhaps <1%. In bone and dentine the amount of organic material, mainly collagen, remains high at about 20–30%. Variabil-

ity within this range depends on the skeletal element involved, i.e., whether dentine, or cancellous or cortical bone.

92

JULIA A. LEE-THORP

Implications for Diagenesis The properties described above have important consequences for the way in which these skeletal tissues are subsequently preserved, and hence their usefulness or otherwise as recorders of dietary signals. Several points from the discussion above are relevant here. It is useful to ask: what are the most important mechanisms or routes for change in buried bones and teeth? One could divide these processes into those with simple addition of new non-apatitic material (various minerals such as pyrites, silicates and simple carbonates) in pores and spaces (Hassan and Ortner 1977), and those related to change within the apatite crystals, usually in the form of recrystallization and crystal growth. The first kind of process has severe implications for alteration of bone and dentine, partly because they are porous materials with high surface area initially and because the approximately 20–30% by volume occupied by collagen is subsequently lost by hydrolysis and/or consumption by bacteria and the void filled by new minerals. Enamel is much denser and contains no pores or Haversian canals; and there is very, little organic material to lose and replace with extraneous material. Cracks are the only interstices available for deposition of material. The second kind of diagenetic process of relevance here is incorporation of exogenous material within crystals either by formation of new apatite entirely, by dissolution and recrystallization, by crystal growth, or by ionic and isotopic exchange within existing crystals. These processes are directly related to crystal properties. Calcium and phosphate ions have an extreme affinity for each other, with a tendency to scavenge and form apatites where these ions are present (Simpson 1972). In life active inhibitory mechanisms prevent wholesale calcification, e.g., presence of ions and ATP (Blumenthal et al. 1977), but these mechanisms cease after death, whereupon conditions favorable for apatite formation and growth emerge. As discussed above, fresh bone apatite is poorly crystalline and highly reactive, properties which imply ready susceptibility to the processes of dissolution, recrystallization and crystal

growth post-mortem and during fossilization. These are the most plausible mechanisms for incorporation of foreign material into the crystal structure. Certainly increased crystallinity is a remarkably consistent observation across a range of studies using either x-ray diffraction (XRD) or infra-red techniques (Blumenthal et al. 1975; Hassan et al. 1977; Schoeninger 1982; Sillen 1986, 1989; Sillen and Morris 1996; Sillen and Parkington 1996; Tuross et al. 1989; Weiner and Bar-Yosef 1990; Weiner et al. 1993; Person et al. 1995), and the

concentration of ions such as fluorine, carbonate, strontium, barium, lead and uranium usually, although not always, increases (Hassan 1975; Baud et al.

1977; Millard and Hedges 1995). Once crystallinity has increased (larger crystals/less distortion), apatite is more stable, therefore it can be predicted that the pace of change should slow. Ionic or isotopic exchange may certainly also occur, although it is likely to be an extremely slow process (on the order

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

93

of thousands of years) as suggested by solid state ion exchange studies (Haul and Stein 1955; Anderson 1969). It should be pointed out, however, that crystal growth is also possible using local, biogenic and not necessarily exogenous material. Ions required for

growth of existing crystals are present in crystal hydration layers and in body fluids, or in addition may occur by the process of Ostwald ripening whereby larger crystals incorporate small ones (Eanes and Posner 1970). This idea is supported by observations where postmortem increases in crystallinity were observed to occur in a wildebeest which was not buried (Tuross et al. 1989), and skeletons sealed above ground in Saharan tombs (Person et al. 1995; Saliège et al. 1995), but minimal change was observed isotopically. With the exception of ionic exchange, all of these processes so important during fossilization of bone mineral, are of greatly reduced importance for enamel apatite because it is already more crystalline and stable. However, this does not mean that important changes do not occur. For instance Michel et al. (1995) have suggested that, based on combined Fourier transform infra-red (FTIR) and XRD Rietveld analysis, for substitution increases in Cervus elaphus (red deer) enamel in Lazaret Cave, France. A recent study of fossil enamel points to the presence of an ionic exchange mechanism (Hedges et al. 1995) using this more sensitive indicator. It was observed that small amounts (about 6%) of “young” carbon could not be eliminated with pretreatment procedures, and although this amount is not detectable for light isotopes, the results suggest recent addition of young carbon, which may imply that the process is continuous (Hedges et al. 1995). One further difference between the tissues should be noted briefly—that of turnover—which holds implications for the nature of the isotopic signal recorded and its interpretation. Bone is constantly resorbed and reformed during life, i.e., it “turns over”, whereas enamel and dentine do not, although

secondary dentine can be later accreted. Enamel and dentine form during a discrete period in the individual’s life. This means that carbon isotope dietary signals in bone, for both collagen and apatite, reflect diet integrated over years, whereas those in enamel and dentine increments reflect diet at time of formation.

Tests for Isotopic Integrity Effects of at least the first type of diagenesis: addition of non-apatitic material, are fairly easily mitigated using appropriate chemical pretreatments, and particularly if the material is a simple carbonate. A fair degree of success has been demonstrated in several studies with the usual procedure being a weak acetic acid wash (Krueger 1991; Lee-Thorp and van der Merwe 1987, 1991; Koch et al. 1997). This protocol also preferentially dissolves away more soluble apatites, including in theory, non-biogenic high-carbonate apatite and the smaller, more reactive biogenic apatite fractions. But where diagenesis also

94

JULIA A. LEE-THORP

leads to more stable apatites through crystal growth or recrystallization and/or incorporation of fluoride (LeGeros and Tung 1983), or ionic/isotopic exchange of structural carbonate ions takes place, these procedures will be less effective.

The question then becomes one of which diagenetic processes are dominant, and at what stages? A more sophisticated buffered acid-based protocol has been developed for sequential dissolution which permits separate examination of phases differing in solubility (Sealy et al. 1991; Sillen 1986, 1989; Sillen and LeGeros 1991), but it has not yet been adapted for examination of light isotope changes because of technical difficulties. A number of tests have been used to evaluate and test the reliability of stable carbon isotope ratios in the mineral phase, sometimes with (apparently) conflicting results, and a great deal of argument has existed around these issues for many years (the history of the debate is reviewed in Lee-Thorp 1989). Most earlier studies concentrated on bone. Following earlier attempts by radiocarbon chemists and suggestions by Land et al. (1980), Sullivan and Krueger

(1981) found that collagen and apatite carbon isotope values for archaeological bones of varying age (within the range of radiocarbon) yielded consistent isotopic differences between the two phases, when subjected to vigorous pretreatment procedures to remove secondary carbonates. Hence, since isotopic values from collagen are generally regarded as sound according to standard criteria for assessment, they concluded that mineral isotope values were reliable if appropriately pretreated (Sullivan & Krueger 1981).

Their conclusions were challenged by Schoeninger & DeNiro (1982) who found fluctuating for another set of archaeological bones, many of them human. There were two main areas of dispute. One related to the effects of different pretreatment procedures. For instance it was suggested that procedures using fairly concentrated acetic acid followed in the latter study would

be ineffective in removing carbonates (Sullivan and Krueger 1983), or that they produced recrystallization artefacts, as demonstrated in several later studies (Lee-Thorp 1989; Lee-Thorp and van der Merwe 1991; Koch et al. 1997). Secondly, it soon emerged that the test is not always appropriate since it changes with diet and trophic level (Krueger and Sullivan 1984; LeeThorp et al. 1989a), and furthermore, it limits the test to samples within the range of collagen survival. A new test devised to overcome these limitations compared bone and enamel apatite carbon isotope values from a large set of browsing animals (with, by definition, consistent diets and therefore correspondingly depleted isotopic signatures) with their modern counterparts, using pretreatment protocols similar to that of Sullivan and Krueger (1981), and found that almost all fossil values were slightly more enriched than modern means (Lee-Thorp and van der Merwe 1987). The enrichment could not be eliminated by standard pretreatments and increased with age (ages from ~1 Ka–3Ma), and hence was interpreted as evidence for diagenesis via contamination by enriched matrix carbonates. However, even after millions of years enamel apatite values

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

95

were clearly identifiable with predicted diet. Furthermore, in Swartkrans Member 1 (~1.8Ma), graminivorous diets for Theropithecus baboons were clearly detectable from their enriched carbon isotope values, which were distinct from depleted, -based diets of Papio baboons in the same period (LeeThorp et al. 1989b). Similar observations using even older material have since confirmed this finding (e.g., Quade et al. 1992; Morgan et al. 1994; Wang and Cerling 1994; Wang et al., 1994; Koch et al. 1992, 1994; Cerling et al., 1998). In addition, Krueger developed further tests: in one test he found that there was little carbonate exchange within the apatite structure in bones immersed in a highly enriched (+9750‰) synthetic groundwater solution; in another test mastodon teeth immersed in sea water for ~10Ka retained terrestrial carbon isotope signatures after appropriate pretreatment (Krueger 1991). In parallel developments, oxygen isotope studies based on the stable phosphate ion in calcified tissues have been found to be more successful using enamel than bone as sample material (Ayliffe et al. 1994; Bryant et al. 1994), and similarly oxygen isotopes from the less stable C—O bond in enamel carbonate seems to be more predictable (Bocherens et al. 1996). Researchers exploring the relationship between Electron Spin Resonance (ESR) and carbonate content in enamel have found that dates are mostly consistent when carbonate levels did not deviate much from biogenic levels (Grün et al. 1990; Rink and Schwarcz 1995). Although the browser isotopic data set provided a sound test for integrity in broad terms, and demonstrated that the approach could work, some minor complicating factors have become apparent that have, or could have, implications for detailed interpretation of the results. One concerns recent changes in atmospheric and the other concerns environmental effects on carbon isotope ratios of plants. The burning of fossil fuel has not only increased concentration of atmospheric it has also led to a marked atmospheric carbon isotope shift over the last century, based on ice core and atmospheric data (Friedli et al. 1986). These trends were confirmed by time-series isotopic data from plants (Marino and McElroy 1991). The net result of the “fossil fuel effect” has been to shift the entire modern food web by at least – 1.5%o over the last 30–50 years. Therefore, in order to calculate pre-industrial “modern” values, 1.5%o should be added to isotopic values of plants and

animals; the value for browsers then becomes –13‰. The effect is illustrated in Fig. 5.1, in a simple flow diagram showing carbon isotopic pathways between atmospheric and plants, and herbivore biological apatite, for present and pre-industrial conditions. It follows from the linkage between recent increases in and depletion of that earlier significant differences between glacial and interglacial periods must also have led to earlier isotopic shifts, which in turn might have affected foodwebs at those times. Marine productivity and conse-

quent

drawdown is widely believed to have dominated atmospheric

(Sarntheim et al. 1988, Falkowski 1997), but the directional effects on

96

JULIA A. LEE-THORP

Figure 5.1. A simple flow diagram of ratios in a modern savanna system (a), and (b) adjusted for a fossil fuel effect of 1.5‰. Average values for and plants, and for browsers feeders) and grazers feeders) are used. Note that for herbivores, apatite values are about 7–8‰ less negative than collagen for the same individual. Carbon isotope ratios are conventionally reported in the notation relative to the PeeDee Belemnite standard as:

are not clear. The shifts may have been relatively small (<1‰) and difficult to detect using the proxy evidence available. Estimates of the magnitude of shifts between the transition from the Glacial Maximum to the present Interglacial vary between +0.3‰ based on ice cores (Leuenberger et al. 1992) and about 1.0‰ based on plants in packrat middens (Marino et al. 1992), i.e., the Glacial Maximum atmosphere was isotopically depleted compared to the Holocene. This topic is discussed further in the Discussion section. In making the reasonable assumption that browser diets should have remained consistent through time, it was also assumed that ratios of their foliage diets would not have changed. This is not the case; environmental parameters such as aridity, osmotic stress, temperature, and irradiance have predictable effects on

ratios of

plants (summarized in

Tieszen 1991). Water stress, for instance, lowers photosynthetic discrimination against resulting in isotopic enrichment. plants are not affected isotopically by environmental parameters although their distributions are (Ehleringer et al. 1997). Under conditions of increasing aridity, the climate variable of most likely importance in African sites, plants should become slightly more enriched on average. Therefore one should be cautious about use

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

97

of a universal mean for browsers; however, since the effects are small, and there is little evidence of aridity in the sites examined here, the problem is unlikely to impinge upon the model as broadly defined. Notwithstanding these caveats, comparison of the values for fossil browsers with the adjusted pre-industrial mean indicates that most fossil values fall within the mean standard deviation for at least 100 Ka, and only in the earlier (Ma) time range is there a more significant alteration (Fig. 5.2).

Isotopic Integrity of Fossil Bones and Teeth in South Africa We have previously assumed that the principle diagenetic process is attributable to exchange or interaction of (depleted) browser apatite values with (enriched) sedimentary matrix values (Lee-Thorp and van der Merwe 1987). Another possibility is ionic or isotopic exchange with soil however in most cases soil values will be closely related to matrix carbonate values. If the former is the case, one would expect enriched grazer values (near 0%o)

Figure 3.2. values of fossil browser apatites plotted against age for sites including Die Kelders, Klasies River Mouth I, Sterkfontein and Makapansgat. Enamel is represented by solid squares, bone apatite by solid circles. The mean value standard deviation for modem browsers, represented by dashed lines, has been shifted by +1.5%o as described in the text. Adapted from Lee-Thorp and van der Merwe (1987).

98

JULIA A. LEE-THORP

to remain essentially the same, but where deposit carbonates are depleted, grazer apatite values should become significantly more depleted, i.e., a shift of the same order of magnitude as observed for browsers, but of opposite sign. Does this occur? The purpose of this paper is to examine these trends more closely by examining both browsing and grazing animal isotope values in sites of different ages, and with different matrix carbonate values ranging from enriched to depleted. In warm savanna regions, such as occurs over much of southern Africa, browsers and grazers have entirely different carbon isotopic signatures, because browsers eat foliage from -depleted trees and shrubs, while grazers eat grasses which are less depleted (Smith and Epstein 1971). Isotopic values of these diets are reflected in the animal’s bones and teeth, which subsequently become fossilized within archaeological or paleontological sites. The combined sample set allows a test of the model which predicts that diagenesis will take the form of alteration of biogenic isotope values towards those of the matrix (Fig. 5.3). The main assumption in this model is that matrix carbonate in solution is the principle agent for alteration of carbonates in biological apatites. Although the liquid and solid carbonate phases differ isotopically (the latter is enriched by about 8%o relative to former, under equilibrium conditions), the isotopic relationship between phases is well-defined (Emrich et al. 1970); hence values obtained from solid are at least representative of the dissolved phase. As shown in the model in Fig. 5.3, isotopic contamination of

Figure 5.3. Model describing expected carbon isotopic shifts in direction of matrix values for browsers and grazers. Large arrows represent large shifts, small arrows the converse.

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

99

Figure 5.4. Map of South Africa showing location of sites mentioned in the text.

browser apatite is best observed in sites with enriched matrix values, but grazers in these sites have apatite values similar to those of the matrix. For grazers, if alteration towards matrix values occurs, it is best observed in sites where matrix carbonates are depleted (Fig. 5.3). Since these conditions cannot be met in one site, a range of sites with associated fauna of varying age has been chosen. Site locations are shown in Fig. 5.4. Sites with enriched matrix carbonates include Swartkrans and Makapansgat where calcified breccias derive from pre-Cambrian marine dolomites (Brain 1988; MacFadden et al. 1979; Maguire et al. 1980) and Die Kelders where deposit carbonates derive from the marine limestone Bredasdorp Formation or quantities of shell midden (Tankard and Schweitzer 1974; Butzer 1979a). At coastal Klasies River Mouth (KRM), carbonates percolate into the caves from (marine) shell-rich aeolian deposits in the overlying hill (Butzer 1979b; Deacon and Geleijnse 1988). Finally Border Cave (Butzer et al. 1978; Beaumont et al. 1978) is a site with more depleted matrix values, although the deposit is carbonate-poor and probably formed from oxidized soil organic matter. These sites cover a great age range and may be either palaeontological accumulations or archaeological living sites. Makapansgat Member 3 is the oldest at about 3 Ma (MacFadden et al. 1979), while Members 1–3 at Swartkrans span a period of about 1.8 to 1.0Ma (Brain et al. 1988). Both have faunal material accumulated in now-collapsed karstic caverns or shelters, and

100

JULIA A. LEE-THORP

most of the earlier material accumulated as a result of carnivore activity (Maguire et al. 1980), but at least part of Swartkrans Member 3 is believed to represent a hominid living site (Brain et al. 1988). Time ranges of the archaeological cave sites Klasies River Mouth, Border Cave and Die Kelders are broadly similar in that they contain Middle Stone Age (MSA) cultural material dating between ~200 and ~25Ka, and evidence for the presence and behavior of anatomically modern Homo sapiens. For KRM, occupation seems to have been concentrated in the Last Interglacial and early Glacial, while more occupation at Border Cave and especially Die Kelders seem to fall within the last glacial period. These three sites also have Holocene-age capping deposits, and enormous gaps in their sequences. For instance at KRM a hiatus exists between about 70 and 4.5Ka (Singer and Wymer 1981).

METHODS Where possible, animals known to have consistent dietary habits of either browsing or grazing were chosen; decisions were based both on observations of the dietary habits of modern animals as reported in the literature (Smithers 1983), and/or from extensive isotopic surveys (Lee-Thorp 1989). But optimal selection was not always possible, and some compromises had to be made. For instance, a site might lack reliable browsers because the vegetation did not support them and/or humans may not have selected them, and in that case a less reliable browser had to be used. The older the site, the less the fauna resembled modern forms, hence decisions about diets of long extinct forms were made on the basis of factors such as tooth morphology,

according to conclusions of paleontologists. This was a particular problem at Makapansgat, because the Early Makapanian faunal assemblage contains few modern forms (Vrba 1983). A list of species and specimens is given in Table 5.1. Analytical methods followed those described elsewhere (Lee-Thorp 1989). Samples were first cleaned of adhering matrix material; enamel was

separated manually from dentine and cementum using a small hobby drill, jewelers side-cutters and scalpel, and for bone, small pieces were removed. After milling in a Spex mill, samples were pretreated with a dilute solution of sodium hypochlorite (~1.7% v/v) and dilute acetic acid (1M) in turn, to eliminate organic residues such as humates, and contaminating calcium carbonates or soluble apatites, respectively. CO2 was produced by hydrolysis in 100% at 25°C according to standard methods (McCrea 1950), cryogenically distilled and clean, dry introduced into a VG 602E Micromass for isotope ratio measurement. According to standard practice, isotope ratios are reported in the notation relative to the PeeDee Belemnite standard (given in Fig. 5.1). Precision as determined from replicate measurements was <0.1%o.

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

101

Table 5.1. List of all specimens reported in this study indicating sites; species; tissue: bone, enamel, dentine or tooth (enamel & dentine), or matrix carbonates; stratigraphic context; approximate age according to available dating estimates and stratigraphy; and Specimens have been assigned UCT Archaeometry Research Unit numbers. Stratigraphic context follows publications referenced in the text. Ages for the Border cave strata follow ESR determinations of Grün et al. 1990,

which are considerably younger than radiocarbon determinations for the upper strata, but are used here because they have greater time depth. Results for the sites of Die Kelders, Swartkrans, Klasies River Mouth, Makapansgat and a few Border Cave specimens have been published previously in Lee-Thorp and van der Merwe (1987, 1994), new results are marked*.

102

JULIA A. LEE-THORP

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

103

RESULTS The results for all sites are given in Table 5.1, and are best considered by dividing sites into three groups according to isotopic nature of the matrix: (i) sites with most isotopically enriched matrix carbonates (Die Kelders and Swartkrans), (ii) sites with rather less enriched carbonates (Klasies River Mouth and Makapansgat), and finally (iii) a site with depleted deposit values (Border Cave). This is summarized in Fig. 5.5. The division also fortuitously provides a range of age depths in two categories. As indicated in Table 5.1, many of these data have been published elsewhere, but the purpose for which they are considered in combination here has not been previously attempted.

Sites with Enriched Matrix Carbonates Deposit carbonates in both Die Kelders and Swartkrans are derived origi-

nally from marine sources and hence have correspondingly enriched isotopic values. Those in Die Kelders have not been measured directly. Several determinations of calcite inclusions and breccia in Swartkrans cluster near –1.3%o (Table 5.1). Hence the distinction between browser apatite and matrix carbonates at these sites is large, of the order of –11%o. There are two complications in this sample set. First, very few reliable browsing herbivores were available due to dominance of the bovid assemblage by grazers. The only animal in the sample known from modern observations to be a reliable browser is the klipspringer (Oreotragus oreotragus) (Smithers 1983), and in the absence of any others, kudu (Tragelaphus strepsiceros), which is not an obligatory browser, is used. Second,

all samples from Die Kelders were bone from a small solitary browser, Cape grysbok (Raphicerus melanotis); no tooth enamel samples were available.

At Die Kelders there is little to distinguish the archaeological browser bone apatite values from the corrected modern mean of –13%o. One sample (UCT 704) is more enriched at –10.1%o, however the possibility exists that a few of these specimens are a related, more opportunistic species, the steenbok (R. campestris), which eats some grass. In the older site, Swartkrans, enamel values for “browsers” cluster near –11% o , where bone data exist, they tend to be more scattered. Enamel values for grazers in Member 1 cluster near –1%o, compared to mean values for modern grazers in a C4 environment of –0.5%o (Lee-Thorp and van der Merwe 1987), and a fossil fuel-corrected effect of +1%o (Fig. 5.1). Several of the grazers in Member 2 have slightly more negative values. Overall, browser and grazer enamel values at Swartkrans are probably altered by about +2%o, and –1 to –2%o, respectively.

Sites with Slightly Less Enriched Matrix Carbonates These sites have deposit carbonates slightly less enriched than the previous group: –3.2 ± 2.1%o at Klasies River Mouth (where matrix values are

104

JULIA A. LEE-THORP

Figure 5.5. Summarized data Tor browsers and grazers, expressed as dark shaded and crosshatched boxes, respectively, incorporating means and standard deviations, from the three groups of sites plotted against time: (a) Die Kelders and Swartkrans, (b) Klasies and Makapansgat, and (c) Border Cave. Typical matrix values are shown as light shaded rectangles.

inhomogenous), and –4.3%o at Makapansgat (Table 5.1). At Klasies River Mouth there were few grazers, and in any case the present mix of and vegetation seems to have persisted into the Pleistocene and present (Sealy 1996), so that grazers could not provide the required “endpoint”. The only animals sampled at Klasies River Mouth were bushbuck (Tragelaphus scriptus)

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

105

and Raphicerus sp., where the latter may include both R. melanotis (Cape grysbok, an obligate browser) and R. campestris (steenbok, predominantly a browser). Makapansgat Member 3 is by far the oldest site and the Early Makapanian faunal assemblage consists of many unrecognizable ancient forms; hence apart from Giraffa camelopardalis (giraffe), preliminary browser/grazer distinctions were made on the basis of palaeonlological attributions. For Klasies, although most values for both enamel and bone apatite fall within one standard deviation of the mean of (corrected) modem browser values (Fig. 5.5), some bone specimens fall outside this range. These enriched specimens suggest that a limited degree of equilibration with matrix carbonates has taken place, although inclusion of a limited amount of grass in the diet is a plausible alternative explanation for UCT #1025, Raphicerus sp. which as noted above could be the more opportunistic species, Raphicerus campestris. Isotopic differences between modern and fossil browsers, and between enamel and bone, are more marked at Makapansgat. Enamel values for the various presumed browsers, mainly giraffids and Cephalopus pricei (Table 5.1) fall between –9.8 and –11.5%o, but bone apatite values for two of the same C.

pricei individuals are –3.7 and –7.4%o. The difference between browser enamel values and the corrected modern mean represents an enrichment of approximately 3%o, which according to the diagenesis model outlined above implies that approximately 20% of original enamel carbonate has been altered. The equid and Hipparion specimens are both equids/grazers, and have values of +0.2%o. The isotopic difference between browser and grazer enamel remains of the order of 10%o. As noted elsewhere, because the faunal material was buried deep within a cave system permeated by bicarbonate/carbonate saturated waters (Wells 1971), these specimens are regarded as a worst-case scenario with conditions most amenable to matrix carbonate contamination (Lee-Thorp 1989, Lee-Thorp and van der Merwe 1987).

Site with Depleted Matrix Carbonates Unlike all the other sites discussed in this paper, the deposits in Border Cave are extremely low in carbonates, and a single value of –13.0%o was extracted with difficulty from matrix adhering to a bone sample from Level 1WA near the top of the sequence. Such a depleted value, similar to carbonates in soils derived from vegetation suggests that the small amount of carbonate is probably derived from oxidation of organic material in the deposit. The site and its contents are unusual in other respects; preservation of organic material such as grass bedding is exceptional even at time depths of 40 Ka, mobility of elements such as uranium into teeth appears to be minimal implying continuous aridity with no penetration of water (Grün et al. 1990), and

temperatures are believed to have remained fairly constant at about 16–18°C (Beaumont et al. 1978). There were few browsers in the faunal assemblage,

106

JULIA A. LEE-THORP

and some less reliable, opportunistic browsers were also included (eland, Taurotragus oryx), as well as Potomochoerus porcus (bushpig), which is an omnivore but at least concentrates on -derived foods (Smithers 1983). The chronology based on ESR measurements of tooth enamel has been followed here because it is more complete, although these dates are almost certainly too young compared to the radiocarbon chronology (Grün et al. 1990; Beaumont et al. 1978). At Border Cave, browser values range between –10.1 and –13.3%o. One outlier at +1.8%o, identified as T. oryx has been excluded from subsequent consideration, as this animal was patently a grazer and was probably misidentified. Another T. oryx specimen was slightly more enriched than expected, but included in subsequent calculations. The remaining animals, including the suids, fell within the corrected mean for modern animals. The grazers are more instructive. All grazer values are >0%o, and some are very enriched indeed, beyond the range of corrected modern grazers (mean There is no evidence of any shift towards matrix values, or any other sources (such as atmospheric There appears to be no unidirectional relationship with time, although there is a hint of an alternating pattern which bears further investigation (see Fig. 5.6). The average difference between browsers and grazers is +13.8%o, almost identical to that observed between the modern counterparts (see Fig. 5.1). Two possibilities for the observed enriched values for many of the grazers, and the often concomitant but slight enrichment of browsers (Fig. 5.6), present themselves. Either these shifts represent atmospheric enrichment shifts at particular periods during the last ~100Ka represented (the Late Pleistocene), or a hitherto unknown or unrecognized fractionation process has taken place during burial and fossilization. The former hypothesis could be tested by comparison of observations from a larger sample set from the site, with concentration and carbon isotope data from Antarctic ice-core records or high resolution marine isotope records.

DISCUSSION AND CONCLUSIONS When carbon isotope data from faunal bone and enamel apatite of up to about 100–120 Ka in age are compared with modern data corrected for the fossil fuel effect, they correspond well with predicted ranges according to diet. All of these sites have well-preserved skeletal material, and almost all have highly calcareous deposits, with the exception of Border Cave where exceptional preservation is ascribed to long-term aridity. Deposit carbonates, which can be readily mobilized, are assumed to be the major contaminating medium for introduction of exogenous carbonate into apatite structure, although it is also possible that exchange may occur with Carbonates in many of the sites discussed are relatively enriched, and thus form a sharp contrast with

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

Figure 5.6.

107

data for browsers and grazers from Border cave, plotted against age (vertical

scale). The shaded areas represent the

values of modern grazers and browsers. Filled squares

and circles are tooth enamel; the open square is bone. The chronology obtained from ESR measurements of faunal enamel has been used; hence the sequence appears somewhat younger than

outlined previously (Grün et al. 1990; Beaumont et al. 1978). The sequence is in fact longer, but no isotopic data are available beyond Stratum 1GBS.LRA (~80 Ka).

expected depleted values for browser apatites. In Border Cave, the site with more depleted deposit values where a sharper contrast with grazers exists, likewise no evidence for alteration or exchange towards matrix values is apparent. The lack of any evidence for exchange coupled with immobility and low concentration of carbonate in the deposit perhaps suggests that the exchange model is not applicable in this case. Such circumstances are rare, but Saliège et al. (1995) found no evidence for exchange in ratios and

in bone mineral from 1,000–5,000 year old skeletons sealed in dry Saharan tombs.

108

JULIA A. LEE-THORP

Many of the Border Cave grazers are unusually enriched (up to +3.3‰). These shifts seem to be matched by slight contemporary positive shifts for

which could have been ascribed to environmental effects on plants except that the microfaunal and other evidence argues against this

(Avery 1992). These shifts may point to evidence for atmospheric

enrich-

ment during periods within the last Glacial period. Although the few existing studies of carbon isotopic composition of the last Glacial atmosphere have indicated more depleted values (Leuenberger et al. 1992; Marino et al. 1992), it must be remembered that these observations pertain to the Last Glacial Maximum. The LGM is not representative of the entire Glacial period. It has been suggested that earlier periods within the last Glacial, possibly during early Oxygen Isotope Stage 3 at about 45–60 Ka, saw most intense marine

productivity (Summerhayes et al. 1995), and depending on equilibria related to carbon burial rates and water column conditions, it is possible that atmospheric became enriched at that time. This suggestion remains speculative until more observations from marine and terrestrial contexts are forthcoming. In the million year range, however, there is clearer evidence for isotopic alteration which also apparently increases with age. Unfortunately, at present there are few reliable browser and grazer data beyond the Late Pleistocene, i.e., in the Middle Pleistocene, which would allow determination of the periods during which bone apatite and enamel begin to deviate. The indications are that the timing should be different. Material from the site of Florisbad falls within the late Middle Pleistocene (~ 125–200 Ka), but at present carbon isotope data are only available for three species of uncertain and/or opportunistic diets (Brink and Lee-Thorp 1992), so that extent or direction of isotopic alteration is impossible to separate from dietary vagaries. Browser enamel values at Swartkrans (~1.5–1.8 Ma, depending on the Member), on average are slightly more enriched than expected, by about 2‰, while grazer values are altered by perhaps +1%o. These results support the idea that isotopic alteration is in the direction of matrix values for both browsers and grazers, hence diagenetic carbonate in the apatite structure derives from deposit carbonates. As a result of these shifts the difference between browsers and grazers is diminished from 14‰ (modern) to 11‰. By this stage bone apatite, although pretreated with standard acetic acid procedures, may be significantly altered but to variable and thus unpredictable degrees. At the older site of Makapansgat, bone apatite is almost completely altered and one sample is indistinguishable from deposit carbonate values.

Browser enamel remains isotopically distinct, but is altered by +3‰ on average whereas grazer values are similar to those at Swartkrans, near 0‰. The difference between browsers and grazers is diminished to 10‰. Because Makapansgat breccia carbonates are more depleted than Swartkrans (–4‰ compared to –1‰), one might have expected grazer values to have moved

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

109

further in the negative direction. This apparently uneven alteration could possibly be explained by enrichment of the flora due to environmental effects, such as water or temperature stress (Tieszen 1991). However this seems implausible since by all accounts the environment during the Makapansgat Member 3 period was humid and fairly closed (Vrba 1983). Another possibility is that the entire foodweb at Makapansgat was more positive in the first

place, i.e., originally grazers may have been +2‰, browsers –12‰, and atmospheric nearer –5‰, for instance. Recent studies of African aeolian dust records in marine cores have pinpointed a major global climate shift near 2.8 Ma (DeMenocal 1996), elsewhere attributed to about 2.6 Ma on faunal grounds (Vrba 1988), which seems to be synchronous with major changes in mammalian faunal assemblages including hominids. The effect, if any, on atmospheric concentration and is not known. The Makapansgat Member 3 fauna clearly existed before this major change. The data available here are too few to allow anything more than speculation; further terrestrial and marine isotopic data from the periods prior to, and post the 2.6–2.8Ma shift might provide firmer indications. The results presented here confirm that determination of unknown diets at great time depths, using carbon isotope analysis of enamel, is entirely feasible, as shown previously (Lee-Thorp et al. 1994; Koch et al. 1994; Bocherens et al. 1996), although it is advisable to determine the end-member carbon isotopic values of the spectrum using known browsers and grazers. Studies of even older enamel of Miocene age (Quade et al. 1992; Morgan et al. 1994; Wang et al. 1994) indicate that isotopic values for animals with diets do not alter beyond –10‰, suggesting that the material at Makapansgat has reached a kind of diagenetic endpoint at least in terms of carbon isotopes. The data also confirm that bone apatite from these Pliocene/Upper Pleistocene sites is too badly altered to be used for such purposes at time depths of millions of years. The results for fauna dating to within the late Pleistocene, however, show that in addition to enamel, bone apatite is also surprisingly

well-preserved isotopically in this time range, at these sites. It is puzzling because many studies (e.g., Schoeninger and DeNiro 1982; Koch et al. 1990; Wright and Schwarcz 1996) have demonstrated substantial alteration even in quite young bone apatite specimens—some of which are only hundreds or a few thousand years old, and certainly recent compared to those reported here. One exception is the Saharan study mentioned above where biogenic ratios and radiocarbon dates closely comparable to dates obtained from the bone collagen were obtained in exceptionally arid burial conditions (Saliège et al. 1995). Could the crucial difference be that the specimens used in the former studies are on a pathway to destruction rather than fossilization, whereas those from Die Kelders, Klasies and Border Cave are buried in environments conducive to fossilization with minimal chemical alteration? This

110

JULIA A. LEE-THORP

material might have survived for millions of years longer if not excavated. The minimal isotopic alteration implies that processes such as crystal growth and recrystallization leading to frequently observed increased crystallinity have not led to (detectable) incorporation of exogenous carbonates in the crystal structure. At Border Cave an increase in crystallinity has been inferred from infrared splitting indices (Sillen and Morris 1996). Therefore structural changes may have merely involved the rearrangement (Ostwaldian ripening) and/or incorporation of in situ “raw material”, as suggested in the discussion above. How then does one reconcile the observation that in the much older material substantial isotopic alteration has occurred in bone apatite but not enamel? One possibility is that in contrast to enamel, and in spite of its increased stability over the formerly reactive state, fossil bone apatite crystals remain more soluble with incorporation of relatively more isomorphic substitutions (e.g., carbonates, strontium, magnesium, transition metals), which make it susceptible to later cycles of dissolution and recrystallization. The slower processes of solid-state ionic and isotopic exchange continue to operate in both tissues, but again bone apatite remains more susceptible to exchange than the denser, more crystalline, enamel. On the basis of the data presented here together with consideration of the implications of the enamel radiocarbon data (Hedges et al. 1995) I would suggest that the solid-state processes of ionic and isotopic exchange are the main pathway for alteration of fossil enamel. Since these are exceedingly slow processes, enamel remains

almost untouched, in terms of carbon isotopes, over time scales of millions of years.

ACKNOWLEDGMENTS I dedicate this contribution in fond memory of Hal Krueger. He was a fine scientist who influenced profoundly the new directions in isotopic studies of diet, and did this with flair and determination in the face of resistance. I am very grateful for his advice and support over the years. In the past we disagreed mildly about bone apatite survival and I think he would have been pleased with this paper. I also thank Andrew Sillen, Paul Koch and Nikolaas van der Merwe for many helpful discussions about diagenesis, preservation, atmospheric effects and related matters. Mark Pollard first drew my attention to the possibility that atmospheric changes might influence interpretation of the Makapansgat data. I am grateful to the excavators and curators who have permitted sampling and analysis of the faunal assemblages, including (in no particular order) Dr. C.K. Brain, Prof. H.J. Deacon, Dr. M. Avery, Dr. G. Avery, Mr. P. Beaumont, Prof. J. Kitching, Dr. J. Maguire, Dr. J.F Thackeray and Dr. V. Watson. Funding for this research came from the Foundation for Research Development, South Africa, and the University Research Council, University of Cape Town.

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

111

REFERENCES Ambrose, S.H. 1998 Prospects for stable isotopic analysis of Later Pleistocene hominid diets in West Asia and Europe. In Akazawa, T., Aoki, K. and Bar-Yosef, O., eds., Neanderthals and

Modem Humans in West Asia. New York, Plenum Press: 277–289. Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relationship of carbon isotope ratios of whole diet and dietary protein to those of bone collagen and carbonate. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 1–37. Anderson, T.F. 1969 Self-diffusion of carbon and oxygen in calcite by isotope exchange with carbon

dioxide. Journal of Geophysical Research 74: 3918–3932. Avery, D.M. 1992 The environment of early modern humans at Border cave, South Africa: micromammalian evidence. Palaeogeography, Palaeoclimatolology Palaeoecology 91: 71– 87. Ayliffe, L.K., Chivas, A.R. and Leakey, M.G. 1994 The retention of primary oxygen isotope compositions of fossil elephant skeletal phosphate. Geochimica et Cosmochimica Acta 58: 5291–5298. Baud, C.S., Bang, S. and Very, J.M. 1977 Minor elements in bone mineral and their effects on its solubility. Journal de Biologie Buccale 5: 195–207. Beaumont, P.B., de Villiers, H. and Vogel, J.C. 1978 Modern man in sub-Saharan Africa prior to 49,000 years B.P.: a review and evaluation with particular reference to Border Cave. South

African Journal of Science 74: 409–419. Bell, L.S. and Lee-Thorp, J.A. in press Advances in the study of human skeletal remains: a joint perspective. In Cox, ed., Grave Concerns, Council for British Archaeology. Besha, K., Key, C., Glimcher, M., Schimizu, M. and Griffin, R.G. 1990 Solid state carbon-13 and proton NMR studies of carbonate-containing calcium phosphates and enamel. Journal of Solid State Chemistry 84: 71–81. Betts, F., Blumenthal, N.C. and Posner, A.S. 1981 Bone mineralization. Journal of Crystal Growth 53: 63–73. Biltz, R.M. and Pellegrino, E.D. 1977 The nature of bone carbonate. Clinical Orthopaedics and Related Research 129: 279–292. Blumenthal, N.C., Belts, F. and Posner, A.S. 1975 Effect of carbonate and biological macromolecules on the formalion and properties of hydroxyapatite. Calcified Tissue Research 18: 81–90. — 1977 Stabilization of amorphous calcium phosphate by Mg and ATP Calcified Tissue Research

23: 245–250. Bocherens, H., Koch, P.L., Mariotti, A., Geraads, D. and Jaeger, J.-J. 1996 Isotopic biogeochemistry of mammalian enamel from African Pleistocene hominid sites. Palaios 11: 306–318.

Boskey, A.L. 1981 Current concepts of the physiology and biochemistry of calcification. Clinical Orthopaedics and Related Research 157: 165–195. Boyde, A. 1967 The development of enamel structure. Proceedings of the Royal Society of Medicine

60: 923–933. Brain, C.K. 1988 New information from the Swartkrans cave of relevance to robust australopithecines. In Grine, F.E., ed., Evolutionary History of the Robust Australopithecines. New York, Aldine de Gruyter: 405–426. Brain, C.K., Churcher, C.S., Clark, J.D., Grine, F.E., Shipman, P., Susman, R.L., Turner, A. and Watson, V. 1988 New evidence of early hominids, their culture and environment from Swartkrans Cave. South African Journal of Science 84: 828–835. Brink, J.S. and Lee-Thorp, J.A. 1992 The feeding niche of an extinct springbok, Antidorcas bondi (Antilopinae, Bovidae), and its palaeoenvironmental meaning. South African Journal of Science 88: 227–229.

112

JULIA A. LEE-THORP

Bryant, J.D., Luz, B. and Froelich, P.N. 1994 Oxygen isotopic composition of fossil horse tooth phosphate as a record of continental palaeoclimate. Palaeogeography, Palaeoclimatology, Palaeoecology 107: 303–316. Butzer, K.W. 1979a Preliminary notes on the geology of Die Kelders. Annals of the South African Museum 78: 225–228. — 1979b Sediment stratigraphy of Middle Stone sequences at Klasies River Mouth. South African Archaeology Bulletin 33: 141–151. Butzer, K.W, Beaumont, P.B. and Vogel, J.C. 1978 Lithostratigraphy of Border Cave: a Middle Stone Age sequence beginning at c. 195,000 B.P. Journal of Archaeological Science 5: 317–341. Cerling, T., Harris, J.M., Ambrose, S.H., Leakey, M.G. and Solounias, N. 1998 Dietary and environmental reconstruction with stable isotope analyses of herbivore tooth enamel from the Miocene locality of Fort Ternan. Journal of Human Evolution 33: 635–650. Chickerur, N.S., Tung, M.S. and Brown, WE. 1980 A mechanism for the incorporation of carbonate into apatite. Calcified Tissue International 32: 55–62. Deacon, H.J. and Geleijnse, VB. 1988 The stratigraphy and sedimentology of the main site sequence, Klasies River, South Africa. South African Archaeological Bulletin 43: 5–14. DeMenocal, P. 1996 Plio-Pleistocene African climate. Science 270: 53–59. Driessens, F.C.M., van Dijk, J.W.E. and Borrgreven, J.M.P.M. 1978 Biological calcium phosphates and their role in the physiology of bone and dental tissues. 1. Composition and solubility of calcium phosphates. Calcified Tissue Research 26: 127–137. Eanes, E.D. 1979 Enamel apatite: chemistry, structure and properties. Journal of Dental Research Special Issue B: 829–836. Eanes, E.D. and Posner, A.S. 1970 A note on the crystal growth of hydroxyapatite precipitated from aqueous solutions. Materials Research Bulletin 5: 377–384. Ehleringer, J.R., Cerling, T.E. and Helliker, B.R. 1997 C4 photosynthesis, atmospheric and climate. Oecologia 112: 285–299. Eisenmann, D.R. 1985 Enamel structure. In Ten Cate, A.R., ed., Oral Histology: Development Structure and Function, 2nd Edition. Toronto, C.V Mosby Co.: 198–217. Emrich, K., Emhalt, D.H. and Vogel, J.C. 1970 Carbon isotope fractionation during the precipitation of calcium carbonate. Earth and Planetary Science Letters 8: 363–371. Falkowski, P.G. 1997 Evolution of the nitrogen cycle and its influence on the biological sequestration of in the ocean. Nature 387: 272–275. Friedli, H., Lotscher, H., Oeschger, H., Siegenthaler, U. and Stauffer, B. 1986 Ice core record of the record in the past two centuries. Nature 324: 237–238. Grün, R., Beaumont, P.B. and Stringer, C. 1990 ESR dating evidence for early modern humans at Border cave in South Africa. Nature 344: 537–539. Hassan, A.A. 1975 Geochemical and Mineralogical Studies on Bone Material and their Implications for Radiocarbon Dating. Unpublished Ph.D. thesis, Southern Methodist University, Texas. Hassan, A.A. and Ortner, D.J. 1977 Inclusions in bone material as a source of error in radiocarbon dating. Archaeometry 19: 131–135. Hassan, A.A., Termine, J.D. and Haynes, C.V. 1977 Mineralogical studies on bone apatite and their implications for radiocarbon dating. Radiocarbon 19: 364–374. Haul, R.A.W. and Stein, L.H. 1955 Diffusion in calcite crystals on the basis of isotope exchange with carbon dioxide. Faraday Society Transactions 51: 1280–1290. Hedges, R.E.M., Lee-Thorp, J.A. and Tuross, N. 1995 Is tooth enamel carbonate a suitable material for radiocarbon dating? Radiocarbon 37: 285–290. Koch, PL., Behrensmeyer, A.K., Tuross, N. and Fogel, M.L. 1990 Isotopic fidelity during bone weathering and burial. Annual Report of the Director, Geophysical Laboratory, 1989–1990. Washington, D.C., Carnegie Institution of Washington: 105–110. Koch, P.L., Fogel, M.L. and Tuross, N. 1994 Tracing the diets of fossil animals using stable isotopes. In Lajtha, K. and Michener, B., eds., Stable Isotopes in Ecology and Environmental Science. Oxford, Blackwell: 63–92.

