The Pennsylvania State University the Graduate School

The Pennsylvania State University

The Graduate School

College of the Liberal Arts

URBAN POPULATION DYNAMICS IN A PREINDUSTRIAL NEW WORLD CITY:

MORBIDITY, MORTALITY, AND IMMIGRATION IN POSTCLASSIC CHOLULA

A Dissertation in

Anthropology

Meggan M. Bullock Kreger

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

August 2010 The dissertation of Meggan M. Bullock Kreger was reviewed and approved* by the following:

Kenneth Hirth Professor of Anthropology Dissertation Advisor Chair of Committee

George Milner Professor of Anthropology

James W. Wood Professor of Anthropology and Demography

Brian Hesse Professor of Anthropology, Jewish Studies, and Ancient Mediterranean Studies Director of the Jewish Studies Program

Nina Jablonski Professor of Anthropology Head of the Department of Anthropology

*Signatures are on file in the Graduate School.

ABSTRACT

It has long been argued that preindustrial cities were unhealthy environments that facilitated the transmission of infectious disease. As a result, Early Modern London and other

Old World cities were assumed to have had such high mortality rates that deaths greatly surpassed births, resulting in negative intrinsic population growth. Consequently, preindustrial cities were believed to be ―population sinks‖ dependent on immigration from rural areas to increase in size. While this vision of preindustrial cities has gained widespread acceptance, it has not gone unchallenged, on both theoretical and evidentiary grounds.

The current investigation contributes to this debate by examining urban population dynamics in the prehispanic New World urban center of Cholula, Puebla. New World cities differed significantly from those of the Old World, not just in terms of their epidemiological environments, but also in terms of their social, political, and economic organization. A paleodemographic and paleopathological study of 309 Postclassic human skeletons from Cholula was combined with strontium and oxygen isotope analyses in order to characterize morbidity, mortality, and immigration to this prehispanic Mesoamerican city. Several new methodological approaches, including the use of transition analysis, a parametric model of mortality, and a multistate model of health were incorporated into the paleodemographic and paleopathological analyses. Strontium and oxygen isotope studies allowed immigrants in the population to be identified so that their impact on demographic processes in Cholula could be considered. The results of this investigation suggest that the cultural and epidemiological environments of

Cholula contributed to the formation of urban demographic patterns in this New World city that differed from those observed in the Old World.

iii

TABLE OF CONTENTS

LIST OF FIGURES……………………………………………………………………………viii

LIST OF TABLES……………………………………………………………………………....xi

ACKNOWLEDGEMENTS…………………………………………………………………...xiii

CHAPTER I: CULTURAL EVOLUTION AND HUMAN HEALTH………………………1 The Anthropological Debate……………………………………………………………....3 The original affluent society..…………………………………………………...... 3 The transition to agriculture…………………………………………………….....5 The urban revolution……………………………………………………………....6 Challenges to the Theory……………………………………………………………….....8 The evidence…………………………………………………………………...... 8 The theoretical foundation…………………………………………………….....10 The Current Investigation………………………………………………………...... 13 The Organization of the Current Study………………………………………………...... 18

CHAPTER II: URBAN POPULATION DYNAMICS IN PREINDUSTRIAL OLD WORLD CITIES……………………………………………………………………………….21 Early Studies of Mortality in Cities……………………………………………………...22 Reconstructing Life and Death Using Parish Records…………………………………..23 Population Dynamics in Preindustrial Old World Cities………………………………...27 Natural Decrease in Early Modern Cities………………………………………………..30 Mortality in Preindustrial Cities………………………………………………………….33 Infant and childhood mortality…………………………………………………...33 Adult mortality…………………………………………………………………...38 Fertility in Preindustrial Cities…………………………………………………………...39 Immigration………………………………………………………………………………40 Living Conditions in Preindustrial Cities………………………………………………..43 What Can We Conclude about Population Dynamics in Preindustrial Cities?...... 45

CHAPTER III: THE NEW WORLD CITY OF CHOLULA………………………………47 Geography and Climate………………………………………………………………….48 The Prehistory of Cholula in Brief……………………………………………………….52 The Political and Social Organization of Cholula……………………………………….63 The altepetl…………………………………………………………………….....63 Social organization……………………………………………………………….68 Cholula as a Market Center………………………………………………………………76 Cholula as a Religious Center……………………………………………………………77 Not All Preindustrial Cities are Created Equal…………………………………………..78 Mesoamerican urbanism………………………………………………………....78 Cholula in Comparison to Preindustrial Old World Cities………………………………85

Water contamination……………………………………………………………..86 Sewage…………………………………………………………………………...88 Garbage disposal…………………………………………………………………89 Markets…………………………………………………………………………..91 Traders and pilgrims……………………………………………………………..93

CHAPTER IV: THE CHOLULA OSTEOLOGICAL SAMPLE………………………….95 Provenience of the Cholula Osteological Sample………………………………………..96 Preservation Biases……………………………………………………………………..102 Underrepresentation of infants………………………………………………….102 Special mortuary treatments……………………………………………………103

CHAPTER V: RECONSTRUCTING THE PALEODEMOGRAPHY OF CHOLULA..118 A Crash Course in Traditional Paleodemographic Methodology………………………119 Paleodemographic Age-at-Death Distributions………………………………………...121 Age determination in skeletal remains………………………………………….121 Paleodemographic age-at-death distributions and the principle of uniformitarianism……………………………………………………………….123 Reconstructing mortality in skeletal assemblages……………………………...127 Estimating the age-at-death distribution………………………………………..130 Parametric models of mortality…………………………………………………134 The Paleodemography of Cholula……………………………………………………...137 The Siler Model Applied………………………………………………………...... 154 Infant and Childhood Mortality………………………………………………………..159 Young Adult Mortality…………………………………………………………………164 Male and Female Mortality……………………………………………………………..167 A Word about Nonstationarity………………………………………………………….171 A Word about Fertility………………………………………………………………….172

CHAPTER VI: A PALEOPATHOLOGICAL ANALYSIS OF CHOLULA……………174 Inherent Problems of Interpreting Pathological Lesions in Skeletal Remains…………175 Sensitivity and specificity………………………………………………………175 The paradoxical nature of skeletal lesions……………………………………...176 Hidden heterogeneity…………………………………………………………..177 Selective mortality……………………………………………………………...177 Responses to the Osteological Paradox………………………………………………...180 Proposed Solutions to the Osteological Paradox……………………………………….182 Pathological Lesions in the Cholula Assemblage………………………………………186 Porotic hyperostosis and cribra orbitalia………………………………………..186 Enamel hypoplasias…………………………………………………………….196 Proliferative lesions…………………………………………………………….204

CHAPTER VII: MIGRATION IN PREHISPANIC MESOAMERICA………………….217 Possible Explanations for Low Mortality in Young Adults…………………………….218 The lack of epidemic diseases…………………………………………………..218 The scale of migration and the demographic characteristics of immigrants…...219

What is an Immigrant?………………………………………………………………….221 Temporary Immigrants and Visitors to the City………………………………………..223 Pilgrims…………………………………………………………………………223 Merchants and vendors…………………………………………………………226 Rotational labor obligations…………………………………………………….227 Group Migrations……………………………………………………………………….227 Accounts of population origins………………………………………………....228 War and natural disasters……………………………………………………….233 Established colonies…………………………………………………………….234 Ethnic barrios…………………………………………………………………...234 Individual Migrations…………………………………………………………………...238 Political motivations……………………………………………………………239 Marriage………………………………………………………………………...239 Economic factors……………………………………………………………….243 Slavery………………………………………………………………………….245 Migration in old age…………………………………………………………….246 Isotopic evidence of individual migrations……………………………………..246 How would Immigrants have been Received in Cholula?...... 247

CHAPTER VIII: STRONTIUM AND OXYGEN ISOTOPE ANALYSES………………250 The Study of Migration through Strontium and Oxygen Isotopes……………………..251 Strontium isotopes……………………………………………………………...251 Oxygen isotopes………………………………………………………………..254 Isotopic Studies of Migration in Mesoamerica…………………………………………255 Sampling Procedure…………………………………………………………………….258 Tooth enamel…………………………………………………………………...258 Bone……………………………………………………………………………258 Sampling strategy………………………………………………………………259 Analysis of samples…………………………………………………………….261 Limitations and Confounding Factors………………………………………………….262 Limitations common to both strontium and oxygen isotopes…………………..262 Problems associated with strontium isotope analysis…………………………..264 Problems associated with oxygen isotope analysis……………………………..265 Results of Isotopic Analysis…………………………………………………………….266 Variation in the sample…………………………………………………………266 Variation by tooth type…………………………………………………………271 Identifying local values…………………………………………………………274 What Do the Results Suggest about Immigration to Cholula?…………………………285 The scale of immigration to Postclassic Cholula……………………………….285 Mortuary contexts of potentially nonlocal individuals…………………………286 Characteristics of Possible Immigrants…………………………………………………287 Temporal differences…………………………………………………………...287 The demographic characteristics of potential immigrants……………………...289

CHAPTER IX: CONCLUSIONS AND FUTURE RESEARCH…………………………302 Avenues of Future Research……………………………………………………………311

Paleodemographic issues……………………………………………………….312 Immigration……………………………………………………………………..315 Paleopathological issues………………………………………………………..320

BIBLIOGRAPHY……………………………………………………………………………..325

APPENDIX A: THE PROVENIENCE OF SKELETONS………………………………..357

APPENDIX B: AGE AND SEX DATA…………………………………………………….364

APPENDIX C: SEX DETERMINATION………………………………………………….371

APPENDIX D: JUVENILE AGE ESTIMATION…………………………………………375

APPENDIX E: SCORING PROCEDURES FOR TRANSITION ANALYSIS………….378

APPENDIX F: TEST OF PROFICIENCY…………………………………………………382

APPENDIX G: TRADITIONAL AGING METHODS……………………………………384

APPENDIX H: CRIBRA ORBITALIA AND POROTIC HYPEROSTOSIS……………388

APPENDIX I: ENAMEL HYPOPLASIAS…………………………………………………395

APPENDIX J: PROLIFERATIVE LESIONS……………………………………………...402

APPENDIX K: STRONTIUM AND OXYGEN ISOTOPE RESULTS…………………..409

vii

LIST OF FIGURES

Figure 1-1 The archaeological site of Cholula…………………………………………………..2 Figure 1-2 The Great Pyramid of Cholula……………………………………………………...15 Figure 1-3 Map of Mesoamerica………………………………………………………………..16 Figure 1-4 Cranium from Cholula with dental mutilations……………………………………..17 Figure 2-1 Baptisms and burials in London, 1540-1840……………………………………….32 Figure 2-2 Baptisms and burials in York, 1560-1700…………………………………………..32 Figure 3-1 Map of the geographical features of Puebla………………………………………...49 Figure 3-2 Map of the major rivers of Puebla…………………………………………………..50 Figure 3-3 Map showing the location of Cholula………………………………………………51 Figure 3-4 The Great Pyramid and the Patio of the Altars……………………………………..56 Figure 3-5 Construction phases of the Great Pyramid………………………………………….57 Figure 3-6 Map of Cholula from the Relaciones Geográficas, 1581…...... 62 Figure 3-7 The hueyaltepetl of Cholula………………………………………………………...65 Figure 3-8 Modern redrawing of the map from the Cholula Codex……………………………66 Figure 3-9 Map of the Basin of Mexico………………………………………………………...75 Figure 4-1 Photo 1 showing the general area of the habitational units…………………………98 Figure 4-2 Photo 2 showing the general area of the habitational units…………………………98 Figure 4-3 Photo of habitational unit…………………………………………………………....99 Figure 4-4 Drawing of habitational unit………………………………………………………..99 Figure 4-5 Map of the habitational zone………………………………………………………100 Figure 5-1 Age-at-death distributions of two modern anthropological populations…………..125 Figure 5-2 Age-at-death distribution of the Hadza……………………………………………125 Figure 5-3 Age-at-death distribution of three paleodemographic populations………………..126 Figure 5-4 Supraorbital ridge………………………………………………………………….138 Figure 5-5 Gonial angle of the mandible……………………………………………………...138 Figure 5-6 Mastoid process……………………………………………………………………138 Figure 5-7 Occipital protuberance…………………………………………………………….138 Figure 5-8 Ventral arc and subpubic concavity……………………………………………….139 Figure 5-9 Ischiopubic ramus, the greater sciatic notch, and the preauricular sulcus………...139 Figure 5-10 Coronal pterica…………………………………………………………………...142 Figure 5-11 Lambdoidal asterica……………………………………………………………...142 Figure 5-12 Sagittal obelica…………………………………………………………………...142 Figure 5-13 Interpalatine and zygomaticomaxillary………………………………………….142 Figure 5-14 Features of the pubic symphysis used in transition analysis…………………...... 143 Figure 5-15 Features of the iliac portion of the sacroiliac joint used in transition analysis…..144 Figure 5-16 Age-at-death distribution of Cholulteca II skeletons…………………………….147 Figure 5-17 Age-at-death distribution of Cholulteca III skeletons……………………………147 Figure 5-18 Cholula age-at-death distribution constructed using transition analysis…………148 Figure 5-19 Comparison of uniform and archaeological prior distributions………………….148 Figure 5-20 Age-at-death distributions constructed using traditional aging methods………...149 Figure 5-21 Age-at-death distribution from Serrano (1973)…………………………………..149 Figure 5-22 Modal ages at death in different types of societies………………………………152 Figure 5-23 Siler hazard for Cholula………………………………………………………….156

viii

Figure 5-24 Survival curve for Cholula……………………………………………………….157 Figure 5-25 PDF for Cholula………………………………………………………………….158 Figure 5-26 Hazard for individuals over age one……………………………………………..161 Figure 5-27 Juvenile age-at-death distribution………………………………………………..162 Figure 5-28 Cholula age-at-death distribution by sex…………………………………………168 Figure 5-29 Male and female Gompertz-Makeham hazards………………………………….170 Figure 6-1 Usher’s multi-state model of mortality……………………………………………183 Figure 6-2 Active cribra orbitalia……………………………………………………………...187 Figure 6-3 Active porotic hyperostosis………………………………………………………..188 Figure 6-4 Porotic hyperostosis on the occipital and parietal…………………………………189 Figure 6-5 Lesions of cribra orbitalia…………………………………………………………191 Figure 6-6 Lesions of porotic hyperostosis on the parietal……………………………………194 Figure 6-7 Lesions of porotic hyperostosis on the occipital………………………………….194 Figure 6-8 Enamel hypoplasias………………………………………………………………..197 Figure 6-9 Enamel hypoplasias on the maxillary incisor………………………………...……200 Figure 6-10 Enamel hypoplasias on the mandibular canine…………………………………..201 Figure 6-11 Enamel hypoplasias on the mandibular first molar………………………………201 Figure 6-12 Enamel hypoplasias on the mandibular second molar…………………………...202 Figure 6-13 Examples of proliferative lesions………………………………………………...205 Figure 6-14 Proliferative lesions of the humerus……………………………………………...208 Figure 6-15 Proliferative lesions of the radius………………………………………………...208 Figure 6-16 Proliferative lesions of the ulna…………………………………………………..209 Figure 6-17 Proliferative lesions of the femur………………………………………………...209 Figure 6-18 Proliferative lesions of the tibia………………………………………………….210 Figure 6-19 Proliferative lesions of the fibula………………………………………………...210 Figure 7-1 Maps of the Basin of Mexico……………………………………………………...229 Figure 7-2 Map of the Puebla-Tlaxcala region………………………………………………..232 Figure 8-1 Strontium isotope values for sites in Mesoamerica………………………………..253 Figure 8-2 Strontium values arranged in ascending order…………………………………….270 Figure 8-3 Oxygen values arranged in ascending order………………………………………270 Figure 8-4 Adjusted oxygen isotope values arranged in ascending order…………………….273 Figure 8-5 Scatter plot of strontium data……………………………………………………...276 Figure 8-6 Scatter plot of unadjusted oxygen data……………………………………………277 Figure 8-7 Scatter plot of adjusted oxygen data………………………………………………277 Figure 8-8 Scatter plot of strontium against unadjusted oxygen values………………………278 Figure 8-9 Scatter plot of strontium against adjusted oxygen values…………………………278 Figure 8-10 Normal Q-Q plot of strontium values……………………………………………279 Figure 8-11 Normal Q-Q plot of unadjusted oxygen values…………………………………..280 Figure 8-12 Normal Q-Q plot of adjusted oxygen values……………………………………..281 Figure 8-13 Age distribution of those identified as possible immigrants……………………..293 Figure C-1 Features of the skull and scoring system used to determine sex………………….372 Figure C-2 Features of the pelvis used to determine sex……………………………………...373 Figure C-3 Features of the pubic symphysis used to determine sex…………………………..374 Figure D-1 Dental development in juveniles………………………………………………….376 Figure E-1 Screen shot of the ADBOU age estimation program……………………………..380 Figure E-2 Screen shot of the ADBOU age estimation program……………………………..381

Figure G-1 Todd’s ten stage method of aging the pubic symphysis………………………….384 Figure G-2 McKern and Stewart’s three component method of aging the pubic symphysis…385 Figure G-3 Brooks and Suchey’s six phase method of aging the pubic symphysis…………..386 Figure G-4 Lovejoy et al.’s method of aging the auricular surface…………………………..387

LIST OF TABLES

Table 2-1 Juvenile mortality in four London parishes…...... 35 Table 2-2 Juvenile mortality in Geneva in the seventeenth century……………………………35 Table 2-3 Juvenile mortality in four English villages…………………………………………..36 Table 2-4 Juvenile mortality in two parishes in York…………………………………………..36 Table 3-1 The cabeceras and barrios of Cholula……………………………………………….67 Table 3-2 Tribute and labor obligations owed to four tlatoani in Tepeaca……………………..73 Table 5-1 Distribution of Cholula skeletons by time period, sex, and age……………………145 Table 5-2 Number of individuals with each age group………………………………………..146 Table 5-3 Siler parameter estimates for Cholula………………………………………………155 Table 5-4 Siler parameter estimates for individuals over age one…………………………….161 Table 5-5 Female Gompertz-Makeham parameter estimates…………………………………169 Table 5-6 Male Gompertz-Makeham parameter estimates……………………………………169 Table 6-1 Cases of cribra orbitalia…………………………………………………………….191 Table 6-2 Parameter values for cribra orbitalia………………………………………………..192 Table 6-3 Cases of porotic hyperostosis………………………………………………………193 Table 6-4 Parameter values for cases of porotic hyperostosis on the parietal………………...195 Table 6-5 Parameter values for cases of porotic hyperostosis on the occipital……………….195 Table 6-6 Cases of enamel hypoplasias on each tooth………………………………………..200 Table 6-7 Parameter values for the maxillary incisor…………………………………………202 Table 6-8 Parameter values for the mandibular canine………………………………………..203 Table 6-9 Parameter values for the mandibular first molar…………………………………...203 Table 6-10 Parameter values for the mandibular second molar………………………………204 Table 6-11 Cases of proliferative lesions……………………………………………………...207 Table 6-12 Parameter values for the humerus………………………………………………....211 Table 6-13 Parameter values for the radius……………………………………………………211 Table 6-14 Parameter values for the ulna……………………………………………………...212 Table 6-15 Parameter values for the femur……………………………………………………212 Table 6-16 Parameter values for the tibia……………………………………………………..213 Table 6-17 Parameter values for the fibula……………………………………………………213 Table 8-1 Median strontium isotope values by region………………………………………...252 Table 8-2 Descriptive statistics for strontium…………………………………………………267 Table 8-3 Descriptive statistics for oxygen……………………………………………………268 Table 8-4 Descriptive statistics for adjusted oxygen isotope values…………………………..273 Table 8-5 Extreme values of strontium………………………………………………………..281 Table 8-6 Extreme values of oxygen…………………………………………………………..282 Table 8-7 Extreme values of adjusted oxygen………………………………………………...282 Table 8-8 Individuals designated as possibly nonlocal………………………………………..284 Table 8-9 Individuals designated as questionable……………………………………………..284 Table 8-10 Characteristics of the collection, the isotope sample, and nonlocal individuals…..288 Table 8-11 Weighted isotope values by time period…………………………………………..289 Table 8-12 Weighted isotope values by age category………………………………………….292 Table 8-13 Pathologies in possible immigrants……………………………………………….299 Table A-1 Excavation unit in which each burial was found…………………………………..357

Table B-1 Age and sex of each skeleton………………………………………………………364 Table D-1 Juvenile age determination based on epiphyseal closure…………………………..377 Table D-2 Juvenile age determination based on epiphyseal length…………………………...377 Table E-1 Scoring procedures for transition analysis…………………………………………379 Table F-1 Results of a test of accuracy in applying transition analysis……………………….383 Table H-1 Cribra orbitalia and porotic hyperostosis…………………………………………..388 Table I-1 Enamel hypoplasias…………………………………………………………………395 Table J-1 Proliferative lesions………………………………………………………………...402 Table K-1 Strontium and oxygen isotope values……………………………………………...409 Table K-2 Strontium values in ascending order……………………………………………….411 Table K-3 Oxygen values in ascending order…………………………………………………413 Table K-4 Adjusted oxygen values in ascending order……………………………………….414

xii

ACKNOWLEDGEMENTS

I would like to show my appreciation to all of those individuals who have contributed in some way to the completion of this dissertation. Special thanks are owed to my dissertation committee members. Their suggestions, comments, and advice have greatly improved the final product. Ken Hirth provided invaluable guidance, not only on the dissertation itself, but also on navigating the waters of Mexican anthropology. I would also like to thank the other individuals who have so generously offered their professional expertise on the project. The Dirección de

Antropología Física (DAF) of the Instituto Nacional de Antropología e Historia (INAH) allowed me to study and photograph the Cholula skeletal material. José Concepción Jiménez (DAF) facilitated the collection of data from the Cholula skeletons, and David Volcanes (DAF) helped take the pictures of the Cholula skeletal material that are used in the dissertation. Edson Chávez helped take the pictures of the other bones that appear in the dissertation. Thanks also go out to

Sharon DeWitte for providing me with examples of her mle programs and to Darryl Holman for his incredibly helpful advice on writing programs in mle. The strontium and oxygen isotope analyses were done by the Laboratory for Archaeological Chemistry at the University of

Wisconsin-Madison, and James Burton and Doug Price graciously answered all of my questions about isotope analysis. I am indebted to my family, David Kreger, and Edson Chávez for their support and encouragement during my years in graduate school.

This project was funded by grants from the Foundation for the Advancement of

Mesoamerican Studies, Inc., the Wenner-Gren Foundation, and the Research and Graduate

Studies Office of the Pennsylvania State University, and by a Hill Fellowship and a Sanders

Award from the Department of Anthropology of the Pennsylvania State University.

xiii

CHAPTER I

Cultural Evolution and Human Health

“Be happy now, for you shall be long dead.” (Scottish proverb)

Preindustrial urban population dynamics are a controversial topic in both anthropology and historical demography. A number of anthropologists have concluded that the health of past human populations was negatively affected by changes that accompanied the development of greater social, political, and economic complexity (see, for example, Cohen and Armelagos

1984; Cohen 1977, 1989, 2007; Swedlund and Armelagos 1990; Steckel and Rose 2002). The transition to agriculture and the urban revolution are presented as two great stages of cultural evolution that caused dramatic declines in the health of humankind. Preindustrial urban populations are, therefore, assumed to have experienced elevated levels of morbidity and mortality attributable to the high population densities and unsanitary living conditions considered to be an inherent part of city life.

A number of early studies of the demography of preindustrial European cities seemed to lend credence to this school of thought (Wrigley 1969; Perrenoud 1975, 1978; DeVries 1984;

Finlay 1981a, 1981b). These investigations concluded that Old World cities were ―population sinks,‖ in which deaths regularly exceeded births (Wrigley 1969). The high mortality and low fertility levels that were observed led to speculation that population growth in cities occurred solely as the result of immigration from rural areas. While this vision of cities as graveyards has gained widespread acceptance, it has not gone unchallenged, on both theoretical and evidentiary grounds (Wood et al. 1992; Wood 1998; Sharlin 1978, 1981; Galley 1995, 1998; Landers 1991,

1993).

1 The current investigation contributes to this debate by examining urban population dynamics in the prehispanic New World urban center of Cholula, Puebla. New World cities differed significantly from those of the Old World, not just in terms of their epidemiological environments, but also in terms of their social, political, and economic organization. A paleodemographic and paleopathological study of 309 Postclassic human skeletons from Cholula was combined with strontium and oxygen isotope analyses in order to characterize morbidity, mortality, and immigration to this prehispanic Mesoamerican city and to identify how both the cultural and epidemiological environments of Cholula contributed to the formation of urban demographic patterns. In order to better understand the nature of the debate over urban population dynamics, we should first examine in more detail the postulated relationship between health and cultural evolution.

Figure 1-1: The archaeological site of Cholula. A colonial church sits atop the Great Pyramid.

2 The Anthropological Debate

The original affluent society

In 1972, Marshall Sahlins published Stone Age Economics, in which he claimed that hunter-gatherers were the ―original affluent society‖ because, although they had little in the way of material belongings, they were able to obtain what they needed to survive without a great deal of exertion. By the 1960s and 70s, a number of anthropological studies of hunter-gatherers had been carried out that challenged previously held views of these groups as technologically and materially impoverished people constantly living on the edge of survival. These investigations indicated that hunter-gatherers had considerable amounts of leisure time and that they were often very selective in regards to dietary choices. This revised perspective on foraging reversed a long-standing bias in anthropology linking cultural evolution with progress and soon gave rise to a school of thought that claimed that the transition to agriculture and the urban revolution were two giant steps backwards, at least in terms of human health (Cohen 1977, 1989; Cohen and

Armelagos 1984; Swedlund and Armelagos 1990; Steckel and Rose 2002; Polgar 1972; McNeill

1979; Cockburn 1971).

At the heart of this theoretical paradigm is the belief that infectious disease is less prevalent in hunter-gatherer bands than in more sedentary populations as a result of lower population densities and higher mobility. Hunter-gatherers are exposed to zoonoses when they butcher or otherwise come into contact with wild animals, and they suffer from chronic infections, like yaws and tuberculosis, with low virulence and long periods of infectivity.

However, parasite burdens are thought to be low because populations are mobile, and bands are not large enough to maintain crowd diseases at an endemic level. Hunter-gatherers are also

3 assumed to be better nourished as a result of greater consumption of animal protein and a more varied diet (Cohen 1977, 1989).

A number of nutritional, medical, and demographic studies have been cited in support of the position that hunter-gatherers are generally very healthy populations. Investigations (Lee

1972; Tanaka 1976) of !Kung bands concluded that they spent comparatively little time acquiring food, and they exploited only a fraction of the animals and edible plants at their disposal. Yet, their intake of vitamins and minerals met recommended requirements, as did their level of protein consumption. Medical studies of the !Kung have also been used to bolster the claim that hunter-gatherers are well-nourished populations with few infectious diseases

(Truswell and Hansen 1976; Hitchcock 1989; Tobias 1976). While these populations were found to have low body weights and little body fat, they were not deficient in any nutrients, nor did they suffer from diseases, such as anemia and scurvy, caused by malnutrition. Hookworms were the only parasitic infestation of note, although a variety of infectious diseases including streptococcus, gastrointestinal and respiratory infections, tuberculosis, and yaws were endemic in the population. Demographic studies (Howell 1976a, 1979) of the !Kung concluded that the infant mortality rate was only 15% to 21%, with mobile !Kung populations having lower infant mortality than settled cattle post bands. Approximately 8% of the population was over age 60 and the life expectancy at birth was 30 to 35 years old. Based on these data, hunting and gathering is often presumed to represent the golden age of human health, during which time human populations were generally well-nourished and morbidity and mortality was low.

4 The transition to agriculture

The transition to agriculture involved not only a revolution in the means by which humans procured food, but also new developments in the organization of human societies. A number of characteristics of this new way of life are thought to have precipitated a decline in the quality of human diets and an increase in morbidity and mortality. Associated with the transition to agriculture were the concomitant sedentism of populations and the establishment of permanent communities. As people began congregating in villages, population densities increased and sanitation became problematic, both of which were conducive to the transmission of parasites and pathogens (Cohen 1977, 1989; Cohen and Armelagos 1984; Swedlund and Armelagos

1990).

Furthermore, the restricted variability of food supplies in agricultural populations was assumed to have caused nutritional inadequacies and, consequently, reduced immunocompetence, which heighten susceptibility to infectious diseases (Cohen 1977, 1989). In agrarian economies, cereal grains or starches comprise the bulk of the diet. Many of these staple crops are low in protein and other nutrients. Maize, for example, contains phytates that inhibit the absorption of iron and calcium in the gastrointestinal tract, and it is lacking in iron, niacin, and the amino acids lysine and tryptophan (Cohen 1989: 59)1. Rather than exploiting a variety of plants and animals, agricultural populations depended heavily on a single staple crop, which could have resulted in nutritional deficiencies. Because of the greater number of infectious diseases present in human populations and conditions that facilitated the transmission of said infections, adherents of this theory argue that morbidity and mortality must have increased.

1 Although treating maize with lime, as was done in prehispanic Mesoamerica, does improve its nutritional profile by increasing calcium and free niacin content and improving protein quality (Latham 1997).

5 Studies of human skeletal remains are the principal evidentiary source cited as proof of this theoretical position. In 1984, Cohen and Armelagos published Paleopathology at the

Origins of Agriculture, an edited volume containing paleopathological and paleodemographic investigations comparing morbidity and mortality in hunter-gatherer and agricultural populations. The majority of these studies found that agricultural populations showed a higher frequency of pathological lesions, assumed to be indicative of greater morbidity, and lower life expectancy than hunter-gatherer groups. Consequently, the editors of the volume concluded that the transition to agriculture was indeed the stimulus for a deterioration in the health of humanity.

The urban revolution

The process of urbanization was seen as the next step along the path of declining health for human populations, as the formation of cities exacerbated many of the conditions accompanying agriculture that were deemed to be detrimental to humankind’s well-being (Cohen

1989; Storey 1985, 1992). With the evolution of urban centers, humans formed ever-larger settlements. Higher population densities were associated with overcrowded housing, which in turn is assumed to have increased the prevalence of respiratory infections and tuberculosis. The increase in population size is also thought to have resulted in greater exposure to human wastes, which facilitated the transmission of parasites and gastrointestinal infections. Preindustrial cities are said to have had contaminated water supplies, and sewage was not always disposed of in an effective manner, both of which are argued to have increased the number of infectious diseases.

Furthermore, epidemic or crowd diseases, such as smallpox and measles, are only able to establish themselves in a population once a particular threshold population size has been reached.

As these diseases are typically viral in origin and are associated with high mortality rates,

6 infected individuals either die or gain lifetime immunity to subsequent infection. Therefore, a sufficient number of previously unexposed individuals must enter the population for the disease to maintain itself. The rise of cities provides the conditions necessary for these diseases to become established.

Urbanism also brought with it a change in the economic structure of societies.

Occupational specialization became commonplace and urban residents, for the most part, ceased to engage in food-producing activities. As a result, the population of cities became dependent on a market system to obtain basic necessities. In years of poor harvests, grain prices could increase dramatically, jeopardizing food supplies and threatening the nutritional state of the populace

(Cohen 1977, 1989). The reliance on a market economy also promoted long-distance trade, which resulted in new diseases, such as cholera and the Black Death, being introduced into urban populations.

A few paleopathological and paleodemographic studies (Storey 1985, 1992; Márquez et al. 2002; Brothwell 1994; Lewis et al. 1994; Cohen 1989; Márquez and Hernández, eds. 2006) have been cited in support of the stance that urbanization led to a further deterioration in human health, but investigations into the demography of preindustrial Old World cities also appear to bolster this theory (Wrigley 1967, 1969, 1987; DeVries 1984; Perrenoud 1975, 1978; Graunt

1662; Finlay 1981a, 1981b). Many of these early studies into the historical demography of Early

Modern London and other European cities relied upon parish records of baptisms and burials as proxies of vital events. As burials were typically found to have exceeded births, it was surmised that the environs of preindustrial cities were grossly unhealthy and resulted in negative population growth.

7 Challenges to the Theory

Objections to the above school of thought have focused on its theoretical shortcomings as well as problems with the evidentiary sources cited in support. The theory that cultural evolution invariably leads to a decline in the human health condition, in fact, conflates a rather large number of significant changes that accompanied the transition to agriculture and the urban revolution, each of which may have influenced health in very different ways. Moreover, it creates typologies of hunter-gatherer, agricultural, and urban societies by reducing them to a handful of salient characteristics that are assumed to be shared by all societies on that rung of the ladder of cultural evolution. In doing so, this paradigm ignores cultural variability as well as potential variation in cultural responses to the problems that are said to accompany societal evolution.

The evidence

The evidentiary basis of the theory under consideration is not nearly as airtight as supporters would like to claim. In fact, challenges have been raised to the ethnographic, osteological, and demographic studies that have been levied as proof of this paradigm. While some ethnographic investigations of hunter-gatherer population have shown them to have healthy constitutions, other studies of these groups have yielded somewhat less-favorable results.

Silberbauer (1972), studying the Gwi !Kung, found that their diet was likely deficient in fat and some water-soluble vitamins and that they occasionally suffered periods of famine. Wilmsen

(1982) concluded that the !Kung diet was variable over the course of the year and that in particular seasons they ate considerably less than the required 2000 calories per day.

8 With respect to mortality in hunter-gatherer populations, Harpending and Wandsnider

(1982) found that infant mortality rates of the settled cattle post !Kung were half those experienced by their mobile counterparts. Furthermore, Harpending (1976) has noted that although cattle post !Kung populations may appear to be sicker, they actually have lower mortality rates, indicating that frailer individuals are better able to survive. Roth (1985) has presented information on eight hunter-gatherer groups with sedentary counterparts. His data indicate that in six of the eight cases considered, fertility increased and mortality declined with sedentism.

The majority of the paleopathological studies cited in support of this theory have assumed a direct relationship between the frequency of skeletal lesions and the health of populations, so high frequencies of pathological lesions are considered to be indicative of poor health. Unfortunately, skeletal lesions may not be so readily interpretable. Wood et al. (1992) note that when populations vary in their frailty, skeletal samples will be affected by selective mortality. Since skeletal lesions take time to form, the frailest individuals will die before they produce lesions. The presence of pathological lesions in a skeleton may, therefore, actually indicate an individual who was able to survive a disease process longer. A straightforward interpretation of these lesions might not be warranted, calling into question the empirical data supporting this theory. While this caveat certainly does not disprove the idea that cultural evolution is linked with increased morbidity and mortality, it does cast doubt on whether the skeletal data should necessarily be interpreted in the manner that Cohen and Armelagos (1984) suggest. Furthermore, as will be discussed later, changes in fertility may have a greater effect on the age-at-death distribution than changes in mortality. As a result, the decline in life expectancy that reportedly occurred with the adoption of agriculture, according to many of the studies in

9 Cohen and Armelagos (1984), might actually indicate an increase in fertility. These methodological challenges, therefore, render suspect much of the skeletal data cited in support of this school of thought.

The demographic studies of preindustrial Old World cities that lend credence to the argument have also been met with challenges. These studies, as well as the objections to them, will be discussed in greater detail in Chapter II. However, suffice it to say that subsequent investigations, particularly of Early Modern London, have found that the high mortality and low fertility rates noted by Wrigley (1967, 1969) and others was, in fact, confined to particular demographic groups. While young children and rural immigrants experienced high mortality rates, owing primarily to epidemic diseases, mortality rates for other urban residents have been found to be comparable to those of rural areas (Sharlin 1978, 1981; Galley 1995, 1998; Landers

1991, 1993). These findings suggest that preindustrial cities were not necessarily the population sinks they have been assumed to be.

The theoretical foundation

One of the primary assumptions of the theory that cultural evolution precipitates a deterioration in human health is the idea that increased exposure to infectious disease necessarily translates into increased mortality. Proponents of this theory are, quite likely, correct in their assertion that sedentism and increasing population densities result in the evolution of new infectious diseases and permit them to become endemic in a population. Epidemic diseases that require a certain threshold population size in order to maintain themselves are unquestionably an example of this.

10 The issue under contention is whether being exposed to a greater number of infectious diseases necessarily translates into greater morbidity and mortality. A number of studies have demonstrated that the nutritional status of the host is an important variable that should not be overlooked in the equation (Scrimshaw 1977; Scrimshaw and Tejada 1970; Scrimshaw et al.

1968; Santos 1994). Malnutrition compromises immunocompetence, resulting in increased morbidity. Thus, a well-nourished individual is less susceptible to infectious diseases.

Scrimshaw and Tejada (1970), for example, indicate that Guatemalan children provided with nutritional supplements showed reduced rates of infectious disease, although they received no additional medical care.

In preindustrial societies, mortality crises, which are sharp spikes in the number of deaths in a population, often resulted from outbreaks of infectious disease that followed a severe food shortage, suggesting that malnutrition causes greater susceptibility to infection (Galloway 1985,

1988; Dyson 1991; Duncan et al. 1993; Mielke et al. 1984). A study of smallpox epidemics from 1550 to 1800 in London and Penrith, a rural market town, demonstrates a correlation between grain prices and outbreaks of the disease, with high grain prices being associated with an increase in mortality from infectious disease (Duncan et al. 1993). Similarly, the periodicity of smallpox outbreaks in Aland, Finland, from 1751-1890 has been linked to seasonal food shortages (Mielke et al. 1984).

The demographic transition, the shift in the demography of modern populations that is associated with a drop in fertility and an even larger decline in mortality, has been attributed by some to improvements in nutrition. Thomas McKeown (1976) has argued that improvements in medicine and public sanitation do not sufficiently explain the decrease in mortality that occurred with the demographic transition and that the improved nutritional status of populations must

11 have played a significant role in the reduction of death rates from infectious disease. Subsequent research has bolstered McKeown’s position. At the time of the transition, technological advances led to greater production and, thus, falling food prices (Hartwell 1961; Beaver 1975;

Chesnais 1992). Increasing heights and weights corroborate these improvements in the quantity and the quality of food consumed (Waaler 1984; Fogel 2004).

What these data indicate, therefore, is that increased exposure to infectious disease does not lead to increased mortality if the host mounts a sufficient immune response, and the immune response of said host is conditioned upon his or her nutritional status. A study by Dirks (1993) using the Human Relation Area Files, suggests that hunter-gatherer and agricultural populations experienced famines with equal frequency. Consequently, there is little reason to assume that hunter-gatherers necessarily enjoyed greater food reliability or lower rates of malnutrition than agriculturalists. If hunter-gatherers and agriculturalists were, on average, equally prone to crises, mortality rates from infectious disease might have been similar.

Wood et al. (1992), Wood (1998), and others have further argued that increased exposure to a larger number of infectious diseases does not necessarily produce a proportional increase in the mortality rate because different causes of death ―compete‖ with one another, with the frailest individuals being killed off regardless of the number of diseases present. In other words, a certain segment of the population, as a result of genetic, developmental, or environmental factors, is susceptible to falling ill and dying of infectious disease. If Infectious Disease A is the only infectious disease present in the population, all of these frail individuals will die of that infirmity. However, if Infectious Disease B is also present in the population, some of these frail individuals will die of Disease A, and some will die of Disease B, but the mortality rate will not double. Comparing mortality rates from infectious diseases in two hunter-gatherer and two

12 agricultural populations, Wood and Milner (1994) found a similar percentage of deaths

(approximately 70%) from infectious disease in all populations.

Finally, an alternative theoretical model has been proposed by Wood (1998) that describes the relationship between health and economic change. Wood argues that well-being, which he defines as being positively associated with fertility and negatively associated with mortality, should be similar among groups that have reached an equilibrium population size, in which they are at a subsistence level of production (one in which each producer in the population can just survive and reproduce himself). If innovation in a productive system occurs, per capita food supply will increase, resulting in increased well-being and, consequently, population growth. As the population continues to grow, per capita income will decline, as will well-being, until a new equilibrium is reached. According to this model, there should be no difference in the health of hunter-gatherers, agriculturalist, or urban societies that have reached this equilibrium- level of well-being.

The Current Investigation

These challenges to the theoretical and empirical foundations of the argument that morbidity and mortality are inversely related to cultural evolution indicate that this theory has not been proven conclusively. Investigations into the health of preindustrial societies are needed that correct methodological problems associated with previous studies and that take into consideration how environmental and cultural variability may have shaped the morbidity and mortality experience of past populations.

13 Demographic studies of mortality in preindustrial New World cities could provide important insights into the effects of urbanism on the health of human populations. The model of human health discussed above assumes all preindustrial cities shared certain problems, namely high population densities, contaminated water supplies, unsanitary living conditions, and a precarious food supply. However, this vision of cities is based on preindustrial European cities.

In order to demonstrate that characteristics of urban societies resulted in increased mortality, it must first be established that all preindustrial cities, in fact, shared these features. Significant differences in political, economic, and social organization existed in Old and New World cities that could have had dramatic implications for how problems associated with urban living were addressed.

Furthermore, New World urban centers were faced with a very different epidemiological environment than Old World cities. Given that New World populations were spared the devastating effects of epidemic diseases until the arrival of Europeans, the population dynamics of New World urban centers are a means to study urban mortality in the absence of infections such as smallpox and measles. Thus, preindustrial cities in the Americas afford an opportunity to directly test how significant epidemic diseases were in shaping the mortality patterns of preindustrial Old World cities. While some investigators (for example, Storey 1985, 1992;

Márquez et al. 2002; Márquez and Hernández, eds. 2006) have made groundbreaking attempts to address the issue of urban health in the New World using skeletal material, they have faced the methodological challenges mentioned above, particularly in regards to aging adult skeletons, reconstructing mortality, and interpreting skeletal lesions.

In order to better understand how urbanism affected New World populations, 309

Postclassic human skeletons from the archaeological site of Cholula (see map in Figure 1-3)

14 were studied using a variety of new paleodemographic and paleopathological techniques that address concerns raised about traditional osteological methods. Located in the state of Puebla in

Central Mexico, Cholula had a long history of urban occupation. During the Classic period (AD

200-800), Cholula developed rapidly into a politically important ceremonial center that dominated the Puebla-Tlaxcala region (Müller 1973). In the Epiclassic period (AD 800-900), changes in the ceramic tradition and the style of burials indicate that another ethnic group may have populated the site at this time, possibly following a conquest or political takeover at the end of the Classic (Müller 1973: 21). The Early Postclassic (AD 900-1325) ushered in a new era of expansion for Cholula, and by the Late Postclassic (AD 1325-1521) the city reached its maximum size with an urban population estimated to have been between 30,000 and 50,000 individuals in an area of 8 km2 (Müller 1973; Sanders 1971).

Figure 1-2: The Great Pyramid of Cholula.

Figure 1-3: Map of Mesoamerica (Vaillant 1944: ii). North is to the right.

16 The skeletal sample from Cholula used in this analysis consists of 78 Cholulteca II (AD

900-1325) skeletons and 231 Cholulteca III (AD 1325-1500) skeletons excavated during the

1967-1970 field seasons of the Proyecto Cholula. These burials were recovered from beneath house floors and plazas of a low-status Postclassic residential area that overlay Classic Period ceremonial structures near the Great Pyramid (López et al. 1976). The Cholula skeletons were scored for both demographic characteristics and pathological lesions, such as porotic hyperostosis and cribra orbitalia, enamel hypoplasias, and proliferative lesions, so as to evaluate how the urban environment affected morbidity and mortality in this New World population. As immigration strongly influenced the demography of Old World cities, strontium and oxygen isotope analyses were also completed to identify immigrants in the Cholula sample. The results of the paleodemographic, paleopathological, and isotopic studies were then considered in conjunction in order to make some broad generalizations about urban population dynamics in

Cholula. Social, political, and economic features of the culture of Postclassic Cholula influenced population dynamics in the urban center and appear to have resulted in demographic patterns that differ somewhat from those observed in preindustrial Old World cities.

Figure 1-4: The cranium of an individual from Cholula with dental mutilations. Photo taken with the permission of the Dirección de Antropología Física (DAF) of the Instituto Nacional de Antropología e Historia (INAH).

17 The Organization of the Current Study

In evaluating how population dynamics in Cholula compare with those of Old World cities, it is first necessary to assess the merits of the widespread belief that preindustrial

European cities were graveyards. In Chapter II, I discuss in detail the theoretical arguments underpinning this belief, as well as some early studies that appeared to corroborate high mortality levels in urban areas. I also present a number of more recent studies that indicate that urban population dynamics in preindustrial Old World cities are significantly more complex than previously thought and cannot adequately be summed up with the ―population sink‖ model proposed by one demographer (Wrigley 1967). Furthermore, I discuss the significance of cultural factors in shaping the demographic regimes of preindustrial Old World cities and the implications they entail for developing global theories about how urbanization may have affected human mortality.

The focus of Chapter III is Cholula, the urban center that is the subject of the current study. Here, I offer a brief description of the social, political, and economic organization of this

New World city and consider how some of these cultural features may have helped shape the population dynamics of Cholula. In particular, I address some of the factors cited as causes of elevated mortality in preindustrial European cities and explore whether similar issues existed in prehispanic Mesoamerican urban centers.

In Chapter IV, I present the archaeological context of the skeletal collection used in the investigation and examine possible selectivity biases that may affect the representativeness of the sample. I focus primarily on ethnohistorical sources describing Nahua burial practices that

18 suggest that cremation may have been a common mortuary treatment in Postclassic Central

Mexico.

In Chapter V, I present the paleodemographic data collected from the Cholula population.

As methodological problems have hindered previous investigations of archaeological populations, I present the aging technique, transition analysis (Boldsen et al. 2002), that is being used in the current study and discuss how it addresses some of the methodological challenges that arise in reconstructing mortality from skeletal samples. Furthermore, I demonstrate how a mortality model, the Siler model (Gage 1988, 1989, 1990), is superior to life tables in examining mortality in past populations. The demographic patterns observed for Cholula are then compared to the mortality experience of preindustrial Old World cities in order to determine if similarities exist.

In Chapter VI, I discuss how the presence of various pathological lesions of the skeleton in the Cholula population influenced the risk of death. As methodological problems have also plagued studies of morbidity in past populations, consideration is given to the limitations that confront paleodemographers when interpreting pathological lesions of the skeleton. The Usher model (Usher 2000), which addresses some of these concerns, is then used to determine how conditions such as porotic hyperostosis, enamel hypoplasias, and proliferative lesions of the skeleton may have affected the mortality hazard in this population.

Chapter VII continues the investigation into the demography of Cholula with special attention being paid to the issue of immigration, as immigrants featured so prominently in the demographic experience of London and other preindustrial European cities. This topic is first approached from a cultural perspective, and ethnohistoric and archaeological sources are used to speculate about the demographic characteristics of immigrants to Cholula.

19 In Chapter VIII, the results of strontium and oxygen isotope analyses that were undertaken to identify immigrants in the Cholula skeletal collection are presented. Demographic characteristics of those identified as possible immigrants are then considered to understand basic patterns of migration to the city and how they may have been shaped by the culture of

Postclassic Cholula. I then consider ways in which immigrants may have contributed to fertility and mortality in the city.

Finally, in the last chapter, I draw some conclusions about whether demographic patterns in Cholula were comparable to those of preindustrial Old World cities and consider how cultural differences between Old World cities and this New World urban center may have resulted in distinct patterns of fertility and mortality. I then discuss whether a generalized theory defining urban population dynamics is warranted. I also offer some potential avenues for future research, to further clarify urban population dynamics in Cholula and other Mesoamerican cities.

20 CHAPTER II

URBAN POPULATION DYNAMICS IN PREINDUSTRIAL OLD WORLD CITIES

“More were buried than were born within the city walls and in the surrounding suburbs. Without a steady stream of immigrants many, perhaps most, towns before the 19th century would have lost population.” (Wrigley 1987: 136-137)

“As for unhealthiness it may well be supposed that although seasoned bodies may and do live near as long in London, as elsewhere, yet newcomers and children do not….” (Graunt 1662: 45- 46)

Demographic studies of Old World cities have provided much of the fodder for the widespread belief that preindustrial urban centers were cesspools of death and disease.

However, a closer look at the demography of Old World cities, most notably London during the

Early Modern period2, reveal that high mortality was confined to particular demographic groups, and that individual culture histories played a significant role in shaping demographic rates.

Furthermore, although generalizations about the character of preindustrial cities are based largely on European urban centers, these caricatures fail to take into account measures that were put in place to ameliorate the negative conditions associated with urban life. Thus, even preindustrial

Old World cities do not conform to the bleak descriptions of urban centers posited by the theory that health declines with cultural evolution. In order to fully understand the problematic nature of these claims, I will consider a number of investigations of Old World cities in more detail in this chapter.

2 There is some debate over the chronological limits of the Early Modern period. It extended from approximately 1500 to 1800. Most of the investigations of preindustrial population dynamics in English cities focus on this time period for four reasons: (1) substantial urbanization occurred during this time, (2) English parishes began keeping records of baptisms, marriages, and burials starting in 1538, (3) until the mid-seventeenth century, the Black Death dominated demographic events, and (4) it proceeds the beginnings of industrialization (Galley 1998).

21 Early Studies of Mortality in Cities

In 1662 John Graunt presented Natural and Political Observations Made upon the Bills of Mortality, one of the first demographic investigations of vital events in London, which was based upon aggregative data on burials and baptisms contained in the London Bills of Mortality.

Among his pioneering assertions was the claim that burials regularly exceeded baptisms in the city, resulting in a natural decrease3 in the population, stymied only by the significant numbers of immigrants entering London from surrounding rural areas. Graunt attributed the mortality rate in

London, which he found to be much higher than the surrounding rural countryside, to unhealthy conditions in the city, but he recognized that immigrants and children were more likely to be affected by high mortality levels than other demographic groups. Furthermore, Graunt astutely inferred that low fertility in the city, particularly among immigrants, also contributed to the natural decrease he observed in London.

Graunt’s study was merely the first in a long line of demographic investigations of Old

World preindustrial cities that suggested that mortality was density dependent and that urban areas experienced higher death rates because conditions in the city fostered the transmission of disease. In the eighteenth century, a German chaplain named Sussmilich came to a similar conclusion after comparing the records of burials and baptisms in several European cities, and in the nineteenth century, the British demographer William Farr concluded that the size of a community directly affected its mortality rate (cited in Finlay 1981a: 8; Galley 1998: 10). These early studies have been profoundly influential in shaping modern studies of urban demographics, although their conclusions have not always been supported by subsequent research.

3 The term natural decrease refers to negative intrinsic population growth or, in other words, when mortality exceeds fertility.

22 Historical demographers investigating the hypothesis that preindustrial cities had higher mortality rates than rural areas require accurate counts of vital events such as births, deaths, and marriages in order to reconstruct the demography of these populations. While a civil register was created in England in 1837, and death registers were kept in France, Italy, and parts of Germany even earlier, rarely do demographers have at their disposal such detailed data on historical populations (Wrigley and Schofield 1981). Rather, demographic information must be extrapolated and inferred indirectly from other documents such as parish records, Bills of

Mortality, tax lists, censuses, court depositions, and Freemen’s records, to name a few. These sources of data have limitations, which will be addressed below, and in most cases cannot be used to construct the age-specific fertility and mortality rates for all age groups that are necessary to definitively address this question. Furthermore, little can be said about rates of migration, which, as we will see, are also a significant feature of the demography of Early Modern cities.

That said, historical records are often the best or the only source of data available for reconstructing the demography of past populations, and the information that they can provide is invaluable as long as their limitations are sufficiently understood. Therefore, before discussing the conclusions of various studies regarding the demography of Early Modern cities, it is first necessary to explain the means by which these demographic events are reconstructed and the potential problems associated with each.

Reconstructing Life and Death Using Parish Records

Historical demographers often heavily utilize parish registers in the estimation of vital statistics. Although these registers do not provide direct counts of births and deaths, they do record baptisms and burials, which can be used as proxies for the vital events of interest. While

23 these records are essential to historical demography, they are not without challenges for the demographer, given that the information contained within these documents varies greatly in quality and accuracy (Wrigley and Schofield 1981, 1983). Some of the individuals charged with keeping parish records were very conscientious about their work, but others were less so, resulting in many incomplete registries. Seldom do parishes have complete sets of records, and, in general, urban parish registers were not as well maintained as those of rural parishes.

Many individuals were excluded from parish records, both intentionally and unintentionally. Of particular concern with regards to the accuracy of ecclesiastical registers has been the issue of infants dying before baptism. Even Graunt noted this issue in his study of

London demography. Infant mortality is typically high in preindustrial societies, and during some time periods, infants were not necessarily baptized immediately at birth. Death prior to baptism meant that the infant was not included in baptismal records. In some cases unbaptized infants were also refused burial in consecrated ground, resulting in their exclusion from burial records as well and, therefore, their complete invisibility in historical documents.

However, studies of English parish records have indicated that underregistration of baptisms was not a significant problem until the eighteenth century (Wrigley 1975; Finlay 1978).

Prior to that time, children were typically baptized on the Sunday following their birth, unless they happened to be particularly sickly. Calculations by Wrigley (1975) indicate an underregistration rate of 5% or less before 1700. Between 1700 and 1750, when the interval between birth and baptism increased, underregistration of births reached only 7.5%.

Religious nonconformists were also sometimes excluded entirely from Anglican parish registers, but the fact that many nonconformist groups kept their own records ameliorates this problem somewhat. In London, for example, parish records kept prior to 1645 would not have

24 been significantly affected by this issue as, in addition to nonconformist groups maintaining their own registers, demographic events were often recorded in parish records as well (Finlay 1978:

96-97). After the rise of religious independency and the fragmentation of the Church that accompanied it, however, parish records in London became less reliable (Finlay 1978: 96-97).

In other places, such as the Netherlands, the diversity of religious beliefs meant that there was no single authority in charge of registering baptisms and burials, a fact which has had a limiting effect on the reconstruction of demographic events in that country (Finlay 1978: 97).

Despite these concerns, studies have demonstrated that the information contained in parish records is, by and large, accurate (Finlay 1978; Wrigley 1975). Perhaps a greater challenge to reconstructing the demography of preindustrial cities is the fact that the nature of the data limits the amount of demographic information that can be gleaned from them. In using parish records to reconstruct demographic events, two types of data analysis are possible: aggregative analysis and family reconstitution (Wrigley and Schofield 1981). Aggregative analysis refers to the use of total numbers of burials and baptisms. Early studies attempting to determine the demographic regimes of Early Modern cities often used such comparisons. The total number of burials for a given period was compared with the total number of baptisms. If burials were greater than baptisms, natural decrease was assumed.