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS

113

Koch, P.L., Tuross, N. and Fogel, M.L. 1997 The effects of sample treatment and diagenesis on the isotopic integrity of carbonate in biogenic hydroxylapatite. Journal of Archaeological Science

24: 417–429. Koch, P.L., Zachos, J.C. and Gingerich, P.D. 1992 Correlation between isotope records in marine

and continental carbon reservoirs near the Paleocene/Eocene boundary. Nature 358: 319–322. Krueger, H.W. 1991 Exchange of carbon with biological apatite. Journal of Archaeological Science 18: 355–361. Krueger, H.W. and Sullivan, C.H. 1984 Models for carbon isotope fractionation between diet and bone. In Turnland, J.F. and Johnson, P.E., eds., Stable Isotopes in Nutrition. ACS Symposium Series 258. Washington, DC, American Chemical Society: 205–222. Land, L.S., Lundelius, E.L. and Valastro, S. 1980 Isotopic ecology of deer bones. Palaeogeography,

Palaeoclimatology, Palaeoecology 32: 143–151. Lee-Thorp, J.A. 1989 Stable Carbon Isotopes in Deep Time. The Diets of Fossil Fauna and Hominids. Unpublished Ph.D. dissertation, University of Cape Town. Lee-Thorp, J.A., Sealy, J.C. and van der Merwe, N.J. 1989a Stable carbon isotope ratio differences

between bone collagen and bone apatite, and their relationship to diet. Journal of Archaeological Science 16: 585–599. Lee-Thorp, J.A., van der Merwe, N.J. and Brain, C.K. 1989b Isotopic evidence for dietary differences between two extinct baboon species from Swartkrans. Journal of Human Evolution

18: 183–190. Lee-Thorp, J.A. and van der Merwe, N.J. 1987 Carbon isotope analysis of fossil bone apatite. South

African Journal of Science 83: 71–74. —1991 Aspects of the chemistry of modern and fossil biological apatite. Journal of Archaeological Science 18: 343–354. Lee-Thorp, J.A., van der Merwe, N.J. and Brain, C.K. 1994 Diet of Australopithecus robustus at Swartkrans deduced from stable carbon isotope ratios. Journal of Human Evolution 27: 361–372. LeGeros, R.Z. 1981 Apatites in biological systems. In Pamplin, B., ed., Progress in Crystal Growth and Characterization. Vol. 4. New York, Pergamon Press: 1–45. —1983 Ultrastructural properties of human enamel apatite. In Lazzari, E., ed., Handbook of

Experimental Aspects of Oral Biochemistry, Florida, CRC Press: 159–179. LeGeros, R.Z. and Tung, M.S. 1983 Chemical stability of the carbonate- and fluoride-containing apatites. Caries Research 17: 419–429. Leuenberger, M., Siegenthaler, U. and Langway, C.C. 1992 Carbon isotope composition of atmospheric during the last ice age from an Antarctic ice core. Nature 357: 488– 490. MacFadden, P.L., Brock, A. and Partridge, T.C. 1979 Palaeomagnetism and the age of the Makapansgat Limeworks site. Earth and Planetary Science Letters 44: 373–382. Maguire.J.M., Pemberton, D. and Collett, M.H. 1980 The Makapansgat Limeworks grey breccia: hominids, hyenas, hystricids or hillwash? Palaeontologio Africana 23: 75–98. Marino, B.D. and McElroy, M.B. 1991 Isotopic composition of atmospheric inferred from carbon in C4 plant cellulose. Nature 349: 127–131. Marino, B.D., McElroy, M.B., Salawitch, R. and Spaulding, W.G. 1992 Glacial to interglacial variations in the carbon isotopic composition of atmospheric Nature 357: 461–466. McCrea, J.M. 1950 On the isotopic chemistry of carbonates and a paleotemperature scale. Journal

of Chemical Physics 18(6): 849–857. Michel, V., Ildefonse, P. and Morin, G. 1995 Chemical and structural changes in Cervus elephus tooth enamels during fossilization (Lazaret Cave): a combined IR and XRD Rietveld analy-

sis. Applied Geochemistry 10: 145–159. Millard, A.R. and Hedges, R.E.M. 1995 The role of the environment in uranium uptake by buried bone. Journal of Archaeological Science 22: 239–250.

114

JULIA A. LEE-THORP

Morgan, M.E., Kingston, J.D. and Marino, B.D. 1994 Carbon isotopic evidence for the emergence

of plants in the Neogenc from Pakistan and Kenya. Nature 367: 162–165. Nelson, D.G.A., Featherstone, J.D.B., Duncan, J.E and Cuttress, T.W. 1982 Paracrystalline disorder of biological and synthetic carbonate-substituted apatites. Journal of Denial Research

61: 1274–1281. Pellegrino, E.D. and Biltz, R.M. 1968 Bone carbonate and the Ca to P molar ratio. Nature 219: 1261–1262. Person, A., Bocherens, H., Saliège, J.-F, Paris, F., Zeitoun, V. and Gérard, M. 1995 Early diagenetic evolution of bone phosphate: an x-ray diffractometry analysis. Journal of Archaeological

Science 22: 211–221. Posner, A.S. 1985 The mineral of bone. Clinical Orthopaedics 200: 87–99. Poyart, C.F., Freminet, A. and Burseaux, E. 1975 The exchange of bone in vivo . Respiratory Physiology 25: 101–107. Quade, J., Cerling, T.E., Barry, J.C., Morgan, M.E., Pilbeam, D.E., Chivas, A.R., Lee-Thorp, J.A. and van der Merwe, N.J. 1992 A 16-Ma record of paleodiet using carbon and oxygen isotopes in fossil teeth from Pakistan. Chemical Geology (Isotope Geoscience) 94: 183–192. Rey, C., Renugopalakrishnan, V., Shimizu, M., Collins, B. and Glimcher, M.J. 1991 A resolutionenhanced Fourier transform spectroscopic study of the environment of the ion in the mineral phase of enamel during its formation and maturation. Calcified Tissue International 49: 259–268. Rink, W.J. and Schwarcz, H.P. 1995 Tests for diagenesis in tooth enamel: ESR dating signals and carbonate contents. Journal of Archaeological Science 22: 251–255. Saliège, J.-F., Person, A. and Paris, F. 1995 Preservation of

original ratios and

dating of

the mineral fraction of human bones from Saharan tombs, Niger. Journal of Archaeological Science 22: 301–312. Sarntheim, M., Winn, K., Duplessey, J.-C. and Fontugne, M.R. 1988 Global variations of surface ocean primary productivity in low and mid latitudes: influence on reservoirs of the deep ocean and atmosphere during the last 21,000 years. Paleoceanography 3:

361–399. Schoeninger, M.J. 1982 Diet and evolution of modern human form in the Middle East. American Journal of Physical Anthropology 58: 37–52. Schoeninger, M.J. and DeNiro, M.J. 1982 Carbon isotope ratios of apatite from bone cannot be used to reconstruct diets of animals. Nature 297: 577–578. Sealy, J.C. 1996 Seasonality of rainfall around the Last Glacial Maximum as reconstructed from carbon isotope analyses of animal bones from Nelson Bay Cave. South African Journal of Science 92: 441–444. Sealy, J.C., van der Merwe, N.J., Sillen, A., Kruger, F.J. and Krueger, H.W. 1991 as a dietary indicator in modern and archaeological bone. Journal of Archaeological Science 18: 399–416. Sillen, A. 1986 Biogenic and diagenetic Sr/Ca in Plio-Pleistocene fossils of the Omo Shungura formation. Paleobiology 12: 311–323. —1989 Diagenesis of the inorganic phase of cortical bone. In Price, T.D., ed., Bone Chemistry and Past Behaviour. Cambridge, Cambridge University Press: 211–229. Sillen, A. and LeGeros, R. 1991 Solubility profiles of synthetic apatites and of modern and fossil bones. Journal of Archaeological Science 18: 385–397. Sillen, A. and Morris, A.G. 1996 Diagenesis of bone from Border cave; implications for the age of the Border cave hominids. Journal of Human Evolution 31: 499–506. Sillen, A. and Parkington, J.E. 1996 Diagenesis of bones from Eland’s Bay Cave. Journal of Archaeological Science 23: 535–542. Simpson, D.R. 1972 Problems of the composition and structure of the bone minerals. Clinical Orthopaedics and Related Research 86: 260–286. Singer, R. and Wymer, J. 1982 The Middle Stone Age at Klasies River Mouth in South Africa. Chicago, University of Chicago Press.

PRESERVATION OF BIOGENIC CARBON ISOTOPIC SIGNALS Smith, B.N. and Epstein, S. 1971 Two categories of

115

ratios for higher plants. Plant

Physiology 47: 380–384. Smithers, R.H.N. 1983 The Mammals of the Southern African Sub-region. Pretoria, University of Pretoria. Sullivan, C.H. and Krueger, H.W. 1981 Carbon isotope analysis of separate chemical phases in modern and fossil bone. Nature 292: 333–335.

__1983 Carbon isotope ratios of bone apatite and animal diet reconstruction. Nature 301: 177–178. Summerhayes, C.P., Kroon, D., Rosell-Mele, P. Jordan, R.W., Schrader, H.-J., Hearn, R., Villeneuva,

J., Grimalt, J.O. and Eglington, G. 1995 Variability in the Benguela Current upwelling system over the past 70,000 years. Progress in Oceanography 35: 207–251. Tankard, A.J. and Schweitzer, F.R. 1974 The geology of Die Kelders cave and environs: a palaeoenvironmental study. South African Journal of Science 70: 365–369. Tieszen, L.L. 1991 Natural variations in the carbon isotope values of plants: implications for archaeology, ecology, and paleoecology. Journal of Archaeological Science 18: 227–248. Tiezsen, L.L. and Fagre, T. 1993 Effect of diet quality and composition on the isotopic composition of respiratory bone collagen, bioapatite and soft tissues. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 121–155. Tuross, N., Behrensmeyer, A.K., Eanes, E.D., Fisher, L.W. and Hare, P.E. 1989 Molecular preservation and crystallographic alterations in a weathering sequence of wildebeest bones. Applied Geochemistry 4: 261–270. Van Klinken, G.J., Richards, M.P. and Hedges, R.E.M. this volume Stable isotopic variations in past European human populations: environmental, ecophysiological and cultural effects. In Ambrose, S. and Katzenberg, M.A., eds., Close to the Bone: Biogeochemical Approaches to Paleodietary Analysis in Archaeology. New York, Plenum Press. Vogel, J.C. 1982 Isotopic evidence for the past climates and vegetation of Africa. Bothalia 14: 391–394. Vrba, E. 1983 Biostratigraphy and chronology, based particularly on Bovidae, of southern African hominid-associated assemblages: Makapansgat, Sterkfontein, Taung, Kromdraai, Swartkrans; also Elandsfontein (Saldanha), Broken Hill (now Kabwe) and Cave of Hearths. In de Lumley, H. and de Lumley, M.A., eds., Proceedings of Congress International de

Paléontologie Humaine, Vol. 2. Nice, Union Internationale des Sciences Préhistoriques et Protohistoriques, Nice: 707–752. —1988 Late Pliocene climatic events and hominid evolution. In Grine, F.E., ed., Evolutionary History of the Robust Australopithecus. New York, Aldine de Gruyter: 405–426.

Wang, Y. and Cerling, T.E. 1994 A model for fossil tooth and bone diagenesis: implications for paleodiet reconstruction from stable isotopes. Palaeogeography, Palaeoclimatology, Palaeoecology 107: 596–606. Wang, Y., Cerling, T.E. and MacFadden, B.J. 1994 Fossil horses and carbon isotopes: new evidence for Cenozoic dietary, habitat, and ecosystem changes in North America. Palaeogeography, Palaeoclimatology, Palaeoecology 107: 269–279. Weiner, S. and Bar-Yosef, O. 1990 States of preservation of bones from prehistoric sites in the Near East: a survey. Journal of Archaeological Science 17: 187–196. Weiner, S., Goldberg, P. and Bar-Yosef, O. 1993 Bone preservation in Kebara Cave, Israel, using

on-site Fourier transform infrared spectrometry. Journal of Archaeological Science 20: 613–627. Wells, A.W. 1971 Cave calcite. Studies in Speleology 2: 129–148. Wheeler, E.J. and Lewis, D. 1977 An X-ray study of the paracrystalline nature of bone apatite. Calcified Tissue Research 24: 243–248. Wright, L. and Schwarcz, H. 1996 Infrared evidence for diagenesis of bone apatite at Dos Pilas: paleodietary implications. Journal of Archaeological Science 23: 933–944.

This page intentionally left blank

Chapter

6

Tooth Oxygen Isotope Ratios As Paleoclimate Monitors In Arid Ecosystems MARGARET J. SCHOENINGER, MATTHEW J. KOHN, AND JOHN W. VALLEY ABSTRACT We report the oxygen stable isotope ratios in tooth enamel of several desertadapted species from east of Lake Turkana in northern Kenya to assess their potential as monitors of temperature and humidity in paleoclimate studies. The data are considered in relation to two recently proposed models of in large mammals. The first model scales physiological parameters pertaining to oxygen inputs and outputs according to body mass for animals. The second model considers these parameters independent of body size. With one exception the data show an inverse relation between body size and as expected but the range of values is five times that predicted by the body size model. This is attributed to the lack of surface water drinking by these species. Two species (dik dik and oryx) have which meet predictions of the second model when unique behavioral patterns are considered. Such species would not be useful in paleoclimate studies because the behavioral patterns would not be known in fossil species. In contrast, Burchell’s zebra would be useful because as an obligate drinker, the in this species has a consistent relation with Grant’s gazelle may serve as a monitor of humidity but further study is required to demonstrate this conclusively. Biogeochemical Approaches to Paleodietary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

117

118

MARGARET J. SCHOENINGER ET AL.

INTRODUCTION The Plio-Pleistocene of Africa exhibits the greatest radiation of hominid species that ever occurred throughout the evolution of our lineage. Experts disagree on the exact numbers of genera and species, but beyond the concerns about accuracy in taxonomy lies the basic question of why this radiation occurred. Early suggestions invoked differences in dietary strategies, where the large bodied species were vegetarian and the smaller species were omnivorous human ancestors. For this reason, much of the early research about early hominids focused on diet reconstruction and methods for achieving such reconstructions. More recently, Elizabeth Vrba argued that “marked environmental change . . . probably caused by a global reduction in temperature and associated alterations in rainfall . . . caused the evolutionary changes in hominids that are observed in eastern and southern Africa” (Vrba 1985:63, 1995; also Foley 1993). Support for this proposal comes from pollen studies, oceanic sediment composition, aeolian records, and deep ocean oxygen isotope curves dating to the late Pliocene and the Pleistocene Epochs (PlioPleistocene) between 3.0 and (Bonnefille 1995; deMenocal 1995; Shackleton 1995). Such changes would, necessarily, present early hominids with a new suite of diet options (Schoeninger 1995). Even so, high latitude deep ocean oxygen isotope curves do not always agree with the continental record (Burckle 1995) and sediment composition may not apply directly to continental interiors where the fossil changes took place. In order to propose reasonable diet options for early hominids, specific information on climate in the African continental interior is critical. For this reason, climate reconstruction is the focus of the present study which may otherwise seem out of place in a book on paleodiet.

Some fifty years ago, a classic theoretical paper (Urey 1947) proposed that the oxygen isotope ratio in various minerals depends on two variables: the local environmental temperature and the oxygen isotopic ratio

in the environmental water. The proposal was verified experimentally in carbonates (Epstein et al. 1951, 1953; McCrea 1950) where the oxygen isotope ratio is dependent on the temperature of formation and the ratio in source water. On continents, the source water is often surface water that comes from rain. In general, the ratios in surface waters (referred to as meteoric waters) and in precipitation vary predictably along latitudinal and temperature lines (Dansgaard 1954, 1964), but worldwide studies of the oxygen isotope ratio distribution in rainfall reveal that variability within the system is

quite high. Significant seasonal variations occur (published in the Technical Reports Series of the International Atomic Energy Agency in Vienna) resulting from multiple factors in addition to seasonal variation in temperature (Dansgaard 1964). As a result, the source water oxygen isotopic composition for continental carbonates is often unknown. Obviously, it is not possible to

calculate both water oxygen isotopic composition and temperature from the

TOOTH OXYGEN ISOTOPE RATIOS

119

single measurement of the oxygen isotope ratio in the carbonate of an animal’s shell or a soil nodule (Pearson and Coplen 1978). The variation in surface water appeared to obviate any attempt at climate reconstruction in continental interiors in the absence of a means of estimating source water oxygen isotopic composition. Suggestions that phosphatic minerals in mammals could be used, however, revived the interest in climate reconstruction in continental interiors. Aquatic, cold-blooded animals like fish have body temperatures and body water oxygen isotopic compositions that are directly dependent on the water in which they live. For these animals, a commonly used equation describes the relationships among temperature, water oxygen isotopic composition and phosphate oxygen isotopic composition as (Longinelli and Nuti 1973; verified by Kolodny et al. 1983, among others):

Oxygen isotope ratios are presented in permil

using the d notation

where

The standard is Standard Mean Ocean Water [SMOW]. The equation is based on the phosphate oxygen isotopic compositions of marine and lacustrine shells and fish bones, and the temperature and oxygen isotopic composition of the surrounding water. As with carbonates, there is a temperature dependence for the oxygen in phosphates (Tudge 1960; Longinelli 1965; Longinelli and Nuti 1968; Longinelli et al. 1976) but because mammals have constant body temperatures, their phosphate oxygen isotope ratios should indicate source water oxygen isotopic composition. Thus, if the source water for mammalian biogenic phosphates could be related to surface water and rain, the of bone and tooth enamel could be useful in paleoclimate studies. Further, given that mammals are relatively slow-growing, seasonal variation might either be smoothed out (averaged) in bone or it might be preserved as growth zonation in teeth. Experimental studies with 18O-enriched water showed that the in

mammalian bone phosphate and in blood varied in direct association with drinking water oxygen isotopic composition (Fig. 6.1, based on data in Luz and Kolodny 1985). Further, the in human and pig blood collected across

Europe correlated linearly with the oxygen isotope ratios of rain (Fig. 6.2, based on data in Longinelli 1984) although the two species show different slopes. Across a series of animals from the arctic to the equator there is a general correlation between the oxygen isotope composition of bone or tooth enamel phosphate and local meteoric (i.e., surface) water (Fig. 6.3). There is, however, a great deal of scatter about the line which appeared to eliminate the

120

MARGARET J. SCHOENINGER ET AL.

Figure 6.1. in drinking water compared with that in rat body water and in bone phosphate. Body water and bone phosphate values plot with slopes less than 1.0 reflecting the input from other oxygen sources (data from Luz and Kolodny 1985).

use of mammal bone or tooth phosphate values in paleoclimate studies with any level of accuracy. Clearly, particular species effects had to be sorted out. Two predictive models based on detailed assessments of each species’ oxygen mass balance have recently been proposed (Bryant and Froelich 1995; Kohn 1996). The former scales physiological parameters according to body mass for animals with body sizes It is based on a series of regression equations showing allometric relations between body size, on the one hand, and metabolic rate, water flux, evaporative water loss, oxygen intake (from ingested water, metabolism of food, and air), and oxygen outflow (water vapor, liquid water, and breath ), on the other. The second model emphasizes the variation around these regression equations. It considers diet, the amounts of fecal and urinary water loss, heat loss mechanisms (sweating vs. panting), and the ratio of daily water consumption to daily energy requirements (Nagy and Peterson 1988) as variables which must be evaluated independent of body size. Body water oxygen isotopic composition in mammals is a function of the oxygen mass balance of the specific animal. In general, it can be related to three major oxygen sources: air oxygen, drinking water, and food (both the

TOOTH OXYGEN ISOTOPE RATIOS

121

Figure 6.2. in rainwater compared with that in human and pig blood and bone phosphate. The values in tissue track that in rainwater, indicating that both food and drinking water have values that are dependent on rainwater values (data from Longinelli 1984).

free water in food and metabolic water), as well as to three major oxygen losses: breath liquid water (sweat, urine, fecal water), and water vapor. The general correlation between the oxygen isotope composition of bone or tooth enamel phosphate and surface water shown in Fig. 6.3 results from the dependence of both plant and surface water oxygen isotopic compositions on average rainwater oxygen isotopic composition, and the fact that most animals derive the bulk of their oxygen from food and water sources rather than air. The data plot with a slope less than 1 because ~20% of each animal’s oxygen is derived from air, which has a constant oxygen isotopic composition, independent of local water oxygen isotopic composition. There are, however, at least four aspects of variation in the biosphere which can affect this correlation and, as such, could account for the variation in the data. The two predictive models differ in the emphasis placed on each of these aspects. First, animal values are not expected to vary predictably with rainfall and surface water in cases where animals obtain the majority of their body water from plants and where plant values vary

independently of surface water values. For example, within Australia there is little continental variation in rainfall values and little surface water for

122

MARGARET J. SCHOENINGER ET AL.

Figure 6.3. in surface water (data from Gat 1980) compared with that in biogenic phosphates of various mammals from many world regions (data shown in squares are from conventional phosphate analyses: Longinelli 1984; Luz et al. 1984; Ayliffe and Chivas 1990; D’Angela and Longinelli 1990; Yoshida and Miyazaki 1991; Huertas et al. 1995). For comparison, plotted as filled circles, are values for tooth enamel analyzed by laser fluorination (six humans, one shark, and one wolf: Kohn et al. 1996).

drinking, yet, there is a wide range in animal

values (Fig. 6.4, data from Ayliffe and Chivas 1990). The animal values vary with average humidity reflecting the negative correlation between leaf values and humidity. The data suggest that a change in relative humidity of 10% affects phosphate by approximately In contrast, white-tailed deer from across North America (Fig. 6.4, data from Luz et al. 1990) show only a weak association between bone phosphate and humidity due, in large part, to the almost variation in surface and rainwater across the region (Luz et al. 1990). When the deer data were normalized to surface water values, the linear pattern was more obvious (Kohn 1996). Second, there could be differences in animal between species that are due to dietary differences. plants consistently have a higher in cellulose than coexisting plants because plants continue to transpire late in the day when it is dry (and their leaf water is high), whereas plant evapotranspiration slows or shuts down (Sternberg 1989; Yakir 1992). In arid settings, the oxygen isotopic composition difference can be as large as (Sternberg et al. 1984) whereas in cooler, moister settings it can be

TOOTH OXYGEN ISOTOPE RATIOS

Figure 6.4. Percent average humidity compared with

123

in bone phosphate. The strong asso-

ciation between the two variables across the species of Australian marsupials (data from Ayliffe and Chivas 1990) and North American white-tailed deer (data from Luz et al. 1990) indicates that their body water oxygen isotopic composition is strongly determined by the in their diet since plant varies in negative correlation with humidity.

or less (e.g., Epstein et al. 1977). When present, this difference could be transmitted through diet to different species. Grazers in regions dominated by grasses could differ from browsers. The of bone apatite structural carbonate among sympatric animals showed this trend with the exception of hippos (Koch et al. 1990).

A third aspect concerns heat loss mechanisms. Most animals lose heat either by sweating or by panting. Doubly labeled water studies in humans (Schoeller 1988; Wong et al. 1988) indicate that respiratory waters (mouth

and nasal) have values that can be to lower than sweat. Thus, animals that pant could be expected to increase their body water compared to animals that sweat. Fourth, body size is one of the major adaptations of any animal species. It affects all aspects of an animal’s life including locomotion, diet, and reproduction. As such, it should also affect an animal’s water balance due to such general considerations as the surface to volume ratio during heat production,

storage, and loss. Obviously, the average species body size is important in considering oxygen inputs and outputs.

124

MARGARET J. SCHOENINGER ET AL.

Data on modern mammals from the desertic region east of Lake Turkana in northern Kenya were used (Kohn et al. 1996, 1997) in testing one of the predictive models (Kohn 1996). The model underpredicted the values for some of the species although increasing transcutaneous water losses to levels measured under experimental conditions brought the model and observed data into compliance. It is significant that both models (Bryant and Froelich 1995; Kohn 1996) used largely temperate species for developing and for testing. In fact, temperate species have been the focus of most studies so far. In temperate regions, water is often plentiful and for much of the year the main concern is heat conservation rather than heat loss. In tropical deserts, the daily temperature range far exceeds the average annual range and surface water is restricted or unavailable for much of the year. Here we reconsider data from Lake Turkana in the context of adaptations required in deserts and add new data on cold-blooded organisms from the region.

ANIMAL SAMPLES, CHARACTERISTICS OF SAMPLING LOCALITY, AND ANALYTICAL METHODS Bone and tooth enamel from modern animals were collected in 1984 and 1993 from skeletons exposed on the surface in Sibiloi National Park, located on the east shore of Lake Turkana in northern Kenya. In addition to its interest as the site of numerous fossil hominid discoveries, the Turkana area provides an ideal controlled situation for the present study. The park is a circumscribed area surrounded by human pastoral groups and the nondomestic fauna remain to a great extent within its confines. Water sources are limited to the lake, ephemeral streams, a limited number of waterholes, and the plants eaten by the animals. The streams last on the order of days and in dry years do not flow at all. The non-domestic animals from which the bone and enamel were collected likely obtained most of their drinking water from the lake itself. Domestic animals entered the park in 1984 during a severe drought. Their drinking water sources may have varied widely.

The Water Source Lake Turkana is situated in a desert area where evaporation could produce surface and shallow water enriched in and thus have values which fall to the right of the meteoric water line (originally noted by Craig 1961). Aspects unique to the lake, however, suggest this may not be a problem in the Turkana region. Published data of Lake Turkana chemistry (Yuretich and Cerling 1983) indicate that the lake is well mixed. Although the surface water temperature has been measured at 29°C, thermoclines are ephemeral with the water usually remaining at 25°–26°C, suggesting rapid turnover. The source of the lake’s water is limited compared with other systems as approximately

TOOTH OXYGEN ISOTOPE RATIOS

125

80–90% of the total water in the lake comes from the Omo River to the north, which receives its water from a single source, i.e., rain on the Ethiopian plateau. The annual precipitation near the lake is less than 180mm per year. Studies on lakes, in general, indicate that “if the lake is large compared to the annual flow of water through it, . . . seasonal variations in precipitation will be smoothed out, and the lake will have a nearly constant isotopic composition” (Pearson and Coplen 1978:332). In combination, the mixing of Lake Turkana, the limited inflow sources, the lack of significant outflow, and the constancy of evaporative regime are the most likely explanations for the stability of the lake water oxygen isotope ratios measured in three different years in 1975, in 1977, and in 1980; data from Cerling et al. 1988) a range of only These values represent the average of a small seasonal range of 5 to (Cerling, cited in Johnson et al. 1991).

Animal Parameters Average body size (Sachs 1967), physiology and dietary pattern are known for most of the animals analyzed in this study. The relevant material is summarized below. Among the mammals, the dikdik (Rhynchotragus guentheri) is one of the smallest ruminant species (5 kg) and, as such, focuses on browse which is higher in protein compared to grass. Grant’s gazelle (Gazella granti) is an order of magnitude larger (50 kg) and feeds on a mixture of leaves and grass. Topi (Damaliscus corrigum) are larger (120kg) and are grazers. Oryx (Beisa oryx) and Burchell’s zebra (Equus burchelli) are both grazers where the former averages about 200kg and the latter is about 250kg. The species also vary in drinking habits, cooling mechanisms, and water excretion rates. The small dikdik, medium sized Grant’s gazelle, and large oryx are all drought tolerant and drink very little (Taylor 1969; Maloiy 1973; Maloiy et al. 1988; Hoppe et al. 1975; Spinage et al. 1980). Dikdik are territorial and inhabit areas that contain no surface water throughout the majority of the year. Grant’s gazelle will drink when water is available and some groups inhabit regions of the park that border the lake while others are far away from the lake. In contrast, oryx stay well away from the lake and thus drink very little, if at all. Burchell’s zebra drink water everyday. Topi commonly live in mixed herds with zebra where they have access to water on a daily basis, although their water dependency is not known. In terms of oxygen output, dikdik pant to cool themselves and their moist, flexible noses also provide relatively large surface areas for heat dissipation. Grant’s gazelle may pant under some conditions but also conserve body water by not panting and permitting body temperature to rise under conditions similar to those east of Lake Turkana (Taylor 1970a). Both species eliminate little urinary water. Burchell’s zebra and oryx sweat to cool themselves and zebra urinate freely. Oryx apparently stop sweating around 40°C (Taylor 1970b) and both Grant’s gazelle and oryx can tolerate a rise in body temperature up to 46.4°C for several hours with no negative effects

126

MARGARET J. SCHOENINGER ET AL.

(Taylor 1970b). Burchell’s zebra will move to water whereas the oryx survives in the absence of surface water. In addition to these non-domestic species, samples were collected of domestic goats during 1984. Goats drink freely and are moved daily to water by their herders. The goats that died east of Lake Turkana formed their tooth enamel in another area that may have had a value quite different from that of the lake. Samples were also taken of crocodile. Crocodiles sun themselves along the lakeshore, opening their jaws to cool. Other than this behavioral variable, the crocodiles were expected to have values that reflected the Lake Turkana waters at Lake Turkana temperatures.

Skeleton weathering ranged up to stage 3 (0–7 year exposure; Behrensmeyer 1978) which should not have affected tooth isotope compositions (Koch et al. 1990). Analytical Problems and Solutions Besides the problems associated with biological variation, the difficulty of sample preparation and analysis has hampered the application of phosphate oxygen stable isotope ratios to paleoanthropology. Biogenic phosphate is a

hydroxyapatite which includes small amounts of and that substitute for and OH groups. Thus, there are three important oxygen reservoirs: phosphate, structural carbonate, and hydroxyl oxygen. Previous analytical approaches have focused on extracting oxygen preferentially

from either the phosphate groups or the structural carbonate. Phosphate oxygen is widely viewed as being the most resistant to diagenesis, but primary structural carbonate in both the phosphate and hydroxyl positions can also be preserved (Lee Thorp and van der Merwe 1987). The original method of phosphate preparation involved extracting the phosphate and reprecipitating it as a bismuth phosphate (Tudge 1960). Alternatively, it is precipitated as a silver phosphate (Wright and Hoering 1989) which involves fewer steps and, more importantly, silver phosphate is not hygroscopic (as is bismuth phosphate) which minimizes the potential for contamination by atmospheric water. A greater hindrance for paleoclimate studies, however, is that the traditional method required reduction in an all metal vacuum line at high temperature (in externally-heated nickel reaction vessels) with bromine pentafluoride a highly reactive gas (Clayton and Mayeda 1963). Handling this material in anything other than a dedicated geochemistry laboratory has proven extremely difficult and dangerous (Chivas 1984). Two alternate methods have recently been developed and both are used in the present study. A laser probe analytical method provided the majority of the oxygen isotope data (see Kohn et al. 1996 for details on testing and developing the method). Laser probes were originally developed for the stable isotope analysis of silicates, oxides, and sulfides in crystalline rocks (Crowe

TOOTH OXYGEN ISOTOPE RATIOS

127

et al. 1990; Sharp 1990) and thus far, the method does not appear applicable to bone or dentine samples which contain high amounts of organic material. Enamel, on the other hand, gave highly precise results. Powdered enamel (2 mg) was pretreated overnight in NaOCl to oxidize interstitial organic material, ground, and loaded into a 72-pit nickel sample block along with other samples and standards. Samples and standards were pretreated with at room temperature overnight to eliminate adsorbed water and organic residues. Sample reaction and analysis the next day was accomplished by individually heating each sample with a focused carbon dioxide laser in the presence of to liberate oxygen cryogenic distillation, conversion of to using a heated graphite rod, and analysis by mass spectrometry (Valley et al. 1995). Analytical reproducibility was (1s) and yields for phosphate standards were typically about 80%. But standard oxygen isotopic compositions are systematically 1.6 to lower than accepted values and a correction is applied to raw laser data (Kohn et al. 1996). With the correction, the laser data are consistent with data produced by conventional methods (Fig. 6.3). The laser probe technique described here differs from one that melts the enamel in vacuo with a laser to produce a small amount of gas (Sharp and Cerling 1995; Cerling and Sharp 1996). In that method, the C is surely derived from structural carbonate, but the oxygen reservoir tapped by this method is uncertain. The second analytical method uses a combustion system (O’Neil et al. 1994) in place of reaction with This method was used for the crocodiles because they were represented by very thin caps of enamel. The enamel was powdered and sieved (20 mg), pretreated in NaOCl to oxidize organic material and then precipitated as silver phosphate. Approximately 10–20 mg of silver phosphate were mixed with powdered graphite in quartz tubes, evacuated and sealed. Combustion at 1,200°C was followed by rapid cooling in water which prevents isotopic fractionation between the produced and the residual solid in the tube. Analyses of separate aliquots from the same sample typically showed precisions of to with 2 to 4 repetitive analyses even though yields are on the order of 25%. Tooth enamel precipitated as silver phosphate and reacted via this method gave values similar to corrected laser data from aliquots of the same tooth enamel (Table 6.1). Thus, the data from the crocodile samples are considered directly comparable with the data from the mammal tooth enamel.

RESULTS AND DISCUSSION Within several of the species there is significant interindividual variation (Table 6.2). This is due, in large part, to a lack of control for tooth type. Several recent studies (Bryant et al. 1996; Fricke and O’Neil 1996; Stuart-Williams and

MARGARET J. SCHOENINGER ET AL.

128

Schwarcz 1997; Kohn et al. 1997) report zoned variation within certain teeth which correlates with seasonal variation in rainfall Intratooth variation reached almost in some teeth from east Turkana (Kohn et al. 1997). While this isotopic variation reflects important climate information regarding seasonal variation in rainfall, this topic is not pursued here and the interested reader is directed to the literature cited above.

The species from Lake Turkana show average species values (Table 6.2) that range from a low of (goat, ) and (crocodile, ) to a high of nearly (Grant’s gazelle, ). When the results are plotted against body size (Fig. 6.5), the 5 kg dikdik shows values that overlap the 250 kg zebra. If the dikdik are eliminated, there is a negative association between body size and tooth enamel values as expected, but the range of average values is

rather than

as predicted by the body-size based model

(Bryant and Froelich 1995). In part, this is due to the ways in which desertadapted mammals obtain body water. Many of them violate one of the assumptions of the body-size based model, i.e., that 80% of ingested water comes from

drinking (Bryant and Froelich 1995). Among the species at east Turkana, dikdik, gazelle, and oryx obtain far less, if any, of their ingested water through drinking. A majority of their body water comes from free water or metabolic water obtained from the plants that they consume. Thus, species diet is a critical variable. When the data are plotted against the which provides dietary information (Schoeninger 1989), there appears a general tendency for an increase in with increased dependence on grass among non-

TOOTH OXYGEN ISOTOPE RATIOS

129

browsing species (Fig. 6.6). In contrast, the browsing dikdik show no apparent correlation between the and enamel values. Dikdik and oryx have similar average values but very different bone collagen values, and

Grant’s gazelle (a browser-grazer) has the least negative

values. This

130

MARGARET J. SCHOENINGER ET AL.

Figure 6.5. of tooth enamel phosphate versus body size (log scale) for Kenyan fauna analyzed in this study. With the exception of the dikdik, there is a general association between the two variables. In contrast to the body-size model (Bryant and Froelich 1995) which predicts a range of values close to however, the measured range in values for species averages is

This and the anomalous values for the dikdik reflect physiological and behavioral adaptations by these desert adapted species.

suggests that a difference in values between and plants, if present, is not the primary variable affecting animal values. For some species, the data fit the Kohn model (Kohn 1996). Zebra have relatively low values and they drink water (the lowest of all oxygen sources at Lake Turkana), sweat to cool (sweat has relatively high

), and

urinate more (urine has relatively high ) than many arid-adapted animals. In other words, the input oxygen is low in and the output oxygen is high in resulting in zebra body water that is, presumably, low in Superficially, topi should be similar to zebra because they are grazers that spend most

of their time near the lake. Yet, the values for topi are more positive that those for zebra. More information on topi physiology is required in order to be certain of the significance of this offset. In contrast to expectations, dikdik have very low values even though they obtain the majority of their water from plants. In such an area, even plants should have higher values than that of the lake. Dikdik pant to cool (low ) which should also result in elevated values. But dikdik also decrease activity during the heat of the day, remaining

TOOTH OXYGEN ISOTOPE RATIOS

131

Figure 6.6. of tooth enamel phosphate versus of bone collagen for Kenyan fauna analyzed in this study. There is a general association between consumption of grasses and animal values among nonbrowsing species. Within the sample of dikdik there is no clear correlation between the two variables and dikdik, zebra, and oryx have similar average values with extremely different bone collagen

values. Night feeding by the oryx, sedentary daytime behav-

ior by the dikdik, and obligate drinking by the zebra account for the similarity in values in these species. With the exception of the zebra’s obligate drinking, these behavioral variables could not be predicted in fossil species.

in shade, thereby minimizing the need to use body water for cooling purposes. They urinate relatively little, which also lowers body water turnover. The wide intraspecies range of variation suggests that there is variation in the input oxygen. Since the majority of the body water in these animals comes from the plants they eat, this, in turn suggests variation in plant water Some desert plants apparently rehydrate their leaves during the night when temperatures fall markedly and relative humidity rises (Taylor 1969). Air water vapor, taken up by these leaves, has lower than surface water as described by the equation (Bottinga 1969):

where T refers to the average air temperature (in °K). At night, the water vapor in air could have values as low as assuming a night temperature of

29°C. Normally, leaf water

shows far greater dependence on the

of

132

MARGARET J. SCHOENINGER ET AL.

surface water than of air water vapor and follows the relationship (Craig and Gordon 1965):

But, in the case of desert plant leaves hydrating at night, there is no water in soil and, presumably, surface water has less input than indicated by the equation. If surface water is removed from the equation, in leaf water is closer

to assuming a night humidity of 75%. This represents an extreme and the point of emphasis is that leaf water could have values at night that are much lower than during the day. Therefore, if animals feed at night, they can be expected to have lower values than predicted based only on diurnal feeding. Such behavior may account for the

values of the oryx which are

lower than zebra. Oryx feed at night (Walther 1978) as well as during the early morning and late evening and it is estimated that they can meet basic water requirements by feeding on hydrated plant leaves (Taylor 1969). This feeding pattern would result in low and low An additional consideration is that oryx do not maintain constant body temperature (Taylor 1970a, 1970b). This physiological adaptation contradicts assumptions of both models because under such conditions water is not evaporated through the skin, lost by sweat, or lost through mouth or nose at the rates used in the models. Total oxygen output is reduced by restricting water loss. Altering body temperature, alone, is not expected to affect the values of the animal’s phosphate markedly. For example:

where the water refers to body water and the temperature refers to body temperature (Longinelli and Nuti 1973). Assuming an average body temperature of 39°C (6 hours at 44°C and 18 hours at 37°C) and a body water value of over surface water, Kohn 1996) predicts a of compared with for 37°C. Oryx will cool down below 37°C during the night (Taylor 1970b) and it is possible that the average body temperature is actually closer to the normal 37°C. Each 1°C increase in body temperature changes the predicted by approximately Grant’s gazelle have the highest values. Superficially this appears due to their particular form of drought adaptation where they obtain most of their body water from plants and they urinate relatively little. But, unlike the oryx, which feeds totally on plants, Grant’s gazelle depend on a mix of and plants, as indicated by their bone collagen values. Unlike the dikdik, they are active throughout the day, which suggests that the majority of their body water comes from daytime plant water. It is likely that such plant water has higher values than that taken by night feeders. If correct, it

TOOTH OXYGEN ISOTOPE RATIOS

133

suggests that Grant’s gazelle values may monitor humidity. Obviously, further study is required. Some of the species data from Turkana can be compared to those from other areas. Several studies have reported data on non-domestic individuals from various species of Equus (see Fig. 6.7, data from Bryant et al. 1994, 1996; Huertas et al. 1995; Sanchez Chillon et al. 1994; and this study). A recent data compilation (Huertas et al. 1995) resulted in the following regression equation:

This equation predicts a value of for the zebra at Turkana assuming an average value of for Lake Turkana water. This predicted value is less than the actual value of Given the variation in methods of sample preparation and analysis, variation between bone and tooth enamel (Stuart-Williams and Schwarcz 1997), and uncertainty in surface water oxygen isotopic composition, these values are extraordinarily close. Alternatively, if the equation is solved for using the actual value of the Turkana zebra,

Figure 6.7. of tooth enamel or bone phosphate in various species of Equus plotted against of surface water. All individuals analyzed were non-domestic (Bryant et al. 1994, 1996; Huertas et al. 1995; Sanchez Chillon et al. 1994; this study). The east Turkana zebra plot as expected and the regression equation (Huertas et al. 1995) predicts the Turkana zebra value within

Given the scatter in the data this supports the conclusion that equid species oxygen isotope composition provides an excellent monitor for surface water oxygen isotopic composition.

134

MARGARET J. SCHOENINGER ET AL.

the predicted value for Lake Turkana is The values reported range from to (Cerling, cited in Johnson et al. 1991). Using the calculated value for lake water and the measured value for crocodile in the equation (Longinelli and Nuti 1973):

an average annual temperature of 30.1°C is predicted. The average annual temperature at east Turkana is 29.3°C. Using the measured average value of for in Lake Turkana and the measured value for the crocodiles predicts an average temperature of 24.5°C. The average water temperature of Lake Turkana is 25–26°C. This exercise illustrates that the critical factor is not the actual body temperature of mammals, which with their minimal variation affect the outcome minimally, rather, it is the prediction of water values. In contrast, a similar comparison made with data from two domestic goats collected in the Turkana region does not produce as close a fit (see Fig. 6.8, data

Figure 6.8. of tooth enamel or bone phosphate in various species of Capra plotted against of surface water (Huertas et al. 1995; this study). Individuals analyzed were both domestic and non-domestic. The east Turkana goats have lower values than expected based on the

Lake Turkana water oxygen isotopic composition. The goats collected at east Turkana did not originate there but were brought into Sibiloi National Park in 1984 by their owners in order to obtain water from the lake. Their tooth enamel

reflects water sources other than Lake Turkana.

TOOTH OXYGEN ISOTOPE RATIOS

135

from Huertas et al. 1995). The regression equation produced from a least squares fit of data from of bone phosphate of various species of Capra is:

This equation predicts a value of for goats in the Turkana region whereas the average of two individual goats is actually When the actual value is used to calculate the value of the result is The most likely explanation lies in the true origin of the goats recovered at Turkana. These are domestic animals that were brought into the park during a severe drought in 1984. As such, their body water at the time of tooth formation probably bore no relation to Lake Turkana water.

SIGNIFICANCE These data show that both models identify important variables that affect

and

in mammals. Both serve to identify the dikdik as

an outlier which may be explained by their sedentary daytime pattern. On the other hand, the body-size model (Bryant and Froelich 1995), which may reliably predict animal in temperate, well-watered regions, does not predict in these desert-adapted species. The second model (Kohn 1996), by emphasizing animal physiology independent of body size, serves to identify species with different sensitivities to climatic parameters. This, in conjunction with considerations of behavior, indicate that certain species are probably not useful for monitoring paleotemperature because their is not tied, in a consistent way, to The oryx, for example, can conserve body water and tolerate a wide range of temperature and water stress by nighttime feeding, increased body temperatures, and termination of sweating or panting. When these specific physiological and behavioral attributes are known (as in living species) the coincides with model predictions. Such information, however, cannot be determined for fossil species.

On the other hand, this study also identifies certain species (e.g., Burchell’s zebra) that should be useful for monitoring paleotemperature and others (e.g., Grant’s gazelle) that may be useful for monitoring humidity. The data in the present study and worldwide data on non-domestic equids, considered in conjunction with the model (Kohn 1996), suggest that a critical variable is obligate drinking. Zebra species monitor in arid regions like the Turkana basin because their is tied in a consistent manner to Obligate drinking can be assumed for all species of Equus. Even the Grevy’s zebra (Equus grevii), which inhabits the driest regions of all

equid species, must drink even though it is not required to every day. At based studies of temperature to single species

present, limiting the

136

MARGARET J. SCHOENINGER ET AL.

where drinking patterns are known (e.g., Fricke et al. 1995; Stuart-Williams

and Schwarcz 1997) is the best strategy. Additionally, the gazelle, as a waterindependent species that is active and that feeds diurnally, may prove useful in monitoring humidity stability or change through time when normalized with zebra This must be demonstrated, however, through sampling of gazelle species in regions with different water and average humidity. The Lake Turkana area in East Africa is particularly appropriate for future paleoclimate research. In the Turkana basin, the water sources to the lake and to the river that at times replaced it during the Plio-Pleistocene (Feibel 1988), are relatively constant. Thus, a change in lake water oxygen isotopic composition indicated by a change in zebra values should reflect a change in rainwater oxygen isotopic composition which, in turn, reflects a change in average temperature. In the present study, the zebra data predict a value of for the lake water. Using this value with values

of the crocodile tooth enamel predicts a temperature of 30.1°C which is slightly higher than the actual average annual temperature of 29.3°C. The intraspecies variation is too large, at present, to predict temperature accurately to better than 1°C but with carefully controlled sampling of single teeth and of single areas within a tooth, however, this variation should be decreased considerably with the expectation that accuracy will improve. It is clear that more samples of target species (zebra, gazelle, and crocodile) from the region must be analyzed before this is certain. The goal, i.e., a quantitative determination of past climate from fossil tooth enamel, is worth the effort because such determination should allow assessment of how climate change affected the radiation of early hominid species (Vrba et al. 1995). ACKNOWLEDGMENTS We thank the editors for including our manuscript even though we were unable to attend the conference. We thank M. Spicuzza for maintaining the laser extraction lines, J. O’Neil, H. Fricke, and R. Blake for help with the silver phosphate precipitation and analysis, J. Farquhar, T. Chacko, and Y. Kolodny for providing standards, and J. Banfield, K. Barovich, L. Baumgartner, S. Baumgartner, G. Dentine, K. Edwards, M. Jackson, D. Kantor, M. Kelley, K. Murphy, D. Valley, M. Valley, and M. Wilhite for donating or helping obtain teeth for analysis. H. Bunn and R. Gilliam verified sample ID’s. Comments by three anonymous reviewers greatly improved the manuscript. Particular thanks go to the National Museums of Kenya and Harvard Summer Field School, Dr. H. Merrick, Dr. Meave Leakey, and Aila for their help with the East African collection and research. Funded by NSF grants EAR 9316349 (MJK), BNS 8509753 (MJS), and EAR 9304372 (JWV), DOE grant FG02-93ER14389 (JWV), and Wenner Gren Foundation for Anthropological Research grant 5615 (MJS).

TOOTH OXYGEN ISOTOPE RATIOS

137

REFERENCES Ayliffe, L.K. and Chivas, A.R. 1990 Oxygen isotope composition of the bone phosphate of Australian kangaroos: potential as a palaeoenvironmental recorder. Geochimica et Cosmochimica Acta 54: 2603–2609. Behrensmeyer, A.K. 1978 Taphonomic and ecologic information from bone weathering. Paleobiology 4: 150–162. Bonnefille, R. 1995 A reassessment of the Plio-Pleistocene pollen record of East Africa. In Vrba, E.S., Denton, G.H., Partridge, T.C. and Burckle, L.H., eds., Paleoclimate and Evolution. New

Haven, CT, Yale University Press: 299–310. Bottinga, Y. 1969 Calculated fractionation factors for carbon and hydrogen isotope exchange in the system calcite-carbon dioxide-graphite-methane-hydrogen-water vapour. Geochimica et

Cosmochimica Acta 33: 49–64. Bryant, J.D. and Froelich, P.N. 1995 A model of oxygen isotope fractionation in body water of

large mammals. Geochimica et Cosmochimica Acta 59: 4523–4537. Bryant, J.D., Froelich, P.N., Showers, W.J. and Genna, B.J. 1996 Biologic and climatic signals in the oxygen isotope composition of Eocene-Oligocene equid enamel phosphate. Palaeogeography, Palaeoclimatology, Palaeoecology 126: 75–89. Bryant, J.D., Luz, B. and Froelich, P.N. 1994 Oxygen isotopic composition of fossil horse tooth

phosphate as a record of continental paleoclimate. Palaeogeography, Palaeoclimatology, Palaeoecology 107: 210–212. Burckle, L.H. 1995 Current issues in Pliocene paleoclimatology. In Vrba, E.S., Denton, G.H., Partridge, T.C. and Burckle, L.H., eds., Paleoclimate and Evolution. New Haven, CT, Yale University Press: 3–7. Cerling, T. and Sharp, Z. 1996 Stable carbon and oxygen isotope analysis of fossil tooth enamel using laser ablation. Palaeogeography, Palaeoclimatology, Palaeoecology 126: 173–186. Cerling, T.E., Bowman, J.R. and O’Neil, J.R. 1988 An isotopic study of a fluvial-lacustrine

sequence: the Plio-Pleistocene Koobi-Fora sequence, East Africa. Palaeogeography, Palaeoclimatology, Palaeoecology 63: 335–356. Chivas, A.R. 1984 Some comments and further recommendations on a report by O’Donnell

and Waugh concerning an explosion involving bromine pentafluoride at the Geology Department, University of Queensland. Adelaide: The Australian National University Memorandum. Clayton, R.N. and Mayeda, T.K. 1963 The use of bromine pentafluoride in the extraction of oxygen from oxides and silicates for isotopic analysis. Geochimica et Cosmochimica Acta 27: 43–52. Craig, H. 1961 Isotopic variations in meteoric waters. Science 133: 1702–1703. Craig, H. and Gordon, L.I. 1965 Deuterium and oxygen-18 variation in ocean and the marine

atmosphere. In Tongiorgi, E., eds., Proceedings of a Conference on Stable Isotopes in Oceanographic Studies and Poleotemperatures. Spoleto, Italy, Consiglio Nazionale Delle Richerche: 9–130. Crowe, D.E., Valley, J.W. and Baker, K.L. 1990 Micro-analysis of sulfur-isotope ratios and zonation by laser microprobe. Geochimica et Cosmochimica Acta 54: 2075–2092. D’Angela, D. and Longinelli, A. 1990 Oxygen isotopes in living mammal’s bone phosphate: further results. Chemical Geology 86: 75–82. Dansgaard, W. 1954 The O18-abundance in fresh water. Geochimica et Cosmochimica Acta 6: 241–260. ___1964 Stable isotopes in precipitation. Tellus 16: 466–468.

deMenocal, P.B. 1995 Plio-Pleistocene African climate. Science 270: 53–59. Epstein, S., Buchsbaum, R., Lowenstam, H. and Urey, H.C. 1951 Carbonate-water isotopic tem-

perature scale. Bulletin of the Geological Society of America 62: 417–425. Epstein, S., Buchsbaum, R., Lowenstam, H.A. and Urey, H.C. 1953 Revised Carbonate-water isotopic temperature scale. Bulletin of the Geological Society of America 64: 1315–1325.

138

MARGARET J. SCHOENINGER ET AL.

Epstein, S., Thompson, P. and Yapp, C.J. 1977 Oxygen and hydrogen isotopic ratios in plant cellulose. Science 198: 1209–1215. Feibel, C. 1988 Paleoenvironments of the Koobi Fora Formation, Turkana Basin, Northern Kenya. Ph.D. dissertation, Salt Lake City: University of Utah. Foley, R. 1993 The influence of seasonality on hominid evolution. In Ulejaszek, S.J. and Strickland, S.S., eds., Seasonality and Human Ecology. Cambridge University Press: 17–37. Fricke, H.C. and O’Neil, J.R. 1996 Inter- and intra-tooth variation in the oxygen isotope composition of mammalian tooth enamel: some implications for paleoclimatological and paleobiological research. Palaeogeography, Palaeoclimatology, Palaeoecology 126: 91–99. Fricke, H.C., O’Neil, J.R. and Lynnerup, N. 1995 Oxygen isotope composition of human tooth enamel from medieval Greenland: linking climate and society. Geology 23: 869–872. Gat, J.R. 1980 The isotopes of hydrogen and oxygen in precipitation. In Fritz, P. and Fontes, J., eds., Handbook of Environmental Isotope Geochemistry. Amsterdam, Elsevier: 21–47. Hoppe, P.P., Johansen, K., Maloiy, G.M.O. and Musewe, V. 1975 Thermal panting reduces oxygen uptake in the dikdik, Rhynchotragus kirki. Acta Physiologica Scandinavia 95: 9A–10A. Huertas, A.D., Iacumin, P., Stenni, B., Chillon, B.S. and Longinelli, A. 1995 Oxygen isotope variations of phosphate in mammalian bone and tooth enamel. Geochimica et Cosmochimica Acta 59: 4299–4305. Johnson, T.C., Halfman, J.D. and Showers, W.J. 1991 Paleoclimate of the past 4000 years at Lake Turkana, Kenya, based on the isotopic composition of authigenic calcite. Palaeogeography,

Palaeoclimatology, Palaeoecology 85: 189–198. Koch, PL., Behrensmeyer, A.K., Tuross, N. and Fogel, M.L. 1990 Isotopic fidelity during bone

weathering and burial. Annual Report of the Director, Geophysical Laboratory, 1989–1990: Washington D.C.: Geophysical Laboratory, Carnegie Institution of Washington: 105–110. Kohn, M.J. 1996 A predictive model for animal explaining old studies and designing new ones. Geochimica et Cosmochimica Acta 60: 4811–4829. Kohn, M.J., Schoeninger, M.J. and Valley, J.W. 1996 Herbivore tooth oxygen isotope compositions:

effects of diet and physiology. Geochimica et Cosmochimica Acta 60: 3889–3896. ___1997 Variability in herbivore tooth oxygen isotope compositions: reflections of seasonality or developmental physiology? Isotope Geoscience (in press). Kolodny, Y., Luz, B. and Navon, O. 1983 Oxygen isotope variations in phosphate of biogenic apatites, I. Fish bone apatite—rechecking the rules of the game. Earth and Planetary Science Letters 64: 398–404. Lee-Thorp, J. and van der Merwe, N.J. 1987 Carbon isotope analysis of fossil bone apatite. South

African Journal of Science 83: 712–715. Longinelli, A. 1965 Oxygen isotopic composition of orthophosphate from shells of living marine organisms. Nature 207: 716–719. ___1984 Oxygen isotopes in mammal bone phosphate: A new tool for paleohydrological and paleoclimatological research? Geochimica et Cosmochimica Acta 48: 385–390. Longinelli, A., Bartelloni, M. and Cortecci, G. 1976 The isotopic cycle of oceanic phosphate, I. Earth and Planetary Science Letters 32: 389–392. Longinelli, A. and Nuti, S. 1968 Oxygen isotopic composition of phosphorites from marine formations. Earth and Planetary Science Letters 5: 13–16. ___1973 Revised phosphate-water isotopic temperature scale. Earth and Planetary Science Letters 19: 373–376. Luz, B., Cormie, A.B. and Schwarcz, H.P. 1990 Oxygen isotope variations in phosphate of deer bones. Geochimica et Cosmochimica Acta 54: 1723–1728. Luz, B. and Kolodny, Y. 1985 Oxygen isotope variations in phosphate of biogenic apatites, IV. Mammal teeth and bones. Earth and Planetary Science Letters 75: 29–36. Luz, B., Kolodny, Y. and Horowitz, M. 1984 Fractionation of oxygen isotopes between mammalian bone-phosphate and environmental drinking water. Geochimica et Cosmochimica Acta 48: 1689–1693.

TOOTH OXYGEN ISOTOPE RATIOS

139

Maloiy, G.M.O. 1973 The water metabolism of a small East African antelope: the dik-dik.

Proceedings of the Royal Society, London B 184: 167–178. Maloiy, G.M.O., Rugangazi, B.M. and Clemens, E.T. 1988 Physiology of the dik-dik antelope. Comparative Biochemistry and Physiology 91A: 1–8. McCrea, J.M. 1950 On the isotopic chemistry of carbonates and a paleotemperature scale. Journal of Chemical Physics 18: 849–857. Nagy, K.A. and Peterson, C.C. 1988 Scaling of water flux rate in animals. University of California

Publications in Zoology 120: 1–172. O’Neil, J.R., Roe, L.J., Reinhard, E. and Blake, R.E. 1994 A rapid and precise method of oxygen isotope analysis of biogenic phosphate. Israel Journal of Earth Sciences 42: 203–212. Pearson, F.J. and Coplen, T.B. 1978 Stable isotope studies of lakes. In Lerman, A., ed., LakesChemistry, Geology, Physics. New York, Springer-Verlag: 325–339. Sachs, R. 1967 Liveweights and body measurements of Serengeti game animals. East African Wildlife Journal 5: 24–36.

Sanchez Chillon, B.S., Alberdi, M.T., Leone, G., Bonadonna, F.P., Stenni, B. and Longinelli, A. 1994 Oxygen isotopic composition of fossil equid tooth and bone phosphate: an archive of difficult interpretation. Palaeogeography, Palaeoclimatology, Palaeoecology 107: 317–328. Schoeller, D.A. 1988 Measurement of energy expenditure in free-living humans by using doubly labeled water. Journal of Nutrition 118: 1278–1289. Schoeninger, M.J. 1989 Prehistoric human diet. In Price, T.D., ed., Chemistry of Prehistoric Human Bone. Cambridge, Cambridge University Press: 38–67. ___1995 Stable isotope studies in human evolution. Evolutionary Anthropology 4: 83–98.

Shackleton, N.J. 1995 New data on the evolution of Pliocene climatic variability. In Vrba, E.S., Denton, G.H., Partridge, T.C. and Burckle, L.H., eds., Paleoclimate and Evolution. New Haven, CT, Yale University Press: 242–248. Sharp, Z.D. 1990 A laser-based microanalytical method for the in situ determination of oxygen

isotope ratios of silicates and oxides. Geochimica et Cosmochimica Acta 54: 1353–1357. Sharp, Z.D. and Cerling, T.E. 1995 A laser GC-IRMS method for in-situ carbon and oxygen isotope analyses of carbonates and phosphates. Geological Society of America, Abstracts Program 27: A255. Spinage, C.A., Ryan, C. and Shedd, M. 1980 Food selection by the Grant’s gazelle. African. Journal of Ecology 18: 19–25. Sternberg, L.d.S.L. 1989 Oxygen and hydrogen isotope ratios in plant cellulose: mechanisms and applications. In Rundel, P.W., Ehleringer, J.R. and Nagy, K.A., eds., Stable Isotopes in Ecological Research. New York, Springer-Verlag: 124–141. Sternberg, L.O., DeNiro, M.J. and Johnson, H.B. 1984 Isotope ratios of cellulose from plants having different photosynthetic pathways. Plant Physiology 74: 557–561. Stuart-Williams, H.L.Q. and Schwarcz, H.P. 1997 Oxygen isotopic determination of climatic variation using phosphate from beaver bone, tooth enamel, and dentine. Geochimica et

Cosmochimica Acta 61: 2539–2550. Taylor, C.R. 1969 The eland and the oryx. Scientific American 220: 89–95. __1970a Dehydration and heat: effects on temperature regulation of East African ungulates. American Journal of Physiology 219: 1136–1139. __1970b Strategies of temperature regulation: effect on evaporation in East African ungulates.

American Journal of Physiology 219: 1131–1135. Tudge, A.P. 1960 A method of analysis of oxygen isotopes in orthophosphate-its use in the measurement of paleotemperatures. Geochimica et Cosmochimica Acta 18: 81–93. Urey, H.C. 1947 The thermodynamic properties of isotopic substances. Journal of the Chemical Society 1947: 562. Valley, J.W., Kitchen, N., Kohn, M.J., Niendorf, C.R. and Spicuzza, M.J. 1995 UWG-2, a garnet standard for oxygen isotope ratios: strategies for high precision and accuracy with laser heating. Geochimica et Cosmochimica Acta 59: 5223–5231.

140

MARGARET J. SCHOENINGER ET AL.

Vrba, E.S. 1985 Ecological and adaptive changes associated with early hominid evolution. In

Delson, E., eds., Ancestors: The Hard Evidence. New York, Alan R. Liss: 63–71. ___1995 The fossil record of African antelopes (Mammalia, Bovidae) in relation to human evo-

lution. In Vrba, E.S., Denton, G.H., Partridge, T.C. and Burckle, L.H., eds., Paleoclimate and Evolution. New Haven, CT, Yale University Press: 385–424.

Vrba, E.S., Denton, G.H., Partridge, T.C. and Burckle, L.H. 1995 Paleoclimate and Evolution. New Haven, Yale University Press.

Walther, F.R. 1978 Behavioral observations on oryx antelope (Oryx beisa) invading Serengeti National Park, Tanzania. Journal of Mammalogy 59: 243–260. Wong, W.W., Cochran, W.J., Klish, W.J., Smith, E.O.B., Lee, L.S. and Klein, P.D. 1988 In vivo isotope-fractionation factors and the measurement of deuterium- and oxygen-18-dilution spaces from plasma, urine, saliva, respiratory water vapor, and carbon dioxide. American

Journal of Clinical Nutrition 47: 1–6. Wright, E.K. and Hoering, T.C. 1989 Separation and purification of phosphates for oxygen isotope analysis. Annual Report of the Director, Geophysical Laboratory, 1989–1990: Washington D.C.: Geophysical Laboratory, Carnegie Institution of Washington 137–141. Yakir, D. 1992 Variations in the natural abundance of oxygen-18 and deuterium in plant carbohydrates. Plant, Cell, and Environment 15: 1005–1020. Yoshida, N. and Miyazaki, N. 1991 Oxygen isotope correlation of cetacean bone phosphate with environmental water. Journal of Geophysical Research 96: 815–820. Yuretich, R.F. and Cerling, T.E. 1983 Hydrogeochemistry of Lake Turkana, Kenya: mass balance and mineral reactions in an alkaline lake. Geochimica et Cosmochimica Acta 47: 1099–1109.

Chapter

7

Quantifying Histological and Chemical Preservation in Archaeological Bone SUSAN PFEIFFER

AND

TAMARA L. VARNEY

ABSTRACT Researchers in archaeological bone chemistry have suggested that the histological preservation of bone may mirror the preservation of its organic framework, comprised primarily of collagen. Such a relationship may provide a criterion for sample selection prior to chemical analysis. This proposed relationship was explored using a study sample comprised of nine human femoral cross-sections from a sandy lake shore, site of the Snake Hill military cemetery (A.D. 1814) at Fort Erie, Ontario, Canada. Eight of the sections showed three distinct layers of cortical preservation that ranged from good to extremely poor. One section was consistently well preserved throughout the cortex. The collagen component of the bone was extracted using a modified Longin (1971) method from each of the three layers of preservation and was evaluated with regard to collagen yield, carbon and nitrogen concentration, carbon to nitrogen atomic ratio, amino acid profile, and amino acid yield. Previous work had shown that mineral components of the bone were relatively unmodified across the entire cross-sections regardless of histological preservation. In the results of the current analysis, histological preservation fails to show a reliable correlation with collagen integrity. There are no correlations among the various evaluation criteria. Amino acid profiles show a non-homogeneous composition for the extracted “collagen”. A substantial amount of non-proteinaceous material of unknown composition is also present. Histological preservation is not a Biogeochemical Approaches to Paleodietary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

141

142

SUSAN PFEIFFER AND TAMARA L. VARNEY

reliable indicator of collagen integrity, and should not be used as a criterion

of sample selection for chemical analysis.

INTRODUCTION

Both histological and stable isotope analyses of archaeological bone are dependent upon the state of tissue preservation. In order for either type of analysis to yield reliable results, the biological characteristics of the bone tissue must be preserved. The histologist needs to see the structures of cortical bone, unobscured by diagenetic destruction or alteration. The isotope chemist needs to measure elements that are unmodified by postmortem additions or deletions. Accurate results are contingent, in part, on choosing good samples. However, the present criteria for assessing bone preservation are not completely satisfactory; a single reliable criterion has been elusive. The initial goal of this research program was to identify relatively simple criteria that could distinguish histologically well preserved bone from tissue that would not yield visible structures. The histological study of archaeological bone holds promise as a source for research insights. Microscopic structure of cortical bone has been proposed as a source of information about age at death, pathological conditions and biomechanical patterns in life. However, research has progressed slowly. One frequent problem is that histological structures cannot be seen clearly in carefully prepared cortical thin sections because diagenetic changes have obscured or destroyed the original features. Our study sample is derived from human bones buried in very sandy, moist soil for 173 years at the Snake Hill site, on the northeast shore of Lake Erie, near Fort Erie, Ontario, Canada. The cortical bone tissue showed variable but frequently poor histological preservation despite homogeneity in the burial environment. The variability was unexpected due to the homogeneity of the burial sample, the temperate climate, relatively neutral soil pH, and the relatively short period of interment. As the work progressed, we found that we were asking questions that were complementary to those being asked by archaeologically oriented bone chemists. These questions focused on the relationship between the diagenetic processes acting on the inorganic and organic phases of the bone. This relationship has yet to be elucidated. Our study was designed to investigate the relationship between histological preservation and collagen integrity of archaeological bone. Since the structure of bone is determined in part by the organization of the collagen fibrils, degradation of this framework might result in a simultaneous degradation of the structure which that framework supports. Hence, one might expect the histological structure of bone to collapse if the organic framework were degraded. If the histological preservation of bone were to mirror the preservation of its collagen in a consistent fashion, this relationship would provide a useful criterion for sample selection in stable isotope analyses.

PRESERVATION IN ARCHAEOLOGICAL BONE

143

Our investigation of the above-proposed relationship did not demonstrate a simple criterion for screening samples. Good histological preservation does not serve as an indicator of well preserved collagen. Complimentary to our findings is a recent study by Hedges et al. (1995) that found that histological preservation and protein content in prehistoric bone are independent.

The bone tissue used in our study is relatively young in comparison to that used by Hedges and Millard (1995), yet the same relationships are noted. Our results indicate that certain elemental ratios and yield values cannot be consistently correlated with good collagen yields. Ambrose and Norr (1992) noted that while the proteinaceous material extracted from prehistoric bone may be comprised mainly of collagen, small amounts of other substances such as noncollagenous bone proteins, post-mortem contaminants, and inorganic residues may also be present. Therefore, they preferred to refer to the proteinaceous material as “collagen”. Our research elaborates on this theme, demonstrating the impact of the non-collagenous components on archaeological interpretation of bone chemistry.

BACKGROUND TO THE STUDY The War of 1812 was fought between the British and the Americans, along the Canadian-American border, from the Atlantic seaboard to the Great Lakes region. Fort Erie, Ontario, Canada was occupied by the Americans during the summer and fall of 1814. Cortical bone tissue used in this study was derived from the femoral midshafts of American soldiers, killed between June and October, 1814, in battle or in siege-related activities focusing on possession of the Fort. The dead soldiers were interred near the surgeon’s tent at the site of Snake Hill, where they were discovered and disinterred in 1987 (Pfeiffer and Williamson 1991). Both the physical evidence and the historical documentation are consistent with the interpretation that all were soldiers who were buried immediately after death with little or no funerary preparation. Protocol during the American occupation of Fort Erie dictated that wounded American soldiers would be ferried back across the Niagara River to Buffalo, New York; even those who were killed were often returned to American soil for burial. The positions of buttons and bandage pins suggest that the individuals interred on the Fort Erie side had sustained injuries from shot or shrapnel such that injury and loss of blood hastened their death. Medical waste pits, containing the remains of amputated arms and legs, were also excavated but none of the samples in this study are from amputated limbs. All the samples here are from relatively intact young men, tall in stature, probably all from the American Northeast. A study of their collagen and isotope values, done to contribute to ascertainment of nationality, demonstrated a relatively broad range of variability (Katzenberg 1991). This variability was interpreted as being consistent with their probable origins

144

SUSAN PFEIFFER AND TAMARA L. VARNEY

throughout the American Northeast (Whitehorne 1991). A small number of O isotope values are also available (Schwarcz et al. 1991). In 1990 the remains

of the soldiers were reinterred in American soil. The analyses described here are of small tissue samples; some adjustments to sampling protocol have been necessary as we have, in some cases, exhausted the available bone tissue. The burials were located on a sandy soil beach, within approximately 25 meters of the Lake Erie shore near its confluence with the Niagara River (Fig. 7.1). As the region of investigation was stripped of overburden, various relic topsoils were noted, overlain by windblown sands. Grave shafts were identifiable through the presence of organic soil mixed with sterile sand. The burials themselves were in sterile sediments, predominantly sand (Thomas and Williamson 1991). The five relatively orderly rows were at a depth just above the current water table. A septic tank and drainage pipe which served a summer cottage had been installed above Burials 8, 10 and 11. The Lake Erie

north shore drainage basin has been used for relatively intensive agriculture, horticulture and human habitation in recent times. Although the soil in which the soldiers were buried is primarily sand, the bones were exposed to frequent ground water fluctuations and substantial phosphate and organic run-off. The soil pH values of soil samples taken from the burials were neutral to slightly alkaline, ranging from 6.9 to 7.8. There is no patterning between these pH values and the position of the sample relative to the cottage septic system, the shoreline, or any other discernible feature of physical geography.

Figure 7.1. Site map. Snake Hill. The well-preserved specimen, Burial 8, is to the left of the septic

pipe. Other burials sampled in this study are marked with circles. The Lake Erie shore line is to the right, ca. 25m. (Modified from Pfeiffer and Williamson 1991).

PRESERVATION IN ARCHAEOLOGICAL BONE

145

Midshaft cross sections of all available femora (n = 23) were initially prepared for histological study. Most of the specimens showed a soft, “chalky” consistency with an intact subperiosteal layer, which tended to break away from the midcortical material. There was no histological evidence for fungal ingrowth from either subperiosteal or endosteal surfaces. Foci of destruction are consistent with descriptions of bacterial intrusion (Fig. 7.2, Pfeiffer 1995). The majority of the cross-sections exhibited a concentric pattern of differential histological preservation in which three distinct layers of preservation could be discerned. These layers were also visible to the naked eye in the form of slight color variations, with the middle layer being slightly darker. The

layers ranged in thickness from 0.5 mm to 4 mm, with the middle layer being the thickest (Fig. 7.3). Only one specimen (Burial 8) yielded a high quality thin section across its full cortical width. Proximity to the septic system is associated with both excellent preservation (Burial 8) and poor preservation (Burials 10 and 11). Subsequent investigation into what features might distinguish the femur from Burial 8 initially focused on the mineral fraction of a selection of Snake Hill femora. Energy dispersive x-ray microanalysis (JEOL JSM-35C SEM equipped with a TN-5500 X-ray analyzer) demonstrated consistent calcium to

Figure 7.2. Microradiograph of Burial 30, periosteal surface at top. Snake Hill bone microradiographs show radiolucent cavities around Haversian canals surrounded by hypermineralized rims,

which Baud and Lacotte (1984) argue are characteristic of bacterial colonization.

146

SUSAN PFEIFFER AND TAMARA L. VARNEY

PRESERVATION IN ARCHAEOLOGICAL BONE

147

phosphorus ratios across the cortical cross-sections of the bones, including

well preserved and poorly preserved regions. The presence of trace elements (>5 ppm) did not correspond to histological preservation. There was no difference among apatite spectra when studied using x-ray diffraction, nor was there any difference in soil pH (mean = 7.3, s.d. = 0.22). Microradiographic images were compared for differences in density using gray level comparisons in a video image analysis system (Jandel Java). There were no significant differences in gray levels, but the images did demonstrate a pattern of modification consistent with the activity of microorganisms (Pfeiffer 1989). Noting that macroscopically Burial 8 was the only specimen to appear “greasy,” we extracted the lipid fraction using a modified Bligh and Dyer (1959) method (as modified by DeNiro and Epstein 1978 and White 1991). Burial 8 showed a significantly lower lipid yield than other samples from a subset of six of the other burials (1.75 mg/g versus 3.5 to 7.2 mg/g)

(Pfeiffer 1995). This outcome was consistent with a hypothesis that structural deterioration could have been a byproduct of microorganism activity. The higher lipid content in the poorly preserved tissue suggests that those lipids are primarily extrinsic, that is, that they were produced by bacteria and/or fungi. As the food

source for such microorganisms, the protein within the bone may have been substantially altered in concert with the microstructure deterioration. The quantification of the changes to the organic fraction became our next focus of research.

MATERIALS AND METHODS We began our study of the organic fraction with the hypothesis that there would be a relationship between the preservation of collagen and preservation of histological features (as per Schoeninger et al. 1989). Samples for collagen extraction were taken from each preservational layer of the eight femoral cortices. The layer closest to the periosteal surface was denoted the outer layer, while that adjacent to the medullary canal was denoted the inner layer. The layer in-between was then referred to as the middle layer. The periosteal and endosteal surfaces of the cross-sections were manually scrubbed and left intact for sampling. The outer and inner layers of most of the bone samples fall into the category of bad histological preservation which corresponds to a histological preservation index of 0–1, as per Hedges et al. (1995), or less than 15% intact bone (grade 3, as per Cook 1992). The middle layers show good preservation

that correspond to a histological preservation index of 3–5, as per Hedges et al. (1995), or greater than 67% of the structures are intact (grades 0–2, as per Cook 1992). The two exceptions to this pattern are Burials 8 and 30 which exhibit good and bad histological preservation, respectively, across their entire cross-

148

SUSAN PFEIFFER AND TAMARA L. VARNEY

sections. Since three layers could still be discerned in the sample from Burial 30, it was partitioned as such, while the sample from Burial 8 was processed whole. The eight samples are from burials distributed across the excavated area. In addition to these partitioned samples, bone powder that had the lipids removed was also available from three of the samples in the partitioned set (Burials 1, 8 and 20), as well as two other samples (Burial 17 and 24). These samples included bone from the entire cross-section. The “collagen” component of each bone sample was extracted following the technique of White (1991) which modifies Longin’s (1971) method by: a) using a gentler acid treatment of repeated soaking in 0.25 N HC1 as per Chisholm et al. (1983), and b) refluxing at a lower temperature of 58°C as per Brown et al. (1988). This gentler sample treatment was applied to minimize protein and large peptide degradation while maximizing product yield. While Sealy’s “pseudomorph” method (1986) has been preferred by various researchers, our division of the poorly preserved bone into layers often yielded mainly bone powder so the pseudomorph approach could not be consistently applied. Product yield expressed as a percentage of dry bone powder weight was calculated. Evaluation of the extracted collagen was by assessment of collagen yield, carbon and nitrogen concentrations, carbon to nitrogen atomic ratios, amino acid profiles, and amino acid yield. These features have been suggested, and some have been widely used, as evaluation criteria for sample suitability in studies involving light stable isotope analysis of archaeological bone. Approximately 5 mg of each “collagen” sample was analyzed for total carbon and nitrogen content using a Carlo-Erba CHN analyzer in the laboratory of Dr. A. Oakes, University of Guelph. From absolute C and absolute N values the C: N atomic ratios were calculated. Approximately 10 mg of each “collagen” sample was analyzed for its amino acid composition by High Performance Liquid Chromotography (HPLC) using the Waters PICOTAG system and MAXIMA 820 computer software in the laboratory of Dr. T. E. Graham, University of Guelph. The PICOTAG system allows the derivitization of proline and hydroxyproline without additional oxidative procedures and the samples are stable for long periods of time when freeze dried (Heinrikson and Meredith 1984). The high sensitivity of the system allows analysis to the picomole concentration. The Guelph system is regularly used for analysis of amino acid composition of human muscle biopsy samples. The PICOTAG derivitization system had not been previously used for archaeologically derived collagen samples at the time our analysis took place, but discussions with the manufacturer and current users identified no potential difficulties. Also Gurley et al. (1991) used a similar system in their analysis of the proteinaceous extract from a dinosaur fossil. The data were expressed both as residues per thousand and as dry weight relative to “collagen” dry weight. The standard was bovine achilles tendon (Sigma). Methods are discussed more fully by Varney (1994).

PRESERVATION IN ARCHAEOLOGICAL BONE

149

RESULTS

Criteria for Sample Acceptability In the following description and discussion of our results, various criteria derived from the literature are used to determine whether or not each sample is of adequate preservation to allow it to be confidently included in a stable isotope study. The values applied in the various criteria have been found to be associated with archaeological bone collagen that retained an isotopic composition that was reflective of its diet, while the majority of samples that had values outside of the criteria did not retain an isotopic composition that reflected diet. The values for these criteria may vary slightly depending upon the collagen extraction methodology used, and such criteria are not exact. In this study samples that fall within the criteria values are deemed “acceptable,” and those that do not are deemed “unacceptable.”

“Collagen” Yield All 22 samples (Table 7.1) produced yields within the range (5–25%) that is generally associated with archaeological bone collagen that retains an isotopic signal that is reflective of its diet (Ambrose and Norr 1992, Schoeninger et al. 1989). Our samples show yields ranging from 7.9 to 21.4% (mean 15.4%, s.d. 4.03). While the highest yield comes from the worst preserved bone, yields from the partitioned samples do not show any statistically significant differences when compared by grades of histological preservation or by layers of bone (one-way ANOVA). The three lipid-removed samples show a trend toward higher yields, although they

are not directly comparable because they are not divided into outer, middle and inner layers.

C : N Ratios DeNiro (1985) suggested that the range of 2.9 to 3.6 for atomic C : N

ratios be used as a criterion for selecting suitable samples for stable isotope analysis, since he found samples with C : N ratios outside this range did not retain an isotopic composition reflective of their diet. When this criterion is applied in this study, seventeen of twenty-two ratios are acceptable (Table 7.1). The five aberrant values are from bone that showed poor histological preservation: four from outer layers, one from an inner layer. In all five cases the C : N ratios are high. The C : N ratios for the three samples that had the lipids removed fall between 3.2 and 3.3 (the value for modern collagen). Some of the partitioned samples that correspond to the same cross-sections (Burial 1 and 20) show C : N values outside of the acceptable range.

150

SUSAN PFEIFFER AND TAMARA L. VARNEY

PRESERVATION IN ARCHAEOLOGICAL BONE

151

C and N Concentrations Ambrose (1990, 1993) found that for most samples well-preserved “collagen” had at least 3% carbon and 1% nitrogen. If this criterion is applied in this study, two of our 22 samples show aberrant values for carbon and one of 22 is aberrant for nitrogen, with three of the samples containing 1% nitrogen (Table 7.1). Of the three samples where we can compare bone with lipidremoved and lipids intact, both C and N concentrations are within acceptable limits in the lipid-removed sample, but substantially lower in most layers with lipid intact. The concentrations of both C and N are low, but within

acceptable criteria, in the sample with well preserved histology (C = 5.11%, N = 1.77%).

Yield of Amino Acids When the yield is expressed in terms of nmol of amino acids relative to the dry weight of the collagen standard sample, the results indicate a very poor yield from the archaeological samples. The mean amino acid

yield of the three standard replicates, 14,910 nmol/mg (s.d. 1432), is much greater than the amino acid yields of any of the archaeological samples. Without lipid removal, they yield from 3 to 17% amino acids by weight relative to the yield of the standard. Removal of lipids increases the amino acid yields to values up to 27% of the standard value. Values for individual amino acids remain proportional. In recent work published by Tuross et al. (1994), only samples yielding amino acids weighing 95% of the standard were included in analysis. Clearly none of these samples meet this criterion.

Relationships Among “Collagen” Preservation Criteria None of the elemental data show any correlation with “collagen” yield. When compared to yield, the correlation coefficients are non-significant for C : N (r = 0.4), %C (r = 0.4) and %N (r = 0.2). All samples with low elemen-

tal values are from layers of bone that showed poor histological preservation. While low values tend to come from bone with poor histological preservation, many samples from bone of equally poor preservation produced acceptable values. Lipid removal generally improved both the yield characteristics and elemental values.

Amino Acid Analysis In comparison with values from the literature, the majority of the amino acid values for our collagen standard (3 replicates) are within acceptable

152

SUSAN PFEIFFER AND TAMARA L. VARNEY

ranges. However, aspartic acid and glutamic acid show comparatively low proportions. Aspartic acid is present at about one quarter the expected value (0.8 vs. 4%) and glutamic acid is present at about one-half the expected value

(3% vs. 7.8%). Results from the archaeological samples follow this same pattern of underrepresentation. All the archaeological samples yielded amino acid profiles that are generally consistent with fresh collagen. Two known breakdown products of collagen amino acids, alpha-amino butyric acid (aaba) and ornithine (orn) (Baker 1991, 1992), are detectable in the standard and in all of the archaeological samples. Five archaeological samples show higher levels of aaba than the standard. These include four outer layers and Burial 8. Several archaeological samples show levels of proline that are higher than expected. Converting amino acid values into Z-scores expressed relative to the mean of the three standard replicates, samples from the outer layers clearly deviate the most from expected and they show the greatest variation across samples (Table 7.2). Neither Students’ t-tests nor one-way ANOVA reveal any significant differences in the proportions of glycine, proline, hydroxyproline, alanine, arginine, glutamic or aspartic acid by grade of histological preservation or by layer of bone.

Internal Consistency To investigate the potential variability of repeated analysis, seven aliquots of a single sample of “collagen” from Burial 8 were subjected to HPLC analysis (Fig. 7.4). The amino acid mean values of these seven runs conform very closely to the expected values for fresh collagen. However, the coefficients of variation are much higher than those of the collagen standard. This suggests that the extracted product is not homogeneous in its composition.

DISCUSSION The Snake Hill sample was chosen for this research because the cortical bone had been buried for a known length of time in relatively homogeneous

PRESERVATION IN ARCHAEOLOGICAL BONE

153

Figure 7.4. Amino acid profiles for seven aliquots from “collagen” of Burial 8.

soil. The femoral midshafts were similar in cortical thickness, being from large, young adult males. Our research demonstrates that while one may obtain a

“collagen” yield of 20% from archaeological bone, very little of that yield may actually be protein. This mirrors the findings of other researchers who have used different collagen extraction and purification procedures (e.g., inclusion of NaOH treatment; Schoeninger et al. 1989). Some of our “collagen” samples

had a composition that was less than 10% amino acids. The compositional nature of the remaining substance is unknown, but the amino acid analysis indicates that it is not proteinaceous. The amino acids which are retained appear to be a very close approximation to fresh collagen in both the amino acids represented and their proportions.

In addition, our results suggest that removal of lipids improves both yield characteristics and elemental characteristics. Recent work by Liden et al. (1995) suggests that the methanol-chloroform method used here is more effective than other methods, such as treatment with NaOH solution, or the maintenance of an acidic environment and ultrafiltration of products during

collagen extraction. It is speculated that the presence of lipids in archaeological bone samples may interfere with the acid hydrolysis of protein during

154

SUSAN PFEIFFER AND TAMARA L. VARNEY

collagen extraction, which would be exacerbated in samples containing lesser amounts of collagen. Although humic acids were not removed through NaOH treatment, the samples with aberrant elemental values had counterparts with good elemental values once the lipids had been removed. Histological tissue preservation does not signal preservation of bone suitable for chemical analysis of the collagen fraction. Our histologically well preserved specimen was not distinct from the poorly preserved samples in its “collagen” yield characteristics or its amino acid yield characteristics (Table 7.3). The loss of protein from archaeological bone appears to be independent of the loss of histological structures, a finding that is also supported in a study by Hedges et al. (1995). Our findings are consistent with those of Hedges et al. (1995). Interpreting our results in the context of the model proposed by Collins et al. (1995), there is an argument for an enzymatic pattern of biodegradation. They observed a direct correlation between loss of tensile strength and loss of collagen in instances of chemical degradation, but no such correlation in enzymatic degradation conditions. Although we did not quantify this characteristic, the well-preserved specimen had substantially more tensile strength than the other chalky, easily fragmented specimens, yet collagen quality was similar. The degradation process which they postulate focuses on physicalchemical changes including the depolymerization of collagen by chemical hydrolysis of bone peptides and the dissolution of those polypeptide fragments through an insufficiency of available H bonds. Our collagen yield results may better fit an enzymatic pattern of degradation. One particularly unexpected result is the high variability observed among aliquots from a single “collagen” sample, derived from apparently homogeneous cortical bone. The variability in amino acid profiles within one bone may be the direct result of collagenase-producing soil-borne pioneer bacteria that are responsible for initial colonization (Child et al. 1993), or the organisms that succeed them (Child 1995). Despite variation in histological preservation, samples from all layers of the cross-sections produced amino acid profiles that were similar to that of modern collagen. Therefore, there does not appear to be a relationship between amino acid composition and histological preservation. However, the outer layer of the bone cross-sections had amino acid compositions that deviated the most from the expected values, and showed the greatest variation across samples. This greater susceptibility to post-mortem change is no doubt due to the outer cortex being in direct contact with the burial environment. Our finding of greater post-mortem change to the “collagen” in the outer layers emphasizes the need to completely remove the external surfaces of archaeological bone samples that are destined for isotopic or amino acid analysis. Although our results for the inner layers are somewhat better, it may be prudent to remove both periosteal and endosteal cortical surfaces. The low yields of aspartic and glutamic acid which limit our research interpretations are most likely due to some aspect of laboratory method

PRESERVATION IN ARCHAEOLOGICAL BONE

155

Table 7.3. Summary of yield characteristics expressed relative to the standards recommended. Key: O: outer layer, M: middle layer, I: inner layer, L: lipids removed (includes material from entire cross-section of bone), +: meets criterion,

–: fails to meet criterion.

associated with the hydrolysis of the samples prior to amino acid analysis. These two amino acids have been found to be particularly sensitive at that stage (Hunt 1985, Hughes and Frutier 1990, West and Crabb 1990). Other researchers have successfully used the same derivitization methodology to analyze archaeological and fossil material (Gurley et al. 1991 and Liden et al. 1995).

Performance measured through the currently proposed benchmarks for identifying good “collagen” does not correlate well with amino acid yield or

156

SUSAN PFEIFFER AND TAMARA L. VARNEY

the presence of degradation products. While the removal of lipids improved

the amino acid yield of all samples, there is clearly a great deal of material in “collagen” which is not actually collagen. Furthermore, that material is not homogeneous in its distribution throughout the proteinaceous extract. This variability may reflect the cumulative action of both biodegradation and chemical hydrolysis followed by subsequent leaching. CONCLUSIONS Subsequent research may demonstrate that if isotopic analysis is done on extracted amino acids, even bone with poor yield characteristics may produce a valid isotopic dietary signal. Conversely, our research demonstrates that the organic fraction of bone is certainly not stable after burial, for even a relatively

short period of interment, in moist, sandy conditions with a high, variable

water table. In addition, our results illustrate that histological preservation does not signal excellent organic preservation. On the basis of our finding that the outer layer of cortical bone seems to have undergone greater post-mortem alteration than the remainder of the cross-section, we recommend that the external surfaces of bone samples destined for isotopic or amino acid analysis be completely removed. Research into bone diagenesis will continue to contribute to the accuracy and precision of archaeological bone chemistry. Not surprisingly, both biological and physical factors appear to influence the structure and composition of bone from organic burial environments. The relative importance of these two factors in a particular environment may determine the appropriateness of those bone samples for chemical analysis. The best methods for identifying and quantifying these diagenetic changes remain to be established. ACKNOWLEDGMENTS We sincerely thank a number of colleagues at the University of Guelph for their support, noting especially T. Graham, P Sathasivam, and A. Oakes. Access to the Snake Hill material was through our colleagues at Archaeological Services, Inc., Toronto, Ontario, with thanks to R. Williamson.