In practice, this type of analysis yields only limited demographic information. Without knowing the underlying size and age and sex structure of the population, demographic rates, particularly age-specific demographic rates, cannot be reconstructed, which impairs interpretations. For example, if we were to find that the number of deaths doubled over a particular time period, it could indicate an increase in mortality, or it could indicate a doubling in the size of the population. Ages are not typically provided in parish records (Wrigley and

25 Schofield 1981). If ages are given, they may be inaccurate or reflect only broad estimates of the age of the individual when the event occurred. Although censuses that record the size and age structure of the population can be used in conjunction with parish records to calculate demographic rates, the first census of this type was not completed in England until 1801

(Wrigley and Schofield 1981).

A method referred to as back projection uses aggregative data to reconstruct demographic rates by beginning at a point in time in which the size and age structure of the population is known. The number of vital events can then be used to determine underlying fertility and mortality rates, and these can be projected back in time to recreate the age structure of the population a certain number of years previously. However, this method requires that the demographer assume migration rates. Since rates of migration into cities are generally unknown, an incorrect assumption could seriously alter the estimation of other demographic rates. As a result, this method is not advocated for use in urban areas (Wrigley and Schofield 1981, 1983).

Family reconstitution is another method of reconstructing demographic data from parish records (Wrigley 1966; Wrigley and Schofield 1981). In this case, individuals are followed from baptism through marriage, through the baptisms of their children, to their burial, creating records of the demographic life histories of individuals. Since individuals are followed throughout their lives, the ages at which all of these events occur is also known. This is a particularly time consuming process, but the advantage over aggregative analysis is that age-specific rates can be calculated because the age of individuals, as well as the age structure and size of the population at risk, is known (Wrigley 1966).

Although family reconstitution works well in small rural parishes, it can be problematic in urban parishes because of the large number of individuals involved. In addition, high rates of

26 migration in urban parishes result in individuals being observed for only part of their lives, making inter-generational links difficult (Finlay 1978: 111-112). Therefore, their ages may not be known (if they were born elsewhere) or they may pass out of observation before they die (if they migrate out of the parish). Urban studies have, therefore, relied on partial family reconstitutions in which such inter-generational links are not attempted. As a result, the ages of many adults may not be known, making constructions of age-specific fertility and mortality rates impossible. In these cases, only infant and childhood mortality is calculated and marital fertility is estimated from birth intervals. Therein lies one of the primary obstacles to definitively addressing whether preindustrial cities experienced negative population growth due to unhealthy living conditions: Age-specific fertility and mortality rates cannot be reconstructed for some segments of the population, including immigrants, which results in demographers having to extrapolate from the available data.

Population Dynamics in Preindustrial Old World Cities

The demographer Wrigley (1967; 1969) has been one of the most prominent proponents of Graunt’s early assertion that mortality in cities was greater than fertility. He argues that for the population of cities to have increased, immigration must have been substantial. Wrigley attributes the high mortality in cities to a number of factors reminiscent of those presented in the anthropological debate regarding health and cultural evolution:

(1) Preindustrial Old World cities and the surrounding populations to which they were linked

created a large enough network to maintain crowd diseases such as measles and

smallpox. Since epidemic diseases typically cause rapid death or, should the individual

27 survive, grant lifetime immunity to subsequent infection by that particular pathogen, a

sufficient number of new susceptibles must enter the population, either through birth or

immigration, for the disease to become endemic. The evolution of cities enables these

diseases to thrive because they supply the requisite population size.

(2) The large populations and crowded living conditions in cities facilitated the transmission

of infectious diseases. Airborne or droplet infections like tuberculosis and other

respiratory diseases are common causes of morbidity and mortality in developing urban

areas because of the density of habitation. Diseases that rely on fecal-oral transmission

are aided by the contaminated water supplies and unsanitary conditions assumed to be

present in preindustrial cities.

(3) As inhabitants of preindustrial cities in Europe did not generally produce their own food,

they were dependent on the market. In years of poor harvests, Wrigley argues that high

grain prices, combined with declining real wages for occupational specialists who are

unable to sell their goods, resulted in urban residents being unable to purchase sufficient

food. This reduction in food supplies for city dwellers would have led to malnutrition,

which in turn would have made urban residents even more susceptible to succumbing to

infectious disease.

(4) Trade and displacements caused by war or famine facilitated the spread of diseases into

cities, as urban areas were economic centers. Traders attempting to sell their wares in

28 cities were often feared as a potential source of epidemic infections brought from other

locations.

Thus, Wrigley presents preindustrial cities as having such high mortality rates that they are only able to sustain their numbers through large-scale rural-to-urban migration, making them demographic ―sinks‖ for the surrounding rural population (1969). DeVries (1984) echoes

Wrigley’s assertions and proposes a ―Law of Urban Natural Decrease,‖ stating that intrinsic negative population growth meant that cities increased in size solely through substantial numbers of immigrants.

In 1978, Sharlin registered his objection to Wrigley’s theory by arguing that the natural decrease observed in cities was largely attributable to the high mortality rates of immigrants.

Many rural migrants were poor, and, therefore, may have been forced into overcrowded, unsanitary conditions once they reached the city. Their low socioeconomic status may have led to many of them being malnourished and, hence, more susceptible to many infectious diseases.

Most importantly, these individuals often lacked previous exposure to the crowd diseases that were endemic in cities. In addition to their high mortality rates, Sharlin argues that migrants contributed little to fertility in preindustrial cities and were unable to replace themselves. He bases his conclusions on the assumption that migrants were only in the city temporarily and did not marry or reproduce there. As migrants were largely apprentices and servants, cultural and sometimes legal restrictions also dissuaded them from marrying while in the city.

To support his argument, he compares citizens (whom he equates to permanent residents) and non-citizens (whom he equates to migrants) from the city of Frankfort-am-Main, Germany.

His data indicate that while baptisms surpass burials among citizens, the opposite is true for non- citizens. Therefore, he argues, migrants were the primary reason for the apparent natural

29 decrease in cities. It was quickly pointed out that Sharlin had no firm basis for his distinction between permanent residents and migrants and that he lacked sufficient data to conclusively prove his point (Finlay 1981b; Sharlin 1981).

In order to truly understand the urban demographic landscape in preindustrial cities, an understanding of age-specific mortality and fertility rates is required, as is knowledge about the extent of in- and out-migration, and the contribution of immigrants to fertility and mortality in the city. Subsequent investigations into urban population dynamics have been met with challenges related to the limitations of the available data, but some progress has been made in deciphering demographic events in preindustrial cities.

Natural Decrease in Early Modern Cities

While Wrigley (1969) and Sharlin (1978) debated over the reasons for high mortality in

Early Modern London, they were in agreement that natural decrease characterized the demography of this preindustrial city. However, can we conclude, as DeVries (1984) has, that negative intrinsic population growth is an inherent feature of urban life? Demographic studies of eighteenth-century Amsterdam (DeVries 1984) and seventeenth-century Geneva (Perrenoud

1975, 1978) indicate that natural decrease did occur in these cities. However, aggregative data on a number of English cities and towns suggest that while natural decrease occurred in London and other urban locals during some periods of time, natural increase was not impossible. Van der Woude (1982), in fact, suggests that the idea that preindustrial cities were population sinks has been largely based on the demography of London during a very limited period and that it is

30 unlikely this snapshot of the demographic history of one city applies to all preindustrial cities across time and space.

In an examination of the validity of the idea that natural decrease epitomized preindustrial urban population dynamics, Galley (1998: 15-18), relying on the Bills of Mortality, identified three phases in the demographic regime of London during the period 1540 to 1840 (see

Figure 2-1). Until the middle of the seventeenth century, outbreaks of the plague resulted in periodic spikes in mortality in the city, followed by increases in fertility and in-migration. As a result, numbers of births and deaths appear to be fairly balanced in between plague episodes.

However, after the last great mortality crisis caused by the plague in 1665, the number of deaths in London became more constant, and deaths exceeded births until 1790. After 1790, births outnumbered deaths, and the city experienced natural increase. Thus, while natural decrease did occur in London, it was confined to a particular time period and did not comprise the whole of the demographic history of the city.

Similar variability was also noted in other parts of England. York experienced natural increase from the mid-sixteenth until the early seventeenth century, when burials begin to equal baptisms, and by the second half of the seventeenth century, the demography of York was characterized by natural decrease (see Figure 2-2) (Galley 1995, 1998). While natural decrease was typical of some English cities and towns, such as Norwich, Rye, Plymouth, Crediton, and

Maldon, during at least part of the period in question, other cities seem to have experienced natural increase when the plague was not dominating mortality. Colchester, Hull, Worcester, and Reading all experienced natural increase in non-plague years until the early or mid- seventeenth century, and Ipswich and Exeter maintained a favorable balance of births to deaths

Figure 2-1: Baptisms and burials in London from 1540-1840. (Galley 1998: 16)

Figure 2-2: Baptisms and burials in York from 1560-1700. (Galley 1995: 452)

32 until sometime in the eighteenth century (cited in Galley 1998: 18-19). As Galley (1998: 19-

20) states:

It would therefore appear that there was no simple urban demographic pattern at least with respect to natural change. Before 1800 it is certain that universal urban natural decrease did not occur, and by the nineteenth century natural increase had become virtually universal.

Of course, the limitation of such aggregative-level analyses is that the reasons for natural increase or decrease are completely unknown. While the natural decrease observed for London between the mid-seventeenth century and the end of the eighteenth century could have been the result of increases in mortality, it could also have been caused by decreases in fertility, among other things. Studies relying on partial family reconstitution offer insights into some of the age- specific vital rates that contributed to these aggregate-level patterns

Mortality in Preindustrial Cities

Infant and early childhood mortality

Infant and early childhood mortality rates can be calculated from urban parish registers through the method of partial family reconstitution, and studies of preindustrial cities have generally shown mortality in this age group to have been high. DeVries’ reconstruction of the demography of Amsterdam reveals that the surplus of burials over baptisms in the city is primarily the result of elevated infant and early childhood mortality rates, with almost 30% of infants dying before they reached the age of one (DeVries 1984). Perrenoud (1975, 1978) reports similar rates of infant mortality in seventeenth-century Geneva, although he does note that mortality differed by social class (Table 2-2).

33 With a population of 200,000 by the seventeenth century, London was one of the largest preindustrial cities and effectively dominated both the demographic and socioeconomic landscape of England during the Early Modern period (Jones and Judges 1935). Consequently, several studies of urban demography have centered on the city. An investigation of London conducted by Finlay (1981a) compared the demography of four urban parishes of different socioeconomic statuses during the sixteenth and seventeenth centuries. He made two significant findings regarding infant and childhood mortality in London. First, juvenile mortality was much higher in the city than in other parts of England, but it was not as high as mortality rates reported for Geneva (Tables 2-1 and 2-2). Second, the two poorer parishes had higher mortality rates than the two wealthier parishes, with St. Mary Somerset, a poor riverside parish, reporting the greatest juvenile mortality.

While, in general, infant and childhood mortality was higher in London than in a number of English towns, when the socioeconomic status of the individual parishes is taken into account, a slightly different picture emerges. Although the two poor parishes had significantly higher infant and childhood mortality rates than towns, the two wealthy parishes had infant mortality rates comparable to those experienced in villages (Table 2-3). However, childhood mortality in the two wealthier parishes is much higher, and the mortality rate for children ages one to fourteen in the wealthier parishes greatly surpasses that of towns. Finlay speculates that the low infant mortality in the two wealthier parishes may be the result of infants from this social class being sent to wet nurses outside of the city, which would cause their deaths to go unrecorded in parish registers. In addition, he suggests that environmental variability within the city itself might have influenced mortality rates, as considerable differences exist in the number of deaths experienced by St. Mary Somerset and Allhallows London Wall, the other poor parish. St. Mary

34 Somerset was near the Thames, which served as both a sewer and a source of water, perhaps explaining the higher mortality rates that affected the residents of this parish.

Table 2-1: Juvenile mortality rate (deaths per 1000) in four London parishes. St. Peter Cornhill and St. Michael Cornhill were wealthy parishes, while St. Mary Somerset and Allhallows London Wall were poorer parishes (Finlay 1981). Note that the highest infant (265/1000) and early childhood (246/1000) mortality rates were in St. Mary Somerset, and the lowest infant (105/1000) and early childhood (148/1000) mortality rates were in St. Peter Cornhill. Age St. Peter St. Michael St. Mary Allhallows Cornhill, 1580- Cornhill, 1580- Somerset, 1605- London Wall, 1650 1650 1653 1570-1636 0 105 139 265 184

1-4 148 156 246 178

5-9 46 83 93 104

10-14 31 34 69 56

Table 2-2: Juvenile mortality rate (deaths per 1000) in Geneva in the seventeenth century. (Perrenoud 1978) Age 1625-1649 1650-1674 1675-1699

0 271 265 246

1-14 386 339 346

35 Table 2-3: Juvenile mortality rate (deaths per 1000) in four English villages (Finlay 1981: 85). Compare to juvenile mortality rates in London (Table 2-1), Geneva (Table 2-2), and York (Table 2-4). Age Bottesford, Colyton, Colyton, Shepshed, Terling, Terling, 1600-1649 1538-1599 1600-1649 1600-1699 1550-1624 1625-1699 0 160 120-140 126-158 126 128 124

1-14 154 124 176 119 142 143

Table 2-4: Juvenile mortality rate (deaths per 1000) in two parishes in York. (Galley 1998) Age St. Martin Coney St., St. Michael le St. Michael le 1561-1700 Belfrey, 1565-1602 Belfrey, 1600-1720 0 257 234 233

1-4 180 121 198

5-9 95 65 41

10-14 58 -- --

Landers (1991, 1993) study of eighteenth-century vital registers kept by the London

Quakers reveals a similar pattern of mortality. Infant and childhood mortality is extremely high, principally between the ages of one to four. In an examination of the seasonality of infant burials, Landers notes that gastrointestinal infections were a significant cause of death for infants, especially in the first month of life, suggesting that artificial feeding may have been a common practice. Respiratory infections also contributed to the mortality rates of older infants.

The excess childhood mortality that both Landers and Finlay calculate for London is particularly notable in comparison to rates in the rest of the country. As the Quaker registers provide information on cause of death, Landers was able to determine that smallpox was largely

36 responsible for the excess mortality observed in children. In addition, like Finlay, he observed that environmental variation within the city, namely the quality of the water supply, seemed to be connected to mortality rates.

Although York, with a population of approximately 12,000, was significantly smaller than London, it also experienced high early childhood mortality from the mid-seventeenth century to the early-eighteenth centuries, during the time that the city was undergoing natural decrease (Table 2-4). Galley (1995, 1998) notes that while infant and adult mortality appears to remain stable from the mid-sixteenth century to the eighteenth century, childhood mortality skyrocketed in the second half of the seventeenth century. Information on cause of death suggests that smallpox and measles were responsible for the majority of deaths, which is consistent with the stability observed in infant and adult mortality. The protective effects of breastfeeding insulate infants from viral diseases of this nature, and adults remain largely unaffected because they would have been exposed to these illnesses during childhood. While infant and childhood mortality rates exceeded those of small rural villages, they were lower than those of London, suggesting that mortality may, in fact, have a density-dependent component.

Galley proposes, based on the seasonality of burials, that the high infant mortality rate (260 per

1000) was due, at least in part, to gastrointestinal infections. Like Landers, he speculates that denying infants colostrum, artificial feeding, or giving infants water may have contributed to this trend.

37 Adult mortality

Typically, it is not possible to reconstruct age-specific mortality rates for adults in urban parishes because the age at death of adults is not recorded in ecclesiastical records. Historical demographers using family reconstitution can determine the age at death of an individual who was baptized and buried within the same parish, but the high rates of migration in urban parishes mean that many adults did not die in the parish in which they were born. In rare instances, ages at death are recorded in parish registers, but they may represent only broad estimates of age, requiring that caution be exercised when relying on them. As a result, comparatively little is known about adult mortality in preindustrial cities. In Amsterdam, high mortality rates appear to have been confined to children under ten years old (DeVries 1984: 193). For the London

Quakers, Landers (1993: 157-159) was able to make some calculations of adult mortality by considering only the burials of married individuals and assuming both a worst-case and best-case scenario for those who passed out of observation. As only married people are included in this analysis, it precludes the reconstruction of mortality for adolescents and young adults. Landers reports that while life expectancy at birth for the London Quakers was significantly lower than in

England as a whole, life expectancy at age 30 was similar, suggesting that adult mortality in

London was not appreciably different than in the rest of the country. For York, Galley (1998:

108-113) examined adult mortality for St. Michel le Belfrey between 1571 and 1586, as age at death was reported in the registers for this parish during that time period. The age distribution of deaths indicated that 15% of the population lived past the age of 60. By comparing the age-at- death distribution with a model life table, he inferred that life expectancy at age 20 was 55.

Furthermore, he notes that the proportion of adult burials was stable from 1571 to 1700,

38 suggesting that adult mortality did not increase significantly during this period, despite the fact that York moved from a regime of natural increase to one of natural decrease.

Information on cause of death, available in some registers, provides revealing, albeit indirect, data on mortality in adolescents and young adults. In considering the distribution of smallpox deaths by age for the London Quakers, Landers (1993: 152-156) discovered that 25 to

35% of smallpox burials involved adolescents and young adults. After age 30, that number declined substantially. Moreover, Landers was unable to link any of the adolescent or young adults who died from smallpox to an entry in the birth registers. He, therefore, surmises from the apparent lack of previous exposure to this epidemic disease among those in this age group that these individuals must have been immigrants. In York, only a small percentage of smallpox deaths occurred in individuals over age ten, but Galley (1998: 105-106) indicates that most of these deaths were of individuals age 15 to 24, again suggesting that immigrants to York were susceptible to this disease. Other studies have also suggested a high mortality rate for immigrants based on the fact that as much as 15% of apprentices in London died before they completed their service (Rappaport 1989: 313; Finlay 1981a: 66-67).

Fertility in Preindustrial Cities

Although high mortality rates have received much of the blame for periods of natural decrease in preindustrial cities, declines in fertility also played an important role. Finlay (1981a) reconstructed marital fertility rates in London from birth intervals and found that marital fertility was high in the city and rates of illegitimacy were low. However, he proposes, based on a study of the comparative ages of marriage for immigrants and native residents, that immigrant women

39 married later and had fewer children than London-born women. Consequently, total fertility in

London was not high enough to surpass the high mortality rates, resulting in natural decrease.

Galley (1995, 1998) argues that in York, fertility must have declined during the second half of the seventeenth century to account for the surplus of burials over baptisms. As both levels of marital and illegitimate fertility were constant over the period 1561 to 1700, the natural decrease that York experienced in the latter part of the seventeenth century must have been caused by a reduction in the percentage of married individuals. In fact, the sex ratio during this period is unbalanced, with more females than males. Galley suggests that among younger adults, higher levels of female in-migration may have caused an even greater discrepancy between males and females in the young adult ages. As a result, women in York may have had fewer opportunities to marry. Galley adds that a decline in the economy in the latter seventeenth century might have also contributed to a reduction in the number of marriages as a result of an economic inability to establish new households.

Immigration

While immigrants figured prominently into the fertility and mortality experience of preindustrial English cities, comparatively little is known about immigration during the Early

Modern period because of the invisibility of immigrants in most of the available demographic records. Migration into preindustrial cities has been classified into two broad categories: subsistence migration and betterment migration (Luu 2005; Clark and Souden 1988; Whyte

2000). Subsistence migrants were generally among the poorest of the poor, and frequently became vagrants, prostitutes, or petty criminals upon arriving in the city. These individuals were

40 often motivated to migrate by a combination of push and pull factors. The lack of employment opportunities in their native community, famine, or even war might have forced them from their homes. Cities, as economic and political centers, provided glimmers of hope as places to find work or to avail themselves of poor relief. Betterment migrants were typically inspired less by desperation and more by the promise of economic opportunities that cities offered. These individuals often moved to the city to work as apprentices or servants. While they might choose to return to their birthplace once their period of service had been completed, some also settled permanently in their new homes. Obviously, the above classification does not capture all of the potential migration that occurred during the Early Modern period, but it provides a general framework for understanding the most common reasons for moving to preindustrial cities and the demographic profile of the majority of migrants.

While information about migration must be inferred indirectly from the available sources, rates of immigration to Early Modern London appear to have been substantial. At the beginning of the seventeenth century, apprentices made up approximately 15% of the population of the city, and a significant number of these individuals are believed to have been immigrants (Finlay

1981a: 66-67). By the beginning of the eighteenth century, this figure had been reduced to only

5% of the population, but rates of immigration still appear to have been impressive. Based on the age structure of the city at this time, nearly 60% of London’s population was between 20 and 49 years old, and a reconstruction of net migration indicates a substantial stream of individuals in their teens and twenties entering the city (Landers 1993: 181-183).

Detailed information on the sex structure of migration is not available for the Early

Modern period, but some general patterns can be observed. The migration of women, in particular, is difficult to quantify, as they tended to work as servants upon arriving in the city

41 and, thus, frequently do not appear on the few sources, such as apprenticeship or Freemen’s registers, documenting male migrants. However, comparisons of the sex ratio of burials or of the living population, as calculated from parish registers, can provide information into patterns of migration. In the mid seventeenth century, the London Bills of Mortality indicate a sex ratio of

113 males per 100 females, indicating that males were migrating in larger numbers (Finlay

1981a: 140). This is consistent with the figures cited by Finlay that indicate a large percentage of apprentices in the population around this time. By the end of the seventeenth century, however, the sex pattern of migration appears to have changed. In London, as well as Ipswich,

Lichfield, Bristol, Gloucester, and other English cities, the sex ratio indicates a surplus of females (Finlay 1981a: 140-141; Galley 1998: 25-26). This shift may reflect the fact that female servants were in demand, or it may indicate that male labor was in greater demand in the countryside (Galley 1998: 25-26). The immigration of families or of vagrants or otherwise marginalized individuals went largely unrecorded.

Clearly, immigration factored heavily into preindustrial urban population dynamics. The economic structure of preindustrial English cities supported high rates of immigration and influenced the demographic characteristics of migrants. These factors, in turn, had ramifications for mortality and fertility in urban areas. As many immigrants would have lacked previous exposure to epidemic diseases such as smallpox, the epidemiological environment of cities resulted in elevated mortality rates in this segment of the population. Given that immigrants were often met with less than adequate living conditions and little familial support upon arriving to the city, they may have been particularly susceptible to succumbing to the urban disease pool.

Furthermore, the sex pattern of migration affected fertility by creating an imbalance in males and females and, thereby, influencing the proportion of those married. Economic or legal

42 impediments to marrying prior to finishing periods of service may have also had repercussions for fertility. While Sharlin’s assertion (1978) does not appear to be completely true, immigrants in London during particular time periods contributed to the high mortality rates in the city while contributing less to fertility than native residents. However, given that social and economic factors featured heavily in creating these dynamics, this scenario is unlikely to apply all preindustrial cities across time and space.

Living Conditions in Preindustrial Cities

While living conditions in preindustrial cities were undoubtedly less than hygienic, attempts were made to deal with some of the sources of infectious disease cited by Wrigley

(1969). Beginning as early as the thirteenth century, London tried putting in place measures to clean the streets and dispose of garbage (Weinstein 1991). In the mid-seventeenth century, 400 individuals were hired as scavengers and rakers to remove refuse. In 1671, the Sewers and Street

Act was implemented to help ensure that streets were free of garbage, that sewers to drain excess water were in good repair, and that the streets were well maintained. Although the Thames, polluted by certain industries and contaminated by sewage, was a major source of water in Early

Modern London, some private companies also offered water piped in from springs outside of the city. Of course, the fees charged for piped water meant that it was unavailable to many

Londoners. While concern clearly existed over these health issues, London was hampered from effectively dealing with them by a division in its administrative responsibilities. This lack of centralized authority meant that it was difficult to enforce measures aimed at improving sanitation and living conditions in the city.

43 The governing bodies of preindustrial Old World cities also attempted to head off mortality crises by limiting the spread of infectious disease, particularly during epidemics and outbreaks of the plague, by forbidding traders, vagrants, and refugees from entering cities. Trade was, in fact, a well-recognized risk to cities during outbreaks of the plague. Galley (1998: 77) reports that during 1563-64, 1579-80, and 1603-1604 the city council of York implemented restrictions on trading with areas infected with this disease in an attempt to protect both its citizens and the economic prosperity of the city. Unfortunately, such restrictions were not always successful, as traders often disregarded threats of prosecution and found ways to bring their goods into the cities despite these measures. While York successfully managed to avoid the plague during 1563-64 and 1579-80, they were not as fortunate during the outbreak that occurred at the beginning of the seventeenth century (Galley 1998: 77-78).

There was also concern with ensuring that citizens were fed. In general, England had a very comprehensive system of poor relief in the sixteenth and seventeenth centuries (Solar

1995). Because it was financed by local income taxes, English poor relief was widely and consistently available. The guidelines regarding eligibility were fairly broad, allowing the elderly, widows, the sick or disabled, and the seasonally unemployed to collect. Other European cities also offered poor relief, but their systems were not necessarily as generous or well- organized.

During times of famine, the governments of cities would intervene to halt the rising costs of grain and to provide additional relief to the poor. Facing a potential famine in 1588, the

Council of York sought to avert disaster by preventing food shortages in the city. They put into effect laws to maintain affordable grain prices. They provided food to the old and infirm, and those who were unable to work. The unemployed were provided with jobs. Migrants seeking

44 charity were turned away from the city, and other measures were put in place to maintain the supply of grain to York (Galley 1998: 88-89). The Council’s response to this looming economic crisis was typical in the city during the late sixteenth and seventeenth centuries, and the government of York clearly attempted to prevent mortality crises from befalling its citizens

(Galley 1998: 90).

What Can We Conclude about Population Dynamics in Preindustrial Cities?

Although early studies regarding population dynamics in preindustrial cities assumed that high mortality and natural decrease were inherent features of urban demography, further investigations have revealed this to be an overly-simplistic vision, primarily influenced by a limited knowledge of London’s demography during a particular point in time. In fact, not all preindustrial European cities experienced consistent natural decrease. London itself did experience periods of natural increase, suggesting that the ―Law of Urban Natural Decrease‖ is very much inaccurate. Infant and childhood mortality rates were high in cities, principally as a result of infant feeding practices and the effects of epidemic diseases, particularly smallpox.

Rural immigrants to cities also experienced high mortality rates, resulting from their lack of immunity to this virulent disease. However, mortality rates were also affected by socioeconomic class and environmental variation within the city.

Several investigations have also demonstrated that fertility and migration were extremely important in shaping the urban demographic regime. During periods of natural decrease, reductions of fertility were just as responsible as increased mortality for creating negative intrinsic population growth. High rates of migration into cities also had a profound influence on

45 population dynamics, as the contribution of these individuals to fertility as well as mortality could significantly affect demographic rates.

While epidemiological and environmental factors are important in shaping the demography of urban populations, social, economic, and political variables are equally influential and should not be discounted. Cultural preferences resulted in some infants in Early

Modern England not being breastfed, which in turn led to higher infant mortality rates. Social and economic characteristics of English society resulted in large numbers of young adults migrating to London, they influenced how the immigrants were received once they arrived in the city, and they reduced fertility levels of immigrants. Therefore, many of the important features of the demography of preindustrial London are dependent upon cultural factors particular to

English society during that period in history. The governments of preindustrial European cities also attempted to take measures to improve living conditions and to avert mortality crises.

Sometimes these provisions were successful and sometimes they were not, but clearly the political system in place influenced demographic events.

Given the information above, why would we expect all preindustrial cities to share a single demographic regime when cultural factors play such an influential role in shaping patterns of fertility and mortality? New World cities faced very different environmental, epidemiological, political, social, and economic conditions that could have affected demographic events in very different ways. While Old and New World cities may share certain fundamental characteristics, such as high population densities (and even this is debatable), are those few similarities so significant to shaping morbidity and mortality that they override all the differences? In the following chapter, I consider what Postclassic Cholula would have been like, and how those characteristics might have influenced its demographic regime.

46 CHAPTER III

THE NEW WORLD CITY OF CHOLULA

“…es la ciudad más hermosa de fuera que hay en España….” (Cortés 1985, Segunda Carta: 111)

As Cortés and his army marched across Mexico, they were alerted by the inhabitants of

Tlaxcala to the nearby enemy city of Cholula, which the Tlaxcalans claimed was aligned with

Moctezuma and planning an ambush of the Spaniards. Although aware of the trap the Cholulans were setting, Cortés chose to proceed to the city and assure their loyalty to the Spanish throne.

In his letters to the king of Spain describing Cholula’s betrayal and the massacre that ensued in the city, Cortés gave the Old World its first glimpse of the thriving metropolis that was Cholula.

Given its status as a pilgrimage center, one of the most striking features of the city was the sheer number of temples – more than 430 by Cortés’ count. According to the conquistador’s estimates, some 20,000 houses were located in the heart of the city and an equal number were scattered throughout the surrounding countryside. Cortés proposed to the king that Cholula, because of its suitable pastureland, would be a promising place for the Spanish to settle in the New World.

While Cortés was impressed by the beauty of the city, he did not fail to notice its shortcomings.

Although a significant amount of arable land was irrigated, the area was so heavily populated that Cholula had many poor people who often lacked food (Cortés 1985, Segunda Carta: 105-

113).

In Chapter III, I briefly outline the prehistory of Cholula for those readers unfamiliar with

Mesoamerican archaeology. Then, I summarize what is known of Cholula’s political, social, and

47 economic organization at the time of Conquest, focusing on those issues that are especially pertinent to urban population dynamics. Finally, I discuss how Cholula and other Mesoamerican urban centers compare to those of the Old World and what these differences might imply about the theory that preindustrial cities were necessarily unhealthy environments.

Geography and Climate

Cholula, nestled within a valley in the modern state of Puebla, sits at crossroads between the Basin of Mexico, the Valley of Oaxaca, and the Gulf Coast, a fact that certainly played a role in its development as a major prehispanic mercantile center (see Figure 3-1, 3-2, and 3-3).

Bordered on the west by the Sierra Nevada and on the east by the volcano La Malinche, the

Puebla-Tlaxcala Valley covers the western and central portions of the state of Puebla and the southern part of the state of Tlaxcala. The floor of the valley is an expansive plain crossed by a number of small rivers, streams, and tributaries, the most significant being the Atoyac River, which is fed by runoff from the snow-capped peaks of Iztaccihuatl, La Malinche, and the active volcano Popocatépetl (Bonfil 1973: 19). These abundant water sources permitted irrigation of the valley floor, and the convergence of several of these streams in the northeast of Cholula formed a shallow lake, which may have been used for chinampa4 agriculture (Mountjoy and

Peterson 1973; Bonfil 1973: 22; Messmacher 1967; McCafferty 2001).

Cholula has a temperate climate with an average temperature of 18 to 20°C and typically no more than 20 to 40 days of frost per year. Like the nearby Basin of Mexico, the Puebla-

Tlaxcala Valley experiences a wet season that lasts from May to October. During these months,

4 Chinampas are a form of intensive irrigation agriculture practiced in Mesoamerica. Raised beds are created in shallow lakes or swamps by piling up rich organic earth from the lake bottom into rectangular fields separated by canals. Because of the organic matter in the soil and the abundant water, chinampas are highly productive.

48 Cholula receives an average of 800 to 900mm of rainfall per year (Bonfil 1973: 22-23). The temperate climate and plentiful water resources ensured that the region was highly productive in prehispanic and colonial times (Bonfil 1973: 22-24; Cortés 1985, Segunda Carta: 105-113).

Figure 3-1: Map showing the major geographical features of Puebla. (Lomelí 2001: 21)

Figure 3-2: Map showing the major rivers of Puebla. (Lomelí 2001: 24)

Figure 3-3: Map of the state of Puebla showing the location of Cholula. (Lomelí 2001: 27)

51 The Prehistory of Cholula in Brief

Compared to the abundant ethnohistoric accounts of Tenochtitlan-Tlatelolco, precious few detailed descriptions of Contact-era Cholula exist, and most of those focus on its importance as a pilgrimage center. While Cholula is assumed to have played an important role in the political landscape of prehispanic Central Mexico during both the Classic and Postclassic

Periods, archaeological reconstructions of prehispanic Cholula are also lacking for the most part.

Several large-scale multidisciplinary projects have been carried out in the Puebla-Tlaxcala region, including the Tehuacan-Archaeological Botanical Project directed by Richard MacNeish, the Proyecto-Arqueológico Puebla-Tlaxcala led by Peter Tschohl, and the Proyecto Cholula, directed first by Miguel Messmacher and later by Ignacio Marquina (Plunket and Uruñuela 2005:

91). With the exception of MacNeish’s project, surprisingly few syntheses of the data gathered in these archaeological projects have been published, and much of the archaeological information collected from Cholula excavations exists only in technical reports (Hernández et al.

1998; López et al. 2002a, 2002b; Plunket et al. 1994; Plunket and Uruñuela 1993). As a result

―research of the last 50 years has not resulted in a clear, synthetic understanding‖ of the archaeology of the area, and ―consequently, Cholula – one of the major cities of Mesoamerica – has not played a significant role in any anthropological consideration of the prehispanic past‖

(Plunket and Uruñuela 2005: 90-91). Some researchers (see, for example, Uruñuela et al. 1998;

Uruñuela and Plunket 1998, 2001, 2002, 2005, 2007; Plunket and Uruñuela 1998a, 1998b,

1998c, 2002, 2003, 2005; MaCafferty 1996, 2000, 2001, 2007) are, however, beginning to construct broader theoretical frameworks for the evolution of the Puebla-Tlaxcala region in general, and Cholula, in particular.

52 Cholula has an extensive history of occupation, beginning in the Preclassic and continuing to present day. The site was first inhabited in the Middle Preclassic (500-200 BC) when two small villages were established near bodies of water on the eastern side of the modern city (Müller 1973: 19-20). In general during the Preclassic period, numerous agricultural villages with planned ceremonial centers were forming in the Puebla-Tlaxcala region. Social stratification also began to emerge during this time period, as did the beginnings of settlement hierarchies (García Cook and Merino Carrión 1987: 158-162). Tetimpa, an agricultural village to the northeast of the volcano Popocatépetl, has been the subject of numerous archaeological investigations, which have contributed greatly to our knowledge of village and household organization during this period (Uruñuela et al. 1998; Uruñuela and Plunket 1998, 2001, 2002,

2007; Plunket and Uruñuela 1998a, 1998c, 2002, 2003, 2005). The houses at the site follow a standardized pattern of a central platform with two smaller structures flanking a courtyard. The modular nature of these structures easily accommodated population growth, and the fact that the domestic compounds show such minimal variation suggests that this arrangement may have had social or ideological importance. Mortuary patterns and grave goods from the site indicate that the village was composed of partilineages, so the house compounds at Tetimpa may reflect the social organization of the village (Uruñuela et al. 1998; Uruñuela and Plunket 1998, 2001, 2002,

2007; Plunket and Uruñuela 1998a, 1998c, 2002, 2003, 2005).

Relatively little is known about Preclassic Cholula itself, but settlement surveys of the

Preclassic site conducted by McCafferty (1996a, 2001: 285) indicate that the center covered approximately 2 km2 during this time period and that monumental architecture was already present. In fact, by the Terminal Preclassic (100 BC-AD 200) construction of several important architectural features of Cholula, including the first stage of the Great Pyramid and the Edificio

53 Rojo, seem to have already been initiated (Figures 3-4 and 3-5) (Noguera 1956; Müller 1973:19-

20). During the Terminal Preclassic, a drastic change occurred in the settlement pattern of the

Puebla-Tlaxcala region and many of the ceremonial centers were abandoned. Around the same time, the population of Cholula increased significantly, likely augmented by the former inhabitants of these sites, and Cholula became politically dominant in the region. McCafferty

(2001: 285) argues that perhaps the construction of the Great Pyramid, which held enormous ideological significance, enabled Cholula to establish itself as a link between the heavens and the underworld and, thus, allowed the center to dominate its neighbors. Plunket and Uruñuela

(2005) have suggested that a more immediate cause of the growth of the center may have been a volcanic eruption in the middle of the first century AD that led to the abandonment of the northwestern portion of the Puebla-Tlaxcala region and the southeastern sector of the Basin of

Mexico. As a result of this event, they speculate that some 50,000 survivors may have immigrated to the emerging urban centers of Cholula and Teotihuacan. In order to clarify how construction episodes at the Great Pyramid were related to this natural disasters and the subsequent demographic upheaval, Plunket and Uruñuela (2005) are reexamining the construction sequence of the monument and attempting to more firmly establish the chronology of the site through radiocarbon dates.

Cholula developed rapidly into an urban center during the Classic period (AD

200-800) and continued its dominance of the Puebla-Tlaxcala region (Müller 1973: 20; Uruñuela and Plunket 2005b). Settlement surveys5 indicate that the site had grown to at least 4 km2 by this time and had an estimated population of 20,000 to 25,000 (McCafferty 1996a: 304). A considerable amount of monumental construction occurred at Classic period Cholula,

5 A surface survey by Peterson (1987) indicates that by 500 AD the site was somewhat larger, covering 6 km2 with a population of 30,000-60,000.

54 particularly at the Great Pyramid, which was expanded in two major stages (McCafferty 2000,

2001). While Cholula’s relationship with Teotihuacan during the Classic Period is an issue of considerable importance for the reconstruction of the political landscape, archaeological research into this question has not yet clarified the nature of the interactions between these two urban centers. Although Cholula shared some iconographic and stylistic elements with Teotihuacan, as well as some aspects of material culture, significant differences also existed between the two sites, which have led some researchers to conclude that Cholula made attempts to differentiate itself from the more powerful city (Paddock 1987; McCafferty 2000). Alternatively, it has also been suggested that perhaps the nature of the relationship between Teotihuacan and Cholula changed over the course of the Classic (Uruñuela and Plunket 2005; Plunket and Uruñuela

1998b). For example, the first stage of the Great Pyramid features talud-tablero architecture, a style strongly associated with Teotihuacano culture. However, the Cholula pyramid and the city itself have a different orientation than Teotihuacan. Furthermore, the second construction phase of the Great Pyramid used an architectural style unique to Cholula, in which all four faces of the structure were made up of steps leading to the top of the monument (McCafferty 2000: 344-345).

Interestingly, the third building phase of the Cholula pyramid, believed to have occurred after the decline of Teotihuacan, again features talud-tablero construction (McCafferty 2000: 346-347).

Figure 3-4: Picture showing the southern face of the Great Pyramid of Cholula and the Patio of the Altars.

Figure 3-5: Drawing of the construction phases of the Great Pyramid (McCafferty 2001:284). Note that north is to the left. The Patio of the Altars is immediately to the south of the pyramid and the Edificio Rojo is to the north. A dating project by archaeologists at the Universidad de las Américas is currently using radiocarbon dates to more firmly establish the chronology of the pyramid.

In terms of ceramics, stylistic similarities have been identified in Cholulan and

Teotihuacano pottery, particularly during the Early Classic when affinities were strong enough to lead Noguera (1954) to speculate a shared ethnic identity between the two sites. McCafferty

(2000: 348-350) echoes Noguera in his interpretation of artifacts recovered from R-106, a

Classic period house in Cholula. He suggests that the domestic refuse recovered from the site potentially indicates a common ethnic origin for the two populations, given the similarities in

57 household ritual, ceramic styles, and obsidian resources. However, he also emphasizes that important differences exist between households within the two centers that should not be ignored. R-106 is a single-family dwelling as opposed to the so-called apartment compounds at

Teotihuacan, and mortuary practices differ between the two sites, possibly suggesting ideological differences in family structure or lineage organization. In addition, certain religious items, fundamental to Teotihuacan’s state religion, are absent from Cholula (McCafferty 2000: 349-

350).

If its relationship with Teotihuacan is poorly defined, Cholula’s fate after the fall of the

Basin of Mexico’s urban giant is even more so (Uruñuela and Plunket 2005: 310-318; Plunket and Uruñuela 2005: 103). Based on settlement survey results from the Proyecto Cholula, many researchers have concluded that Cholula suffered a significant decline of its own in the Late

Classic and Epiclassic periods, with substantial decreases in the size of the site (Peterson 1987;

Müller 1973; Mountjoy 1987; Durmond and Müller 1972). However, other archaeologists

(Sanders 1971, 1989; Sanders et al. 1979; McCafferty 2000; Brumfiel 2005) have suggested that

Cholula continued to be an important regional center and may have even been able to expand its influence into the southern Basin of Mexico following the decline of Teotihuacan. McCafferty

(2000: 350) argues that monumental construction continued at the site in the Epiclassic, with the addition of the third phase of the Great Pyramid, which reached 350m on each side and 65m in height during this period. He also suggests that the Patio of the Altars, a large plaza on the south side of the Great Pyramid, which is so named for the three altars found there, was also constructed during the Epiclassic (Figures 3-4 and 3-5) (McCafferty 2001: 292-294).

Excavations and radiocarbon dates from the Universidad de las Américas support the idea that Cholula did, indeed, suffer some sort of decline starting in the middle of the seventh century

58 AD (Plunket and Uruñuela 2005; Uruñuela and Plunket 2005: 311). The radiocarbon data would seem to indicate that Classic period artifacts predate this time period. Excavations have also demonstrated that Classic period deposits were covered by a layer of sterile volcanic ash, on top of which Early Postclassic ceramics were found (Plunket and Uruñuela 2005: 103). Plunket and

Uruñuela (2005: 103) have proposed that a second eruption of Popocatépetl in the latter part of the Classic period may have contributed to the decline of the site. They note, however, that

McCafferty’s (2001) assertion that monumental construction continued at the site in the

Epiclassic is not necessarily incompatible with the theory that Cholula was in decline at this point, as it has not been argued that the city was completely abandoned (Uruñuela and Plunket

2005: 312).

Changes in stylistic and iconographic elements at Cholula at the close of the Classic period may also indicate that the Epiclassic was ushered in by a change in the ruling ethnic group

(McCafferty 2000, 2001). Gulf Coast, Mixteca Alta, and Maya influences are evident in murals and architectural elements from the Patio of the Altars, and imitations of Gulf Coast pottery appear at Cholula at this time (McCafferty 2000: 350-351; McCafferty 2001: 294-296). In addition, a change in funerary customs can also be seen in Cholulteca I (AD 800-900) burials from the site, which largely consist of cremations (López et al. 1976), and osteological studies have demonstrated that differences exist in the physical traits of the Classic and Postclassic populations at the site (López and Salas 1989). Ethnohistoric sources do mention a takeover of the site by the Olmeca-Xicallanca, who were affiliated with Gulf Coast cultures (Ixtlilxochitl

1975-1977: 530-531). Cacaxtla, a nearby site, may provide further evidence of an Olmeca-

Xicallanca intrusion into the valley at the end of the Classic period. A mural at the site with strong Mayan influences has been interpreted as a representation of the dual rulership structure

59 of the Olmeca-Xicallanca (Paddock 1987). However, it has been debated whether Cacaxtla represents a site established by the ruling ethnic group of Cholula after the supposed collapse of the city, or a site established by the Olmeca-Xicallanca prior to their takeover of Cholula

(Paddock 1987; Garcia Cook and Merino Carrión 1990; McCafferty 2000; McCafferty and

McCafferty 1994). Clearly, additional research into this period in Cholula’s history is necessary to firmly establish what was occurring during the Classic to Postclassic transition. Uruñuela and

Plunket (2005: 314-319) have already begun this work by attempting to sort out the Epiclassic ceramic chronology and by proposing questions that must be addressed to determine what transpired in the Puebla-Tlaxcala region at this time.

In the Postclassic period (AD 900-1521) Cholula reached its maximum size and enjoyed a resurgence as a politically powerful and culturally influential city (Müller 1973). However, significant turmoil plagued the urban center with another apparent change in the ruling group around the twelfth century. While archaeological research at the site has not yet been able to clarify much of the prehistory of the city, the dominant ethnic group at the time of the Spanish arrival, the Nahuatl-speaking Tolteca-Chichimeca, had their own historical accounts of their arrival Cholula and its subsequent takeover. According to the Historia Tolteca-Chichimeca

(Berlin and Rendon 1947; Kirchhoff 1947), the Tolteca-Chichimeca, along with the Nonoalca-

Chichimeca and the seven Chichimec tribes of Chicomoztoc, originally inhabited Tollan, a city in the northern part of the Basin of Mexico that is typically identified as Tula. When political unrest gripped the city, the Nonoalca-Chichimeca left the region, followed some fifteen years later by the Tolteca-Chichimeca. The Tolteca-Chichimeca migrated to Cholula, which was under the control of the Olmeca-Xicallancas. The Olmeca-Xicallancas allowed the migrants to settle in the city; however, they oppressed and enslaved the new arrivals. Unhappy with their

60 subjugation, the Tolteca-Chichimeca rebelled, overthrowing the Olmeca-Xicallanca and establishing themselves as the ruling ethnic group in Cholula.

Now referring to themselves as Chololteca, the Tolteca-Chichimeca soon faced a challenge to their dominance in the region when allies of the Olmeca-Xicallanca began to wage war on the new rulers. In response, the Chololtecas appealed to the seven Chichimec tribes still residing in Chicomoztoc to aid them in their conquest of their hostile neighbors. The remaining

Chichimecs obligingly immigrated to the region and helped defeat the allies of the Olmeca-

Xicallanca, and for their troubles, they were awarded land in nearby areas.

Archaeological evidence from Postclassic Cholula does substantiate a takeover of the site by a different ethnic group around the end of the Early Postclassic. The religious focus of the city shifted away from the Great Pyramid, and a new ceremonial center dedicated to

Quetzalcoatl was constructed at this time. While the Great Pyramid remained important as a shrine to the rain deity, it was no longer the primary focus of ritual activity (McCafferty 2000:

356-358). Along with this shift in religious ideology, archaeological data from the Postclassic are suggestive of conflict (McCafferty 2000: 356-357). Some sculptures from the final construction phases of the Patio of the Altars were intentionally destroyed. In addition, UA-1, an

Early Postclassic house was burned and a significant number of projectile points were found during the excavation. McCafferty (2000: 356-357) also posits that the Great Pyramid, which is missing part of its facade, may have been desecrated as a symbolic means of demonstrating the defeat and subjugation of the city.

Figure 3-6: Map of Cholula from the Relaciones Geográficas, 1581. (Rojas 1985)

62 The Political and Social Organization of Cholula

The altepetl

The primary unit of political organization in prehispanic Central Mexico was the altepetl, a Nahuatl word meaning ―water-mountain.‖ Although this word was frequently translated as

―town‖ or ―city‖ by the Spanish conquerors of the region, the term is actually infinitely more complex and encompasses not only a form of political organization but also a social unit with ideological and religious significance (Fernandez and Garcia 2006: 13). The two words that compose the term altepetl – ―water‖ and ―mountain‖—are fundamental symbols in creation stories that explain population origins in prehispanic ideology. Thus, the altepetl is an organizational manifestation of these myths that define the relatedness and unity of a population.

As a political and territorial unit, the altepetl refers to the territory and commoners belonging to one particular ruler (tlatoani) and his royal household (Hirth 2003: 61). While an altepetl might include both urban and rural settlements, prehispanic Mesoamericans did not draw a clear distinction between the two, certainly not in the way that Western cultures distinguish city and countryside. Urban settlements were viewed as the location where the tlatoani resided, but they were not conferred a special status in terms of settlement hierarchy due merely to their greater population density. Instead, cities were one part of the greater whole that was the altepetl

(Hirth 2003).

In some cases, multiple altepetl joined together to create a confederation of sorts referred to as a complex altepetl or hueyaltepetl. Such confederations were likely fairly common in

Central Mexico, as many urban settlements had a segmental form of political organization at the time of Spanish arrival. While Old World cities typically acted as single, integrated entities,

63 prehispanic Mesoamerican cities were often divided into multiple parts, each of which maintained a certain degree of autonomy (Hirth 2003: 66-68). Such was the case with Cholula.

The city of Cholula was the center of a hueyaltepetl whose approximate boundaries at the time of

Contact are shown in Figure 3-7, a map featured in the Cholula Codex (see also Figure 3-8 for a redrawing of this map) (González-Hermosillo and Reyes García 2002). According to this early colonial document, the city of Cholula was subdivided into six cabeceras, or independent administrative units. Not only did each of the cabeceras exercise authority over a number of barrios (calpulli) within the city itself, but they also oversaw the administration of settlements and territory outside of the urban nucleus (see Table 3-1 for a list of cabeceras) (González-

Hermosillo and Reyes García 2002).

Figure 3-7: The boundaries of the hueyaltepetl of Cholula, as shown in the Cholula Codex. Note that north is to the left. (González-Hermosillo and Reyes García 2002)

Figure 3-8: A modern redrawing of the map featured in the Cholula Codex showing the boundaries of the altepetl. Note that north is to the left. (González-Hermosillo and Reyes García 2002)

66 CABECERAS Barrios (calpulli) CABECERAS Barrios (calpulli) Tequepa Tequepa (Tecpan) San Juan Coacocongo (Tecpan) (continued) San Pedro Xahulxutla Cemotuntlica Xuteco (Xiuteco) Ostuma (Ostoman) Santa Maria Ocotlan Tlaquipaque (Tlacpac) Tlaxcoaque (Tlachcoac) Tianqueznauaque Ticoman (Tianquiznauac) Tulapustla (Tollan Pochtlan) Acachuysco (Acahuitzco) Panchimalco Tuspa (Tochpa) Calmecaque (Calmecac) San Pablo Mexico Santiago Yzquitlan Cuxpango (Tochpanco) Coquilaqui (Tzocuilac) Tecaman Izquentla Caotlan (Tzautlan)

Coamilco San Andres Tequepan Colomusco (Tecpan Collomochco) Xilnasco (Xeluasco) Coaco (Coac) Cuytlisco (Cuitlixco) Matalcingo (Matlaltzinco) Cuymencon Xicotongo (Xicotenco) Qualmehuaca Xalotle San Juan Tequepan (Tecpan) Aquiaguaque (Aquiauac) Papalutla (Papalotla) Tepetitlan Cuaque (Coac) Tlascalancingo (Tlaxcalantzinco) Caqualga (Tzacualpa) Table 3-1: A list of the cabeceras and barrios of Cholula. (González-Hermosillo and Reyes García 2002)

In Central Mexico, each altepetl was headed by a tlatoani, or king. While a tlatoani could be a supreme ruler of a state, as in the case of Texcoco or Tenochtitlan, more often multiple tlahtohqueh (plural of tlatoani) ruled jointly (Hicks 1986: 41-43). Given the segmentary nature of hueyaltepeme (plural of hueyaltepetl), joint rule is an expected feature of

Mesoamerican political organization. Cholula, in fact, had two supreme tlahtohqueh, referred to at the tlalquiach and the aquiach, who, in addition to their political role, had religious duties associated with the temple of Quetzalcoatl. This form of dual leadership is not typical of other

Central Mexican polities and may have its origins in Gulf Coast cultures (Hicks 1986: 43;

Paddock 1987). Interestingly, when a tlalquiach or an aquiach died, they were replaced by the most senior of the priests devoted to the temple of Quetzalcoatl. As all of the priests of this

67 religious order were nobles selected from a particular barrio of Cholula, Tianquiznahuac, it would seem that both of the supreme rulers were also of this barrio (Rojas 1985: 129-130).

Social organization

Postclassic Cholula was a stratified society made up of nobles, commoners, and slaves.

Only a brief overview of Cholulan social structure will be provided here, but it should be noted that describing the social organization of Central Mexico in terms of the three broad categories mentioned above – nobles, commoners, and slaves— glosses some of the nuance and complexity of Postclassic society. Some individuals of noble birth were forced to till their own lands, and some commoners were able to achieve significant status in appointed posts such as tribute collectors or work supervisors, or as a result of their achievements in warfare (Carrasco 1971;

Hicks 1986: 38). The following description is, thus, intended as a summary of the major features of Cholula’s social organization and not an exhaustive description.

Nobles (singular : pilli, plural: pipiltin) were, by virtue of their birth, members of a noble house (referred to as a teccalli in the Puebla-Tlaxcala region) headed by a teuctli, or lord. Each noble house had lands associated with it, which were sometimes dispersed throughout the altepetl rather than being located in the immediate vicinity of the teccalli (Hicks 1986: 38-39).

The teuctli (plural: teteuctin) distributed these lands to his dependent kin and to commoners

(singular: macehualli; plural: machualtin) attached to the teccalli. Every pilli was a member of a teccalli and held lands only by virtue of his membership in a noble house (Lockhart 1992: 102).

A tlatoani, as head of a noble household, would similarly have had commoner subjects who were his dependants. Teccalli were ranked among themselves, and teteuctin were subjects of a

68 particular tlatoani to whom they owed allegiance and tribute (Hicks 1986: 39; Lockhart 1992:

106-107).

While consanguinial ties linked the pipiltin of each noble house, no such lineage affiliations existed between pipiltin and macehualtin (Hicks 1986: 45-46). The word macehualli seems to refer to commoners in general; however, the type of political and economic subjection that macehualtin experienced varied depending upon whether they were calpulli members or mayeque. Calpulli were constituent units of the altepetl that held lands for exclusive use by their members (Lockhart 1992: 16-17). Macehualtin who belonged to the calpulli had hereditary usufruct rights to land that they farmed for their own subsistence. These usufruct rights were passed down patrilineally, typically from father to son, and as a result, some commoners within a calpulli might have access to more lands than other members. Those individuals who had insufficient or unproductive land could rent land from other calpulli in some parts of Central

Mexico (Carrasco 1971: 355; 364-365).

The calpulli had certain tribute obligations to the altepetl in the form of goods, military service, and draft labor6 (Hicks 1991: 203-204; Lockhart 1992: 17). Calpulli could be rural hamlets or barrios within a larger town or city, but they typically formed a residential group of some sort (Hicks 1991: 204). In general, members of a calpulli were not linked by kinship ties, although, in some cases, they may have shared a common ethnic affiliation (Lockhart 1992: 16-

17). Indeed, each of the cabeceras of Cholula is thought to have been founded by a particular

6 Hicks (1991) makes a distinction between ―gifts‖ and tribute payments. He argues that two kinds of subjugation existed in Aztec Mexico, political subjection and tributary subjection. In political subjection, commoners gave voluntary ―gifts‖ to their lords, and their lords were obligated to reciprocate by providing some benefit to their subjects. In tributary subjection, goods and services were forcibly extracted from a dependent population: It was a relationship based on exploitation as opposed to reciprocity. Hicks characterizes the relationship of dependency between the altepetl and commoners of the calpulli as one of political subjection, and, therefore, he refers to this payment of goods and services as gifts instead of tribute.