REFERENCES Ambrose, S.H. 1990 Preparation and characterization of bone and tooth collagen for isotopic analy-

sis. Journal of Archaeological Science 17: 431–451. __1993 Isotopic analysis of paleodiets: methodological and interpretive considerations. In Sandford, M.K., ed., Investigations of Ancient Human Tissue: Chemical Analyses in Anthropology. Langhorne, PA., Gordon and Breach Science Publishers: 59–130.

PRESERVATION IN ARCHAEOLOGICAL BONE

157

Ambrose, S.H. and Norr, L. 1992 On stable isotopic data and prehistoric subsistence in the Soconusco region. Current Anthropology 33: 401–404. Baker, B.J. 1991 Collagen preservation and composition in ancient Nubian skeletal remains.

American Journal of Physical Anthropology, Supplement 12: 46–47. __1992 Collagen composition in human skeletal remains from the NAX cemetery (A.D. 350–550)

in Lower Nubia. Ph.D. dissertation, University of Massachusetts, Amherst, MA. Baud, C.-A. and Lacotte, D. 1984 Étude au microscope électronique á transmission de la colonisation bactérienne de l’os mort. Comptes Rendus de L’ Académic des Sciences 298: 507–510. Bligh, E.G. and Dyer, W.J. 1959 A rapid method of total lipid extraction and purification. Canadian Journal of Biochemical Physiology 37: 911–917. Brown, T.A., Nelson, D.A., Vogel, J.S. and Southern, J.A. 1988 Improved collagen extraction by modified Longin method. Radiocarbon 30: 171–177. Child, A.M. 1995 Towards an understanding of the microbial decomposition of archaeological bone in the burial environment. Journal of Archaeological Science 22: 165–174. Child, A.M., Gillard, R.D. and Pollard, A.M. 1993 Microbially-induced promotion of amino acid racemization in bone: isolation of the microorganisms and the detection of their enzymes. Journal of Archaeological Science 20: 159–168. Chisholm, B.S., Nelson, D.E., Hobson, K.A., Schwarcz, H.P. and Knyf, M. 1983 Carbon isotope measurement techniques for bone collagen: notes for the archaeologist. Journal of Archaeological Science 10: 355–360. Collins, M.J., Riley, M.S., Child, A.M. and Turner-Walker, G. 1995 A basic mathematical simulation of the chemical degradation of ancient collagen. Journal of Archaeological Science 22: 175–183. Cook, M.A. 1992 Age-at-death estimation in human skeletal remains: using histomorphometry. M.Sc. thesis, University of Western Ontario, London, ON, Canada. DeNiro, M.J. 1985 Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to paleodietary reconstruction. Nature 317: 806–809. DeNiro, M.J. and Epstein, S. 1978 Influence of diet on the distribution of carbon isotopes in

animals. Geochimica et Cosmochimica Acta 42: 495–506. Fairgrieve, S.I. 1993 Amino acid residue analysis of Type I collagen in human hard tissue: an assess-

ment of cribra orbitalia in an ancient skeletal sample from tomb 31, site 31/435-D5-2, Dakhleh Oasis, Egypt. Ph.D. dissertation, University of Toronto. Gurley, L.R., Valdez, J.G., Spall, W.D., Smith, B.F. and Gillette, D.D. 1991 Proteins in the fossil bone of the dinosaur, Seismosaurus. Journal of Protein Chemistry 10: 75–90.

Hedges, R.E.M., Millard, A.R. and Pike, A.W.G. 1995 Measurements and relationships of diagenetic alteration of bone from three archaeological sites. Journal of Archaeological Science 22: 201–209. Heinrikson, R.L. and Meredith, S.C. 1984 Amino acid analysis by reverse-phase high-performance liquid chromotography: Precolumn derivatization with phenylisothiocyanate. Analytical Biochemistry 136: 65–74. Hughes, G.J. and Frutier, S. 1990 Amino acid analysis: protocols, possibilities, and pretensions. In Fini, C., Floridid, A., Finelli, V.N. and Wittman-Leibold, B., eds., Laboratory Methodology in Biochemistry: Amino Acid Analysis and Protein Sequencing. Boca Raton, FL, CRC Press, Inc.: 44–61.

Hunt, S. 1985 Degradation of amino acids accompanying in vitro protein hydrolysis. In Barrett, G.C., ed., Chemistry and Biochemistry of the Amino Acids. London, Chapman and Hall: 376–398. Katzenberg, M.A. 1991 Analysis of stable isotopes of carbon and nitrogen. In Pfeiffer, S. and Williamson, R.F., eds., Snake Hill: An Investigation of a Military Cemetery from the War of 1812. Toronto, Dundrun Press: 247–255. Liden, K., Takahasi, C. and Nelson, D.E. 1995 The effects of lipids in stable carbon isotope analy-

sis and the effects of NaOH treatment on the composition of extracted bone collagen. Journal of Archaeological Science 22: 321–326.

158

SUSAN PFEIFFER AND TAMARA L. VARNEY

Longin, R. 1971 New method of collagen extraction for radiocarbon dating. Nature 230: 241–242. Pfeiffer, S. 1989 Characterization of archaeological bone decomposition in a sample of known length of interment. American Journal of Physical Anthropology 78: 283. __1995 An exploration of possible relationships between structural and chemical decomposition of bone. Proceedings of the I World Congress on Mummy Studies. Museo Arqueologico y Etnografico doe Tenerife, Islas Canarias (Spain): 549–558. Pfeiffer, S. and Williamson, R.F., eds. 1991 Snake Hill: An Investigation of a Military Cemetery from the War of 1812. Toronto, Dundurn Press. Schoeninger, M.J., Moore, K.M., Murray, M.L. and Kingston, J.D. 1989 Detection of bone preservation in archaeological and fossil samples. Applied Geochemistry 4: 281–292. Schwarcz, H.P., Gibbs, L. and Knyf, M. 1991 Oxygen isotopic analysis as an indicator of place of origin. In Pfeiffer, S. and Williamson, R.E, eds., Snake Hill: An Investigation of a Military Cemetery from the War of 1812. Toronto, Dundurn Press: 263–268. Sealy, J. 1986 Stable carbon isotopes and prehistoric diets in the South-western Cape Province, South Africa. Cambridge Monographs in African Archaeology 15, BAR International Series 293. Thomas, S.C. and Williamson, R.F. 1991 Archaeological Investigations. In Pfeiffer, S. and Williamson, R.F., eds., Snake Hill: An Investigation of a Military Cemetery from the War of 1812. Toronto, Dundrun Press: 70–159. Tuross, N., Fogel, M.L., Newsom, L. and Doran, G.H. 1994 Subsistence in the Florida Archaic— The stable isotope and archaeobotanical evidence from the Wendover site. American Antiquity 59: 288–303. Varney, T. 1994 Characterization of “collagen” in histologically modified bone derived from an archaeological context. MSc Thesis, University of Guelph, Guelph, ON, Canada. West, K.A. and Crabb, J.W 1990 Performance evaluation: automatic hydrolysis and PTC amino acid analysis. In Vallafraca, J.J., ed., Current Research in Protein Chemistry: Techniques,

Structure, and Function. San Diego, Academic Press: 37–48. White, C.D. 1991 Isotopic Analysis of Multiple Human Tissues from Three Ancient Nubian Populations. Ph.D. dissertation, University of Toronto. Whitehorne, J. 1991 Fort Erie and U.S. operations on the Niagara Frontier, 1814. In Pfeiffer, S. and Williamson, R.F., eds., Snake Hill: An Investigation of a Military Cemetery from the War of 1812. Toronto, Dundrun Press: 25–60.

Chapter

8

The Use and Abuse of Trace Elements for Paleodietary Research JAMES H. BURTON

AND

T. DOUGLAS PRICE

ABSTRACT Recognition among bone-chemistry researchers that strontium enters bone in proportion to dietary levels has resulted in widely accepted yet erroneous inferences about the relationships among various elements in bone and past diet. One such inference is that more of any element in the diet translates directly to more of that element in bone. If an element is not biogenically incorporated within bone, or if biological levels are metabolically controlled, then that element will not reflect diet. A second erroneous inference is that strontium can be used to measure the dietary plant/meat ratio. Sr/Ca ratios in meat are generally lower than those of plants, but meat is also low in calcium and hence has little effect on the composition of bone. Plants, on the other hand, contribute substantially to bone composition. Variations in the strontium levels of bone thus more likely reflect differential consumption of plants rather than trophic position. Although efforts to determine plant/meat ratios from strontium and to draw dietary inferences from elements other than strontium and barium have not been successful, this failure has been due to inappropriate expectations, not to a failure of bone strontium to reflect diet.

Biogeochemical Approaches to Paleodietary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

159

160

JAMES H. BURTON AND T. DOUGLAS PRICE

INTRODUCTION Although the chemical analysis of bone has gained widespread recognition as a tool for paleodietary and paleoenvironmental research, the primary focus of such research is now mainly isotopic. Trace-element compositional

studies have fallen into disfavor, a consequence of disappointing attempts to deduce prehistoric behavior from such analyses. These disappointments, however, are largely the historical result of applying overly enthusiastic and simplistic ideas when drawing inferences from bone compositions. These ideas, in turn, stem from an accumulation of methodological missteps, individually quite minor but cumulatively significant enough to warrant serious reevaluation of the trace-element approach.

Most archaeologists who have explored the utility of strontium, the primary focus of paleodietary studies, are aware that the theoretical underpinning of their research stems from the thermonuclear tests of the Cold War,

specifically from the global redistribution of radioactive Sr-90 produced during these tests. The fact that strontium is chemically similar to calcium and accumulates in calcium-rich substances, e.g., mother’s milk and growing bones, led to widespread alarm that an entire generation could be at risk from radiationinduced diseases of the bone. This alarm prompted a quarter century of intense studies of strontium metabolism in humans and into its distribution in food webs (Pecher 1941; Harrison 1955; Alexander et al. 1956; Wasserman et al. 1956, 1957, 1958; Engström 1957; Odum 1957; Menzel and Heald 1959; Gran 1960; Holmberg 1960; Barnes et al. 1961; Comar et al. 1961; Harris et al. 1961; Ichikawa 1963; Lengemann 1963; Lough et al. 1963; McClellan 1964; Bronner and Lemaire 1969; Martin 1969; Rosenthal et al. 1970, 1972; Wolf 1973; Bratter et al. 1977; Rosenthal 1981; Elias et al. 1982). It was quickly recognized that virtually the entire body-burden of strontium (>90%) is incorporated into the skeleton (Hodges et al. 1950; Comar et al. 1952, 1955, 1956, 1957; Hartsook et al. 1956; Turekian and Kulp 1956; Thurber et al. 1958; Burton and Mercer 1962; Comar 1963; Baud et al. 1968; Schroeder et al. 1972). These studies also revealed, however, that strontium levels decrease as one moves up through the food chain. Because strontium does not pass through the gut wall as efficiently as calcium, the Sr/Ca ratio of an animal’s tissues is lower than that of the food it consumes, with tissue Sr/Ca ratios typically being about 20% of the dietary ratio (Elias et al. 1982). This discrimination, sometimes called “biopurification”, means that an herbivore has in its bones about 1/5 the Sr/Ca ratio of it’s diet and a carnivore feeding exclusively on that herbivore would have in its bones 1/5 the Sr/Ca ratio of the herbivore tissues, or only about four percent the Sr/Ca ratio of the herbivore diet. Because there is additional discrimination against strontium in the mammary gland, milk also has Sr/Ca ratios that are substantially lower than dietary levels (Wasserman et al. 1958). Recognition of this biopurification of calcium in the food-chain, coupled with the cessation of above-ground thermonuclear tests, largely

TRACE ELEMENTS FOR PALEODIETARY RESEARCH

161

dissipated concern about radiostrontium, but gave birth to the use of strontium analyses to assess positions in the food chain (Toots and Voorhies 1965). Archaeologists, recognizing that Sr levels reflected trophic position, began trying to apply bone Sr measurements to address important paleodietary issues, the relative consumption of meat versus plants in particular (Brown 1973, 1974; Kavanagh 1979; Sillen 1981, 1986; Price and Kavanagh 1982; Schoeninger 1982) By comparing Sr levels in human bones to levels in carnivores and herbivores, researchers began to quantitatively infer the ratio of plants to meat in prehistoric human diets. Strontium was also applied in a relative fashion, comparing average Sr levels among sets of samples (e.g. social classes and genders) to deduce which group consumed the most meat. Although numerous studies failed to observe expected trophic-level patterns in strontium data (Schoeninger and Peebles 1981; Decker 1986; Blitz 1995), this remains today the most common manner in which archaeologists try to apply trace element data. About the time archaeologist began using strontium as a proxy for the amount of meat versus plants in the diet, a thesis by Gilbert (1975) suggested that bone zinc levels could also reflect plant/meat ratios, although in an opposite sense of meat being high in zinc while plants are generally lower. This was followed by the use of. other elemental studies that sought to interpret a high or low concentration of virtually any element through comparison to high or low abundances in particularly dietary items (Beck 1985; Francalacci and Borgognini Tarli 1988; Buikstra 1989; Whitmer et al. 1989; Lidén 1990; Weydert 1990). Thus a situation arrived where, for example, sodium was claimed to reveal “sex roles”, and copper was given as evidence for the intentional cultivation of maggots as a “prelude to the invention of agriculture” (Arrhenius 1990). A few other studies (Hancock et al. 1989, 1994) actually tested the utility of some of these elements and, finding none, brought wholesale discredit to the trace-element approach. So now we have a research method that continues to be applied in a routine, if not appropriate, manner by archaeologists while much of the archaeometric community that created the methodology has abandoned it.

USE OF NON-ALKALINE EARTH ELEMENTS The methodological misstep most responsible for this interpretive chaos began with the inclusion of elements other than strontium and barium. Although it required extensive research to reveal the bone-diet connection for

strontium, this was reduced, evidently for simplicity, and not altogether without validity, to the idea that plants are higher in strontium, than is meat, and that bones reflect this dietary difference. Then with Gilbert’s thesis suggesting that zinc apparently also reflects this, the concept as transcribed to

162

JAMES H. BURTON AND T. DOUGLAS PRICE

archaeologists was that bones chemically reflect the chemistry of the diet. Lost

in this translation, however, was the critical requirement, established earlier for strontium, that the element must be incorporated within bone (and if bone ash is the analyte, then the element of interest should specifically be incorporated within the hydroxyapatite lattice of bone mineral, rather than in the

organic phase). Ezzo (1994a, 1994b), who recently reviewed this “you are what you eat” acceptance of other elements, enumerated specific criteria that should be attributes of any inorganic paleodietary indicator. The element of interest should not only be incorporated within bone mineral but should also be relatively free from metabolic control (i.e., biological levels should not be homeostatic and thus independent of diet), and biological levels should considerably exceed levels that can be anticipated from post-depositional processes, which is a problem even for strontium. Although the elements that have been analyzed by archaeologists for paleodietary purposes span the periodic chart

(Whitmer et al. 1989; Sandford 1993), Ezzo’s criteria essentially constrain elements with real paleodietary potential to just strontium barium, and possibly lead. We can improve the current situation by simply omitting these other elements from our analyses, but we would nonetheless forfeit useful information. While non-alkaline earth elements are not likely to inform us about prehistoric diet, the fact that they are not biologically incorporated within bone

mineral makes them useful as diagenetic indicators (Price 1989; Price et al. 1992; Ezzo 1994a). The use of non-alkaline-earth elements such as zirconium to assess diagenetic contamination is actually one of the earliest applications of such elements (Katzenberg 1984). There are many ways that bones can be contaminated, e.g., mechanically or physically by soil, chemically by soluble salts, biologically by microorganisms that precipitate metallic oxides on bone surfaces, and through recrystallization of bone hydroxyapatite. Because measures for any one of these modes of contamination does not necessarily imply anything about other modes, analyses of multiple elements can be fruitful even though they cannot directly reveal past diets. Analyses of bones from Paloma, a preceramic site of coastal Peru (Reitz 1988; Quilter 1989), illustrate this utility. Barium measurements revealed that marine resources comprised the dominant source of dietary calcium at Paloma (Burton and Price 1990). Examining barium data as a function of age revealed significantly higher barium levels in infants and children than in adults. While it’s tempting to interpret this as a dietary change with age, examination of other

elements reveals a high correlation between barium and manganese, i.e., evidence of post-mortem contamination by hausmannite, a manganese-barium oxide precipitated on the surfaces of bone by microbial action (Parker and Toots 1972). The increase of barium with decreasing age is actually a systematic artifact of the efficiency of cleaning the bones. The femora of adults were large enough to thoroughly remove exterior and interior surfaces. The

TRACE ELEMENTS FOR PALEODIETARY RESEARCH

163

exterior surfaces of pre-teen femora could be mechanically removed, but the interior medulliary surfaces were more difficult; paper-thin bones from infants could not be mechanically cleaned. This systematic contamination due to the inefficiency of the cleaning process was only evident through monitoring elements other than those of dietary interest. Any paleodietary inference drawn in the absence of such data would likely be in error.

STRONTIUM AS A MEASURE OF THE PLANT/MEAT RATIO

Omitting elements other than Ba and Sr as direct paleodietary indicators is critically necessary, but is not, in itself, sufficient. A second significant misstep in trace-element research was inferring that bone strontium is a proxy for the dietary plant/meat ratio. The recognition of the trophic biopurification of calcium has been adopted as a tool for assessing one’s position in a foodchain continuum, often as a more or less quantitative measure of a plant/meat ratio of the diet. The major error in using bone Sr/Ca as a proportional measure of the plant/meat ratio is that the Sr/Ca ratio of the diet, itself, is not proportionally related to the plant/meat ratio (Burton and Wright 1995; Burton 1996).

In a simple, three-component food chain (plant/herbivore/carnivore) without compositional variability within individual levels, the carnivore will indeed have less bone strontium than will the herbivore. Problems arise however when one tries to apply this on a continuous scale, i.e., as a measure of intermediate trophic positions. Examining the composition of the total diet as a function of varying plant/meat ratio is highly instructive. Total diet Sr/Ca ratios, calculated from typical plant and meat compositions (Watt and Merrill 1963; Shacklette 1980; Elias et al. 1982), are shown in Figure 8.1. The Sr/Ca ratio of a binary diet of meat and amaranth, which has Ca and Sr levels typical of leafy vegetables, is shown as a function of increasing meat in the diet. Although pure plant and pure meat diets have, ipso facto, appropriate end-member compositions, the dietary Sr/Ca ratio for mixed diets is not linearly related to the plant/meat ratio or even particularly sensitive to it. Notice that the amount of meat can change from virtually nil to nearly 90% of the diet by weight without visibly affecting the dietary Sr/Ca ratio. Leafy vegetables such as amaranth have more than ten times as much calcium and strontium as meat. In other words, 100 grams of leafy vegetables have a greater effect on diet compositions, and, hence, upon bone, than does a kilogram of meat. Plants, themselves, vary greatly in their mineral content and correspondingly in their ability to affect dietary Sr/Ca. It is similarly instructive to examine the effect upon diet composition of changing the kinds of plants, while holding other parameters constant. Figure 8.2 shows the Sr/Ca ratio of a binary diet as a function of the calcium content of the plant, i.e., as a function of the type of plant being consumed. Figure 8.2 was generated using a

164

JAMES H. BURTON AND T. DOUGLAS PRICE

Figure 8.1. Dietary Sr/Ca as a function of the plant/meat ratio.

constant Sr/Ca ratio of 0.00790 for the plants and fixing the amount of meat, with Sr/Ca constant at 0.00158, arbitrarily at 50% of the diet by weight. Observe that a change from leafy vegetables to grains, in contrast to changing the plant/meat ratio, has a substantial effect on dietary Sr/Ca. The decreasing dietary Sr/Ca ratio with decreasing plant mineral content is toward the Sr/Ca of meat, and away from that of the plants. The relative mineral contribution of meat to the diet is increasing although the absolute amount of meat is, in

this example, constant. If there were no meat in the diet, the graphical relationship would, in this hypothetical example, be flat. Conversely, the effect of changing plants would be somewhat more pronounced if there were more meat in the diet, appearing as lower Sr/Ca ratios for diets having plants with the lowest calcium. In Figure 8.2, for the purpose of examining the effect of variable calcium, the Sr/Ca ratios of all the plants were set to the same fixed value (0.00790). In nature this situation is rarely likely to exist because there is intrinsic variability among different plant species (Runia 1987, 1988). The fact that bone represents a long-term (ca. 7 years) dietary average somewhat obviates this source of variability. To the extent, however, that the dietary selection of plants varies systematically, the dietary Sr/Ca ratio will reflect this difference. Consistent dietary differences in the selection of plants are likely to affect bone Sr/Ca, especially since plants will usually be the dominant source of calcium. It should be noted that one of the earliest paleodietary applications of bone

TRACE ELEMENTS FOR PALEODIETARY RESEARCH

165

Figure 8.2. Dietary Sr/Ca as a function of the calcium content of plants.

strontium, that of Toots and Voorhies (1965), used these compositional differences among plants to assess not the dietary plant/meat ratio but the degree of browsing versus grazing. It is this phenomenon due to shifting plants that we are likely to observe in comparative Sr/Ca studies of bone, rather than the amount of meat in the diet, simply because plants have a much greater impact on diet composition. A study by Ezzo et al. (1995) illustrates this sensitivity to vegetation. In an earlier study of bone samples from late preagricultural, early contact, and late contact period sites from the Georgia Bight, isotopic and other archaeological data show that there was an increased reliance upon maize (Larsen et al. 1992). The trace-element data of Ezzo et al. (1995) show substantially less barium in the agricultural periods, indicating an increased contribution of bone-forming minerals from marine resources. Isotopic data are inconsistent, however, with any increase in seafood. The contradiction is resolved by recognizing that the trace-elements reflect not the total diet, but the main dietary sources of calcium. Although seafood consumption was not greater in the agricultural societies, the relative contribution of bone-forming minerals from seafood increased because low-calcium maize had replaced high-mineral plants. This is exactly analogous to the example of Figure 8.2 where a shift to maize decreases the mineral contribution from plants, this change in dietary composition being reflected in the bones.

This sensitivity to vegetation also appears in the data from our study of modem plants and animals. More than a thousand samples spanning a wide range of species of both plants and animals were selected from a five-county area in northeastern Wisconsin. While herbivores have significantly lower

166

JAMES H. BURTON AND T. DOUGLAS PRICE

Figure 8.3. Bone Sr/Ca of mammals from northern Wisconsin.

Sr/Ca ratios than plants, and carnivore Sr/Ca ratios are lower than those of herbivores, there is substantial variability among the plants and correspondingly among various herbivores and carnivores (Figure 8.3). For example, beavers and hares, both pure herbivores, had 1100ppm and 310ppm bone strontium, respectively. If one were to use existing trophic-level concepts, taking the highSr beaver as representing a diet of only plants and the low-Sr bobcat (avg. = 90ppm) as a diet of only meat, one might then infer intermediate diets for intermediate Sr levels. A linear model would imply that hares have a diet of 80% meat. However, the hares do have Sr/Ca ratios that are 1/5 that of grass, and the bobcats have Sr levels only slightly higher than 1/5 that of the hares. Beavers are herbivorous, but consume the bark of shrubs and trees which have substantially elevated strontium. Although the mammalian strontium levels do not map dietary plant/meat ratios, they do accurately reflect actual dietary Sr/Ca ratios.

CONCLUSION Although there are many failed attempts to determine even relative measures of meat consumption, the disappointment is not due to any failure of strontium to reflect diet but to an unwarranted expectation that bone strontium should necessarily reflect meat consumption. The frustration with efforts to draw paleodietary inferences stems from simplistically equating Sr/Ca ratios with plant/meat ratios and to the inappropriate use of elements not

TRACE ELEMENTS FOR PALEODIETARY RESEARCH

167

biologically incorporated within bone in proportion to dietary levels. In spite of some contemporary applications, the repeatedly confirmed theoretical foundations for inferring that bone Sr/Ca ratios reflect diet are sound. Although much analytical work is required, even on a case-by-case basis, to understand the sources of variability in Sr/Ca ratios and to draw appropriate archaeological inferences, this should not be seen as a problem but as an opportunity. The examples discussed above suggest useful directions for future

research involving trace element analysis of bones. Specifically, the effects of developmental age and other factors (e.g., porosity, mineralization) that may lead to differences in surface area of specimens should be considered. Diagenetic effects should be monitored by analysis of a suite of elements whose abundances are not controlled by dietary abundances (e.g., Mn, Zr, etc.). Finally, although alkaline elements such as Sr and Ba are most likely to reflect the Sr/Ca and Ba/Ca levels of the diet, omnivores such as humans are likely to obtain the majority of these elements from plants rather than from animals. Therefore for accurate diet reconstruction it is necessary to determine the total abundance of Ca as and the Sr/Ca and Ba/Ca ratios of the plant and animal resources that were potential dietary staples. The effects of culinary practices

on elemental abundances (Burton and Wright 1995; Katzenberg et al. this

volume) must also be evaluated.

ACKNOWLEDGMENTS This research was funded in part by grants from the National Science Foundation (SBR-9307301 and SBR-9409834) to T. Douglas Price and James

H. Burton. We would also like to thank James Ashbrenner of the Wisconsin Department of Natural Resources and William Middleton for help obtaining modern biological samples and Cynthia Hallock, Steven Wernke, Joel Glick, Kim Pettis, Amanda Park, Kaaren Brauner, Michelle Ninneman, and Tanya Zhong for assistance with sample preparation and analysis. Paloma samples were provided by Robert Benfer.

REFERENCES Alexander, G.V., Nusbaum, R.E. and MacDonald, N.S. 1956 The relative retention of strontium and calcium in bone tissue. Journal of Biological Chemistry 218: 911–919. Arrhenius, B. 1990 Trace element analyses of human skulls. In Arrhenius, B., ed., Laborativ Arkeologi. Stockholm, Stockholm University: 15–19. Barnes, D.W.H., Bishop, M., Harrison, G.E. and Sutton, A. 1961 Comparison of the plasma concentration and urinary excretion of strontium and calcium in man. International Journal of Radiation Biology 3: 637–646. Baud, C.A., Bang, S., Lee, H.S. and Baud, J.P 1968 X-ray studies of strontium incorporation into bone mineral in vivo. Calcified Tissue Research Supplement 2: 6.

168

JAMES H. BURTON AND T. DOUGLAS PRICE

Beck, L.A. 1985 Bivariate analysis of trace elements in bone. Journal of Human Evolution 14: 493–502. Biltz, J.A. 1995 Dietary Variability and Social Inequality at Monte Alban, Oaxaca, Mexico. Ph.D. dissertation, University of Wisconsin, Madison. Bratter, P.D.G., Lausch, J. and Rosick, U. 1977 On the distribution of trace elements in human skeletons. Journal of Radioanalytical Chemistry 37: 393–403.

Bronner, F. and Lemaire, R. 1969 Comparison of calcium kinetics in man and rat. Calcified Tissue Research 3: 238–248. Brown, A.B. 1973 Bone Strontium Content as a Dietary Indicator in Human Sheletal Populations. Ph.D. dissertation, University of Michigan, Ann Arbor. 1974 Bone strontium as a dietary indicator in human skeletal populations. Contributions to Geology 13: 47–48. Buikstra, J.E., Frankenberg, S., Lambert, J.B. and Liang Xue 1989 Multiple elements: multiple expectations. In Price, T.D., ed., The Chemistry of Prehistoric Human Bone. Cambridge,

Cambridge University Press: 155–210. Burton, J.D. and Mercer, E.R. 1962 Discrimination between strontium and calcium in their passage

from diet to the bone of adult man. Nature 193: 846–847. Burton, J.H. 1996 Trace-elements in bone as paleodietary indicators. In Orna, V. ed., Archaeological Chemistry. Washington, D.C., American Chemical Society: 327–333. Burton, J.H. and Price, T.D. 1990 The ratio of barium to strontium as a paleodietary indicator of consumption of marine resources. Journal of Archaeological Science 17: 547–557. Burton, J.H. and Wright, L.E. 1995 Nonlinearity in the relationship between bone Sr/Ca and diet: paleodietary implications. American Journal of Physical Anthropology 96: 273–282. Comar, C.L. 1963 Some over-all aspects of strontium-calcium discrimination. In Wasserman, R.H., ed., The Transfer of Calcium and Strontium Across Biological Membranes. New York, Academic Press: 405–418. Comar, C.L., Lotz, W.E. and Boyd, G.A. 1952 Autoradiographic studies of calcium, phosphorus, and strontium distribution in the bones of the growing pig. American Journal of Anatomy 90: 113–125. Comar, C.L., Whitney, I.B. and Lengemann, F.W. 1955 Comparative utilization of dietary 90Sr and calcium by the developing rat fetus and growing rat. Proceedings of the Society for Experimental Biology 88: 232–236. Comar, C.L. and Georgi, J. 1961 Assessment of chronic exposure to radio-strontium by urinary assay. Nature 191: 390–391. Comar, C.L., Russell, R.S. and Wasserman, R.H. 1957 Strontium-calcium movement from soil to man. Science 126: 485–496. Comar, C.L., Wasserman, R.H. and Nold, M.M. 1956 Strontium-calcium discrimination factors in the rat. Proceedings of the Society for Experimental Biology 92: 859–863. Decker, K.W. 1986 Isotopic and chemical reconstruction of diet and its biological and social dimensions at Grasshopper Pueblo, Arizona. Paper presented at the 51st Annual Meetings of the Society for American Archaeology, April 1986, New Orleans. Elias, R.W., Hirao, Y. and Patterson, C.C. 1982 The circumvention of the natural biopurification of calcium along nutrient pathways by atmospheric inputs of industrial lead. Geochimica

et Cosmochimica Acta 46: 2561–2580. Engström, A., Björnerstedt, R. and Clemedson, C.J. 1957 Bone and Radiostrontium. Stockholm, Almqvist & Wiksell. Ezzo, J.A. 1994a Putting the “Chemistry” back into archaeological bone chemistry analysis: modeling potential paleodietary Indicators. Journal of Anthropological Archaeology 13: 1–34. 1994b Zinc as a paleodietary indicator: an issue of theoretical validity in bone chemistry analysis. American Antiquity 59: 606–621. Ezzo, J.A., Larsen, C.S. and Burton, J.H. 1995 Elemental signatures of human diets from the Georgia Bight. American Journal of Physical Anthropology 98: 471–481.

TRACE ELEMENTS FOR PALEODIETARY RESEARCH

169

Francalacci, P. and Borgognini Tarli, S. 1988 Multielementary analysis of trace elements and preliminary results on stable isotopes in two Italian prehistoric sites. Methodological aspects. In Grupe, G. and Hermann, B., eds., Trace Elements in Environmental History. Berlin, Springer-Verlag: 41–52.

Gilbert, R.I. 1975 Trace Element Analyses of Three Skeletal Amerindian Populations at Dickson Mounds. Ph.D. dissertation, University of Massachusetts, Amherst.

Gran, F. 1960 Studies on calcium and strontium-90 metabolism in rats. Acta Physica Scandinavia 48. Hancock, R.G.V., Grynpas, M.D., Åkesson, K., Obrant, K.B., Turnquist, J. and Kessler, M.J. 1994 Baselines and variabilities of major and trace elements in bone. In Lambert, J. and Grupe, G., eds., Prehistoric Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 189–201. Hancock, R.G.V., Grynpas, M.D. and Pritzker, K.P.H. 1989 The abuse of bone analyses for archaeological dietary studies. Archaeometry 31: 169–180. Harris, R.S., Moor, J.R. and Wanner, R.L. 1961 Calcium metabolism of the normal rhesus monkey. Journal of Clinical Investigation 40: 1766–1774. Harrison, G., Raymond, W.H.A. and Tretheway, H.C. 1955 The metabolism of strontium in man. Clinical Science 14: 681–695.

Hartsook, E., Cowan, R.L., Chandler, P.T. and Whelan.J.B. 1956 The relative retention of strontium and calcium in bone tissue. Journal of Biological Chemistry 218: 911–919. Hodges, R.M., MacDonald, N.S., Nusbaum, R., Stearns, R., Ezmirlian, F., Spain, P. and McArthur, C. 1950 The strontium content of human bone. Journal of Biological Chemistry 185: 519–524.

Holmberg, B., Nelson, A. and Wallgren, E. 1960 The transfer of Sr-90 from mother to fetus in mice. Radiation Research 12: 167–172. Ichikawa, R.A.Y.E. 1963 Strontium-calcium discrimination in rats feeding on rice, milk and commercial diets. Health Physics 9: 717–720. Katzenberg, M.A. 1984 Chemical analysis of prehistoric human bone from five temporally distinct populations in southern Ontario. National Museum of Man, Mercury Series, No. 129, Ottawa, Archaeological Survey of Canada. Kavanagh, M. 1979 Strontium in Bone as a Dietary Indicator. Master’s thesis. University of Wisconsin, Madison.

Larsen, C.S., Schoeninger, M.J., van der Merwe, N.J., Moore, K.M. and Lee-Thorp, J. 1992 Carbon and nitrogen isotopic signatures of human dietary change in the Georgia Bight. American

Journal of Physical Anthropology 89: 197–214. Lengemann, F.W. 1963 Overall aspects of calcium and strontium absorption, In Wasserman, R.H., ed., Transfer of Calcium and Strontium Across Biological Membranes. New York, Academic Press: 85–96. Lidén, K. 1990 A diet study from the Middle Neolithic site Ire. In Arrhenius, B., ed., Laborativ Arkeologi. Stockholm, Stockholm University: 21–28. Lough, S., Rivera, J. and Comar, C.L. 1963 Retention of strontium, calcium, and phosphorus in human infants. Proceedings of the Society of Experimental Biological Medicine 112: 613–636. Martin, A. 1969 Content and distribution of stable strontium and 226Ra in human skeletons from Wisconsin descendants 1957–1961. British Journal of Radiology 42: 295–298. McClellan, R.O. 1964 Calcium-strontium discrimination in miniature pigs as related to age. Nature 202: 104–106. Menzel, R.G. and Heald, W.R. 1959 Strontium and calcium contents of crop plants in relation to exchangeable strontium and calcium in the soil. Proceedings of the Soil Science Society of

America 23: 110–112. Odum, H.T. 1957 Biogeochemical Deposition of Strontium. Institute of Marine Science Publications No. 4, University of Texas, Austin: 39–106.

170

JAMES H. BURTON AND T. DOUGLAS PRICE

Parker, R.B. and Toots, H. 1972 Hollandite-coronadite in fossil bone. American Mineralogist 57: 1527–1530. Pecher, C. 1941 Biological investigations with radioactive calcium and strontium. Proceedings of the Society for Experimental Biology Medicine 46: 86–91. Price, T.D. 1989 Multi-element studies of diagenesis in prehistoric bone. In Price, T.D., ed., The Chemistry of Prehistoric Human Bone. Cambridge, Cambridge University Press: 126–154. Price, T.D., Biltz, J., Burton, J.H. and Ezzo, J. 1992 Diagenesis in prehistoric bone: problems and solutions. Journal of Archaeological Science 19: 513–529. Price, T.D. and Kavanagh, M. 1982 Bone composition and the reconstruction of diet: examples from the midwestern United States. Mid-Continental Journal of Archaeology 7: 61–70. Quilter, J. 1989 Life and Death at Paloma. Iowa City, University of Iowa Press. Reitz, E. 1988 Faunal remains from Paloma, an Archaic site in Peru. American Anthropologist 90: 310–322. Rosenthal, H.L. 1981 Content of stable strontium in man and animal biota. In Skoryna, S.C., ed., Handbook of Sable Strontium. New York, Plenum Press: 503–514. Rosenthal, H.L., Chochran, O.A. and Eves, M.M. 1972 Strontium content of mammalian bone,

diet, and excreta. Environmental Research 5: 182–191. Rosenthal, H.L., Eves, M.M. and Cochran, O.A. 1970 Common strontium of mineralized tissues from marine and sweet water animals. Comparative Biochemistry and Physiology 32: 445–450. Runia, J.T. 1987 Strontium and calcium distribution in plants: effect on paleodietary studies. Journal of Archaeological Science 14: 599–608. 1988 Discrimination factors on different trophic levels in relation to the trace element content in human bones. In Grupe, G. and Herrmann, B., eds., Trace Elements in Environmental History. Berlin, Springer-Verlag: 53–66. Sandford, M.K. 1993 Investigations of Ancient Human Tissue: Chemical Analyses in Anthropology. Langehorne, PA., Gordon and Breach Scientific Publishers.

Schoeninger, M.J. 1982 Diet and the evolution of modern human form in the Middle East. American Journal of Physical Anthropology 58: 37–52. Schoeninger, M.J. and Peebles, C.S. 1981 Effect of mollusc eating on human bone strontium levels.

Journal of Archaeological Science 8: 391–397. Schroeder, H.A., Tipton, I.H. and Nason, A.P. 1972 Trace metals in man: strontium and barium. Journal of Chronic Diseases 19: 545–571. Shacklette, H.T. 1980 Elements in Fruits and Vegetables from Areas of Commercial Production in the Conterminous United States. U.S. Geological Survey Professional Publication, No. 1178. Washington, DC, U.S. Government Printing Office. Sillen, A. 1981 Strontium and diet at Hayonim Cave. American Journal of Physical Anthropology 56: 131–137.

1986 Biogenic and diagenetic Sr/Ca in Plio-Pleistocene fossils of the Omo Shungura Formation. Paleobiology 12: 311–323. Thurber, D.L., Kulp, J.L., Hodges, E., Cast, P.W. and Wampler, J.M. 1958 Common strontium content of the human skeleton. Science 128: 256–257. Toots, H. and Voorhies, M.R. 1965 Strontium in fossil bones and the reconstruction of food chains. Science 149: 854–855. Turekian, K.K. and Kulp, J.L. 1956 Strontium content of human Bones. Science 124: 405–407. Wasserman, R. and Comar, C.L. 1956 Carbohydrates and gastrointestinal absorption of radiostron-

tium and radiocalcium in the rat. Proceedings of the Society for Experimental Biological Medicine 101: 314–317. Wasserman, R., Comar, C.L. and Papadopoulou, D. 1957 Dietary calcium levels and retention of radiostrontium in the growing rat. Science 126: 1180–1182. Wasserman, R.H., Lengemann, F.W. and Comar, C.L. 1958 Comparative metabolism of calcium and strontium in lactation. Journal of Dairy Science 41: 812–821.

TRACE ELEMENTS FOR PALEODIETARY RESEARCH

171

Watt, B.K. and Merril, A.L. 1963 Composition of Foods. U.S. Department of Agriculture, Handbook No. 8. Washington, U.S. Government Printing Office. Weydert, J. 1990 Elemental Analysis of Bone for Ancient Diet Reconstruction. Unpublished Ph.D. dissertation, Northwestern University. Whitmer, A.M., Ramenofsky, A.F., Thomas, J., Thibodeaux, L., Field, S.D. and Miller, B.J. 1989

Stability or instability. The role of diffusion in trace element studies. Archaeological Method and Theory 1: 205–273. Wolf, N., Gedalia, I., Yariv, S. and Zuckermann, H. 1973 The strontium content of bones and teeth of human foetuses. Archives of Oral Biology 18: 233–238.

This page intentionally left blank

Chapter

9

Modeling Protein Diagenesis in Ancient Bone: Towards a

Validation of Stable Isotope Data GISELA GRUPE, ASTRID BALZER, AND SUSANNE TURBAN-JUST

ABSTRACT Modern mammalian bone was experimentally inoculated with soil bacteria to induce protein diagenesis. The protein relics extracted from these bones were compared with degraded protein from archaeological human bone. Protein identification was done by gel electrophoresis and amino acid analyses, whereby non-proteinogenetic amino acids were especially monitored. We present a model for microbial collagen break-down and the resulting consequences for stable isotopic data and thus paleodiet reconstruction. While bacteria produce a shift in towards lighter values as a result of selective amino acid consumption, an overall enrichment of in the range of a trophic level effect was detected. A method for an estimation of in vivo values in collagen is suggested by correction of diagenetically altered amino acid profiles. Since a shift in should largely be the result of peptide bond cleaving, no such estimations are possible. Biogeochemical Approaches to Paleodietary Analysis , edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

173

174

GISELA GRUPE ET AL.

INTRODUCTION The research potential of stable isotope analyses from archaeological bone collagen is enormous, especially for the reconstruction of paleodiet and subsistence strategies. The chemical components of bone tissue are capable of telling life histories, and certain aspects of these histories, including diet, give clues to the major determinants of population development in the past (DeNiro 1985; Ambrose 1986). Type I collagen, which is the interstitial matrix protein of mammalian bone, is protected by the bone’s mineral matrix, an almost insoluble hydroxyapatite. Thus, collagen is able to survive very long inhumation periods, even in unfavorable soil conditions (Grupe 1995). However, this does not necessarily mean that intact molecules are preserved. Since matter is recycled and bone protein decomposes with time and in relation to soil conditions and time since burial, bone collagen lose its structural properties and suffers change in its amino acid composition. Undoubtedly, collagen protection by bone mineral depends on both the preservation of the hydroxyapatite and non-collagenous proteins (NCP’s), which are responsible for the protein/mineral bonds. It has been claimed that NCP’s, due to their acidic nature, tend to be better preserved than collagen (Masters 1987).

Unlike bone mineral, which decomposes mainly through physical and chemical soil properties, protein suffers largely from microbial decomposition. While collagen is also hydrolyzed at acidic pH after the protecting mineral layers have been solubilized, it is nonetheless actively degraded by soil microorganisms which utilize bone protein as their nutrient substrate. Previous experiments have demonstrated how soil fungi and soil bacteria invade bone interstices, destroy protein/mineral bonds and metabolize specific amino acids (Grupe et al. 1993). Since these phenomena belong to the initial stages of bone decomposition, microbial invasion is responsible for the further fate of a buried bone. The most important experimental result derived was that soil microorganisms induce selective collagen breakdown. One should assume that microbial degradation is likely to change the stable carbon and nitrogen isotope ratios of extractable bone collagen in relation to the amount and type of preserved/metabolized amino acids. Depending on their biochemical pathway, every amino acid has its own isotopic signature; thus, the collagen and values are dependent on their amino acid composition. Dietary behavior can only be deduced from the consumer’s isotopic composition as long as there is no decomposition artifact present. It was the aim of this research to experimentally find out how soil bacteria affect bone collagen, how they change the

collagen’s amino acid composition (Macko et al. 1983) and which deviations from the original isotopic signature may be expected in archaeological bone specimens that have experienced microbial invasion (Fig. 9.1). What does “modeling protein diagenesis” mean? The ultimate goal would be the full understanding of biochemical processes, but the probability

TOWARDS A VALIDATION OF STABLE ISOTOPE DATA

175

Figure 9.1. Thin section of compact bone from Schleswig (cf. Table 9.1), after application of a cell stain (methylene blue). Besides the bone cells, exogenous cellular material in the hollow spaces (shown by arrows) indicates previous microbial invasions.

of attaining this goal is low. Rather, it was our aim to validate isotopic data from bone collagen, and in so doing, seek to answer the following questions. What do microorganisms do to a bone? How does their action influence and values? Is it possible to sort out the biological and the diagenetic signals? Is it possible to estimate the native d value from a degraded protein? The answers could offer a safer framework in which to interpret archaeometric data. While the aim of this decomposition research becomes clear, so do the limitations. For instance, a single amino acid will have varying isotopic ratios due to its origin, depending on whether it is cereal, fish, milk or meat. Also, values might vary simply due to position effects. It seems obvious that one will not be capable of revealing all these details in retrospect, but the knowledge of the amount of uncertainty with regard to stable isotope ratios of degraded collagen in itself will be beneficial in preventing misinterpretation of archaeometric data.