69 ethnic group (refer back to Table 3-1 for a list of cabeceras and calpulli in Cholula) (Carrasco

1971: 366). Carrasco (1971: 368) suggests that in rural areas kin relationships may have existed between calpulli members simply as a logistical byproduct of a small insulated population.

Calpulli in urban areas were less likely to be united by such lineage affiliations, although one particular ethnic group might have dominated.

The internal organization and structure of calpulli is not well understood and variation undoubtedly existed. Calpulli were subdivided into wards, and each ward was headed by an individual responsible for collecting tribute, organizing draft labor, and otherwise overseeing the functions of the calpulli (Hicks 1991: 204; Lockhart 1992: 17). In some cases, such as in

Tenochtitlan-Tlatelolco, it appears that particular calpulli were associated with the production of certain crafts, but that does not seem to have been universally true. Zorita (1994), in fact, states that calpulli included members of many different occupations.

Other commoners had no such hereditary usufruct rights to land. These individuals were known as mayeque (Zorita 1994; Lockhart 1992; Reyes García 1988a). Mayeque were associated with particular teccalli, or noble houses. While ideological distinctions regarding land rights divided mayeque from calpulli members, in practice, there seems to have been fairly minimal differences in the lives of commoners that were the dependants of a lord and those commoners that held calpulli lands. Mayeque had duties and obligations that were much the same as those of calpulli members. They were expected to provide tribute in material goods, participate in corvee labor drafts and military endeavors, and occasionally serve in the house of the lord. Hicks (1991: 203) suggests that calpulli members may have had lighter obligations than mayeque but states that the evidence is inconclusive. Mayeque also seem to have had plots

70 of land that were, on average, smaller than those of calpulli members, although calpulli lands were more variable in size (Lockhart 1992: 96-99; Hicks 1991: 202).

The relationship between teccalli, or noble houses, and calpulli is somewhat unclear. In western Central Mexico, it appears that teteuctli could be both the head of a noble house and the political leader of a calpulli (Lockhart 1992: 105). In the Puebla-Tlaxcala region the teccalli and the calpulli seem to be very separate institutions, at least superficially. Teteuctin typically claimed to Spanish authorities that they did not bear the same obligations to the altepetl as did the calpulli. However, documents from Tlaxcala indicate that teteuctin did, indeed, have to pay tribute to the altepetl, and that, therefore, the mayeque on their estates also indirectly provided financial support to the altepetl (Lockhart 1992: 106-107). In Tlaxcala, Huexotzinco, and other states in the eastern part of Central Mexico, Lockhart (1992: 107) has documented several instances in which teteuctin are named as the head of calpulli.

That being said, in the Puebla-Tlaxcala region, the teccalli had clearly overtaken the calpulli as the primary form of social, political, and economic organization by the time of the conquest (Lockhart 1992; Reyes García 1988a; Hirth 2003; Olivera 1978). Mayeque outnumbered calpulli members in much of this region. In Cuauhtinchan just after the Spanish conquest, at least 57.5% of tribute payers, and probably more, were mayeque as opposed to calpulli members (Reyes Garcia 1988a: 122), and in Tecali, as well, most commoners appear to have been dependants of a teccalli (Olivera 1978: 174-175; Lockhart 1992: 106). Calpulli were rare and largely appear to have descended from Tolteca-Chichimeca groups that migrated from

Cholula to the Cuauhtinchan region in the thirteenth century (Hicks 1982: 244; Reyes García

1988a: 80-100).

71 Tribute collection in the Puebla-Tlaxcala region was based upon the cuadrilla system, which had three administrative levels (Hirth 2003). Groups of 20 households, known as centecpantin, formed the basis of the system and were administered by an individual known as a centecpanpixqui. A group of five centecpantin constituted the second level of the cuadrilla system, and this level was overseen by a macuiltecpanpixqui. At the top of the hierarchy was a calpixque who coordinated the tribute payments and labor obligations of the 100-household groups. A teccalli or altepetl employed these administrative officials to take care of the bureaucratic side of tribute collection (Hirth 2003). See Table 3-2 for a table showing the labor and tribute obligations owed to four tlatoani in Tepeaca shortly after the conquest (Reyes Garcia

1988a: 109).

Table 3-2: Tribute and labor obligations owed to four tlatoani in Tepeaca. (Reyes García 1988a: 109)

73 It appears that a fair amount of mobility was commonplace for all macehualtin. Rulers sometimes resettled both calpulli members and mayeque in new lands (Hicks 1991: 205). For example, the ruler of Texcoco, Nezahualcoyotl, sent groups of emigrants to the Calpollalpan region (refer to the map in Figure 3-9 for sites in and around the Basin of Mexico) (Hicks 1982:

243). Macehualtin were also free to move to other locations provided they could find another lord willing to accept them (Hicks 1991: 204-205). In fact, one means by which commoners might have become mayeque and lost their hereditary rights to land was by moving to new regions (Lockhart 1992: 99). Ethnohistoric reports from both Tlatelolco and Texcoco indicate that commoners from other areas settled in those centers and were provided land (Hicks 1982:

243). The migration of entire calpulli was not an uncommon occurrence in prehispanic times.

Ethnohistoric accounts of both the Tolteca-Chichimeca of Cholula and the Mexica of

Tenochtitlan indicate that they migrated to their respective homelands sometime in the

Postclassic period. Alva Ixtlilxochitl (1975-1977, II: 32, 36; Hicks 1982: 243) states that immigrant groups to Acolhuacan were also granted lands. The issue of mobility in prehispanic

Central Mexico will be discussed further in Chapter VII.

Figure 3-9: Map of the Basin of Mexico. (Gibson 1967: 19)

75 Cholula as a Market Center

Given its strategic geographical location, it is not surprising that Cholula had developed into one of the major market centers in Central Mexico by the time of the Spanish Conquest.

Only limited descriptions of the Cholula market itself exist. For example, Durán (1971: 278) states that the market at Cholula was known as a place where ―jewels, precious stones, and fine featherwork‖ could be acquired. While data on the Cholula market is lacking, we can extrapolate from more elaborate accounts of other Central Mexican markets to gain a better understanding of what it might have been like. Some smaller markets were held only every five or 20 days, but larger markets met more regularly. Thus, the Cholula market was likely a daily occurrence (Hicks 1986: 52). Cortés, in his description of the nearby market of Tlaxcala, states that

Hay en esta ciudad un mercado en que casi todos los días hay en el de treinta mil ánimas arriba, vendiendo y comprando…. (1985: 104).

There is in this city a market in which almost every day there are upwards of thirty thousand souls, selling and buying….

The conquistador further describes the Tlaxcala market as selling all manner of things from clothing and jewelry, to firewood and medicinal herbs (Cortés 1985: 104). Although Cortés’ estimate of the number of individuals visiting the market on a daily basis is quite likely exaggerated, it does suggest that the markets in Central Mexican urban centers were very busy places. Descriptions of the markets in Tenochtitlan-Tlatelolco similarly recount the great variety of products and services offered in the marketplace, and the large crowds of people who visited each day. Bernal Díaz del Castillo (1996: 216-218), describing the market in Tlatelolco, tells of a well-ordered plaza in which gold and silver, precious stones, cloth, sandals, animal skins, pottery, fruits and vegetables, salt, stone tools, meat and poultry, live animals, cooked food, and

76 slaves were sold, among other things. Presumably, the market of Cholula would have offered a similar variety of goods for sale, thereby attracting residents of the city, as well as individuals from surrounding areas and merchants and their entourages from all over Mesoamerica.

Cholula as a Religious Center

In addition to being renowned as a market center, Cholula was a site of religious pilgrimage for peoples from all across prehispanic Mesoamerica (Olivera 1970). The patron deity of the city was Ehecatl-Quetzalcoatl, a feathered-serpent god associated with wind. In the

Historia Tolteca-Chichimeca, Quetzalcoatl is presented as a gentle and humane god, opposed to human sacrifice (Berlin and Rendon 1947). Described as a priestly figure, he is also associated with civilization and learning, and is credited with turning the Tolteca-Chichimeca into great craftsmen skilled at metallurgy and lapidary work (Berlin and Rendon 1947; Torquemada 1976,

III: 42-43). Quetzalcoatl is also identified as the patron deity of merchants (Durán 1971, Chapter

VI: 128-136).

At the time of the Spanish Conquest, the main ceremonial center of Cholula had moved away from the Great Pyramid and a pyramid dedicated to Quetzalcoatl was the primary focus of religious activity in the city. Rulers from other altepeme would travel to Cholula to be confirmed in their offices. As part of this investiture, they swore obedience to Quetzalcoatl and presented him with valuable offerings (Rojas 1985: 130-131). Not only did nobles make pilgrimages to the shrine of Quetzalcoatl, but commoners also seem to have traveled to make offerings to the god. Rojas (1985: 131-132), in fact, describes Cholula as the Mecca of the New

World in his 1581 account of the city. According to Durán (1971, Chapter VI: 135),

77 Quetzalcoatl was the patron deity of a number of ailments, and those afflicted with these conditions provided offerings to the god and prayed for a return to good health.

Not All Preindustrial Cities Are Created Equal

Models of urban health such as those proposed by Wrigley (1967) or Cohen (1989) make certain assumptions regarding the nature of preindustrial cities, namely that all preindustrial cities had high population densities, contaminated water supplies, unsanitary living conditions, and unreliable food supplies. While many preindustrial cities would have shared certain functional and logistical challenges as a result of a large number of people in a restricted area, including issues related to sanitation and ensuring sufficient food supplies for the urban population, cultural responses to these problems may have varied considerably. Therefore, it is a useful proposition to examine these generalizations in order to ascertain if they are, indeed, applicable to New World cities, in general, and to Cholula, in particular. While Cholula was quite likely comparable to Old World preindustrial cities in some fundamental ways, there were undoubtedly essential differences between Old and New World cities that could have resulted in distinct patterns of morbidity and mortality.

Mesoamerican urbanism

Of particular concern is whether Cholula and other New World urban centers can even be fairly classified as cities comparable to those of the Old World. Although it would seem that cities should be easy enough to recognize in the archaeological record, this has definitely not been the case in Mesoamerica. The issue has, in fact, resulted in a contentious debate within the

78 archaeological community (see, for example, Haviland 1970; Marcus 1983; Blanton 1981;

Sanders and Webster 1988; Chase et al. 1990; Hirth 2003), at the heart of which is a disagreement over how densely settled a site must be in order to qualify as a city.

While a number of different definitions of city have been proposed, Weber’s definition of what constitutes a city is among one of the most well-known and often cited (1958: 54-55). He suggests that cities are large, densely-populated settlements with significant economic functions and some degree of political autonomy from the state. He adds that the populations of cities are characterized by their heterogeneity, with people of different social classes, ethnicities, and occupations living in proximity. Wirth (1938) offers a similar definition, albeit somewhat more focused on the quantitative aspects of urbanization, such as the size and density of the population. As these definitions of city are largely derived from the development of European urbanization, they correspond quite closely to the generalizations of preindustrial cities posed by

Wrigley (1967) and Cohen (1989); however, they bear little resemblance to preindustrial cities in many parts of the ancient world. Sjoberg (1960), in fact, has argued that Weber’s definition, in particular, cannot be applied to preindustrial cities that developed outside of the Western tradition. He offers a somewhat broader interpretation, emphasizing the role of preindustrial cities as the centers of states. Sjoberg, therefore, allows for the possibility that preindustrial cities may have had significant political and religious functions in addition to, or perhaps even instead of, economic ones.

Other researchers have similarly proposed alternative definitions of cities that attempt to account for the variety of urban settlements observed in Mesoamerica as well as other parts of the ancient world (see, for example, Fox 1977; Sanders and Webster 1988). However, the general lack of consensus over the very definition of a city highlights the problem with the

79 argument that preindustrial cities would have experienced increased levels of morbidity and mortality as a result of high population densities. While some Mesoamerican urban centers, such as Tenochtitlan-Tlatelolco and Teotihuacan, were densely-settled economic and administrative centers, similar to the characterization of preindustrial cities posited by Wrigley (1967) and

Cohen (1989), others, such as Copan, had much lower population densities and were primarily devoted to political and religious activities (Sanders and Webster 1988).

This variation in Mesoamerican urban centers has been attributed to two primary factors: agricultural practices and energetics (Drennan 1984; Sanders and Webster 1988; Sanders and

Santley 1983). The more densely-settled centers in Mesoamerica were located in highland areas, while more dispersed settlement systems were typical of lowland areas. Sanders and Santley

(1983) suggest that climatic variations in highland and lowland areas contributed to the adoption of different agricultural practices, which dictated the potential for large population agglomerations to form. Lowland areas in Mesoamerica, for the most part, are tropical areas with abundant rainfall and vegetation. Slash-and-burn agriculture was traditionally practiced in these regions. As this form of agriculture requires that land be allowed to lie fallow for some period of time, more land is required per family, resulting in a more dispersed settlement pattern.

The highlands, in contrast, have a somewhat more arid climate and receive less annual rainfall, which makes irrigation agriculture beneficial. As irrigation agriculture is generally more productive, more people can be supported with less land under cultivation. Of course, in the

Basin of Mexico, the lake system permitted chinampa agriculture, which is exceptionally productive.

The relative lack of energy sources in Mesoamerica, namely draft animals, also had implications for the development of urban centers, as it limited the transport of food and raw

80 materials from the hinterlands. In areas with nearby rivers or lakes, such as the Basin of Mexico, canoe transport, in conjunction with the productiveness of chinampa agriculture, made it possible to support a large number of nonfood producers in Tenochtitlan-Tlatelolco. In fact, the Mexica capital had one of the largest and densest settlements in all of the New World. In areas in which water transport was not possible, human porters had to be used to carry food and other goods.

Sanders and Santley (1983) and Drennan (1984) note that the use of the human back as the primary energetic source significantly limits the potential size of an urban center because food can only be brought from so far away before the costs of transport (feeding the porter, for example) exceed the load the porter is able to carry. They further speculate that urban centers that had to be supplied exclusively overland would have reached a maximum population of only

25,000 to 60,000 because of the higher costs of transporting staple foods (Sanders and Santley

1983: 288). In the lowlands, the larger tracts of land required by slash-and-burn agriculture increased the distances that had to be traversed, thus increasing transport costs and contributing to more dispersed settlement patterns.

Urban centers in Mesoamerica, even those in highland areas, differed markedly in their population densities, economic organization, and urban-rural relationships. Tenochtitlan-

Tlatelolco with a population of 160,000 to 200,000 people at a density of 12,000 to 17,000 per km2, comes closest to approximating the characteristics of Old World cities such as London

(Calnek 1972). Its strategic location on the lake system and its use of chinampa agriculture allowed for a large number of economic specialists – nonfood producers, in other words—to be supported. Consequently, the city was dependent upon a market where these individuals could exchange their products and obtain the goods they required.

81 Teotihuacan, a Classic period center in the Basin of Mexico, is located some 12 km from the lake system that supplied Tenochtitlan-Tlatelolco. As a result, food supplies to the city would have been transported using a combination of water (via the lake system) and land routes

(Sanders and Santley 1983). While the city was not as large as Tenochtitlan-Tlatelolco, it did reach the respectable size of 125,000, with a population density of 7,000 people per km2

(Sanders and Santley 1983). What is most notable about Teotihuacan is its economic organization, which resulted from a political decision early on in its history. By around AD 100,

Teotihuacan implemented a policy of forced resettlement7. As much as 85% of the population in the Basin of Mexico was concentrated in the city, and another 10% lived in small settlements within 15 km of the site. Only 6% of the residents of the city appear to have been nonfood producers during this period (Sanders and Santley 1983: 261). Later in the site’s history, the rural population grew, but two-thirds of the residents of the Teotihuacan continued to farm land outside the city, at least part-time (Sanders et al. 1979: 109). Consequently, inhabitants of

Teotihuacan would not have been as dependent upon the market as residents of Tenochtitlan-

Tlatelolco or preindustrial European cities. This economic organization also raises the question, if the urban population of Teotihuacan did, in fact, experience natural decrease as a result of high morbidity and mortality in the urban environment, as models of urban health would claim, would immigration from the relatively small rural population have been sufficient to account for the population growth the city experienced over this period?

Texcoco, also in the Basin of Mexico, provides another example of how the political organization of a Mesoamerican urban center affected its population density. Hicks (1982) indicates that the city of Texcoco had a population of 100,000 spread out over 80 km2. The

7 Plunket and Uruñuela (2005) suggest that the eruption of the volcano Popocatepetl in the first century AD may have also contributed to the growth of Teotihuacan, as it resulted in the mass migration of survivors from the southeastern Basin of Mexico and the northwestern part of the Puebla-Tlaxcala region.

82 estimated boundaries that Hicks presents as the limits of the ―city‖ most likely represent the limits of the entire altepetl as well. While the area around the royal palaces was densely settled, houses in the rest of Texcoco were widely scattered. The reason for this extremely dispersed settlement system seems to be the political segmentation of the city and the calpulli form of social organization. Texcoco was divided into six sections, each of which had its own noble lineage. These sections shared a common politico-religious center where the ruler’s palace and other civic and religious structures were located. The macehualtin of the city were provided with land by their respective nobles in exchange for fulfilling their tribute obligations. These lands were located within the city itself, as well as in the surrounding region. In other words, residential areas of Texcoco were clustered around cultivated fields within the boundaries of the city proper. Just as in the case of Teotihuacan, the fact that the city residents themselves were farmers indicates that the city was not as dependent as their Old World counterparts on a market system for basic food supplies.

Lowland Maya centers were similar to Texcoco in that population densities were low, owing to the fact that some cultivated fields were located within the city itself. Tikal, for example, had an estimated population of approximately 55,000 individuals within an area of about 120 km2 (Chase et al. 1990). Even at its center, where population densities were highest, they only reached approximately 700 people per km2 (Haviland 1970). While Maya sites functioned as political and religious centers and were home to the majority of the elite members of society, they usually had limited economic functions. Economic specialization was restricted to full-time specialists at the urban center who manufactured elite goods and part-time rural specialists (Sanders and Webster 1988). Large market places such as those in Tenochtitlan and

Teotihuacan were typically not found in lowland centers.

83 Recently, several authors (Hirth 2003; Fernández and García, eds. 2006; Marcus 1983), in light of this observed variation in Mesoamerican urban centers, have proposed that models of

European urbanism are inappropriate for understanding Mesoamerican cities and that we should instead look towards an emic perspective for further insights into the nature of urbanism in this region. Hirth (2003) argues that indigenous groups in Mesoamerican did not conceive of a dichotomy between rural and urban areas, at least in a political sense8. Rather, the term altepetl, as mentioned previously, encompassed a ruler and all of his territory. The ―city‖ was simply the location where the ruler lived, and as a result, elites tended congregate in these areas as well.

This perspective on Mesoamerican cities requires that cities not be looked at in isolation, and suggests that they can only be understood in regional terms in which the entire polity is considered. The fact that indigenous populations of Mesoamerica conceived of urbanism in such fundamentally different terms had implications for not only the sociopolitical organization of cities, but also their economic organization. As a result, Mesoamerican urban centers did not always resemble their counterparts in the Old World. This variability in urban centers calls into question the typology of cities as densely-settled places dependent upon their hinterlands for food. These differences in Old and New World cities could have produced very distinct patterns of morbidity and mortality. In fact, within Mesoamerica itself, population dynamics may have varied considerably among urban centers. Clearly, significantly more research into the demography of prehispanic New World urban centers is necessary to adequately address how the evolution of cities affected human morbidity and mortality, if such generalizations can even be made. The current study of Cholula is but one step towards this goal.

8 Lockhart (1992) states that in Central Mexico, indigenous words did exist to describe more densely settled areas; however, these words are never used in a political context.

84 Cholula in Comparison to Preindustrial Old World Cities

How, then, does Cholula compare to preindustrial Old World cities? Are there enough similarities that we could expect comparable demographic patterns to have been operating in this

Mesoamerican city? While not as large or as densely settled as Teotihuacan or Tenochtitlan-

Tlatelolco, Cholula does seem to have been more similar in its internal organization to these New

World metropolises than to Lowland Maya urban centers. Cortés’ estimate of 20,000 houses within the city center would suggest a population of some 100,000 people within the nucleus of

Cholula, with an equal number, according to him, scattered throughout the surrounding countryside (1985). Other chroniclers such as Francisco de Aguilar (1954:29) and Torquemada

(1943, I: 281) give similarly high estimates for the population of the city at the time of Conquest:

Aguilar indicates that there were 50,000 to 60,000 houses within the city, and Torquemada, perhaps relying on Cortés, gives an estimate of 20,000 houses in Cholula. Given the general unreliability of population estimates from ethnohistoric sources, these figures are almost certainly grossly exaggerated. In contrast, Cervantes de Sálazar (1914: 241), also referring to the population at the time Spanish-contact, gives an estimate of 200,000 people for the entire altepetl of Cholula, which, in his description, includes 28 different sites. Rojas (1985: 126) states that at the time of his writing of the Relacion de Cholula in 1581, the city, which had been struck by at least two major epidemics in the years following European contact, had a population of only

9,000. He adds that the people of the city claimed that the pre-Contact population of Cholula was more than 40,000, and he suggests that that figure seems accurate given the number of prehispanic buildings still standing. Sanders (1971) agrees with the estimates of Rojas and

85 suggests a population of 30,000 to 50,000 people within the 8 km2 of the city just prior to

Spanish arrival.

Using 40,000 as an estimate of the population, the population density of the site would have been approximately 5000 people per km2. However, the center of Cholula was probably much more densely settled than the outskirts of the city. Sanders (1981: 189-190) states that, in general, smaller urban settlements in the Basin of Mexico, with populations of 10,000 to 30,000 inhabitants, had a settlement pattern which included concentric zones of habitation. While the centers of the cities were densely settled, these population nuclei were surrounded by less concentrated zones of houses with small land holdings. Janet Anderson (1987: 185) proposes such a model for Cholula, and Cortés’ comment that ―the city is very fertile with many small holdings‖ (cited by Anderson 1987: 185) supports this theory. This type of settlement pattern would, of course, result in some parts of the city being more densely populated than others.

Water contamination

In addition to the sheer number and density of people in a city, water contamination is typically cited as one of the primary sources of infectious disease in urban areas. Finlay (1981a), for example, indicates that residents of parishes near the Thames had higher mortality rates those of other parishes because they drew their drinking water from the polluted river. One of the most obvious means of avoiding contaminated water supplies in preindustrial societies, and one that was employed in both the Old and the New Worlds, was to rely partially on alternative sources of hydration. Fermented beverages such as ale and wine were consumed in, by today’s standards, rather large quantities. In Mesoamerica, aguamiel (sap from the maguey plant) and pulque (aguamiel that has been allowed to ferment) were likely important sources of hydration as

86 well as calories and nutrients. Rojas (1985:134), in fact, mentions that pulque was one of the beverages that Cholulans typically consumed.

Efforts do seem to have been made in prehispanic urban centers to limit the use of contaminated water supplies. In Mesoamerica, it is likely that collected rainwater was a common source of potable water, at least in the rainy season. In Classic-period Teotihuacan, evidence of wells has been found in habitational units, although the nearby San Juan and San Lorenzo Rivers or various springs may have also been sources of potable water (Escalante Gonzalbo 2004a: 61).

The city of Tenochtitlan-Tlatelolco was located in the middle of a lake system; however, the salinity of Lake Texcoco, while reduced from its natural state by a dyke allowing fresh water from Lake Xochimilco to enter, makes it unlikely that this water was used for drinking purposes.

Instead, at the time of the Spanish conquest, fresh drinking water was transported to the city via an aqueduct from Chapultepec. Interestingly, this aqueduct was not constructed until the mid- fifteenth century, so it is unclear what the source of potable water was before its construction

(Sanders 2000: 357). It seems unlikely that it would have been directly from the lakes, given their salinity, but during the Spanish siege of the city when the water supply was interrupted, residents did drink the lake water. Coincidentally, they were also hit with an outbreak of dysentery (Sahagún 1950-1969). Some homes in the city did have wells for household use, but again, as the water supplying the wells came from the lakes, it does not seem to have been a source of drinking water (Ortiz de Montellano 1993: 156-157).

In Cholula, Rojas (1985: 126) mentions several sources of potable water that were in use in the city in 1581. A number of shallow wells were located throughout the urban center, but evidently this water was brackish and heavily mineralized. A fountain in the plaza, which brought water from a source the northwest, a league and a half away, was the primary water

87 supply for the indigenous inhabitants of Cholula, although Rojas states that this water was also saline and mineralized. The Spanish, therefore, chose to get their water from an unspecified source outside the city that had water of a better quality. Rojas also comments on several springs on one side of the city that were used for wash water, but they do not seemed to have been sources of potable water. While these were the water sources utilized in 1581, the population of

Cholula had declined dramatically by that time. However, excavations of habitational areas in

Cholula have confirmed the existence of wells associated with residences, and aerial photographs and excavations have identified an aqueduct from the northwest that carried water to the city from the nearby snow-capped volcanoes (Matos 1967).

Furthermore, it should be noted that in preindustrial London, private companies provided drinking water brought from outside the city and residents were expected to pay for access themselves, which could be costly. As a result, many simply could not afford the expenditure and resorted to the Thames. The aqueducts supplying Mesoamerican cities were likely built with draft labor, and there is no indication that urban residents were charged for their use beyond their expected tribute obligations. Thus, the political economy of preindustrial cities could have dictated residents’ access to potable water.

Sewage

Exposure to human waste is another concern in urban areas because many parasites and infectious diseases are spread via fecal-oral transmission. With so many people living in a restricted area, finding a means of disposing of sewage becomes a real challenge9. In

9 The most significant problem with sewage disposal would be preventing water supplies from becoming contaminated. If water sources were coming from outside the city, this issue would have been less pressing.

88 preindustrial Old World cities, human wastes were often disposed of in the streets or kept in boarded up cesspools beneath the floors of houses (Landers 1993). How, then, did prehispanic

Mesoamerican cities resolve this issue? In Teotihuacan, it is unknown how human waste was disposed of. While a drainage system does exist within the city, this construction was used to direct the flow of rainwater and does not seem to have been used for sewage (Escalante

Gonzalbo 2004a: 58, 60). Tenochtitlan-Tlatelolco, on the other hand, was known to have had a system of public latrines located along the canals of the city. Canoes were strategically placed within these latrines to collect excrement, and then it was taken to be processed and sold as fertilizer (Escalante Gonzalbo 2004b: 208; Ortiz de Montellano 1993: 156-157). While this system of waste disposal certainly must have aided in keeping the city clean, the use of human excrement as a fertilizer would have greatly increased exposure to certain pathogens and parasites for those involved in agricultural activities10. In Cholula, the method of disposing of human waste is unknown, although presumably human excrement was used as fertilize, in which case it was likely collected in some way, either at the household level or the community level, as in Tenochtitlan-Tlatelolco. Alternatively, sewage could have been disposed of in unoccupied house lots, depending upon the density of habitation.

Garbage disposal

The accumulation of garbage is also cited as a potential source of disease exposure in preindustrial cities, as garbage attracts insects and vermin that are carriers of some illnesses. At

Teotihuacan, some patio areas with dirt floors within the residential compounds are speculated to

10 Of course, it would seem that urban and rural inhabitants would face exposure to pathogens via this route, both from contact with human excrement in the soil and from consumption of food grown under these conditions.

89 have been used to store garbage, but it is unknown what measures, if any, were in place to keep public streets and plazas free of refuse (Escalante Gonzalbo 2004a: 58). In Tenochtitlan-

Tlatelolco, steps were taken to ensure adequate means to dispose of garbage, and the Spanish, in fact, commented on the cleanliness of the city11 (Escalante Gonzalbo 2004b: 208). Public workers collected garbage from around Tenochtitlan-Tlatelolco and kept the canals free of silt and debris, and receptacles for burning trash were also located around the city (Escalante

Gonzalbo 2004b: 208; Ortiz de Montellano 1993: 156-157). While methods of garbage disposal in Cholula are not known, Torquemada mentions that the urban center was clean and orderly

(1943 I: 281-282). Given that Tenochtitlan-Tlatelolco had public workers to dispose of refuse, it is quite possible that Cholula did as well. The residents of Cholula may have also buried or thrown trash in house lots, in a manner similar to that observed in Teotihuacan.

While measures appear to have been in place to deal with garbage and sewage disposal, we should consider how strictly these regulations would have been enforced. London also attempted to improve the level of cleanliness of the city, but they were thwarted in their efforts by the lack of a centralized authority to actively enforce policies to keep the streets free of waste and debris. A similar problem may have existed in Mesoamerican cities, particularly given the segmentary nature of many of them. The altepetl of Tenochtitlan-Tlatelolco appointed judicial officials to maintain order and workers to help preserve the cleanliness of the city, but what, if any, punishments were in place for transgressions is unclear. In general, Tenochtitlan-Tlatelolco is known to have punished relatively minor offences rather severely—with mutilation, enslavement, or death (Escalante Gonzalbo 2004b: 210-211). However, whether littering was

11 Cleanliness is obviously a culturally relative construct. From the perspective of a modern industrialized nation, it seems more likely that Spanish cities were exceptionally dirty than that Mesoamericans cities were exceptionally clean.

90 perceived as sufficiently damaging to Mexica society to warrant a truly deterrent punishment is unlikely, as waste was said to have been left in the street at the close of the daily market

(Escalante Gonzalbo 2004b: 208).

Markets

Wrigley (1967, 1969) has argued that a market economy contributes to high levels of morbidity and mortality in urban communities because in years of poor harvests, grain prices increase, and city inhabitants are unable to purchase sufficient food to stave off malnutrition. As a result, they become increasingly susceptible to infectious disease. However, as in Old World cities, where the government frequently intervened during crisis periods to protect the supply of grains and offer poor relief, the rulers of prehispanic cities also attempted to buffer their citizens against economic hardship.

It has been suggested that in Tenochtitlan-Tlatelolco, a certain portion of tribute payments coming into the city was sold on the market to keep prices of foodstuffs at levels that craftsman in the city could afford and to assure a steady supply of both raw materials and basic necessities, thereby supporting economic specialization (Brumfiel 1980; Hicks 1986: 52-53). It is unknown if this also occurred in Cholula, but if it did it would have helped to stabilize market prices for urban residents. In addition, prehispanic states used tribute payments, in part, to stock emergency food stores that could be redistributed to the population during famine years (Berdan

1976: 189-190). For example, native histories tell of a particularly devastating drought that hit the Basin of Mexico from 1450-1454. Moctezuma I, ruler of Tenochtitlan-Tlatelolco opened the state storehouses in an attempt to alleviate the mounting toll on the starving population. While the government was able to provide food to carry the population through a couple of years, the

91 stores did run out before the drought was over. As a result, some people left the city, and many others probably died of starvation or disease. The state-sponsored chinampa construction that occurred shortly thereafter was an effort to protect against future droughts (Quiñones Keber

1995).

More significantly, differences in the economic organization of Old World cities and prehispanic Central Mexican urban centers would have factored into the effects of market fluctuations on urban residents. As mentioned, Sanders and Santley (1983) and Sanders et al.

(1979) have proposed that in Classic-period Teotihuacan, many residents of the city were farmers who resided within the urban center but travelled to work fields in the hinterlands, and in

Texcoco, fields were located within the city itself (Hicks 1982). Hicks (1982), in fact, suggests that most of the cities in prehispanic Central Mexico would have had an internal organization and settlement density similar to that of Texcoco; however, we currently have very little information on settlement patterns within Cholula or other similarly sized urban centers. If agricultural fields were, indeed, located within the city of Cholula as well, a significant portion of the urban population would not have been dependent on the market for their food supply.

While the densely-settled Tenochtitlan-Tlatelolco did have a large proportion of economic specialists who would have been dependent on the market, small chinampa plots were kept by residents of the city. These plots were insufficient to supply the bulk of the urban diet, but they could have potentially supplied fresh produce12 and may have buffered city inhabitants somewhat in cases of serious market disruptions.

12 In fact, many of the urban chinampa plots in Tenochtitlan-Tlatelolco were planted in flowers at the time of the Conquest, so they do not typically appear to have contributed significantly to the diet.

92 Traders and pilgrims

Wrigley (1967) also mentions the effects of trade on the spread of infectious disease within urban populations. In Old World cities, visitors were of concern, most especially during epidemics, because they introduced diseases into the urban environment. Preindustrial European cities would, in some cases, prohibit outsiders from entering their walls during outbreaks of plague to prevent the spread of the disease. Given the importance of the Cholula market, it likely attracted significant numbers of people from outside the city. Merchants came to the market to buy and sell their wares, and people from surrounding towns and rural areas also came to conduct business there. As a result, the presence of the market would have resulted in large numbers of transient visitors to the city. Cholula’s importance as a religious center is potentially of interest in this regard as well. Significant numbers of pilgrims visited the city, some of whom were motivated to undertake their journey by their already poor health. Given that epidemic diseases were not present in the prehispanic New World, however, these traders and pilgrims may have had minimal impact on the epidemiological environment of the city. Any diseases these individuals would have carried to Cholula would likely have already been endemic in the urban center. In fact, there are no ethnohistoric sources that mention prehispanic cities being closed to trade due to fears of disease.

In this chapter, I have provided a general overview of Cholula and other New World cities in order to evaluate whether Mesoamerican urban centers can fairly be described as having had high population densities, contaminated water supplies, unsanitary living conditions, and unstable food supplies. The New World differed from the Old World not only epidemiologically, but also in regards to the environment and the social, economic, and political

93 systems in place. Given this degree of variability, a substantial amount of research will have to be done before it can be demonstrated that all preindustrial urban centers shared the same basic patterns of morbidity and mortality in spite of tremendous underlying cultural differences. If we consider the logistical challenges of urbanism that result from a large number of people living in a restricted area, it is clear that cultural responses to these problems can vary greatly and that social, political and economic factors dictate how a society chooses to address these issues. In the following chapter, we will look at the paleodemography of Cholula, as reconstructed from a skeletal collection excavated from the site, and begin to evaluate how the culture of Postclassic

Cholula, along with the urban environment, may have shaped population dynamics in the city.

94 CHAPTER IV

THE CHOLULA OSTEOLOGICAL SAMPLE

“Cuando morían, enterrábanlos....” [Relación de Cholula, 1581 (Rojas 1985: 133)]

While historical documents have provided significant insights into urban population dynamics in Old World cities, written demographic information sufficiently detailed to permit the reconstruction of vital events in prehispanic Mesoamerican societies is simply not available.

As New World populations experienced a profoundly different epidemiological regime than Old

World populations, an understanding of morbidity and mortality in these populations could provide a wealth of information on the nature of epidemic diseases. Furthermore, demographic investigations of New World cities offer a means of examining how the interplay of cultural and biological factors affects human health. Human skeletal remains provide a possible avenue by which some of these questions regarding urban population dynamics can be addressed, but they are not without significant shortcomings and obstacles. Paleodemography simply does not have sufficient resolution to capture the details of human mortality and fertility that are observable in demographic studies of historic and modern populations. That being said, it does allow us a glimpse of the general demographic characteristics of past populations and is, therefore, a worthwhile pursuit as long as its limitations are clearly understood.

In Chapter IV, I introduce the skeletal collection from Cholula that was used in the current investigation. I briefly discuss the archaeological context of the burials and what little is known about the living conditions of these individuals. I then address the sample biases that might have contributed to the formation of this collection, focusing particularly on ethnohistoric

95 accounts of Nahua burial practices in Central Mexico. Based on this discussion, I consider whether this collection of skeletons is sufficiently unbiased to discern information on urban population dynamics in New World cities.

Provenience of the Cholula Osteological Sample

During the 1967-1970 excavation seasons of the Proyecto Cholula, 68 Cholulteca II (AD

900-1325) and 278 Cholulteca III (AD 1325-1500) burials were recovered from the southern and western sides of the Great Pyramid (Lopez et al. 1976). During the Postclassic period, the Great

Pyramid had fallen into disuse as a focus for ritual activity, and a low-status residential zone was constructed in its shadows (Figures 4-1 and 4-2). Human burials were found near altars and beneath the floors of these habitational units and patios13. While most of the burials located near altars, particularly the Central Altar, were sacrificial burials, small groups of skeletons interred under house floors and patios appear to represent familial groups who resided in the zone. For the purposes of this project, ceremonial burials and corporal segments14, which were sometimes used as offerings, were not included in any of the paleodemographic, paleopathological or

13 The skeletons excavated during the Proyecto Cholula have been the subject of a number of previous osteological, paleodemographic, and paleopathological studies including Lagunas (1973,1994), López (1973), Lopez et al. (1970, 1976), López and Salas (1989), Serrano (1972, 1973), Hayward (1986), Mansilla (1978), Marquez et al. (2002), and Carmargo et al. (1999) among others. Many of these studies use different subsets of the burials, so the results cannot be compared directly, but readers should refer to the above mentioned sources for further information on health and mortality in the population of Cholula derived using a variety of methodological techniques. The study of Marquez et al. (2002) is particularly interesting in that it uses the health index established by Steckel and Rose (2002) to analyze the collection. The current study relies on new methodological techniques to analyze this population. Uruñuela (1989) and McCafferty (1992) present skeletons from Cholula excavated from different residential areas. Readers may wish to compare features of these burials with those excavated during the Proyecto Cholula.

14 Corporal segments refer to human body parts, frequently hands and feet in the case of Cholula, that were used as offerings in some of the burials. They were not included in the current analysis.

96 isotopic analyses, as many of these individuals were sacrificial victims who may not have been residents of Cholula.

Relatively little information has been published on the habitational area from which the burials were recovered (Messmacher 1967; C. Hernández 1970; Noyola 1992). Figures 4-3 and

4-4 show the habitational unit excavated by Messmacher (1967)15. The map in Figure 4-5 shows the areas in which burials were found, and Appendix A lists the particular excavation unit in which each skeleton was located. Unfortunately, this habitational area was reburied and is no longer visible at the archaeological site. However, Sanders (2006, personal communication), who witnessed some of the excavations of the Proyecto Cholula, has described the habitational units he observed as being very much like the apartment compounds of Teotihuacan – a maze of small, crowded rooms. While the economic endeavors of the residents of Tlajinga 33 at

Teotihuacan were clearly discernable from the artifacts, little information points to the economic pursuits of the inhabitants of these units. White chert and obsidian workshops were found in the habitational area, and two ovens, possibly used for ceramics or metallurgy, were associated with the residences. These features suggest that at least some residents were involved in craft production, but how many individuals would have been employed in these activities is unknown.

The nature of the residences, as well as the paucity of grave goods, does indicate that these individuals were of low status. Offerings typically consisted of domestic artifacts such as lithic material and ceramics, but shell, animal remains (particularly dogs), and human bone and corporal segments were also included in some burials. Most individuals were buried with few to no items (López et al. 1976).

15 The habitational area excavated by Messmacher (1967) was found in the same general vicinity and dates to the same time period as those habitational units excavated by Marquina (1970). Skeletons excavated during Messmacher’s direction of the project were not included in the current study because documentation of the burials was not as detailed as for those burials excavated during the 1967-1970 field seasons. However, Messmacher (1967) provides information about the habitational units from this zone that is not available in Marquina (1970).

Figure 4-1: The archaeological site of Cholula as seen from on top of the Great Pyramid looking toward the south and showing the general area of the Postclassic habitational zone. The habitational units were reburied and are not visible in the above picture of the modern site.

Figure 4-2: The archaeological site of Cholula as seen from on top of the Great Pyramid looking toward the south and showing the general area of the Postclassic habitational zone. The habitational units were reburied and are not visible in the above picture of the modern site.

Figure 4-3: Picture of the habitational unit excavated by Messmacher. A well is visible. (Messmacher 1967)

Figure 4-4: A drawing of the habitational unit excavated by Messmacher. (Messmacher 1967)

Figure 4-5: Map showing part of the habitational zone. Red arrows indicate habitational areas (labeled “Zona Habitacional” in Spanish). (Marquina 1970)

100 As individuals were interred in and around their residences over hundreds of years, the burial situation was complex. This osteological collection includes both primary and secondary burials, and many of the burials include multiple individuals. The secondary burials likely represent primary burials that were disturbed by later interments in the area and then reburied.

Some of the multiple burials may, in fact, be individuals who died at the same time and were buried together; however, in other cases, it appears that individual burials were so close together that the excavators were simply unable to distinguish between them in the field and, therefore, excavated them as a single feature. Taking into account the multiple burials and the exclusion of ceremonial burials, a total of 78 Cholulteca II and 231 Cholulteca III skeletons were included in the analyses presented here. Preservation of the material varies from good to very poor, but most of the skeletons are in fair condition. Dating of the skeletal material was based on stratigraphic data as well as ceramics associated with the burials (López et al. 1976).

In the rest of the dissertation, I will refer to the results of the paleodemographic and paleopathological analyses in reference to their implications for the ―population of Cholula.‖

However, the reader should keep in mind that this is merely shorthand and that the skeletal collection under study represents only one habitational zone within the urban center and not the entire Postclassic population of Cholula. Variation in demographic processes due to status differences and microenvironmental differences within the city undoubtedly existed. As most of the inhabitants of Cholula were of low status, the patterns of morbidity and mortality observed in the skeletons from this residential zone are quite likely representative of much of the commoner population of Cholula. However, as in Old World cities, individuals of higher-status, because of factors such as greater access to resources and reduced exposure to environmental hazards, would have had levels of fertility and mortality that differed from those of lower socioeconomic

101 groups. Similar paleodemographic studies of skeletal material from other neighborhoods and of other socioeconomic groups in Cholula must be completed before we can assume that we have a complete picture of urban population dynamics for the population as a whole.

Preservation Biases

Excavated skeletal collections are typically the end result of a number of selection processes that render a biased sample. Special mortuary treatments, differential preservation of skeletons, and incomplete recovery of skeletal material can all affect the final composition of the osteological collection under study (Milner et al. 2000: 473-475). As the particular biases present in any given osteological collection affect the types of research questions that can be addressed using that material, the researcher must carefully consider the selection processes that may have affected his or her sample.

Underrepresentation of infants

One bias present in practically all skeletal collections is the underrepresentation of infant remains. Infants are often subjected to special burial treatments that result in their remains being separated from those of adults. Consequently, infant skeletons may go unexcavated. Moreover, their smaller bones typically do not preserve as well as those of adults (Gordon and Buikstra

1981), and their remains may be missed by excavators inexperienced in identifying the bones of children. Given how important infant mortality was in preindustrial population dynamics, the underrepresentation of juveniles is an unfortunate, but nearly universal, problem. The Cholula

102 collection is no exception. The issue of infant underenumeration in the Cholula sample will be addressed in more detail in the following chapter.

Special mortuary treatments

Biased preservation caused by special mortuary treatments is also of some concern in

Postclassic Cholula given ethnohistoric documents that indicate cremation was the typical means of disposing of the dead in the Nahua populations of Central Mexico. A number of researchers, primarily referencing the works of Fray Bernardino de Sahagún and Friar Diego Durán, have argued that mortuary treatments were dictated by the cause of death of the individual. They claim that cremation was the most common funerary treatment and that only select individuals were interred (Nagao 1985; Ortiz de Montellano 1990; López Austin 1988). However, some investigators (Smith 1992:259, 367-369; Smith 1996: 141-143; Brundage 1985) have challenged this interpretation, instead suggesting that ethnohistorical sources describing the funerary practices of elites in urban contexts have been broadly applied to all Postclassic Nahua populations without sufficient archaeological evidence to support such a conclusion. If cremation were the prescribed treatment of the dead in Postclassic Cholula, it would, in fact, result in a biased sample of intact skeletal remains consisting exclusively of individuals who died under very limited circumstances. Any conclusions drawn from such a sample about general morbidity and mortality in the population as a whole would be invalid. Consequently, it is imperative to examine whether the mortuary treatments most commonly described in the ethnohistoric literature were part of the funerary practices of the Postclassic inhabitants of

Cholula.

103 Fray Bernardino de Sahagún, a Franciscan friar who investigated prehispanic customs of the indigenous people of Central Mexico in order to facilitate their conversion to Christianity, included descriptions of Nahua mortuary practices and conceptions of the afterlife in two of his most notable works: the Florentine Codex and Primeros Memoriales. Other Spanish missionaries, such as Friar Diego Durán, described Nahua funerary customs as well as the burial ceremonies of rulers. Two native colonial period documents, the Codex Magliabechiano and the

Codex Ixtlilxochitl, provide pictorial representations of the funerals of individuals of different social strata, and several mestizo chroniclers also briefly mention Nahua burial customs (Nagao

1985: 37).

While these ethnohistoric sources contain valuable information on many aspects of

Nahua culture that are difficult to address archaeologically, they are not without their own problems and biases. First, these sources tend to focus disproportionately on the elite sector of society. Sahagún commissioned as his assistants young native noblemen from Tlatelolco and relied on older members of the nobility as his informants (Baird 1988: 15). Although little is known of Durán’s native informants, at least some of the information he included in his Historia de las Indias de Nueva España e Islas de la Tierra Firme was obtained from Nahuatl texts, which themselves would have been written and illustrated by members of the nobility (Colston

1988: 59-60). Even the mestizos who chronicled native religious beliefs were the children of nobles (Nagao 1985: 37). Consequently, the extent to which these sources represent the beliefs and practices of the broader society is unclear. Second, many of these sources describe mortuary practices in the Basin of Mexico, particularly in the Mexica capital of Tenochtitlan-Tlatelolco.

Although these descriptions are frequently assumed to be applicable to all Nahuatl-speaking

104 groups in Central Mexico, archaeological evidence has not necessarily borne out this supposition.

The Nahuas recognized at least five separate afterworlds, each of which was associated with one of the cardinal directions. According to Sahagún’s (1950-1969) description of Nahua funerary practices, the treatment of the body after death depended upon the cause of death itself and the particular underworld for which the individual was thought to be destined. Most individuals in Nahua society were bound for the underworld of Mictlan, a foreboding place associated with the north in which the souls of the deceased dined on pus and fetid beetles and endured icy winds that cut like obsidian blades (Sahagún 1997, II: 6). In Mictlan, the souls of the dead had to pass through nine hells, a journey which would last four years and which subjected the dead to numerous trials and tribulations. In the ninth and final hell, the soul either disappeared or was allowed to rest for all eternity (Sahagún 1997, II: 6).

Most interpretations of the ethnohistoric sources agree that those who went to Mictlan were cremated. Sahagún states that the corpse was burned after it had been properly adorned with the ornaments necessary to complete the trials the soul would have to face in the afterlife

(1952: 42). The body was bound in blankets or papers and then the corpse was cremated along with a yellow dog, which was to help the deceased cross a deep river in the first level of this underworld. Once the body had been reduced to ash, the remains were gathered up and placed in a jar. This jar was then buried in a temple or in the individual’s house, depending upon the status of the deceased (Sahagún 1952: 43).

Warriors who died in battle, merchants, and sacrificial victims could expect a far more glorious afterlife than that experienced by those in Mictlan. These souls were destined for the eastern paradise of Tonatuihichan, or the ―House of the Sun‖ (Sahagún 1952: 47-48; Nagao

105 1985: 39). To be admitted into the House of the Sun was a reward for an honorable and esteemed death: These individuals had died in the service of the sun god, and were, therefore, reaping the benefits of their devotion (Ortiz de Montellano 1990: 49). Sahagún writes of this paradise that

They share a life of continued delight with him [the Sun],… never do they experience any pain or sorrow, because they live in the mansion of the Sun, where there is an abundance of delights. (1969: 140)

For four years, these souls accompanied the sun each day from sunrise to mid-day on his journey across the sky. After four years, the souls of these individuals returned to earth as hummingbirds or butterflies (Ortiz de Montellano 1990: 49; Sahagún 1952: 47-48).

The treatment of the body of those individuals destined for Tonatuihichan seems to be somewhat variable. In the case of pochteca (merchants) or warriors who died far from home, the bodies frequently could not be recovered. Dead warriors were reportedly burned on the battlefield, and an arrow was taken from each corpse to give to the family as a substitute for the fallen hero (López Austin 1988: 321). An effigy of the warrior was made and then cremated as if it were the body (López Austin 1988: 321; Sahagún 1957:69-70 cited in Nagao 1985: 39).

Merchants who died at home may have been cremated, or, according to Sahagún, their bodies may have been carried to a nearby mountain, where they were left to the animals (Nagao 1985:

39). For the most part, the bodies of sacrificial victims seem to have been buried or otherwise discarded without cremation (Sahagún 1952: 43).

The western paradise, also associated with the sun, was the home of the souls of women who had died in childbirth. In Aztec ideology, childbirth was analogous to going to battle; therefore, women who died giving birth went to join the sun just as warriors did (Ortiz de

Montellano 1990: 50). These women accompanied the sun from mid-day to sunset as it moved across the western horizon (Ortiz de Montellano 1990: 50). However, women who died in

106 childbirth were not fated to become hummingbirds or butterflies, as were the souls of warriors who died in combat. Rather, after four years of accompanying the sun, they became cihuateteo, frightful ―women-goddesses‖ with a skull for a head and claw-tipped hands and feet, who returned to earth on particular days of astrological significance. To encounter one of these cihuateteo was a sign of ill fate to come, particularly for women and children (Ortiz de

Montellano 1990: 50).

According to Sahagún, the bodies of women who died in childbirth were buried without cremation. In fact, since the bodies of these women were said to have certain magical powers that could protect warriors in battle, young men often tried to steal the right arm of the deceased woman. Consequently, her husband or other kinsmen had to guard the body for four nights to protect the corpse from mutilation by these would be thieves (1969: 161-162). At one point,

Sahagún even suggests that the body parts most desired by grave robbers were removed prior to burial so as to discourage those seeking these magical talismans (1969: 161-162). The bodies of these women were supposedly buried in the temple of the Cihuapipiltin (Sahagún 1969: 161-

162).

Tonacacuauhtitlan, reserved especially for the souls of children, was a verdant paradise where fruits, flowers, and trees flourished (Sahagún 1969: 115; 1997:178). As these souls were considered to be pure and innocent, they were to be afforded another opportunity to return to the earth: They would be transformed into a new human race when the present one was finally destroyed (Ortiz de Montellano 1990: 66). Accounts differ as to the age of the children destined for Tonacacuauhtitlan. Some sources suggest that only suckling infants are admitted into this paradise, since only they were so pure as to not have eaten corn or to have engaged in sexual activities (Ortiz de Montellano 1990: 66). Sahagún does say about this afterworld for children

107 that it was a place for ―he who died when he was a rather young child, and indeed still a babe in the cradle,‖ lending credence to the idea that only infants were admitted into Tonacacuauhtitlan

(Sahagún 1997: 178). In addition, this paradise is sometimes referred to as Chichiualcuauhco, or

―The Place of the Nursemaid Tree,‖ because the souls of infants suckled from a tree with udder- like leaves filled with milk (Ortiz de Montellano 1990: 66-67).

Other sources have claimed that anyone, regardless of age, who was considered pure and chaste, might be admitted into this afterworld (López Austin 1988: 314). Brundage (1985: 193) seems to have extended this to anyone who was unmarried. Although Sahagún does recount the story of a man who was allowed into Tonacacuauhtitlan, the point of the story was that this individual was exceptional in his purity, not that anyone who was celibate was permitted entry

(1969:115). Those individuals destined for this heaven were buried, supposedly near granaries

(Sahagún 1969: 116; Nagao 1985: 40).

Admission into the fourth paradise, Tlalocan, was by invitation of the rain god Tlaloc only. As Tlaloc was the god of agriculture as well as the rain deity, this afterworld was conceived of as a place of abundance where no one ever suffered (Sahagún 1952: 45). Tlaloc chose those individuals whom he wanted to serve him in the afterlife by afflicting them with certain diseases, by striking them with lightening, or by drowning them. Thus, Sahagún writes

And there went those who had been struck by thunderbolts, and those who had been submerged in the water, and those who had died in the water. And those who were leprous, or affected by venereal sickness, or by boils, or by the itch, or by sores, or by the gout; and those full of tumors, [those] they took [there], and those swollen by the dropsy, [who] thus died. (1952: 45)

The conditions that Sahagún cites as being associated with Tlaloc are somewhat problematic.

First, no agreement exists as to how to translate many of these diseases from the original Latin or what conditions they actually represent (see Footnote 5 in Sahagún 1952: 45). Most frequently in the literature, leprosy, running sores, and various unspecified skin conditions are assumed to

108 be the diseases that assure admission into Tlalocan (see Nagao 1985: 40; Brundage 1985: 190).

Second, the Florentine Codex was written decades after European contact; therefore, some diseases of European origin may have been incorporated into this aspect of Nahua religious ideology. For example, leprosy was not a New World disease, so Postclassic populations could not have been afflicted with it. Ethnohistoric sources imply that entry into Tlalocan was considered to be unusual, suggesting that the diseases associated with the rain deity were not common causes of death.

Those that were chosen by Tlaloc were said to have been buried rather than cremated.

Their faces were painted and they were dressed to resemble the rain god. Images of mountains were placed in the graves, and in their hands they held a large staff of wood, reminiscent of the lightning bolt staff carried by the deity (Sahagún 1952: 45; Nagao 1985: 36). Although Sahagún does not specify where these bodies were buried, Nagao (1985: 40) speculates that they may have been placed near mountains or perhaps temples dedicated to Tlaloc.

One of the most contentious issues surrounding the mortuary practices of the Nahuas is whether the funerary customs described in the ethnohistoric texts applied to both elites and commoners. Sahagún (1952: 43) specifically states that his descriptions of cremations applied to both lords and peasants. However, some scholars have questioned whether that was truly the case (Brundage 1985; Smith 1992, 1996). It seems virtually certain that Aztec kings were cremated, since a number of ethnohistoric accounts describe in detail specific funerary ceremonies of several different rulers (Durán 1964; Alvarado 1943). In all of these descriptions, the body of the ruler is burned, along with the hearts of a number of sacrificial victims, and the remains were then buried. Cremated remains have been found in elite contexts. In the excavations of the Templo Mayor, eight mortuary deposits consisting of urns and cremated

109 human remains were recovered. These remains are assumed to be those of high-ranking members of Nahua society because of their placement in a temple (López Luján 1994: 238; Matos 1984).