MATERIALS AND METHODS Macerated femora from modern marten (Martes martes, an omnivorous mustelid) were cut into pieces, sterilized by irradiation at 25kgy (kilograys)

176

GISELA GRUPE ET AL.

and inoculated with three species of soil bacteria: Alkaligenes piechaudii, Bacillus subtilis and Pseudomonas fluorescens. These three bacterial species were chosen because they all show collagenase activity. Alkaligenes and Pseudomonas are gram-negative and utilize amino acids as both carbon and nitrogen sources. Alanine (ala), aspartic acid (asp), glutamic acid (gly) and proline (pro) are preferentially metabolized for carbon. Bacillus subtilis is gram-positive and prefers glucose and as carbon and nitrogen source. It is one of the most important decomposers of dead organic matter (Starr et al. 1981). The samples were kept for six to nine months in the dark in optimum bacterial growth temperature and aerobic conditions to enhance the rate of protein degradation. It has been shown previously that these bacteria are capable of growing on bone as their only substrate at temperatures as low as 4°C; higher temperatures only change the rate but not the mode of bacterial break-down (Grupe et al. 1993). The bacteria were raised on nutrient broth

prior to inoculation, during which time the growth medium was enriched in some instances by one weight percent of powdered wheat, maize or beans in order to check whether bacteria would leave the isotopic signal of their source within the infected bones. To enable comparison to this experimental approach, archaeological human bones of various ages and soil properties (Table 9.1) from the Anthropological Collection in Munich were analyzed. All German skeletal series come from humic soil with, neutral to slightly basic pH. The samples from Turkey, Syria, coastal Peru and Egypt have been buried in dry, sandy soils. Soil samples from most of the excavation sites were available and bone sample

TOWARDS A VALIDATION OF STABLE ISOTOPE DATA

177

numbers ranged from 5 to 35 individuals per series. Collagen extracted from modern human bone served as control. Most of the specimens were analyzed in duplicate, thus more than 200 bone samples were analyzed in this study. Non-mineral-bound, non-collagenous proteins, mineral-bound, noncollagenous proteins and collagens were separated and purified from the homogenized bone samples according to a three-step extraction protocol using guanidine-HCL and EDTA (Grupe 1995). Protein purification by desalting and dialysis had to be modified in relation to bone preservation. Thus, desalting

over 10PDG columns (Biorad) was performed once or twice, and dialysis against distilled water was performed for one to three days. Lyophilized (freeze-dried) NCP’s and collagen were forwarded to gel electrophoresis (Desaga) and the molecular range of the extracted proteins were determined on 10% SDS-PAGE gels (Serva). Serum NCP’s and collagen bands were stained with Coomassie Blue R250, while mineral-bound NCP’s were stained with silver nitrate (Grupe and Turban-Just 1996). Amino acid profiles of all three protein fractions were determined with an HPLC-supported amino acid analyzer (Alpha-plus, Pharmacia) equipped with a Li-system. A variety of non-proteinogenic amino acids (Table 9.2), of which the majority were of

microbial or plant origin, were routinely monitored as checks for microbial contamination. and values were determined by mass spectrometry (MM 903, VG Isogas), equipped with a CN elemental analyzer (Roboprep CN, Europa Scientific). As control for sample purity, C : N ratios from the elemental analyser were compared with C: N ratios as calculated from the amino acid profile of the same sample.

178

GISELA GRUPE ET AL.

RESULTS The bacterial organic matter contained high amounts of aspartic acid (asp), threonine (thr), glutamic acid (glu) and especially alanine (ala), which is a constituent of bacterial cell walls. Amino acids not present in mammalian collagen, e.g., muramic acid (mura), pimelic acid (pim), sarcosine (sarc), carnitine (carn), also occurred, but in lower quantities. Table 9.3 shows a selection of representative amino acid profiles (all raw data from the inoculation experiments are as listed by Balzer, 1994) and indicate efficiency of the collagen extraction procedure. Since commercial type 1 collagen contained a lot of impurities (Table 9.3), freshly prepared collagen extracts from modern human and marten bones were used as internal laboratory standards. It should be noted that marten collagen differs in certain aspects from the amino acid profile of human collagen in that marten “collagen” (the insoluble residue

obtained by the extraction protocol) is low in asp and hydroxyproline (OHpro), but exceptionally high in thr and pro. This was the case for all marten bone specimens and the result was stable even after duplicate analyses. The result may partly be explained by the fact that the specimens came from free-ranging marten were captured in winter (Grupe and Kruger 1990), during which the marten metabolism is at rest and hydroxylation of pro may be reduced. The low value of the marten collagen (Table 9.4) is also probably due to this physiological state of the organism, but we cannot exclude “in vitro” decomposition, since the specimens had been sampled some years before this study was carried out. However, the specimens had been kept in the deep freeze in sealed bags after sterilization. Although our substrate differs from collagen in its amino acid composition, it is still suitable for study of change in composition after bacterial breakdown. The collagen extracts from the inoculated marten bones differ from the control with regard to their amino acid composition (Table 9.3), but no nonproteinogenetic amino acids were detected. Non-proteinogenetic amino acids were, however, present in the extract of soluble serum proteins, indicating that the exogenous biomass was leached from the infected bones by our extraction protocol. Although it is not possible to determine whether a specific amino acid (e.g., ala) in the collagen extract is of collagenous or bacterial origin, we are confident that the collagen residue was as “clean” as possible and that bacterial remnants, if at all still present in the extract, are negligible. Stable carbon and nitrogen isotope ratios of collagen showed alteration after experimental protein degradation. The bacterial colonies were in isotopic equilibrium with their growth substrate (nutrient broth) prior to inoculation, with a slight trophic level effect for both carbon and nitrogen (Table 9.4). The infecting bacterial biomass differed significantly from the collagen isotopic signature of the control specimen, which showed a value typical for a consumer in a plant community, and low due to physiological reasons. An induced microbial degradation, independent of the composition of the

TOWARDS A VALIDATION OF STABLE ISOTOPE DATA

179

bacterial growth medium prior to inoculation, resulted in a depletion of collagen extract of up to 2.4‰ in and an enrichment of up to 3.6‰ with in (Table 9.4), indicating that the bacteria produced a trophic level effect in stable nitrogen isotope ratios. They did not, however, leave their own isotopic

signatures in the degraded collagen since this would have led to corresponding enriched stable carbon isotope ratios (Balzer 1994).

180

GISEIA GRUPE ET AL.

Table 9.4. C:N molar ratios (calculated and measured), total C and N content and stable carbon and nitrogen isotope data from bacteria, their growth medium (nutrient broth), and from collagen (infected and non-infected marten bone). The bacteria for inoculation were raised on nutrient broth (nb), with/without additives. nd: not determined.

The detailed results of the analyses of the excavated human bones and soil samples from the respective archaeological sites are the subject of a forthcoming paper (Balzer et al. 1997, Turban-Just, in prep.). Therefore, only the major topics relating to the experimental approach shall be considered here. As expected, the integrity of the collagen extracts varied according to burial conditions and time, and despite some intragroup variability, patterns were group specific. Comparison of collagen and serum protein amino acid profiles revealed that solubilized collagen was detectable among the serum proteins as a result of degradation. This also holds for well-preserved and even partly mummified specimens from coastal Peru (Fig. 9.2), which is no longer surprising considering the rapid break-down of protein by bacteria in vitro. Thus, most protein bands of the G-extract after gel electrophoresis also showed collagenous bands at 160 and 80kDa in addition to the lighter serum proteins like albumin and a2-HS-glycoprotein. As a rule of thumb, better preserved collagen led to higher amounts of solubilized collagen remnants in both NCPextracts of samples from humic soils. Among the German skeletal series, the oldest one, Bell Beaker, closely resembled the badly preserved samples from dry, desert soils, in that badly preserved collagen did not necessarily leave

TOWARDS A VALIDATION OF STABLE ISOTOPE DATA

181

Figure 9.2. Amino acid content (mol %) in collagen extract (dark columns) and serum proteins (light columns) from a skeleton from coastal Peru. Due to a prevalence of degraded and soluble

collagen, the nonmineral-bound protein fraction shows a collagen amino acid profile.

traces of solubilized collagen remnants among the NCP’s. This result roughly describes the successive degradation of bone collagen: denaturation, solubilization and leaching from the specimen.

In general, it was possible to group the skeletal series into categories according to the state of collagen degradation: nearly intact (Steinhöring); excess of OHpro (Weichering), ascribed by Elster (1990) as “super-collagen”; specific loss of amino acids (Minshat Abu Omar, Grape 1995); and no preserved collagen but relics of only certain amino acids (Tell ed Der). Specific loss of certain amino acids from collagen was an important result of previous break-down experiments (Grupe et al. 1993). Of all samples analyzed, 54 (~25%) showed complete loss of certain amino acids (Fig. 9.3). Two aspects of this result are worth considering. First, among amino acids that have the same number of carbon atoms, those which occur in lowest amounts in collagen are lost first. Second, calculation of the weighted mean reveals that only amino acids with higher numbers of carbon atoms are more frequently lost than the mean. Since non-proteinogenetic amino acids, including those of microbial origin, were abundant among the NCP-extracts from the archaeological specimens (Fig. 9.4), it is highly probable that the presence of microbial biomass in the specimens is related to the state of collagen preservation, indicating that the samples have suffered from previous biogenetic decomposition. Since amino acids were preferentially lost from collagen as the number of carbon atoms increases, it seems noteworthy that high carbon amino acids should be a valuable energy source for decomposing microorganisms.

182

GISELA GRUPE ET AL.

Figure 9.3. Percentage of non-detectable amino acids in collagen extracts from archaeological human skeletons. Numbers on top of the columns indicate number of carbon atoms per amino acid. XW = weighted mean of % loss (cf. text). Only high-carbon amino acids are more frequently lost than the average.

Figure 9.4. Non-proteinogenetic amino acids in NCP’s from all archaeological human bones. The values are the averages of all archaeological human bone samples taken together (cf. Table 9.1). Dark columns: mineral-bound NCP; light columns: serum proteins. Most non-proteinogenic amino acids are extracted with the serum proteins.

DISCUSSION AND CONCLUSIONS

Biogenetic Collagen Degradation The results from the decomposition experiments clearly support

the previous hypothesis of selective biogenetic collagen break-down by

TOWARDS A VALIDATION OF STABLE ISOTOPE DATA

183

microorganisms (Grupe et al. 1993). Archaeological human bones that showed signs of microbial invasion at any time during burial had collagen amino acid profiles comparable to those from the experiments. Soil microorganisms with collagenase activity metabolize specific amino acids not only with regard to their tissue demands, but also with regard to the energy yield. Thus, amino acids with high numbers of carbon atoms will be lost preferentially from the bone’s matrix. Loss of amino acids means destruction of the collagen molecule, and since hydroxylated amino acids, which are important for the stabilization of both the helix and the fibril, are also affected (Alberts et al. 1990; Fig. 9.3), the molecule must have lost its structural properties. How is it then possible to extract “collagen” with a still “collagenous” amino acid profile, and how can such a protein produce bands with the proper molecular weight range on the gel? A tentative explanation is that this phenomenon is the result of transpeptidization reaction occurring under certain conditions: in systems where water concentration is low and where protease activity has led to an accumulation of protein break-down products, the enzymes are capable of reversing their usual catalytic function. In a stereospecific reaction, no further cleaving, but rather, repolymerization and build-up of new molecules take place. This system is well established in food chemistry (Nash et al. 1974, Weegels et al. 1995). Such a situation may well occur within the microsystem “bone” during the decomposition process, and the new polymers might be built under the exclusion of those amino acids which are metabolized by the protease producers (Balzer et al. 1997). All these processes must lead to an alteration of stable isotopic ratios of the extractable protein. Since this has been shown for experimentally degraded bone, the same should hold for archaeological human bone. Thus, it is highly probable that amino acid profiles from excavated bones with reduced numbers and altered concentrations of amino acids would give isotopic signals which

no longer reflect the subsistence behavior of the individual, but rather the bone’s diagenetic history. Since microbial break-down is even capable of inducing a “trophic level effect” in the reconstruction of paleodiet from bone may be very misleading. The application of an adequate collagen extraction protocol from bone excludes exogenous biomass. Therefore, altered isotope ratios are a product of microbial metabolism, and further cleaning procedures will not improve the results. While degraded collagen should become enriched in 15 N due to cleavage of peptide bonds (Bada et al. 1989), depleted values should rather be the result of changed amino acid composition. Due to their biochemical pathways, the values of the single amino acids differ from the respective value of the complete collagen molecule. As a rule of thumb, those amino acids with long biochemical pathways are more depleted in 13C, while those with short pathways have “heavier” isotopic signatures (Schmidt, pers. comm.). Thus,

alterations of amino acid composition should result in corresponding altered values, whereby, in principle, both a shift to lighter or heavier signatures should be possible.

184

GISELA GRUPE ET AL.

Problem Solution and Perspectives Since it is possible to differentiate well-preserved from badly preserved collagen through amino acid analysis and gel electrophoresis, it is also possible to determine which bone samples are likely to give erroneous isotopic ratios. At least for it should be possible to estimate the in vivo isotopic signature by correcting the changed amino acid concentrations of the “collagen” extract. This way, a reasonable approach to the reconstruction of paleodiet should be possible. Amino acid profiles are routinely reported as amino acid concentrations in mol %. Thus, a significant depletion in a particular amino acid raises the percentage of the remaining ones. At the same time, the percentage of contributed light and heavy carbon to the protein is altered. We tried to reestimate the original signature from the experimentally degraded marten bone proteins by adding missing or subtracting surplus amino acids, using as comparison an intact, unaltered collagen from the control specimen of noninfected marten bone. The contribution in % of any mol of a certain amino acid was then calculated, using the estimations of single amino acid as presented by Ambrose (1993; after Hare Estep 1983). This correction procedure was as follows: The amino acids threonine, glutamine, proline, glycine and alanine showed the largest deviation with regard to their abundance in the degraded proteins compared to the control with its unaltered composition. These five amino acids contribute to about 87mol% of the total amino acid profile, or 60% of collagen carbon. The correction was carried out by estimating the contribution to collagen by the amount of carbon atoms involved. For instance, in case of a higher abundance of a “light” amino acid in the altered protein, the degraded protein will also show a lower total

value.

Likewise, a higher abundance of “heavier” amino acids will cause a higher value for the degraded protein. For all of the five amino acids mentioned, the relative contribution to the total value is calculated according to the number of carbon atoms involved. These weighted figures are added or subtracted from the total value (according to whether a certain amino acid shows a higher or lower abundance in the altered protein). As a result, the native, unaltered of the total protein can be estimated. In the meantime,

ongoing work has determined and of single amino acids from samples of infected marten bone collagen. The deviation from the collagen isotopic ratio largely confirms the data presented by Ambrose (Turban-Just, in prep.). Table 9.5 shows that for most specimens, such a re-estimation of the

original isotopic signal from a degraded collagen works fairly well. Unquestionably, there are limitations to such a procedure, and because

calculations for two specimens failed, they lead to an even higher deviation from the control signal. These limits have to be defined by future work. The model experiment is only an approach to collagen diagenesis in buried bone. In nature, more than one bacterial species feeds from bone protein, and a

TOWARDS A VALIDATION OF STABLE ISOTOPE DATA Table 9.5. Re-estimation of in vivo

185 from

experimentally degraded collagen. * Correction failed. nd = not determined.

succession of certain types of soil microorganisms exists as well (Grupe et al. 1993). Therefore, whether even larger deviations from the in vivo isotopic signature may occur after biogenetic decomposition is still unclear. In addition, there are at present no means for a re-estimation of changed stable nitrogen isotope ratios because they are not caused by a simple shift in the collagen’s amino acid decomposition.

Conclusions The results of the decomposition experiments clearly show selective biogenetic collagen break-down by microorganisms. Amino acids with high numbers of carbon atoms are lost preferentially from the bone’s matrix. Archaeological human bones with signs of microbial invasion had collagen amino acid profiles comparable to those from the experiments. Loss of amino acids means destruction of the collagen molecule, but it is still possible to extract “collagen” with a “collagenous” amino acid profile and with the proper mol-

ecular weight range. We propose the tentative explanation that enzymes capable of reversing their usual catalytic function can repolymerize protein break-down products that have accumulated in the bone. If the new polymers are built without the amino acids that are selectively metabolized by the protease producers, then this must lead to an alteration of stable isotope ratios of the extractable protein. Since isotopic alteration has been shown for experimentally degraded bone, the same should hold for archaeological human bones that have amino

186

GISELA GRUPE ET AL.

acid profiles with reduced numbers and altered concentrations of amino acids. They may have isotopic signals that reflect their diagenetic histories rather than their subsistence behaviors. NOTE ADDED IN PROOF: This manuscript had been submitted shortly after the presentation of the paper at the Fourth Advanced Seminar on Paleodiet, 1994. Ongoing research, especially stable isotope analysis of single amino acids from inoculated and non-inoculated marten bones (same specimens as in this paper) further and strongly support our conclusion that bacterial modifica-tion causes substantial shifts in “collagen” stable isotope ratios (Balzer et al. 1997).

ACKNOWLEDGMENTS This research project was financially supported by the BMBE We are most indebted to Prof. Dr. H.-L. Schmidt, Technical University of Munich, for the measurement of stable isotope ratios and his advice and comments. Text edited

by Siew Eiselt.

REFERENCES Alberts, B., Bray, D., Lewis, J., Raff, M., Keith, R. and Watson, J.D. 1990 Molekularbiologie der Zelle.

Weinheim, VCH. Ambrose, S.H. 1986 Stable carbon and nitrogen isotope analysis of human and animal diet in

Africa. Journal of Human Ecology 15: 707–731. —— 1993 Isotopic analysis of paleodiets: Methodological and interpretative considerations. In Sandford, M.K., ed., Investigations of Ancient Human Tissue. Chemical Analyses in Anthropology. Langhorne, PA., Gordon & Breach: 59–130. Bada, J.L., Schoeninger, M.J. and Schimmelmann, M. 1989 Isotopic fractionation during peptide bond hydrolysis. Geochimica et Cosmochimica Acta 53: 3337–3341. Balzer, A. 1994 Experimenteller Abbau von Knochenprotein. Diplomarbeit (unpublished thesis), München. DeNiro, M.J. 1985 Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation of paleodietary reconstruction. Nature 317: 806–809. Elster, H. 1990 Age determination of fossil bone by amino add racemization. Ph.D. dissertation, Weizmann Institute, Rehovot. Grupe, G. and Krüger H. 1990 Feeding ecology of the stone and pine marten revealed by element analysis of their skeleton. Science of the Total Environment 90: 227–240. Grupe, G., Dreses-Werringloer, U. and Parsche, F. 1993 Initial stages of bone decomposition: Causes and consequences. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 257–274. Grupe, G. in press Preservation of collagen in bone from dry sandy soil. Journal of Archaeological Science. Grupe, G. and Turban-Just, S. in preparation Proteins in archaeological human bone. Macko, S.A., Estep, M.L.F., Hare, PE. and Hoering, T.C. 1983 Stable nitrogen and carbon isotopic composition of individual amino acids isolated from cultured microorganisms. Carnegie Institute of Washington Yearbook 82: 404–410.

TOWARDS A VALIDATION OF STABLE ISOTOPE DATA

187

Masters, P.M. 1987 Preferential preservation of noncollagenous protein during bone diagenesis: Implications for chronometric and stable isotopic measurements. Geochimica et Cosmochimica Acta 51: 3209–3214. Starr, M.P., Stolp, H., Trüper, H.G., Balows, A. and Schlegel, H.G. 1981 The Prokaryotes: A Handbook on Habitats, Isolation, and Identification of Bacteria. Vol. I. Berlin, Springer-Verlag. Turban-Just, S. in preparation Biogene Dekomposition von Knochenprotein. Ph.D. dissertation, München. Weegels, P.L., Orsel, R., van de Pijpekamp, A.M., Lichtendonk, W.J., Hamer, RJ. and Schofield, J.D. 1995 Functional properties of low Mr wheat proteins. 11. Effects on dough properties Journal of Cereal Science 21: 117–126.

This page intentionally left blank

Chapter

10

Some Biochemical Aspects of Carbon Isotopic Paleodiet Studies HENRY P. SCHWARCZ ABSTRACT Carbon isotope ratios of preserved human tissues (collagen and carbonate in bone/tooth apatite) are used as indicators of the sources of nutrients in human populations. We discuss the possible roles of some biochemical phenomena in determining the partitioning of isotopically labeled nutrients in these tissues. Respiration converts all nutrient-derived carbon into which is irreversibly fractionated with respect to bicarbonate at the time of formation. Together with poorly-known fractionation between apatite and solution, this leads to the observed c. 11‰ fractionation between apatite and diet. Experimental studies show that exogenous amino acids (AAs) are preferentially “routed” to collagen, even though enzymes for synthesis of non-essential (ne) AAs (making up c. 80% of collagen carbon) exist in all cells. Assimilation of protein apparently leads to inhibition and suppression of endogenous synthesis of neAAs. Some natural populations of humans with lower protein intakes may not experience this inhibition, leading to more endogenous neAA synthesis from carbohydrate and lipid carbon sources, approaching linear mixing. This is suggested by high values in some consuming populations. Correction of their diets for routing would lead to nutritionally inadequate (maize-rich) diets. Biochemical pathways for synthesis of some neAAs from lipids are partially blocked; this may account for the apparent difference in spacing of apatite and collagen between carnivores and herbivores. Biogeochemical Approaches to Paleodietary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

189

190

HENRY P. SCHWARCZ

INTRODUCTION The use of stable isotopic analyses of preserved human tissues to reconstruct paleodiet is based on the principle that “you are what you eat” (DeNiro and Epstein 1978). In applying stable isotope methods to the determination of paleodiet, a number of theoretical issues have arisen concerning the interpretation of the data. This paper will draw attention to some biochemical concepts bearing on these theoretical problems. It should not be surprising that biochemical principles would be relevant to these issues. The study of isotopic paleodiet has tended to treat the human body as a “black box” containing a number of discrete compartments (collagen, bone carbonate, etc.), and partitioning of isotopes between these compartments has been treated as an equilibrium state of the organism, albeit a poorly understood one. The biochemical processes that govern the partitioning of nutrients and their metabolic products amongst these compartments must be consistent with the biochemical

activities of these systems. Here I shall consider only a few of these processes and suggest how their actions may affect the isotopic behavior of the compartments. Further studies along this path may eventually allow us to better understand the relationship between diet and isotopic composition human tissues. The biochemical concepts discussed in this paper can be found in most introductory texts on biochemistry (e.g., Stryer 1988; Lehninger et al. 1993). This paper will address three subjects: a) carbon isotopic composition of bone carbonate; b) the possible role of lipid metabolism in determining fractionation between bone carbonate and collagen and its apparent trophic signature; and c) concepts of “routing” of carbon isotopes in the body, especially from protein foods to collagen. The discussion will be limited to carbon isotope ratios although we appreciate that ratios, which are also very useful in determining trophic levels and sources of protein, present similar biochemical problems.

SOME BASIC BIOCHEMICAL PRINCIPLES RELEVANT TO PALEODIET Stable isotope partitioning in human tissues is fundamentally governed by metabolic biochemistry. Let us first consider the question of pathways of nutrients. We shall be particularly concerned with the extent to which all ingested nutrients follow equivalent pathways in metabolism. That is, are some nutrients preferentially utilized (or barred from utilization) in particular metabolic pathways? Ingested nutrients are degraded into their constituent components by enzymes in the digestive system. Starches are digested to sugars (largely glucose); proteins are hydrolyzed into amino acids; fats are broken

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

191

down into glycerol and fatty acids. In addition we assimilate small amounts of vitamins and minerals (especially phosphate), which we can disregard in this discussion. The digested constituents are assimilated by the gut and enter

various metabolic pathways. Eventually, some of the atoms from these constituents are used in anabolic processes to form tissues, including the collagen and carbonate of bone.

Respired

and Bone Carbonate

The fate of a particular constituent (glucose, fat or amino acid) can be quite variable depending on the particular metabolic pathways that are available to it. Ultimately, however, one fact is certain: virtually all assimilated carbon atoms leave the human body in the same state: as carbon dioxide respired by the lungs. A small fraction (less than 1%) leave as solid or liquid wastes (urea, degraded blood cells, bile, sloughed skin, etc.). Respired is therefore a sample of all the ingested carbon atoms, weighted according to their proportions in the diet. The respired is in isotopic equilibrium with blood bicarbonate which at 37°C, should be enriched by about 6.4‰ with respect to produced during respiration (Mook et al. 1974). If we assume

that is also in isotopic equilibrium with the carbonate component of bone mineral, and that the isotopic fractionation between this component and dissolved is constant in all humans, then the value of bone carbonate should be a measure of the weighted average of all foods in the diet. This interpretation is in conflict with some published discussions of the source of bone carbonate, which suggest that bone carbonate gives us a sample of “energy sources” (Krueger and Sullivan 1984), by which is meant carbohydrates and fats (e.g., Ambrose and Norr 1993).1 It should be clear from the previous discussion that very little partitioning of carbon sources can occur. Bicarbonate in the blood is derived from essentially all the carbon atoms in the diet, including proteins, in proportion to their abundance in the diet; all foods are “energy sources”. Minor deviations from this arise due to (1) Tissue 1

(Editor’s note:) Secondary sources may not accurately and precisely convey the primary litera-

ture. Krueger and Sullivan’s (1984) original formulation of the model of the relationship between the value of diet and bone apatite carbonate defined the energy component of the diet as including lipids, carbohydrates and proteins that are not used for tissue synthesis. Krueger and

Sullivan actually proposed that most, if not all carbon, from all dietary macronutrients, including protein, is used for energy metabolism. Ambrose and Norr (1993) did not explicitly include protein in the energy component of the diet in describing the original Krueger and Sullivan model (mea culpa). However, Ambrose and Norr (1993) and Tieszen and Fagre (1993) both concluded that carbonate carbon fit the linear mixing model because both controlled diet experiment programs clearly demonstrated that all dietary macronutrient fractions, including protein, served as energy sources and all contributed to the isotopic composition of bone carbonate. We hope this note prevents further perpetuation of incomplete representation of the original model.

192

HENRY P. SCHWARCZ

growth: During active growth in immature individuals, some of the carbon intake is channeled into growing tissues and is not immediately balanced by catabolic oxidative processes. This may preferentially deplete the carbon pool

to the extent that of the added tissue differs from that of the total diet. On the other hand, mature humans are at steady state, that is, the amount of C atoms coming in as food is equal to the number being excreted as and other waste products, As long as these excreta are not isotopically fractionated with respect to total diet, the of each of them should be equal to that of the food. (2) Fractionation in excreta: The C content of sloughed skin, urea, sweat and feces may be isotopically fractionated with respect to the diet. In order to maintain isotopic mass balance (input) (output)), the of respired might be offset from that of total diet. This effect is however quite small due to the small proportion of C that is not excreted as and can be completely neglected here. Therefore, to a very good

approximation,

of respired

I shall discuss how this relates to

should equal

of the total diet. Later

of bone carbonate.

Proteins and Amino Acids Paleodiet studies have focused on the analysis of collagen, due to its ability to survive in ancient bone. Like all proteins, collagen is composed of amino acid (AA) units present in relatively constant proportions characteristic of the specific protein. The isotopic composition of a sample of collagen is the weighted average of the values of each of the constituent amino acids. There are two important classes of AAs: essential (eAA) and nonessential (neAA). Non-essential amino acids are capable of being synthesized within the body from other biochemical precursors. For example, glycine is derived from serine (another amino acid). For many AAs there is more than one possible biosynthetic pathway. Glycine can also be synthesized from

ammonia, and methylene tetrahydrofolate. The value of a non-essential amino acid is determined by the isotopic composition of the substrates used in its synthesis but, especially where a large pool of substrate is present, the product AA may be isotopically fractionated with respect to its precursor due to kinetic isotope effects. Hare et al. (1991) found large differences between constituent AAs of proteins. Although neAAs can be synthesized by the body, they need not be. Like eAAs, they are made available to the body through digestion of proteins, and can directly enter the blood stream from the liver whence they can be transported to cell-sites for assembly into proteins. Later we shall see that this has significant implications in our understanding of paleodiet. Essential AAs cannot be synthesized by the body. Therefore their isotopic composition is directly inherited from the ingested foods. Indeed, AAs which are essential to humans are generally also essential to the herbivores from

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

193

which humans obtain much of their protein. Therefore, the carbon skeleton of these AAs is passed on intact from the plants on which the herbivores fed. The essential AAs are probably not isotopically fractionated during these transfers, due to the relatively high molecular weights of the AAs.

Biochemistry of Collagen Collagen is principally composed of three non-essential AAs (Table 1): glycine, proline, and hydroxyproline, the carbon atoms of which are derived from proline. eAAs contribute 22.3% of the carbon atoms in the collagen structure. The neAAs making up the remaining 78% of carbon atoms can be built up from almost all the dietary sources of carbon. In the following discussion we shall consider the isotopic consequences of some biochemical constraints on the way in which neAAs are generated indigenously (inside the animal which we are studying). We should be aware, however, that a significant amount of these neAAs are also generated exogenously, and consumed “preformed” from the diet. There is clearly a strong tendency in some dietary situations for preferential incorporation of such pre-formed, exogenous neAAs

to be incorporated into newly formed protein (including collagen). This has been impressively demonstrated by the animal-feeding experiments of Ambrose and Norr (1993) and Tieszen and Fagre (1993). Later we shall discuss the isotopic consequences and causes of this phenomenon. For the moment we shall consider only the consequences of endogenous neAA synthesis to the isotopic relationship between diet and collagen and, in particular, possible isotopic differences between carnivores and herbivores.

Metabolism of Fats: Barriers to Collagen Synthesis Part of the diet consists of fats, which are triglycerol esters of fatty acids (FAs). The FAs from digestion of ingested fats can be metabolized in a variety

of pathways. Fragments of the original FAs are preserved in these processes and can be utilized in the biosynthesis of other molecules. It is important to note that, during metabolism, almost all FAs are broken down into two-carbon units. The only exceptions are FAs with odd numbers of carbon atoms; these are relatively rare in the diet. It can be shown further that there is a partial barrier to the incorporation of FA-derived carbon into the amino acids which constitute collagen. Let us consider, first of all, the synthesis of glycine. As noted earlier, some glycine can be synthesized (in the liver) from ammonia and methylene tetrahydrofolate. One of the two carbon atoms in the product is derived from and therefore could come from oxidation of FAs. Most of the body’s supply of glycine is however synthesized from serine, also utilizing methylene tetrahydrofolate. Most of the C atoms in serine come from 3-phosphoglycerate, which is derived from glucose, as a part of the glycolytic pathway. However, glucose

194

HENRY P. SCHWARCZ

cannot be synthesized from FAs, because the 2-carbon units produced by metabolism of FAs cannot be used in the glycogenic cycle. This effectively blocks the entry of carbon atoms from FAs by this pathway. FAs are, however, metabolized to acetyl-coA which enters the citric acid cycle. They can thus transfer their carbon atoms to both glycine and proline, but with relatively low yield relative to glucose. The principal significance of this process (acetyl-coA production and subsequent cycling) to the animal is the production of energy, through the oxidation of the FAs. These mechanisms for the synthesis of glycine present a partial barrier to the movement of FA carbons into this molecule, the most abundant AA in collagen. On the other hand, proline is synthesized from -keto glutarate which can be freely derived from either carbohydrates or FAs; thus the synthesis of proline does not present a barrier to entry of FA-derived carbons into collagen. Similar arguments can be made for the synthesis of the other neAAs in collagen. Alanine, for example, is derived from pyruvate and thus dominantly from carbohydrates (via glucose). For each of the neAAs there are “preferred” sources of carbon atoms, with the dominant source coming through the glycolysis pathway and therefore deriving initially from glucose. Since FAs cannot be converted to glucose, the result is that there is a partial barrier to the utilization of FA-derived carbon atoms in the synthesis of many of the AAs. The story is further complicated by the existence of multiple pathways for the biosynthesis of most neAAs. Nevertheless, some pathways are preferred over others and there is no question that fat-derived C atoms are significantly

blocked from utilization in synthesis of some AAs. This has a distinct isotopic consequence for the composition of proteins synthesized from these amino acids. To the extent that FAs differ isotopically from the remainder of the diet, this difference is under-represented in proteins partly built up of AAs that cannot be synthesized from an FA precursor. We could represent this difference by a ratio the fraction of FA-derived C atoms in the AA pool from which proteins are made, divided by the fraction of FA-derived C atoms in the total diet.

where C(x) is the atomic fraction of C in the AA pool or the diet that is derived from a given nutrient x: FA = fatty acids; CA = carbohydrates and PR = protein. Then represents the extent to which FAs (as compared to carbohydrates or proteins) are barred from synthesis of AAs (which we could call “antirouting”). From the previous discussion it is clear that since metabolism of both carbohydrates and proteins leads to generation of glucose which

can be used to make any neAA, whereas FA metabolism yields only 2-carbon units which enter the citric acid cycle as acetyl-coA, but do not lead to the

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

195

formation of glucose, one of the principal precursors of glycine. One consequence of this preferential routing will be seen when we discuss the trophic level effect in apatite-collagen fractionation. We would also expect to find that amino acids which could be derived in part from FA precursors would exhibit lower ratios with respect to remainder of the AA population. Hare et al. (1991), studying values of individual AA’s separated from pig bones, found that proline and hydroxyproline were –2.0 and –1.4%o with respect to the total collagen; only valine was lighter than these two AA’s (–6.2%o w.r.t. total collagen). In their studies of mice fed on isotopically controlled diet, Tieszen and Fagre (1993) found very little change in of collagen as a result of changing the of dietary lipid (from to This is also consistent with the notion that lipid is partially blocked from contributing to synthesis of some amino acids. However, as we shall see later, their observation may be more closely linked to the preferential use of exogenous amino acids in protein synthesis (“routing”).

Metabolism of Protein-Derived Amino Acids Ingested proteins are enzymatically broken down into amino acids in the gut, the AAs are absorbed from the intestines and transported to the liver. From there they can be directed to growing tissues via the into the blood. AAs are also constantly entering the blood as a result of enzymatic degradation of pro-

teins in the body. Each amino acid has its own metabolic pathway but the ultimate products are limited to two categories of biochemical substances. Some AAs including leucine and lysine are metabolized to ketones (e.g., acetoacetate and acetyl coA). Their C atoms therefore have fates similar to those derived from FAs. The remainder of the AAs are converted into various components in the metabolic cycle of carbohydrates, including -keto glutarate and pyru-

vate. The carbon skeletons of these AAs are thus ultimately converted to glucose or made directly available to energy production. The carbon atoms of eAAs are also recycled in this fashion. Thus it is possible for C atoms from

one ingested AA to make their way into the synthetic path of any other neAA, leading to “scrambling” of the C atoms from protein sources. During these transformations, the nitrogen atoms are transferred to form glutamate which acts as a carrier of amine groups for the synthesis of new nonessential amino acids, or in the transamination of the carbon skeleton of

some eAAs. Nitrogen Balance and Growth When an organism is in steady state, the rates of production and consumption of biomolecules are equal, and there is no net addition to the body

on a time scale longer than a few hours (e.g., temporary storage of glucose as glycogen or FAs as fat globules). This is an adequate description of a mature

196

HENRY P. SCHWARCZ

adult but does not describe very well an animal which is actively growing, or is in certain disease states. Before reaching maturity, animals are actively synthesizing new protein which must accumulate in various tissues in order to allow the animal to increase in size. Some nutrients (carbohydrates, fats) are turned over faster while the there is an increase in the mass of others, especially proteins needed to construct cells, and calcium and phosphate in the form of bone mineral. A similar situation prevails when the animal is recovering from trauma or some diseases: again, repair of injured tissues leads to a higher demand for nitrogen. The need for additional protein in a growing animal generates a nitrogen deficit which is met by assimilation of larger amounts of amino acids. This condition is referred to as a positive nitrogen balance. The amino acid needs of animals in various types of growth- or disease-related stress, are different from those of mature animals. The additional needed amino acids are partly derived in an intact state from ingested protein. This is inevitably true for the essential amino acids, but under conditions of positive nitrogen balance, neAAs will also tend to be conserved and utilized directly. This is because the normal biosynthetic pathways for creation of these AAs are not adequate to supply the needs of the growing organism. This condition of dependence on external supplies of neAAs has been identified in immature humans and pigs (Ball et al. 1986; Hunt and Groff 1990). Three categories of amino acids have been recognized by Hunt and Groff: 1. Indispensable AAs: eAAs which may be assimilated whole (e.g., lysine, threonine) or as skeletons to be transaminated (e.g., histidine); 2. Condition-

ally indispensable AAs: These can be interconverted from other amino acids already present, such as cysteine and tyrosine; and 3. Acquired indispensable AAs: These may behave as essential AAs in immature animals, in states of metabolic disorder or during severe stress. Hunt and Groff (1990) identify cysteine and tyrosine in the latter category, but it appears that other amino acids can also be conditionally indispensable in this way. For example, Ball et al. (1986) find that proline and possibly arginine are essential AAs in young pigs. In other immature animals not yet investigated it is possible that yet other AAs may be conditionally indispensable.

These categories of variable dispensability should be contrasted with the eAA/neAA distinction. The latter specifies which AAs the animal can, in principle, synthesize from other precursor species (glucose, acetyl-coA, etc.). If a neAA is also conditionally indispensable, then at some times (e.g., in immaturity) an animal will require an external supply of the particular AA even though at other times (as a healthy, mature animal) it is able to synthesize internally its entire requirements of this AA. When an external supply is needed, this neAA will behave like an EAA. Isotopically, we would expect to see direct “routing” of this AA to growing tissues in an immature animal, whereas in a mature animal the animal would synthesize the same AA from a

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

197

pool of nutrients whose isotopic composition was derived from the total diet (except for FA barriers as described above).

Preferential Use of Exogenous neAAs: Inhibition of AA Synthesis Even in a mature, healthy animal at steady state, some fraction of all of the imported neAAs are also used in protein synthesis. From the above arguments we might expect that this would be a small fraction except for in the state of conditional indispensability. Experimental studies showed, however

that protein synthesis in normal, well-fed mice and rats showed a strong preference for utilization of pre-formed neAAs assimilated from the diet (Ambrose and Norr, 1993; Tieszen and Fagre, 1993). Thus, even though local synthesis of AAs in specific cells would be triggered by the needs of protein synthesis in the cell, there is clearly a countervailing tendency to make use of externally derived AA’s. Thus, osteoblasts, which are presumably capable of synthesizing amounts of glycine and proline adequate for use in collagen synthesis, nevertheless were shown to synthesize collagen that was strongly labeled with exogenous neAAs.

We can understand the origin of this effect if we look more closely at the biochemistry of AA synthesis. To a first approximation one could account for this preferential use of pre-formed AA’s on the basis of simple energetic arguments. In order to synthesize an AA from its constituents, it is necessary to expend some energy. If the same AA is already available in the extracellular fluid (ECF) then it would clearly be energetically advantageous to use this source of the AA in preference to synthesizing it within the cell. But this argument does not identify the mechanism by which this selective use of exogenous AA’s is achieved. This can be accounted for through the inhibition of the synthesis of the AA by one of two possible mechanisms. First, it is known that synthesis of all AA’s occurs through a series of

steps, etc., each catalyzed by a distinct enzyme, and each of which transfers a product to the next step until, at the last step, the prescribed AA is produced (Fig. 10.la). Once any material has passed through called the “committed step”, the remaining steps must act to convert their respective substrates unless some alternative pathway for the product is available. It has been found that, in general, the action of the enzyme controlling the committed step is inhibited by the presence of significant intracellular concentrations of the end product, in this case, an amino acid (Umbarger, 1978). The synthesis of some AA’s occurs in branched pathways in which more than one product can be formed following the committed step. In this case the inhibition may occur at a later step (Fig. 10.lb); other more complex modes of inhibition can also occur. A second form of inhibition could occur whereby the product AA acts to inhibit the synthesis of the enzyme associated with the committed step in

198

HENRY P. SCHWARCZ

Figure 10.1. Schematic diagram showing inhibition of synthesis of amino acids: a) single chain: inhibition occurs when enzyme controlling committed step is inhibited by increasing concentrations of product branched chain: inhibition of by increased concentration of occurs at a post-branching step while permitting continued production of product of other branch In general, each step is controlled by a single enzyme.

AA biosynthesis. At an ever deeper level it is possible that high concentrations of the product AA can inhibit expression of the gene for formation of the RNA needed for enzyme synthesis. This form of suppression of gene action over many generations could lead to loss of the gene, causing the AA to become an essential amino acid.

These inhibitory mechanisms all have similar consequences for the isotopic composition of animal proteins. We would expect that when a given neAA is present at sufficiently high concentration in the animal’s diet, the majority of molecules of this AA that are incorporated into protein will come from the diet, while a smaller fraction will be synthesized in the cell. This mechanism would generate the phenomenon which has been described as selective transfer (“routing”) of AA’s from the diet to newly synthesized protein, as reported in the animal feeding experiments. This should be contrasted with an earlier model which has been widely used in the isotopic paleodiet literature, in which it is assumed that all assimilated carbon atoms were equally available to contribute to the synthesis of neAAs, and therefore to the isotopic labeling of these AA’s. This is sometimes referred to as the “scrambled egg”

isotope model (e.g., Ambrose and Norr 1993), and was described by this author (Schwarcz 1991) as the linear mixing model. We shall now discuss the isotopic consequences of these various phenomena and how they would be

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

199

detected in paleodiet data: lipid antirouting; inhibition of neAA synthesis; and linear mixing.