Seven of these mortuary offerings were found in the part of the structure dedicated to

Huitzilopochtli, and all of these deposits were oriented toward the west, both of which suggest an association with the sun deity. Furthermore, four of these mortuary deposits were found at the base of a pedestal that would have held an image of the god (López Luján 1994:224-225), echoing Alvarado’s description of the placement of the cremated remains of rulers. While some of these cremations could represent Aztec kings, others could be members of the royal family, or even priests or officers who died in battle (López Luján 1994: 238).

In addition to the cremations found in Tenochtitlan, cremated human remains, as well as burials, have been recovered from excavations of the ceremonial center of Tlatelolco (Nagao

1985: 36-37). Cremated remains placed in clay urns were also found in the main temple at

Tenayuca (Nagao 1985: 36-37). As with the cremations from Tenochtitlan, these mortuary deposits are assumed to represent members of the nobility because of their location in elite contexts. In the Tehuacan Valley, Sisson (1973, 1974) encountered a large number of cremations in urns dating to the Postclassic clustered around the bases of temples.

Some disagreement exists as to whether the majority of rulers and other elites were truly destined for Mictlan, as some sources imply. According to Sahagún, ―all of those who died here on earth, who died only of sickness – the chiefs, the vassals‖ (1952: 39) would find themselves in this cold, dark, northerly realm of the dead. Other sources, however, have suggested that

Mictlan would not have been inhabited by the souls of deceased nobles. Brundage (1985: 189) contends that this undesirable underworld was the final resting place for only the souls of commoners and that the four paradises were reserved for members of the nobility. Furthermore,

110 he suggests that commoners were buried rather than cremated. He bases this assertion on the idea that cremation allowed the soul to be carried to heavens, or the House of the Sun.

Therefore, only warriors and members of the nobility were cremated. Those who would be bound to the earth even in death received a burial. This included those who would reside in the mountainous home of Tlaloc, the souls of those women who would inhabit ―the place of the sunset,‖ and those fated for Mictlan (Brundage 1985: 193). Unfortunately, Brundage does not cite any ethnohistoric sources to support his interpretation, so the origin of his assertion is unclear.

Brundage does raise an interesting question about Nahua ideology. The Nahuas had a stratified social structure with limited mobility between social classes. Yet Sahagún suggests that in death some basic equality existed among people regardless of their social status in life.

Although this certainly a possibility, it seems highly unlikely that members of the nobility, particularly kings, would have faced the icy winds of Mictlan in the afterlife rather than basking in splendor in the House of the Sun. In fact, social stratification in Aztec society was justified by an ideology that stated that elites were spiritually superior to commoners (Ortiz de Montellano

1990: 53). Since the gods chose people to serve them in the various paradises, it would make sense that the nobility would be preferentially selected for these honors. Friar Mendieta wrote the following passage concerning the afterlife, which supports Brundage’s assertion that status could have played a role in an individual’s final destination:

The people of Tlaxcala believed that the souls of lords and heads of state became mists, clouds, birds of fine feather, and other things, and also precious stones of great worth. And that the souls of common people became weasels, stinking beetles, animals that emit stinking urine, and other thieving animals. (1971: 97)

In Primeros Memoriales, Sahagún relates a Nahuatl myth that supports the idea that paradises were reserved primarily for the elite. The story involves a noblewoman who goes to Tlalocan

111 upon her death. While she is there, she sees a number of other souls, all of whom are identified as having been members of the nobility in life. She encounters no one who is identified as a commoner (1997: 181-183). Thus, there appears to be some basis for the argument that elites and commoners experienced different fates after death, and, therefore, different funerary treatments.

Other ethnohistoric accounts seem to suggest that cremation as a mortuary practice was fairly limited in general. An anonymous account describing the burial ceremonies of nobles, merchants, and commoners written in the sixteenth century makes no mention of cremation at all. Rather, the only differences between the funerary treatment of these three groups is that the nobles were buried in tombs along with sacrificed slaves, merchants were buried with trade goods, and commoners were buried with food (cited in Smith 1992: 367). The Relaciones

Geográficas pertaining to Central Mexican sites also make no mention of cremation as a typical funerary practice (Smith 1992: 367).

Cremated remains are, for the most part, uncommon in nonelite contexts in Postclassic

Central Mexico (Evans 1988: 15; Smith 1992: 259, 1996: 151; Healan 1989; Ceja 1987; P.

Hernández 2006; Salas Cuesta 1982; Zacarias 1975; Monzón 1989; Uruñuela 1989, although see

López 1972 for an exception pertaining to Cholula). Postclassic period burials from residential contexts are also thought to be uncommon (Smith 1992, 1996; McCafferty and McCafferty

2006) because household excavations that have been widely published in the American literature have recovered relatively few skeletons. The osteological material that has been found in these excavations is also notable for consisting almost exclusively of the skeletal remains of children.

Two skeletons recovered from under housefloors at Cihautecpan in the Basin of Mexico were both juveniles – one an adolescent female and the other a child estimated to be between 6 to 18

112 months old (Evans 1988: 125, 185). At Capilco and Cuexcomate in Morelos, Smith excavated fourteen burials of infants and children found under housefloors and adjacent residential areas

(1992: 259; 1996: 142). In the Tehuacan Valley, Sisson also reports infant burials under house floors (1973, 1974), and excavations of an Early Postclassic residence at Cholula, the UA-1 domestic compound, yielded eighteen burials, seventeen of which were juveniles (McCafferty and McCafferty 2006).

The fact that most of the known burials recovered from Postclassic sites are juveniles has frequently been used to support the assertion that the Nahuas only buried select groups of people, children among them. However, the discovery of juveniles under housefloors is actually inconsistent with Sahagún’s reports. According to the ethnohistoric sources, only suckling infants were buried. While some of the excavated juvenile remains are those of infants, others belong to older children and adolescents.

Smith (1992: 369; 1996: 142-143) maintains that one possible explanation for the limited recovery of adult remains from household excavations is that the Nahua buried only juveniles under housefloors and deposited of their adult dead in cemeteries16 that have not yet been located by archaeologists. Excavations at Terrace 85 at Xochicalco uncovered three adult burials dating to Early and Late Aztec periods (Smith 1992: 369; 1996: 143). Smith suggests that the

Epiclassic site may have had ritual importance to the Postclassic inhabitants of the area and that, therefore, they used it as a cemetery for the remains of adults (1996: 143).

It has been suggested that the skeletal sample from Cholula under study in the current project also represents a cemetery (McCafferty and McCafferty 2006; McCafferty 2007)17.

16 A cemetery is recognized archaeologically as a well-defined area in which burials are formally arranged.

17 In fact, McCafferty and McCafferty (2006) take up Smith’s argument that infants and young children were buried in houses, which would account for the low numbers of infants in López et al.’s (1976) published lists of the burials

113 However, this seems unlikely given the reported archaeological context. Although the findings of the Proyecto Cholula have been poorly published for the most part, there are reports that clearly refer to the excavation of low-status Postclassic habitational units from around the Great

Pyramid (Messmacher 1967; Hernández 1970). The physical anthropologists who excavated these burials have explicitly stated in a number of publications (López et al. 1976; López et al.

2002; Lagunas 1994) that these skeletons are from under the housefloors and adjacent areas of habitational units18:

Los entierros fueron hallados en estrecha asociación con vestígios de pisos de habitaciones y de pequeños audoratorios de la época…. (López et al. 2002: 56)

The burials were found in close association with the remains of housefloors and small altars from this time period….

Thus, while examples of Postclassic cemeteries may, in fact, exist, the burials excavated from the

Proyecto Cholula cannot be counted among them19.

Furthermore, it should be noted that excavations and salvage projects undertaken by the

Instituto Nacional de Antropología e Historia and other Mexican institutions have recovered a much larger number of Postclassic skeletal remains, many of them from residential areas.

Nonelite burials of both children and adults have been excavated at Postclassic Teotihuacan

(Monzón 1989), Tenochtitlan (Salas Cuesta 1982), Tlatelolco, Azcapotzalco (Ceja 1987),

from the Proyecto Cholula. However, infant underenumeration is typical of virtually all skeletal samples, so the fact that the burials from the Proyecto Cholula show a low percentage of infants, as per Hayward (1986), is not surprising and may result purely from poor preservation and recovery of infant remains. More importantly, the archaeological context of these skeletons suggests that they do not represent a cemetery.

18 More recent excavations in the fields to the northeast of the Great Pyramid have also uncovered the remains of Postclassic houses overlying Classic period structures, providing more evidence that the area around the pyramid was a residential zone during that time period (López et al. 2002a, 2002b).

19 In the case of both the UA-1 domestic compound (McCafferty 1992) and a Postclassic house at Teotihuacan (Monzón 1989) some of the reported burials from these habitations were deposited after the abandonment of the house. It may be the case that using abandoned houses to bury the dead was a common practice in Postclassic Central Mexico.

114 Xochimilco (Hernández 2006), and Cholula (Uruñuela 1989)20 as well as other sites. If we consider the age and sex distribution of Xochimilco and Tenochtitlan, two of the larger collections of Nahua populations, there does not appear to be a significant bias that would suggest widespread cremation with only select individuals being buried (Salas Cuesta 1982; P.

Hernández 2006). As unweaned children and women dying in childbirth reflect particular demographic groups, we would expect to see a discrepancy in the sex balance and an abundance of young children and women of childbearing age in the skeletal remains of a population that adhered to the mortuary practices set out in Sahagún’s description. Those dying of conditions associated with the Tlaloc would presumably include both males and females of a variety of ages, although the very young and the very old might be overrepresented, depending upon the underlying causes of the infections said to be associated with this deity. The age and sex distributions of Xochimilco and Tenochtitlan are typical of paleodemographic samples and show no such biases (P. Hernández 2006; Salas Cuesta 1982).

What, then, can we conclude about Nahua burial practices? It appears that Sahagún’s description of Aztec mortuary practices may not be applicable to the whole of Nahua society

(Smith 1992:259, 367-369; Smith 1996: 141-143; Brundage 1985), although the reason or reasons for this discrepancy are not entirely clear. Smith (1992, 1996) has suggested that

Sahagún’s description applied primarily to the urban area of Tenochtitlan-Tlatelolco and that rural locations did not frequently cremate the dead. However, burials from low-status residences at Tenochtitlan, Tlatelolco, and Xochimilco, all urban locations, would seem to suggest that status was the most significant variable affecting mortuary practices.

20 This skeletal collection from Cholula is not large, as it includes only 38 individuals from both Classic and Postclassic contexts, but it is from a residential area and includes both subadults and adults (Uruñuela 1989). Other Postclassic skeletons from non-residential contexts have also been excavated at Cholula (Romero 1937; Suárez 1990).

115 Looking more closely at the mortuary practices discernable from the Cholula burials being investigated in the current project, some ideological elements found in ethnohistorical accounts do seem to manifest themselves. For example, most of the primary interments were found facing north or northeast, and the remains of dogs were found with some burials, both of which suggest an association with Mictlan (López et al. 2002). However, it does not appear that cremation was the primary method of disposing of the dead at Cholula during the time periods under study

(Cholulteca II and Cholulteca III). Indeed, as the Cholula collection comes from low-status habitational units, it may be the case that Brundage’s position that commoners were destined for

Mictlan and were, therefore, buried, is correct.

Interestingly, during Cholulteca I (AD 800-900), the majority of the human remains recovered by the Proyecto Cholula were cremations in urns (López et al. 1976), but only a small percentage of the burials from Cholulteca II and III were cremations (approximately 3% of the non-ceremonial burials21). Archaeological and osteological evidence does suggest that there was a change in the ethnic group inhabiting the site at the end of Cholulteca I, which would account for the change in burial practices. If this is true, the ethnic group that inhabited the site by the latter part of the Postclassic, thought to be the Nahua, does not appear to have practiced widespread cremation, suggesting that differential burial practices are not a significant biasing factor in this sample.

Rojas (1985), in fact, explicitly states that the indigenous people of Cholula buried their dead and makes no mention at all of cremation. Furthermore, as will be demonstrated in the following chapter, an examination of the age and sex distribution of the Cholula population

21 Of those cremations from Cholulteca II and III that could be sexed, all were male. These cremated individuals may, therefore, have been either individuals of some status within the habitational unit, or they might have been warriors who died in battle.

116 indicates an approximately equal number of males and females (Table 5-1 and 5-2), and an age- at-death distribution within the boundaries of what would be expected for a preindustrial population (Figure 5-18). There is no overrepresentation of young children or reproductive-age females that would be caused by the burial practices described in ethnohistoric accounts. As the

Cholula collection does not appear to be biased in this respect, it is an appropriate sample to address the issue of urban population dynamics in New World cities.

In this chapter, I have considered the archaeological context of the skeletal collection to be used in the paleodemographic and paleopathological analyses of Cholula. I have also discussed the potential biases that may have affected the composition of the sample, with particular emphasis on funerary customs in Nahua society. Archaeological and paleodemographic evidence would suggest that ethnohistoric accounts of Nahua burial practices are likely inaccurate, at least with respect to certain segments of the population. For the Cholula collection, it would appear that cremation was not a common practice among the Postclassic inhabitants of this residential area and that we can, therefore, draw some general conclusions about morbidity and mortality from the analysis of these skeletal remains.

117 CHAPTER V

RECONSTRUCTING THE PALEODEMOGRAPHY OF CHOLULA

“…hay muchos indios que pasan de cien años.” (Rojas 1985: 135)

As the Cholula skeletal assemblage does not appear to suffer from the effects of selectivity biases apart from the underrepresentation of infants, the sample is an appropriate object of study to examine urban population dynamics in low-status residents of the city. The term paleodemography refers to the reconstruction of demographic characteristics of a population using human skeletal remains. Traditionally, paleodemographers have estimated the ages at death of individuals22 and the sex of adults in a collection of skeletons in order to identify general features of the mortality experience of the population.

In Chapter V, I discuss the problems associated with paleodemography, as well as some of the solutions that have been proposed to resolve them. In particular, I present a method of aging adult skeletal remains, referred to as transition analysis, that addresses some of the osteological challenges facing paleodemographers and a mortality model, known as the Siler model, that can be used to reconstruct the hazard of dying in paleodemographic samples. These methods are then applied to the Cholula skeletal sample in order to examine mortality in this population and to understand some of the basic features of the demography of this Postclassic

New World urban center.

22 While paleodemographers have traditionally estimated the ages at death for individual skeletons in a sample as a first step to reconstructing mortality, recent theoretical advancements in the field suggest that this approach may not be the most methodologically sound (Hoppa and Vaupel 2002). Further discussion of this theoretical reorientation in paleodemography, referred to as the Rostock Manifesto, will be presented later in the chapter.

118 A Crash Course in Traditional Paleodemographic Methodology

The following section is intended to orient readers who may be unfamiliar with basic paleodemographic concepts. The emphasis here is on giving readers enough background information that they are able to understand problems that will be raised later on in the chapter with some of these methodological approaches. The discussion below includes only the most basic of information and does not reflect all the techniques employed in paleodemography or the challenges faced by paleodemographers.

Paleodemographers working in recent decades have typically begun their reconstructions of the mortality of past populations by estimating the age at death of individual skeletons in the sample (see Footnote 17 for a cautionary note). In children, who are still growing and undergoing developmental changes, age determination is based on the timing of certain developmental markers. The formation and eruption of teeth takes place at fairly constant rate across populations and is, therefore, one of the most frequently used means of estimating the age of juveniles. Some parts of the skeleton, referred to as epiphyses, remain unfused during childhood to allow for growth. The timing of the fusion of various epiphyses can also be used in juvenile age estimation. The length of the diaphysis, or shaft, of long bones, obviously increases throughout childhood and adolescence as growth occurs. As stature is variable across populations, this means of age estimation has some disadvantages, but it may be used to estimate the ages of juveniles if other types of skeletal data are unavailable. In adults, skeletons can be aged by looking at the rate of degenerative changes in nonmobile joints such as the cranial sutures, the sacro-iliac joint (the joint between the sacrum and the pelvis, or innominate), the pubic symphysis (the joint between the left and right innominates), or the sternal end of ribs.

119 The age estimates for individual skeletons are then usually aggregated to produce an age-at-death distribution that shows the number of individuals that died at each age or in each age interval.

As the sex of individuals may influence their risk of morbidity or mortality, separate age- at-death distributions are often constructed for adult males and females in order to compare them. In light of the fact that the female pelvis must accommodate childbirth, particular characteristics of male and female pelvises make them readily distinguishable. Because human males and females are sexually dimorphic, that is, they differ physiologically in terms of size and muscle mass, the sex of adult skeletons can also be determined from features of the skull or from general differences in robusticity. However, the pelvis is the most reliable indicator of the sex of an individual. The sex of juveniles is generally not determined from skeletal remains, as the degree of sexual dimorphism before the onset of puberty is not sufficient to reliably differentiate males and females.

These data on age and sex have then been used to make further inferences about mortality in past populations. Traditionally, paleodemographers have calculated abridged life tables, a method of analysis widely used in demography, to summarize various aspects of mortality with respect to age (Hinde 1998: 30). Abridged life tables group individuals into five-year age intervals and then provide information, such as the age-specific mortality rate for each age interval, the probability of surviving from birth to exact age a, and the life expectancy at age a

(Hinde 1998: 30-48). While this approach to reconstructing mortality in skeletal populations characterizes the majority of paleodemographic studies, both past and current, this methodology has been widely criticized for reasons that will be explored below.

120 Paleodemographic Age-at-Death Distributions

Age determination in skeletal remains

In living populations, mortality rates are often used to estimate the overall health of a population, since malnutrition and disease increase the risk of death (Roberts and Manchester

1997: 27-28). However, calculating age-specific mortality rates for osteological samples has proven to be considerably problematic. One issue with which paleodemographers must contend is inaccuracies in age determination. Juvenile age determination is based on changes in the skeleton associated with growth, such as dental development or the fusion of epiphyses, which occur at a fairly regular rate across populations. Therefore, juveniles can be aged within a limited range of error.

In contrast, adult age estimation is much less precise because, with a few exceptions23, it is based on the timing of degenerative changes. As the rate of bone degeneration varies depending upon genetic factors, environmental variables, and physical activity levels, adult ages are significantly more difficult to determine than those of juveniles. Traditional means of aging adult skeletal remains include methods based on cranial sutures, the pubic symphysis, the auricular surface of the ilium, or the sternal rib ends. Tests of the accuracy of these methods have yielded less-than-favorable results. The aging method developed by Todd (1921), which relies on changes in the pubic symphysis, was found to overage younger individuals (Brooks 1955). Tests of McKern and Stewart (1957), a method of aging based on the pubic symphysis that was developed using the Korean War dead, indicate a high degree of inaccuracy with a strong bias

23 Some epiphyses fuse in early adulthood. The medial clavicle, for example, typically fuses sometime in the early 20s in males, although in some cases the epiphysis may remain unfused until around age 30 (McKern and Stewart 1957). Similarly, the ventral rampart, a ridge of bone that begins to develop on the ventral border of the pubic symphysis in early adulthood, reflects a process of bone formation rather than degeneration.

121 toward underaging individuals (Meindl et al. 1985a). Lovejoy et al.’s method (1985) of aging adult skeletons, based on degenerative changes in the auricular surface of the ilium, was found to overage younger individuals and underage older individuals by a significant amount (Murray and

Murray 1991).

While variations in the rate of degenerative changes contribute to the imprecision of these methods, Bocquet-Appel and Masset (1982) have raised concerns that aging techniques are also biased by the age-at-death distribution of the reference collection from which they were developed. As a result, the age-at-death distribution of a target sample (a collection under study) may mimic that of the reference collection used to construct the aging method. Konigsberg and

Frankenberg (1992; 1994; 1997) have explained mathematically why this occurs. When aging a skeleton from a target sample, we are looking for the probability that a skeleton is age α given that it has age indicator c [P(α│c)]. However, what is observed in the reference sample when constructing an aging method is the probability than an individual has age indicator c given that he or she is age α [P(c│α)]. These two probabilities are not equal to each other, but there is a mathematical relationship between the two that is specified by Bayes’ Theorem. In order to obtain the desired probability, P(α│c), researchers developing aging techniques generally rely on an inverse regression, regressing age on the age indicator. Mathematically, the assumption is made when deriving P(c│α) from P(α│c) that the age distribution of the target population is the same as that of the reference population. The result of this assumption is that the age distribution of the target population is biased in the direction of the reference population. For example,

McKern and Stewart’s age determination method based on the pubic symphysis was developed using the Korean War dead. This reference collection consists almost entirely of young males.

Consequently, paleodemographic age-at-death distributions constructed using this method will

122 show much higher young adult mortality than if they were constructed using a different aging technique.

Paleodemographic age-at-death distributions and the principle of uniformitarianism

Paleodemographic age-at-death distributions have long been criticized for their deviations from known human mortality patterns24. In general terms, human populations across time and space seem to share certain commonalities in regards to the age-specific force of mortality. In preindustrial societies, mortality is high in infants and young children but then falls off rapidly to a lifetime low in late childhood and early adolescence. Mortality begins to rise slightly at that point, but remains low throughout early and middle adulthood. It then begins to accelerate rapidly at later ages (Wood et al. 2002; Weiss 1973; Keyfitz and Fleiger 1971). A comparison of adult mortality in a wide range of societies indicates that in most populations the majority of adult deaths occur after the age of approximately 50 (Gurven and Kaplan 2007).

Populations do show variation in terms of the levels of excess mortality in early childhood or senescence, and the timing of the decline of juvenile mortality or the rise of senescent mortality may differ (Wood et al. 2002: 138). However, a general conformity to the basic pattern described above exists.

Paleodemographers, on the other hand, have typically reported unexpectedly low infant mortality and adult age-specific mortality that is abnormally high and that accelerates rapidly from age 15 to 50, resulting in few individuals surviving into old age (age-at-death distributions from skeletal samples, as well as from modern anthropological populations can be compared in

24 See Milner et al. (1989) for one exception. The Norris Farms age-at-death distribution is in line with expected patterns of human mortality.

123 Figures 5-1, 5-2, and 5-3). Howell (1976b, 1982) and Bocquet-Appel and Masset (1982) note that age-at-death distributions constructed from skeletal remains typically look significantly different than those produced by historical demographers, even though the historical and archaeological samples may not be that far removed in time.

While some anthropologists have argued that paleodemographic reconstructions reveal real differences in past human mortality experiences (Lovejoy et al. 1977), the striking similarities observed in skeletal age-at-death distributions across time and space make this position suspect. Furthermore, Howell (1976b, 1982) has postulated that if the mortality patterns observed in paleodemographic samples were accurate, they would have created a number of cultural challenges for the societies involved. Such rapidly accelerating mortality rates in adults would have resulted in burdensome workloads for surviving adult members of the population, large numbers of orphaned children, and insufficient knowledge to ensure cultural continuity due to the absence of elders. As this mortality pattern has not been observed in any well-documented historical or modern populations, paleodemographic age-at-death distributions violate uniformitarian principles. The principle of uniformitarianism in paleodemography is the idea that ―the human animal has not basically changed in its direct biological response to the environment,‖ although the rate of performance of biological processes may vary over time

(Howell 1976b: 26). Consequently, sample biases, such as poor preservation and recovery of infant remains, and methodological errors that affect the accuracy of adult age estimates, are most likely to blame for the unique appearance of most paleodemographic reconstructions.

124 Age-at-Death Distributions for Two Modern Anthropological Populations 60

30 !Kung Percentage 20 Yanomamo 10

0 0-5 5-15 15-25 25-35 35-45 45+ Age at death

Figure 5-1: The age-at-death distributions of two modern anthropological populations (Milner et al. 1989). Note that the majority of deaths occur before the age of 5 and after the age of 45. This pattern contrasts sharply with paleodemographic age-at-death distributions.

Hadza Adult Age-at-Death Distribution 60

30 Number 20

0 16-20 21-30 31-40 41-50 50+ Age at death

Figure 5-2 : The age-at-death distribution of the Hadza, another modern anthropological population. Note that most adult deaths occur after the age of 50. (from Blurton et al. 2002)

125

Three Paleodemographc Adult Age-at-Death Distributions 45 40 35 30 25 20 Libben 15 Teotihuacan 10 Xochimilco

Percentageof Adult Deaths 5 0 16-20 21-30 31-40 41-50 50+ Age at Death

Figure 5-3: Age-at-death distributions of three paleodemographic populations. Libben is a pre-Contact North American Indian site, Teotihuacan is a Classic Period urban center in the Basin of Mexico, and Xochimilco is a Postclassic urban center in the Basin of Mexico. Note that few adults live past the age of 5025. (Lovejoy et al. 1977; Storey 1992; P. Hernández 2006)

25 The age categories ―45+‖ and ―50+‖ in the graphs in Figures 5-1, 5-2, and 5-3 reflect that fact that many traditional aging methods have a terminal age category that is approximately 50+, as osteologists have long assumed that older adults could not be aged accurately.

126 Reconstructing mortality in skeletal assemblages

Until relatively recently, paleodemographers relied on the construction of life tables, which summarize information such as age-specific mortality rates, survivorship, and life expectancy, to study mortality in skeletal samples. However, in order to calculate these estimates, paleodemographers have assumed that the population under study was not only stable, but also stationary. A stable population is characterized by closure to migration, constant age-specific fertility and mortality rates, a stable age distribution, and a constant growth rate. A stationary population meets all of the conditions for stability, plus it has a growth rate equal to zero.

Paleodemographers in the past typically assumed stationarity because it made their jobs a great deal easier. In a stationary population, the proportion of individuals in each age category is constant over time, so abridged synthetic cohort life tables can be constructed based solely on the estimated ages at death for the skeletons. The ages at death are assumed to be equal to the dx column in the life table, and then, with a few minor manipulations, the other estimates contained in the life table can also be calculated. Using this information, paleodemographers frequently relied heavily on the calculation of life expectancy at birth –which, in a stationary population, is conveniently equal to the mean age at death of the skeletal sample-- to compare mortality in past societies. Life tables can also be calculated for stable populations in which the growth rate is not equal to zero, but this requires that paleodemographers have information on the growth rate, which traditionally has meant that they must rely on archaeological estimates or their best guess.

It is not necessarily unreasonable to assume that a population is stable. Lotka (1907, 1922) has demonstrated that it only requires about 100 years for a population to reach stability given constant age-specific fertility and mortality rates. Furthermore, even if perturbed from stability

127 by a major crisis, populations require only a few generations to regain their former structure

Weiss (1973) has also indicated that if a skeletal collection is deposited over several hundred years, stochastic variations should be smoothed out.

One potential impediment to assuming stability, however, is migration. Stable populations are, by definition, closed to migration. In smaller scale societies, in which migration is likely to be confined to the reciprocal exchange of mates, it has been demonstrated that this restriction can be relaxed because no net migration occurs, but many urban societies may have experienced high rates of immigration. For example, in medieval urban Danish parishes, Boldsen (1984) estimates that immigration may have been as high as 40%, and Wrigley (1967) has estimated that approximately one out of every six people living in England during the Early Modern period lived in London at some point in their lives. In general, the effects of migration on the paleodemography of past populations have received very little attention. Fortunately, however, demographic research has demonstrated that the assumption of stability is not terribly restrictive.

Most populations do, indeed, appear to conform to stable models, even when significant migration is taking place (Milner et al. 2000: 480; Keyfitz 1977: 89-92; A. Lopez 1961: 66-68;

Coale 1972: 117-161).

Stationarity is a more problematic assumption. It has been argued that groups in the past would have had very low growth rates, and that therefore, an assumption of stationarity is reasonable. This would, in fact, be very convenient for paleodemographers, as the age-at-death distribution would be solely a product of mortality, making the calculation of life tables a methodologically-sound endeavor. Although on a global scale over long periods of time, growth rates may have been close to zero, we cannot assume that such was the case for any individual population over a more limited period of time. Moreover, at some points in prehistory, such as

128 the transition to agriculture, populations were clearly not stationary (Johansson and Horowitz

1986: 237).

In nonstationary populations, changes in fertility rates have a more profound effect on the age-at-death distribution than do changes in mortality rates, since the number of deaths in each age category is a product of both the age-specific mortality rate and the number of people alive in each particular age group (Sattenspiel and Harpending 1983; Johansson and Horowitz 1986).

To illustrate how nonzero population growth can affect the age-at-death distribution, let us first imagine a population in which positive population growth is occurring because of increased fertility rates. In a growing population, more people are alive in each successive birth cohort and, therefore, the number of people that are at risk of death at each age increases (Milner et al.

2000: 480-481). Since each new birth cohort is larger than the last, more and more individuals are entering the mortality sample, even if the mortality rate itself remains constant. As infant and early childhood mortality is high in preindustrial societies, individuals ―pile up‖ at the youngest ages of the age-at-death distribution. The age-at-death distribution for a growing population will, thus, have a high proportion of young individuals. If stationarity is assumed to analyze this hypothetical population, then it will appear that mortality has increased. Conversely, if a population is declining in size, there will be fewer people alive in each successive cohort, and there will be a greater proportion of adults in the skeletal sample. An assumption that this declining population is stationary will result in the paleodemographer concluding that mortality has decreased. It is therefore, dangerous to assume stationarity without substantial evidence to support the claim. Many of the paleodemographic studies of the transition to agriculture have indicated that life expectancy declined with the adoption of agriculture, based on an assumption of stationarity. However, it may be the case that this apparent decline in life expectancy actually

129 represents a positive intrinsic population growth rate, not an increase in mortality (Wood et al.

1992). Since stationarity cannot be assumed for most populations, the calculation of life tables becomes considerably more challenging.

Estimating the age-at-death distribution

What, then, can be done to correct these methodological problems? Bocquet-Appel and

Masset (1982), in fact, suggested that the challenges faced by the discipline were insurmountable and that the pursuit of demographic data from skeletal samples was a foolhardy endeavor at best.

Since their initial criticism of paleodemography, however, a number of more optimistic suggestions have been made regarding how to correct some of these methodological difficulties.

In 1999 and 2000, two workshops were held at the Max Planck Institute for Demographic

Research in order to evaluate various methodological approaches for estimating adult ages and reconstructing paleodemographic profiles. At the conclusion of the first workshop, a theoretical framework for paleodemography, referred to as the Rostock Manifesto, was agreed upon by the participants. The Rostock protocol suggests four steps for addressing the criticisms that have been leveled at paleodemography (Hoppa and Vaupel 2002). First, working with known-age reference collections, osteologists must develop age indicator stages that are more reliably related to chronological age. Second, methods must be developed to estimate P(c│α), the probability of observing a particular age indicator stage c given some age a, from the osteological data collected from the known-age reference sample. Third, osteologists and paleodemographers must recognize that what we are really interested in is P(α│c), the probability that a skeleton is age a given that age indicator c has been observed, not P(c│α),

130 and that P(α│c) must be calculated from P(c│α) using Bayes’ Theorem. However, the use of

Bayes’ Theorem first requires knowledge about f(a), the distribution of ages at death in the target population. The fourth step, therefore, must be to find a way to either model or estimate f(a).

Note that the Rostock protocol calls for estimating f(a), the age-at-death distribution of the target sample, before estimating the individual ages at death in the skeletal collection. So how can paleodemographers accomplish this seemingly counter-intuitive task? First, the conditional probability that an individual has some age indicator given that he or she is age a

P(c│α) must be estimated from a known age-at-death reference sample. A parametric model of mortality is then selected to represent the age-at-death distribution of the target population. The

Siler model, which will be presented momentarily, is but one example of a model that could serve just such a purpose. The observed frequencies of skeletal indicators of age from the target sample, combined with the information about P(c│α) from the reference sample, allow the parameters of the mortality model to be estimated using the following likelihood equation:

│a) Pr(a) da

where Pr (ci) refers to a vector of age indicator stages observed in the target population, a refers

* to age, Pr (ci│a) refers to the probability observed from the reference collection that a skeleton has a particular age indicator ci at age a, and Pr (a) is the age-at-death distribution of the target population. Individual age-at-death estimates, as well as their error distributions, can then be produced using Bayes’ Theorem (Hoppa and Vaupel 2002; Wood et al. 2002; Love and Müller

2002; Holman et al. 2002).

131 While this is perhaps the most theoretically-sound means of reconstructing paleodemographic profiles, it may not be practical in all cases. As the number of age indicators used increases, the number of parameter values to be estimated also increases (Holman et al.

2002). Not only is a significant amount of computer power necessary for this endeavor, but a sufficient sample size is also required. While it is not clear exactly how many skeletons are necessary, many osteological samples are likely to be too small to support the estimation of f(a) as outlined in the Rostock Manifesto (Boldsen et al. 2002: 77). The sample from Cholula being used in this analysis, which consists of only 280 skeletons with associated aging information, may not be large enough to allow for compliance with the Rostock protocol. Consequently, an alternative approach to reconstructing the age-at-death distribution must be used instead.

Transition analysis, developed by Jesper Boldsen and George Milner (Boldsen et al.

2002), is an adult aging method that addresses the problem of age mimicry of the reference sample, although it does not calculate f(a) from the target population itself, as the Rostock protocol recommends. It does, however, offer a means of improving age estimates that is still applicable to the small sample sizes with which paleodemographers often have to work.

Transition analysis relies on age-related information collected from five different cranial sutures, the pubic symphysis, and the iliac portion of the sacroiliac joint using the scoring system established by Boldsen et al. (2002). These data are then entered into a computer program, which combines this osteological information in a statistically meaningful way, and calculates a maximum likelihood estimate of age using Bayes’ Theorem and either a uniform prior distribution of f(a) or an external f(a).

The uniform prior probability assumes that all ages are equally likely to be represented in the age-at-death distribution of the population. Clearly, this is not a theoretically-valid

132 assumption, as some age groups are much more likely to be represented in a skeletal sample than others. Furthermore, it places disproportionate weight on the oldest ages at the uppermost part of the age-at-death distribution (Boldsen et al. 2002: 77-78). While assuming a uniform prior probability avoids injecting information about the age-at-death distribution of the reference sample used to develop the aging method, it does not completely resolve the problem, as it still influences the age-at-death distribution of the target population. However, the other option available in the transition analysis program is to use an external f(a) for an archaeological population, which is based on seventeenth-century rural Danish parish records (Boldsen et al.

2002:88). External priors can be derived from historical documents or other sources independent of the target population under study and, if they are appropriately chosen, they can provide better age estimates than will be obtained from the use of a uniform prior. Unfortunately, data of this type do not exist for Cholula, and paleodemographers must be cautious in applying external f(a)’s to a target population without firm evidence that both populations would have similar age- at-death distributions. Clearly, neither the use of the uniform prior nor the external prior is the ideal solution to the problems associated with age estimation, but they provide an alternative approach when small sample sizes preclude the implementation of the full Rostock protocol.

One of the most significant benefits of transition analysis is the improved estimates of age for older adults. In the past, this has been a problematic issue in paleodemography, and many methods of adult age estimation have an open-ended terminal age category, such as 55+.

While these open-ended categories acknowledge the difficulties of accurately aging older individuals, they also limit the information regarding senescent mortality that can be obtained from the skeletal sample. The transition analysis program includes age information from various parts of the skeleton, which aid in aging adults. In particular, characteristics such as posterior

133 iliac exostoses, which are more likely to be present in the oldest individuals, address deficiencies in traditional aging methods in estimating the ages of very old members of the population.

In order to test the validity of transition analysis, Milner (2010) applied the method to

239 skeletons from the Bass Collection, assuming both the uniform prior and the archaeological prior distributions. He found that age estimates for adults produced using the full method, which includes the cranial sutures, the pubic symphysis, and the iliac portion of the sacroiliac joint, tended to underestimate the ages of those adults over 60. Up until age 40, there was little diffence in the age estimates produced using the archaeological distribution and those produced using the uniform prior distribution. After age 40, the discrepancy between age estimates from the archaeological and uniform prior distributions was approximately five years or less. When tested separately, the pubic symphysis performed best, while the cranial sutures produced the least accurate and least precise results. Interestingly, the width of confidence intervals for age estimates increased from the fifties through the seventies, but after that they began to decrease.

While transition analysis represents an improvement over traditional methods of age estimation, it still lacks the degree of accuracy and precision we would ultimately like to have in estimating adult ages.

Parametric models of mortality

Numerous attempts have also been made to find alternate methods to analyze mortality in skeletal samples (Milner et al 1989; Konigsberg and Frankenberg 1992; Gage and Dyke 1986;

Gage 1988). One promising solution for studying mortality in past populations is the use of parametric models to understand the age-at-death distribution of the population under study.

134 Parametric models of mortality are mathematical equations that include a number of parameters whose values must be estimated from the empirical data, such as the age-at-death distribution of the skeletal sample. These parameter estimates can then form the basis of inferences about mortality in the population under study. One such mortality model is the Siler model, which consists of three additive components:

-β t β t = α1e 1 + α2 + α3e 3

where t refers to the age-at-death, and h(t) refers to the hazard of dying at some age t. The first

-β t component (α1e 1 ) captures the high mortality of infancy and early childhood, which then declines rather rapidly. The second component (α2) represents a constant ―baseline,‖ or age-

β t independent, mortality, and the third component (α3e 3 ) is senescent risk, or the increasing risk of death with age (Gage and Dyke 1986; Gage 1988). In these equations, α1,2,3 and β1,3 are parameters, or constants, that are to be estimated from the skeletal sample itself. The Siler model can thereby provide us with a means to estimate h(t) for skeletal samples, which is age- specific mortality, given survival to age t. We can also find two probabilities mathematically related to h(t): S(t), or survivorship, is the probability of surviving from birth to age t, and f(t), the probability density function, is the unconditional probability of dying at age t rather than at some other age.

The Siler model does have some limitations in that it does not capture the ―accident hump‖ associated with late adolescence and early adulthood, nor does it capture the mortality slowdown that occurs at the oldest ages (Wood et al. 2002). It is uncertain, however, that these are universal features of human mortality, and paleodemographic samples are too imprecise to

135 capture these features even if they do exist. In general, therefore, the Siler model does a reasonably good job of modeling human mortality.

Gage (1988; 1989; 1990; 1994; Gage and Dyke 1986) has done an extensive amount of research concerning the Siler model and its application to anthropological populations, and

Wood et al (2002) have also discussed the use of this model, as well as others, in the construction of age-at-death distributions. The primary benefit of using a parametric mortality model over life tables in paleodemography, besides the fact that life tables require much larger sample sizes and more information about the target population than skeletal samples can provide, is that paleodemographers must only assume that the population under study is stable, not necessarily stationary. Because of ―weak ergodicity,‖ most populations do, in fact, approximate the conditions of stability, even if they are open to migration or fertility and mortality rates are changing (Wood et al. 2002: 135-136; A. Lopez 1961:66-68).

The parameters of the Siler model (as well as other mortality models) can be calculated using maximum likelihood estimation. The primary objective of a maximum likelihood analysis is to find the parameters for a given model that make the observed data most likely. Darryl

Holman (2002) has created a computer program called mle, which uses an iterative technique to generate estimates of the parameters once the appropriate likelihood equations have been programmed. Usher (2000: 22-23) tested the ability of mle to correctly estimate the parameters of the Siler model using simulated data and found that the program was, in fact, able to capture the parameter values, even with sample sizes as small as 50.

Parametric models of mortality have now been used in several paleodemographic studies.

Usher (2000) fit the Siler model to paleodemographic data collected from the medieval Danish cemetery of Tirup using mle, but as the age-at-death distribution of Tirup has some unexpected

136 features, she concluded that the Siler model did not fit the Tirup data well. Dewitte and Wood

(2008) also used the mle program to fit the Siler model to paleodemographic data from both the

East Smithfield Black Death cemetery and two ―normal mortality‖ Danish cemeteries, and determined, based on the results, that older individuals were at a higher risk of death from the disease, which is consistent with information from historical documents.

The Paleodemography of Cholula

As part of the paleodemographic analysis of the Cholula assemblage, data on age and sex were collected from the skeletons (see Appendix B for data regarding age and sex in Cholula).

Sex was assigned to adults in the Cholula collection based upon an evaluation of the skull and pelvis using criteria established by Phenice (1969) and Buikstra and Ubelaker (1994). Features of the skull, including the supraorbital ridge, the mastoid process, the occipital protuberance, the gonial angle, and the mental eminence were subjectively scored, as were characteristics of the pelvis such as the ventral arc, the subpubic concavity, the ischiopubic ramus, the greater sciatic notch, and the preauricular sulcus (see Figures 5-4 to 5-9 and Appendix C). Overall robusticity of the skeleton was also considered. As the pelvis, and particularly the pubic symphysis, is the most accurate means of determining sex, preference was given to indicators of sex from the pelvis in all ambiguous cases (Buikstra and Ubelaker 1994). Sex determination was not attempted in juveniles (individuals less than 15 years of age) as sexual dimorphism sufficient to differentiate males and females does not develop until puberty (Buikstra and Ubelaker 1994).

137

Figure 5-4 Figure 5-5

Figure 5-6 Figure 5-7

Figure 5-4: Picture showing the supraorbital ridge, a ridge of bone above the orbits, which is more pronounced in males than females.

Figure 5-5: Picture showing the gonial angle of the mandible. In males, the gonial angle tends to be a right angle, while in females it is more oblique.

Figure 5-6: Picture showing the mastoid process on the temporal bone. As the mastoid process is a muscle attachment area, it tends to be larger in males than in females.

Figure 5-7: Picture showing the occipital protuberance. Again, as a muscle attachment area, the occipital protuberance is more pronounced in males than females. Photo taken with permission of the DAF of the INAH.

138 Ventral arc Subpubic concavity

Figure 5-8: Picture showing the ventral arc and the subpubic concavity of the pelvis. The ventral arc is a ridge of bone on the ventral face of the pubic symphysis that is present in females. The subpubic concavity refers to the shape of the inferior surface of the pubic symphysis. In females this area is typically concave, while in males it is convex.

Preauricular sulcus

Ischiopubic ramus

Greater sciatic notch

Figure 5-9: Picture in which the ischiopubic ramus, the greater sciatic notch, and the preauricular sulcus can be seen. The ischiopubic ramus is narrow in females and wide in males. The greater sciatic notch is wider in females than in males. The preauricular sulcus is a groove just anterior to the auricular surface. The preauricular sulcus is deeper and more pronounced in females than in males.

139

Age determination in juveniles was based on dental development, the union of epiphyses, and, when necessary, diaphyseal length. As rates of dental development seem to be relatively invariant across populations (Smith 1991), tooth formation and eruption were preferentially used when possible. Ages were assigned to juveniles in accordance with the criteria established in

Ubelaker (1989) (see Appendix D). Fetal and neonate ages were established using measurements of available cranial elements and long bones (Fazekas and Kosa 1978). The midpoint of juvenile age estimates was used in the construction of the age-at-death distribution.

Adult ages were estimated using transition analysis. Cranial suture closure and characteristics of the pubic symphysis and the iliac portion of the sacroiliac joint were scored according to the protocol outlined by Boldsen et al. (2002) (Figures 5-10 to 5-15 and Appendix

E). This information was then entered in the ADBOU Age Estimation software and maximum likelihood estimations of age based on both a uniform prior and an archaeological distribution were generated (see Appendix E for an example). The point estimates of age generated using the uniform prior distribution were used in the construction of the age-at-death distribution for

Cholula26. Prior to beginning my study of the Cholula skeletal material, I performed a test of my proficiency with the method using a sample of known-age pubic symphyses. The results of this test are presented in Appendix F. Separate age-at-death distributions were generated for the

Cholulteca II and Cholulteca III skeletons to determine if significant changes in mortality occurred over time (see Table 5-1 and 5-2 and Figures 5-16 and 5-17 below). As no statistically

26 It should be noted here that using the point estimates of age ignores the error associated with aging skeletons. The error associated with aging young adults tends to be much smaller than the error associated with aging older individuals. I have chosen to present the age distribution of Cholula in 5 year increments to emphasize the overall pattern of the distribution, but doing so in no way accurately reflects the error of the age estimates. An alternative would be to present the age-at-death distribution in one-year increments; however, I feel that it is much more difficult for the non-specialist reader to interpret the overall pattern from the peaks and valleys of this type of presentation.

140 significant differences exist between the two distributions, they were combined into a single age- at-death distribution for Cholula (see Table 5-2 and Figure 5-18 for the combined age-at-death distribution of Cholula using a uniform prior probability and Figure 5-19 for a comparison of the uniform prior and archaeological distributions). From an archaeological perspective, combining

Cholulteca II and III is, perhaps, the best course of action as questions have been raised about the validity of the ceramic chronology used to distinguish these two phases (McCafferty 1996b)27.

In addition to using transition analysis, the Cholula skeletons were also aged using Todd

(1921), McKern and Stewart (1957) (males only), Gilbert and McKern (1973) (females only),

Brooks and Suchey (1990), and Lovejoy et al. (1985) for comparative purposes (see Appendix G for a brief overview of these methods). These traditional aging methods28 suffer from the age mimicry problem that Bocquet-Appel and Masset (1982) criticized. The age-at-death distributions constructed using these aging methods appear below (Figure 5-20).

27 Several different ceramic chronologies have been elaborated for Cholula, none of which are closely associated with radiocarbon dates (Lind 1994; Plunket 1995; Suárez 1994; McCafferty 1992). However, a recent dating project by archaeologists at the Universidad of the Américas is more firmly establishing the chronology of the site and may offer some clarifications about the dating of ceramics (Plunket and Uruñuela 2005; Uruñuela and Plunket 2005). In the near future, it may be possible to revisit the dating of these burials to more narrowly establish their temporal provenience.

28 Throughout the rest of the dissertation, the aging methods of Todd (1921), McKern and Stewart (1957), Gilbert and McKern (1973), Brooks and Suchey (1990), and Lovejoy et al. (1985) will be referred to collectively as traditional aging methods because their use results in age mimicry of the reference collections on which their development was based.

141

Coronal pterica Lambdoidal asterica

Figure 5-10 Figure 5-11

Sagittal obelica Zygomaticomaxillary Interpalatine

Figure 5-12 Figure 5-13

Figures 5-10: Picture showing coronal pterica, one of the sutures using in transition analysis

Figure 5-11: Picture showing lambdoidal asterica, one of the sutures used in transition analysis.

Figure 5-12: Picture showing sagittal obelica, one of the sutures used in transition analysis.

Figure 5-13: Pictures showing the interpalatine and zygomaticomaxillary sutures, two of the cranial sutures used in transition analysis. All photos taken with permission from the DAF of the INAH.

142

Superior apex

Dorsal symphyseal margin

Ventral symphyseal margin

Figure 5-14 : Picture showing the face of the pubic symphysis. The features of the pubic symphysis used in transition analysis are labeled. Symphyseal relief and symphyseal texture are assessed over the entire face. Readers should consult Appendix E for scoring procedures.

143

Superior posterior iliac surface Posterior iliac area

Inferior posterior iliac surface

Superior auricular surface

Apical auricular surface Inferior auricular surface

Figure 5-15: Picture showing the features of the iliac portion of the sacroiliac joint used in transition analysis. Readers should refer to Appendix E for scoring procedures.

144

Table 5-1: Distribution of Cholula skeletons by time period, sex, and age.

Age Cholulteca II Cholulteca III Total

Infants 6 26 32

Children (1-14) 17 78 95

Adult Males 27 57 84

Adult Females 26 57 83

Adults 2 13 15 (Sex Undetermined) Total 78 231 309

145

Table 5-2: Number of individuals within each age group. Note that only individuals who could be aged are included here.

Age Cholulteca II Cholulteca III Total

0 6 26 32

1-5 13 54 67

6-10 3 18 21

11-15 4 15 19

16-20 3 7 10

21-25 4 4 8

26-30 1 5 6

31-35 4 1 5

36-40 2 5 7

41-45 0 5 5

46-50 3 14 17

51-55 6 17 23

56-60 8 14 22

61-65 7 11 18

66-70 2 9 11

71-75 3 2 5

76+ 3 1 4

Total 72 208 280

146 Cholulteca II 14

6 Number 4

Age at death

Figure 5-16: Age-at-death distribution of Cholulteca II skeletons constructed using transition analysis and assuming a uniform prior probability.

Cholulteca III 60

30 Number 20

Age at death

Figure 5-17: Age-at-death distribution of Cholulteca III skeletons constructed using transition analysis and assuming a uniform prior probability.

147 Cholula Age-at-Death Distribution 80

Number 30

40 75 15 20 25 30 35 45 50 55 60 65 70

------

76+

36 71 11 16 21 26 31 41 46 51 56 61 66 Age at death

Figure 5-18: Cholula age-at-death distribution constructed using transition analysis, assuming a uniform prior probablility and combining Cholulteca II and III.

Cholula Adult Age-at-Death Distribution 70

30 Uniform Prior Archaeological Prior 20

Percentageof Adult Deaths 10

0 15-20 21-30 31-40 41-50 50+ Age-at-Death

Figure 5-19: Graph comparing the age-at-death distributions of Cholula constructed using a uniform prior distribution and an archaeological prior distribution. Assuming the archaeological distribution produces a slightly higher modal age of death in the 60s.

148 Age at Death Distribution of Cholula Using Traditional Aging Methods 50 45 40 35 30 Todd 25

Number 20 McKern and Stewart 15 Brooks and Suchey 10 Lovejoy et al. 5 0 16-20 21-30 31-40 41-50 50+ Age at death

Figure 5-20: Graph comparing the age-at-death distributions of Cholula constructed using various traditional aging methods.

Previous Age-at-Death Distribution for Cholula (Serrano 1973) 160 140 120 100 80

Number 60 40 20 0 18-20 21-35 36-55 55+ Age-at-death

Figure 5-21: Graph showing an age-at-death distribution of the Cholula population constructed by a different researcher using traditional aging methods. This researcher included sacrificial burials, which partially accounts for the very large number of young individuals. Note that there are very few individuals over age 55 (based on Serrano 1973).

149 Recall from earlier in the chapter that paleodemographic age-at-death distributions have been called into questions because most indicate patterns of mortality not found modern or historical populations. When traditional aging methods are used to estimate age in the Cholula sample, mortality peaks in adults in the fourth and fifth decades of life and comparatively few people survive into old age. A paleodemographic reconstruction of the Cholula population done by another researcher (Serrano 1973) using traditional aging methods is shown in Figure 5-21.

Note that in this distribution as well, almost no one survived past the age of 55. The adult age-at- death distributions for Cholula constructed using traditional aging methods look very much like other paleodemographic age-at-death distributions (refer back to Figure 5-3) and very unlike age- at-death distributions from modern anthropological populations (refer back to Figures 5-1 and 5-

2).

The argument has been made that the extremely high young adult mortality noted in most paleodemographic samples is, in fact, a reflection of a real human mortality pattern that existed in past populations. Lovejoy et al. (1977) have suggested that in the past, humans were exposed to much lower levels of infectious disease. As a result, frail individuals were more likely to survive childhood, but their reduced immunocompetence made them vulnerable as adults. This explanation is thoroughly unconvincing (Howell 1982). If the low levels of infectious disease

(which is a questionable assumption in and of itself) did not kill frail individuals as children, why would their mortality suddenly increase as adults in the face of the same low levels of infectious disease?

Paine (2000), using simulated data, attempted to determine if mortality crises, such as epidemic diseases or severe famines, could produce this unusual mortality pattern. He found that an isolated crisis that kills indiscriminately would produce such a pattern of high young adult

150 mortality. However, this pattern would shortly be erased from the osteological record by a return to a normal mortality regime, as has been demonstrated by numerous study of population stability (Weiss 1973; Lotka 1907, 1922; A. Lopez 1961). It would require repeated crises to maintain this age-at-death distribution in the small sample sizes typically available to paleodemographers.

Paine (1997) has also tested the suggestion that high rates of immigration into a population could account for the high numbers of young adults in skeletal collections. As immigrants tend to be young adults, these individuals might account for the unusual mortality observed in paleodemographic samples. Paine’s simulated data, however, showed that high rates of immigration resulted in an increase in the proportion of juveniles in the mortality sample, attributable to the reproductive contributions of the childbearing-age immigrants, but resulted in little change in the number of deaths of young adults. Paine does make the assumption in his simulation that immigrants and native residents have the same mortality levels, which may not be justified. However, if immigrants are the reason for high young adult mortality in paleodemographic populations, one has to wonder where all these immigrants are coming from.

Why are there so many immigrants in all populations that paleodemographers have studied, regardless of the size or level of complexity of the society? Why do paleodemographers never find the populations of origin of all these immigrants?

Modern and historical demographers have studied a wide range of societies across time and space, and human mortality conforms to certain general patterns with few deviations. As infants have been exposed to fewer selective factors than other age groups in the population, we would expect the greatest variation in frailty, and, therefore, the highest mortality to occur in these individuals. As the frailest individuals are selected out of the population by disease,

151 malnutrition, or other circumstances, mortality should decline considerably, since the individuals left alive have demonstrated a certain amount of immunocompetence. As individuals continue to age, degenerative diseases and declining immunocompetence begin to take their toll. These are, in fact, the most salient features of the human mortality profile -- high infant mortality that declines rapidly, low mortality in young adults, and increasing mortality in old age – that have been observed over and over again by demographers.

Figure 5-22: Graph comparing the adult modal ages at death in different types of societies. (Gurven and Kaplan 2007)

Note that the age-at-death distribution of the Cholula population constructed using

McKern and Stewart’s method (1957) shows a particularly young modal age at death. If traditional aging methods are sufficiently accurate and are capturing a true feature of human

152 mortality, why does McKern and Stewart yield a modal age at death ten years younger than the other methods? Given the composition of the reference sample used to develop this method

(young men who died during the Korean War), it is absolutely predictable that it would yield a lower modal age at death. Numerous studies have now demonstrated both theoretically and empirically that traditional aging methods are seriously flawed, and that the paleodemographic profiles constructed from them produce abnormal patterns of human mortality that cannot be adequately explained, rendering them highly suspect. We must assume, therefore, that what is being observed in paleodemographic age-at-death distributions constructed using these problematic aging methods is the result of methodological error rather than a true variation of human mortality.

Transition analysis, while not completely resolving all the methodological problems of adult age estimation, does avoid the issue of age mimicry of the reference collection and it improves age estimates for older individuals. The age-at-death distribution for Cholula constructed using transition analysis (under the assumption of a uniform prior probability29) shows a modal age of death in the sixth decade. Mortality in young adults is low. In addition, more than half of those individuals who live to adulthood survive past the age of 50, and a few individuals live into their 70s and beyond. The age-at-death distribution of this population is no longer outside the range of known human mortality patterns. Other than the underrepresentation of infants likely caused by preservation and recovery biases, mortality in Cholula conforms to what is known about human mortality from historical and anthropological societies.

29 The age estimates produced assuming a uniform prior distribution were used in all of the analyses presented in this dissertation. While age estimates were also generated using the archaeological prior for comparative purposes, as shown in Figure 5-19, the archaeological prior age estimates were not used in any analysis. This decision was made because I have no basis for assuming that the age-at-death distribution of the Cholula population would be similar to that of a rural Danish parish. However, the uniform and archaeological priors actually produced very similar age-at-death distributions.

153 Some anthropologists are sure to be skeptical of the age-at-death distribution of Cholula constructed using transition analysis because of their belief that people in the past did not live into old age. In response, I would first argue that people in traditional and historical societies that have been studied by demographers do occasionally live into their 70s and 80s. A comparison of mortality in different types of societies has shown that the modal age of death for adults is nearly 70 and that it is similar in all preindustrial populations studied, regardless of complexity (see Figure 5-22 above) (Gurven and Kaplan 2007). Second, I would also like to point to a quote from Gabriel de Rojas, written about the indigenous population of Cholula less than a century after many of the individuals in the collection under study were buried: ―…hay muchos indios que pasan de cien años.‖ (There are many Indians that are more than 100 years old.) (1985: 135). While I would not make the claim that Cholulans kept perfectly accurate counts of their ages, I would argue that the above quote strongly suggests that it was not uncommon for ancient Cholulans to live into old age.

The Siler Model Applied

The mle program was used to estimate the parameters of the Siler model from the observed age-at-death distribution of Cholula. Both the simplex and simulated annealing methods of maximization were tested with minimal to no differences in the results30. The estimated parameter values are presented in Table 5-3, and Figures 5-23, 5-24, and 5-25 show the hazard of dying, the survival curve, and the probability density function (PDF) generated

30 The mle program includes four methods of maximizing the likelihood function, each of which has its own strengths and weaknesses. The simplex and simulated annealing methods are two of the ways in which the parameter values that maximize a particular function can be found.

154 using the estimated parameter values. The α1, β1, and β3 parameter estimates were significantly different from zero (p<0.05).

The Siler model does assume that age data are drawn from a stable population, so the assumption of stability will be made in the following discussion. While it is unwise to assume that a paleodemographic sample was drawn from a stationary population, the assumption of stability is less restrictive. A stable population, as discussed earlier, is technically defined as one that had unchanging fertility and mortality rates and that is closed to migration. However, ―even when fertility and mortality rates are changing and substantial migration is taking place, most human populations still closely approximate a stable age distribution at any given time‖ (Milner et al. 2000). The exception to the above statement occurs when a population is subjected to a catastrophic event, such as the introduction of an epidemic disease previously unknown in the population, which significantly alters demographic rates (Milner et al. 2000). The archaeological context of the Cholula skeletons argues strongly against the collection being the result of a catastrophic event. Furthermore, the Cholula age-at-death distribution would seem to corroborate that this skeletal collection represents an attritional, rather than a catastrophic assemblage (Margerison and Knusel 2002).

Table 5-3: Siler parameter estimates for Cholula. Loglikelihood = -1132.401, AIC = 2274.8019 Parameters Parameter estimates S.E.

α1 0.1282 0.0184

β1 0.2366 0.0556

α2 0.0000 0.0120

α3 0.0061 0.0064

β3 0.0356 0.0108

155

Cholula Hazard of Dying 0.25

0.2

0.15 h(a) 0.1

0.05

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Age

Figure 5-23: Siler hazard constructed for Cholula using the parameter estimates in Table 5-3.

156

Cholula Survival Curve 1

0.9

0.8

0.7

0.6

0.5 S(a) 0.4

0.3

0.2

0.1

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Age

Figure 5-24: Survival curve for Cholula constructed using Siler parameter estimates from Table 5-3.

157

Cholula PDF 0.16

0.14

0.12

0.1

0.08 f(a)

0.06

0.04

0.02

0 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 84 88 92 96 100 Age

Figure 5-25: PDF of Cholula constructed using Siler parameter estimates from Table 5-3.

158 Infant and Childhood Mortality

While juvenile mortality is an important component of any human mortality profile, it is especially important for the current study as infants and children in preindustrial Old World cities had particularly high mortality, which typically exceeded mortality levels of juveniles in smaller towns and villages (refer back to Tables 2-1, 2-2, and 2-3 for infant and childhood mortality rates from London, Geneva, and York. Table 2-4 shows juvenile mortality rates for four English villages for comparison.). Landers (1993: 139-148) argues that high infant mortality in London in the seventeenth and eighteenth centuries was, in part, caused by a high incidence of water- and food-borne infections within this age group, suggesting that the practice of artificial feeding – partially or completely supplementing infants’ diets with food and water – was a widespread practice. Epidemic diseases also factored into high infant and early childhood mortality rates in Old World cities, as shown by studies of York (Galley 1998) and the London

Quakers (Landers 1993). Children in York between the ages of one to four seem to have been particularly affected by outbreaks of diseases such as smallpox that occurred every three to four years (Galley 1998: 104-108).

Artificial feeding was unlikely to have been widely practiced in Cholula. On the contrary, ethnohistoric sources indicate that children frequently breastfed up to age three or four

(Sahagún 1997), which would have offered infants some protection against infections, particularly gastrointestinal infections caused by contaminated food or water. Epidemic diseases would not have figured in to the mortality experience of children at Cholula. However, we must be careful about thinking that mortality from these causes of death can simply be

―subtracted out‖. Landers (1993: 154), for example, calculates age-specific mortality rates for

159 children eliminating the effects of smallpox, which results in a seemingly significant drop in infant and childhood mortality rates. What we must remember, though, is that many of these children were frail individuals who would have died from some other cause in the absence of epidemic disease. We cannot assume a priori, therefore, that infant and childhood mortality would have necessarily been lower in Cholula than in preindustrial Old World cities if competing causes of death are taken into account.

Unfortunately, infant mortality is very difficult to assess in paleodemographic populations because of the preservation and recovery biases discussed earlier. Looking at the

Cholula age-at-death distribution, we would expect infant mortality to be higher than it is given the levels of early childhood mortality, suggesting that infants are, in fact, underrepresented in the collection. One of the attractive features of using parametric models of mortality is that they can be used to estimate the number of ―missing‖ infants in a skeletal collection (Usher 2000). If we use the α1 parameter estimate as a guide, we would expect infants to account for 12.8% of the sample rather than 10.7%, indicating that there are 6 infants ―missing‖ from the collection.

However, this would be a very small number of missing infants, and given the less than stellar preservation of the collection and complicated burial situation, the Siler model constructed for

Cholula still appears to grossly underestimate missing infants.

In order to test if the low number of infants is biasing the estimates of the juvenile mortality component in the Siler model, I estimated the Siler parameters again, this time omitting children under one year of age. The parameter estimates are presented in Table 5-4, and a comparison of the two Siler hazards is shown in Figure 5-26. Removing infants from the model has little effect on the parameter estimates or the hazard.

160 Table 5-4: Siler parameter estimates for individuals over age 1. Loglikelihood = -1035.006, AIC = 2080.0114 Parameters Parameter estimates S.E.

α1 0.1148 0.0237

β1 0.2092 0.0661

α2 0.0000 0.0127

α3 0.0057 0.0063

β3 0.0365 0.0113

A Comparison of the Siler Hazard Constructed 0.25 for All Individuals with the Siler Hazard Constructed for Those Over Age One 0.2

0.15

h(a) All Skeletons 0.1

Over 1

0.05

Age

Figure 5-26: A comparison of the hazard for all ages in the Cholula collection with the hazard for only those individuals over age one.

161

Cholula Juvenile Age-at-Death Distribution 35

Number 15

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Age

Figure 5-27: Juvenile age-at-death distribution for Cholula by single-year increments.

162 The fact that the hazard of dying for juveniles in the Cholula collection falls off so slowly in early childhood is somewhat unusual (see Figure 5-26 and 5-27). Typically, mortality declines rapidly after infancy, with older juveniles experiencing fairly low mortality rates. Given that the ages of juveniles can be estimated with minimal error, it is highly unlikely that methodological problems are to blame for the abundance of older juveniles. It is possible that population growth might be responsible for the pattern of juvenile mortality in the Cholula collection. As discussed previously, the population growth rate can have a profound effect on the age-at-death distribution because the number of individuals entering the mortality sample is, in part, dependent upon the number of individuals at risk of death. Later on, I will discuss what must be done in future studies of Cholula in order to calculate an estimate of the growth rate.

Because of the likely underrepresentation of infants, it is difficult to speculate how infant mortality might compare to Old World cities. Based upon high perinatal and infant mortality at

Tlajinga 33 at Teotihuacan, Storey (1985, 1992) concludes that New World cities were unhealthy places. However, the Tlajinga 33 paleodemographic information must be viewed with some caution. First, preservation of perinatal and infant skeletons was exceptionally good in this collection as a result of the cultural practice of burying very young individuals in ceramic urns

(Storey 1992: 77). While an absolutely wonderful feature of this skeletal sample, the extraordinary preservation, and therefore, much higher numbers of perinatal and infant skeletons than are found in other paleodemographic collections, should not be confused with higher mortality. Furthermore, the population growth rate of Tlajinga 33 is unknown. The large numbers of perinatal and infant skeletons may be the result of high fertility as opposed to high mortality, making it difficult to conclude that Teotihuacan was necessarily a population sink dependent upon immigration from rural areas.

163 Young Adult Mortality

Another important feature of the demography of preindustrial Old World cities was the contribution of young adult immigrants to fertility and mortality rates. Immigration was a particularly significant feature of the demography of London because rates of immigration to the city were so high. While it is extremely difficult to quantify the number of immigrants that would have been present in London in the seventeenth century, Finlay (1981: 66-67) estimates that apprentices comprised 15% of the population and that a high percentage of apprentices were immigrants from other regions. He indicates that up to 15% of apprentices died before finishing their seven-year term. Finlay also provides data suggesting that aliens, defined as individuals coming from other countries, made up another approximately 1% to 5% of the population of

London from the mid-sixteenth to the mid-seventeenth centuries (1981: 68-69). Other types of immigrants, for example, those individuals who came to London from rural areas to work as servants, would have also been present in the population. For eighteenth-century London,

Landers (1993: 181-183) discovered that the age distribution of the population was heavily weighted towards young adults. Furthermore, estimating net immigration as a percentage of initial cohort size, he found very high rates of immigration in teenagers and young adults.

When data from preindustrial Old World cities regarding deaths from epidemic diseases are considered, a significant percentage of smallpox burials were found to be of young adults.

Landers (1993: 153-154) reports that in the latter part of the seventeenth century, 31% of smallpox deaths experienced by the London Quakers were of adolescents and individuals in their twenties. As native residents would have been exposed to the disease in childhood, this significant percentage of deaths from the disease in individuals aged 10 to 29 suggests that rural

164 immigrants were succumbing. In fact, when mortality levels in London are compared with national levels, little difference in the two exists above the age of 30, underscoring the importance of both epidemic disease and immigration in shaping the demography of the city

(Landers 1993: 160-161).

As immigrant mortality was such a significant component of the mortality of Old World cities, its potential contribution to the demographic regimes of New World cities should be considered. Clearly epidemic disease played a large role in the mortality experience of immigrants to London. However, as immigrants are generally difficult to identify in parish records and other documents used by historical demographers, determining whether immigrants experienced greater mortality than native residents apart from that which was caused by epidemic disease has not been possible in most preindustrial European cities. Immigrants to

Cholula would not have faced the threat of epidemic diseases, but they would have been exposed to infections that were endemic in the urban center. How would immigrants have fared in this

New World urban city?

In her study of mortality in Teotihuacan, Storey (1992), in fact, speculates that large numbers of young adults in the skeletal series might possibly be the result of immigration to the city. Based upon the age-at-death distribution for Cholula, however, mortality in young adults appears to be low (Figure 5-18). The estimate of the age-independent component of mortality for the Siler model is zero31, suggesting that mortality associated with young and old age is vastly more important in the Cholula sample. The age-at-death distribution, therefore, would

31 The baseline component of mortality in the Siler model is typically difficult to estimate even in modern populations with much larger sample sizes because age-independent mortality is low. The fact the α2 parameter is estimated as zero is a reflection of the fact that the low mortality in this age group and interference from the β1 and α3 parameters is affecting the ability of the mle program to capture this parameter value. Therefore, the α2 parameter value should not be taken as an indication that mortality in this age group is actually zero.

165 seem to indicate a potential deviation from the demographic experience of London, as there is no evidence of increased levels of young adult mortality in Cholula that might be attributable to appreciable numbers of immigrant deaths.

However, it should be emphasized that the low young adult mortality observed for

Cholula has several possible explanations. First, it might be a result of significantly lower levels of immigration to this New World city than was experienced in London. Given the very different economic system in place in Cholula, this explanation is quite possible. Moreover, as mortality would tend to be low in native resident of the city, it would probably require a considerably higher mortality rate in immigrants, combined with high rates of immigration, for the age-at-death distribution to be affected. Second, it could result from immigrants to Cholula coming from different demographic groups than was the case for preindustrial European cities.

Immigrants to London were typically young single adults, but this may not have been true in

Postclassic Cholula32. Finally, it might also be the result of immigrants having mortality rates similar to those of the native residents of Cholula. Let us return to Paine’s (1997) simulation of the effects of immigration on the age-at-death distribution of a population. When immigrants have mortality rates that are identical to those of the native population, the young adult portion of the age-at-death distribution is essentially unaffected by their presence in the population. As immigrants to Cholula would not have been confronted with the epidemic diseases which contributed so significantly to immigrant deaths in preindustrial Old World cities, they may not have experienced higher mortality than native residents upon arriving to the city.

Given the potential importance of immigration in shaping urban population dynamics, these issues will be considered in greater detail in Chapters VII and VIII. In Chapter VII, I will

32 There is, however, a very strong age-dependent pattern of migration across populations that will be discussed in Chapter VIII (Rogers and Castro 1984). Immigration does typically peak in young adulthood.

166 consider the demographic characteristics of immigrants to Cholula and their possible reasons for migrating. In order to further explore the impact of immigration of the paleodemographic age- at-death distribution of Cholula, strontium and oxygen isotope studies were undertaken to identify immigrants, determine the scale of immigration to the city, and compare the mortality experience of immigrants and native residents. The results of this study will be presented in

Chapter VIII.

Male and Female Mortality

Although most anthropological populations have a fairly balanced sex ratio, early paleodemographic studies often showed an overabundance of males in skeletal collections

(Weiss 1973). In the past, osteologists favored cranial features as a means of determining sex, but this practice can lead to classifying all but the most gracile of individuals as males. Although osteologists continue to consider cranial features and general robusticity when determining sex, in recent decades they have begun putting greater emphasis on characteristics of the pelvis, which has a higher degree of accuracy than sexing techniques based exclusively on cranial morphology. Some bias does still exist in sexing skeletons, particularly when poor preservation of pelvic bones necessitates a sex determination based solely upon the skull or overall robusticity of the individual. Younger males, who may be more gracile, can be misclassified as females, and older females, who may develop more masculine cranial features, may be misidentified as males

(Meindl et al. 1985b; Walker 1995). However, in general, a high degree of accuracy exists in determining sex in skeletal remains.

167 When sexing individuals in the Cholula collection, characteristics of the pelvis were prioritized over features of the skull or postcranial skeleton. The pelvis is extremely dimorphic in this population; therefore, there are likely to have been few errors in sexing when the pelvis was available. Few individuals were ambiguous with regards to sex. Males and females are almost equally represented in the sample.

In order to determine if differences existed in adult male and female mortality in Cholula, a Gompertz-Makeham mortality model, which is essentially the Siler model without the juvenile component of mortality, was fitted to the observed age-at-death distributions. Juveniles (those younger than 15) were excluded from this analysis as the skeletons of children cannot be reliably sexed. The age-at-death distributions for males and females are shown in Figure 5-28. The parameter estimates for the Gompertz-Makeham model are in Tables 5-5 and 5-6, and the hazards constructed from these estimates are shown in Figure 5-29. For both males and females, only the β3 parameter estimate was significantly different from zero.

14 Cholula Age-at-Death Distribution by Sex

6 Males Number 4 Females

0 15 16-20 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 71-75 75+ Age

Figure 5-28: Graph showing the Cholula age-at-death distribution by sex.

168

Table 5-5: Female Gompertz-Makeham parameter estimates. Loglikelihood = -322.4182, AIC =650.83643

Parameter Parameter Estimates Standard Error

α2 0.000 0.0118

α3 0.0058 0.0065

β3 0.0372 0.0113

Table 5-6: Male Gompertz-Makeham parameter estimates. Loglikelihood = -307.2023, AIC = 620.40468

Parameter Parameter Estimates Standard Error

α2 0.000 0.0187

α3 0.0070 0.0114

β3 0.0326 0.0160

169

Male and Female Hazards 0.3

0.25

0.2

0.15 h(a) Male Female 0.1

0.05

39 60 15 18 21 24 27 30 33 36 42 45 48 51 54 57 63 66 69 72 75 78 81 84 87 90 93 96 99 Age

Figure 5-29: Male and female hazards produced using Gompertz-Makeham estimates.

170 In the vast majority of historic and modern populations, males have greater mortality than females, particularly during early adulthood and senescence, as the result of both biological and cultural factors (Keyfitz and Flieger 1971). In Cholula, males do have excess mortality at advanced ages, which conforms to expectations. The Gompertz-Makeham curves for males and females appear to be practically identical at young adult ages. However, the α2 parameter is very difficult to estimate even with much larger sample sizes (Darryl Holman, personal communication). Consequently, the fact that the age-independent component of mortality appears to be equal in males and females in the fitted Gompertz-Makeham model is essentially meaningless. Particularly given the very small sample sizes in the young adult age range, the model is simply unable to capture any difference in mortality between the sexes that might exist.

A Word about Nonstationarity

An estimation of the growth rate of the Postclassic population of Cholula is important to the study of urban population dynamics not only because it permits the issue of nonstationarity to be addressed, but also because preindustrial cities have typically been characterized as having such high levels of mortality that deaths surpassed births, resulting in natural decrease (Wrigley

1969). Wood et al. (2002) have proposed an equation to derive the population growth rate at the same time that the age-at-death distribution of the population is estimated, following the tenets of the Rostock protocol. As mentioned, the sample size of the Cholula skeletal collection is likely insufficient to support the estimation of so many parameters, so the growth rate for the population could not be estimated in the current study. Future investigations aimed at calculating the growth rate for Cholula will be discussed further in Chapter IX.

171 Archaeological data can provide an alternative source of information on population growth rates, but no archaeological growth rate estimates exist for Cholula. However, archaeological data do suggest that the site was growing over the course of the Postclassic and had attained its maximum size by the time of the Spanish Conquest (Sanders 1971; Durmond and

Müller 1972; Müller 1973; Uruñuela and Plunket 2005: 313). While it is possible, therefore, that the so-called Law of Urban Natural Decrease did not apply to Postclassic Cholula, a positive growth rate, in and of itself, does not tell us about age-specific fertility or mortality rates nor does it tell us how the fertility or mortality of immigrants contributed to population dynamics in the city. Determining the scale of immigration, the demographic characteristics of immigrants, the conditions they might have faced in the city, and their mortality experience is a necessary step to sorting out urban population dynamics in Postclassic Cholula. In subsequent chapters, I will attempt to address some of these issues.

A Word about Fertility

Fertility was clearly a significant component of the demography of preindustrial Old

World cities. Galley (1998) indicated that periods of natural decrease in York were the result of decreases in fertility combined with increases in mortality. Finlay (1981) concluded that immigrant women in London had lower marital fertility rates than native residents, which contributed to the natural decrease he postulates for the city. Unfortunately, while fertility has a profound effect on age-at-death distributions, little useful information about fertility can be gleaned from paleodemographic samples (Milner et al. 2000), so it is impossible to address

172 whether similar patterns of fertility existed in Cholula. In Chapter VII, however, I will consider how immigrants might have contributed to fertility in the city.

In this chapter, we have seen some basic features of the paleodemography of Postclassic

Cholula. Notably, the age-at-death distribution constructed for Cholula indicates that the majority of individuals who survived into adulthood lived past the age of 50 and some even survived into old age. Poor preservation of infant remains precludes drawing any inferences about infant mortality, but it would appear that juvenile mortality drops off rather slowly.

Cholula may have experienced positive population growth based upon the limited archaeological data available, although, from the paleodemographic data alone, we can say little about how immigrants may have contributed to fertility and mortality in the city. This New World urban center does not seem to have experienced the elevated mortality rates in young adults found in preindustrial European cities such as London. As the high death rates of immigrants to Old

World cities are responsible for this observed demographic pattern, it is necessary to address in more detail the issue of immigration to the urban center of Cholula, which will be done in

Chapters VII and VIII. However, in the following chapter, I will first consider morbidity in the city and what we can discern from pathological lesions identified in the Cholula osteological sample.

173 CHAPTER VI

A PALEOPATHOLOGICAL ANALYSIS OF CHOLULA

“The timing of death, like the ending of a story, gives a changed meaning to what preceded it.” (Mary Catherine Bateson, With a Daughter’s Eye, 1994)

Paleopathology refers to the study of pathological lesions on skeletal remains, often with the intention of determining how healthy a population was. Although paleopathological lesions are one of the few means available to determine if the urban residents of Cholula were, in fact, subject to nutritional stress or a heavy disease burden, these skeletal lesions are difficult to interpret because mortality is selective and individuals are heterogeneous in their frailty, two issues which will be discussed in more detail below. A means of dealing with the issue of selectivity now exists, but the problem of hidden heterogeneity has yet to be overcome, although proposals have been made regarding potential ways of addressing this obstacle to paleopathological research (Boldsen 1991, 1997; Milner et al. 2000; Usher 2000; Boel 2001).

Unfortunately, forming conclusions about the health of a population based on pathological lesions observed in skeletons is still not possible; however, a first step towards this goal is to determine how pathological lesions influenced mortality in a given population.

Studies that have attempted to determine how skeletal lesions influence the risk of death have shown that the relationship between particular pathological lesions and mortality may be population specific (Usher 2000; Ferrell 2002; Dewitte 2006). Thus, it is imperative that the exact nature of this relationship be made explicit for every skeletal collection under study. It is in this spirit that the interaction between mortality and certain pathological lesions is being investigated in the Cholula skeletal sample. While understanding how these lesions are related

174 to mortality will not permit us to address how urbanism affected the health of the residents of

Cholula at this point in time, it is a positive step in that direction. Therefore, in the following chapter, I will examine how porotic hyperostosis and cribra orbitalia, enamel hypoplasias, and proliferative lesions of the skeleton influenced the risk of dying for the residents of Postclassic

Cholula.

Inherent Problems of Interpreting Pathological Lesions in Skeletal Remains

Sensitivity and specificity

In diagnosing disease, epidemiologists consider several measures of the accuracy of the diagnostic criteria being used, two of which are the sensitivity and the specificity of the indicator under consideration. The term sensitivity refers to the probability that a particular symptom will be observed or a test result will be positive if the disease in question is, in fact, present. The term specificity refers to the probability that a symptom will be absent or a test result will be negative if the disease in question is absent (Grobbee and Hoes 2009). In paleopathological studies, therefore, the sensitivity of a pathological lesion is the probability that an individual suffering from the particular disease under study displays a particular skeletal pathology, while the specificity of a pathological lesion is the probability that a particular skeletal pathology is absent in an individual who is not suffering from the disease under study.

In general, pathological lesions of the skeleton display low sensitivity and low specificity, making them poor diagnostic indicators. Skeletal lesions take time to form. Consequently, pathological lesions of the skeleton are insensitive indicators of disease because they develop in only a small percentage of cases, typically those that are chronic or advanced (Aufderheide and

175 Rodrigeuz-Martin 1998). In fact, most diseases do not affect the skeleton at all. In addition, the skeleton can respond to disease and malnutrition in only a limited number of ways, namely through the resorption or proliferation of bone. As a result, skeletal lesions are nonspecific indicators of disease because many maladies cause similar pathological lesions (Ortner 2003).

Much of the time, it is not possible to make a diagnosis based upon observed pathological lesions, although in some cases, the pattern of bones affected, combined with information about the type of lesions present (proliferative versus osteolytic), indicate a particular disease process at work.

The paradoxical nature of skeletal lesions

In 1992, Wood et al. published an article entitled ―The Osteological Paradox,‖ in which they discussed how the problems of hidden heterogeneity and selective mortality can result in equivocal interpretations of pathological lesions. Although other researchers (Ortner 1991; Cook and Buikstra 1979; Goodman and Armelagos 1988; Stuart-MacAdam 1988) had pointed out that skeletal lesions take time to form and that they, therefore, represent survival of a disease process, at least for a time, the implications of this realization had not been made explicit in the literature.

Wood et al. argued that as a result of the related issues of hidden heterogeneity and selective mortality, skeletal lesions bear a complex relationship to health.

176 Hidden heterogeneity

Individuals vary in their underlying frailty as a result of genetic, developmental, environmental, and socioeconomic factors (Vaupel and Yashin 1985; Vaupel et al. 1979; Alter and Riley 1989). Frailty in this sense refers to biological characteristics of individuals that are associated with their relative risk of falling ill or dying (Vaupel 1990). The fact that individuals vary in terms of frailty should be readily apparent if we consider modern populations (Weiss

1990). For example, the lowest socioeconomic classes in modern societies typically have higher rates of morbidity and mortality than the most privileged groups. While some variation in frailty is attributable to factors such as age and sex that are observable, other differences in frailty in a population result from less obvious and less measurable causes, such as genetic variation in susceptibility to diseases, differences in nutrition, or even differential exposure to environmental stresses. These unseen differences in the risk of death are what is referred to as hidden heterogeneity. When heterogeneity in frailty is hidden, it is exceedingly difficult to control for it in paleopathological analyses of skeletal remains. However, failure to take into account hidden heterogeneity in susceptibility to disease or death may result in erroneous interpretations of skeletal lesions, as will be illustrated in an example below.

Selective mortality

As a result of heterogeneity in frailty, both seen and unseen, mortality is selective, meaning that the frailest individuals in a population have the greatest risk of death (Wood et al.

1992). In other words, a skeletal sample is not a random sample of the living population from

177 which it was drawn. Within any birth cohort, those individuals with higher frailty are more likely to die and enter the mortality sample than others of their sex and age. Furthermore, those individuals in a population with higher frailty are likely to be selected out of the population at younger ages than those with lower frailty.

As a result of selective mortality, pathological lesions in skeletal remains cannot necessarily be interpreted in a straightforward manner. Because skeletal lesions take time to form, frailer individuals may die before infection or malnutrition can leave their marks.

Individuals must have at least some degree of immunological or genetic resistance to these insults in order to survive long enough for paleopathological indicators of stress to form.

However, an individual who develops a pathological lesion clearly has a different health status than someone who was immunologically able to avoid or suppress the insult before it resulted in skeletal involvement, or someone who was not exposed to the stress at all (Wood et al. 1992).

Prior to Wood et al.’s article, most paleopathological analyses had assumed that skeletal populations with higher frequencies of pathological lesions were less healthy than those that had lower frequencies of such lesions. However, the Osteological Paradox suggests that it might also be the case that populations with higher frequencies of lesions consisted of somewhat healthier individuals who lived at least long enough for the skeletal lesion to form. Populations with lower frequencies of lesions could consist of particularly healthy individuals that were able to avoid or fight off the disease before it could impact the skeleton, or they could consist of particularly unhealthy individuals that died before the disease could result in skeletal involvement.

To illustrate how hidden heterogeneity and selective mortality can confound the interpretation of skeletal lesions, Wood et al. (1992) provide an example of three subgroups in a

178 population exposed to a disease process. The first subgroup consists of particularly frail individuals who die of the condition before skeletal lesions can form; the second subgroup consists of somewhat healthier individuals who are able to live at least long enough for bony lesions to develop; and the third subgroup consists of individuals who have no skeletal lesions because they did not contract the disease. If these individuals are all buried in the same cemetery, it appears that there are only two groups: a healthy group that does not have skeletal lesions, and an unhealthy group that does. Therefore, as Wood et al. point out, better health may result in worse skeletons.

To further complicate matters, pathological lesions of the skeleton can be classified as active, meaning that the disease process was ongoing at the time of death, or healed, indicating that the individual suffered from the disease at some point during life but that the disease process had resolved itself prior to death. Individuals with healed lesions would seem to be healthier or less frail than those individuals who died with active skeletal lesions. However, as Alter and

Riley (1989) indicate, repeated insults to health may increase an individual’s frailty and have long term effects on the risk of death. Thus, although the individual with healed skeletal lesions survived that particular insult, his or her overall health may have been negatively impacted.

While the presence of pathological skeletal lesions, for the most part, does indicate something about the health of the individual, the message being conveyed is quite complicated.

Without some way to model the distribution of frailty in the living population or to determine how the presence of skeletal lesions relates to mortality, the frequency of pathological lesions in a population is essentially uninterpretable.

179 Responses to the Osteological Paradox

While some paleopathologists have responded favorably to the fact the Wood et al. brought into the open paleopathology’s dirty little secret (Ortner 1991), others (Goodman 1993;

Cohen 1994, 1997) have questioned the validity of the Osteological Paradox. In part, this controversy stems from the fact that Wood et al. suggested that the many paleopathological studies cited in support of the argument that the transition to agriculture led to a decline in health could just as easily be interpreted to the contrary, as the data are ambiguous. Cohen (1994,

1997) and Goodman (1993), who are advocates of the theory that cultural evolution led to a decline in health, have been adamant that Wood et al. overstated their case.

Cohen (1994, 1997) in particular has objected to the theoretical premise presented by

Wood et al. (1992) by stating that most deaths in past societies were the result of random chance and, thus, not influenced by selective mortality or hidden heterogeneity. Therefore, on average,

Cohen argues, skeletal samples should be representative of the populations from which they came. However, Cohen’s claims are problematic in two regards. First, his assertion that most deaths in traditional populations were accidental is simply untrue. Most deaths in the past were due to infectious diseases, which could potentially result in paradoxical skeletal lesions (Wood and Milner 1994; Buikstra 1997). Second, hidden heterogeneity and selective mortality can affect even accidental deaths. Milner et al. (1991) found a significant number of deaths in the

Norris Farm population due to violence inflicted in small-scale raids by enemy groups. Many of the individuals who died violent deaths had a pre-existing condition or disability that would have made it difficult for them to escape. Hence, their deaths were influenced by selective mortality.

180 Both Cohen (1994, 1997) and Goodman (1993) have also suggested that osteologists are interested in the quality of life of past populations and not just mortality. They argue that skeletal lesions indicate that a population was under stress, but they do not necessarily have any influence on the risk of death. But as Wood and Milner (1994) have pointed out, if a skeletal lesion has absolutely no effect on mortality, how can it be reflecting information about health?

In a recent edited volume, Steckel and Rose (2002) suggest that Wood et al. have failed to demonstrate that selective mortality causes a significant bias in the formation of skeletal samples. Instead of addressing the concerns raised by the Osteological Paradox, Steckel and

Rose (2002) have attempted to standardize the evaluation of health in past societies by introducing a ―health index,‖ which permits investigators studying health in different populations to apply the same criteria in assessing morbidity and mortality33. They contend that the methodology of using multiple skeletal lesions to assess health transcends the potential biases outlined in the Osteological Paradox.

The volume presents paleopathological and paleodemographic studies of societies in the

Western hemisphere over a 12,000 year time span. While many of the populations that had the lowest scores on the health index were agricultural populations, there was, in fact, a great deal of variability in the results of the investigations and some populations scored contrary to the expectations of the authors. One of the studies (Higgins et al. 2002) presented in the volume strongly suggests that the health index is not a solution to the problems of hidden heterogeneity and selective mortality. A collection of skeletons from a poorhouse scored very high on the index, and it is suggested that this may be a case in which the Osteological Paradox is at work.

However, the fact that this population scored high on the health index clearly indicates the index

33 Cholula is, in fact, one of the populations studied using the health index. Readers should refer to Márquez et al. (2002).

181 itself does nothing to address the methodological problems raised by Wood et al. (1992). While it is certainly a worthwhile endeavor to standardize the the study of skeletal remains to make results comparable across populations, other methodological approaches must be used to address selective mortality and hidden heterogeneity directly.

Proposed Solutions to the Osteological Paradox

In order to more directly address the concerns associated with the Osteological Paradox, a number of individuals including Wood et al. (1992), Boldsen (1991, 1997), and Boel (2001) have looked at the relationship between skeletal lesions and mortality to determine whether selectivity is at work. Wood et al. (1992) demonstrated that individuals from an Illinois skeletal collection with porotic hyperostosis and cribra orbitalia died at earlier ages than those without the condition. Boldsen (1991) established that heavy tooth wear was related to an elevated risk of death in the Tirup skeletal collection. Boldsen (1997) also examined the age-specific pattern of caries in males and females from this population. He found that in males, the frequency of carious lesions was fairly constant throughout adulthood, while in females, caries were much more frequent among women in their mid to late reproductive years. He suggests that the effects of childbearing and breastfeeding increased calcium depletion, resulting in more caries, and that the women who survived into old age were those with fewer carious lesions because these individuals had fewer children or perhaps a better diet. Boel (2001), also studying the Tirup population, found that in females, bone mineral density increased in individuals living into old age, indicating that these individuals were less frail than those dying at earlier ages.

182 Bethany Usher (2000) has presented a multistate mortality model to address the issues of hidden heterogeneity, selective mortality, and nonstationarity. As skeletal collections are typically biased samples, it is impossible to extrapolate the frequency of pathological lesions in the living population directly from their frequency in the skeletons under study; therefore, paleodemographers need a means of ―working backwards‖ from the biased skeletal sample to reconstruct the living population that produced it. The Usher model is just such a tool, and, thereby provides a means of making inferences about the frequency of lesions in the living population, the distribution of frailty with respect to age, and the relationship between pathological lesions and the risk of death.

Figure 6-1: Usher’s multi-state model of mortality (Usher 2000)

The full model consists of 4 states: well, ill, healed, and dead, as shown in Figure 6-1.

―Well‖ refers to having no pathological lesions on the skeleton, ―ill‖ refers to having active lesions, and ―healed‖ refers to having healed pathological lesions. Of course, as many diseases do not leave lesions on the skeleton, the model cannot test the effects of those illnesses. Rather, the model is evaluating the effects of lesions that can be observed by paleodemographers.

183 Newborns enter the population in the ―well‖ state through a renewal process that takes into account the growth rate, thereby addressing the issue of nonstationarity. Each newborn is randomly assigned a frailty level (z) from a distribution of frailty. As they get older, individuals can either move into the ―ill‖ state or they can die without ever becoming ―ill.‖ If they become

―ill,‖ individuals can die from that state, or they can move into the ―healed‖ state and eventually die from there. Transitions between the living states are dependent upon the individual’s frailty, as are the hazard rates for moving from a living state to the dead state. The hazards of passing from the ill and healed states to the dead state are modeled as being proportional to the hazard of moving from the well to the dead state. This allows the issue of selective mortality to be addressed. Obviously, individuals are observed by the paleodemographer only once they have entered the ―dead‖ state, so the age at death of each individual, as well as the distribution of active and healed lesions in the skeletal sample provide the information necessary to determine the parameters of the model using maximum likelihood estimation.

Unfortunately, the full four-state model is very complicated and includes a significant number of parameters that must be estimated. Consequently, it has not yet been fully tested.

However, Usher (2000) did implement a reduced stationary three-state version of the model and was able to detect the impact of selective mortality on several different types of skeletal lesions.

The reduced model includes only the ―well,‖ ―ill,‖ and ―dead‖ states. No distinction is made between active and healed lesions. Usher used the Siler model as the hazard of dying from the

―well‖ state. The hazard of becoming ―ill‖ at some age a was modeled as a constant, k1, meaning that the risk of developing a lesion was assumed to be constant across all ages. This assumption is simplistic, and its implications will be discussed later in the chapter. The hazard of dying once ―ill‖(k2) was modeled as proportional to the baseline risk of death from the ―well‖

184 state, so the risk of death is either increased or decreased, depending upon the effect of the pathological lesion. The likelihood of dying with a skeletal lesion is, therefore, defined by the equation below, whose derivation can be found in Usher (2000: 24-29):

( ) d

where the subscripts refer to the transition between stages and a refers to age.

The parameters of the Siler model, as well as the k1 and k2 constants are estimated using a maximum likelihood estimation program (mle). A value of k2 significantly greater than one indicates that the presence of a pathological lesion increases the risk of death. A value of k2 significantly less than one indicates that the pathological lesion reduces the risk of death, and a k2 value of one indicates the lesion has no effect on mortality. In tests of the reduced model using simulated data, Usher found that k1 and k2 values were captured by the maximum likelihood estimation program, but the Siler parameter values, when part of this more complicated model, were not as well-estimated. For this reason, the Siler model was fit as a separate exercise in the previous chapter, and only the k1 and k2 values will be considered in determining the effects of various pathological lesions on mortality in the Cholula population.

Although more work must be done to resolve this issue completely, some progress has been made. Usher’s approach is one of the most promising available at this time for analyzing osteological data from the Cholula skeletal collection.

185 Pathological Lesions in the Cholula Assemblage

A number of pathological lesions were recorded in the Cholula collection including porotic hyperostosis and cribra orbitalia, enamel hypoplasias, and proliferative lesions. These pathological lesions were chosen, in part, because they are among the most commonly recorded lesions in skeletal collections, and they occur with sufficient frequency in past populations to ensure enough observations of the condition to determine how their presence affects mortality.

For all lesions considered, no distinction was made with regard to sex. Rather, adult males and females were combined into one sample. This was done because these pathologies are present in children as well as adults, yet the sex of children cannot be reliably determined from skeletal remains. Eliminating children and dividing males and females would significantly reduce the sample size. However, it should be remembered that pathological lesions may differentially affect the sexes, a possibility which should be addressed if sample sizes permit. Furthermore, no distinction was made between healed and unhealed lesions, although survival of the stress episode, as indicated by healing of the lesion, is likely to tell us something about selective mortality.

Porotic hyperostosis and cribra orbitalia

Porotic hyperostosis is characterized by porous lesions on the frontal, parietal, and occipital bones of the skull. The underlying condition that produces the lesions frequently produces similar porosities on the superior parts of the orbits, a condition referred to as cribra orbitalia (see Figures 6-2, 6-3, and 6-4). Stuart-Macadam (1987) conducted a radiograph study

186 of living populations, in which she linked porotic hyperostosis and cribra orbitalia to anemia. In response to the anemia, the body attempts to produce more red blood cells in the cranial diploë.

The diploë expands and puts pressure on the outer table of the skull, causing it to thin and resulting in the porous appearance (Wright and Chew 1999: 925). As red marrow is not present in the cranial bones of adults, porotic lesions as an active response to anemia are limited to juveniles, although healed lesions may be present on adult skeletons (Stuart-Macadam 1985).

Evidence of healing or remodeling of the bone indicates that the individual survived the stress episode.

In New World skeletons, iron-deficiency anemia, as opposed to genetic anemias, have traditionally been assumed to be the cause of porotic hyperostosis and cribra orbitalia, as genetic anemias are not typically found in Native American populations. The iron-deficiency anemia that has been cited as the cause of these skeletal lesions is thought to be the result of several factors, most likely working in tandem, including malnutrition, gastrointestinal infections that inhibit iron absorption in the intestines, and parasites that both inhibit iron absorption and result in intestinal bleeding (El-Najjar et al. 1976; Stuart-Macadam 1987; Mensforth et al. 1978; Holland and O’Brien 1997; Palkovich 1987: 528-529).

Figure 6-2: Active cribra orbitalia on the superior orbits of a juvenile from Cholula. Photo taken with permission from the DAF of the INAH.

187

Figure 6-3: Active porotic hyperostosis on the cranium of a juvenile from Cholula. Photo taken with permission from the DAF of the INAH.

188

Figure 6-4: Porotic hyperostosis on the occipital and parietal bones of a juvenile from Cholula. Photo taken with permission from the DAF of the INAH.

A more recent study suggests that while genetic anemias may, in fact, be responsible for porotic hyperostosis, iron-deficiency anemia is an unlikely culprit (Walker et al. 2009). Instead, this investigation posits that megaloblastic anemia, caused by vitamin B12 deficiencies, is responsible for most cases of porotic hyperostosis observed in skeletal material. As B12 is primarily found in foods of animal origin, a diet with little meat consumption, particularly if combined with intestinal parasites, could result in such deficiencies. Furthermore, the authors of this study argue that cribra orbitalia reflects both B12 deficiencies and scurvy.

For the purposes of analyzing lesions of porotic hyperostosis, an attempt at diagnosing the underlying condition was not made because of the tentativeness of such diagnoses and the fact that nutritional deficiencies do not typically occur in isolation. Rather, if an individual is

189 deficient in one nutrient, they are likely to be deficient in others as well, so it is possible for B12 deficiencies, scurvy, and iron-deficiency anemia to all be present. Therefore, all cranial lesions involving significant porosity were included in the following analysis.

Regardless of the cause, Stuart-Macadam (1988) has suggested that porotic hyperostosis may, in fact, indicate a healthy response to nutritional deficiencies. In other words, it is possible that sicker individuals would die before their bodies could mount such a defense. However, in her analysis of the Tirup skeletal sample, Usher (2002) found that individuals with cribra orbitalia in the medieval Danish village were almost five times more likely to die than those without the lesions. Wood et al. (1992) and Dewitte (2006) report similar findings in regards to the relationship between porotic hyperostosis and mortality. These data, therefore, support the idea that this condition reflects poor health, but this connection must also be established in the

Cholula skeletal sample.

The state of preservation of the orbits, frontal squama, parietal, and occipital bones in the

Cholula collection was assessed following Buikstra and Ubelaker (1994) (see Appendix H for the presence or absence of this pathology in each skeleton). Lesions were recorded as present or absent, and, if present, as active or inactive. In addition, the severity of lesions was scored as described by Usher et al. (2000:143). As cribra orbitalia is a bilateral condition, good preservation (75-100% complete) of at least one of the orbits was sufficient for inclusion in the analysis. There were no cases in which both orbits were well preserved that cribra orbitalia was recorded on one side, but not the other. Healed and unhealed lesions were combined34. Table 6-

1 indicates the number of cases of this pathology observed in the Cholula collection. While the graph (Figure 6-5) suggests that individuals with this condition died at younger ages than those

34 As cribra orbitalia and porotic hyperostosis form in childhood, children included in the analysis may have healed or unhealed lesions, but all adults have healed lesions.

190 without the lesions, the maximum likelihood analysis (Table 6-2) indicated no increase in mortality from cribra orbitalia (p>0.05), likely because of the limited number of cases observed.

The high standard error would also suggest that the small percentage of individuals with cribra orbitalia caused statistically insignificant results.

Table 6-1: The number of cases of cribra orbitalia in the Cholula collection.

Present Absent No information

Orbits 12 113 154

Cribra Orbitalia

50 45 40 35 30 25

Number 20 Absent 15 10 Present 5 0

Age

Figure 6-5: Number of individuals in each age group with lesions of cribra orbitalia.

191

Table 6-2: Parameter values for cribra orbitalia. LL= -526.3233, AIC= 1066.6466

Parameter Parameter estimate Standard error

α1 0.1437 0.0260

β1 0.1942 0.0620

α2 0.0000 0.0130

α3 0.0021 0.0033

β3 0.0605 0.0217

k1 0.0040 0.0030

k2 3.2901 3.6820

The effects of porotic hyperostosis on the risk of death were assessed for the parietal and occipital bones35 separately, and only bones with good preservation (75-100% complete) were included. As porotic hyperostosis is also a bilateral condition, either the left or right parietal was included depending upon which was better preserved. There were no cases in which both parietals were relatively complete that porotic hyperostosis was observed on one side and not the other. The decision to treat the bones separately was simply a means to increase the number of observations that could be included in the analysis (see Table 6-3 for the sample size of each analysis and the number of observed cases of porotic hyperostosis on each bone) and was not motivated by any theoretical assumptions that the presence of porotic hyperostosis on one cranial bone might affect mortality differently than its presence on another bone. However, it is interesting to note that porotic hyperostosis is much less frequently observed on the frontal squama in the Cholula population than on the occipital or the parietal, and the individuals with

35 The frontal bone was not included in the statistical analysis as so few individuals (only three) had frontal lesions.

192 frontal involvement tend to have more severe lesions on other parts of the cranium as well. This perhaps indicates that frontal involvement occurs only after the nutritional deficiency has been present for some time, which could, indeed, have implications for the relationship between the location of the lesion and what the pathology indicates about mortality. Unfortunately, so few individuals were observed with frontal lesions that statistical analysis of this bone was not possible, so this hypothesis could not be tested. Alternatively, it may be the case that frontal lesions are caused by a different nutritional deficiency than lesions on the parietal and occipital, perhaps scurvy. This is, in fact, suspected for at least one of the individuals with frontal lesions, as sphenoidal and zygomatic involvement is also present (Ortner 1999, 2003).

The presence of porotic hyperostosis does appear to increase the risk of death in the

Cholula population. Based on solely the graphic distribution of the prevalence of the pathological lesion by age (Figures 6-6 and 6-7), we can see that the pathology is much more common in younger individuals than in older individuals. While the results of the likelihood analysis (Tables 6-4 and 6-5) appear to concur that porotic hyperostosis increases the risk of death tenfold or more, the results were not statistically significant for either the parietal or the occipital (p>0.05), possibly due to fairly limited number of observations of the pathology.

Table 6-3: Cases of porotic hyperostosis.

Present Absent No information

Parietal 17 164 98

Occipital 11 143 129

193 Porotic Hyperostosis on the Parietal 80

Number 30 Absent

20 Present

Age

Figure 6-6: Number of individuals in each age group with lesions of porotic hyperostosis on the parietal.

Porotic Hyperostosis on the Occipital 60

Number Absent 20 Present 10

Age

Figure 6-7: Number of individuals in each age group with porotic hyperostosis on the occipital.

194

Table 6-4: Parameter values for the parietal. LL= -789.5922, AIC = 1593.1843

Parameter Parameter estimate Standard error

α1 0.1144 0.0189

β1 0.1772 0.0492

α2 0.0000 0.0776

α3 0.0021 0.0019

β3 0.0538 0.0010

k1 0.0035 0.0010

k2 9.9999 6.5231

Table 6-5: Parameter values for the occipital. LL= -663.4312, AIC = 1340.8625

Parameter Parameter estimate Standard error

α1 0.1070 0.0226

β1 0.2338 0.0753

α2 0.0001 0.0109

α3 0.0044 0.0046

β3 0.0422 0.0110

k1 0.0025 0.0009

k2 9.9999 10.0446

195 Enamel hypoplasias

Enamel hypoplasias are pathological lesions that appear on the teeth as transverse lines or grooves, primarily on the buccal surfaces (Figure 6-8). These grooves form during the development of the teeth, when some sort of stress, be it nutritional deficiencies or disease, interrupts enamel formation for an extended period of time (a few weeks to two months)

(Roberts and Manchester 1997: 58-59; Skinner and Goodman 1992; Goodman et al. 1984). As these lesions form during the process of tooth development, they are indicative of childhood illnesses or malnutrition. As with porotic hyperostosis and cribra orbitalia, enamel hypoplasias may be observed in adults, but they do not occur after the process of crown formation has ended.

Furthermore, enamel hypoplasias always represent survival of a stress episode, as the grooves result from enamel formation being interrupted and then beginning anew. Unlike with most other pathological lesions, the age at which enamel hypoplasias form can be determined. As teeth develop at fairly set rates across populations, the location of the enamel hypoplasia on the tooth crown can be used to estimate the age of the individual at the time that enamel formation was interrupted.

196

Figure 6-8: Enamel hypoplasias present on the mandibular canines of an adult male from Cholula. Photo taken with permission from the DAF of the INAH.

Four permanent teeth were included in the present study: the maxillary first incisor, the mandibular canine, and the first and second mandibular molars36. In each instance, the left tooth was scored, unless it was not present, in which case the right tooth was used (see Appendix I for a list of the presence or absence of enamel hypoplasias in each skeleton). Deciduous teeth were not included because few had enamel hypoplasias. This finding is fairly typical as much of the deciduous dentition forms prior to birth when the fetus is buffered against stress by the mother

(Goodman et al 1984: 26). Enamel hypoplasias were identified based upon a macroscopic examination of the tooth and/or by running a dental tool over the surface of the tooth to determine if a notable indentation was present (following Usher 2000: 179). Each tooth was scored as absent, present with no enamel hypoplasias, or present with at least one enamel hypoplasia (Usher 2000: 69-70). For each existing enamel hypoplasia, the distance from the cemento-enamel junction to the most occlusal aspect of the defect was also measured in order to

36 The mandibular third molar was also scored, but it was not included in the statistical analysis due to the smaller number of observations.

197 determine the age at which the defect occurred (following Goodman and Rose 1990). These measurements will be used in future studies of the Cholula population, as will be discussed later.

For the present analysis, individuals were identified as ―sick‖ if at least one enamel hypoplasia was present on the particular tooth being analyzed (Table 6-6).

Tooth wear and tooth loss are concerns in the Cholula population, particularly with older individuals. Requiring that 100% of the tooth be present for inclusion in the analysis resulted in small sample sizes consisting exclusively of juveniles, which would have biased the results. As enamel hypoplasias do not typically occur in cuspal enamel (Goodman and Armelagos 1985), a decision was made to include in the current study those teeth in which at least two-thirds of the crown was present in order to maximize the sample size37. As can be seen in the graphs (Figures

6-9, 6-10, 6-11, and 6-12) indicating the distributions of enamel hypoplasias by age, tooth wear and tooth loss resulted in fewer observations of older individuals, even allowing for the loss of up to one-third of the tooth. The extent of tooth loss and tooth wear does not become significant until around age 50, however, so the effects of hypoplasias on the risk of death should still be discernable. Analyses were also run requiring that 75% of the tooth be present and that only

50% of the tooth be present, and results similar to those discussed below were obtained.

Previous studies examining the relationship between enamel hypoplasias and mortality have resulted in inconsistent findings. While Goodman (1996) established that the existence of enamel hypoplasias on the maxillary central incisors, the mandibular canines, or the maxillary first molars was linked to an increase in mortality, Usher (2000) found that the presence of enamel hypoplasias on only the first molar was associated with an elevated risk of death. She, in

37 The amount of the tooth crown present was determined by scoring each tooth as being complete or as showing wear and by measuring each tooth crown of each individual. The heights of those tooth crowns that showed no wear were averaged for each tooth type. Individuals who had a crown height that was at least two-thirds of this average were included in the present study.

198 fact, reports that hypoplasias on the incisors were linked to a reduction in mortality in the Tirup population. These conflicting results may indicate that populations differ in their response to stress because of genetic or environmental variability. Consequently, examining the relationship between the presence of particular pathological lesions and the risk of death in each population is a necessary step in understanding health in past populations.

The results of the likelihood analyses for the maxillary incisor, the mandibular canine, and the first and second mandibular molars are presented below (Tables 6-7, 6-8, 6-9, and 6-10).

The results indicate that enamel hypoplasias on the incisors, and the mandibular first and second molars are associated with a statistically-significant increase in the risk of death (p< 0.05). The k2 constant for the mandibular canine indicates that there was no statistically-significant change in the risk of death associated with the presence of enamel hypoplasias (p>0.05). As the sample size for the canine was similar to that of the others, it does not appear that an insufficient sample size or a limited number of observations of the pathological lesion is to blame for the lack of an effect on mortality. Referring to the distribution of canine enamel hypoplasias by age (Figure 6-

10), the pathology occurs at a fairly consistent frequency in all age groups, supporting the results of the likelihood analysis that indicate that presence of enamel hypoplasias on the canine has no effect on mortality. In fact, very few individuals in the population did not have at least one enamel hypoplasia on the canine. Interestingly, Usher (2000) also found that enamel hypoplasias on the canine had no effect on the risk of death in the Tirup population.

199

Table 6-6: Cases of enamel hypoplasias.

Tooth Present Absent No information

Maxillary incisor 58 50 172

Mandibular canine 94 15 170

Mandibular first molar 44 76 160

Mandibular second molar 41 66 172

Maxillary Incisor 18 16 14 12 10 8 Number Absent 6 Present 4 2 0

Age

Figure 6-9: Number of individuals by age group with at least one enamel hypoplasia on the maxillary incisor.

200 Mandibular Canine 20 18 16 14 12 10

Number 8 Absent 6 Present 4 2 0

Age

Figure 6-10: Number of individuals in each age category with at least one enamel hypoplasia on the mandibular canine.

Mandibular First Molar 20 18 16 14 12 10

Number 8 Absent 6 Present 4 2 0

Age

Figure 6-11: Number of individuals in each age category with at least one enamel hypoplasia on the mandibular first molar.

201

Mandibular Second Molar 20 18 16 14 12 10

Number 8 Absent 6 Present 4 2 0

Age

Figure 6-12: Number of individuals in each age category with at least one enamel hypoplasia on the mandibular second molar.

Table 6-7: Parameter values for maxillary incisor. LL = -540.9260, AIC = 1095.8521

Parameter Parameter estimate Standard error

α1 0.0095 0.5436

β1 0.0020 0.0930

α2 0.0000 0.5434

α3 0.0008 0.0039

β3 0.0541 0.0456

k1 0.0175 0.0029

k2 9.9999 4.5807

202

Table 6-8: Parameter values for mandibular canine. LL = -518.4486, AIC = 1050.8971

Parameter Parameter estimate Standard error

α1 0.0083 0.0068

β1 0.0866 0.0899

α2 0.0000 0.0032

α3 0.0002 0.0004

β3 0.0700 0.0188

k1 0.0346 0.0067

k2 8.8819 5.2601

Table 6-9: Parameter values for mandibular first molar. LL = -602.7428, AIC = 1219.4856

Parameter Parameter estimate Standard error

α1 0.0183 0.0479

β1 0.0505 0.1771

α2 0.0000 0.0598

α3 0.0026 0.0105

β3 0.0477 0.0374

k1 0.0117 0.0022

k2 7.3197 3.3072

203

Table 6-10: Parameter values for mandibular second molar. LL = -537.6822, AIC = 1089.3644

Parameter Parameter estimate Standard error

α1 0.0000 0.2583

β1 0.2603 63432.3473

α2 0.0000 0.0082

α3 0.0042 0.0043

β3 0.0405 0.0104

k1 0.0098 0.0012

k2 9.9999 4.9857

Proliferative lesions

Skeletal indicators of nonspecific infections were also assessed. As the skeleton can only respond to stress in a limited number of ways, namely through the proliferation or resorption of bone, diagnosing particular infectious diseases from skeletal remains is often impossible.