IMPLICATIONS FOR ISOTOPIC PALEODIET STUDIES

Bone Carbonate as an Estimate of

of Total Diet

Ambrose and Norr (1993) and Tieszen and Fagre (1993) have shown of carbonate in bone apatite is the “most accurate measure of the whole-diet composition” (Ambrose and Norr 1993: 28). The actual of total diet is related to that of apatite by an isotopic offset (fractionation) which Ambrose and Norr estimate to be Other estimates range from for small mammals on controlled diets (DeNiro and Epstein 1978) to for large herbivores on natural diets (Lee-Thorp et al., 1989). The origin of this offset is of some concern to us here. We can only use as a measure of total diet if we know and also know that this fractionation is a constant, at least for a given species, and does not itself depend on the quality of the diet. Let us then consider what factors might determine the value of The carbonate molecules entering the structure are presumably derived from dissolved inorganic carbonate species in the cell. At the pH of intracellular fluids, the dominant species would be Most of this bicarbonate is produced intracellularly by hydrolysis of metabolic although it is likely that can also be transported through the cell membrane from the ECF by appropriate ion-pumps. In any living organism, respiration is constantly producing a large amount of carbon dioxide. Even though almost all of this oxidized carbon is transported away from the cells as this species is continuously enriched in with respect to the primary by the equilibrium fractionation factor of Fractionation between bicarbonate and bone carbonate is likely to be quite small by analogy with the C isotopic fractionation between carbonate minerals (aragonite, calcite) and dissolved (e.g., Grossman and Ku 1986). The sum of these two fractionations is equal to or slightly smaller than the observed fractionation between

that

diet and carbonate of between 9.5 and Given our present uncertainty in the magnitude of the fractionation between dissolved and crystal-bound carbonate species, the observed fractionation seems to be well explained by this model. Nevertheless, there is some question as to the constancy of both between species and within one species. We do not expect that will vary much as a function of diet since blood bicarbonate equilibrates with the entire pool of respired carbon, whatever its source. We cannot be so certain that whatever mechanism causes this fractionation is itself independent of such factors as rate of bone growth, remodeling, stress, etc. Given the very large magnitude of for large herbivores, for other

200

HENRY P. SCHWARCZ

mammals), we cannot be sure that environmental effects on might not be correspondingly large, and could lead to equivalent errors in our estimates of

The Trophic Level Effect in Bone Carbonate Several authors, beginning with Krueger and Sullivan (1984), have noted a smaller fractionation between collagen and apatite in carnivores as compared to herbivores. These authors presented a model to account for this, and the observation was confirmed in studies by Lee-Thorp and van der Merwe (1991) of populations of carnivores and herbivores that differed widely in their intake of and foods. Here 1 shall show how this effect might be accounted for as a consequence of partial blocking of AA synthesis from lipids. Another factor which has been suggested as leading to this shift arises from the fact that most of the herbivores whose collagen and carbonate has been studied were ruminants. Metges et al. (1990) have noted that bacteria present in the stomachs of some ruminants generate significant amounts of methane, which has a very low value (typically for methane produced by in vitro fermentation of a substrate). As a result, the of respired is significantly heavier (l3C-enriched) compared with total diet. This could account for the apparent increase in through an increase in of the apatite. This is a possible alternative explanation for an apparent trophic-level effect, one that would not be useful for comparison between carnivores and omnivores (partial herbivores) unless the latter were also converting significant amounts of assimilated carbon into methane. Comparative analyses of carnivores and herbivores, however, do not appear to support this model. Lee-Thorp et al. (1989) studied populations of carnivores and herbivores (several species of each) from the southwestern Cape region of South Africa. This region is dominated by a flora which would have been the nutrient base for the herbivores and therefore also the carnivores which were eating them. If the herbivores had been significantly enriched in as a result of methane production then we would expect that they would exhibit significantly higher values than the carnivores. Figure 4 in their paper shows that there is no significant difference in of the carnivores and herbivores, whereas the collagen of the carnivores is significantly enriched in Most accounts of the larger in carnivores have attributed this effect to higher proportion of lipids in the diet of carnivores. This arises because carnivores obtain all or most of their nutrition from the flesh of other animals, a significant part of which is composed of lipid. By contrast, lipids make up a much smaller fraction of the total carbon pool in the diet of herbivores, particularly ruminants which get much of their energy from digestion of cellulose. Humans who selectively use seeds and grains as food sources obtain a

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

201

rather higher proportion of their carbon intake from plant-derived lipids, but still a much smaller fraction than would be obtained from animal flesh. Lipids from both plants and animals are isotopically lighter than the whole plant by 3 to (DeNiro and Epstein 1978). The argument was made by Sullivan and

Krueger (1984) that lipids would contribute selectively to the “energy food” portion of the diet, which is the source of carbonate molecules. This would therefore lower

and therefore decrease

We can now appreciate that this explanation is incorrect, because the “energy food” for an animal is all of its diet and not just carbohydrates and lipids. Therefore we should not expect any selective offset due to the presence of lipids in the flesh of herbivores. Indeed, in general, the average of total consumable herbivore tissues (flesh, lipids, etc.) is very close to that of the diet, and we might not expect any difference in the isotopic composition of the collagen or carbonate of a consumer of pure plants as opposed to a consumer of the flesh of eating herbivores. We must seek elsewhere for the cause of the trophic level effect on

As noted above, there is a biochemical barrier to the use of fatty acids in the synthesis of some AAs, which I expressed quantitatively as the ratio the proportion of FA-derived C atoms that can be converted to AAs. Although we do not know this precise ratio (because of the multiple biosynthetic paths for

all neAAs), we can assume some value a little less than 1.0. For the purposes of this discussion we can initially assume that the offset between collagen and the total AA pool in the body is some constant, say the same as the fractionation between collagen and diet usually assumed for animals in the wild. Figure 10.2 shows schematically the effect of the inhibition of AA synthesis from FAs, and contrasts the values of a herbivore and a carnivore both subsisting on a population of plants Figure 10.2a shows how of collagen (co) and apatite carbonate (ap) are related to

of diet in a pure herbivore. Collagen is diet, while of apatite is about is about

flesh of a

heavier enriched) than the heavier than the diet. As a result,

Figure 10.2b shows the case of a carnivore consuming the

eating herbivore (e.g., that in Figure 10.2a), but with

(no barrier to AA synthesis from FAs). Now the carnivore obtains all of its nutrient intake from a mixture of animal proteins (PR) and lipids (LI), whose values differ as shown in Figure 10.2b. The outcome is the same as in Figure 10.2a, since all of the C atoms in both LI and PR components have equal chance of being made up into collagen-building neAAs. In addition, 22% of the collagen is indirectly derived from plants as essential amino acids (which have been recycled through the herbivore’s flesh without being isotopically fractionated). Finally, if we set (Fig. 10.2c), the pool of neAAs available to produce collagen excludes some part of the complement of C atoms derived from lipid. Since these are isotopically lighter, the result is that the AA pool is made somewhat heavier. The fraction of eAAs used to make collagen and their

202

HENRY P. SCHWARCZ

Figure 10.2. Schematic diagram showing how restricted conversion of fatty acids to amino acids influences the fractionation between collagen and of bone apatite: LI = lipid component, PR = protein, T = total isotopic composition; component of apatite, a) Herbivorous diet plants only); b) Carnivorous diet, assuming (no barrier to fatty acid conversion to AAs); c) Carnivorous diet, assuming note that carbonate-collagen fractionation is smaller.

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

203

isotopic composition remains the same as in the previous cases. We assume that the fractionation between the AA pool and collagen is still therefore the collagen synthesized from this pool is heavier than in the previous case. On the other hand, of carbonate that is derived from the total nutrient (protein + lipid) is the same as in the herbivorous case (or for the hypothetical carnivore where The net result is that is smaller for the carnivore (with For an omnivore, the magnitude of the decrease would be proportional to the proportion of flesh in the diet. If we knew the actual value of it would be possible in principle to estimate the magnitude of the resultant trophic level effect. It is interesting to note that the trophic shift in appears to be largely or entirely due to a shift in the fractionation between collagen and diet, which increases with increasing proportion of lipid in the diet, whereas remains constant. This is consistent with the data from Lee-Thorp et al. (1989) cited above. Since the lipid fraction is assumed to come principally

from meat, varies with extent of meat consumption. This also shows that we should not assume a constant value for (e.g., as has been commonly done in many paleodiet studies. The lipid anti-routing mechanism leads to a trophic effect on of collagen, although most researchers do not observe a clear offset with trophic level (see, e.g., Schwarcz and Schoeninger 1992). In marine animals, however, a trophic effect of between 1.0 and per level has been observed (Peterson and Fry 1987). To further complicate the matter, preferential assimilation of exogenous amino acids (“routing”) should also lead to such a trophic level effect, as we shall now discuss.

Isotopic Consequences of Inhibition of neAA

Synthesis (“Routing”) As noted above, there is strong experimental evidence to suggest that

endogenous production of non-essential amino acids is inhibited in animals consuming food that provides significant levels of these AAs (Ambrose and Norr 1993; Tieszen and Fagre 1993). These observations would have various isotopic consequences that have been discussed by these authors. They lead to an isotopic paleodiet model in which we expect to find that dietary protein is over-represented in the carbon isotopic composition of all body proteins, and specifically in collagen, a phenomenon which has been also referred to as routing of exogenous AAs to protein. The over-representation of dietary protein would be detected when compared with estimates of contributions to diet by various foods as based on since the latter should give an unbiased estimate (except in those cases where gut methanogenesis might be invoked to account for enrichment of the ECF pool of bicarbonate). Some paleodiet studies, particularly of consumers of plants such as maize, provide an interesting contrast with the expectations of this model. For

204

HENRY P. SCHWARCZ

example, in studies of late Classic and Post-Classic Maya from Belize (White and Schwarcz 1989; White et al. 1994) many individuals exhibited values Assuming a linear mixing model (Schwarcz, 1991), these

would correspond to diets consisting of maize. This should be close to the upper limit of possible maize dependence, since maize is not a complete nutrient and serious nutritional deficiencies would result if even higher levels of maize dependence had occurred (White and Schwarcz, 1989). No values greater than 75% apparent maize use have been recorded. The studies of Ambrose and Norr (1993) on tracing uptake of isotopically labeled dietary constituents showed that a large fraction (up to 80%) of the C atoms in collagen came from dietary protein and not from carbohydrate. Some of the dietary protein of the Belize Maya was presumably derived from the flesh of various herbivores (peccaries, turkeys, etc.) whose bones are found in waste accumulations at sites, most of which are

consumers as shown by isotopic analy-

ses of bone collagen (White and Schwarcz, 1989). A linear mixing model would presumably under-represent the contribution of this labeled component. Therefore, a calculation of the percent in the diet based on a linear mixing model would underestimate the proportion of maize in the diet compared with a “routing model”. Consequently, an even larger fraction of food must be present in the total diet to cause the total collagen to have such high values, since proteins (assumed to be dominantly based) would be over-represented in collagen (Figure 10.3). Indeed, above a certain assumed level of protein routing, the diet of an individual with would appear to consist of more that 100% maize. That is to say that there must be some upper limit to the possible of collagen from a consumer of mixed

animal tissues and plants. At Maya sites such as Lamanai, linearmixing estimates of dietary % appear to be close to the upper limit of nutritional acceptability. Therefore, a paleodiet model based on an assumption of strong inhibition of endogenous neAA production (routing) would seem to be unacceptable for this population. A partial solution to this dilemma could be that a large proportion of the protein-rich foods (meat, eggs) consumed by these people came from animals that were themselves fed a diet. We know that dogs typically share the same diet as humans (Katzenberg 1989; Cannon et al. 1999) and are important components of the diet in some sites (e.g., Cuello: Hammond 1991; van der Merwe et al., this volume). It is unlikely that all the meat consumed by Maya peoples was derived from pure consumers, however, as we have evidence for at least some based animal bones that are presumed to be waste from food preparation. This should a subject of future study to test for the degree of domestication (and consequent feeding on maize) of meat-supplying animals such as turkeys. Notwithstanding the possibility that some based meat was consumed, the existing evidence for at least some based meat in the diet of Mesoamerican (and other) peoples suggests that a routing model may lead

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

Figure 10.3. Schematic graph of %

in diet as a function of

205

of collagen, for a population

consuming a mixture of plants, and flesh of consuming herbivores. If routing occurs, then a smaller proportion of the C atoms in collagen are derived from the foods (for any given ratio in diet) and a given value therefore corresponds to a higher proportion of food in the total diet. For sample a a linear mixing model gives 20% whereas the routing model would give 42% Upper limit of possible values is lower (more negative) for routing models: for sample linear mixing gives 60% while no solution is possible for this routing model.

to surprisingly high estimates of consumption levels. Therefore, in spite of the strong evidence from two sets of laboratory experiments that inhibition of AA synthesis can occur in laboratory animals, it is possible that these effects may be less pronounced in some human populations to which we would like to apply these data. I have interpreted these laboratory data as indicating that endogenous AA synthesis is suppressed by the presence of relatively high concentrations of the respective neAAs in the ECF. From this we might then infer some general expectations about the manifestation of this phenomenon in “wild” populations. Specifically, we might expect to see these inhibition effects most strongly exhibited in populations which were consuming high levels of meat-derived protein which would provide a balanced supply of all the neAAs required for endogenous protein synthesis. These people (or animals) would display values strongly biased towards the values of the protein source, while sources of carbohydrate and lipid would be significantly underrepresented. By contrast, populations in which protein was largely derived from plants and was generally in low supply would approach the linear-mixing model in which there was a random assignment of C atoms from all foods,

206

HENRY P. SCHWARCZ

with the exception of lipids. But, since lipids generally make up a comparatively small proportion of plant-derived foods as compared to animal tissues, the effects due to partially blocked utilization of lipid-derived C atoms might also be negligible. These effects should be studied both experimentally and by study of natural populations. Interestingly, it appears that the routing of exogenous neAAs should also contribute to the previously discussed trophic level effect on The effect is quite analogous to that which was described for the blockage of lipid uptake into AAs (Figure 10.4). Let us assume that an individual lives in a dominant region in which they have the option of eating either plants directly, or the tissues of consuming herbivores. We assume that of the total assimilable tissues of the herbivore is close to that of the plants themselves. As before (Figure 10.2), and by mass balance,

Figure 10.4. Effect on apatite-collagen isotopic fractionation due to inhibition of amino acid production and preferred use of exogenous amino acids. Carnivore and herbivore, both based on plants, have similar bulk isotopic composition of total edible tissues (T), leading to similar for apatite carbonate (AP). Collagen (CO) of carnivore is more enriched in than that of herbivore, because of preferential utilization of amino acids derived from protein (P) of herbivore flesh in construction of carnivore’s proteins. assimilated carbon.

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

207

would be higher than of the total diet. A pure carnivore would obtain the majority of the neAAs needed for protein synthesis directly from herbivore flesh, and this protein would therefore also acquire a value higher than that of the base of the trophic system plants). In contrast, a pure herbivore, eating a diet poorer in protein, would construct its proteins for C atoms derived from all the constituents of the plant foods: carbohydrate, lipid and protein. The resultant protein would be lower in On the other hand, since the total C atom populations from both herbivore tissues (protein + lipid) and from total plant food are closely similar, therefore of the herbivore and carnivore would be similar. The net result would be a smaller in carnivores than in herbivores.

CONCLUSIONS Some aspects of the biochemistry of metabolic processes affecting nutrients appear to have significant consequences for the expected behavior of stable carbon isotopes as tracers of diet. Specifically, we have seen that

the simple model of a total “scrambling” of carbon atoms during endogenous biosynthesis is inconsistent with the expected pathways of some nutrients, whereas other isotopic records in ancient human tissues can be adequately accounted for by this model. The applicability of the linear-mixing model is seen most prominently in the interpretation of the of bone apatite which has been shown to represent the total diet, rather than being derived from “energy foods”, as was previously proposed by some authors. Although should represent total diet, the isotopic fractionation between this component and total diet appears to be somewhat variable, suggesting that more definite knowledge about this fractionation is needed if we are to use as an index of total dietary values. On the other hand, the “scrambled” model of carbon sourcing does not seem to be applicable when we consider the metabolic fate of fatty acids. We find that there are partial barriers to the movement of FA-derived carbon atoms into the synthesis of proteins. This partial restriction leads us to expect a trophic level effect in the fractionation between collagen and bone apatite or respired of which apatitic carbonate is a sample. The magnitude of the fractionation depends on two separate fractionation factors which cannot be disentangled by analyses of bone samples alone. The other example presented of a non-scrambled distribution of isotopes involves the synthesis of collagen. For a mature animal at steady state, we might expect extensive atomic scrambling in the sense that most of the nonessential amino acid content of this protein (78% of its carbon atoms) can be synthesized from the general pool of glycogenic substrates that arise from metabolism of all sugars and fats, although the pathway from fats is restricted

208

HENRY P. SCHWARCZ

as noted above. Very convincing experimental evidence for routing of amino acids to sites of protein synthesis has now been obtained by two studies, on rats and mice respectively (Ambrose and Norr 1993; Tieszen and Fagre 1993). I have shown how these results may be accounted for by inhibition of endogenous synthesis of amino acids when the amino acid is already present at or

above some critical level in the cell. We would expect to observe this inhibition in animals receiving adequate amounts of protein in their diet, which

would be broken down in the liver into the respective amino acids and made available to the cells. This would result in bone collagen (and other surviving proteins) giving analyses that were strongly biased towards the values of the protein sources (commonly the flesh of animals), and less representative of the total dietary intake of carbon. On the other hand, data from some isotopic paleodiet studies of maize consumers suggest that the isotopic composition of bone collagen could not have been strongly biased in favor of based herbivore flesh, if this was indeed a major source of dietary protein. I suggest that in natural human populations living at low protein intake levels, inhibition of endogenous AA synthesis is not as active as in the laboratory studies described above, and the isotopic composition of bone collagen comes closer to the linear mixing model, or at least lies somewhere between these two models. These are only a few specific examples of paleodietary consequences of biochemical pathways. Paleodiet researchers should probably try to enlist the aid of metabolic biochemists in a search for other possible consequences of differential metabolic pathways, internal recycling of metabolites, etc. Furthermore, many of these problems will become clearer as we begin to have access to isotopic analyses of individual amino acids or even specific carbon atoms at sites on individual AAs.

ACKNOWLEDGMENTS The author has benefited greatly from discussions with Professors Karl Freeman, Evert Nieboer and Stephanie Atkinson on biochemistry and nutrition. Boaz Luz clarified my ideas about carbon isotope fractionation in blood. Research was supported by a grant from the Social Sciences and Humanities Research Council. The author thanks Shannon Coyston, Lori Wright, Chris White and Stanley Ambrose for useful comments and discussions.

REFERENCES Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relation of the carbon isotope ratios of whole diet and dietary protein to those of collagen and carbonate, In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone—Archaeology at the Molecular Level. Berlin, Springer-Verlag: 1–37.

BIOCHEMICAL ASPECTS OF CARBON ISOTOPIC PALEODIET STUDIES

209

Ball, R.O., Atkinson, J.L. and Bayley, H.S. 1986 Proline as an essential amino acid for the young pig. British Journal of Nutrition 55: 659–668.

Cannon, A., Schwarcz, H.P. and Knyf, M. 1999 Marine-based subsistence trends and the stableisotope analysis of dog bones from Namu, British Columbia. Journal of Archaeological Science 26: 399–407.

Chisholm, B.S., Nelson, D.E. and Schwarcz H.P. 1982 Stable-carbon isotope ratios as a measure of marine versus terrestrial protein in ancient diets. Science 216: 1131–1132. DeNiro, M..J. and Epstein, S. 1978 Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta 42: 495–506. Grossman, E. and Ku, T.-L. 1986 Oxygen and carbon isotope fractionation in biogenic aragonite: temperature effects. Chemical Geology (Isotope Geoscience Section) 59: 59– 74. Hammond, N. 1991 Cuello: an Early Maya Community in Belize. New York, Cambridge University Press. Hare, P.E., Fogel, M.L., Stafford, T.W., Jr., Mitchell, A.D. and Hoering, T.C. 1991 The isotopic composition of carbon and nitrogen in individual amino acids from modern and fossil proteins. Journal of Archaeological Science 18: 277–292. Hunt, S. and Groff, J.L. 1990 Advanced Nutrition and Human Metabolism. St. Paul, West Publishing Co. Krueger, H.W. and Sullivan, C.H. 1984 Models for carbon isotope fractionation between diet and bone. In Turnlund, J.R. and Johnson, P.E., eds., Stable Isotopes in Nutrition. ACS Symposium Series 258. Washington, DC, American Chemical Society: 205–222.

Lee-Thorp, J.A. and van der Merwe, N J. 1987 Carbon isotope analysis of fossil bone apatite. South African Journal of Science 83: 712–715. Lee-Thorp, J.A., Sealy, J.C. and van der Merwe, N.J. 1989 Stable carbon isotope ratio differences between bone collagen and bone apatite and their relationship to diet. Journal of Archaeological Science 16: 585–599. Lehninger, A.L., Nelson, D.L. and Cox, M.M. 1993 Principles of Biochemistry, 2nd ed. New York, Worth. Metges, C., Kempe, K. and Schmidt, H.-L. 1990 Dependence of the carbon isotope contents of breath carbon dioxide, milk, serum and rumen fermentation products on the value of food in dairy cows. British Journal of Nutrition 63: 187–196. Mook, W.G., Bommerson, J.C. and Staverman, W.H. 1974 Carbon isotope fractionation between dissolved bicarbonate, and gaseous carbon dioxide. Earth and Planetary Science Letters 22: 169–176. Peterson, BJ. and Fry, B. 1987 Stable isotopes in ecosystem studies. Annual Review of Ecology and

Systematics 18: 293–320. Schwarcz, H.P. 1991 Some theoretical aspects of isotope paleodiet studies. Journal of Archaeological Science 18: 261–275. Schwarcz, H.P and Schoeninger, M. 1991 Stable isotope analyses in human nutritional ecology, Yearbook of Physical Anthropology 34: 283–321. Stryer, L. 1988 Biochemistry, 3rd ed. New York, W. H. Freeman. Umbarger, H.E. 1978 Amino acid biosynthesis and its regulation. Annual Review of Biochemistry 47: 533–606.

Tieszen, L. and Fagre, T. 1993 Effect of diet quality and composition on the isotopic composition of respired , bone collagen, bioapatite and soft tissues. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone—Archaeology at the Molecular Level. Berlin, Springer-Verlag: 121–155. White, C.D. and Schwarcz, H.P. 1989 Ancient Maya diet as inferred from isotopic and chemical analyses of human bone. Journal of Archaeological Science 16: 451–474. White, C.D., Healy, P.F. and Schwarcz, H.P. 1993 Intensive agriculture, social status and Mayan diet at Pacbitun, Belize. Journal of Anthropological Research 49: 347–375.

This page intentionally left blank

Chapter 11

“Consider a Spherical Cow . . .”—on Modeling and Diet ROBERT E.M. HEDGES

AND

GERT J. VAN KLINKEN

ABSTRACT This paper explores how models may be developed to account for the relationship between the stable isotope composition of a body tissue of an organism and its diet. The main approach taken is to express this relationship as an explicit equation, or a “DIFF”, and then to show how the values of such a DIFF can be evaluated from published experimental data. These values can be expected to have a much wider meaning than a simple encapsulation of a particular experimental design. As a main example, we show how the values may be used to construct a metabolic model in which the synthesis of non-essential amino acid for collagen construction can be treated. A second example is to show how the evaluation, in terms of diet, of the spacing between collagen and carbonate may be put on a rigorous basis. A second kind of model is briefly treated, based on isotopic mass balance arguments, and it is shown that large isotopic discrimination during methanogenesis in ruminants may account for data trends when comparing herbivores and carnivores. A third class of model is sketched at the level of biochemical flows, where some fundamental points are made concerning points where the isotopic composition of metabolites may be altered. The relevance of this to nitrogen isotopic enrichment is considered. Biogeochemical Approaches to Paleodietary Analysis, edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

211

212

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

1. INTRODUCTION In this paper we explore how palaeodietary studies might benefit from the deployment of models of the fluxes of organic matter in feeding systems. A main aim in palaeodietary studies is the reconstruction of ancient diets on the basis of stable isotope measurements of surviving body tissues. While it is recognized that stable isotopic compositions depend on biochemical processes at the molecular level, a description at the level of diet, or even of comparative metabolism, presupposes a huge amount of simplification. Here we suggest how models may be developed to bridge this gap, and to suggest hypotheses

for future testing. The title, with its type-example from a physics text book, recognizes that a modeling approach is necessarily a risky simplification. And cows are briefly singled out later (Section 3.2.2), because their peculiar digestion can be shown to have interesting consequences at a rather simple level of modeling. Dietary studies, based on stable isotope values of animal tissues and foodstuffs, can be successful only insofar as the complexities of molecular metabolism manage to give rise to relatively simple relationships in what is measured. Many of these relationships have been explored, and have in part been described by observational regularities, and in part summarised by physiological or biochemical explanations. Some questions can and have now been clearly posed: for example: Why is collagen isotopically heavier (in carbon) than most other protein?

What determines the difference in between collagen and bone carbonate apatite (bioapatite) for animals with different diets or digestive systems? How much protein in an animal is made exclusively from protein in its diet? What is the basis of the increase in in protein of animals relative to that of their diet?

Answers to these are likely to involve explanations which invoke models of a more or less explicit and non-trivial nature. Furthermore, as techniques

develop, isotopic data are becoming available from a wider range of molecular types, for examples, individual amino acid species and lipids such as cholesterol, and these will require greater sophistication in interpretation. Modelling demands a careful definition and precise use of terms, and we have tried to define these with greater rigor than is customary. We suggest the shorthand words isodiscrimination and isofractionation to describe respectively the degree to which the rate of chemical reactions isotopically discriminate, and the extent to which the isotopic composition has been altered as a result of isodiscrimination or of a redistribution of components (such as amino acids between different proteins).

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

213

Modeling, as it is Understood in this Paper For clarity, it is useful to distinguish two meanings of the word “model”. “Modeling” can mean to formally represent experimental data; or it can mean to study a system-analogue (the “model”) which itself is a simplified representation of reality. In straightforward cases the two meanings can be so similar as to be more or less equivalent. This paper develops both approaches; that is, a formal representation of the data, here called the Dietary Isotopic Fractionation Function (DIFF) in Section 2 of the paper, and the construction and operation of system-analogues, here called flow-models (Sections 3 and 4). We aim to show that by casting the data into the more formal representation of a DIFF, it may more easily and directly be related to the behaviour of possible

flow-models of the animals’ metabolites. This then helps to bridge the gap between isotopic signals originating from biochemistry and observations from individuals and populations in ecosystems.

Plan of this Paper The main part of the paper consists in first formulating a DIFF for a data set of body component isotopic compositions when fed on known diets, next developing a flow-model appropriate to this data set whose behaviour can then be compared with the DIFF Many of the issues raised in this process help to provide an understanding for the wider questions posed in the introduction. The specific application described here is concerned with the data obtained from controlled feeding experiments on rats and mice as reported by Ambrose and Norr (1993) and by Tieszen and Fagre (1993). Few other data exist in such a relatively complete form. Ambrose’s data in particular were collected to decide the question of the extent to which the carbon atoms of body tissue protein comes exclusively from the carbon in dietary protein (“direct routing”), or from the total dietary carbon (“scrambling”). We first show how these data may be represented in a generalized and mathematically useful form (i.e., as a DIFF, especially in Section 2.4). Next, we describe how a feeding organism can be represented by a flow-model (Section 3), and in Section 4 we demonstrate one which explicitly considers carbon metabolic fluxes, and can be evaluated in terms of the DIFF of Section 2.4. This leads to a quantitative

physiological interpretation (however crude it may be) for the data, in terms of the relative contribution of non-protein in the diet to the body protein carbon.

A simple application of the flow-model suggests why bioapatite carbonate values should directly follow the values for the diet as a whole. Experimentally, this observation seems to be generally confirmed (Ambrose

and Norr 1993; Tieszen and Fagre 1993). In terms of the flow-model, it is

214

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

possible to see under what conditions this relationship might not apply (a particular example is that where significant quantities of methane are also produced, as in ruminants). A DIFF can be written for the measurement of for bioapatite carbonate, and this can be combined with the DIFF for collagen to provide an expression for the collagen—bioapatite carbonate spacing in mammals, in particular showing how the spacing varies according to the diet (see Section 4.3). This goes a considerable way to “explaining” the observation that the spacing differs between herbivore and carnivore. Two further examples are treated in terms of corresponding flow-models and their related DIFFs in Section 5. These are the relationship between dietary and body protein nitrogen and the influence of water stress, and the usefulness of DIFFs for discussing fractionation within ecosystems for both N and C. THE REPRESENTATION OF ISOTOPIC FRACTIONATION BETWEEN DIET AND BODY COMPONENT This section develops a general approach to data representation and describes some of its features.

The Dietary Isotopic Fractionation Function (DIFF) Any quantitative relationship between body isotopic composition and dietary isotopic composition can be regarded as a DIFF, which therefore needs to be flexible, but also must be precise. The cliché “you are what you eat is potentially a DIFF, but only when “you” and “what you eat” are properly defined. It is useful to impose a general form on the DIFF Before doing so, it is easier to present a simple and specific case, which we shall also go on to use to represent part of the data presented by Ambrose and Norr (1993) and Tieszen and Fagre (1993). A DIFF relates a specified body component isotopic composition to the various compositions of a specified and complete set of dietary components. The complete diet must be accounted for in the DIFF, but it can be partitioned in any way that seems sensible; for example, into individual amino acids; or into protein, carbohydrate and lipid, etc. Before proposing a general formulation, we illustrate a DIFF with a simple case, in which we consider bone collagen as the body component, and the diet as containing just two components, protein and non-protein. In the notation, B stands for body values, and D for diet values; suffixes distinguish the particular component specified. Thus stands for the collagen and for the dietary protein and non-protein values. We can write the DIFF as:

“CONSIDER A SPHERICAL COW. . .”—ON MODELING AND DIET

215

Where additional symbols d, f() and F require further explanation: d

F

= Any change in isofractionation to the dietary component resulting from its uptake and subsequent conversion to the body component. It is not directly measurable. = The proportion of carbon in the specified dietary component. The

sum of all F values—two in equation (1)—must add up to 1. f(F) = A mathematical function (which is chosen to be “suitable”) which “weights” the contribution of the protein or the non-protein to the particular body component. In general, f(F) will be a function of F. As with F, the weighting function f() must also be normalized. That is, so that the weighting is fully accounted. The point of the weighting function can be illustrated by considering the two extremes of complete scrambling (for which f( ) = 1), and complete routing

(for which If we choose to make we have constructed a convenient function that can be adjusted to take intermediate values (for this particular DIFF), where may take any values between 0 (where f(F) = 1: scrambling) and –1 (where f(F) = 1/F: direct routing). Later (Section 4) we use an alternative formulation, in which

and in place of w the varying parameter is X. (In this case, X is seen to have a physiological interpretation.) In words, the DIFF equation states that the observed collagen values are determined by both the dietary protein (with additional isofractionation), and the non-protein in the diet (with a potentially different isofractionation), by an amount that may vary with the fraction of each component in the diet.

Note that if

and

were both set equal to

and if the f(F) were set

= 1, this would re-state the original DIFF “you are (i.e., your collagen is) what you eat (protein and non-protein) plus five permil”. We discuss how this equation represents the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data set in Section 2.4. Here we go on to develop a more general expression with wider application and generalized notation.

The full generalization can be written as a set of equations of the type:

It may be useful to reiterate and expand the more formal definition of the terms.

216

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

The value of the j-th component of any set of body components. For example, bone collagen; bone-apatite carbonate; total body lipid, total body glycine; muscle glycine; etc. Each particular choice will have its own distinct equation. The set need not be complete— unlike D. The value of the i-th component in the diet. This diet must be summed over all components, so that all the possible sources are included. For example, it might be protein, carbohydrate and lipid; protein and non-protein; or broken down into much greater detail. The value of the isofractionation incurred when the atoms of dietary component i (i.e., of are used to construct body tissue component The proportion of carbon (or nitrogen) by weight in the i-th component of the diet. F is dimensionless, between 0 and 1, and is constrained by the normalization condition: which says that all the diet components have been included. The mathematical function of F, and acting on F, which weights the dietary contribution away from mass proportionality when atoms from are used to construct body component (shortened to f()). f() is constrained by the normalization condition: which says that there is no loss or gain overall in weighting the diet components. Although this formulation appears complicated at first sight, it seems to be the simplest possible set of equations that can usefully represent experimental data. We use bold type when discussing any non-specific set of components. The advantages of this approach are: (1) The use of the set of equations forces one to be explicit about what measurements are being related together, and to what part of the diet.

(2) The task of reconstructing the diet can be defined as that of finding D and F, by first evaluating or estimating the matrix com-ponents and and applying these values to measurements of B. (3) The form of the DIFF is not committed to data for any particular set of components in diet or in body tissue.

From the experimental point of view, two important features need emphasis. First, the diet can be split into as many components as are appropriate to the problem, provided the list is complete. Secondly, the molecular level at which the components are categorized, either for the diet or for the

body parts, is entirely open to choice, within the bounds set by the experimental data. Ideally, an extensive set of DIFFs could be set up at the molecular level (on individual amino acids, for example), from which the behavior at the level of proteins could be derived. However, the data to do this are simply

“CONSIDER A SPHERICAL COW. . .”—ON MODELING AND DIET

217

not available. In fact a DIFF implies considerably more experimental values

than has been common; in particular, the number of values for and increases as the multiplication of the number of dietary components with the number of body components analyzed. It is probably most useful, at present, to work at the level of molecular categories: protein, carbohydrate and lipid, since these tend to have relatively stable isotopic relationships. Therefore we could envisage a reasonably standard formulation of the DIFF as follows:

where PRO, CHO and LIP refer to protein, carbohydrate and lipid. Equation (4) refers to collagen. Another equation would refer to, for example, bioapatite carbonate. While the six unknown values for d and f may seem daunting, present usage corresponds to the simplifying approximation of assuming all the d values are the same, and of putting all f() = 1. Evaluating the DIFF at least makes these assumptions explicit. Once evaluated, d and f() values should hold over a usefully wide range of fields of study. However, any DIFF does have implied assumptions. For the type formulated here—proposed because it seems the simplest available—it is only useful if d and f() can be regarded as basically stable entities, and having some degree of generality. It relies on a reasonably sensible partition of the diet into components, and choice of body-part components to relate to them. For example, attempts to evaluate d and f() for body protein in terms of a diet partitioned as carbohydrate and non-carbohydrate will very likely to produce incoherent results. To this extent, the use of a DIFF does already imply some kind of “model” and, seen in these terms, must be chosen to be consistent with physiology.

Diet Reconstruction and the DIFF Here we evaluate what is implied in diet reconstruction. From equation 3,

The task of diet reconstruction is to solve this equation for values of on the basis of measurements of Assume that we already know the relevant values of and A simple case, often implicit in past diet reconstructions, is to assume that all where j corresponds to collagen, and that all If we consider the diet

and

to be partitioned into n components, it is specified by 2n – 1 values (the n of and the n – 1 values of and these can only be evaluated from

218

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

measurements of 2n – 1 body components. Therefore even a two component diet, e.g. protein and non-protein, requires three body isofractionation values for its full specification. At present, for archaeological material, it is only realistic to make two measurements for carbon (protein and bioapatitecarbonate). Therefore, even at this simplest level, additional assumptions or knowledge must be incorporated in order to specify the diet. For example, the value of one of the dietary components may be known, or estimated, from other evidence. Independent evaluation from the DIFF for nitrogen might help here, since this can be reasonably approximately treated as a single dietary component, that of all protein, and one body component, that of collagen, values for most proteins are approximately similar, despite the differences in the corresponding values of the amino acids. Therefore for dietary protein is easily evaluated from collagen, provided is known. It is generally taken to be about Knowing the for dietary protein may provide additional information for the of the protein. These considerations show precisely how valuable further measurements of body components might be in helping to specify paleodiets. Possibilities include single amino acids, or lipids such as cholesterol (Hare et al. 1991; Stott et al. 1997). The relevant and values must be determined, and it must be shown that the additional components do give independent information. In any case, it is clear that additional assumptions are always going to be needed in applications to archaeological data. Other constraints may also be applicable—for example, constraining the total diet to be isocaloric (Little and Little 1997).

Physiological Interpretations and the DIFF We have noted that the relevant physiological, or biochemical, description is encapsulated in the d and f() elements. These can be evaluated from body component measurements on organisms with known diets. The simpler case of nitrogen, with only one significant dietary component, has been mentioned in Section 2.2. Measurements of bone collagen give a good representation of body tissue nitrogen as a whole. Thus while

necessarily = 1. Also,

is directly measurable. It is a well known

fact that for many mammals, for nitrogen equals approximately This has been attributed to the animals excretory physiology (Ambrose 1993), and in Section 5.2 we outline a possible model that treats this. While it is apparently valid to take as a constant value for many mammals for many diets, there appears to be a variation in between animals whose

diets differ in degree of water stress, which is presumed to be due to the physiological adaptation of the animal to the environmental stress (Heaton et al. 1986; Ambrose 1991). Therefore for is only approximately a constant, and is altered by a specific set of physiological adaptations. This is a particularly interesting and clear cut case, and shows that measurements of body tissue isotopic composition can also provide indications about past

“CONSIDER A SPHERICAL COW. . .”—ON MODELING AND DIET

219

physiological, and by implication, environmental conditions, as well as about the dietary input. Similar effects might occur under protein deficiency, for example. Since f() has been defined as a function of F, the potential for variation

in dietary component routing with dietary composition is explicitly included. For the carbon “protein routing” question, to which we now turn, the degree to which non-protein carbon in the diet is incorporated in body protein probably (and not surprisingly) depends on the amount of protein in the diet. It might also depend on other features, for example energy expenditure. We hope to show that the DIFF formulation provides an acceptable simplification of the question, in line with the experimental data. Of course real understanding would need to consider individual amino acids; body protein carbon in all essential amino acids must come only from protein in the diet, while nonessential acids need not. They would have different f() functions. DIFFs could be evaluated at this level of analysis once sufficient data on individual amino acids are available. Nitrogen routing becomes an interesting issue at the level of individual amino acids. For example, for threonine, seems to be about (Hare et al. 1991) while is probably If essential amino acids in the diet limit the uptake of non-essential amino acids, is surplus non-essential amino acid nitrogen routed directly to excretion, or does it join a transamination pool? These issues will be addressed in a more focused way through the use of flow-models in Sections 3 and 4.

Evaluating DIFF for Data of Ambrose and Norr (1993) and Tieszen and Fagre (1993) The data sets presented by Ambrose and Norr (1993) and Tiezsen and Fagre (1993) record analyses of rodents fed on diets in whose dietary components have been analyzed. Thus D and F are given for the diets, (generally for

the protein, carbohydrate and lipid components; sometimes in more detail), and have been varied so as to relate to the corresponding change in B. Mea-

surements were made on animals thought to have reached an overall steady state. While there are many other aspects from which the data sets can be

viewed, here we are concerned to fit a two-state DIFF for two measurements: of collagen and of bioapatite carbonate. The approach is outlined here with fuller details in Appendix A. We take collagen first, and use the same notation as Equation (1) in Section 2.1 for the DIFF That is, we describe the combined data sets (six diets from Ambrose and Norr 1993, eight from Tieszen and Fagre 1993) with the equation:

220

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

Using the data set, we attempt to determine values for and which predict values closest to the actual measured values from the diets. We take and as constant over the data set. Essentially, the difference between (predicted) and (measured) is evaluated for all possible combinations of and w, for all the diets, and that combination of and w which minimizes both the mean and the variance of the difference (B(est)-B(meas)) is selected. We find that: and predict all the measured to within a mean difference of less than and an average standard deviation of about This is not necessarily the optimum fit, but it is satisfactorily close. If or are changed by more than from +5 and +2 respectively, the prediction is substantially worse. That generally takes a value shows that dietary protein is more strongly represented than the dietary average in collagen. As mentioned already,

would correspond

to protein only being represented, to the average diet being represented in collagen. From Fig. All.2, it is clear from the best-fit values of f() that rat and/or mouse metabolism is close to, but distinguishable from, direct routing. It is an interesting question, whether other mammals are similar. The D1FF gives an explicit and quantitative form to the degree of routing. This is given a quantitative interpretation in terms of the flows of metabolites in Section 4. Much of the fractionation in can be attributed to increasing the concentration of the isotopically heavy glycine in the synthesized collagen. It is interesting that appears to have a distinctly lower value. This observation may not stand up to more detailed data, but it is consistent with predictions made from the simple flow-model in Section 4.1. It is also in line with the observation that synthesized lipid is generally isotopically lighter than synthesized protein. It shows how the DIFF formalism can raise penetrating questions from quite simple existing data sets. Turning now to the bioapatite carbonate, a DIFF can be written in a way analogous to that for collagen:

where

stands for the bioapatite carbonate measurement. Here, refer to the fractionation between ingested protein (non-protein) and carbonate. In fact, we find that (bioapatite carbonate) in the Ambrose and Norr (1993), Tieszen and Fagre (1993) data set is well described by the equation:

i.e., or, putting f() as then In comparison with collagen, there is complete scrambling of the dietary carbon to bioapatite-carbonate. The DIFF takes an almost trivial form here. The

“CONSIDER A SPHERICAL COW. . .”—ON MODELING AND DIET

221

formulation of is unnecessarily complicated here, but is retained for consistency. Without using the formalism of the DIFF, both Ambrose and Norr (1993) and Tieszen and Fagre (1993) have pointed out from their experimental data

that the carbonate has a constant spacing from the average diet. Since the conversion of protein, or non-protein, to bioapatite-carbonate is metabolically more complicated than, for example, carbohydrate to protein, a simple physiological explanation may not be obvious. However, the interpretation of the DIFF does becomes clear in the light of a very simple flux model in Section 3. Both equations should be applicable to data beyond the immediate data set described above. Ambrose has provided two further sets of body tissue measurements (pers. comm.), which should be predictable from the known diets using the DIFFS with the values given above. The satisfactory agreement shows that the DIFFs are providing a useful description of how isotopic signatures are related to diet, at least for rats.

FLOW MODELS OF FEEDING At its simplest, a dietary system can be represented as below: DIET [D, F]

{ORGAN1SM[B, d, f()]} Excreted and respired components [X]

To a degree, every tissue component B within the organism is part of a pool and is in metabolic exchange with other pools. Any tissue not in

exchange, for example hair, is essentially an excreted component. In the absence of detailed information of fluxes and topologies for each component, overall equations of isotopic mass balance become particularly useful. That is, equations accounting for the integrated total flux for each isotope from input (diet) to output (excretion) of the organism can be written. However, progress is helped enormously if the approximation of a steady state can be used, since less information is needed. Fluxes do not have to be integrated over a life

222

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

history. Such an approximation must be justified by considering the time scales of the various processes involved. If an organism is not gaining weight, or the weight gain is very small compared to the total mass-flow, the system as a whole is in an overall steady state, assuming an unchanging diet, and the inputs are in isotopic mass balance

with the outputs. This situation generally applies, at least to the larger and adult mammals, and leads to a prima facie case for the steady state approximation. Nevertheless, the temporal variation in diet, in excretion, and in many components of the body’s metabolic pools (e.g., free amino acids) may be very rapid (hours or less), while other body tissue components may have long turnover times (e.g., years, for bone collagen). Measurement of the longer turnover time body tissues should be reflecting the economy of the system as averaged over a long part of its history. Any analysis must address this issue by considering whether the time-variations in either diet or metabolic turnover are at such a frequency and magnitude as to invalidate a steady state approximation for the particular measurements.

Overall Isotopic Mass Balance In a steady state, there can be no net isofractionation in the excreted components (by definition); the mass-weighted sum of all the values for the excreted material must equal that of the diet. It so happens that what is

excreted very often is dominated by one component.

Carbon Isotope Mass Balance For humans, about 90% of carbon ingested is excreted as respired The remainder is mainly bacterial and desquamated cells and complex catabolites in the feces, while urine accounts for only 1 or 2%. Therefore most of the isotopic mass-balancing must come from the and assuming that fecal C does not have markedly different values from the mean diet and this appears to be the case from published information on animals, respired should have the same as the mean diet. This does not seem to have been verified experimentally; such data as exist (DeNiro and Epstein 1978; Tieszen and Fagre 1993) on non-ruminants do suggest the possibility that respired is isotopically lighter than the diet. This requires closer investigation. The case of ruminants, where additional excreta are involved, is taken up in Section 3.2.1.

Nitrogen Isotope Mass Balance For humans, 80–90% of ingested nitrogen is excreted as urea. Minor contributions include ammonia and creatinine (in urine), nitrogen compounds in feces, sweat, and loss of protein in skin cells, hair, nails, etc. So, as with C, the

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

223

main excreted product should have a similar value to that of the diet. Measurements of urinary N in relation to the N in diet are very sparse (Steele and Daniel 1978; Sutoh et al. 1987), especially since measurements on lactating animals must take account of the substantial N flux (up to 20% in cows) into the (excreted) milk. Again, they indicate a difference, but again, the evidence is still far too sparse (i.e., essentially results on two pigs only) to provide more than an incentive for better data. These considerations show that the isofractionation found in body tissues cannot be simply related to the isofractionation in excreta. Body tissues do not figure in any overall mass balance based on a steady state. There seems

to be some confusion on this point in the literature, where the body tissue and the excreta are often taken together to provide a mass balance. This way of thinking perhaps stems from an attempt to integrate the isotopic mass fluxes; in which case, the body mass must be balanced against all the accumulated excreta during growth (and put equal to all the accumulated diet). In a steady state, body tissue isofractionation values depend on how different biochemically fractionating steps are coupled together (see Section 3.4). Nevertheless, overall considerations of steady state inputs and outputs enable some interesting and important points to be made.