Periostitis and osteomyelitis are general terms that refer to proliferative changes in bone caused by systemic or localized infections (Figure 6-13). Periostitis occurs when a bacterial infection or traumatic injury causes an inflammatory response in the periosteum, the thin membrane covering the bone. The inflammation stimulates osteoblasts to lay down new bone, which is porous in appearance. Localized infections may result in unilateral skeletal involvement, while systemic infections tend to cause bilateral proliferative lesions. Osteomyelitis is a more severe form of

204 bone infection in which bacteria enter the medullary cavity, again as a result of either a systemic infection or direct injury to the bone. Osteomyelitis can be differentiated from periostitis by one or more of the following characteristics: cloacae (which allow for the discharge of pus), involucrum, expansion of the marrow cavity, and general distortion of shape of the bone (Ortner and Putschar 1981). Unlike the other pathological lesions considered here, proliferative lesions may occur at any time during life. Healing of the lesions indicates that the individual survived the disease episode. No distinction was made between periostitis and osteomyelitis in the current investigation. Instead, all proliferative lesions were grouped together.

Figure 6-13: Examples of proliferative lesions in three individuals from Cholula. Photos taken with permission from the DAF of the INAH.

205 Macroscopic examination of the skeleton was used to assess the presence or absence of skeletal lesions associated with infections (Table 6-11). The degree of preservation of all the bones of the skeleton was recorded following Buikstra and Ubelaker (1994). Proliferative lesions were scored for presence or absence according to Usher (2000:143). For the purposes of the likelihood analysis, no distinction was made between active and healed lesions, nor was the severity of the lesion considered; however, these data were collected and will be considered in future studies. The relationship between proliferative lesions and risk of death was assessed for all long bones. To be included in the analysis, the diaphysis of the bone had to be well-preserved

(75%-100%). Other bones were not included in the analysis of proliferative lesions due to the relatively small number of observations of this pathology elsewhere on the skeleton.

The distributions of proliferative lesions by age for each long bone are presented in

Figures 6-14, 6-15, 6-16, 6-17, 6-18, and 6-19, and the results of the likelihood analyses are shown in Tables 6-12, 6-13, 6-14, 6-15, 6-16, and 6-17 (see Appendix J for a list of the presence or absence of proliferative lesions for each skeleton). These findings indicate that proliferative lesions on the ulna, femur, and tibia are associated with an increased risk of death (p<0.05).

Similar results were reported by Usher (2000) in her analysis of proliferative lesions of the femur. The estimates of the k2 parameter for the humerus, the radius, and the fibula do not indicate a significant increase in the risk of death as a consequence of proliferative lesions on these bones (p>0.05). The fact that very few proliferative lesions were observed on the humerus and the radius likely explains why the maximum likelihood analyses fail to show a change in mortality associated with this pathology: There are simply not enough data to discern how mortality is affected. The distribution of the pathology by age (Figure 6-14) would seem to suggest that individuals with proliferative lesions on the humerus did die at younger ages;

206 however, the age distribution of proliferative lesions on the radius is more ambiguous (Figure 6-

15). The maximum likelihood results for the fibula appear to be reflecting a genuine phenomenon, as a fair number of observations were recorded for this bone (Table 6-11 and

Figure 6-19). Proliferative lesions on the fibula apparently have little effect on mortality.

Table 6-11 : Cases of proliferative lesions on the long bones.

Bone Present Absent No information

Humerus 7 152 120

Radius 7 135 137

Ulna 22 129 128

Femur 30 124 126

Tibia 76 97 105

Fibula 48 109 121

207 Proliferative Lesions of the Humerus 50 45 40 35 30 25

Number 20 Absent 15 Present 10 5 0

Age

Figure 6-14: Number of individuals in each age category with proliferative lesions on the humerus.

Proliferative Lesions of the Radius 50 45 40 35 30 25

Number 20 Absent 15 Present 10 5 0

Age

Figure 6-15: Number of individuals in each age category with proliferative lesions of the radius.

208 Proliferative Lesions of the Ulna

50 45 40 35 30 25 Absent

Number 20 Present 15 10 5 0

Age

Figure 6-16: Number of individuals in each age category with proliferative lesions of the ulna.

Proliferative Lesions of the Femur 50 45 40 35 30 25

Number 20 Absent 15 Present 10 5 0

Age

Figure 6-17: Number of individuals in each age category with proliferative lesions on the femur.

209 Proliferative Lesions of the Tibia 60

Number Absent 20 Present 10

Age

Figure 6-18: Number of individuals in each age category with proliferative lesions of the tibia.

Proliferative Lesions of the Fibula 40

Number 15 Absent

10 Present

Age

Figure 6-19: Number of individuals in each age category with proliferative lesions of the fibula.

210

Table 6-12: Parameter values for humerus. LL = -681.9573, AIC = 1377.9145

Parameter Parameter estimate Standard error

α1 0.0864 0.0162

β1 0.1396 0.0533

α2 0.0000 0.0118

α3 0.0016 0.0020

β3 0.0677 0.0163

k1 0.0015 0.0015

k2 4.8135 6.4000

Table 6-13: Parameter values for radius. LL = -618.3896, AIC =1250.7792

Parameter Parameter estimate Standard error

α1 0.0782 0.0157

β1 0.1311 0.0553

α2 0.0000 0.0118

α3 0.0013 0.0017

β3 0.0698 0.0175

k1 0.0016 0.0007

k2 2.7835 1.2501

211

Table 6-14: Parameter values for ulna. LL = -671.2828, AIC =1356.5657

Parameter Parameter estimate Standard error

α1 0.1239 0.0230

β1 0.2469 0.0700

α2 0.0000 0.0090

α3 0.0021 0.0026

β3 0.0591 0.0173

k1 0.0053 0.0012

k2 2.1940 0.6883

Table 6-15: Parameter values for femur. LL = -716.2528, AIC = 1446.5057

Parameter Parameter estimate Standard error

α1 0.0897 0.0181

β1 0.2099 0.0719

α2 0.0000 0.0102

α3 0.0038 0.0044

β3 0.0419 0.0118

k1 0.0064 0.0015

k2 4.5353 2.0501

212

Table 6-16: Parameter values for tibia. LL = -838.8199, AIC = 1691.6397

Parameter Parameter estimate Standard error

α1 0.0861 0.0157

β1 0.1991 0.0503

α2 0.0000 0.0054

α3 0.0016 0.0018

β3 0.0504 0.0122

k1 0.0171 0.0030

k2 5.1573 2.0230

Table 6-17: Parameter values for fibula. LL = -730.8417, AIC =1475.6834

Parameter Parameter estimate Standard error

α1 0.0799 0.0145

β1 0.1409 0.0539

α2 0.0000 0.0096

α3 0.0010 0.0014

β3 0.0726 0.0188

k1 0.0129 0.0021

k2 1.5347 0.4176

213 In general, it appears that the majority of pathological lesions tested in the Cholula population were associated with an increase in the risk of death. While the k2 estimates for some pathological lesions were not found to cause a statistically significant increase in mortality, these results must be viewed tentatively as insufficient sample sizes or the small number of individuals who were observed to have lesions may have affected the outcome of these analyses. The likelihood results for the canine, when viewed in conjunction with the distribution of canine enamel hypoplasias by age, does suggest that enamel hypoplasias on this particular tooth do not affect the risk of death in the Cholula population. Similarly, proliferative lesions on the fibula do not appear to increase the risk of death.

Furthermore, it is important to note that the assumption that is made in the Usher model, by virtue of the way the k1 and k2 parameters are specified, is that all individuals, regardless of age, are at equal risk of acquiring and dying from a particular pathology. The results of the likelihood analyses discussed above, therefore, indicate the relationship between pathological lesions and mortality at the aggregate level. However, age-related variations in selective mortality are almost a certainty. For example, examining the distribution of proliferative lesions on the long bones, it is notable that relatively few infants were observed to have this pathology.

Quite probably, most of these individuals, as they are likely among the frailest in their cohort, simply die of the infection before the lesions have a chance to form. The presence of proliferative lesions indicative of infection in an infant may, therefore, indicate a somewhat

―healthier‖ (relatively speaking) individual, who was, at least, able to survive long enough for the lesions to develop. However, in other age groups the relationship between proliferative lesions and the risk of death may not be the same.

214 Some studies have, in fact, indicated age-related differences in the risk of death associated with pathological lesions. Goodman (1996) found that in the Hamann-Todd collection those individuals that had enamel hypoplasias that formed from ages one to two died at older ages than those who had hypoplasias that formed from ages two to four. Ferrell (2003) found that accentuated striae, a type of enamel microdefect, were generally associated with an increase in the risk of death, except from ages two to four, when they were associated with a reduced risk of death. Clearly, it is necessary for paleodemographers to study not only aggregate-level relationships between pathological lesions and mortality, but also the age-related pattern of risks associated with particular pathological conditions. Enamel hypoplasias afford just such an opportunity, as the age at which the defect formed can be estimated. In future studies of the Cholula assemblage, an examination of the age at formation of enamel hypoplasias and the associated risk of death will be completed to determine if the age at which the stress occurred had an effect on mortality in this population.

In this chapter, the Usher reduced model of health was used to look at the relationship between several pathological lesions and the risk of death for the Postclassic population of

Cholula. Enamel hypoplasias on the incisors and first and second mandibular molars were shown to be associated with increased mortality, as were proliferative lesions on the ulna, the femur, and the tibia. Enamel hypoplasias on the canine had no effect on mortality, nor did proliferative lesions on the fibula. The parameter estimates for other pathologies are more equivocal because of small sample sizes or small numbers of observations. More research must be done in order to model underlying frailty in populations and, thereby, resolve the Osteological

Paradox. While we cannot yet draw any conclusions about the health of the Postclassic residents

215 of Cholula, being able to specify the relationship between observed pathological lesions and the risk of death in this population brings us one step closer to that goal.

216 CHAPTER VII

MIGRATION IN PREHISPANIC MESOAMERICA

“8 casas de chichimecos, 6 casas de otomíes, 5 casas de pinome (chochos), que se han juntado, personas reunidos, maceualli pobres, que han venido de tierras lejanas” [quote referring to the grouping of immigrants in a town in the Puebla-Tlaxcala region (Reyes García 1988a: 110)]

The paleodemographic age-at-death distribution for Cholula indicates that this New

World urban center might have had patterns of mortality that differed from those of preindustrial

Old World cities. In contrast to London, mortality levels in young adults in Cholula was low, suggesting that the scale of immigration, the demographic characteristics of immigrants, or the different epidemiological experience that immigrants faced upon arriving to Cholula could have resulted in fundamentally distinct urban population dynamics. In the current chapter, I explore in greater detail these possible hypotheses, and I consider how political, social, and economic variables may have shaped the process of migration in prehispanic Mesoamerica. In particular, I focus on archaeological and ethnohistoric sources that provide clues as to how common migration might have been in the prehispanic past, who these migrants were, why they chose to migrate, and how they would have been incorporated into their newly-chosen communities.

217 Possible Explanations for Low Mortality in Young Adults

The lack of epidemic diseases

The differences in young adult mortality observed in Early Modern London and

Postclassic Cholula could be attributable to a variety of factors. Perhaps the most obvious explanation is that epidemic diseases, which were largely responsible for the high death rates of immigrants to preindustrial European cities, were not part of the epidemiological regime of New

World urban centers. Consequently, immigrants to Cholula would not have been faced with virulent diseases to which they had no previous exposure. On this front, however, we must consider that migration involves several selection processes that should be weighed when evaluating how the urban environment might have affected immigrants. Individuals who relocate to an urban environment to pursue economic opportunities, in many cases, choose to move because they have little economic security in their community of origin. Thus, they immigrate in hopes of bettering their lots in life. If these immigrants were immunologically weakened as a result of malnutrition related to their impoverished states, they might have been at higher risk of death from the bacterial or viral infections that would have been present in

Cholula, even though these diseases were not epidemic in nature.

On the other hand, the very act of migrating from one locale to another is itself a selection process in which frailer individuals are likely to be excluded before they reach their destinations. Particularly in Mesoamerica, where forms of transportation were limited, immigrants would have had to walk to their new communities carrying their belongings. Sickly individuals would have been less likely to even begin the trip or would have succumbed along

218 the way. Those individuals who arrived in the city, while perhaps economically disadvantaged, had already proven themselves to have a certain degree of hardiness just by virtue of the fact that they undertook and survived the journey.

It is, therefore, unclear how immigrants to Cholula would have fared in an epidemiological sense. While they were not exposed to novel diseases, competing causes of death could have resulted in increased mortality in immunologically-vulnerable individuals.

However, given that immigrants who arrived in the city were potentially less-frail individuals simply by virtue of having succeeded in their endeavor, the infectious diseases present in

Postclassic Cholula may have posed little threat.

The scale of migration and the demographic characteristics of immigrants

In addition to the epidemiological differences between the Old and New Worlds, rates of immigration and the demographic characteristics of immigrants to preindustrial European cities were largely determined by economic and social conditions particular to those societies.

Economic growth in London during the Early Modern period resulted in increased economic opportunities for apprentices and servants. Consequently, young people moved to the city in large numbers to take advantage of these jobs. As was discussed previously, these individuals met with social obstacles that may have reduced their immunocompetence, thereby lowering their resistance to infectious disease and increasing their mortality rates. Furthermore, legal prohibitions restricting apprentices from marrying simultaneously reduced their fertility rates.

It is necessary to consider how common migration would have been in New World urban centers, such as Cholula. While economic and social impetuses to migrate surely existed in

219 prehispanic Mesoamerica, they were not on par with those of London. The fact that a significant portion of the populations of most prehispanic Mesoamerican cities engaged in agricultural activities would have given rural immigrants fewer incentives to move to urban areas. It is also not clear how immigrants to prehispanic cities would have been incorporated into craft production activities, provided that this was even a common occurrence. As a result of the political economy of Central Mexico, the influx of migrants into Cholula may have been considerably less than that experienced in preindustrial European cities. If that is the case, their mortality and fertility rates would have had less impact on the overall demographic experience of

Cholula.

In addition, the possibility exists that immigrants to Cholula came from different demographic groups than London immigrants and were subjected to very different conditions upon arriving in the city. If immigrants to Cholula were not primarily young adults, the low mortality observed in this demographic in the Cholula population is uninformative about the mortality experience of immigrants to the urban center. Studies of migration in a wide variety of societies across time and space indicate that migrants are typically young adults moving to pursue economic opportunities or relocating to marry (Rogers and Castro 1984). To a lesser extent, juveniles who move with their parents, and older individuals who go to live with their children or other family members in their advancing age also make up part of the flow of migrants. While it would be highly unusual if young adults did not comprise the bulk of immigrants to Cholula, the topic should be explored before it is simply assumed to be true.

Furthermore, the conditions immigrants faced upon arriving to the city should be considered in an attempt to subjectively evaluate how their frailty may have been impacted by their status as an

220 immigrant. Drawing from ethnohistorical accounts and archaeological evidence, I will attempt to address these questions in the following sections.

What is an Immigrant?

In considering how the urban environment may have affected immigrants to Cholula, it may be useful to first ponder who immigrants to Cholula may have been, where they came from, why they chose to migrate, and how they would have been received in the city. Of course, precious little direct evidence exists to answer these questions with confidence. However, we can extrapolate from ethnological, ethnohistorical, and archaeological sources to offer some speculative answers.

Even before considering who these immigrants might have been, we must first consider what an immigrant is from a cultural perspective. In modern demographic studies, migration is loosely defined as a change in a person’s usual place of residence (Hinde 1998: 191). This definition is of limited utility as people may choose to relocate to a new home permanently, or they may stay only temporarily. They may migrate over very long distances or very short distances. Of course, moves over especially short distances (within the same community, for example) are typically excluded from the definition of migration as they involve no real change in social structure, support networks, or economic opportunities. Rather, migration is implicitly assumed to involve crossing a real or perceived boundary, particularly a political boundary38

(Hinde 1998: 191). Furthermore, distinctions are often made among the terms immigrant, in-

38 Political boundaries in the modern world are typically well-defined, but such is not the case for the ancient world. As a result, perceptions of who an immigrant was may have varied. Later in the chapter, I will consider how the political unit of the altepetl might have influenced who was perceived as an immigrant in Postclassic Cholula.

221 migrant, emigrant, and out-migrant. Immigration is movement into a population that includes the crossing of an international boundary while in-migration takes place within national boundaries. Similarly, emigration is movement out of a population that includes the crossing of an international boundary while out-migration occurs within the boundaries of a nation (Hinde

1998: 191).

Transferring these concepts to archaeological populations is problematic. The above definitions rely heavily on political boundaries, which can generally only be approximated by the archaeologist and which may not have even been clearly defined by the societies involved. For the purposes of this project, we are interested in all individuals who moved from outside the urban zone of Cholula into the city, and especially in those individuals who emigrated from rural areas, as they would have been most likely to be negatively affected by the urban epidemiological regime. Of course, what is meant by the urban zone of Cholula is problematic in and of itself, as the altepetl form of organization that existed in Mesoamerica made no clear distinctions, at least in a political sense, between urban and rural. Therefore, the urban zone of

Cholula would not have had a marked political boundary – the standard that is often used in modern demographic studies of migration.

From a purely practical perspective, it makes little real difference how migration or the urban zone is defined. Isotopic methods of identifying migrants simply do not allow for such precision. While our primary interest may be individuals who moved from rural areas, the current resolution of strontium and oxygen isotope analyses is not sufficient for us to be able to identify immigrants in such narrow terms. Further discussion of this issue can be found in the following chapter; however, suffice it to say that methodological limitations of strontium and

222 oxygen isotope analyses impose a definition of immigrant that does not necessarily correspond to the individuals we are interested in for the purposes of the current investigation.

While our above etic definition of migration is fairly restricted in its usefulness because of these methodological limitations, an emic perspective on immigration to Postclassic Cholula could provide insights into the way in which immigrants were incorporated into society. How immigrants were treated and whether they had a social support network could have affected their frailty, which, in turn, would have impacted their levels of morbidity and mortality. I will come back to this point throughout the rest of the chapter in considering whether immigrants would have been openly welcomed in Cholula.

Temporary Immigrants and Visitors to the City

As an urban center, Cholula would have attracted a number of foreign visitors who came to the city to worship at the temple of Quetzalcoatl or to participate in the markets. These individuals are better classified as visitors to the city rather than immigrants, as they resided in

Cholula only temporarily. However, we should briefly consider who these visitors were and whether they are likely to be represented in the Cholula skeletal sample because, from an isotopic perspective, these individuals will be indistinguishable from true immigrants.

Pilgrims

Cholula, as a religious center, attracted a large number of pilgrims who came to worship at the temple of Quetzalcoatl. We know relatively little about those who made pilgrimages to

223 Cholula, other than the fact that they came from all walks of life. Nobles from other settlements were said to visit the shrine to pay homage to the god and to be confirmed as rulers of settlements within Cholula’s sphere of influence (Rojas 1985:130-131). However, commoners also seem to have brought offerings to the temple of Quetzalcoatl in order to show their devotion, as Rojas (1985: 131-132), writing in the sixteenth century, states that Cholula was to the indigenous groups of the region what Mecca is to Muslims. Among the pilgrims were individuals with ―las bubas, y del mal de los ojos, y del romádico y tosse‖ (tumors, diseases of the eyes, and certain respiratory infections and coughs), as Quetzalcoatl was associated with these particular infirmities. The annual ceremony to the god even included performances in which diseased individuals were impersonated. Prayers were made on behalf of these people, and the afflicted came to the temple to leave offerings to the god in the hopes that their suffering would be alleviated (Durán 1971, Chapter IV: 128-136; Motolinía 1971: 339-344).

These pilgrims may have traveled alone or, probably more commonly, in groups from other settlements, both near and far. On this point, Rojas comments only that Indians ―de toda la tierra‖ (―from all the land‖) came to Cholula to demonstrate their obedience to the god (1985:

131). It is likely that people of all ages were present among the pilgrims, but as significant travel on foot was likely required, less mobile individuals such as the very young or the very old may have been fewer in number. How long these individuals would have remained in Cholula and how the city provided accommodations for them is unclear. Motolinia refers to houses and rooms in Cholula maintained by different provinces where pilgrims from those regions could stay while in the city:

…y cada provincia tenía sus salas y casas dentro de Chololla, donde se aposentaban. (1971: 70-71)

…and each province had its rooms and houses within Cholula, where they took up lodgings. (author’s translation)

224

However, it is unknown if these houses were exclusively for use by nobles or if similar accommodations were made for commoners. It is possible that commoners found refuge with relatives or members of their altepetl or ethnic group or that some residents of the city offered hospitality for a fee. In Tenochtitlan-Tlatelolco, ethnohistoric sources make reference to hostels in the city (Calnek 1982: 291), and similar arrangements may have been available in Cholula.

Commoners may have also simply established temporary camps.

The fact that Cholula was a pilgrimage center has potential implications for patterns of morbidity and mortality in the city. In the Old World, during outbreaks of diseases like the plague, there was considerable concern about individuals entering the city who might have been exposed to infection in other regions. Wrigley, in fact, cites displacements and trade as potential sources of disease in urban areas (1967). It seems that pilgrims would be in a similar category in that they could be possible carriers of diseases acquired from other regions. However, the primary concern in the Old World was that visitors to the city might transmit epidemic diseases.

In Mesoamerica, given the absence of epidemic diseases, any infections the pilgrims might have had were likely already endemic in Cholula. Therefore, infectious disease outbreaks precipitated by pilgrims were probably not a concern in this New World city.

If a significant number of very old or very young individuals were among the pilgrims to

Cholula, their greater frailty may have contributed to mortality rates in the city. The difficulties of the journey or the living conditions experienced by pilgrims could have potentially made them even more vulnerable to endemic infections found in Cholula. On the other hand, as mentioned previously, the strenuousness of the pilgrimage may have limited the inclusion of these age groups. As some pilgrims traveled to Cholula to be cured of various maladies, they, too, may have had a higher mortality rate than other immigrants as a result of their preexisting condition.

225 An already weakened immune system caused by their affliction, or perhaps greater underlying frailty, may have also increased their susceptibility to acquiring other infections in the city. It is possible that some pilgrims may be represented in the skeletal sample under study if pilgrims were taken in by residents of Cholula. However, this number is likely to be small, and should not pose a serious bias to the current study.

Merchants and vendors

The market was also a likely cause of individuals coming to the city, at least temporarily.

Merchants and traders would have frequented the city of Cholula bringing wares to the market.

Individuals from rural areas may have also made temporary visits to the city in order to sell crops or other goods. In rural Copan, Sanders and Webster (1988) found that peasants engaged in part- time industries such as the production of manos and matates. This part-time production was evident from prehispanic times to the present, and modern day peasants would travel substantial distances to sell their goods. Some of the numerous vendors that are said to have been in the prehispanic markets were, therefore, likely to have been rural residents who came to the city for brief intervals in order to sell their products or to purchase needed items. Again, these individuals would be represented in the skeletal sample under study only if they found temporary housing with the residents of Cholula, and, even then, they are unlikely to be represented in sufficient numbers to affect the current investigation.

226 Rotational labor obligations

As part of their tribute obligations, macehualtin took turns performing domestic tasks for their lord. In the house of Francisca de la Cruz of Tepeaca, for example, each week, eight men and eight women were required to provide services in the lord’s house (Reyes García 1988a:

108-109). As the nobility typically lived in the city, or the center of the altepetl, macehualtin may have had to travel from the countryside to the city center in order to serve in this capacity.

How long they would have remained in the noble’s house is unknown and may have varied depending upon the particular tecalli and the distance the macehualtin had to travel. These macehualtin might have temporarily resided in the city in order to fulfill their tribute obligations.

However, given that the Cholula skeletal collection comes from low-status residences, commoners who perished in the city while fulfilling their rotational labor obligations are probably not present within the sample.

Group Migrations

Much of the information we have on immigration in Mesoamerica in general, and

Cholula in particular, comes from ethnohistoric sources discussing group migrations. The histories of many ethnic groups include stories of migration, and while these stories may be mythologized, at least in part, they may also contain some elements of truth (Beekman and

Christensen 2003; Smith 1984). Regardless of the accuracy of the accounts, however, native histories of migrations reveal something about prehispanic motivations for migration and the incorporation of migrants into a new homeland. Group migrations in the ethnohistoric accounts

227 are largely influenced by push factors in the place of origin, such as political discord, wars, famine, or natural disasters. Nevertheless, there are also stories of groups being given land in exchange for their services as warriors. For the most part, these group migrations involved travelling significant distances before the immigrants finally settled in their new home. Along the way, they typically stopped at a number of different sites at which they stayed only temporarily39.

Accounts of population origins

The settlement of the Mexica in Tenochtitlan is perhaps one of the most well-known stories of migration in prehispanic Mesoamerica (Smith 1984). According to the written histories of the Mexica, they, along with a number of other Nahuatl-speaking groups, migrated south into the Basin of Mexico sometime around the twelfth or thirteenth century. They traced their origins to a place referred to as Aztlan, an island in the middle of a lake, believed by some archaeologists to be located in West Mexico. Eventually, they began their move southward, guided by their idol Huitzilopochtli, but when they arrived in the Basin of Mexico and attempted to settle in Chapultepec, they were attacked by their neighbors, the Tepanecs of Azcapotzalco

(Figure 7-1). The Mexica were then forced to ask the ruler of Colhuacan for land. Their request was granted, but after a disastrous incident in which the Mexica sacrificed the daughter of the ruler of Colhuacan, they fled to an unoccupied island in the middle of Lake Texcoco, where they founded Tenochtitlan.

39 Some of the place names given in prehispanic migration accounts have, in fact, been matched up to towns and cities that existed in early colonial documents (Smith 1984; Kirchhoff 1947).

228

Figure 7-1: Maps of the Basin of Mexico (Gibson 1967).

229 Like the Mexica, the dominant ethnic group in Cholula, the Tolteca-Chichimeca, traced their origins to a mythical homeland in the north from which they migrated a few centuries before the arrival of the Spanish. The Historia Tolteca-Chichimeca (Berlin and Rendon 1947), which seems to have been written in the Cuauhtinchan region near Cholula (Figure 7-2), provides a detailed account of the journey of the Tolteca-Chichimeca from their departure from

Colhuacatépec-Chicomóztoc, believed to be near Tula (Figure 7-1), to their arrival in Cholula in the twelfth century. According to this account, their migration lasted many decades, with the emigrants temporarily residing at a number of different places for a period of between six and eleven years. When they finally arrived in Cholula, they were forced into servitude by the ruling ethnic group, the Olmeca-Xicallanca. The Tolteca-Chichimeca soon grew weary of their lowly status and decided to overthrow their oppressors. After defeating the Olmeca-Xicallanca, they expelled the ruling group from the city and took political control of the religious center.

Five years after their triumphant takeover of the Cholula, the Tolteca-Chichimeca faced a new threat. Tribes allied to the Olmeca-Xicallanca began a war against the new rulers of the city. The Tolteca-Chichimeca asked the seven Chichimec tribes with whom they had lived in

Colhuacatépec-Chicomóztoc to immigrate to Cholula in order to help them defeat their enemies.

The text claims the group took a mere ten days to arrive in Cholula, although they passed through many places on their way. In any case, the Toltec-Chichimeca were victorious with help from their Chichimec allies, and as a reward for their service, the Cholulteca gave the Chichimec tribes lands in the modern state of Puebla, where they established the independent kingdom of

Cuauhtinchan (Figures 7-1 and 7-2) in the twelfth century.

The Historia Tolteca-Chichimeca (Berlin and Rendon 1947, Paragraphs 322-327) also tells of another ethnic group that settled in the Puebla-Tlaxcala region. The Mixteca-Popolloca

230 arrived in Cacauilotlan soon after the Chichimecas established themselves in Cuauhtinchan.

There they were warmly received and given women by the Chimalpanea Xacomolca. They remained in Cacauilotlan for 35 years before immigrating to Oztoticpac, where they lived for 86 years before finally settling in Tecamachalco (Figure 7-2) (Reyes García 1988a: 56-58).

Both the Historia Tolteca-Chichimeca (Berlin and Rendon 1947, Paragraphs 337-338) and the Manuscript of 1553 (Reyes García 1988b, Paragraphs 51-79, 181-184: 85-86, 98; Reyes

García 1988a: 62-66) describe a similar migration of calpulli from Cholula to Cuauhtinchan approximately 60 years after the Chichimecs established themselves in the region (Figure 7-2).

According to these texts, 25 Tolteca-Chichimeca calpulli left Cholula in the middle of the thirteenth century following an attack on the city by the Huexotzinco in which the temple of

Quetzalcoatl was defaced and crops were destroyed. As a result of the famine that ensued, the calpulli decided to abandon Cholula and seek their fortunes in nearby Cuauhtinchan. The priests of the calpulli migrated first, and they were then followed by the macehualtin, or commoners.

The most powerful ruler, or tlahtoani, in Cuauhtinchan was head of the lineage that had received six Cholulteca women when the Chichimeca helped Cholula defeat their enemies. Consequently, he welcomed the calpulli as long lost relatives. They were provided with land but freed of all tribute obligations. They were also given women to marry. The Cholulteca calpulli remained in

Cuauhtinchan for 240 years before emigrating to Tepeyacac.

231

Figure 7-2: Map of the Puebla-Tlaxcala region. (Lomelí 2001: 40)

232 Wars and natural disasters

Migrations provoked by war seem to have been a fairly common cause of group migrations in prehispanic Mesoamerica. A number of immigrants to Tenochtitlan-Tlatelolco are said to have moved to the city as war refugees. Both the Huexotzinco and the Cuauhquecholteca immigrated to the city after the Tlaxcalans attacked their homelands (Figure 7-2). Similarly, the population of Colhuacan resettled in Tenochtitlan, Texcoco, and Cuauhtitlan after abandoning their city in the fourteenth century (Figure 7-1) (Calnek 1982: 289-290). It appears that war refugees were readily incorporated into their new communities in many cases. When the Mexica fled to Colhuacan after their defeat in Chapultepec, they received land and intermarried with native residents. Many, in fact, chose to stay in Colhuacan even after the founding of

Tenochtitlan (Calnek 1982: 289).

Large scale migrations also occurred as the result of natural disasters. In the fifteenth century, a severe drought hit the Basin of Mexico and lasted a number of years. While the government of the city attempted to mediate the effects of the agricultural shortfall on the population, ultimately the state food stores ran out before the drought was over. Native documents indicate that many people left the city and that some members of the elite migrated to the Gulf Coast region (Quiñones Keber 1995). Archaeological evidence also indirectly points to the possibility of mass migrations occurring as the result of natural disasters. The first century

AD eruption of the volcano Popocatépetl in the Puebla-Tlaxcala region appears to have resulted in a depopulation of the area. It has been suggested that the population growth that occurred in

Cholula and Teotihuacan around the same time was the result of the former inhabitants of sites

233 affected by the eruption relocating to these urban centers (Plunket and Uruñuela 2005; Uruñuela and Plunket 2005).

Established colonies

In addition to these stories of voluntary migrations by ethnic groups, there are also accounts of conquering polities establishing colonies in dependent cities, although this appears to have been a relatively uncommon occurrence. For example, the Aztec sent colonies into

Oztoma, Toluca, and Totonacapan in order to help maintain control of these subject areas

(Carrasco 1971: 361). After Tepeyacac (Figure 7-2) was conquered by the Aztec king

Moteuczoma Ilhuicamina, a noble named Coacuech was sent to serve as a governor of the town, and inhabitants of the Basin of Mexico were resettled to the region, as were the Cholula calpulli that had immigrated to Cuauhtinchan (Carrasco 1999: 289). Similarly, the Texcoco king

Nezahualcoyotl sent colonists drawn from each of the six sections of the city to settle the newly- established colony of Calpollalpan (Figure 7-1) (Hicks 1982: 237).

Ethnic barrios

Ethnohistoric and archaeological accounts of ethnic neighborhoods provide indirect information about the movement of people and groups in prehispanic Mesoamerica. Of course, many of these neighborhoods would have been established by both group and individual migrations, and there may have been streams of migrants moving back and forth between the homeland and the enclave. Sahagún (1950-69, Book 9: 80) describes a number of different

234 ethnic barrios in Tenochtitlan which were associated with the production of particular crafts.

The lapidaries, for example, were said to have come from Xochimilco, while the manuscript painters were Mixtecos (Calnek 1982: 289). Many of the founding members of these ethnic groups would have had to migrate substantial distances to establish themselves in the Mexica city.

Teotihuacan also seems to have had ethnic barrios, and a significant amount of information has been gleaned about prehispanic migration from the archaeological and osteological material excavated in these neighborhoods. The Oaxaca Barrio of Teotihuacan, for instance, has a ceramic industry, architectural features, and burials styles that all point to the fact that the inhabitants either originated in Oaxaca or had significant contact with the region.

Oxygen isotope data collected from skeletons excavated in the barrio provide evidence that residents of this neighborhood interacted both with their homeland of Oaxaca and with other diaspora communities and that immigration occurred among these locations (White et al. 2004a).

The Merchants’ Barrio of Teotihuacan is also thought to be an ethnic enclave of individuals originally from the Gulf Coast, again because of the presence of foreign architectural forms, burial practices, and ceramic styles (Spence 1996). Strontium analysis of the Merchants’ Barrio demonstrated that inhabitants were emigrating from two different areas of the Gulf Coast, thus confirming archaeological evidence linking the compound with Gulf Coast cultures (Price et al.

2000).

Six ethnic groups were known to have immigrated and settled in the Acolhuacan area

(Figure 7-1). Two of these groups, the Chimalpaneca, and the Tlailotlaque, arrived in the first half of the fourteenth century when the ruler Quinatzin was in power, and the other four, the

Mexitin, the Tepaneca, the Colhuaque, and the Huitznahuaque, immigrated during the reign of

235 Techotlalatzin in the latter part of the fourteenth century (Carrasco 1971: 366). Each of these groups was given land and allowed to establish themselves in Texcoco and other towns in the region. While these two rulers assigned the newcomers a place to settle, it was Netzahualcoyotl that allowed each ethnic group to construct its own ceremonial center in the mid-fifteenth century (Hicks 1982: 236). At the time of the conquest, Texcoco was divided into six sections which bore the names of the ethnic groups mentioned above (Hicks 1982: 236).

Additional evidence of group migrations comes from a 1554 document, in which immigrant groups of macehualtin in Amozoc in the Puebla-Tlaxcala region (Figure 7-2) are briefly mentioned:

8 casas de chichimecos 6 casas de otomíes 5 casas de pinome (chochos) que se han juntado personas reunidos maceualli pobres que han venido de tierras lejanas (quoted in Reyes García 1988a: 110)

8 houses of chichimecos 6 houses of otomies 5 houses of pinome (chochos) that have been brought together reunited people poor maceualli that have come from far-away lands (author’s translation)

These individuals had been made mayeques of a lord in Amozoc and, therefore, owed him tribute. Several different ethnic groups seem to have been grouped together for organizational purposes. While the text does not mention the reasons for these migrations, it does indicate the way in which these ethnic groups were managed upon their arrival to Amozoc. These immigrant groups appear to have been given land by a lord in the region in exchange for fulfilling certain tribute obligations.

236 Evidence of an ethnic ward or section also exists for the city of Cholula. An ethnic group named the Colomochca was said to have migrated to Cholula from Cuauhtinchan (Berlin and

Rendon 1947, Paragraph 334). While five of the six cabeceras, or wards, in Cholula have names associated with Tolteca-Chichimeca groups that immigrated to the city, a cabecera with the name Colomochco appears in several Spanish documents, including the Suma de Vistas

(Carrasco 1971: 62; Reyes García 1988a: 60-61). Little is known about the establishment of the group in Cholula, but based on the fact that a cabecera is named after them, they seem to have been provided with land.

What then can we conclude about group migrations in prehispanic Central Mexico? The migrations often seem to have involved movements over long-distances, but some short-distance migrations are recorded as well, as with the Cholulteca calpulli and the Colomochca. In many cases, the long-distance migrations occurred as a series of short-distance migrations from one place to another. In general, the newcomers, if not exactly welcomed with open arms, do, in fact, seem to have been incorporated into a lord’s house as macehualtin or mayeque. From the perspective of the nobility, welcoming immigrants and providing them with usufruct rights to land would have had economic benefits, as having more macehualtin would have increased the amount of tribute received by the lords. In turn, economic power could be translated into political power, so a lord could potentially become more influential than other lords in the altepetl by gaining more macehualtin.

237 Individual Migrations

While stories of group migrations are common in ethnohistoric documents, migrations of individuals receive little or no mention. We can draw two possible conclusions about this lack of information on individual migrations. One explanation is that individual migrations were uncommon in the prehispanic past because the political economy of Central Mexico did not provide sufficient economic incentives or opportunities to promote the migration of individuals.

Thus, group migrations were more typical within this cultural framework. As much is still unknown about migration in prehispanic Mesoamerica, we cannot completely discount this possibility. However, group migrations of the sort described in prehispanic Mesoamerican texts are unusual from a demographic perspective. Migrations of individuals or small family groups are significantly more common in societies around the world (Rogers and Castro 1984). The second explanation for the absence of stories of individual migrations in ethnohistoric documents is that these migrations simply went unrecorded. Group migrations have broader ramifications for the societies involved, while individual migrations may not have been perceived as significant enough to warrant mention in indigenous texts. Even in preindustrial European cities, where written documents were more abundant, migrants are largely invisible in the historic record. Archaeological and isotopic data as well as early colonial texts provide limited evidence that individual migrations did occur in prehispanic Mesoamerica.

238 Political motivations

Archaeological evidence suggests that members of the nobility occasionally migrated to other sites to act as rulers, ambassadors or other political figures. Political emissaries from

Teotihuacan are believed to have resided at Kaminaljuyu and Tikal because of the presence of

Teotihuacan-style talud-tablero architecture, burial customs, and material goods at these sites

(Sanders 1977; Spence 1996). In addition, rulers from some Maya centers are recorded as being foreigners. A ruler of Bonampak, for example, is thought to have originated from an unknown site in the Usumacinta River region (Palka 1996). At Palenque, some members of the elite are recorded as being from a foreign site, and the Dos Pilas ruling lineage seems to have been established by nobles from Tikal (Houston 1993: 99-100). At Cholula, the burial of an individual who appears to have been a Mayan ambassador was found in the Classic Period ceremonial center (Suárez 1985). However, as the current study involves low-status individuals, any immigrants in the sample are highly unlikely to have relocated due to political motivations.

Marriage

Marriage is one of the more frequent reasons for migration in traditional societies, and we can surmise that it was also a motivation for movement between sites in prehispanic

Mesoamerica, as marriage alliance were clearly a means of establishing relationships between groups (be they calpulli, ethnic groups, or altepeme) in prehispanic Mesoamerica. Exogamous marriages are well-documented among members of the nobility in prehispanic Mesoamerica.

Epigraphic evidence from the Maya region provides evidence that members of the elite

239 immigrated to other sites in order to marry into the ruling family (Molloy and Rathje 1974), and at Caracol, high-ranking foreign males were even said to have married into the ruling dynasty

(Stone et al. 1985). In Central Mexico, Aztec rulers had numerous wives and concubines who were often acquired from dependent states. Axayacatl, for example, had one wife from Tula and another from Itztapalapa (Carrasco 1971: 370). Lesser members of the nobility also married foreigners in order to promote political alliances or to expand their sphere of political influence, as when nobles from the region of Cuauhtinchan-Tecali married women from Tlatelolco

(Carrasco 1971: 370).

The ethnohistoric accounts of migration in the Puebla-Tlaxcala region, already discussed, mention a number of instances in which wives were given to other ethnic groups as a reward or as a means of ensuring an amicable relationship. For example, the Cholulteca gave wives to the

Chichimeca in the twelfth century when the latter group defeated the Xochimilca and Ayapanca who had come to the aid of the conquered Olmeca-Xicallanca. This exchange of women clearly benefited the Cholulteca some years later when war and famine forced some calpulli to emigrate to Cuauhtinchan. The tlahtoani Teuhctlecozauqui, who had married six women from Cholula, viewed the Cholulteca calpulli as family, and generously bestowed land upon them free of the typical tribute obligations. The Cholulteca were, in turn, given women by the Chichimeca group as a means of cementing the alliance.

Macehualtin also married foreigners, as is demonstrated by Olivera’s (1978) study of sixteenth century marriage records from Tecali, a region near Cholula. From 1583 to 1594, a total of 2132 marriages occurred in Tecali and its 22 dependent towns. In the case of two of these marriages, both the bride and the groom were immigrants, and in 95 (4.44%) additional marriages, one of the spouses, most often the bride (71%), was an immigrant. The fact that most

240 of these immigrant spouses were women concurs with Carrasco’s (1971: 367) assertion that in prehispanic Mexico, there was an ambilocal postmarital residence pattern with a preference for patrilocality.

Although one of the immigrant spouses in the above-mentioned study was from Mexico

City, the majority of the foreign spouses were from the Puebla-Tlaxcala region and belonged to the same ethnic groups that were present locally in Tecali. Most were also from towns or cities with which Tecali had maintained some sort of political alliance, such as Cuauhtinchan (38),

Tepeaca (25), and Cholula (10). Other immigrant spouses, although they came from somewhat more distant locations, were also from towns that maintained an amicable relationship with

Tecali, including Tlaxcala, Tecamachalco, Quecholac, Totomihuacan, and Tehuacan. While the percentage of marriages involving a foreigner are relatively low, Olivera suggests that exogamy would have been more common in prehispanic times, as the political and organizational system imposed by the Spanish interfered with traditional political alliances between indigenous groups and reduced the mobility of the population, both of which contributed to increased numbers of endogamous marriages. Olivera also notes that endogamous marriages were more common in the smaller towns than in the cabecera of Tecali.

Interestingly, when the marriage statistics within Tecali (including the cabecera and its

22 towns) are broken down by social status, marriages between macehualtin were much more likely to involve spouses from different towns than were marriages between pipiltin, or nobles.

Of 35 marriages in which both the bride and the groom were members of the nobility, there were no cases of the spouses being from different towns. This is largely attributable to the fact that the nobility typically resided in the cabecera. Of 1757 marriages between macehualtin in Tecali,

241 490 involved spouses from different towns. In some cases, poor pipiltin married macehualtin, and in one third of these mixed marriages, the spouses came from different towns.

Thus, from the available data, it appears that immigration due to marriage occurred in both the pipiltin and macehualtin populations of prehispanic Central Mexico. In most of these cases, patrilocality prevailed and the woman relocated to the town of her husband. Zorita (1994) reports an average age at marriage of 20 for men in the region, and we can surmise that the average age of marriage for women would have been the same or perhaps slightly younger. This would indicate that immigrants to Cholula would likely include a fair percentage of young women that had relocated to the city in order to marry a local resident.

For the most part, there is little indication that immigrant spouses would not have been welcomed or at least accepted in their new communities, as many seem to have come from allied political states. Furthermore, if they came from an ethnic group that was already well- established in the city or held a position of political power, they may have even enjoyed some degree of status within their new home. On the other hand, if relations between previously friendly ethnic groups suddenly soured, the immigrant spouse may have been confronted with a potentially dangerous situation. The Historia Tolteca-Chichimeca (Berlin and Rendon 1947,

Paragraphs 392-395) tells of a man from Cuauhtinchan who killed his Totomiua wife and a woman who killed her Totomiua husband after hostilities broke out between the Totomiuaques and the people of Cuauhtinchan. Even Totomiuaque children were said to have been stoned to death.

242 Economic factors

In demographic studies of both modern and historic societies, economic motivations often factor heavily into decisions to migrate to urban areas. For prehispanic Central Mexico, we have little information on individuals emigrating to pursue economic opportunities, but whether this reflects a true difference in patterns of migration in prehispanic Mesoamerica or merely a bias in the ethnohistoric record is not clear. In the Old World, many individuals migrated from rural to urban areas in order to pursue apprenticeships in the city. There is virtually no evidence that similar means of learning crafts existed in prehispanic Mesoamerica. Calnek (1982: 297) does mention an individual whom he refers to as an apprentice learning a trade from a craftsman in Tenochtitlan:

Thus, according to our literature, a certain platero (goldsmith or silversmith), who resided with his wife’s family in the barrio named Zacatlan in the great quarter of Atzacualpa, acknowledged the authority of the lords (principales) of the plateros’ guild (which was centered in the barrio of Yopico in Moyotlan) up to the time of his death in 1543. He himself employed an apprentice from Copolco in Cuepopan, who appears as a craftsman in his own right at a later date (AGN Tierras, vol. 30, exp. 1, fols. 14-16, 64). Although little more than two decades had elapsed since the Conquest, there is no indication that this arrangement was considered unusual.

However, details of the arrangement between the platero and his ―apprentice‖ are not known.

Furthermore, the document describing the apprenticeship postdates the Conquest (Calnek

1982:297). Although Calnek does not believe the situation to have been unusual, it may reflect

European influences or result from economic and social disturbances occurring at the time.

Instead, learning a craft specialization seems to have been a skill typically passed down from parent to child in prehispanic Mesoamerica. In Tenochtitlan-Tlatelolco, for example, the existence of ethnic barrios of craft specialists suggests that particular family or calpulli groups specialized in the production of certain products (Sahagún 1950-1969, Book 9: 80; Calnek 1982:

289). It is, therefore, unclear if knowledge of craft production would have been imparted to

243 unrelated immigrants arriving to Mesoamerican cities, but isotopic studies of the Tlajinga 33 apartment compound in Teotihuacan suggest that immigrants were sometimes incorporated into familial groups of economic specialists. The residents of Tlajinga 33 housed a corporate group of low-status craft producers. Although most members of the apartment compound are believed to have been related in some way, oxygen isotope studies have demonstrated that several immigrants also resided in Tlajinga 33, and it has been suggested that these individuals were brought into the compound specifically to work in the craft industry (White et al. 2004b).

Land could have been another economic incentive for migration. Although calpulli members were given usufruct rights to land in exchange for providing tribute and services to their lords, in some cases, individuals did not have sufficient land or sufficiently productive land to meet all of their needs. When calpulli had extra land, they would sometimes rent it to members of other calpulli that did not have enough land for all their members (Carrasco 1971:

355; 364-365). Alternately, individuals could become mayeque of a noble by moving to a region in which they had no calpulli ties. In fact, mayeque may have increased greatly in number over the course of the Postclassic precisely as a result of individuals moving around on the landscape.

Attaining access to land may have been a possible reason for individuals or family groups to move from one location to another. However, this would involve individuals immigrating to urban areas only if the city had a more dispersed settlement pattern, like Texcoco, in which some farmland was within the limits of the city itself, or if it was typical for individuals to reside in the urban center and farm lands in the hinterland.

In the Old World, cities also attracted the indigent because urban governments offered poor relief. As mentioned previously, Cortés comments pointedly on the number of poor in

Cholula, begging in the markets and from the nobles in the streets. While we do not know the

244 origin of these individuals, feasibly some of them could have been immigrants who came to the city specifically because of the potential charitable help that could be found there.

Clearly, little direct information on individuals migrating to pursue economic opportunities exists for prehispanic Mesoamerica. Further investigations into this topic are necessary in order to establish if individuals immigrated to Mesoamerican cities to pursue economic opportunities, and, if so, how they became involved in craft specialization or other economic activities once arriving in the city. In the final chapter, I discuss some possible avenues of future research that could help clarify the nature of migration in prehispanic

Mesoamerica.

Slavery

As slaves were bought and sold on the market, slavery sometimes resulted in individuals being relocated to a new community from their original place of residence. Slaves in prehispanic

Mexico were typically not chattel, but rather individuals who had sold themselves or their children to pay debts, individuals who had been captured in war, or criminals who had been made slaves as punishment for a particular crime. These individuals acquired service obligations, but were otherwise free to pursue their own personal interests. In some cases, slaves maintained their own households, but in other cases they lived in the house of their master. Female slaves were used for grinding corn and weaving, and they sometimes became the concubines of their masters. Male slaves did farming, cut and carried firewood, and served as porters. These individuals were treated as family members, but they could be sold at market if they were not sufficiently obedient (Carrasco 1971: 356). Slaves served in the houses of both nobles and

245 commoners, so it is possible that they could be present in the Cholula skeletal collection under study.

Migration in old age

While most immigrants tend to be young adults, older adults sometimes migrate to live with adult children or other family members in old age. The elderly mother or father of an immigrant to Cholula might have relocated to the city so that the child was better able to care for him or her. Conversely, an individual who immigrated earlier in life might have returned to his or her place of origin in old age to be with family. A possible example of these scenarios in a low-status habitational unit in Teotihuacan is discussed below.

Isotopic evidence of individual migrations

The apartment compound Tlajinga 33 in Teotihuacan has provided some evidence of the immigration of individuals, although the particular reasons for these immigrants resettling in the compound are unknown. Nothing discovered in the archaeological excavations of Tlajinga 33 suggests that this apartment compound housed a significant number of foreigners. However, oxygen isotope analysis of human skeletons recovered from Tlajinga revealed that 29% of the individuals sampled were immigrants (White et al. 2004b). Foreign individuals seem to have come from at least two different regions, one of which may have been Michoacan. Both men and women migrated to the apartment compound. White et al. (2004b) suggest that women may have migrated primarily to marry a resident of Tlajinga and men, to participate in the craft

246 specialization of the compound. However, they also propose alternative scenarios in which men may have married into the compound in order to retain the Tlajinga women or couples may have immigrated together. In addition, one individual, an older woman whose bone values indicated she had spent her adult life elsewhere, may have been a widow moving back to her childhood home or the mother of an immigrant who had established himself or herself in Tlajinga (White et al. 2004b: 186). Immigrants were found in both high and low status burials, but for the most part, they received higher status burial treatments after death, indicating that they attained some degree of importance within the apartment compound. Furthermore, a comparison of diet between immigrants and native inhabitants of the Teotihuacan habitational unit suggests that the diet of the two groups was indistinguishable, another indication that immigrants were well- integrated into Tlajinga.

How would Immigrants have been Received in Cholula?

With respect to the osteological sample from Cholula, there is no evidence from the archaeological excavations of the habitational units that this residential zone represents an ethnic enclave. Unlike the Merchants’ or the Oaxaca Barrios at Teotihuacan, architectural features, ceramics, and mortuary treatments are all consistent with local styles and practices (López at al.

1976). Therefore, any immigrants living in this residential area probably moved as individuals or in small family groups. Marriage, or perhaps the opportunity to participate in craft production or service industries in the city, would have been the most likely reasons for relocating to

Cholula, which would suggest that many of the immigrants came to city as young adults. Most

247 of these individuals would probably have immigrated to Cholula from nearby towns and rural areas, although some might have come from more distant locations.

How would these individuals have been perceived by the native-born residents of

Cholula? We can imagine that those individuals who migrated from distant places, particularly if they were from a different ethnic group, were likely considered to be foreigners or outsiders in some respect. They would have been the subjects of a different ruler and they would have brought with them customs and perhaps a language that differed from those of the Cholulteca.

As the altepetl was the primary political unit, its boundaries may very well have determined who was considered to be a foreigner and who was not. There do seem to have recognized boundaries or limits to the altepetl as the word altepetepantli (―walls of the city‖) and altepequaxochtli (―limits of a country or city‖) indicate (Bernal and García 2006). However, it is unclear how individuals who migrated from nearby locations would have been perceived.

Individuals from rural areas and dependent towns within the altepetl may not have been considered foreigners at all, as they would have been under the political authority of the same ruler as the inhabitants of the city of Cholula. In any case, immigration seems to have been a relatively common occurrence in prehispanic Mesoamerica and immigrants appear to have readily incorporated into their new homes, at least in economic terms. How they were treated on a social level was probably dependent on numerous factors including their place of origin, their ethnic affiliation, their socio-economic status, and their reasons for migrating.

Based upon the ethnographic and archeological evidence available, it would seem that young adults would likely have made up a large percentage of immigrants to Cholula, although family groups might have also been included in the migrant population. While ethnohistoric and archaeological sources of information provide some insight into the nature of migration and the

248 identity of immigrants in prehispanic Mesoamerica, they cannot directly address how immigrants would have contributed to demographic patterns in Postclassic Cholula. Furthermore, a great deal of invisible migration likely occurred in the past that is not discernible in the archaeological record. Stable isotope analysis provides a means of identifying individual migrants through bones and enamel. Therefore, in order to further examine the nature of migration in prehispanic

Mesoamerica and the demographic contribution of immigrants to Cholula, strontium and oxygen isotope analyses were carried out to identify those individuals within the skeletal collection under study who were not native residents of the city. The results of these isotope analyses will be presented in the following chapter.

249 CHAPTER VIII

STRONTIUM AND OXYGEN ISOTOPE ANALYSES

“…de toda la tierra venían….” (Rojas 1985: 131)

While ethnohistoric sources and archaeological remains from various sites in

Mesoamerica indicate that migration was a relatively common feature of the prehispanic past, the development over the last two decades of isotopic techniques to identify immigrants and foreigners has greatly expanded our ability to understand the movement of people on the landscape. Both strontium and oxygen isotopes have been used around the world to track the migration of people and animals in the archaeological record, and, in general, these studies have demonstrated that human populations in the past were, in fact, quite mobile (Hodell et al. 2004;

Hoppe et al. 2003; Burton et al. 2003; Burton et al. 1999; Knudson et al. 2004; Price et al. 2004;

Price et al. 2002; Price et al. 2001; Price et al. 2000; Price et al. 1998; Price et al. 1994;

Schweissing and Grupe 2003a,b; Montgomery et al. 2000; Ezzo et al. 1997; White et al. 2004a;

White et al. 2004b; Wright 2005a,b). In this chapter, I explain how strontium and oxygen isotopes can facilitate the identification of immigrants to Cholula, and I present the results of an isotopic analysis of skeletons from the Cholula collection. I then discuss what these results imply about immigration to the city and how immigrants may have influenced mortality and fertility in Postclassic Cholula.

250 The Study of Migration through Strontium and Oxygen Isotopes

Strontium isotopes

The use of strontium isotopes to identify immigrants is based upon the premise that ratios of strontium isotopes (87Sr/86Sr) vary across geographic areas due to underlying differences in local geologies (Price et al. 2008; Price et al. in press; Price et al. 2002; Burton et al. 2003).

The strontium isotope ratio of a geographical area is essentially transmitted up the food chain from plants to animals and then to humans. Because dietary strontium is incorporated into skeletal tissues, the bones and teeth of individuals reflect the strontium isotope ratio from the area in which they reside. Tooth enamel forms during early childhood, so it reflects the strontium signature of the place where the person spent his or her earliest years. In contrast to enamel, bone regenerates over time, resulting in strontium signatures that, eventually, will come to reflect those of the geographic area in which the individual was residing at the time of death.

By comparing the strontium ratios from the enamel and the bones of a particular individual, or by comparing strontium ratios in the enamel of an individual to a baseline that has been established for a region, immigrants to an area can be identified.

Mesoamerica is fruitful region in which to conduct strontium isotope studies, as the underlying geology across the culture area varies considerably. The Peten and the Yucatan

Peninsula, home of the lowland Maya, consist of a limestone shelf. The oldest limestones in this region are in the southern part of the peninsula, and they get progressively younger as you move northward. Because strontium isotopes in sea water have increased over time, lower strontium ratios are found in the southern portion of the Yucatan, and higher ratios are found in the north of the region (Hodell et al. 2004; Hess et al. 1986; Price et al. 2008). In the Central Highlands,

251 where Cholula is located, the geology is a mixture of volcanic deposits, sedimentary deposits of marine origin, and metamorphic rocks. The Sierra Madre Oriental, a mountain range along the eastern edge of the highlands, is primarily sedimentary rock of marine origin, while the Sierra

Madre Occidental, the mountain range flanking the western portion of the highlands, is volcanic in origin. In general, strontium isotope values increase from west to east (Price et al. 2008).

This high degree of variability in geological formations enables even geographically-close areas to be differentiated isotopically. The Laboratory for Archaeological Chemistry at the University of Wisconsin-Madison is currently involved in a project to map strontium isotope values across

Mesoamerica (Price et al. 2008). Their work thus far is shown in Table 8-1 and Figure 8-1.

Table 8-1: Median strontium isotope values by region. (Price et al. 2008)

252

Figure 8-1: Isotope values for sites across Mesoamerica. (Price et al. 2008)

253 Oxygen isotopes

Oxygen isotopes (18O/16O) have also been used in Mesoamerica to identify immigrants and foreigners at archaeological sites. In human bones and teeth, the ratio of 18O to 16O primarily reflects body water composition, which, in turn, reflects the oxygen isotope ratios of water sources. Water supplies in ancient populations typically came either directly or indirectly from rainfall, and factors such as latitude, elevation, and distance from the ocean can affect the isotopic ratio of rain water (White et al. 2002; White et al. 2004a, 2004b; Price et al., in press).

Consequently, oxygen isotope ratios in human skeletal remains vary according to the geographical provenience of an individual. While oxygen isotope ratios do show promise as a means of identifying immigrants, some problems with the method have yet to be addressed.

Many populations use river water for drinking or agricultural purposes, and the rivers may originate in a different geographical location, which would have an impact on the 18O/16O ratio.

At larger sites, where residents might be obtaining water from multiple springs or rivers, it is unclear how much variation in oxygen isotope ratios should be expected. While the studies thus far have demonstrated that intersite variation is slightly more than intrasite variation, there is still likely to be a great deal of uncertainty in identifying immigrants (Price et al., in press).

Moreover, in some cases, oxygen isotope ratios in human skeletal remains have not matched the anticipated local values when geography is taken into account (Price et al., in press). Although the Laboratory for Archaeological Chemistry at the University of Wisconsin-Madison is currently attempting to map oxygen values for Mesoamerica as well, this database is not as extensive as that for strontium. Due to the ambiguities that still surround oxygen isotopes, strontium isotope analysis appears to be a more robust means of identifying nonlocal individuals.

254 That said, an oxygen isotope analysis was run on the current sample as part of the ongoing mapping project being done at the Laboratory for Archaeological Chemistry. Those values will be presented here along with those of the strontium analysis, but the strontium data will be given more weight.

Isotopic Studies of Migration in Mesoamerica

In Mesoamerica, isotopic studies have been used to verify epigraphic inscriptions, to clarify connections between polities, and to trace the origins of sacrificial victims. A number of sites in the Maya region, including Tikal, Altun Ha, Kaminaljuyu, and Copan are believed to have had ties to Teotihuacan during the Classic Period. Strontium and oxygen isotopes have provided a means of exploring the nature of these relationships (White et al. 2000; White et al.

2001; Wright 2005a, 2005b).

At Tikal, Teotihuacano-style burials, architectural elements, epigraphic inscriptions, and iconographic features all point to the possibility that this Maya city was closely associated with

Teotihuacan (Coggins 1975; Moholy-Nagy 1999). Using strontium isotopes, Wright examined the validity of epigraphic readings from the site suggesting that the ruler Yax Nuun Ayiin I hailed from Teotihuacan and found that the isotopic signature of his tooth enamel was, contrary to expectations, local to Tikal (2005a). A further study of 83 individuals from Tikal revealed that 10% of the sample could be firmly identified as immigrants, while another 4% to 13% possibly had isotopic values outside the range of Tikal (Wright 2005b).

Similar studies in Altun Ha, Kaminaljuyu, and Copan have assessed the extent of

Teotihuacan influence at those sites. White et al. (2001) determined through oxygen isotope

255 analysis that an Early Classic elite male burial (Tomb F/8-1) in Altun Ha, thought to be an individual originally from Teotihuacan, was, in fact, a foreigner, although he was not from

Teotihuacan. At Kaminaljuyu, a site believed to have been an outpost of Teotihuacan (Sanders and Michels 1977), oxygen isotope analysis of burials resulted in the discovery of one individual who spent part of his childhood in Teotihuacan. Several other individuals were also shown to have been immigrants, although they were not from Teotihuacan (White et al. 2000). At Copan, strontium isotopes were used to determine the origins of the ruler Yax Ku’k Mo, who arrived to the site in AD 426. Architectural elements and artifacts in the tomb suggest an association with

Teotihuacan, but isotopic results indicate that he was most likely from the southern part of the

Yucatan (Buikstra et al. 2004; Price et al. 2008: 174-175).

At the Maya site of Palenque, Price et al. (2008: 175-176) examined epigraphic evidence regarding the king Pakal the Great, who is said to have been a native of Palenque who ascended to the throne at the age of 12. Strontium isotope analysis of tooth enamel from Pakal yielded a value significantly higher than the local geological signature. However, further consideration of the site by the authors of the study revealed that the agricultural fields of Palenque were underlain by younger limestone, whose strontium isotope ratio was in line with that of Pakal.

The study, thereby, illustrates the importance of fully understanding both the geology of a site, as well as the sources of the food supply.

Isotopic studies in Central Mexico have primarily focused on Teotihuacan (White et al.

2004a; White et al. 2004b; White et al. 1998; Price et al. 2000). Oxygen isotopes have been used at the site to determine the origins of sacrificial victims buried under the Feathered Serpent

Pyramid (White et al. 2002), and both oxygen and strontium isotopes have been used to confirm immigration and foreign influence in the Merchant’s Barrio and the Oaxaca Barrio (White et al.

256 1998; White et al. 2004a; Price et al. 2000). The oxygen isotope data for the Oaxaca Barrio provided evidence that residents of the barrio interacted both with the homeland and with other diaspora communities and that immigration occurred among these locations (White et al. 2004a).

Strontium analysis of the Merchants’ Barrio demonstrated that inhabitants were emigrating from two different areas of the Gulf Coast, thus confirming archaeological evidence linking the compound with Gulf Coast cultures (Price et al. 2000). Oxygen isotope data from Tlajinga 33 have demonstrated that the apartment compound included a number of immigrants (29% of the sample), and many of these immigrants obtained some social status within the compound (White et al. 2004a).

Several of the above referenced studies do consider the demographic characteristics of immigrants. For example, White et al. (2004b) note that both men and women immigrated to

Tlajinga 33, and that one older woman had immigrated to the apartment compound shortly before her death. In their study of the Oaxaca Barrio, White et al. (2004a) found that both males and females were relocating to the neighborhood and that many individuals immigrated to

Teotihuacan in infancy and early childhood. They also note that some immigrants died shortly after arriving to the city. The current isotope analysis of Cholula, therefore, builds on previous work in Mesoamerica by considering both the demographic characteristics of immigrants to this

New World city and how immigrants may have contributed to fertility and mortality at the urban center.

257 Sampling Procedure

Tooth Enamel

In order to identify immigrants in the Cholula population through strontium and oxygen isotope analyses, enamel and bone samples were collected. Tooth enamel, due to its hardness and density, is particularly useful in isotopic studies because it is less susceptible than bone to postmortem changes in its physical or chemical structure (Kohn et al. 1999; Price et al. 2002).

Therefore, contamination is not a significant problem in isotope studies when enamel samples are utilized. As the tooth enamel forms during childhood, it is also a useful indicator of where an individual spent his or her early years (Price et al. 2002; Burton et al. 2003).

Bone

While bone is more susceptible to postmortem change and contamination, it does provide useful information as to how long an individual has resided at his or her current location. Since bone remodels over time, it will eventually reflect the isotopic signature of the geographic area in which the individual was living at the time of death. However, this process may be prolonged, particularly in the case of thick cortical bone, like that found in the femur, where bone turnover rates are estimated to be approximately 3% per year (Price et al. 2002). Consequently, isotopic ratios in bone, when compared to those in the tooth enamel of the same individual, can theoretically indicate if a person has recently migrated, migrated a short time ago, or migrated in the distant past. If tooth enamel from a skeleton has an isotope ratio that indicates that an individual was born in a different location, but the isotope ratio from cortical bone is consistent

258 with the local signature, it suggests that the individual migrated long before death. If, however, both the enamel and the bone isotope ratios are different from the local signature, a more recent migration is likely. This information is significant to the present study because not only can an individual be identified as an immigrant, but also information about how soon they died after they arrived in the city can be discerned, thus providing a better picture of the timing of immigration and the relationship between immigration and mortality (Price et al. 2002).

Unfortunately, bone, being more porous than enamel, can be contaminated by strontium (as well as other elements) found in the soil in which it is deposited postmortem. Methods such as acid cleaning can help remove these elemental contaminants, and any remaining contaminant would reflect the strontium isotope ratio of the local geology (Grupe et al. 1997; Price et al.

2002). Bone’s susceptibility to contamination does make its use in strontium isotope studies complicated. A strontium value in bone that is different from the local value would indicate a recent immigrant, but a value consistent with the local signature would be equivocal between contamination and an individual who resided at the site for some time. As bone values can potentially provide valuable information regarding the timing of migration, bone samples were collected from 25 individuals from Cholula included in the current isotope study. The analysis of these bone samples has not yet been completed, but the results will be presented in a future publication.

Sampling strategy

A total of 50 enamel samples were analyzed for the current project. Enamel samples (of approximately 30 to 50 mg from at least half of the crown height) were preferentially taken from

259 the lower first molar, although samples from other teeth were substituted as necessary (see

Appendix K, Table K-1 for the sample list). Only permanent teeth were sampled. Samples were not taken from ceremonial burials, corporal segments, or sacrifices.

In order to minimize damage to the osteological collection, the Dirección de

Antropología Física requested that enamel samples be taken only from loose teeth (teeth not in the mandible or maxilla) and that sampling not affect the future research potential of the skeletal collection. I had initially intended to draw random samples proportionally from each age group, with slight oversampling of individuals between the ages of 15 to 30. As immigrants are typically young adults, I wanted to ensure sufficient coverage of this age range to be able to address whether immigrants in these age groups were dying soon after arriving to the city.

However, in order to comply with the conditions established by the Dirección de

Antropología Física, it was not possible to maintain this sampling strategy. Given that samples could only be taken from loose teeth and care had to be taken to ensure that those loose teeth, in fact, came from the skeletons with which they were associated, only a limited number of individuals meeting these criteria, who also had sufficient preservation to determine age and sex, were available. Furthermore, deciduous dentition was not sampled in order to preserve aging data for juveniles, and individuals with few teeth were not sampled in order to ensure that teeth from these individuals would be available for future osteological research, such as metric analyses. Attempts were made to select individuals of both sexes and all ages in order to test a cross-section of the population, but ultimately I had very little choice in selecting which individuals were included in the study. Obviously, this sampling strategy was far from ideal. As a result, some age groups are significantly underrepresented in the sample and other age groups are overrepresented.

260 The primary factors affecting inclusion in the sample were the presence of both aging and, for adults, sexing criteria and the presence of loose teeth, both of which should be independent of whether an individual was an immigrant40. Therefore, while age had an effect on inclusion in the study, a degree of randomness otherwise exists. In other words, individuals were not chosen for sampling based on differential mortuary treatments, exotic grave goods, or any other features that might have marked them as possible foreigners.

Analysis of samples

Samples were then taken to the Laboratory for Archaeological Chemistry at the

University of Wisconsin-Madison for analysis using the procedures described in Price et al.

(2004; 2002; 2001; 2000; 1998; 1994), Wright and Schwarcz (1998), and Wright (in press).

Tooth enamel was mechanically cleaned and separated from dentine using a dental drill fitted with a diamond bit. Strontium samples were sonicated in deionised water and again in 5% ultrapure acetic acid. After being left in the acid overnight, samples were rinsed in deionised water and then put in a drying oven for 24 hours. Samples were subsequently put in sterile silica glass tubes and ashed in a muffle furnace for eight hours at 825°C. Afterwards, samples were placed in ultrapure concentrated nitric acid, dried, and then put in ultrapure 2·5N hydrochloric acid. Strontium was isolated with cation exchange chromatography and analyzed using a thermal ionisation multiple collector mass spectrometer (Price et al. 2004; 2002; 2001; 2000;

40 It could be argued that if immigrants and native residents were buried in different parts of the habitational zone, or if they received different mortuary treatments, differential preservation might be at work, thereby resulting in the over or underrepresentation of immigrants in the sample. However, individuals identified as possible immigrants came from various parts of the habitational zone, and there are no differences in the burial treatments of local and possibly nonlocal individuals.

261 1998; 1994). Oxygen samples were soaked in a 1.5% sodium hypochlorite solution to remove organic material and then in pH4.5 acetic acid to remove diagenetic carbonates. Enamel samples were exposed to phosphoric acid in the Isocarb common acid bath carbonate device and the oxygen isotope ratios of the resultant gas were measured with a mass spectrometer (Wright and

Schwarcz 1998; Wright, in press).

Limitations and Confounding Factors

Limitations common to both strontium and oxygen isotopes

While strontium and oxygen isotope analyses can provide invaluable information on immigration to Cholula, they also come with certain limitations that should be addressed. First and foremost, by virtue of the theoretical basis of the methodology, strontium and oxygen isotopes impose a definition of immigrant that does not necessarily conform perfectly to those individuals of interest to the current investigation. For the purposes of this project, any individual who moved from outside the urban zone of Cholula into the city would be an immigrant, and their response to the urban environment would be pertinent in evaluating demographic patterns in New World cities. However, in strontium and oxygen isotope analyses, an immigrant or foreigner is anyone who comes from a geographic region that has an isotope ratio different from that of the local signature of Cholula.

While distinguishing between broad regions based on strontium values is often possible in Mesoamerica due to the amount of geological variation, individuals who have immigrated from around the vicinity of Cholula may not be readily distinguishable from the permanent

262 residents of the city if the underlying geology of their place of origin is similar. Individuals from the immediate hinterlands of Cholula will likely also be indistinguishable from those in the urban zone because they would have shared a similar resource catchment area. Furthermore, while the inter-quartile ranges of strontium isotope values tend to differ between regions, there are some regions that have the same isotope values and are, therefore, not distinguishable. The ranges of some areas also overlap, again creating confusion (Price et al. 2008). Thus, the resolution of the method is not sufficient to capture all immigrants. Oxygen isotope analysis is even more uncertain in this regard, as less research has been completed as this point establishing site ranges, and the data that do exist show considerable overlap among sites (Price et al. in press).

It will also not be possible to determine whether an individual migrated from a rural or an urban area. The demographic profile of rural immigrants and their reasons for migrating are likely to differ from those of urban immigrants. Furthermore, epidemiological conditions in the urban environment of Cholula could have differentially affected individuals emigrating from rural areas if the disease environment of their place of origin was dissimilar. While rural and urban migrants cannot necessarily be distinguished with isotope data, demographic research has shown that those born in rural areas are more likely to migrate than those born in urban areas, largely because of the perceived economic advantages of cities, and rural immigrants are of particular interest in the current project (Clark and Souden 1988: 19). However, it should also be kept in mind that individuals who migrate may do so multiple times, moving from a rural area, to a small city, and then on to a larger city, for example. If rural immigrants had already resided for some period of time at a smaller urban center before making their way to Cholula, their previous exposure to an urban environment would have implications for how they fared once in the city.

263 Problems associated with strontium isotope analysis

When interpreting strontium isotope results, the potential sources of dietary strontium are an important consideration, as imported food supplies can alter isotopic values in human remains. Strontium isotope ratios in skeletal material reflect local values because, for the most part, the food catchment area of a population is limited to its immediate environs. In

Mesoamerica, even in urban areas in which people are consuming food obtained from a market, the bulk of the diet is likely to reflect the isotopic values of the surrounding countryside because of the high energetic costs of transporting staple foods, like maize, long distances. In isotopic studies of high-status individuals, the impact of imported food sources is of greater concern since these individuals are more likely to have access to exotic goods. As the Cholula skeletons under investigation are of low status, it is extremely unlikely that a substantial portion of their diet is coming from outside the general area.

However, there are two mineral components to the Mesoamerican diet, salt and limestone, that could heavily affect isotopic values and that are often imported from some distance away.

Salt was mined in various parts of Mesoamerica including the Basin of Mexico and the Yucatan.

Sea salt, in particular, has a relatively high strontium content and strontium ratio (Price et al.

2008; Wright 2005b). Sufficient intake of imported salt could affect isotopic values. Similarly, limestone, which was commonly added to maize in prehispanic Mesoamerica, can strongly affect strontium isotope values in skeletal material (Wright 2005b). A deposit of limestone is located just to the south of Cholula and appears to be the most likely source of the mineral for all the city. The effects of salt and limestone on isotope values should not be terribly problematic in the current study, as a large enough sample is being tested that the ―local‖ strontium isotope

264 signature when dietary choices are taken into account should be apparent from the skeletons themselves. In other words, most individuals in this habitational zone are likely eating food from similar sources and in similar quantities, meaning that native residents should have similar isotope values, even if imported items such as salt or limestone cause them to deviate slightly from the local geologic signature.

Problems associated with oxygen isotope analysis

Oxygen isotope analysis is still in its early stages, and it is not yet clear how reliable it is for identifying immigrants. While oxygen isotope ratios do vary by geographic regions, a number of confounding factors influence oxygen ratios in biological tissues and interfere with the interpretation of local values. Both seasonal and annual variations in rainfall can result in greater than expected variations in oxygen isotope ratios in skeletons from a given site, particularly if the skeletal sample spans a long period of time (Price et al., in press). Standing water, such as in lakes or ponds, typically has a higher oxygen ratio due to evaporation of the lighter 16O. If these sources are used as the potable water supply, the oxygen isotope ratios found in the population may deviate from expected values. Similarly, rivers that originate in other geographical locales may affect the oxygen isotope ratios of a population and cause them to diverge from geographically-predicted values (Price et al., in press). A considerable amount of variation in oxygen isotope ratios between different teeth may also exist, even in a single individual. As enamel in individual teeth generally forms over several years, and different teeth form at different times, factors such as breastfeeding, climatic variation over time, and changes in diet or water supply may all affect oxygen isotope values derived from teeth (Price et al., in

265 press). At this time, these factors and their impact on the range of variability between sites are not sufficiently understood for oxygen isotopes to provide easily interpretable data on population movements. Price et al. (in press), in fact, caution that oxygen isotope analysis is still very much in its nascent stages and that its primary usefulness, at this point, may be identifying outlying isotopic values.

Results of Isotopic Analyses

Variation in the sample

Both strontium and oxygen isotope ratios for Cholula are presented in Appendix K, Table

K-141. Intrapopulation oxygen isotope values for animals and some human populations has been found to fall in a fairly narrow range – within 1 o/oo; however, intrapopulation variation in

Mesoamerica is somewhat greater, around 2 o/oo. This greater range of values is thought to result from the inclusion of both elites who ate imported food, as well as immigrants and foreigners in the Mesoamerican samples that have been tested (White et al. 2004). It is also likely that the confounding factors mentioned above are contributing to this variability. While the median values of sites from different regions are typically distinct, there is often considerable overlap in the ranges of oxygen values (Price et al., in press).

41 Hydroxyapatite in skeletons contains oxygen in both phosphate groups and carbonate groups, and both can be used to study migration. When the oxygen in phosphate groups is measured, it is reported relative to the Vienna Standard Mean Ocean Water (VSMOW) and is a value between 10 to 20o/oo in Mesoamerica. When oxygen in carbonate groups is measured, it is reported relative to the Pee Dee Belemnite (PDB) carbonate standard and is a value between -10 to 0o/oo. A rough conversion between the carbonate-derived and phosphate-derived oxygen isotope ratios by subtracting 21o/oo from the phosphate values (Price et al., in press).

266 Strontium isotope values in Mesoamerica have been found to fall into a range of less than

+/-0.001, both when local geological values and intrapopulation values are considered (Price et al. 2008). Larger sites, such as Teotihuacan, Kaminaljuyu, and Tikal, have larger ranges, again, presumably due to the inclusion of nonlocal individuals in the samples tested (Price et al. 2008:

169). The inter-quartile ranges of strontium isotope values tend to be somewhat more discrete than oxygen values, but some overlap does exist among regions (Price et al. 2008).

Table 8-2: Descriptive statistics for strontium.

Descriptive Statistics: Strontium

Statistic Std. Error

Strontium Mean .70566020 .000082378 95% Confidence Interval for Lower Bound .70549457 Mean Upper Bound .70582584 5% Trimmed Mean .70565319 Median .70571000 Variance .000 Std. Deviation .000576644 Minimum .704380 Maximum .707220 Range .002840

Interquartile Range .000690 Skewness .069 .340 Kurtosis .723 .668

267

Table 8-3: Descriptive statistics for oxygen.

Descriptive Statistics: Oxygen

Statistic Std. Error Oxygen Mean -5.6014 .12373

95% Confidence Interval for Lower Bound -5.8500 Mean Upper Bound -5.3528 5% Trimmed Mean -5.6759 Median -5.6500 Variance .765 Std. Deviation .87489 Minimum -6.97 Maximum -2.13 Range 4.84 Interquartile Range 1.02

Skewness 1.550 .337 Kurtosis 4.642 .662

Looking at isotope results in Tables K-1 (Appendix K), 8-2, and 8-3, greater than normal variation is present in both the strontium and oxygen isotope values for Cholula. Oxygen isotope values span a range of 4.84 o/oo, and the range of strontium values is 0.002840. This greater than normal variation in isotope values suggests that some nonlocal individuals are, in fact, included in the sample. The question then becomes how local individuals can be differentiated from immigrants, and unfortunately, there is no easy or foolproof way to define the outer limits of local ranges. To further complicate matters, the individuals who fall at either end of the range

268 of strontium values are not necessarily the same individuals who fall at the outer limits of the range of oxygen values. In a recent study, Price et al. (in press) also found a lack of correspondence between strontium and oxygen isotope data from skeletons from Copan. Are confounding factors causing some of the oxygen isotope values to fall far from the median even though these individuals are native to Cholula? Are individuals who have oxygen ratios at the tails of the distribution but who have strontium values close to the median truly immigrants, perhaps from a locale with a strontium signature similar to that of Cholula? What about the individuals that have a strontium value that appears to be nonlocal, yet an oxygen isotope value close to the median? Research into oxygen isotopes has simply not reached a point where satisfactory answers can be offered for these questions. Consequently, in the following discussion, the strontium results will be given priority, as strontium isotope analysis is better established than oxygen. These analyses should be considered preliminary and tentative until such time as more research is available on the use of oxygen isotopes to identify immigrants in

Mesoamerica and the local isotope values for Cholula are better defined.

269

Strontium Values 0.70750 0.70700 0.70650 0.70600 0.70550 0.70500 0.70450 0.70400 0.70350 0.70300 0.70250

Figure 8-2: Strontium values arranged in ascending order.

Oxygen Values 0.00

-1.00

-2.00

-3.00

-4.00

-5.00

-6.00

-7.00

-8.00

Figure 8-3: Oxygen values arranged in ascending order.

270 Variation by tooth type

While an attempt was made to preferentially sample from the first permanent molar, as it forms within the first few years of life, it was not always possible to do so. As a result, samples were taken from incisors (1), canines (2), premolars (9), first molars (24), second molars (10), and third molars (4). There were no significant differences (p> 0.05) in the mean strontium or oxygen values by tooth. Wright (2005a) also found that tooth position does not seem to have a significant effect on strontium isotope values. Other oxygen isotope studies have found variation in oxygen isotope values of teeth due to trophic level effects from breastfeeding (White et al.

2000; Wright and Schwarcz 1998; Price et al., in press). As the permanent incisors and first molar form prior to the age of weaning in ancient Mesoamerica (around three to four years old), these teeth were found to have, on average, an oxygen isotope ratio 0.7 o/oo higher than the permanent second and third molars, which form after weaning. The formation of premolars and canines begins prior to weaning and continues afterward, so White et al. (2000) suggest that they have an oxygen isotope value approximately 0.35 o/oo higher than second and third molars.

They, therefore, advocate adjusting oxygen isotope values accordingly. However, James Burton

(personal communication, 2010) has cautioned that as neither the exact age of weaning, nor the part of the tooth from which the enamel was taken, is known with certainty, it is not possible to determine if correcting for the effects of breastfeeding is even warranted. Furthermore, the figures cited by White et al. (2000) are merely an average of the difference noted between tooth types and may be influenced by a number of different confounding factors that are still not well understood (J. Burton, personal communication, 2010).

271 That being said, these corrections were applied experimentally to oxygen isotope values strictly to see how variability in the sample was affected. Incisor and first molar values were reduced by 0.7 o/oo and premolar and canine values were reduced by 0.35 o/oo. The adjusted values are presented in Table K-4 (Appendix K), Table 8-4, and Figure 8-4 for those interested in comparing the values with the strontium and unadjusted oxygen data. Adjusting for the effects of breastfeeding did not reduce variability in the sample. On the contrary, the range of oxygen isotope values increased slightly. Interestingly, applying this adjustment did increase the degree of correspondence between strontium and oxygen isotope results. However, given the early state of the research on oxygen isotope values and the fact that no significant differences were noted in the mean oxygen isotope values based on the tooth position, it is perhaps imprudent to place too much faith in these ―corrected‖ values. As it has not yet been demonstrated conclusively that these corrections represent a meaningful and necessary manipulation of the data, I will not be relying on adjusted oxygen values to identify immigrants in the current investigation.

272

Table 8-4: Descriptive statistics for adjusted oxygen isotope values.

Descriptive Statistics: Adjusted Oxygen

Statistic Std. Error Adjusted Oxygen Mean -6.0284 .13035

95% Confidence Interval for Lower Bound -6.2903

Mean Upper Bound -5.7665 5% Trimmed Mean -6.1266

Median -6.0700 Variance .850 Std. Deviation .92169

Minimum -7.22

Maximum -2.13 Range 5.09 Interquartile Range .94

Skewness 2.052 .337 Kurtosis 6.854 .662

Adjusted Oxygen Data 0.00

-1.00

-2.00

-3.00

-4.00

-5.00

-6.00

-7.00

-8.00

Figure 8-4: Adjusted oxygen isotope values arranged in ascending order.

273 Identifying local values

In regards to differentiating local from nonlocal values, a statistical approach to identifying immigrants provides the allure of an objective standard. We could, for example, designate everything within two standard deviations of the mean as ―local,‖ and anything outside of two deviations as ―foreign.‖ Price et al. (1994) suggested this method in an early article on strontium isotopes. Applying this approach, Skeletons 292 and 373-142 would appear to be immigrants based on the oxygen data and Skeletons 151, 245, and 292 would appear to be immigrants based upon the strontium data. The most obvious problem with attempting to identify non-local values in this manner is that even in a sample in which all of the individuals are local, 5% will still fall outside of two standard deviations. Wright (2005), in considering strontium isotope data from Tikal, suggests that using the criterion of two standard deviations from the mean as a way to define local isotope signatures may not be the most appropriate.

For the Cholula sample, using two standard deviations on either side of the mean results in a range of ―local‖ oxygen values of almost 3.5 o/oo – far greater than other documented

Mesoamerican sites. Similarly, two standard deviations yields a strontium range of +/- 0.00115 -

- again, a wider range of variation than has been observed elsewhere in the region. These larger- than-normal ranges would seem to suggest that immigrants are being included within those two standard deviations.

More recent publications of strontium and oxygen isotope results have relied on simple visual analysis of graphical data to spot potential nonlocal values (Wright 2005a; Price et al.

2000; Price et al., in press). A subjective graphical analysis of the Cholula data reveals a number

42 Care should be taken when interpreting the values for Skeleton 373-1. The oxygen signal for this individual was weak, and no strontium value could be determined.

274 of individuals that could have nonlocal strontium and oxygen ratios. Bar graphs of strontium and unadjusted oxygen isotope values are presented in Figures 8-2 and 8-3, scatter plots are shown in

Figures 8-5 and 8-6, Q-Q plots are shown in Figures 8-10 and 8-11, and extreme values are shown in Tables 8-5 and 8-6. A scatter plot of strontium against unadjusted oxygen values appears in Figure 8-8. Adjusted oxygen isotope values are shown in Figures 8-4, 8-7, 8-9, and 8-

12 and Table 8-7 for comparative purposes only.

The graphs of oxygen values indicate two clear outliers: Skeleton 292 and Skeleton 373-

1. Skeleton 292, a 56 year old male, has the highest strontium value (0.70722) as well as the highest oxygen value (-2.13). Both isotope values are so significantly outside the mean and median for this sample that he almost certainly grew up in a foreign location. Looking at the map in Figure 8-1, it would appear that similar strontium values are also found to the south of

Cholula near Oaxaca43. Once further mapping of oxygen values has been completed, it may be possible to verify whether his oxygen signature is, in fact, consistent with that region. Also of note in the case of Skeleton 292 is the fact that his third molar was sampled, indicating that he must have migrated sometime after its formation in adolescence. Skeleton 373-1, a 60 year old female, also has an oxygen ratio that appears to be an outlier (-3.16), but this value should be viewed with some skepticism since the oxygen signal for this individual was weak and no strontium value could be obtained.

Price et al. (in press) and James Burton (personal communication, 2010) suggest that until a better understanding of confounding factors is reached, oxygen isotope data are primarily useful in pinpointing extreme outlying values and should not otherwise be used to identify

43 In fact, Lagunas (1994) mentions in passing that artifacts from the Gulf Coast and Oaxaca were found during excavations of the Proyecto Cholula. However, I have not been able to identify what these artifacts were, where they were found, or to which time period they were dated. No grave goods were buried with this individual, however, so artifacts from Oaxaca were not directly associated with him.

275 nonlocal individuals in the population. Therefore, at this time, the oxygen isotope data will be used only to corroborate that Skeleton 292 is nonlocal and no further attempts will be made to identify immigrants based on the oxygen isotope values. The oxygen isotope data will be revisited when confounding factors are better understood.

Scatter Plot of Strontium Data 60

Individual 20

0 0.70400 0.70450 0.70500 0.70550 0.70600 0.70650 0.70700 0.70750 Strontium Ratio

Figure 8-5: Scatter plot of strontium data.

276 Scatter Plot of Oxygen Data 60

30 Individual 20

0 -8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 Oxygen Ratio

Figure 8-6: Scatter plot of unadjusted oxygen data.

Scatter Plot of Adjusted Oxygen Data 60

Individual 20

0 -8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 Adjusted Oxygen Value

Figure 8-7: Scatter plot of adjusted oxygen data.

277 Scatter Plot of Strontium Against Oxygen 0.70750

0.70700

0.70650

0.70600

0.70550

Strontium Values Strontium 0.70500

0.70450

0.70400 -8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 Oxygen Values

Figure 8-8: Scatter plot of strontium values against unadjusted oxygen values.

Scatter Plot of Strontium Against Adjusted Oxygen

0.70750 0.70700 0.70650 0.70600 0.70550

0.70500 Strontium Values Strontium 0.70450 0.70400 -8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 Adjusted Oxygen Values

Figure 8-9: Scatter plot of strontium values against adjusted oxygen values.

278

Figure 8-10: Normal Q-Q Plot of strontium values.

279

Figure 8-11: Normal Q-Q plot of unadjusted oxygen values.

280

Figure 8-12: Normal Q-Q plot of adjusted oxygen values.

Table 8-5: Extreme values of strontium.

Extreme Values of Strontium

Case Number Value Strontium Highest 1 49 .707220

2 48 .706810 3 47 .706800 4 46 .706540

5 45 .706200 Lowest 1 1 .704380 2 2 .704500 3 3 .704610 4 4 .704650 5 5 .704750

281

Table 8-6: Extreme values of unadjusted oxygen.

Extreme Values of Unadjusted Oxygen

Case Number Value Oxygen Highest 1 50 -2.13 2 49 -3.16 3 48 -4.44 4 47 -4.57 5 46 -4.70

Lowest 1 1 -6.97

2 2 -6.79 3 3 -6.77 4 4 -6.64 5 5 -6.57

Table 8-7: Extreme value of adjusted oxygen

Extreme Values of Adjusted Oxygen

Case Number Value Adjusted Oxygen Highest 1 49 -2.13 2 50 -3.16 3 12 -4.81 4 34 -5.14 5 21 -5.27 Lowest 1 42 -7.22

2 2 -7.21 3 5 -7.20 4 44 -7.12 5 7 -7.06

282 While the amount of variation in the sample would suggest the presence of immigrants, the graphs of strontium values offer no clear indications of where to draw the line between local and nonlocal values. Several values at either end of the range seem suspect because they fall so far from the median, but their continuous distribution makes it difficult to conclude with any degree of certainty that the values are, in fact, nonlocal. In order to better define local strontium values for Cholula, several future studies are planned. First, animal remains (primarily dog) that were found during the excavations of the Proyecto Cholula are currently being analyzed to corroborate local biological strontium values. Second, animal remains will be collected from around the catchment area of Cholula in order to determine the possible range of strontium values that might be found in the native population. This should help clarify whether the degree of variability that is present in the sample is the result of geological variation in the area or the inclusion of immigrants in the sample. Third, limestone from the source just to the south of

Cholula will be tested to establish how local strontium ratios might have been affected by the inclusion of this mineral in the diet. As these analyses have not yet been completed, I will proceed with a very cautious preliminary interpretation of the strontium isotope results with the caveat that these discussions are highly tentative and subject to revision as the results of additional studies become available.

In addition to Skeleton 292, the 56 year-old male mentioned above, Skeletons 213

(0.70654), 178 (0.70680), and 245 (0.70681) are possible nonlocal values on the upper end of the strontium range. On the lower end of the range, no obvious boundary exists, but Skeletons 151

(0.70438), 186 (0.70450), 294 (0.70461), 328 (0.70465), and 301 (0.70475) are possible nonlocal values (Table 8-12), and Skeletons 127 (0.70493) and 302 (0.70505) are questionable (Table 8-

13). If we arbitrarily designate a range of +/- 0.001 around the median (range = 0.70471-

283 0.70671) of the strontium isotope data in keeping with the typically-observed variation in other sites in Mesoamerica, most of these same individuals would fall outside of that range, suggesting that designating the above individuals as possibly nonlocal based upon the graphical analyses is not necessarily an overly liberal approach. Interestingly, the individuals that have the lowest isotope values in the sample fall into the strontium isotope range of the nearby Basin of Mexico, and several also have oxygen isotope values consistent with those found in the Basin.

Table 8-8: Individuals designated as possibly nonlocal based upon graphical analysis of the strontium data.

Time Adjusted Skeleton Age Sex Period ∂18O O 87Sr/86Sr

151 19.63 F II -6.79 -6.79 0.70438

186 40.85 F III -6.51 -7.21 0.70450

294 56.99 F II -6.42 -6.42 0.70461

328 17.5 F III -5.97 -6.32 0.70465

301 35.86 M II -6.50 -7.20 0.70475

213 7 ? III -5.91 -6.61 0.70654

178 48.13 F II -6.05 -6.75 0.70680

245 42.26 F III -4.79 -5.49 0.70681

292 56.65 M II -2.13 -2.13 0.70722

Table 8-9: Individuals designated as questionable based upon graphical analysis of the strontium data.

Time Adjusted Skeleton Age Sex Period ∂18O O 87Sr/86Sr

127 21.71 M II -5.16 -5.86 0.70493

302 31.91 M II -6.36 -7.06 0.70505

284 What do the Results Suggest about Immigration to Cholula?

In the following discussion, the nine individuals from Table 8-8 identified as having possible nonlocal strontium values will tentatively be considered immigrants. Note that this list includes Skeleton 292 who appeared as an outlying value in the oxygen isotope results as well.

As Skeleton 373-1’s outlying oxygen value is potentially the result of a weak signal, she will not be included in any calculations or discussions. Skeletons 127 and 302 (Table 8-13), deemed questionable based on their strontium isotope ratios, will not be included in the calculations below, but will be mentioned in the discussion in regards to how they might impact the interpretation of the data set if they are, indeed, immigrants.

The scale of immigration to Postclassic Cholula

In considering the scale of migration to Postclassic Cholula, it would appear that somewhere between nine and eleven individuals out of a sample of 50 (18-22%) were possible immigrants to the city. While this percentage cannot be interpreted as the percentage of immigrants in the population or the rate of migration, it does provide a very general idea about the scale of migration to Postclassic Cholula. The finding that possibly as much as 22% of the sample consisted of nonlocal individuals is not out of line with the percentage of immigrants identified from low-status contexts in other Mesoamerican urban centers. At Tlajinga 33, the low-status apartment compound in Teotihuacan, 29% of individuals sampled were identified as immigrants, and 30% of a sample of commoner burials from Copan were found to be of nonlocal individuals (White et al. 2004b; Price et al., in press). The isotope data hint that the scale of

285 immigration to the urban center was not insignificant during the Postclassic; however, it must be more firmly established that these individuals were, indeed, immigrants and that other residential areas of Cholula also experienced a similar influx of immigrants.

Mortuary context of potentially nonlocal individuals

Of the eleven individuals that have been identified as possible immigrants, nine are primary burials, one was a secondary burial, and one was a salvage burial44. What is most notable about the mortuary contexts of these individuals is the fact that there is absolutely nothing notable at all. In other words, there are no indications from the mortuary contexts of these individuals that they might be immigrants. They were all low-status individuals, and their burials were very much like those of native residents of Cholula. None of them were buried with any grave goods, but many native residents of Cholula were also interred with few or no offerings. The lack of grave goods does suggest, however, that immigrants to this habitational area enjoyed no special social status, unlike many of the immigrants identified at Tlajinga 33

(White et al. 2004b). It is interesting to note that individuals 301 and 302, who were excavated from nearby areas, both have strontium isotope values that are consistent with a similar place of origin. This could indicate that they lived in the same household during life or that this household had familial or other connections to the foreign region from which these immigrants came. The lack of archaeological indicators of foreign associations is interesting in that it suggests that a significant amount of migration in the prehispanic past may be invisible in the archaeological record.

44 López et al. (1976) do not clarify what a salvage burial is. I imagine that it refers to a burial that was disturbed, perhaps by archaeological excavations, before it could be confidently assessed as primary or secondary.

286 Characteristics of Possible Immigrants

As the number of individuals who have been preliminarily identified as nonlocal is only between nine and eleven, apparent differences between category variables are not statistically significant. General trends will be discussed with the obvious caveat that such a small sample size can provide only suggestions and not definitive conclusions. The uncertainty that exists in identifying nonlocal isotope values and the fact that it was not possible to proportionally sample by age only add to the tentativeness of any observable patterns. Juveniles are significantly underrepresented in the sample and young and middle age adults are overrepresented (Table 8-

10). Weighted data are shown in Tables 8-11 and 8-12 but should still be viewed with a critical eye due to the very small numbers of possibly nonlocal individuals. Consequently, I wish to emphasize that the following discussion should be taken as merely possible patterns that merit further research rather than firm conclusions. I will mention, where appropriate, additional avenues of investigation that could confirm or refute these observations.

Temporal differences

In constructing the age-at-death distribution for Cholula, skeletons from Cholulteca II and

Cholulteca III were combined, partly to increase the sample size and partly because of questions that have been raised about whether the ceramic sequence used to differentiate these two phases is correct. However, it is notable that skeletons dated to Cholulteca II make up a disproportionate percentage of the individuals identified as possible immigrants (Table 8-10, see

Table 8-11 for weighted values). While 34% of the 50 individuals included in the isotope

287 analysis were dated to Cholulteca II, 58% of the individuals tentatively designated as nonlocal are from this time period. Although the potential archaeological problems with distinguishing

Cholulteca II and Cholulteca III make conclusions regarding variability in migration rates over time suspect, it should be at least mentioned as a possibility. Given the archaeological data that suggest that Cholula increased in size from the Early to the Late Postclassic (Sanders 1971;

Müller 1973; Durmond and Müller 1972), high rates of immigration during Cholulteca II could account for some of this population growth. The inclusion in the analysis of Skeletons 127 and

302, the two additional individuals identified as possible immigrants, would only serve to increase this apparently disproportionate number of immigrants in Cholulteca II, as both of these individuals are dated to that time period.

Table 8-10: Characteristics of the collection as a whole, of the isotope sample, and of those identified as possibly nonlocal. Isotope Sr Time Period Collection Sample "Immigrants"

Cholulteca II 25.7% 34% 55.6% Cholulteca III 74.3% 66% 44.4%

Sex (Adults

only)

Male 50.3% 56% 25% Female 49.7% 44% 75%

Age

<15 45.7% 14% 11.1% 15-30 12.5% 28% 22.2% 30-50 8.2% 22% 44.4% 50+ 33.6% 36% 22.2%

288

Table 8-11: Weighted values of local and possibly nonlocal individuals by time period.

Percentage of Weighted Weighted Nonlocals by Time Time Period Local Nonlocal Period Cholulteca II 9.2 3.8 29.20% Cholulteca III 32.77 4.52 12.10%

The demographic characteristics of potential immigrants

In the isotope sample as a whole, males were very slightly overrepresented in relation to females, but six out of the eight adults identified as possible immigrants are female (Table 8-10).

This could, perhaps, reflect that women are immigrating to marry and that a preference for patrilocal postmarital residence patterns existed, which is consistent with ethnohistoric data

(Carrasco 1971; Olivera 1978). It is possible that both men and women may have been drawn to the city by economic opportunities, but, as mentioned in the previous chapter, we know little about what these might have been. Given that Skeletons 127 and 302 are both males, however, this discrepancy in the sex of possible immigrants may not be as pronounced at it seems.

As only enamel samples were included in the current analysis, we do not know at what age nonlocal individuals immigrated to Cholula. However, the presence of a 7 year-old child among the tentatively identified nonlocal values suggests that, at least in some cases, family groups might have immigrated. In an isotope study of the Oaxaca Barrio at Teotihuacan, White et al. (2004a) found considerable relocation of individuals to this ethnic barrio during infancy and early childhood. Economic considerations, in the form of either push (famine, for example) or pull factors, may have been the cause of immigration of family groups, although it is also

289 feasible that these relocating families could have included native Cholulans returning to the city with a foreign spouse and children after some time spent abroad.

If families with young children comprised a significant portion of the immigrants moving to Cholula, they could have had a major impact on the paleodemography of the city. A high rate of immigration may indirectly result in an increase in the number of infants in the mortality sample because immigrants, who are typically young adults, have high fertility rates. They thereby contribute to births, but as infants have the highest mortality rates, this increase in the number of babies born results in a concomitant increase in the number of babies that die (Paine

1997). Families who are immigrating with young children may also contribute to infant and early childhood mortality rates in a somewhat more direct way if said children die after reaching the city. If immigrating families experienced precarious economic conditions and, therefore, precarious nutritional conditions, the children, in particular, would have been vulnerable to infectious disease. Strontium and oxygen isotope studies of the deciduous dentition of children from Cholula could help determine how common migration was in these age groups, as these teeth form prior to birth and during the earliest years of life. It was not possible to sample deciduous dentition in the current study, but this line of inquiry could provide very valuable evidence about the scale of childhood immigration to the site. Also, recall that juvenile mortality fell off rather slowly in the Cholula collection (Chapter V). Identifying nonlocal individuals among these children could give some indication of how heavily immigrants contributed to mortality in these earliest age groups.

The ages at death of adult individuals identified as possible immigrants are interesting and, if they reflect a true pattern, potentially reveal other ways in which sociocultural factors may have helped shape urban population dynamics in Cholula. Table 8-10 shows the

290 percentages of the Cholula collection as a whole, the percentages of the isotope sample, and the percentage of possible immigrants that fall into particular age ranges45. Table 8-12 presents the weighted data, and Figure 8-13 shows the age distribution of the isotope sample compared to the age distribution of those tentatively identified as nonlocal. Individuals identified as possible immigrants appear to fall disproportionately in younger age groups, particularly the 30 to 50 age category, as these individuals comprised 22% of the isotope sample, but 44.4% of possible immigrants. I wish to restate here again, however, that the small sample size precluded statistical analysis of the data, so this difference has in no way been shown to be statistically significant.

We must first consider the possibility that the fact that potential immigrants seem to be disproportionately represented among adults aged 30 to 50 is purely the result of sampling or the means by which local and nonlocal isotope values were designated. The number of possible nonlocal individuals is so small that one individual more or less in any category could produce an entirely different picture. For example, the inclusion in the analysis of Skeleton 373-1, a 60 year old, would reduce this discrepancy. Several other skeletons falling at the ends of the oxygen isotope range were also older individuals. If some or all of these individuals were deemed to be immigrants, it could erase this apparent overabundance of potential immigrants in the 30 to 50 year old age category. In other words, the disproportionate number of possible immigrants in this age group may not reflect reality at all. To be certain that such a pattern truly existed, it would first be necessary to have a clearer idea of how to determine the boundaries of the ranges of local values for both oxygen and strontium data. As mentioned, ongoing isotopic investigations into the site may help clarify who is an immigrant to Cholula and who is not, but

45 These broad age categories were chosen based on the following rationale: Immigrants are most commonly adults in their late teens and twenties, thus the age category 15-30 is meant to reflect this. If individuals are dying soon after reaching the city, they should disproportionately appear in this age category. Mortality in the age range 30-50 is low in the Cholula collection as a whole, but it begins to rise around 50.

291 the results of these analyses are not currently available. Furthermore, a much larger sample of

Cholula skeletons, combined with a proportional sampling strategy, would be needed to identify a sufficient number of immigrants to allow for statistical analysis of the distribution of immigrants by age. Sample sizes could potentially be augmented somewhat by future isotopic studies, as additional skeletal material from Cholula does exist.

If it were demonstrated through further analyses that immigrants are disproportionately falling into the 30 to 50 age grouping, such a finding would have at least two potential explanations. One possible reason for this pattern could be that adult immigrants, in fact, had a lower modal age at death than native adult residents of Cholula. In Early Modern London, mortality rates of young adult immigrants in their teens and early twenties was high due to a lack of previous exposure to the epidemic diseases found in the urban environment. If immigrants to

Cholula were disproportionately represented in the 15 to 30 age category, it could be argued that something similar was occurring in this New World urban center, even in the absence of epidemic disease. However, a true overrepresentation of immigrants in the 30 to 50 age group would suggest a different dynamic at work.

Table 8-12: Weighted data indicating the number of local and possibly nonlocal individuals in each age

category.

Age Weighted Weighted Percentage of Nonlocal Category Local Nonlocal by Age Category < 15 19.56 3.26 14.30% 15-30 5.35 0.89 14.29% 30-50 2.59 1.48 36.36% 50+ 14.88 1.86 11.10%

292 Age Distribution of Isotope Sample and Identified Immigrants 6 5 4 3

Number 2 1 0

Age at Death

Sample Immigrants

Figure 8-13: Graph showing the age distribution of the isotope sample and the age distribution of those identified as possible immigrants.

If it were demonstrated to be true that, of those individuals reaching adulthood, immigrants to Cholula tended to die at younger ages than native residents, why might that be the case? Could the unhealthy environment of the city be the cause? I would, instead, argue that selective factors on migration, as well as perhaps social and economic factors negatively influencing the well-being of immigrants in their new communities, would have been more to blame for the younger modal age at death of adult immigrants than the epidemiological environment of the city itself. We do not know at what age these individuals immigrated to

Cholula, but studies of a wide variety of societies across time and space have demonstrated a very strong age-dependent pattern of migration (Rogers and Castro 1984). In childhood, migration reaches a peak in infancy and then declines until late adolescence, when it begins to rise sharply and reach a second, significantly higher peak, in the early twenties. It then declines

293 again, although in some societies a small, tertiary peak occurs in the sixties. Migrants are overwhelmingly young adults because these individuals have the fewest impediments to moving as well as the strongest incentives to relocate (marriage or economic opportunities). Infants and children may sometimes migrate along with their young adult parents, which creates the secondary peak seen in early childhood. Occasionally older individuals may also migrate to live with family during old age, but by and large, immigrants are in their late teens to early twenties.

Therefore, it seems unlikely that those immigrants who died in their thirties or forties had just arrived in Cholula. More probably, they migrated as young adults or possibly even as children.

The suggestion that many of these individuals immigrated well before death should also be tested, and I will briefly mention some possible avenues of future research that would permit this question to be addressed with more confidence. As bone remodels over time, strontium or oxygen isotope ratios obtained from bone can be compared to those obtained from enamel, which is laid down in childhood, to determine if someone is a recent immigrant. As discussed previously, if enamel values indicate that someone is an immigrant, but bone isotope ratios are consistent with the local signature, it indicates that migration occurred well before death. If both enamel and bone isotope ratios are significantly different from the local signature, it indicates a recent migration.

Bone samples have been collected from several of the individuals identified as possible immigrants and a strontium isotope analysis of this material is planned. The primary obstacle to strontium analysis of bone is contamination from the soil. If bone strontium values of these identified immigrants are significantly different from the local values for Cholula, they were recent immigrants. On the other hand, if bone values are consistent with local levels of strontium, it might be the case that the individual immigrated well before death, or that the bone

294 sample has been contaminated by local soil values. Thus, isotopic analysis of bone could help clarify the issue, provided that all of the values are nonlocal, indicating recent migration.

Otherwise, it will be thoroughly unhelpful.

A second possible line of inquiry would be to test multiple teeth from individuals identified as immigrants to determine if they moved during childhood or adolescence. As the teeth form at different ages, testing later forming teeth, such as the second or third molars could indicate if a person immigrated sometime prior to adulthood. However, if third molar values also indicate a foreign place of origin, it would suggest that the person immigrated after early adolescence. So, for example, if an individual who died at 35 years old was found to have a first molar value that was foreign but a second molar value that was local, it would indicate that the person immigrated in childhood and had resided at the site for a significant period of time before his or her death. If, however, the same individual was found to have first, second, and third molar values that all showed a nonlocal isotopic ratio, it would suggest that the person immigrated in adolescence or later, but we would not know if that event occurred years prior to death or the week prior to death.

These additional lines of isotopic inquiry will be pursued as a means of delimiting when these individuals immigrated to Cholula, and future isotope studies of Cholula skeletal material should also include such investigations. Their potential is somewhat limited, but they might provide a few scraps of additional evidence to address the timing of migration in relationship to the timing of death. However, for now, I will proceed on the assumption that individuals and small nuclear family groups immigrating to Cholula conform to the patterns of age-dependent migration discussed above. Such an assumption is consistent with ethnohistoric and archaeological data on individual migrations discussed in the previous chapter. Assuming, then,

295 that immigrants who died in their thirties and beyond moved to the city in their early twenties or before, the question becomes why these adults would tend to die earlier than non-immigrant adults in the population. If the supposedly unhealthy environment of the city were responsible, why would it have taken so long to kill them? It seems more likely that characteristics of the immigrants themselves would have been related to greater frailty and would have resulted in their early deaths.

Consider that migration is a self-selecting process. One possible explanation for the younger ages at death of adult immigrants to Cholula is that those individuals who migrated were spurred to do so by very poor economic conditions in their place of origin (Luu 2005; Clark and

Souden 1988; Whyte 2000). If these individuals experienced significant nutritional deficiencies in childhood and adolescence due to their impoverished states, such insults could have had long- term consequences with respect to frailty (Alter and Riley 1989). These individuals could have been unhealthy upon their arrival to Cholula46.

Regardless of the state of the immigrants’ health when they arrived in the city, the conditions the immigrants faced in their new community may not have been substantially better than in their homelands. We have no idea what economic activities these individuals might have engaged in upon settling in Cholula, but the low-status habitational units with which they were

46 Alternatively, we should also consider the possibility, presented in the previous chapter, that immigrants to Cholula might have been, on average, healthier individuals because they were able to undertake and survive the journey. Studies of modern immigrants to the United States and Canada have documented what is referred to as the healthy migrant effect, which is essentially the idea that immigrants are typically healthier than non-immigrants in the host country (Singh and Siapush 2001; Hyman 2004; Razum et al. 2000) . This effect has been ascribed to a number of different causes, some of which would not be applicable to Cholula, but one of the proposed causes is that healthier individuals are more likely to immigrate than unhealthy individuals. But if these individuals are healthier on average, why would they have died at younger ages than non-immigrant adults? In modern populations, the healthy migrant effect only seems to apply to recent immigrants. Those who have been in their new homes for some time have no such advantage and their health appears to deteriorate over time (Noh and Kaspar 2003). This apparent decline has been attributed to a number of factors including low socioeconomic status, stress levels associated with acculturation, and less access to health care.