Any Body Tissue in Direct Equilibrium with the Major Excretum Will Reflect the Isotopic Composition of the Diet as a Whole Any body tissue in direct equilibrium with the major excretum should reflect the isotopic composition of the diet as a whole. This seems to be the case for bioapatite carbonate, which is thought to be in equilibrium with plasma bicarbonate, which itself is in equilibrium with respired In fact the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data sets (among others) show clearly that the bioapatite carbonate differs from total diet (or respired by an amount approximating to the equilibrium isotopic fractionation in the system (Mook 1989):

The equilibrium value at 37°C is according to (Mook 1989), while the measured spacing between diet and breath is between 9 and and the value is used in this paper as a round figure. Whether there is significance

that this is between 1 and different from the equilibrium value is not clear to us at this stage. However, the general correspondence between bioapatite carbonate and total diet seems to hold well. It is an interesting point, whether this is necessarily true; for example, respiration may involve additional isofractionation, such as molecular diffusion in the lungs, perhaps dependent on different physiology, although no direct evidence is known so far to support this.

224

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

For nitrogen, where the measured body tissue is protein, there are more metabolic steps from protein to excreted urea, and there is no strong reason to expect a constant difference between protein values and the total diet. Such a relationship could be altered by changes in the flux patterns from protein to urea production. However, in practice an approximately constant difference, as already mentioned in Sections 2.2 and 2.3, has been widely recognized, although in animals whose diet is stressed by limiting water uptake, it may no longer be constant. The issue is an important one, since body protein (bone collagen) values can be interpreted to indicate the degree of carnivory of an animal, but the conditions under which this relationship holds are not well known. Influence of Different Physiological States and Digestive Strategies on Isotopic Mass Balance Important classes of animals, for example ruminants, produce additional excreta which are isotopically significant. The case of cows is interesting, in that the value of breath is found to differ from the diet by being heavier by (Metges et al. 1990). This can be exactly accounted for from the production of isotopically very light through bacterial action during rumination. Some of the physiology involved—for example, the production of isofractionated short chain fatty acids—has also been demonstrated. The methane produced is some depleted in 13C relative to diet, and accounts for about 10% of the total carbon flux (Metges et al. 1990). Methane production is therefore likely to invalidate the values for the D1FF of Equation (7). To a first approximation, would be rather than The same type of argument will also apply to lactation, although the degree of isotopic fractionation is far smaller. Thus one cannot base mass balance arguments on measurements relating cows’ diets with only their urea, since milk production is a significant component in the input/output mass balance.

Digestibility of Dietary Components The question of digestion must also be borne in mind. Isofractionation during metabolism will only apply to food that has been digested. The mass balance equation, and moreover the DIFF, needs to be explicit whether the whole diet or the digested diet is taken, and that account is taken of the excretion of undigested food. If cellulose is not digested, it should not be included in the dietary analysis. More subtle effects must be considered too. For example, flesh protein contains a high proportion of collagen with a significantly different This is digestible after cooking, but may not be digested by carnivores eating uncooked flesh. The ingested protein might then have a

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

225

different

value from the total protein and only this would be reflected in the mean total diet and therefore in the respired and consequently in the bioapatite carbonate.

The Occurrence of Isotopically Discriminating Biochemical Reactions Which chemical reactions within the metabolic pathways are isotopically discriminating is an absorbing question, but not one we take up in this paper. In probably all cases examined, the lighter isotope reacts faster, and discrimination is greater when bond breaking cleaves a relatively small molecular fragment, as in decarboxylations or deaminations. While potentially the number of isodiscriminating reactions is huge, quite possibly the actual number of significant cases may be quite small.

Manifestation of Isofractionating Reactions As mentioned in the Introduction, it is necessary but not sufficient for isotopically discriminating reactions to occur if isofractionation is to be observed in any particular body component. Whether the isofractionation incurred in a particular reaction in the pathway leads to a measurable effect in a body pool component depends on the pattern or topology of the metabolic fluxes. Three basic cases are considered here for illustrative purposes.

The Uni-directional Linear Chain The uni-directional linear chain is:

In steady state, no matter what isofractionation takes place. If only one step is isofractionating, this will be expressed only in the composition of the immediate precursor, somewhat contrary to immediate common sense. Hence if is plasma and X is the known fractionation between them, favoring isotopically lighter C in the will act so that the is maintained at an isotopically heavier level relative to the diet D. The same argument applies at any point in the chain.

The Reversible Linear Chain The reversible linear chain is:

226

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

Here the effects of any one fractionating step can be expressed in a change in isotopic composition in a wider range of body tissue components, including the product as well as the precursor of a (reversible) reaction. The details depend on the explicit model, for example how rates depend on metabolite concentrations. Therefore, where a metabolic pathway is, or becomes, reversible, the effect on isofractionation on measured body components can be more widespread.

The Branching Chain The branching chain is:

where represents a branching into and Fractionation in either of the branching steps 2, 3 or 2, 4 can then produce isofractionation in both of the branches to to in a way which depends on the ratio of flux (or rate constant) between the two branching reactions. This case is treated for lipid metabolism, both in general and for a specific model, which however does not use a steady state assumption, by Hayes (1993). All three cases can be combined to produce more complicated and perhaps realistic topologies. Evaluation of where isofractionation manifests itself in a complex flow-model requires an explicit description of the fluxes,

at steady state, or, alternatively, a prescription of the kinetics and rate constants for the reactions. It is relatively easy to make computer simulations of models involving first order kinetics for example for 20 metabolites. One can use this kind of modeling approach to do two things. First to find the simplest topology that reproduces the observed pattern of fractionation. This should help in understanding the observations. Second, to test ideas of “where the fractionation comes from” in an explicit way, and thereby to eliminate some of the candidate possibilities. In illustration, we consider a very simple model relevant to the D1FF constructed for the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data sets, which focuses on the protein routing question.

EVALUATING AN EXPLICIT FLOW MODEL FOR THE “PROTEIN ROUTING” QUESTION Consider the flow model shown in Fig. 11.1, in which the diet has two components (protein and non-protein), and the excretion has one The

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

227

Figure 11.1. A flow-model scheme for treating the protein routing question. Labels refer to flow rates of carbon. The total carbon flux, into and out of the body, is 1, divided into F (for protein)

and 1 – F for the remainder. The significant relevant internal fluxes are between the “amino acid pool” (coupled to the body protein pool), and the “energy metabolism pool”. The extent to which protein routing is observable in the body protein composition depends on the value of X (See Fig. 11.2). Numbers in {} refer to suggested isotopic fractionations associated with a metabolic path,

which are consistent with the data of the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data set (see Section 4.1).

organism has several body tissue or body pool components, of which two are measured (collagen and carbonate). Of the rest the most important are: the pool of amino acids, and the pool encompassing Krebs cycle and glycolysis activity (labeled as energy metabolism). If the total carbon flux through the system in steady state is = 1 unit (e.g., so many grams C per day), then the protein input (dietary) flux is called F, and the non-protein input is therefore (1 – F). The model has one internal

228

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

and unknown flux, X, which is the carbon involved in synthesizing amino acids from the “energy metabolic pool”. The physiological basis of this model is that both protein and non-protein must be respired, ultimately, as Protein can be formed, and broken down, into amino acids, which are either derived directly from the diet, or may be synthesized (directly or indirectly) from Krebs cycle or glycolysis metabolites. The question of protein routing is essentially a question of evaluating X. If X = 0, the protein is fully routed. If X > 1, the protein is “scrambled”. If all the relevant fluxes are in steady state, we can write equations for each flux and solve for the system as a whole, thereby relating the composition of the collagen to the composition of the diet, to F and to X. Assume, as a first approach, no isotopic fractionation takes place within the organism. We find:

This has the same form as the DIFF already discussed (Equation (1) in Section 2.1), with and and In other words, the flow-model here provides an explicit expression for f(). In fact, rather than fit the exponential function of in the DIFF from the data set, one can try to fit the data to a value of X (see Fig. 11.2). Unfortunately for the analysis, X is not very accurately determined in this way, since measurements at high and at low FP do not sensitively determine X. Nevertheless, it is clear from this figure, together with Fig. Al 1.1, that X has a value of about 0.2 for diets of 20–50% protein, that X is probably smaller for low protein diets, and is essentially undetermined for very high protein diets. Figure 11.2 also shows that a closer fit for a constant X can be achieved if, instead of relating f() to a constant fraction of total ingested carbon, it is related to a constant fraction of ingested energy (non-protein) carbon. The variation of X with F, which would be of considerable physiological interest, is not sufficiently tied down by these data to say more than that, in principle, it could be measured in this way. We emphasize that X has a direct physiological interpretation; it is the rate of amino acid synthesis (relative to the total food input) from the total catabolic carbon pool of the organism. From the same model, we also find:

That is, as noted above, the bioapatite carbonate tracks the total diet.

Dottedcurve: (This is established in the text as the best fit to the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data set). Continuous curve: (The weighting function implying total routing; collagen carbon derives only from diet protein). Lower dashed curve: as calculated from the flow model of Fig. 11.1, where X is a constant fraction of the total dietary

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

229

Figure 11.2. Plot showing how different weighting functions, f(F), change with the fraction, F, of protein carbon in the diet. Compare with Appendix Fig. A11.l.

input. Here the best fit to the dotted curve, and therefore to the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data set, is for

Intermediate dashed curve:

where Y is a constant

fraction of the total non-protein input. The best fit to the dotted curve

(and therefore to the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data set), is for Note that this is a rather better fit than that achieved for a constant fraction of the total dietary input (i.e,. the lower dashed curve).

The “Flow Model” for “Protein Routing”, Including Isofractionation We know, as the measured values of and (from the DIFF in Section 2.4) imply, that isofractionation takes place within the organism. To include fractionation in the model, the fractionating steps must be made explicit. The

simplest arrangement, consistent with observation, is as labeled in Fig. 11.2. That is, a fractionation of

from the amino acid pool to collagen suf-

230

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

faces to account for the “glycine redistribution” (+ here meaning favoring 13C), while a fractionation of between the amino acid and the energy metabolic pools is sufficient to account (in accordance with Section 3.4.2) for the observed additional collagen or protein fractionation giving a total of ca. Because the system has been shown to be mainly protein routing, i.e., X < 1, the reversibility in the flux between the protein and energy pools is small. It is reasonable that the lighter isotope be favored in both directions (between

amino acid and energy metabolic pools), and this in fact reduces the net isofractionation observed between non-protein and collagen to approximately The additional fractionations do not alter the previously evaluated relationship between f() and X. For we take isofractionation at to be which is close to that of the measured equilibrium fractionation. Embedding this in the complete model gives:

so long we can disregard the (physiologically unrealistic) reverse reaction,

Summary of the Outcome of the Flow-Model A simplest possible flow-model (Fig. 11.1) has been formulated to account for collagen and for bioapatite carbonate measurements, under conditions where the protein content in the diet is changing. Its “predictions”, in terms of the Dietary Isotope Fractionation Function (DIFF), have been made explicit. Comparison of the model-derived DIFF with the experimentally evaluated DIFF show that protein routing can be described in terms of the amount of protein synthesized from catabolic activity; that is, in general it is around 20% of the mean carbon flux of the diet. A simple explanation why is also proposed. As for carbonate measurements, the model confirms that bioapatitecarbonate should exhibit a value offset from the average-ingested-diet by an amount equal to the isotopic equilibrium fractionation between carbonate and gaseous

How the Collagen—Bioapatite Carbonate Spacing Varies With Diet This spacing has been measured for a range of mammals in a variety of ecosystems (Krueger and Sullivan 1984; Lee-Thorp et al. 1989). If we presume that the DIFF from the model, or from the Ambrose and Norr (1993) and

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

231

Tieszen and Fagre (1993) data sets, can be used to apply to mammals generally, we can express the spacing as a function of diet for mammals generally. We use the fractionation values of etc., from the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data sets. Combining equations (1) and (7):

Herbivores typically eat diets with high carbohydrate and low protein and lipid. F is about 0.15, and the protein is mainly plant derived and not very different in from the non-protein (i.e., is close to For carnivores, F is typically 0.5 or over, and carbohydrate is low. Animal protein is generally isotopically heavier, while the non-protein is much higher in lipid, so that is generally quite large The spacing, for the two diets is evaluated according to the equation above.

Thus if we take, for herbivores, we obtain

and

If we take, for carnivores,

and we obtain The difference in the spacing arises from the combination of two effects: that of protein routing (no difference is seen if protein is scrambled), and the difference in the spacing between and in herbivore and carnivore diets. In fact, for the rat diets recorded by Ambrose and Norr (1993), there is a clear linearity between and This is not to say that other effects are not also operating. The production of methane by ruminants, discussed in Section 3.2, is one additional influence which would have a similar effect on the spacing, as is apparent by modifying the d term in equation V from +10 to But this approach should enable field data to be interpreted in a unified and simple form. Note the strong emphasis it places on the effects of the values both of the protein and non-protein components on the spacing. It also allows predictions to be made of the spacing in unusual situations, for example, where the protein may be isotopically lighter than the non-protein.

MORE SPECULATIVE APPLICATIONS

Evaluation of More Detailed DIFFs At present, insufficient data are available, but the Tieszen and Fagre (1993) data set does give values for and as well as together with and and and It should therefore be possible to express and in terms of the three main food components. Explicitly,

232

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

and:

Note that f() has simply been evaluated as a constant and, while correct relatively, would require further normalization in practice (that is, Note also that these figures imply strong routing of lipid. This DIFF requires evaluation of the isofractionation and weighting matrix elements (d and f()). This is a question of searching the multidimensional space for the combination of d and f() values which provide a good fit between the DIFF and the observations. We have attempted to do this

from the Tieszen and Fagre (1993) data set, but the data do not contain sufficiently informative relationships to provide clear-cut results, and our results cannot be shown to be the most accurate and unambiguous descriptions of the data. We have taken f() to be a constant, as the simplest possible evaluation. However, it seems worthwhile to present the results (Table 11.2), if only to point to what should be possible. These functions should be regarded in the spirit of “working suggestions”.

We would argue that it is the task of those aiming to reconstruct ancient diet as well as those aiming to understand how isotopic fractionation signatures are produced, to devise experiments and analyze the results in order to fill in the blanks in the matrices.

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

233

Figure 11.3. A flow-model scheme intended to represent relevant nitrogen flows, especially with regard to which flows are reversible. The labeled reactions I, II, III

IV are all potentially iso-

topically fractionating. Because reaction II is not reversible, subsequent fractionations in the excretory pathway should not influence the isotopic composition of the body protein pool.

Nitrogen and Water Stress Figure 11.3 is a flow model representing in extremely simple form the main relevant features of nitrogen metabolism. It is not difficult to propose a sufficient explanation why is isotopically heavier than the diet. We might expect that the net effect of transamination and deamination of amino acids is to remove isotopically lighter N (Macko et al. 1987). That is to say, we may expect that the equilibrium constant for the reaction:

234

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

would favor lighter nitrogen on the ammonium side, and that, in nonequilibrium conditions, rates in both directions would be faster when transferring 14N rather than 13N. If this reaction is always driving to the right, the effect, in a steady state, is to make the amino acid heavier (this follows from Section 3.4.1). If the reaction is near equilibrium, the effect is probably reduced, depending on the fractionation term in the equilibrium constant. The total effect must take account of subsequent fractionating steps which may influence the fractionation of amino acids. If the next step is also fractionating, as is likely, (either active transport of or condensation to carbamyl phosphate), this effect will be added to that of the deamination reaction. But subsequent fractionations will not influence the amino acid isotopic composition unless they are reversibly coupled to amino acid metabolism, which seems inherently unlikely. A possible reversible coupling is via urease in the gut, which may be significant under conditions of starvation. We may then inquire why water stress should alter the fractionation observed in the body protein, or body pool amino acids. The critical part of the pathway for nitrogen is:

If the effect of water stress is to alter regulation of the pathway such that the rate constant for reaction is increased or is decreased (which would have an overall effect of conserving nitrogen), then the fractionation at G can be shown to be thereby increased. At present this is speculative, but in fact explanations for the water-stress effect using flow-models are rather constrained. For example, it is not possible to relate what might happen at the kidneys (e.g., resorption of urea) to the amino acid body pool, since the urea cycle is non-reversible. It should be possible to design experiments that test this suggestion.

“Trophic Levels” It is often found that in the study of an ecosystem the pattern of isofractionation signatures conforms to structural patterns in the ecosystem. Isotopic signatures may be used to provide ecological information without the need or possibility of detail at the dietary or physiological level of description. Such generalizations can bring about clarity and insights. However, there is also a

real danger of confusion, and the phrase “trophic level effect” especially when applied to carbon, seems highly misleading, because carbon can be ingested in so many forms (different components of D), and measured in so many forms (different components of B). Any observed “trophic level effect” is the sum of an undefined set of DIFFS. How does the modeling of an ecosystem differ from that of a single organ-

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

235

ism? Obviously several organisms (or rather populations of organisms) can be linked together in a food chain (or more complicated network). Because of excretion at each stage, the overall mass balance equations lose their simplicity, but each member in the chain can be evaluated in terms of its preceding member. Qualitatively, the main difference is in the definition of diet. This

must now take explicit account of the selectivity of the feeder; that is, which body parts of the organism are chosen as food and digested by its predator. Thus branches have become an essential part of the topology, both due to excretion and to dietary selectivity (see Fig. 11.4).

Figure 11.4. (a) Flow-model scheme for a simple food chain with one predator-prey relationship. See text for discussion, (b) The steps involved whereby atoms from prey collagen (i.e., the diet) may be transferred to a predator’s collagen (i.e., the consumer tissue). Each arrow represents a potential change in carbon isotopic composition, complicating the relationship between prey collagen and predator

236

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

The isotopic compositions at such branches are generally different. The situation is simplest for nitrogen in mammals, since only protein is involved (but might be more complicated for other animals such as insects, where chitin

is an important component), and there is little difference in between different proteins. Therefore dietary selection does not have an effect on the isotopic composition of the diet in this case and can be ignored. We expect (from empirical evidence) the body tissue (protein) nitrogen to be some heavier than the dietary input (or the excretory output). Proceeding along the chain, it is evident, and hardly needs mathematical proof, that at each excretory stage, corresponding to each separate trophic level, the population body tissue values will be successively elevated. For carbon, the model would need to be evaluated with an explicit account of the particular body parts, and their isotopic values, that are digested, together with the relevant DIFF For example, if the DIFF in Section 5.1 is used, the in the predator can be evaluated from the relevant dietary contributions, and would tend to be about

enriched in 13C over the average digested diet. For collagen the enrichment would be greater. Because all the values for are there is a net increase in 13C in an animal’s protein. However, note that the DIFF of Section

5.1 is only tentative. Although the predator protein may be enriched with respect to average diet, it may not be enriched with respect to the average prey, since the diet may be selecting isotopically lighter elements, for example, collagen may not be strongly represented in the digested diet. Several further steps have to be considered if only measurements of predator and prey collagen are concerned (see Fig. 11.4b).

SUMMARY AND CONCLUSIONS We attempt here to understand and predict body isotopic composition through modeling, and show that the observable data need to be set in a suitable framework (the DIFF), so that assumptions can be made more explicit. We evaluate the terms in the DIFF for the combined controlled diet data set of Ambrose and Norr (1993), Tieszen and Fagre (1993) for collagen and bioapatite carbonate, and amino acids from experiments by Hare et al. (1991). We construct and analyze a simple flow model for carbon protein and energy metabolism (Fig. 11.1) that is able to produce the behaviour described by the DIFF The model has physiological consequences: it predicts the level of amino acid synthesis from the “metabolic energy pool”, and it suggests where, within simplified metabolic fluxes, isofractionation is taking place. From this model, DIFFs may be formulated that allow the collagen to bioapatite carbonate spacing to be predicted for a given diet. The prediction is in line with observations of how this spacing differs between herbivores and carnivores, and shows explicitly how protein

“CONSIDER A SPHERICAL COW . . .”—ON MODELING AND DIET

237

routing and dietary differences combine to produce the observed differences in spacing.

A useful simplification in such models is the steady state approximation, which enforces mass balance between input and output. This has some interesting consequences, and in particular predicts that animals with ruminant

digestive physiologies, essentially those that produce significant quantities of isotopically light methane (Metges et al. 1990), will tend to have a larger bioapatite carbonate spacing than those that do not. Finally, an outline of how the same approach to modeling can be applied to questions of nitrogen metabolism and to food chains in ecosystems is presented. We would argue that the value of such an approach lies at least as much in helping to better define issues, and in suggesting new, testable hypotheses, than in providing explanations which may rest on oversimplified assumptions.

ACKNOWLEDGMENTS

We would like to acknowledge helpful and lively discussions with other members of the diet group at RLAHA, notably Tamsin O’Connell, Michael Richards and Sarah Webb.

REFERENCES Ambrose, S.H. 1991 Effects of diet, climate and physiology on nitrogen isotope abundances in terrestrial foodwebs. Journal of Archaeological Science 18: 293–317. 1993 Isotopic analysis of paleodiets: methodological and interpretive considerations. In

Sandford, M.K., ed., Investigations of Ancient Human Tissue, Langhorne, PA., Gordon and Breach Science Publishers: 59–130. Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of bone collagen and carbonate. In Lambert, J.B. and Grupe, G., ed., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 1–37. DeNiro, M.J. and Epstein, S. 1978 Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta 42: 495–506. Hare, P.E., Fogel, M.L., Stafford, Jr., T.W., Mitchell, A.D. and Hoering, T.C. 1991 The isotopic composition of carbon and nitrogen in individual amino acids isolated from modern and fossil proteins. Journal of Archaeological Science 18: 277–292. Hayes, J.M. 1993 Factors controlling I3C contents of sedimentary organic compounds: Principles and evidence. Marine Geology 113: 111–125. Heaton, T.H.E., VogelJ.C, Von la Chevallerie, G. and Collett, G. 1986 Climatic influence on the isotopic composition of bone nitrogen. Nature 322: 822–823. Krueger, H.W and Sullivan, C.H. 1984 Models for carbon isotope fraclionation between diet and bone. In Turnlund, J.S. and Johnson, P.E., eds., Stable Isotopes in Human Nutrition. ACS Symposium Series 258, Washington DC, American Chemical Society: 205–220.

238

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

Lee-Thorp, J.A., Sealy, J.C. and Van der Merwe, N.J. 1989 Stable carbon isotope ratio differences between bone collagen and bone apatite, and their relationship to diet. Journal of Archaeological Science 16: 585–599.

Little, J.D.C. and Little, E.A. 1997 Analyzing prehistoric diets by linear programming. Journal of Archaeological Science 24: 741–749. Macko, S.A., Fogel, M.L., Hare, P.E. and Hoering, T.C. 1987 Isotopic fractionation of nitrogen and carbon in the synthesis of amino acids by microorganisms. Chemical Geology (Isotope Geoscience Section) 65: 79–92. Metges, C., Kempe, K. and Schmidt, H-L. 1990 Dependence of the carbon isotope contents of breath carbon dioxide, milk serum and rumen fermentation products on the 13C value of food in dairy cows. British Journal of Nutrition 63: 187–196. Mook, W.G. 1989 Principles of Isotope Hydrology: Introductory Course on Isotope Hydrology. Amsterdam, Vrije Universiteit. Steele, K.W. and Daniel, R.M. 1978 Fractionation of nitrogen isotopes by animals; as further complication to the use of variations in the natural abundance of 15N for tracer studies. Journal

of Agricultural Science 90: 7–9. Stott, A.W., Davies, E., Evershed, R.E. and Tuross, N. 1997 Monitoring the routing of dietary and biosynthesized lipids through compound-specific stable isotope (delta13C) measurements at natural abundance. Naturwissenschaften 84(2): 82–86. Sutoh, M., Koyama, T. and Yoneyama, T. 1987 Variations in natural 15N abundances in the tissues

and digesta of domestic animals. Radioisotopes 36: 74–77 Tieszen, L.L. and Fagre, T. 1993 Effect of diet quality and composition on the isotopic composition of respiratory bone collagen, bioapatite, and soft tissues. In Lambert, J.B. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 121–155.

APPENDIX The data set actually used for modeling, from controlled diet experiments of Ambrose and Norr (1993) and Tieszen and Fagre (1993), is shown in Table A11.l.

Diets in which

is small do not provide a sensitive indication of

routing or differences in dp and dN.

The DIFF is evaluated as follows: is calculated from equation (A1) using the measured values of D and F from the data set, together with a set of trial values of and f(). and can independently take any value between +6 and –6. f() is taken as the function where w takes any value between 0 and –1.

The mean and variance of the difference between (from table) and is determined for all 14 diets for each trial combination of and and the best values for and w chosen to minimize both the mean and the variance. These values turn out to be and Figure A11.l shows a plot of the difference between the estimated and calculated collagen values for each diet for this particular DIFF, and it can be seen that, except for one point, the others are correctly estimated to within 1 or

“CONSIDER A SPHERICAL COW ... ”—ON MODELING AND DIET

239

Table A11.1. Carbon isotope and protein content of experimental diets used for modeling in this paper. The values are in %° with reference to the PDB standard, except for protein content (by weight) which is in %. The type of diet corresponds to the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data sets.

Figure A11.1. A plot of the difference (residuals) between observed collagen

values and

values calculated from the DIFF for and as a function of the dietary protein carbon content. Due to the combination of composition and manipulated isotopic compositions of the different diets, some diets test the predictions of the DIFF more precisely than

others. These are represented as squares (the remainder are represented as diamonds). Although the difference has been minimized, it is not zero. Nevertheless, and especially for the more reliable rectangular points, the difference is small, for a wide range of diets and collagen values.

Other combinations of

and f(F) give greater residuals.

240

ROBERT E.M. HEDGES AND GERT J. VAN KLINKEN

In fact a somewhat wider range of and values provides a reasonable fit, and what is regarded as the best fit is rather subjective. For example, gives the best average values, but give the least

scatter (but with an offset mean). Furthermore, not all data are consistent (the low protein data seem most difficult to fit). Nevertheless, it is clear that the DIFF gives a reasonable prediction if: f() is taken to be close to direct routing; or and to (and anti-correlated with Figure A11.2 shows how f() values vary with the protein content of the

diet, for and From inspection of the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data sets, we note that data from diets [B, C, D, E, F, G, 5, and 6] are more able to constrain values of d and f() and these points are separately indicated. The smooth curves represent (complete routing), (the best fit to the data), and ( = 1 = scrambling). For high protein diets there is little distinction between the curves. Note that

f() is always greater than 1, implying some protein routing. for

The prediction for the collagen of Ambrose’s diets (Table A11.2) is made and with (i.e., for protein).

Figure A11.2. Plot of f(F) values for data points from the Ambrose and Norr (1993) and Tieszen and Fagre (1993) data sets. f(F) is calculated from the DIFF (Equation I) with and Note that all the points lie between the two functions and corresponding to total protein routing and total scrambling respectively. The best fit is for The rectangular data points are, a priori, regarded as more reliable than the

diamond points in that they constrain f() more precisely.

“CONSIDER A SPHERICAL COW...”—ON MODELING AND DIET

241

In practice, the prediction is slightly closer with and for with (both predictions being within from the measured value). Other combinations generally differ by The prediction for carbonate is made for: and and The results, presented in the main body of this paper (Table 11.1), show close agreement between observed and predicted carbonate and collagen values. While the justification for the values used in the DIFF is not based on rigorous or systematic analysis of the data, it is

adequate to demonstrate the general features, and to suggest that the values obtained be confirmed by further experiment.

This page intentionally left blank

Chapter

12

Controlled Diet and Climate Experiments on Nitrogen Isotope Ratios of Rats STANLEY H. AMBROSE ABSTRACT Nitrogen isotope ratios ( 15 N/ 14 N) increase from plants to herbivores to carnivores and can be used to estimate the degree of carnivory in human diets. Some field studies observe a greater difference in between trophic levels in dry, hot habitats than in wet, cool ones. Two hypotheses have been proposed to explain this variation in difference in between trophic levels. (1) Elevated excretion of 15N-depleted urea in heat/water-stressed animals; (2) recycling of nitrogen on protein-deficient diets. Both predict increased diet-tissue difference under stress. Controlled diet and climate experiments with laboratory rats were performed to test these hypotheses. Litters of six or more rats were raised to maturity on diets with protein from cow milk (casein) with known values. Diets were formulated with 5%, 20% and 70% protein. Animals on 5% protein diets had slow growth rates and low adult body weights, suggesting significant protein stress. Rats were also raised to maturity on 20% and 70% protein diets, at temperatures of 36° and 20°C, with restricted water intake or water provided ad-libitum. There were no significant differences in values of bone collagen for low versus normal or high protein diets, or for water- and/or heatstressed versus unstressed animals. The results suggest either that rats are an inappropriate animal model for such experiments, or that the protein and heat/water stress models of nitrogen isotope variation are incorrect. Biogeochemical Approaches to Paleodietary Analysis , edited by Ambrose and Katzenberg. Kluwer Academic/Plenum Publishers New York, 2000.

243

244

STANLEY H. AMBROSE

Decreases in muscle and increases in hair and collagen

with age

were observed in litters raised on 20% protein. These results provide partial

support for the hypothesis of a relationship between

and age.

INTRODUCTION The step-wise increase in 15N between trophic levels, from plants to herbivores to carnivores, and the enrichment of 15N in marine compared to terrestrial ecosystems, have been used to estimate meat versus plant and/or marine versus terrestrial resource consumption (DeNiro and Epstein 1981; Schoeninger and DeNiro 1984; Minagawa and Wada 1984; Schoeninger 1985; Ambrose and DeNiro 1986; Fogel et al. 1989). If enrichment in 15N between trophic levels from plants to herbivores to carnivores is constant, linear and predictable it can be used to estimate the consumption of plants versus meat. However, estimates of the degree of enrichment in 15N relative to the diet and between trophic levels vary between studies (Ambrose 1991; Minagawa and Wada 1984). The most widely accepted figure for enrichment in between trophic levels, and from diets to tissues such as bone collagen and hair, is to but estimates range from to in different regions (Table 12.1). Nitrogen isotope values in animal tissues may be influenced by climate and/or nutrient composition. In hot, arid environments, soils, plants and animals tend to have higher values (Ambrose 1991; Sealy et al. 1987; Heaton et al. 1986; Vogel et al. 1990). A relationship between rainfall and has been observed for deer in North America, but only among individuals that fed on small amounts of plants (mainly grasses); no relationship was observed for deer that were identified as pure browsers or feeders (Cormie and Schwarcz 1996). In hot, arid environments the diet-tissue nitrogen isotope difference seems to increase, but this pattern is not widely replicated. In semi-arid southern and eastern Africa (Sealy et al. 1987; Ambrose 1991) the diet-collagen difference for herbivores seems greater than 3%0 in more arid environments. Only one study has found a constant difference between diet and tissue nitrogen isotope ratios between habitats (Vogel et al. 1990). In this study of the nitrogen isotopic composition of elephants and their foodwebs, a constant enrichment of 2.1%o was observed at all levels of precipitation. The increase in diet-tissue spacing has been proposed to be caused by the effects of water and heat stress on urinary nitrogen excretion. The model has been described in detail previously (Ambrose 1991) and will be briefly summarized here: Nitrogen is excreted mainly as urinary urea. Its value is substantially (2-5%o) more negative than that of the diet (Steele & Daniel 1978; Yoneyama et al. 1983). Under heat and water stress the concentration

NITROGEN ISOTOPE RATIOS OF RATS

245

of urinary urea and the amounts excreted by many herbivore species increases in response to water conservation in the kidney (Maloiy 1973a, 1973b; Maloiy et al. 1979). In order to satisfy the requirements of isotopic mass balance (the isotopic composition of what is excreted plus that retained must equal that isotopic composition of what is consumed), the elevated excretion of 15Ndepleted urea must be balanced by the retention of 15N-enriched nitrogen, which may then be used for tissue synthesis. Conditions that promote

increased urea excretion should result in increased retention of l5N-enriched nitrogen and thus higher diet-tissue values. The effects of heat and/or water stress can be determined experimentally by raising animals at high temperatures with reduced water intake. This study reports on experiments designed to test the heat and water stress hypothesis.

246

STANLEY H. AMBROSE

Protein stress is an alternative explanation for increased

This

hypothesis is based on the observation that mammals assumed to be under protein stress appear to have higher nitrogen isotope ratios than those receiving adequate protein (Hare et al. 1991; Sealy et al. 1987). This hypothesis posits that when dietary protein is scarce there will be more intensive recycling of scarce nitrogen into proteins. Preferential deamination and transamination of isotopically light amino acids (Macko et al. 1986) should lead to nitrogen isotope enrichment of the remaining tissue. Each cycle should be accompanied by selective retention of 15N-enriched nitrogen. Protein stress and recycling of nitrogen could also have the opposite effect, however. If less 15N-depleted N is excreted as urea, then there should be less overall enrichment in the nitrogen available for tissue synthesis. Moreover, if urea itself is recycled for protein synthesis under protein stress, which

often occurs in herbivores, then the diet-tissue difference should be smaller than in unstressed individuals because urea has a substantially lower value than the diet. The problem of nitrogen isotope fractionation during digestion of proteins, and use of non-protein N in urea and ammonia produced during digestion and metabolism or ingested as salivary or dietary urea is complex. Urea and ammonia both tend to have lower values than their precursors or substrates, and bacteria can use both ammonia and urea as substrates for growth. An in-vitro study of fractionation of nitrogen isotopes by ruminal bacteria (Wattiaux and Reed 1995) shows that when the dietary substrate was amino acids from casein (milk protein) the of bacteria was +1.3 to When ammonium bicarbonate was the substrate, bacterial decreased from to and ammonia increased from -1.4 to +12.7 over six hours. The increase in ammonia during the course of this experiment is expected because bacteria may preferentially metabolize isotopically light ammonia. These results thus satisfy isotopic mass balance requirements. Regrettably, this study did not investigate the relationship between bacterial and urea nitrogen isotope ratios. In-vivo studies of bacteria and protozoa in the cow rumen show that they have values that are lower and higher, respectively, than that of the diet (Sutoh et al. 1987). This result is expected if bacteria used biochemical fractions that have lower values (urea, ammonia) than the bulk macronutrients (proteins, amino acids) that are probably consumed by protozoa. Changes in the value of food in different parts of the digestive system of three goats (Sutoh et al. 1987) reveal the complexity of isotope effects during digestion. In this study the food value was + but immediately rose to high values in the forestomach (rumen, reticulum and omasum: Digesta then dropped at the beginning of the intestines (abomasum: but increased linearly toward the end of the intestines (duodenum: jejuno-ileum: and caecum: Colonic (fecal) matter had the lowest values Consistent isotopic enrichment

NITROGEN ISOTOPE RATIOS OF RATS

247

during digestion may reflect both selective removal of isotopically light nitrogen and the addition of isotopically enriched digestive secretions and digestive tract tissues. Reduction in in the colon may reflect digestion and absorption of the 15N-enriched microbial community before excretion. Comparative studies of nutrient-stressed and unstressed ruminants should be undertaken to assess the relevance of these results for the problem under consideration and perhaps to salvage insight from this complexity. Urea and ammonia are incorporated into tissue proteins in humans (Richards 1972). Studies of protein-stressed humans from Papua New Guinea have demonstrated uptake of 15N-labeled urea, but this inorganic source seems to be used only under nutrient stress (Rikimaru et al. 1985). Human milk contains from 11–52% non-protein N, approximately half of which is urea (Atkinson 1985; Donovan et al. 1986). Most (84%) of the non-protein N is absorbed by nursing infants (Donovan et al. 1986). If the logic of the urea excretion model is correct, and if humans use serum or other sources of urea as a substrate for protein synthesis, then protein-stressed humans should also have reduced diet-tissue The role of urea as an N source should also be considered in interpreting nitrogen isotope changes observed during weaning in

humans (e.g., Fogel et al. 1989; Katzenberg et al. 1993). Absorption of milk

urea may explain why the experimental infant-mother observed by Fogel et al. (1989) is less than 3‰. Convincing experimental evidence for the effect of nutrient stress on tissue has been found only in studies of birds (Hobson and Clark 1992). Crows that were switched from uncontrolled, presumably high protein diets to plant diets 15 days after hatching, had reduced growth rates and lost muscle mass while maintaining bone growth compared to crows fed fish. Crows on plant diets had higher values than those fed fish. The diet-tissue difference for stressed and unstressed crows was particularly substantial for bone collagen and respectively). Careful consideration of this experiment is warranted. Although most vertebrates probably have similar protein synthesis and metabolism pathways, there are substantial differences between birds and mammals in rates of growth and development, and digestive and excretory physiology. For example, birds excrete excess nitrogen as uric acid rather than as urea. The effects of excretion of nitrogen in these forms and their effects on isotopic mass balance should be investigated. Potential differences in bone turnover rates between crows and mammals should also be considered. The nutrient stress hypothesis can be tested by comparing diet-tissue values of animals on low versus normal and high protein diets. Our controlled diet experiments, although primarily designed to trace carbon from dif-

ferent dietary macronutrient fractions (proteins versus carbohydrates, fats and sugars) to animal tissues under different levels of nutrient stress (Ambrose and Norr 1993) may be suitable for testing this hypothesis because they contain diets with 5, 20 and 70% protein by weight.

248

STANLEY H. AMBROSE

Tissue

values have also been hypothesized to increase with age

(Minagawa and Wada 1984), based on the observation by Gaebler et al. (1966)

that “Many amino acids extracted from rats have variable 15N contents during the transaminating process. If nitrogen undergoes fractionation sometime during the transamination or deamination process the value of nitrogen accumulated within animals could be expected to increase with time.” Minagawa and Wada did not conduct extensive experiments on the aquatic organisms in their study and found no correlation between tissue and age in two mussels. Kidneys and hearts of ten cows sampled at ages 27 to 124 months after birth showed no increase in 15N with age (Yoneyama et al. 1983). Feathers of four species of birds showed no increases in 15 N values with age over periods ranging from four to 23 years (Mizutani et al. 1992). The experiments reported below are also suitable for testing the age hypothesis because littermates were sacrificed at three different ages. Our results suggests unforeseen complexities in the relationship between tissue and age.

RESEARCH DESIGN

Eleven controlled diet and environment experiments have been designed in a way that can be used to investigate the effects of protein nutrition and heat and/or water stress on diet-tissue Laboratory rats were raised on purified, pelletized diets in which the isotopic composition of proteins, lipids and carbohydrates were well characterized and their proportions accurately and precisely measured (Ambrose and Norr 1993). Four experiments involved manipulation of temperature and/or water availability. Of these four experiments, one used a diet with high (70%) protein concentrations and heat/water stress (36°C) and three used normal (20%) protein concentrations. Seven experiments were conducted at normal temperature (21°C) with water ad libitum. Of these seven experiments, two used diets formulated with very low protein (5%), three with normal protein and two with high protein concentrations.

Growth Protocol Sperm-positive female Holtzman rats were started, one day after insemination, on the diets that their offspring would consume. Birth occurred 21 days after insemination and weaning occurred 21 days later; sexes were separated prior to sexual maturity. Surplus individuals were sacrificed at weaning. Male and female pairs from each group were sacrificed at 91, 131 and 171 days after birth. Pairs on one diet were also sacrificed at 211 and 251 days after birth. Euthanasia was performed by exposure to Isotopic analyses were

NITROGEN ISOTOPE RATIOS OF RATS

249

performed on mother’s milk, bone collagen of appendicular bones (forelimbs or hindlimbs), muscle and hair (carbon and nitrogen) and bone carbonate (carbon). Normal room temperatures (20°C) were maintained, and food and

water were provided ad libitum. Food consumption was monitored by daily weighing of uneaten food pellets.

Diet Synthesis Seven diets were constructed from purified natural ingredients obtained from either (beet sugar, rice starch, cottonseed oil, wood cellulose, Australian Cohuna brand casein, soy protein or wheat gluten for protein) or foodwebs (cane sugar, corn starch, corn oil, processed corn bran for fiber, Kenya casein for protein) supplemented with appropriate amounts of vitamins and minerals (Ambrose and Norr 1993: Table 3a). The amino acid composi-

tions of wheat gluten and soy protein differ significantly from that of casein (Ambrose and Norr 1993). Diets were formulated at Teklad Premier (Madison, Wl) to produce accurately and precisely weighed combinations of pure

energy with

protein, and

energy with

energy and protein,

protein (Ambrose and Norr

1993). Proteins were provided at very low (5%), normal (20%) and very high (70%) levels in the diet. The plant protein diet was formulated with 10% wheat gluten plus 10% soy protein (diet F´). The expected carbon isotopic composition of the whole diets based upon the proportions and isotopic composition

of their ingredients matched the actual diets almost perfectly (Ambrose and Norr 1993: Table 3b), indicating precise and accurate formulation of diets. Nitrogen isotope analysis of dietary ingredients and whole pelletized diets has been problematic because nitrogen concentrations in some experimental

diets are extremely low. To produce samples large enough for accurate measurement on the Finnegan Delta-E mass spectrometer it was necessary to combust samples up to 110 mg, so replicable results are not yet available for some diets. The dietary nitrogen source is almost entirely from the protein component (casein, or wheat gluten and soy protein), with minor contributions from starches, vitamins and corn bran (Ambrose and Norr 1993: Table 3a). The contribution of the non-protein ingredients to total nitrogen content is higher on the low-protein diets. The degree to which their N is incorporated into collagen is unknown. We have previously demonstrated that most of the carbon

in collagen comes from the protein, even on the 5% protein diets (Ambrose and Norr 1993). Therefore, in this study, values of the protein source alone will be used to estimate diet-to-tissue nitrogen isotope enrichments.

Heat and Water Stress Protocol In order to examine the effects of water and heat stress on diet-tissue nitrogen isotope enrichment, four groups were raised with various

250

STANLEY H. AMBROSE

combinations of water and heat stress, all with food provided ad libitum, on the insemination, growth and sacrifice schedule described above. Individuals were housed in separate cages after weaning in order to prevent individual variation due to food and water competition. Cages were placed on two racks of shelves semi-enclosed by black plastic sheets. Heat was provided by infrared heat lamps placed within the enclosure, but not shining directly on the cages. Heat and water stress conditions were initiated after gestation and lactation were completed (21 days after birth). Body weight, food and water consumption were measured daily. The animals exhibited no signs of discomfort, aggressiveness or lethargy throughout the experiment. Group 9 was raised at room temperature on a 20% protein diet (Diet A: all ingredients 20% protein), with water consumption restricted to provide a target growth rate of 80% of groups on the same diet with water ad libitum. Three groups had a temperature regime of 12 hours at 36°C and 12 hours at 20°C. Group 10 received water ad libitum and diet A. Group 14 (diet A) was water-restricted to maintain 80% of the growth rate of Group 10. This group experienced the greatest stress. Group 11 was water-restricted, but consumed a 70% protein diet (Diet E, energy). Group 5 served as the unstressed control for diet E.

RESULTS AND DISCUSSION

Diet-Tissue Differences in Collagen and Hair Nitrogen Isotopes Since virtually all nitrogen comes from protein, the values of the values. For groups kept at normal room temperatures, the mean value is 3.3 0.18 (Table 12.2). These data agree with previous estimates (Table 12.1). There are no systematic differences in collagen values between 5%, 20% and 70% protein diets. The absence of differences in between groups raised on widely varying amounts of protein does not support the hypothesis that low protein diets increase values. Hair and collagen values are significantly correlated = 0.76). The mean difference between hair and collagen values is insignificant (–0.01 ± 0.29 Hair can thus directly substitute for collagen where seasonal changes in the isotopic composition of diet have not occurred. protein sources is used for estimating

Effects of Heat and Water Stress on Tissue Nitrogen Isotopes The results of heat and water stress experiments (Table 12.3) show that these factors had no significant effects on nitrogen isotope ratios of bone collagen or hair. The mean collagen-diet difference values of the waterrestricted litters ranged from 2.6 for group 10 on diet A (36°, water ad

NITROGEN ISOTOPE RATIOS OF RATS

251

252

STANLEY H. AMBROSE

libitum) to for group 11 on diet E (the highest protein, with highest heat and water stress). The three litters on diet A with different heat and water stress treatments (groups 9, 10 and 14) did not differ significantly in values (2.6–3. The high value for group 11 may be significantly different from the remaining litters but not from the control (group 5). The mean values for the unstressed groups (Table 12.2) ranged from 3.1 to (mean of means = 3.3 ± 0.18). The values of the heat and/or waterrestricted groups (mean of means = 3.1 ± were slightly lower than those of the unstressed low temperature control groups, with the exception of group 11, on the 70% protein diet (diet E), where was Hair values follow the pattern of collagen values very closely. The mean of mean values for heat-stressed rats is 3.0 ± 0.41 That for unstressed rats is 3.4 ± 0.46 The reduced spacing between diet and collagen and diet and hair under environmental stress seems counterintuitive. One possible explanation is that under water and heat stress the rats had inadequate nitrogen for elevated urea excretion and recycled it instead. This explanation would be consistent with the high diet-collagen spacing for group 11 (diet E, 70% protein), where protein nutrition would have been more than adequate for the urea excretion

model. Variable rates of nitrogen excretion in response to manipulation of the nitrogen content of diet have been found in other studies (Livingstone et al. 1962). Another possible explanation is that fractionation was reduced under reduced metabolic rates. Basal metabolic rates of many species are known to drop significantly (the fasting metabolic rate) under nutrient and environmental stress (Maloiy 1973b), while high protein intakes are known to result in elevation of metabolic rates (Schuette et al. 1981; Trilok & Draper 1989a, 1989b). These data do not provide strong support for the heat & water stress/urea recycling model (Ambrose 1991). This model may be incorrect or inapplicable to rats. Were the experimental conditions inappropriate, with temperatures too low and/or protein levels too low or too high? In the heat stress experiments that inspired this research animals were kept at a temperature of 40°C for 12 hours each day rather than 36°C in this study. In our heat and water stress experiments the protein content of the diets were set at 20% and 70%. These are relatively high levels compared to those in herbivore diets. It would be necessary to repeat the experiments with ruminant herbivores or lower protein diets to conclusively determine if rats are an inappropriate model.

Effects of Protein Stress on Tissue Nitrogen Isotopes Protein stress has been proposed to cause tissue nitrogen isotope enrichment (Hare et al. 1991; Hobson and Clark 1992). Diets B and C, with 5% protein by weight, produced offspring that achieved substantially smaller adult weights, clearly demonstrating protein limitation. In many cases the

NITROGEN ISOTOPE RATIOS OF RATS

253

individuals were so small that dermestid beetles had to be used to deflesh

enough skeletal elements for analysis. Their mean values were not significantly different from the animals fed 20% and 70% protein diets in which the proteins had the same isotopic composition. Our results do not support the protein stress model. However, this model may apply in cases where stress is intermittent and results in tissue loss, as observed in the study of crows (Hobson and Clark 1992). Low protein levels throughout life after weaning may have produced overall slow and reduced rate of growth rather than tissue loss. Adult rats fed protein-deficient diets after maturation show systematic losses of nitrogen from most tissues that are in proportion to their turnover rates and masses (Uezu et al. 1983). Perhaps tissue nitrogen isotope enrichment may occur under these conditions. New experiments are needed to evaluate this hypothesis.

Relationship Between Tissue Nitrogen Isotopes and Age Three previous studies investigated but did not find that

increases

with age (Minagawa and Wada 1984; Yoneyama et al. 1983; Mizutani et al.

1992). The design of our experiments, with sacrifice of individuals at three 40day intervals, permits examination of potential relationship of age and in rat hair, flesh and collagen. No statistically significant correlations with age were found for collagen, but in eight of 13 groups the slopes were positive (Fig. 12.1). For hair, 10 of 13 groups had positive correlations, of which four

Figure 12.1. Relationship between collagen values and age of male and female rats raised on controlled diets, and sacrificed at 40 day intervals beginning 91 days after birth.

254

STANLEY H. AMBROSE

were significant at P = 0.003 to 0.072 (Fig. 12.2). Only Group 3, Diet C (5% protein), has a systematically negative slope for age versus collagen and hair The increase in of hair and collagen with age requires explanation. It occurs on both 5% and 20% protein diets. The lower rate of change in collagen compared to hair may be a function of the former’s lower rate of turnover once skeletal maturity is reached. Flesh values decrease systematically with age (Fig. 12.3). The mean difference between flesh and diet is 3.0 ± on day 91, 2.8 ± on day 131 and 2.2 ± on day 171. If muscle has a relatively high turnover rate, then this result is extremely counterintuitive given the assumptions of Minagawa and Wada (1984) that selective catabolism of isotopically light molecules should occur and should cause tissue 15N enrichment rather than depletion.

The mean difference between collagen and flesh

values for the

first sacrificed pairs (91 days after birth) is 0.3 ± while that for third pairs (171 days after birth) is 1.4 ± This result was unanticipated but seems robust. This indicates the relationship between diet and tissue and age is complex and varies between tissues. Future studies of diet-tissue nitrogen isotope spacing will have to consider age effects. This contrasts with carbon isotopes (Ambrose and Norr 1993), where we have observed little increase in with age in the same individuals.

Figure 12.2. Relationship between hair values and age of male and female rats raised on controlled diets, and sacrificed at 40 day intervals beginning 91 days after birth.

NITROGEN ISOTOPE RATIOS OF RATS

Figure 12.3. Relationship between rat flesh

255

values and age of male and female rats raised

on controlled diets, and sacrificed at 40 day intervals beginning 91 days after birth.

Unlike results of studies of human infants and adults (Fogel et al. 1989; Katzenberg et al. 1993), where tissue values decline after weaning, rat hair and collagen values increase with age. The decrease in muscle seems to be offset by the increase in collagen and hair Hair is probably synthesized in part from circulating amino acids derived from soft tissue turnover. If so, it is likely that the isotopically enriched nitrogen pool drawn from flesh contributes to that used to synthesize hair. The experimental animals were not

losing weight after maturity and females were not reproducing, so enrichment in flesh is unlikely to be a product of catabolic loss or accelerated turnover of muscle proteins. In order to fully understand the degree to which there is isotopic segregation among tissues as animals age it will be necessary to determine the total animal mass balance by nitrogen isotope analysis of all major tissue types, biochemical fractions and excreta.

Weanling Diets Is this decrease in analogous to that following weaning observed in human collagen (Fogel et al. 1989; Katzenberg et al. 1993)? In our study there is no systematic difference between weanling and adult collagen (Fig. 12.4). The mean difference is only –0.14 ± The weanling values are, however, less reliable because they are based on fewer individuals. Unfortunately, it is difficult to evaluate the relationship between weanling diet and

256

Figure 12.4. Relationship between rat weanling and adult collagen adult ages are 21 and 91 to 251 days old, respectively.

STANLEY H. AMBROSE

values. Weanlings and

collagen Milk samples were obtained but were extremely small, and insufficient nitrogen was generated for analysis in some litters. Therefore the relationship between the nitrogen isotope ratios of milk and weanling collagen is not yet well-defined.

CONCLUSIONS We conclude that the diet-collagen spacing in nitrogen isotopes of rats is relatively immune to variation due to protein, heat or water stress. The appropriateness of rats as a model for herbivores or humans must, however, be questioned: Rat appendicular bone has a significantly lower turnover and resorption rate than that of humans (Wolfe and Klein 1996), perhaps because rat cortical bone has a relatively low number of Haversian systems (DeMoss and Wright 1997). Collagen may thus be insensitive to 15N enrichment due to protein recycling because of its very slow turnover rate. Changes in values of collagen, hair and flesh with age were unexpected. The relationships between tissue and age are sometimes counterintuitive and contradictory given previous hypotheses. For example, flesh, which has a higher turnover rate than bone, showed a decrease in with age, while collagen and hair showed slight increases. More research into the roles of heat, water and nutrient stress and age, is needed, but the design of experiments will depend upon the development of new theoretical models of metabolic physiology or use of different model animals.

NITROGEN ISOTOPE RATIOS OF RATS

257

ACKNOWLEDGMENTS The experimental research program was supported by grants from the National Science Foundation (USA) BNS 9010937 and SBR 9212466, and from the University of Illinois Research Board. Lynette Norr, Valerie Williams, Jon Getting and Theresa Schober were instrumental in setting up the heat and water stress experiments, caring for the animals and preparing samples.

REFERENCES Ambrose, S.H. 1991 Effects of diet, climate and physiology on nitrogen isotope ratios in terrestrial foodwebs. Journal of Archaeological Science 18: 293–317. Ambrose, S.H. and DeNiro, MJ. 1986 The isotopic ecology of East African mammals. Oecologia

69: 395–406. Ambrose, S.H. and Norr, L. 1993 Experimental evidence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of bone collagen and carbonate. In Lambert, J. and Grupe, G., eds., Prehistoric Human Bone: Archaeology at the Molecular Level. Berlin, Springer-Verlag: 1–37.

Atkinson, S.A. 1985 Characterization of the non-protein nitrogen fraction of human milk: method-

ological considerations. In Jensen, R.G. and Neville, M.C., eds., Human Lactation. Milk Components and Methodologies. New York, Plenum Press: 39–43. Cormie, A.B. and Schwarcz, H.P 1996 Effects of climate on deer bone and lack of precipitation effects on for animals consuming low amounts of plants. Geochimica et Cosmochimica Acta 60: 4161–4166. DeMoss, D.L. and Wright, G.L. 1997 Analysis of whole skeleton 3H-tetracycline loss as a measure of bone resorption in maturing rats. Calcified Tissue International 61: 412–417. Donovan, S.M., Atkinson, S.A. and Lonnerdal, B. 1986 Total nitrogen and non-protein nitrogen balance in preterm infants fed preterm human milk. In Jensen, R.G. and Neville, M.C., eds.,

Human Lactation 2. Maternal and environmental factors. New York, Plenum Press: 603–610. DeNiro, MJ. and Epstein, S. 1981 Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica et Cosmochimica Acta 45: 341–351. Fogel, M.L., Tuross, N. and Owsley, D.W. 1989 Nitrogen isotope tracers of human lactation in modern and archaeological populations. Annual Report of the Director, Geophysical Labora-

tory, Carnegie Institution, Washington D.C. 1988–1989: 111–117. Gaebler, O.H., Choitz, H.C., Vitti, T.G. and Vukmirovich, R. 1966 Isotope effects in metabolism

of 14N and 15N from unlabeled dietary proteins. Canadian Journal of Biochemistry 44: 1249–1257. Hare, P.E., Fogel, M.L., Stafford, T.W., Jr., Mitchell, A.D. and Hoering, T.C. 1991 The isotopic composition of carbon and nitrogen in individual amino acids isolated from modern and fossil proteins. Journal of Archaeological Science 18: 277–292. Heaton, T.H.E. 1987 The ratios of plants in South Africa and Namibia: relationship to climate and coastal/saline environments. Oecologia 74: 236–246. Heaton, T.H.E., Vogel, J.C., Chevallarie, G. and Collett, G. 1986 Climatic influence on the isotopic composition of bone nitrogen. Nature 322: 822–823. Hobson, K.A. and Clark, R.G. 1992 Assessing avian diets using stable isotopes II: factors influencing diet-tissue fractionation. The Condor 94: 181–188. Katzenberg, M.A., Saunders, S.R. and Fitzgerald, W.R. 1993 Age differences in stable carbon and nitrogen isotope ratios in a population of maize horticulturalists. American Journal of

Physical Anthropology 90: 267–281.

258

STANLEY H. AMBROSE

Koch, P. Behrensmeyer, A.P. and Fogel, M.L. 1991 The isotopic ecology of plants and animals in

Amboseli National Park, Kenya. Annual Report of the Director, Geophysical Laboratory, Carnegie Institution, Washington D.C. 1990–1991: 163–171. Kreitler, C.W. 1975 Determining the Source of Nitrate in Ground Water by Nitrogen Isotope Studies. Report of investigations No. 83. Austin, Bureau of Economic Geology, University of Texas. Livingstone, H.G., Payne, WJ.A. and Friend, M.T. 1962 Urea excretion in ruminants. Nature 194: 1057–1058. Macko, S.A., Fogel Estep, M.L., Engel, M.H. and Hare, PE. 1986 Kinetic fractionation of stable nitrogen isotopes during amino acid transamination. Geochimica et Cosmochimica Acta 50: 2143–2146. Maloiy, G.M.O. 1973a The water metabolism of a small East African antelope: the dik dik. Proceedings of the Royal Society, London B184: 167–178. –––– 1973b Water metabolism of East African ruminants in arid and semi-arid regions. Zeitschrift für Tierzüchtungund und Züchtungsbiologie 90: 219–228. Maloiy, G.M.O., Macfarlane, W.V and Shkolnik, A. 1979 Mammalian herbivores. In Maloiy, G.M.O., ed., Comparative Physiology of Osmoregulation in Animals. Vol. 2. London,

Academic Press: 185–209. Minagawa, M. and Wada, E. 1984 Step-wise enrichment of 115N along food chains: further evidence and the relationship between and animal age. Geochimica et Cosmochimica Acta 48: 1135–1140. Mizutani, H., Fukuda, M. and Kabaya, Y. 1992 13C and 15N enrichment factors of feathers of 11 species of adult birds. Ecology 73: 1391–1395. Richards, P. 1972 Nutritional potential of nitrogen recycling in man. American Journal of Clinical Nutrition 25: 615–625. Rikamaru, T., Fujita, Y., Okuda, T., Kajiwara, N., Date, C, Heywood, P, Alpers, M.P. and Koishi, H. 1985 Utilization of urea nitrogen in Papua New Guinea highlanders. Journal of Nutritional Science and Vitaminology 31: 392–402. Rodière, E., Bocherens, H. Angilbault, J-M. and Mariotti, A. 1996 Particularités isotopiques de l’azote chez le chevreuil (Capreolus capreolus L): Implications pour les reconstitutions paléoenvironnementales. Comptes Rendu de 1’Académie des Sciences de Paris 323: 179–185. Schoeninger, M.J. 1985 Trophic level effects on 15N/14N and 13C/12C ratios in bone collagen and strontium levels in bone mineral. Journal of Human Evolution 14: 515–525. Schoeninger, M.J. and DeNiro, M.J. 1984 Nitrogen and carbon isotopic composition of bone collagen from marine and terrestrial animals. Geochimica et Cosmochimica Acta 48: 625–639. Schuette, S.A., Hegsted, M., Zemel, B. and Linkswiler, H.M. 1981 Renal acid, urinary cyclic AMP, and hydroxyproline excretion as affected by level of protein, sulfur amino acids and phosphorus intake. Journal of Nutrition 111: 2106–2116. Sealy, J.C., Van Der Merwe, N.J., Lee Thorp, J.A. and Lanham.J. 1987 Nitrogen isotope ecology in southern Africa: implications for environmental and dietary tracing. Geochimica et Cosmochimica Acta 51: 2707–2717. Steele, K.W. and Daniel, R.M.J. 1978 Fractionation of nitrogen isotopes by animals: a further complication to the use of variations in the natural abundance of 15N for tracer studies. Journal of Agricultural Science, Cambridge 90: 7–9. Sutoh, M., Koyama, T. and Yoneyama, T. 1987 Variations of natural 15N abundances in the tissues and digesta of domestic animals. Radioisotopes 36: 74–77. Trilok, G. and Draper, H.H. 1989a Effect of a high protein diet on acid-base balance in adult rats. Calcified Tissue International 44: 339–342. –––– 1989b Sources of protein-induced endogenous acid production and excretion by human adults. Calcified Tissue International 44: 335–338. Uezu, N., Yamamoto, S., Rikimaru, T., Kishi, K. and Inoue, G. 1983 Contributions of individual body tissues to nitrogen excretion in adult rats fed protein-deficient diets. Journal of Nutrition 113: 105–114.

NITROGEN ISOTOPE RATIOS OF RATS

259

Vogel, J.C., Talma, A.S., Hall-Martin, AJ. and Viljoen, RJ. 1990 Carbon and nitrogen isotopes in elephants. South African Journal of Science 86: 147–150. Wattiaux, M.A. and Reed, J.D. 1995 Fractionation of nitrogen isotopes by mixed ruminal bacte-

ria. Journal of Animal Science 73: 257–266. Wolfe, M.S., Klein, L. 1996 Sex differences in absolute rates of bone resorption in young rats: appendicular versus axial bones. Calcified Tissue International 59: 51–57. Yoneyama, T., Ohta, Y. and Ohiani, T. 1983 Variations of natural 13C and 15N abundances in the rat tissues and their correlation. Radioisotopes 32: 330–332.

This page intentionally left blank

About the Editors Stanley H. Ambrose is Associate Professor and Director of the Stable Isotope Environmental Biogeochemistry Laboratory in the Department of Anthropology, University of Illinois, Urbana. He has used stable isotope methods extensively to investigate East African ecology and the roles of dietary and environmental change in human evolution, and has used using carbon and nitrogen isotopes to stand isotopic variation in modern African ecosystems and controlled feeding and climate experiments. He received his Ph.D. in anthropology at the University of California, Berkeley, in 1984, and was a Foundation for Research into the Origin of Man Postdoctoral Fellow with Michael J. DeNiro, Department of Earth and Planetary Sciences, University of California, Los Angeles. M. Anne Katzenberg is Professor in the Department of Archaeology, University of Calgary, Calgary, Alberta. She is interested in the interaction of diet and health in the past and has worked extensively with stable isotope methods for reconstructing diet, including infant feeding practices in prehistoric and early historic North Americans. She is also interested in the stable isotope ecology of temperate and boreal forest environments where freshwater resources were important in human subsistence. She received her Ph.D. in Anthropology at the University of Toronto in 1983 and was elected President of the Canadian Association of Physical Anthropology in 1994.

261

This page intentionally left blank

Index Acetic acid, 5, 26, 71, 93, 94, 100, 108 Adults, 5, 6, 16, 162, 255 Africa, 30, 40, 46, 48, 50, 68, 90, 96–99, 109, 118, 136, 244 Agave, 7 Agouti, 25 Agricultural, 2, 40, 41, 49, 56, 165 Agriculture, 24, 34, 35, 144, 161

Aguateca, 35 Alanine, 152, 176, 178, 184, 194 Alaska, 66, 69, 71, 75, 81, 82 Albumin, 180

Alpha-amino butyric acid, 152

Arginine, 152, 197 Arid, 46–48, 68, 73, 96, 105, 109, 122, 130, 135, 244 Armadillo, 19, 24, 25, 27, 31 Asia, 40 Aspartic acid, 152, 154, 176, 178

Atmosphere, 26, 90, 96, 108 Atmospheric, 10, 28, 41, 43, 45, 73, 74, 78, 89, 95, 106, 108, 109, 126

Aurochs, 41 Australia, 121, 123, 249

Austria, 54 Azé Cave, 71

Altar de Sacrificios, 34, 35

Amaranth, 163 Amelogenins, 91 Amino acid, 4, 5, 47, 66, 141, 148, 151–156, 173–186, 189–192, 194–199, 202–204, 207–209, 211, 212, 214, 216, 218, 219, 222, 227–229, 233, 234, 236, 246, 248, 249, 255 Ammonia, 45, 192, 194, 222, 246, 247

Ammonium, 45, 234, 246 Anabolic, 51, 191 Anglican, 2, 9 Anthropogenic, 39, 41, 49, 58 Apatite, 19, 26, 27–29, 31, 50, 51, 65–68, 71– 75, 79, 81, 84, 89–100, 103, 105–110, 123, 126, 147, 162, 174, 189, 191, 195, 199–203, 207, 208, 212, 213, 216–221, 223, 225, 228, 230, 236, 237 Apatite-collagen, 50

Archaic, 10 Arctic, 45, 65, 76, 77, 106, 119

Ba/Ca, 167; see also Barium Baboon, 95 Bacteria, 92, 145, 147, 154, 173, 174, 176, 178– 180, 184, 186, 200, 222, 224, 246

Baking Pot, 34, 35 Barium, 92, 159, 161. 162, 165 Barton Ramie, 34, 35

Beans, 34, 176 Bear, 55, 71, 72, 74–77, 79–81, 84

Beaver, 166 Belgium, 54 Belize, 19, 24, 27–35, 204 Bell Beaker, 180 Belleville, 2–4, 7, 8, 10, 15, 19 Bicarbonate, 7, 12, 28, 50, 105, 189, 191, 200, 204, 223

Biopurification, 160, 163 Birth, 3–5, 16, 161, 248, 250, 253–255

Bison, 41, 46, 71 Boar, 41 263

264

INDEX

Bobcat, 166 Bone, 1–6, 8–11, 14, 15, 17–19, 25–29, 31, 33–

35, 40–42, 44, 47–49, 52–55, 57, 65–69, 71, 72, 74–82, 89–95, 97, 98, 100, 101, 103, 105–110, 119–124, 127–129, 131–135, 141–149, 151–156, 159–167, 173–178, 180–189, 190–192, 195, 196, 199, 200, 203–205, 208, 212, 214, 216, 218, 222, 224, 243, 244, 247, 249, 250, 256 Border Cave, 99–101, 103–109 Brain, 47

Bread, 9, 10 Breast-feeding, 2, 15 Breath, 120, 121, 223, 224 Breccia, 99, 103, 108

Bromine pentafluoride, 126 Bronze, 57 Browser, 11, 66, 89, 95–100, 103–109, 123,

129, 244 Bushbuck, 104 Bushpig, 106 C/N ratio, 26, 52, 66 C3, 9, 10, 19, 26–28, 31, 32, 40, 50, 53, 54, 57, 65–69, 73, 74, 83, 94–96, 98, 104–106, 108, 109, 122, 130, 132, 178, 195, 200– 206, 208, 221, 244, 249, 250 C4, 7–9, 19, 26–32, 39, 40, 49, 50, 66, 81, 83, 95, 96, 98, 103–105, 109, 122, 123, 128, 130–132, 189, 195, 200, 204–206, 221, 244, 249, 250 Ca. See Calcium Cahal Pech, 34 Calcium, 12, 17, 91, 100, 145, 159, 160, 162– 165, 196 CAM, 40 Canada, 1, 2, 10, 18, 49, 73, 141–143 Canis, 25 Canopy effect, 28, 41, 52, 58, 74 Capra, 134, 135 Capreolus, 75 Carbohydrate, 7, 8, 47, 48, 67, 189, 191, 194– 196, 201, 204, 206, 214, 216, 217, 219, 221, 231, 247, 248 Carbon, 1–5, 8, 11, 16, 18, 19, 26, 28–31, 34, 39–44, 46–48, 52, 55, 57, 65, 66, 68, 69, 71–74, 78–81, 83, 84, 89–95, 97–99, 103,

105, 106, 108, 109, 118, 119, 127, 141, 148, 150, 151, 174, 176, 178–185, 189–196, 199–201, 204, 207, 208, 209, 211–213,

Carbon (cont.) 215–222, 224, 227–230, 232, 234–236,

239, 241, 247, 249, 254 Carbonate, 7, 12, 28, 50, 65–67, 71, 72, 74, 79, 89–91, 93, 94, 95, 98, 99, 103, 105, 106, 108, 110, 118, 126, 127, 190–192, 199, 201, 204, 208, 214, 218, 220, 221, 223, 230, 236, 241, 249 Carnitine, 178 Carnivore, 44, 46, 47, 49, 50, 54–56, 65, 67, 68, 70–72, 74–77, 79–81, 83, 84, 100, 160, 161, 163, 166, 189, 194, 200–202, 206, 207, 211, 214, 224, 231, 236, 243, 244 Carnivory, 49, 50, 58, 69, 71, 224, 243 Catabolic, 50, 192, 228, 230, 255

Cattle, 27, 55 Cellulose, 5, 122, 201, 224, 249 Cephalopus, 105

Cervids, 65 Cervus, 93 Chamois, 73 Ci/Ca, 42 Climate, 40, 42, 44, 46, 52, 53–56, 58, 78, 90, 96, 109, 117–120, 126, 128, 136, 142, 243, 244 Climatic, 41–44, 46, 52–57, 67, 73, 135 Coastal, 40, 47, 48, 50, 99, 162, 176, 180 Collagen, 1, 4–6, 8–11, 15, 18, 19, 26–35, 40, 42, 47, 50–52, 55, 57, 65–69, 71–84, 90– 94, 96, 109, 128, 129, 131, 132, 141–143, 147–149, 151–155, 173–186, 189–194, 195, 197, 200–205, 207, 208, 211, 212, 214– 222, 224, 227–230, 235, 236, 238–241, 243, 244, 247, 249–256 Conch, 25

Contamination, 52 Continental, 41, 43, 78, 118, 119, 121 Cooking, 1, 2, 4, 7, 12, 14, 16–18, 224

Copper, 26, 161 Coral reef, 27 Corn, 9, 10, 16, 24, 25, 27, 34, 249 Cortical, 91, 141, 142, 145–147, 152–154, 156, 256 Cow, 74, 212, 223, 224, 243, 246, 248 Crocodile, 126–128, 134, 136 Crystalline, 91–93, 110, 126 Crystallinity, 66, 92, 93, 110 Cuello, 19, 24–29, 31, 32, 34, 205 Cultivation, 24, 45, 57, 161 Cysteine, 197

INDEX

Dasypus, 25 Decomposition, 174, 175, 178, 181, 182, 185 Deer, 19, 25, 27, 28, 41, 45, 46, 55, 68, 78, 93, 122, 123, 244 Defatted, 26 Demineralized, 4 Denitrification, 45 Denmark, 40 Dentine, 65, 68, 72, 74, 75, 79, 82, 83, 91–93, 100, 101, 127 Desert, 117, 124, 128, 130–135, 180 Diagenesis, 89–91, 93, 94, 98, 105, 126, 156, 173, 174, 184 Diagenetic, 66, 80, 90, 92, 94, 97, 108, 109, 142, 156, 162, 173, 175, 183, 186 Didelphus, 25 Die Kelders, 97, 99, 100, 101, 103, 104, 109 Diet-tissue, 27, 243, 244, 245, 246–249, 251, 254 Dikdik, 48, 125, 128–132, 135 Dog, 19, 25, 27–29, 31, 32, 35 Domestic, 24, 25, 48, 56, 57, 124, 126, 133– 135 Dos Pilas, 35 Drought, 124, 125, 132, 135

Ectomycorrhizal, 46 Ecuador, 30, 34 Egypt, 176

Eland, 2, 54, 106 Electron spin resonance, 95, 101, 106, 107 Elk, 27, 55 Enamel, 19, 24, 26, 28, 29, 31, 67, 71, 83, 89, 91, 93–95, 100, 101, 103, 105–110, 117, 119, 121, 122, 124, 126–131, 133, 134, 136

England, 40, 54, 57 Enrichment, 31, 42, 45, 47, 48, 65, 67, 68, 72, 74, 75, 79, 81, 82, 94, 96, 105, 106, 108, 109, 173, 179, 204, 211, 236, 244, 246, 249, 252–256 Enzyme, 47, 183, 185, 189, 190, 198, 199 Equid, 105, 133, 135 Escale Cave, 71 ESR. Set Electron spin resonance Ethiopia, 125 Europe, 39, 40, 42, 44, 46–49, 51, 52, 54, 56– 58, 66, 69, 70, 73, 74, 78, 90, 119 Excretion, 46, 47, 48, 125, 219, 221, 222, 224, 226, 235, 243, 244, 245, 247, 252 Experiments, 2, 7, 29, 67, 174, 178, 181, 182,

265

Experiments (cont.) 185, 193, 199, 206, 213, 232, 234, 236, 238, 243, 245, 247, 248, 250, 252, 253, 256 Fat, 24, 26, 47, 190, 191, 194, 195, 196, 208, 247 Fatty acid, 191, 194, 201, 203, 208, 224 Fe. See Iron Female, 16, 17, 19, 32, 248, 253–255

Fermentation, 7, 200 Fertilizer, 45, 49 Fibril, 91, 142, 183 Fish, 25, 27, 40, 48, 49, 52, 56, 58, 119, 175, 247 Flesh, 26, 51, 201, 202, 204, 206, 208, 224, 251, 253–256 florisbad, 108 Fluorine, 92

Fodder, 25 Food chain, 40, 41, 42, 46, 49, 160, 163, 235, 237

Foodweb, 1, 27, 40, 47, 58, 65–67, 69, 73, 95, 109, 160, 244, 249

Forage, 28 Forest, 19, 24, 27, 28, 34, 41, 45, 73, 74, 75 Fossil fuel, 10, 41, 73, 89, 95, 96, 103, 106 Fourier transform infra-red spectroscopy, 93 Fractionation, 27, 42, 47, 50, 106, 127, 189– 191, 195, 199, 200, 202, 203, 207, 208, 212, 214–216, 218, 220, 222–234, 236, 246, 248, 252

France, 41, 42, 54, 66, 71, 93 French, 46, 49, 71 Freshwater, 19, 25, 49, 56 FTIR. See Fourier transform infra-red spectroscopy

Gazella, 125 Gazelle, 48, 117, 125, 128, 129, 132, 135, 136 Gel electrophoresis, 173, 177, 180, 184 Gender, 32, 161 Germany, 54

Gibnut, 25 Giraffa, 105 Giraffe, 105 Glacial, 95, 96 Glucose, 176, 190, 191, 194–197 Glutamic acid, 152, 154, 176, 178 Glutamine, 184

266

INDEX

Glycine, 152, 184, 192–195, 197, 216, 220, 230 Goat, 126, 128, 134, 135, 161, 246 Graminivorous, 95 Grass, 27, 32, 40, 66, 98, 103, 105, 123, 125, 128, 131, 166, 244 Grazer, 11, 66, 89, 96–99, 103–109, 125, 129, 130 Grazing, 27, 98, 100, 165 Great Britain, 2, 71, 78 Greece, 54, 73

Ipomoea, 24

Grysbok, 103,105

KentÕs Cavern, 71, 75, 78, 79 Kenya, 41, 117, 124, 129–131, 249 Kidney, 47, 234, 245

Hair, 3, 4, 11, 14, 18, 25, 31–33, 47, 48, 221, 222, 244, 249–256 Hare, 166

Haversian, 92, 145, 256 Heating, 5, 7, 127 Herbivore, 26, 44, 46–50, 54–56, 58, 65, 67, 68, 71–81, 83, 84, 95, 96, 103,160, 161, 163, 165, 166, 189, 192–194, 199, 200– 202, 204, 206–208, 211, 214, 231, 236, 243–245, 246, 252, 256 Hipparion, 105 Hippo, 123 Historical, 141–143, 145–147, 149, 151, 152, 154, 156 Histomorphometry, 3 Holmul, 35

Holocene, 41, 42, 44, 54, 56, 67, 96, 100 hominid, 100, 109, 118, 124, 136 Horse, 46, 55, 65, 68, 71, 73, 75, 78, 79 Human, 1–3, 5, 6, 9, 10, 14, 15, 17–19, 24, 26– 29, 31, 32–35, 39, 40–42, 44–46, 48–58, 90, 94, 100, 118, 119, 121–123, 124, 141, 142, 144, 148, 160, 161, 167, 173, 176, 177, 178, 180, 182, 183, 185, 189–193, 196, 205, 206, 207, 208, 222, 243, 247, 255, 256 Humic, 4, 26, 52, 154, 176, 180 Humidity, 42, 43, 117, 122, 123, 131–133, 135 Hunt, 28 Hunter-gatherer, 40, 41, 50 Hydroxyapatite. See Apatite Hydroxyproline, 148, 152, 178, 193, 195 Hyrax, 48 Ice core, 41, 73, 95, 96 Industrial effect, 26, 28, 32, 33 Infant, 2, 3, 5, 8, 14–16, 19, 162, 247, 255 Interglacial, 95

Ireland, 2, 54 Iron, 1, 4, 5, 9, 12, 13, 16–18 Iron Age, 57

Isotope enrichment, 31, 246, 249, 252, 253 Italy, 54 Itzan, 35 Juvenile, 31, 32

Kinetic isotope effect, 47, 192 Kinosternon, 25 Klasies River Mouth, 97, 99–101, 103, 104 Klipspringer, 103 Kudu, 103 Lake Turkana, 117, 124–126, 128, 130, 133–136 Lamanai, 32, 34, 205 Last Glacial, 108 Leaching, 12, 156, 181 Lead, 80, 94, 162

Legumes, 43 Leucine, 196 Lichen, 73, 78 Linear mixing, 189, 191, 199, 204, 208 Lipid, 48, 50, 67, 147–149, 151, 153, 155, 156, 189, 190, 191, 195, 199, 200, 201, 202, 203, 206, 212, 214, 216, 217, 218, 219, 220, 226, 231, 232, 248 Lysine, 196, 197

Magdalenska Gora, 40 Magnesium, 12, 17, 110 Maize, 7, 9, 19, 24–26, 28, 29, 32, 34, 35, 40, 45, 51, 165, 176, 189, 204–206, 208 Makapansgat, 97, 99–101, 103–105, 108, 109

Malanga, 24 Male, 16, 17, 19, 32, 153, 248, 253, 254, 255

Malta, 54 Mammal, 10, 11, 46, 54, 65–70, 75, 78–81, 83, 90, 117, 119, 120, 122, 124, 125, 127, 128, 129, 134, 135, 166, 173, 174, 178–200, 214, 218, 220, 222, 230, 236, 246, 247 Manganese, 162 Mangrove, 27 Manihot, 24 Manioc, 24

INDEX

267

Manuring, 49, 56, 57 Marillac, 46, 71, 75, 78, 79 Marine, 19, 26–28, 30, 39–41, 47, 49, 50, 53, 54, 58, 95, 99, 103, 106, 108, 109, 119, 162, 165, 202, 244

Nitrogen (cont.) 218, 219, 222, 224, 233, 234, 236, 237, 243–250, 252–256

Marsupial, 25, 123 Marten, 175, 178, 180, 184, 186 Martes, 175

Non-essential amino acid, 189, 192, 193, 196,

Maya, 19, 24, 25, 28, 32, 33, 35, 49, 204, 205 Mazama, 25 Meat, 1, 9, 10, 12, 14, 17, 24, 25, 29, 30, 31, 32, 35, 40, 51, 55, 57, 159, 161, 163–166, 175, 202, 205, 206, 244 Medieval, 49 Mediterranean, 45, 48 Mesolithic, 40, 41 Metabolism, 40, 46, 47, 50, 120, 160, 178, 183, 190, 191, 194, 195, 208, 212, 220, 224, 226, 227, 233, 234, 236, 237, 246, 247 Methane, 67, 200, 214, 224, 231, 237 Methanogen, 67, 74, 204, 211 Mexicans, 48 Mg. See Magnesium Mialet Cave, 71 Microbial, 45, 67, 162, 173–175, 177, 178, 181, 183, 185, 247 Milk, 15, 19, 47, 51, 65, 74, 160, 175, 223, 224, 243, 246, 247, 249, 256 Millet, 40, 58 Mineral, 1, 3, 12, 71, 89, 90, 92–94, 107–119, 141, 145, 162–165, 167, 174, 177, 181, 182, 191, 196, 200, 249

Mojo Cay, 34 Mollusc, 25 Moose, 73, 74 Mummies, 11

Muramic acid, 178 Muscle, 47, 148, 216, 244, 247, 249, 254, 255 NCP, 174, 177, 180–182 Neanderthal, 49 Neolithic, 41, 45, 49 Netherlands, 54

Newborn, 6 Nitrate, 45 Nitrification, 45 Nitrogen, 2–4, 8, 9, 11, 14, 16, 19, 26, 28–30, 32–35, 39, 40, 43, 45–49, 51, 56, 58, 66– 68, 71, 73, 75, 141, 148, 150, 151, 174, 176, 178, 179, 180, 185, 196, 211, 214, 216,

Nitrogen fixation, 43 Non-collagenous, 4, 143, 174, 177 204, 208, 211, 219 North America, 10, 40, 69, 70, 122, 123, 244 Nubian, 11 Nursing, 8, 14, 15, 68, 247

Odocoileus virginianus, 25 Omnivore, 44, 54, 55, 106, 167, 201, 202 Omnivorous, 70, 74, 118, 175 Omo River, 125 Ontario, 1, 4, 7, 10 Opossum, 25 Oreotragus, 103 Ornithine, 152 Oryx, 106, 117, 125, 128, 129, 131, 132, 135 Ostwald ripening, 93 Oxygen, 95, 117–123, 125, 126, 130–134, 136 Oxygen Isotope Stage, 108 Oyster, 25, 27 Pacbitun, 32, 35 Pachyrrhizus, 7 Packrat, 96 Paloma, 162

Pancreas, 47 Papio, 95 Papua New Guinea, 247

Pastoralists, 50 Pathology, 2 Peccary, 19, 25, 27 Peru, 162, 176, 180 Peten, 34, 35 pH, 71, 142, 144, 147, 174, 176, 199 Phosphate, 91, 92, 95, 117, 119–123, 126–128, 130–136, 144, 191, 196, 234 Photosynthesis, 40, 42, 57 Photosynthetic, 42, 66, 67, 96 Physiology, 14, 75, 80, 125, 130, 135, 217, 218, 223, 224, 247, 256 Pig, 48, 55, 119, 121, 195, 196, 223

Pimelic acid, 178 Pleistocene, 41, 42, 46, 65, 66, 69, 71, 75, 81, 82, 84, 89, 90, 104, 106, 108, 109, 118, 136 Pliocene, 109, 118 Pomacea, 25

268 Portugal, 40, 49 Potomochoerus, 106 Preclassic, 19, 24–29, 31, 32, 34, 35 Preservation, 4, 5, 65–67, 69, 75, 82, 83, 105, 106, 141–143, 145, 147, 149, 151, 152, 154, 156, 174, 177, 181 Proline, 148, 152, 176, 184, 193, 194, 195, 197 Protein, 3, 4, 8, 11, 15, 19, 24, 29, 45, 48, 50, 51, 56–58, 66, 67, 90, 91, 125, 141, 143, 147, 148, 153, 154, 156, 173–178, 180–185, 189–193, 195–197, 199, 202–206, 208, 212–222, 224, 226–231, 233, 234, 236, 239–241, 243, 244, 246–256

Pulses, 43, 45; see also Legumes Pyrite, 92 Radiocarbon, 52, 66, 78, 94, 101, 106, 109, 110 Rain, 16, 24, 27, 46, 47, 73, 99, 118, 119, 121, 125, 128, 136, 144, 164, 216, 240, 244 Raphicerus, 103, 105 Rb. See rubidium

Reindeer, 46, 73, 74, 78 Rhynchotragus, 125 Rodent, 19, 48, 219 Roe deer, 41, 45, 75 Routing, 189, 190, 195, 197, 199, 202, 204– 206, 208, 213, 215, 219, 220, 226–228, 230–232, 237, 238, 240; see also Scrambling, Linear mixing Rubidium, 5, 14, 17, 18 Ruminant, 48, 67, 125, 200, 201, 211, 214, 222, 224, 231, 237, 247, 252 Russia, 54

Sahara, 68, 93, 107, 109 Salinity, 46, 73 Salmon, 48 Sarcosine, 178 Savanna, 27, 66, 96, 98 Scavenge, 24, 25, 28, 92 Scotland, 53, 54 Scrambled egg model. See Scrambling; also see Routing, Linear mixing Scrambling, 196, 199, 207, 208, 213, 215, 220, 240 Seafood, 40, 50, 165 Seasonal, 11, 42, 118, 119, 125, 128, 250 Sedge, 27 Seibal, 34, 35 Shark, 122 Shell, 25, 27, 40, 99, 119

INDEX Shellfish, 25

Siberia, 66, 69–72, 77, 82 Silver, 26, 126, 127 Silver nitrate, 177

Snail, 25 Snake Hill, 141–146, 152 Social class, 161 Sodium, 7, 100, 161

Sodium bicarbonate, 7 Soil, 14, 45, 49, 56, 73, 75, 97, 99, 105, 119, 132, 142, 144, 147, 153, 154, 162, 173, 174, 176, 180, 185, 244 South Africa, 30, 46, 97, 99, 201 Soybean, 45 Spacing, 1, 10, 11, 18, 32, 33, 47, 50, 51, 65, 68, 79, 80, 189, 211, 214, 221, 223, 230, 231, 236, 237, 244, 252, 254, 256 Spain, 53, 54 Spectrometry, 5 Spondylus, 25 Sr. See Strontium Sr/Ca ratio, 159, 160, 163, 164, 166, 167 Stable isotope, 3–5, 8, 9, 11, 14, 19, 39, 42, 50, 66, 117, 126, 142, 148, 149, 174, 175, 185, 186, 190, 211, 212 Staple, 9, 24, 25, 167 Staples, 9 Starches, 190, 249 Staurotypus, 25

Steenbok, 103, 105 Stomata, 42 Stomatal conductance, 42 Strombus, 25 Strontium, 1, 5, 12–14, 17, 18, 92, 110, 159– 163, 165, 166 Subadults, 6 Subarctic, 45 Sugar, 7–10, 27, 190, 208, 247, 249 Sugar cane, 9, 27 Suid, 106 Sunflower, 7 Swartkrans, 95, 99, 101, 103, 104, 108 Sweden, 54

Sweet potato, 24 Swidden, 24 Syria, 176 Taurotragus, 106 Teeth, 28, 31, 65, 66, 69, 73, 75, 79, 82, 89, 90, 92, 95, 98, 105, 119, 128, 136

INDEX

269

Temperate, 41, 65, 66, 69, 75–82, 124, 135, 142

Urea, 46, 47, 191, 192, 222, 224, 234, 243, 244, 245, 246, 247, 252

Temperature, 42, 53, 66, 73, 96, 105, 109, 117–

Urinary, 120,125, 223, 245

119, 124–127, 131, 132, 134–136, 148, 176, 243, 245, 248–250, 252

Urine, 121, 130, 222, 244

Ursid, 71

Teosinte, 24 Terrestrial, 26, 28, 40, 48, 49, 53, 54, 67, 68, 74, 95, 108, 109, 244 Thalassia, 27 Theropithecus, 95 Threonine, 178, 184, 197, 219 Tooth, 19, 26, 28, 29, 31, 65, 68, 72, 79, 100, 101, 103, 106, 107, 117, 119–122, 124, 126– 131, 133–136, 189 Topi, 125, 130

Trace element, 1, 3, 5, 14, 15, 17, 18, 90, 147, 161, 167 Tragelaphus, 103, 104 Transamination, 47, 196, 219, 233, 246, 248 Trophic effect, 31, 202 Trophic level, 14–16, 19, 46–49, 51, 54, 56, 57, 65, 68, 73, 74, 94, 173, 178, 179, 183,

Vegetables, 1, 9, 10, 12–14, 163, 164 Vegetarians, 55, 56 Volatilization, 12, 45 Water stress, 42, 46, 48, 135, 243–245, 248, 250, 252

Water use efficiency, 42 Waterbuck, 48 Water-dependent, 48 Weaning, 3, 15, 16, 31, 74, 75, 82, 247, 248, 250, 253, 255 Wheat, 9, 16, 176, 249

Wildebeest, 48, 93 Wisconsin, 71, 165, 166 Wolf, 46, 55, 71, 72, 74, 122 Wood, 25. 42, 53, 73, 249

190, 195, 201, 202, 204, 206, 208, 234, 236, 243, 244 Tropical, 24, 27, 40, 66, 124

Xanthosoma, 24

Turkey, 176

X-ray diffraction, 92, 93

Turnover, 93, 124, 131, 222, 247, 253–256 Turtle, 19, 25, 27 Tyrosine, 197

X-ray fluorescence, 5

Uaxactun, 35 United Kingdom, 66 Uranium, 92, 105

X-ray, 5, 92, 145, 147

XRD. See X-ray diffraction Zebra, 117, 125, 128, 130, 131–133, 135, 136 Zinc, 5, 12, 14, 17, 161

Zirconium, 162 Zn. See Zinc