296 associated indicate that immigrating to the Cholula did not result in significant improvements in their economic conditions. Moreover, although these individuals appear to have been accepted into the families of native residents, they do not appear to have been accorded any special or privileged status, as demonstrated by their lack of grave goods. While they may have been treated similarly to native residents, it is also possible that they faced some significant social or economic challenges being integrated into their new home. Although economic difficulties might be more readily accepted as an explanation for their higher mortality, social challenges immigrants face should not be dismissed as insignificant when pondering why these individuals might have had lower ages at death than native residents of Cholula. Studies of immigrants in modern societies have indicated that in addition to facing bleak economic conditions, immigrants may have little social support and considerable stress levels related to acculturation, discrimination, language barriers, and changing gender expectations among other things

(Thomas 1995; Ben-Sira 1997). All of these factors could potentially have been applicable to the immigrants of Postclassic Cholula as well. Chronic stress has, in turn, been linked to increased morbidity and mortality (Cohen and Williamson 1991; Cohen et al. 1991; Peterson et al. 1991).

Looking at the pathological lesions that are present in the individuals identified as possible immigrants (Table 8-13; also see Appendices H, I and J for the pathological lesions that could be assessed in each individual), of the eight that could be assessed for enamel hypoplasias

(everyone except 294), seven had enamel hypoplasias on multiple teeth. Only Skeleton 213, the seven year old child, did not have enamel hypoplasias on the permanent incisor and first molar that were present. The presence of this skeletal lesion suggests that these individuals experienced multiple incidences of growth faltering in childhood. Furthermore, of those who had several teeth present for evaluation, four, all of whom died before the age of 50, had enamel

297 hypoplasias on the second and/or third molar, indicating significant stress episodes occurring in late childhood and early adolescence. Recall from Chapter VI that enamel hypoplasias on all teeth except the canine were associated with increased mortality. The fact that these individuals have so many growth disruptions throughout childhood may indicate that they were severely economically disadvantaged growing up. For those who immigrated in late adolescence or later, this may mean that they were, in fact, motivated to immigrate by poor economic conditions in their place of origin. If they immigrated as young children with their parents, it suggests the conditions they experienced in Cholula were bleak.

A number of possible immigrants, particularly those who died at younger ages, had other pathological lesions associated with increased mortality. Skeleton 213 has lesions of porotic hyperostosis on the sphenoid, probably indicating scurvy (Ortner 1999). Proliferative lesions were also common in the potential immigrants. We do not know at what age those lesions occurred, and in most of the cases they are healed, but they are consistent with generally poor health in these individuals. I will point out here, though, that it is not possible to infer anything from comparisons of lesion frequencies between immigrants and nonimmigrants, so I will not attempt to do so. I, therefore, cannot conclude that immigrants were necessarily less healthy than native residents. This overview of the pathological lesions observed on possible immigrants is merely to point out that almost all of them had pathological lesions associated with an increased risk of death. The two immigrants that lived to be the oldest had comparatively fewer pathological lesions.

298

Table 8-13: Pathological lesions observed in individuals identified as possible immigrants.

Time Skeleton Period Age Sex Pathological Lesions Healed proliferative lesions on ulna, tibia, and fibula; Enamel 151 II 19.63 F hypoplasias on canine, M1, and M2

Arthritic lipping of both radii and right ulna; Enamel hypoplasias 178 II 48.13 F on M3

Large dental abscess; Healed proliferative lesions on tibia; 186 III 40.85 F Enamel hypoplasias on incisor and canine

Possible porotic hyperostosis on endocranial surface of 213 III 7 ? sphenoid

Healed depression fracture on occipital; Active and healed proliferative lesions on clavicles; Healed proliferative lesions on ribs, ulna, tibias, and fibulas; Enamel hypoplasias on incisor, 245 III 42.26 F canine, M1, M2, and M3

Dental abscess; Extensive arthritic lipping; Enamel hypoplasias 292 II 56.65 M on incisors and canines

294 II 56.99 F Healed proliferative lesions on tibia

Healed proliferative lesions on tibias and fibulas; Enamel 301 II 35.86 M hypoplasias on incisor

Healed proliferative lesions on femurs, tibias, and fibulas; Healed cribra orbitalia; Enamel hypoplasias on incisor, canine, 328 III 17.5 F and M2

One additional possibility should be mentioned that might explain the apparent overrepresentation of tentatively-identified immigrants in the 30 to 50 year old age category and their apparent underrepresentation in the 50+ age category. Some immigrants to Cholula who survived into old age might have chosen to return to their place of origin to live out their final years. How common such a move would have been in prehispanic Mesoamerica would be

299 difficult, if not impossible, to determine. The isotope study of Tlajinga 33 did find one older woman who had foreign bone values, which would indicate that she moved to the apartment compound shortly before her death (White et al. 2004b). White et al. (2004b) propose that she may have been a widow moving back to her childhood home, although they acknowledge that other scenarios are also possible. It is technically feasible to identify immigrants who have returned to their homeland after living abroad – they should have foreign bone values47 but local tooth values. Thus, we could confirm that such movements did happen in the prehispanic past, even if we were not able to say how common such migrations were.

Although the isotope data presented in this chapter suggest that some immigrants are most likely included in the Cholula collection, identifying nonlocal values is a challenging proposition. Based on a graphical analysis of the strontium isotope data, it would appear that somewhere between nine and eleven individuals were possible immigrants to the site. Further isotopic studies are underway to better define local strontium values and, thereby, determine whether the preliminary interpretations offered here are valid. If it can be demonstrated that these individuals were, in fact, immigrants to the city, the effects of immigration on the paleodemography of Cholula must be considered.

Unfortunately, the small sample size does not allow any definitive conclusions to be drawn, but it does provide some directions for future research that might allow us a better understanding of urban population dynamics in Postclassic Cholula. Specifically, the possibility

47 Contamination of bone would be a potential problem.

300 that families with young children made up a fair portion of immigrants to the site should be explored, as these children may have affected juvenile mortality rates. If so, it would suggest that the demographic characteristics of immigrants to Cholula influenced patterns of mortality in the city. The current isotope study also hints that immigrants might be overrepresented among individuals between the ages of 30 and 50, perhaps indicating that these individuals were frailer than native residents due to either the conditions they experienced prior to immigrating or the conditions they experienced once they arrived in Cholula. Alternatively, this pattern could have been produced by immigrants over 50 returning to their homelands. Should this observation prove to be true, it would underscore the importance of considering the interplay of social, economic, and biological factors in thinking about urban population dynamics.

301 CHAPTER IX

CONCLUSIONS AND FUTURE RESEARCH

Given the complexity of urban demography and the vagaries of studying human skeletal material, a definitive understanding of urban population dynamics in Postclassic Cholula cannot be acquired from a single study. The very nature of paleodemographic samples means that many of the details of the demographic experiences of past populations are simply out of reach.

Moreover, a number of serious methodological problems in paleodemography and paleopathology have yet to be resolved. While significant progress has been made on this front, we are still limited in our pursuit of some lines of inquiry. That being said, by using the methods that are currently available, and by incorporating data about human demography from archaeological, ethnohistoric, and isotopic studies, we can begin the arduous process of teasing out bits of information about the demographic experiences of past populations from skeletal remains.

The current study offers a few new insights into morbidity, mortality, and immigration in

Postclassic Cholula and highlights some avenues of future research that must be pursued in order to clarify demographic patterns in this New World urban center and how they might compare to those of preindustrial Old World cities. Many of the demographic features of preindustrial Old

World cities resulted from the particular culture histories of those societies. Preindustrial urban population dynamics cannot be summed up by considering only the density of the population and the problems that potentially come with numerous people living within a restricted area. Rather, political, social, and economic factors come into play, and must be thoughtfully considered when

302 attempting to reconstruct demographic patterns of preindustrial cities. Prehispanic urban centers in Mesoamerica were different from Early Modern London and other European cities not only in regards to the epidemiological environment, but also with respect to their social, economic, and political organization. Even within Mesoamerica, a great deal of variation existed in urban centers. The reconstruction of urban population dynamics requires that the effects of cultural variables on the demography of a city be taken into account. It is quite possible that there are no overarching patterns of morbidity and mortality that extend to all cities across time and space.

Instead, health in the urban environment may have been affected as much by the culture histories of individual societies as by high population densities.

In Chapter III, I discussed some of the ways in which New World urban centers may have differed from those of the Old World, and I described some of the variation within New

World cities. Particularly in regards to population densities, urban centers in Mesoamerica display a range of settlement types, from the densely-settled Tenochtitlan-Tlatelolco to the more dispersed lowland Maya centers. The more dispersed settlement pattern that existed in many

Mesoamerican cities could have potentially resulted in fewer challenges with regards water contamination and sewage disposal. Moreover, many Mesoamerican urban centers would not have been as dependent on markets because part of the population residing within the city proper was engaged in agricultural activities. With respect to Postclassic Cholula, further archaeological investigation at the site could significantly enhance our understanding of those factors that may have contributed to urban morbidity and mortality. We currently have relatively limited data on issues such as settlement densities, economic specialization and the degree of dependency on the market, and water supplies and waste disposal systems. Additional archaeological information on the urban environment of Cholula would provide a richer context

303 in which to interpret data on morbidity and mortality in the city gleaned from paleodemographic and paleopathological studies.

While our current knowledge regarding the urban environment of Cholula is not as extensive as we might like, the skeletal collection excavated from low-status Postclassic habitational units near the Great Pyramid provides an opportunity to begin an initial assessment of how urbanism influenced morbidity, mortality, and immigration in Cholula.

Paleodemography and paleopathology were used in conjunction with strontium and oxygen isotope analyses to reconstruct demographic features of the Postclassic city. Some of the methodological difficulties that have hindered other paleodemographic studies have been avoided through the use of several new methodological techniques.

In particular, the age-at-death distribution of the population of Postclassic Cholula, constructed using transition analysis, is possibly one of the most significant findings of the current project. Unlike most other paleodemographic age-at-death distributions in which few individuals live into old age, the age-at-death distribution for Cholula indicates that more than half of those who survived to adulthood in the population lived past the age of 50. A comparison of this age-at-death distribution with several constructed using traditional aging methods provides additional evidence that age mimicry of the reference collection, rather than different patterns of mortality in the past, is the primary cause of the supposedly high young adult mortality observed in the vast majority of skeletal collections.

In addition to demonstrating that survival into old age did occur in Cholula, the age-at- death distribution has several features that are of interest in comparing urban patterns of mortality in Old and New World cities. High infant and early childhood mortality were important characteristics of preindustrial European cities. Unfortunately, infant mortality is

304 typically difficult to assess in paleodemographic samples because of the underrepresentation of this age group. The Cholula collection does appear to have fewer infants than would be expected given levels of early childhood mortality. An attempt was made to estimate the number of

―missing‖ infants using the Siler model, but the produced estimation was still lower than expected given the state of preservation of the material in general and the obvious disturbances of older burials by newer interments. Consequently, little can be said about levels of infant mortality in Postclassic Cholula at this time.

However, two other features of the age-at-death distribution of Cholula warrant further examination. Juvenile mortality falls off very slowly, which could possibly be the result of a high population growth rate or the confounding influence of young families immigrating to the city. The fact that mortality is low in young adults is also significant to the current study as high mortality in this age group in other paleodemographic investigations (quite possibly attributable to age mimicry) has been interpreted as being the result of high mortality in young adult immigrants.

The suggestion has been made that preindustrial cities were such unhealthy environments that they had intrinsic negative growth rates and only increased in size due to immigration from the surrounding countryside (Wrigley 1969). Unfortunately, the relatively small size of the skeletal sample under study from these habitational units precluded application of the Rostock protocol and, therefore, an estimation of the growth rate. Consequently, it was not possible in the current investigation to test whether the so-called Law of Urban Natural Decrease applied to

Cholula or whether the slow decline observed in juvenile mortality is the result of a high population growth rate. Archaeological data do suggest that the city was growing over the

Postclassic period and that it was at its maximum size at the time of the Spanish arrival (Sanders

305 1971; Durmond and Müller 1972; Müller 1973; Uruñuela and Plunket 2005). Thus, the limited information that we have on Cholula would seem to indicate that the city was experiencing positive population growth; however, it is not possible to specify whether this growth was attributable to a positive intrinsic growth rate, significant rates of immigration, or both.

Whether or not the Law of Urban Natural Decrease applies to Postclassic Cholula, understanding urban population dynamics requires more than knowledge of the growth rate.

Information about age-specific fertility and mortality rates for both native residents and immigrants is needed as well, as several studies of the demography of preindustrial Old World cities have demonstrated (Galley 1998; Landers 1993; Sharlin 1981). Paleodemography does not provide that degree of resolution, unfortunately, but as immigrants played a fundamental role in the urban population dynamics in Old World cities, strontium and oxygen isotope data were used to obtain more information about how they might have contributed to the demography of this

New World city as well.

The isotope study, in fact, provides possible evidence that immigration was not an insignificant force on the demography of Cholula during the Postclassic. We cannot extrapolate the rate of migration to the city from the isotope results, but a sufficient number of potential immigrants were identified in the sample that we can surmise that immigration was not a rare occurrence. If, in fact, the apparent temporal difference in the number of migrants is confirmed through further studies, immigrants may have even been partly responsible for the growth of

Cholula over the Postclassic period. Thus, the contributions of immigrants to fertility and mortality in the city should be evaluated. In Early Modern London, marital fertility among immigrants was lower than among native residents of the city because the social and legal milieu into which they entered resulted in their marrying later (Finlay 1981). The preliminary isotope

306 data from Cholula suggests that while both men and women migrated to the city, women may have immigrated more frequently. Ethnohistoric sources (Carrasco 1971; Olivera 1978) indicate that patrilocal postmarital residence patterns were preferred in the early colonial period and that the majority of immigrant spouses in Tecali, a town near Cholula, were female. Therefore, many of the female immigrants to Cholula may have moved there for the expressed purpose of marrying. Young couples may also have been among the immigrants to Cholula, as a child is included in the possible nonlocal strontium values. Immigrants to Cholula, therefore, might have contributed appreciably to the crude birth rate of the city.

Just as immigrants to Cholula would have had an effect on fertility, they would also have had an effect on mortality. Paine (1997) has demonstrated that high rates of immigration result in an increase in infants in the mortality sample because of the increase in the number of births.

Immigrants, being primarily young adults, generally have high fertility rates. As infant mortality is high in preindustrial societies, an increase in births translates into an increase in the number of infants that die. As infants are underrepresented in the Cholula osteological collection, we are not able to directly verify if this was the case, but if future studies confirm that young adults and couples were migrating to Cholula, it would be quite reasonable to assume that they were affecting the proportion of infants in the mortality sample.

Juvenile mortality, which typically falls off fairly sharply in early childhood, declines more slowly in the Cholula paleodemographic age-at-death distribution. While it may be related to a high population growth rate, as mentioned, the confounding effects of immigration on this portion of the age-at-death distribution should also be considered. The strontium isotope data suggest that immigrants to the site might have included families with young children, which is consistent with isotopic studies of immigration at Teotihuacan (White et al. 2004b). We do not

307 currently know how common it was for children to immigrate to Cholula because of an underrepresentation of juveniles in the isotope study, but if families frequently moved with infants and young children, these immigrants were almost certainly contributing to the number of infant and early childhood deaths in the city. Additional strontium isotope analysis of the deciduous dentition of juveniles would provide more information about migration in early childhood in prehispanic Mesoamerica, so we could then surmise how immigration affected juvenile mortality in Cholula.

The way in which immigration might have impacted adult mortality rates is far less clear.

In Early Modern London, the mortality rates of adolescent and young adult immigrants was higher than those of native residents of the same ages because many rural immigrants had never been exposed to the epidemic diseases present in the urban environment. The age-at-death distribution of Cholula indicates low mortality in these age groups, but lower rates of migration to Cholula might make higher mortality in immigrants undetectable in the paleodemographic data. Additionally, because mortality in adolescents and young adults is generally very low, immigrant mortality would probably have to have been astronomical to have noticeable effects on the paleodemographic data. The isotope study hinted that immigration to the city, while perhaps not as high as in Early Modern London, should not be dismissed in reconstructing urban population dynamics in Cholula; however, it did not provide enough information to clarify how the mortality rates of adult immigrants might have compared to those of native residents.

It appeared from the ages of individuals tentatively identified as nonlocal that immigrants could be disproportionately represented in the 30-50 year old age group. Of course, given the small sample size, such an observed pattern could be due merely to chance. In addition, the way in which the strontium and oxygen isotope data are interpreted affects who is determined to be an

308 immigrant. Because factors confounding the interpretation of oxygen isotope data cannot be adequately controlled for at this point, a decision was made to limit the use of oxygen isotope values in identifying immigrants. However, several of the individuals at the ends of oxygen isotope range are over 50 years old. If these individuals are indeed immigrants, the apparent overrepresentation of individuals between 30 and 50 years old in the immigrant sample would become considerably less apparent. Clearly, a larger isotope study of skeletons from Cholula is needed, as is a better understanding of what constitutes local isotope values for the site. While greater control over factors that can confound the interpretation of isotope data awaits the research findings of ongoing studies into these techniques, it may be possible to more strictly define local strontium and oxygen isotope values for Cholula. The means by which this objective could be accomplished will be discussed in more detail below.

Another possibility is that the apparent overrepresentation of 30 to 50 year olds among possible immigrants is caused by older immigrants returning their birthplaces. If this was commonplace in prehispanic Mesoamerica—and at the present time we do not know that it was – it would be an additional example of how cultural practices can affect urban demography.

Identifying whether this was a common occurrence in Mesoamerica would be difficult, however.

Individuals who have lived abroad and then returned to their homelands can potentially be identified by comparing the isotopic values of bones and enamel, although contamination can make results equivocal. Of course, individuals who spent their adulthood in Cholula and returned to their place of origin in senescence are not in skeletal collections from Cholula.

Perhaps the only way to assess frequency would be to identify at a variety of sites the number of native residents who spent their lives abroad and then returned in old age in order to generate a broad estimate of how common such an occurrence was.

309 If we assume for the moment that the disproportionate number of individuals between the ages of 30 and 50 among those identified as possible immigrants is truly reflecting a lower modal age at death for adult immigrants in comparison with native adult residents of Cholula, such a mortality pattern could be due to the greater frailty of immigrants. Although we do not know at what age these individuals moved to Cholula, a very strong cross-cultural, age-dependent pattern of migration would indicate that they most likely migrated as children or young adults (Rogers and Castro 1984). I will attempt to substantiate this assumption with further isotopic analysis of bone samples from these individuals, which can indicate if someone has moved recently.

However, contamination can result in strontium isotope studies of bone yielding equivocal results, so it may not be possible to verify that these individuals had resided in Cholula for some time before their deaths. If a large percentage of adult immigrants were moving to Cholula between the ages of 30 and 50, it would be a noteworthy finding in and of itself, as it would indicate a different pattern of migration than has been observed in other societies.

Proceeding, for the moment, with the assumption that these potential immigrants had been at the site for many years prior to their deaths, the fact that migration is a self-selecting process may have played a large role in them having higher age-specific mortality than native residents of Cholula. It could be the case that these individuals were largely motivated to immigrate by extremely poor conditions in their homelands and that health insults experienced in childhood as a result of such deprivation had lifelong effects on their frailty. If, in addition, they faced bleak economic circumstances in Cholula (and it would appear that they did), lack of a social network, or difficulties integrating into their new community, it may have resulted in their earlier deaths. The pathological lesions noted on the skeletons of possible immigrants suggest that these individuals did, indeed, experience stress episodes during childhood severe enough to

310 cause enamel hypoplasias, and most of them also had proliferative lesions on bones that the paleopathological analysis demonstrated to be associated with an increased risk of death.

However, this explanation is highly speculative and hinges upon the assumption that immigrants are disproportionately represented in the 30 to 50 year age group because they truly had higher mortality than native residents, which has yet to be demonstrated conclusively.

It should also be remembered that isotope analysis will not allow us identify immigrants that came from the immediate hinterlands of Cholula, as these individuals would have had a similar food catchment area as the residents of the city, nor will it allow us to determine whether individuals came from rural or urban areas. As rural immigrants may have responded differently to the environment of the city than did urban immigrants, the inability to distinguish between the two groups is, indeed, unfortunate. We also cannot identify with certainty the places of origin of those individuals identified as immigrants. Potentially, how far these individuals migrated could be related to frailty and mortality. Long-distance migrants are less likely to have been very frail individuals, and they quite possibly moved a number of times before arriving in Cholula.

Furthermore, those who immigrated very long distances may have had different motivations for migrating than those who came from more nearby areas.

Avenues of Future Research

Given the complexity of human health, the number of confounding factors at work in an urban environment, and the methodological limitations inherent in paleodemographic investigations, developing a clear picture of population dynamics in preindustrial New World cities will require considerably more research. Although the current investigation has provided

311 some limited insights into urban population dynamics in Postclassic Cholula, no firm conclusions can be drawn about morbidity or mortality or the effects of immigration on the demography of the city. However, several potential avenues of future research, which could contribute to our knowledge of the demography of Postclassic Cholula, have been illuminated.

What do we need to know in order to better understand urban population dynamics and how can we go about obtaining that information using the tools currently at our disposal? I would suggest that attempting to approach the question purely through the study of skeletal material is likely to yield incomplete and unsatisfactory results. While paleodemography and paleopathology is certainly an important part of the equation, research into urban demography in preindustrial New World cities must be a multidisciplinary effort in which archaeology, ethnohistory, and archaeological chemistry also come into play. Below I consider some avenues of investigation that could help us gain a clearer understanding of morbidity, mortality, and immigration in Cholula, as well as other preindustrial New World cities.

Paleodemographic issues

Due to an insufficient number of skeletons with aging information, it was not possible to implement the Rostock protocol in the current paleodemographic investigation of Cholula. From a purely methodological standpoint, an adherence to the Rostock protocol would have two beneficial consequences. First, it would permit the age-at-death distribution of Cholula to be estimated free of any biasing effects that the assumption of a uniform prior distribution might have caused. Producing an age-at-death distribution free of any methodological biases is, of course, essential to comparing mortality across populations. Otherwise, it becomes impossible to

312 sort out true similarities in the mortality experience of past populations from common methodological biases that are influencing the age-at-death distributions in similar ways. If our ultimate goal is to compare mortality across urban centers or across different types of societies, we must ensure that our paleodemographic age-at-death distributions are purely a reflection of the demographic processes that occurred in the population under study for such comparisons to have validity.

Second, adherence to the Rostock protocol would also allow the issue of nonstationarity to be addressed. In other words, the growth rate of Cholula could be estimated from the paleodemographic sample. In order to do this, the following equation is substituted for Pr(a) in the likelihood equation used to estimate the age-at-death distribution of the target population:

where f(a) is the age-at-death distribution of the target population, a is age, and r is the growth rate (Wood et al. 2002). Armed with information about the growth rate, the slow decline in juvenile mortality seen in the Cholula osteological sample would be more readily interpretable.

In addition to providing information necessary to an accurate interpretation of the age-at-death distribution of Cholula, an estimate of the growth rate is significant in evaluating how urban population dynamics in this New World city compare to those of preindustrial Old World cities.

The idea of natural urban decrease has dominated investigations into the demography of preindustrial Old World cities (Wrigley 1967; DeVries 1984; Perrenoud 1975, 1978), but more recent studies have demonstrated how short-sighted this vision of preindustrial urban population dynamics truly is (Galley 1998; Landers 1993). However, as this model has been suggested to

313 apply to the Mesoamerican urban center of Teotihuacan (Storey 1992), its applicability to

Cholula should at least be investigated. A starting point for determining whether natural urban decrease occurred in ancient Cholula would be to verify if the city was, in fact, experiencing population growth over the course of the Postclassic as the limited archaeological data that are available would suggest. In order to realize this goal of estimating the growth rate of Cholula from the osteological data, it will be necessary to increase the size of the skeletal sample. This is a feasible task as additional skeletal material from the Proyecto Cholula does exist that was not included in the current analysis (Ferré and Xalpa 1967). Other skeletal material has also been excavated from residential areas of Cholula (Uruñuela 1989; McCafferty 1992), although these collections are significantly smaller. Further archaeological investigation into population dynamics in Postclassic Cholula would be a significant contribution in this regard as more precise settlement survey data could provide independent estimates of growth rate to corroborate those derived from the paleodemographic data. Unfortunately, modern encroachment is makes more accurate settlement data less and less likely with every passing year. Of course, while the growth rate is an important piece of the puzzle, it does not give us the full picture, as it does not tell us how immigrants contributed to fertility and mortality in Cholula. While discerning age- specific fertility and mortality rates for immigrants is beyond current paleodemographic sample sizes and methodology and, in fact, may never be a possibility, a more general understanding of immigration within prehispanic Mesoamerica would enhance our ability to interpret the paleodemography of New World urban centers.

314 Immigration

In attempting to reconstruct demographic events in preindustrial cities using skeletal collections, it is useful to first consider why knowledge of immigration is necessary to this endeavor, what we would like to know about immigration, and how we can go about learning this information, keeping in mind that methodological or practical constraints may limit our pursuit. Immigration is a significant force in shaping the demography of modern and historic cities, yet immigration has received little attention from paleodemographers (see Boldsen 1984 and Paine 1997 for exceptions), perhaps because it is so difficult to account for in the archaeological record. Advances in isotopic techniques in recent years have drawn the issue of migration into the limelight, but for the most part these techniques have been used to examine issues related to culture contact rather than considering the broader demographic implications of migration. As immigration can have confounding effects on the age-at-death distribution of a population (Paine 1997), interpreting the demography of urban centers will first require an understanding of the scale of migration and the demographic characteristics of migrants. The scale of immigration into urban centers is, of course, important because it gives us some idea of how significantly mortality and fertility in the city would have been affected by incoming migrants. If rates of migration are low, the contribution of immigrants to the demography of the city may be essentially negligible, particularly from a paleodemographic perspective. On the other hand, if more substantial rates of immigration were occurring, migration as a force shaping the demography of New World urban centers cannot simply be ignored.

In considering the impact of immigration on urban population dynamics, knowing the demographic profile of immigrants can aid paleodemographers in determining how immigrants

315 may have contributed to fertility and mortality. For example, were males and females equally likely to migrate, or were immigrants predominantly of one sex? A substantial imbalance in the sex ratio caused by immigration of only one sex into the population could have implications for fertility. Which age groups were most likely to migrate? Did immigrants include mostly unmarried young adults, or were a substantial number of families with young children included among the immigrants? Such information is crucial for considering possible confounding effects of immigration on the age-at-death distribution.

Finally, understanding the reasons that individuals immigrated to the city can help paleodemographers assess the fertility and mortality of immigrants. If individuals were primarily moving to the city in order to marry, their contribution to fertility may differ from individuals who immigrate in order to beg in the streets, for example. Moreover, given that immigration is a selective process, having some knowledge of the reasons that individuals chose to migrate gives the paleodemographer insight into the frailty of these individuals, which, in turn, can be informative about the mortality levels of immigrants.

While the current study provides some clues as to the scale of migration to Cholula and the demographic characteristics of immigrants, further investigations are required to clarify these issues. The ethnohistoric documents and archaeological evidence of migration that were considered in Chapter VII confirm that group migration was fairly common in the prehispanic past, but these sources indicate very little about the migration of individuals or families. It could be the case that individual migrations were, in fact, uncommon in prehispanic Mesoamerica, perhaps owing to the political economy; however, it is also possible that the movement of individuals, particularly low-status individuals, was simply not considered to be worthy of mention and, therefore, went largely unrecorded in ethnohistoric sources. Archaeologically,

316 individual migrations of commoners are also likely to go unnoticed since they may have little or no impact on the archaeological record. Even the burials of these individuals may not bear indications that they are immigrants, as isotopic studies of individuals at both Tlajinga 33 and now Cholula have demonstrated.

Colonial-period documents such as ecclesiastical records may provide additional information about migration due to marriage, as Olivera’s (1978) study of marriages in Tecali demonstrates. Studies of other colonial documents such as tax records, wills, and transcripts of legal proceedings are a potential font of data on individual migrations that has not yet been explored. These documents may be particularly informative about individuals migrating for economic reasons. Furthermore, it may be possible to identify general characteristics of migrants such as sex and age, and whether they are immigrating alone or with family members from such documents. Of course, the caveat with relying on early colonial documents to learn about migration is that Spanish influences and the upheaval of the Conquest may have significantly altered prehispanic patterns.

An assumption is being made here that individuals would have immigrated to prehispanic Mesoamerican urban centers in order to pursue economic opportunities because demographic studies of modern and historical cities indicate that economic considerations are one of the primary reasons individuals choose to immigrate. However, the ethnohistoric sources examined in the current study do not broach this subject. Considering the nature of the political economy in prehispanic Central Mexico, further research into prehispanic and early colonial documents is essential to establish whether individuals would, in fact, have migrated to urban centers to pursue economic opportunities, and, if so, what kind of economic activities they would have engaged in and how they would have become involved in said activities. Particularly in

317 smaller urban centers where much of the population was involved in agricultural activities and craft specialization was pursued on a part-time basis, ideas about immigration established from studies of modern and historical cities may not directly apply.

Strontium and oxygen isotopes provide a means of directly investigating immigration in prehispanic populations, but these methods lack the degree of resolution we would ultimately like to have, and some methodological challenges have yet to be resolved. The variation in the isotopic results for Cholula suggest that immigrants are included in the sample, but to more accurately determine how many individuals are nonlocal, it will be necessary to more narrowly define local isotope values for Cholula. Additional strontium isotope studies of animal remains are underway to identify how much variation in strontium isotope values can be expected in native born residents of the city. Archaeological animal remains recovered during the excavation of the skeletal material used in this study will be analyzed, as will snail shells collected from the catchment area of Cholula. These strontium values should more clearly delimit the local signature for the site, and, thereby, allow immigrants to be identified with greater confidence.

As a result, the scale of migration and the demographic characteristics of immigrants should be clarified somewhat. However, it should be remembered that individuals coming from rural locales within the catchment area of the city will still go undetected in the sample, and the possibility exists that the demographic characteristics of these individuals were different from those individuals immigrating from somewhat farther away.

In order to gather more information about the age at which individuals were immigrating to Cholula, a strontium isotope study of bone samples from individuals identified as potentially nonlocal will be completed. While the possibility of contamination of bone samples limits the utility of this type of analysis, it can indicate if an individual immigrated shortly before his or her

318 death and, thus, provide data on the approximate age at migration. Of greater utility would be an additional strontium isotope study aimed at identifying immigrants among infants and juveniles in the Cholula skeletal sample. Such a study would confirm whether young families commonly immigrated to the city. One cautionary note would have to be observed when interpreting the results of such a study, however. Many individuals who immigrated to Cholula as children would have survived childhood and would not have entered the mortality sample until much later in life. Therefore, the presence of juveniles identified as immigrants would indeed verify that young families were immigrating, but these individuals who died as children would represent only a portion of those individuals who immigrated to Cholula early in life.

Ethnohistoric sources and isotopic analyses offer valuable information on the nature of migration to prehispanic New World cities. Obviously, however, they do not provide the kinds of detailed data on immigrants that paleodemographers would ideally like to have.

Unfortunately, reconstructing demographic processes in past populations without the aid of detailed records of vital events is always going to involve a lack of precision. Ultimately, a comparison of the age-specific mortality rates of native residents and immigrants to Cholula and other New World cities would contribute significantly to understanding the paleodemography of prehispanic urban centers and how they might compare to their Old World counterparts. While such a comparison is not beyond the methodological techniques that are currently available, it would require sufficiently large skeletal samples to be able to identify enough immigrants that the calculation of age-specific mortality rates would be possible. Amassing a skeletal collection of this huge size for a single site is quite unlikely to happen anytime in the near future. A more realistic goal for paleodemography would, therefore, be to focus on identifying general features of immigration into New World urban centers. These data can then be used as the basis for

319 developing models, much like that generated by Paine (1997), to understand how the paleodemographic age-at-death distribution would be influenced by the observed patterns of migration.

Paleopathological issues

Obviously, a substantial amount of work must still be done to understand population dynamics at the urban center of Cholula. One of the largest gaps in our knowledge of morbidity and mortality in Postclassic Cholula is how the pathological lesions observed in the skeletal collection relate to health. An examination of the relationship between particular pathological lesions and the risk of death indicated that the presence of most lesions did, in fact, increase mortality, although there were some exceptions. We still do not know, however, if the presence of healed versus unhealed lesions says something different about health since sample sizes precluded such a comparison in the current study. For example, porotic hyperostosis occurs during childhood, so the few adults in the Cholula collection observed to have porotic hyperostosis have healed lesions. Clearly these individuals survived well beyond the insult that caused the pathology. What does that indicate about the frailty of these individuals? Are they frailer than other members of their cohort because they experienced the illness, or are they less frail because they survived it? We also do not know how the age at which a pathological lesion developed affects mortality. The reduced version of the Usher model measures only the aggregate-level effects of a pathological lesion on the risk of death, but studies have demonstrated that mortality risks may be related to the age at which the insult occurred

(Goodman 1996; Ferrell 2003). Data on the age at which enamel hypoplasias formed in

320 skeletons from the Cholula collection were gathered as part of the current study and will be used in a future investigation to determine the interaction between the age at which the stress occurred and its affect on mortality for this population. However, a comparison of the health of individuals from different urban centers in Mesoamerica, or a comparison of the health of individuals from Cholula with that of residents of smaller, less densely-settled sites awaits the implementation of Usher’s full multi-state model of health.

Archaeology could make a substantial contribution to the study of urban health by focusing on the investigation of particular features of preindustrial societies thought to facilitate the spread of infectious disease. Dense settlements, contaminated water supplies, and inadequate disposal of sewage have been cited as three features of preindustrial cities that result in higher morbidity and mortality. However, as was discussed in Chapter III, it has yet to be demonstrated that these are universal features of preindustrial cities. Archaeological investigations into these issues at Cholula could identify the settlement density of the site and how issues such as the supply of water to the urban population and the disposal of waste were managed. In general, there is much that we still do not know about Cholula and the Puebla-Tlaxcala region, although several archaeologists are currently attempting to rectify this situation (Plunket et al. 1994;

Uruñuela et al. 1998; Uruñuela and Plunket 1998, 2001, 2002, 2005, 2007; Plunket and Uruñuela

1993, 1998a, 1998b, 1998c, 2002, 2003, 2005; MaCafferty 1996, 2000, 2001, 2007).

Many aspects of the demography of Postclassic Cholula require further research. In general, there is evidence that suggests that some elements of the demography of the city may differ from the patterns observed in preindustrial Old World cities and that cultural dynamics played a role in shaping urban population dynamics in this New World city. The demographic characteristics of immigrants, their reasons for migrating, and the way in which they are

321 incorporated into their new home are influenced by social, economic, and perhaps, in some cases, even political factors. For example, if marriage was one of the primary reasons that immigrants came to Cholula, it would have had implications for fertility rates in the city.

Similarly, if many immigrants to Cholula were families with young children, economic or social conditions in prehispanic Mesoamerica must have made such movements feasible, and the demographic characteristics of these individuals, in turn, would have had effects on both fertility and mortality levels the city. Without first identifying how cultural issues would have impacted urban population dynamics, it is difficult to address whether the epidemiological environment of the city was truly causing higher mortality among residents.

Several additional studies of the Cholula skeletal collection used in the current project have been outlined here. These investigations are necessary to clarify patterns of urban demography in this habitational zone. However, it should be noted that variability almost certainly existed in different parts of the city. Large prehispanic skeletal collections from other residential zones of Cholula are not currently available (although see McCafferty 1992 and

Uruñuela 1989 for small collections excavated from residential contexts). If, however, additional skeletal material is excavated from other parts of the city, paleodemographic and isotopic investigations of this material would be required to determine if the patterns noted in the current study also applied to other neighborhoods of prehispanic Cholula.

As prehispanic Mesoamerican cities show considerable variability in population densities, settlement patterns, and economic organization, among other things, paleodemographic investigations of a wide range of urban centers are necessary to gain an accurate understanding of urban population dynamics in the New World. Considerable study of the apartment compound Tlajinga 33 in Teotihuacan has already occurred, and Rebecca Storey has been quite

322 conscientious about updating her work in light of new methodological approaches. However,

Tlajinga 33 is just one of the apartment compounds in the city, so there is a substantial amount of work left to do to understand the demography of this single site. Ultimately, a comparison of

Maya centers with Central Mexican centers such as Tenochtitlan-Tlatelolco, Tula, and

Xochimilco will be essential to identifying whether a common urban demographic regime existed across prehispanic Mesoamerica. Márquez et al. (2002), Márquez and Hernández (eds,

2006), and Camargo et al. (1999) have already begun such work in their comparisons of a number of different Mesoamerican sites of various sizes and densities. The incorporation of new methodological techniques into their studies could verify if the patterns that they have observed are, in fact, accurate.

Human health is affected by a complex interplay of cultural and biological factors. The theory that morbidity and mortality increase as societies become more complex is alluring in its simplicity. Unfortunately, human health may not be so easily summarized by reference to a handful of environmental variables. All cities cannot necessarily be characterized as densely- populated, unsanitary settlements dependent upon market exchange. Many Mesoamerican urban centers, in fact, defy such a description. Even Classic Period Teotihuacan, whose status as a city is not in dispute, does not completely conform to this typology.

Demographic studies (Galley 1998; Landers 1993; Finlay 1981a) of Old World cities have recognized the importance of the economy in influencing rates of immigration and the demographic characteristics of immigrants, the importance of sociopolitical factors in influencing the age of marriage of immigrants, and the importance of ideological beliefs in dictating issues such as breastfeeding practices. All of these cultural factors had biological consequences that, in turn, affected morbidity and mortality in Old World cities. While

323 environmental factors and epidemic diseases were also significant variables in the equation, they were not the sole variables in the equation.

Prehispanic Mesoamerican cities not only had a very different epidemiological environment, but also very different political economies, forms of social organization, and religious and ideological beliefs, all of which would have played some role in the formation of an urban demographic regime. From determining how problems of urban living were mediated to influencing who immigrated to the city, culture simply cannot be overlooked in understanding human demography. Before overarching theories of human health in cities can be demonstrated conclusively, anthropologists must first disentangle the contribution of cultural, biological, and environmental variables in shaping urban population dynamics.

324 BIBLIOGRAPHY

Aguilar, F. de 1954 Relación breve de la conquista de la Nueva España. Porrua e Hijos Sucesores, Mexico.

Alter, G. and J. Riley 1989 Frailty, Sickness, and Death: Models of Morbidity and Mortality in Historical Populations. Population Studies 43: 25-45.

Alvarado, F. 1943 Crónica mexicana. UNAM, Mexico.

Anderson, J. 1987 Archaeological Testimony of the Ancient Cholula Diet. Notas Mesoamericanas 10: 179-187.

Aufderheide, A. and C. Rodrigues-Martin 1998 The Cambridge Encyclopedia of Human Paleopathology. Cambridge University Press, Cambridge.

Baird, E. 1988 Sahagún’s Primeros Memoriales and Codex Florentino: European Elements in the Illustrations. In Smoke and Mist: Mesoamerican Studies in Memory of Thelma Sullivan, edited by K. Dosserand and K. Dankin, pp. 15-40. British Archaeological Reports, Oxford.

Bateson, M.C. 1994 With a Daughter’s Eye: A Memoir of Margaret Mead and Gregory Bateson. Harper Perennial, New York.

Beaver, S. 1975 Demographic Transition Theory Reinterpreted. Lexington Books, Lexington.

Beekman, C. and A. Christensen 2003 Controlling for Doubt and Uncertainty through Multiple Lines of Evidence: A New Look at the Mesoamerican Nahua Migrations. Journal of Archaeological Method and Theory 10: 111-164.

Berlin, H. and S. Rendon, eds. 1947 La Historia Tolteca-Chichimeca: Anales de Quauhtinchan. Antigua Librería Robredo, Mexico.

Ben-Sira, Z. 1997 Immigration, Stress, and Readjustment. Praeger Publishers, Connecticut.

325

Berdan, F. 1976 Organización del tributo en el Imperio Azteca. Estudios de Cultura Nahuatl 12: 185-195.

Bernal García, M.E. and A. García Zambrano 2006 El altepetl colonial y sus antecedentes prehispánicos: contexto teórico- historiográfico. In Territorialidad y paisaje en el altepetl del siglo XVI, edited by F. Fernandez and A. Garcia, pp. 31-113. Fondo de Cultura Económica and the Instituto de Geografía de la UNAM, Mexico.

Blanton, R. 1981 The Rise of Cities. In Archaeology, Supplement to the Handbook of Middle American Indians, edited by J. Sabloff, pp. 392-400. University of Texas Press, Austin.

Blurton Jones, N., K. Hawkes, and J. O’Connell 2002 Antiquity of Postreproductive Life: Are There Modern Impacts on Hunter- Gatherer Postreproductive Lifespans? American Journal of Human Biology 14: 184-20

Bocquet-Appel, J. and C. Masset 1982 Farewell to Paleodemography. Journal of Human Evolution 11: 321-333.

Boel, L. 2001 Adult Bone Mass and Physiological Stress: A Paleoepidemiological Study. Unpublished Dissertation. University of Aarhus, Aarhus.

Boldsen, J. 1984 Paleodemography of Two Southern Scandinavian Medieval Communities. Meddelanden Fran Lunds Universitets Historiska Museum 5: 107-115.

1991 Ageing and Dental Attrition in a Medieval Rural Danish Population. International Journal of Anthropology 6: 217-224.

1997 Estimating Patterns of Disease and Mortality in a Medieval Danish Village. In Integrating Archaeological Demography: Multidisciplinary approaches to Prehistoric Population, edited by R. Paine, pp. 229-241. Center for Archaeological Investigations, Illinois.

Boldsen, J., G. Milner, L. Konigsberg, and J. Wood 2002 Transition Analysis: A New Method for estimating age from skeletons. In Paleodemography: Age Distributions from Skeletal Samples, edited by R. Hoppa and J. Vaupel, pp. 73-106. Cambridge University Press, Cambridge.

Bonfil, G.

326 1973 La ciudad sagrada en la era industrial. UNAM, Mexico.

Brooks, S. 1955 Skeletal Age at Death: The Reliability of Cranial and Pubic Age Indicators. American Journal of Physical Anthropology 13: 567-597.

Brooks, S. and J. Suchey 1990 Skeletal Age Determination Based on the Os Pubis: A Comparison of the Acsadi- Nemeskeri and Suchey-Brooks Methods. Human Evolution 5: 227-238. Brothwell, D. 1994 On the Possibility of Urban-Rural Contrasts in Human Population Paleobiology. In Urban-Rural Connexions: Perspectives from Environmental Archaeology, edited by A. Hall and H. Kenward, pp.129-136. Oxbow Books, Oxford.

Brumfiel, E. 1980 Specialization, Market Exchange, and the Aztec State: A View from Huexotla. Current Anthropology 21: 459-478.

2005 Opting In and Opting Out: Tula, Cholula, and Xaltocan. In Settlement, Subsistence, and Social Complexity: Essays Honoring the Legacy of Jeffrey R. Parsons, edited by R. Blanton, pp. 63-88. Cotsen Institute of Archaeology, University of California, Los Angeles.

Brundage, B. 1985 The Jade Steps: A Ritual Life of the Aztecs. University of Utah Press, Utah.

Buikstra, J. 1997 Paleodemography: Context and Promise. In Integrating Archaeological Demography: Multidisciplinary approaches to Prehistoric Population, edited by R. Paine, pp. 367-380. Center for Archaeological Investigations, Illinois.

Buikstra, J., T.D. Price, L. Wright, and J. Burton 2004 Tombs from the Copan Acropolis: A Life History Approach. In Understanding Early Classic Copan, edited by E. Bell, M. Canuto, and R. Sharer, pp. 185-205. University of Pennsylvania Museum, Philadelphia.

Buikstra, J and D. Ubelaker 1994 Standards for Data Collection from Human Skeletal Remains. Arkansas Archaeological Society, Arkansas.

Burton, J., T. D. Price, L. Cahue, L. Wright 2003 The Use of Barium and Strontium Abundances in Human Skeletal Tissues to Determine Their Geographic Origin. International Journal of Osteoarchaeology 13: 88-95.

Burton, J. T.D. Price, and W. Middleton

327 1999 Correlation of bone Ba/Ca and Sr/Ca due to biological purification of calcium. Journal of Archaeological Science 26: 609-616.

Calnek, E. 1972 Settlement Pattern and Chinampa Agriculture in Tenochtitlan. American Antiquity 37:104-114.

1982 The Internal Structure of Tenochtitlan. In Ancient Mesoamerica, pp. 331-344. Peek, Palo Alto.

Carmargo, L.. L. Márquez, and M. Prado 1999 Paleodemografía del México prehispánico. In Hacia la demografía del siglo XXI, Vol. 3, edited by Benítez and Ornelas, pp. 227-250. Sociedad Mexicana de Demografía, Mexico.

Carrasco, P. 1971 Social Organization of Ancient Mexico. In Archaeology of North Mesoamerica, Part 1, edited by G. Ekholm and I. Bernal, pp. 349-375, Handbook of Middle American Indians, Vol. 10. University of Texas Press, Austin.

1999 The Tenocha Empire of Ancient Mexico: The Triple Alliance of Tenochtitlan, Texcoco, and Tlacopan. University of Oklahoma Press, Oklahoma.

Ceja, M. 1987 Azcapotzalco: una población del Postclásico vista a través de sus enterramientos. Licenciatura thesis, ENAH-INAH, Mexico.

Cervantes de Salazar, F. 1914 Crónica de la Nueva España. The Hispanic Society of America, Madrid.

Chase, D., A. Chase, and W. Haviland 1990 The Classic Maya City: Reconsidering the Mesoamerican Urban Tradition. American Anthropologist 92: 499-506.

Chesnais, J. 1992 The Demographic Transition: Stages, Patterns, and Economic Implications. Oxford University Press, Oxford.

Clark, P. and D. Souden 1988 Introduction. In Migration and Society in Early Modern England, edited by P. Clark and D. Souden, pp. 11-48. Barnes and Noble, New Jersey.

Coale, A. 1972 The Growth and Structure of Human Populations. University Press, Princeton.

Cockburn, T.A.

328 1971 Infectious Diseases in Ancient Populations. Current Anthropology 12(1): 45-62.

Coggins, C. 1975 Painting and Drawing Styles at Tikal. Unpublished Ph.D. Dissertation. Harvard University, Massachusetts.

Cohen, M. 1977 The Food Crisis in Prehistory: Overpopulation and the Origin of Agriculture. Yale University Press, New Haven.

1989 Health and the Rise of Civilization. Yale University Press, New Haven.

1994 The Osteological Paradox Reconsidered. Current Anthropology 35(5): 629-631.

1997 Does Paleopathology Measure Community Health? A Rebuttal of ―the Osteological Paradox‖ and Its Implications for World History. In Integrating Archaeological Demography: Multidisciplinary Approaches to Prehistoric Population, edited by R. Paine, pp. 242-260. Center for Archaeological Investigations, Carbondale.

2007 Ancient Health: Skeletal Indicators of Agricultural and Economic Intensification. University of Florida Press: Gainseville.

Cohen, M. and G. Armelagos, eds. 1984 Paleopathology at the Origins of Agriculture. Academic Press, Orlando.

Cohen, S., D. Tyrrell, and A. Smith 1991 Psychological Stress and Susceptibility to the Common Cold. New England Journal of Medicine 325: 606-612.

Cohen, S. and G. Williamson 1991 Stress and Infectious Disease in Humans. Psychological Bulletin 109: 5-24.

Colston, S. 1988 Fray Diego Durán and His Native Informants. In Smoke and Mist: Mesoamerican Studies in Memory of Thelma Sullivan, edited by K. Dosserand and K. Dankin, pp. 59-68. British Archaeological Reports, Oxford.

Cook, D. and J. Buikstra 1979 Health and Differential Survival in Prehistoric Populations. American Journal of Physical Anthropology 51: 649-64.

Cortés, H. 1985 Cartas de Relación. Dastin, Spain.

DeVries, J.

329 1984 European Urbanization 1500-1800. Harvard University Press, Cambridge.

DeWitte, S. 2006 The Paleodemography of the Black Death 1347-1351. Unpublished Ph.D. Dissertation, Department of Anthropology, The Pennsylvania State University, Pennsylvania.

DeWitte, S. and J. Wood 2008 Selectivity of Black Death Mortality with Respect to Preexisting Health. Proceedings of the National Academy of Sciences 104: 1436-1441.

Diaz del Castillo, B. 1996 The Discovery and Conquest of Mexico. Da Capo Press, New York.

Dirks, R. 1993 Starvation and Famine: Cross-Cultural Codes and Some Hypothesis Tests. Cross Cultural Research 27: 28-69.

Drennan, R. 1984 Long-Distance Transport Costs in Pre-Hispanic Mesoamerica. American Anthropologist 86(1): 105-112.

Duncan, S., S. Scott, and C. Duncan 1993 The Dynamics of Smallpox Epidemics in Britain, 1550-1800. Demography 30 (3): 405-423.

Durán, D. 1963 The Aztecs: The History of the Indies of New Spain. Orion Press, New York.

1971 Books of the Gods and Rites and the Ancient Calendar, translated by F. Horcasitas and D. Heyden. University of Oklahoma Press, Norman.

Durmond, D. and F. Müller 1972 Classic to Post-classic in Highland Central Mexico. Science 175: 1208-1215.

Dyson, T. 1991 On the Demography of South Asian Famines. Population Studies 45: 5-25.

El-Najjar, M., D. Ryan, C. Turner, B. Lozoff 1976 The etiology of porotic hyperostosis among the prehistoric and historic Anasazi Indians of the Southwest United States. American Journal of Physical Anthropology 44: 477-478.

Escalante Gonzalbo, P. 2004a La vida urbana en el periodo Clásico mesoamericano: Teotihuacan hacia el año 600 d.C. In Historia de la vida cotidiana en México,Vol. I, edited by P.

330 Gonzalbo, pp. 41-98. El Colegio de México Fondo de Cultura Económica, Mexico.

2004b La ciudad, la gente, y las costumbres. In Historia de la vida cotidiana en México,Vol. I, edited by P. Gonzalbo, pp. 199-230. El Colegio de México Fondo de Cultura Económica, Mexico.

Evans, S. 1988 Archaeology of the Aztec Village of Cihuatecpan, Mexico: Methods and Results of the 1984 Field Season. In Excavations at Cihuatecpan: An Aztec Village in the Teotihuacan Valley, edited by S. Evans, pp. 50-234. Vanderbilt University Publications in Anthropology, Nashville.

Ezzo, J., C. Johnson, and T.D. Price 1997 Analytical perspectives on prehistoric migration: a case study from east-central Arizona. Journal of Archaeological Science 24: 447-466.

Fazekas, I. and F. Kosa 1978 Forensic Fetal Osteology. Akademiai Kiado, Budapest.

Fernández, F. and A. García 2006 Introducción. In Territorialidad y paisaje en el altepetl del siglo XVI, edited by F. Fernández and A. Garcia, pp. 13-28. Instituto de Geografía de la UNAM, Mexico.

Ferré, R. and F. Xalpa 1967 Ecología humana. In Cholula: Reporte Preliminar, edited by M. Messmacher, pp. 109-133. Editorial Nueva Antropología, Mexico.

Ferrell Thomas, R. 2002 Enamel Defects, Well-Being, and Mortality in a Medieval Danish Village. Unpublished Ph.D. Dissertation, Pennsylvania State University.

Finlay, R. 1978 The Accuracy of the London Parish Registers, 1580-1653. Population Studies 32: 95-112.

1981a Population and Metropolis. Cambridge University Press, Cambridge.

1981b Natural Decrease in Early Modern Cities. Past and Present 92: 169-174.

Fogel, R. 2004 The Escape from Hunger and Premature Death 1700-2100. Cambridge University Press, Cambridge.

Fox, R.

331 1977 Urban Anthropology: Cities in Their Cultural Settings. Prentice-Hall, New Jersey.

Gage, T. 1988 Mathematical Hazard Models of Mortality: An Alternative to Model Life Tables. American Journal of Physical Anthropology 76: 429-441.

1989 Bio-mathematical approaches to the study of human variation and mortality. Yearbook of Physical Anthropology 32: 185-214.

1990 Variation and classification of human age patterns of mortality: analysis using competing hazards models. Human Biology 62: 589-614.

1994 Population variation in cause of death: level, gender, and period effects. Demography 31: 271-296.

Gage, T. and B. Dyke 1986 Parameterizing Abridged Mortality Tables: The Siler Three-Component Hazard Model. Human Biology 58: 275-291.

Galley, C. 1995 A Model of Early Modern Urban Demography. Economic History Review 48: 448-469.

1998 The Demography of Early Modern Towns: York in the Sixteenth and Seventeenth Centuries. Liverpool University Press, Liverpool.

Galloway, P. 1994 Annual Variation in Death by Age, Deaths by Cause, Prices, and Weather in London 1670-1830. Population Studies 39: 487-505.

1988 Basic Patterns in Annual Variation in Fertility, Nuptuality, Mortality, and Prices in Preindustrial Europe. Population Studies 42: 275-304.

García Cook, A. and L. Merino Carrión 1987 Condiciones existentes en la región poblano-tlaxcalteca al surgimiento de Cholula. Notas Mesoamericanas 10: 153-176.

1990 El ―Epiclásico‖ en la región poblano-tlaxcalteca. In Mesoamérica y norte de México: Siglo IX- XII, edited by F. Sodi, pp. 257-280. INAH, Mexico.

Gibson, C. 1967 Los aztecos bajo el dominio español 1519-1810. Siglo Veintiuno, Mexico.

Gilbert, B. and T. McKern

332 1973 A Method for Aging the Female Os Pubis. American Journal of Physical Anthropology 38: 31-38.

González-Hermosillo, F. and L. Reyes García 2002 El códice de Cholula: La exaltación testimonial de un linaje indio. INAH, Mexico.

Goodman, A. 1993 On the Interpretation of Health from Skeletal Remains. Current Anthropology 34: 281-288.

1996 Early life stresses and adult health: Insights from dental enamel development. In Long-term Consequences of Early-Environment: Growth, Development and the Lifespan Developmental Perspective, edited by C. Henry and S. Ulijaszek, pp. 163-182. Cambridge University Press, Cambridge.

Goodman, A. and G. Armelagos 1985 Factors Affecting the Distribution of Enamel Hypoplasias within the Human Permanent Dentition. American Journal of Physical Anthropology 68: 479-493.

1988 Childhood Stress and Decreased Longevity in a Prehistoric Population. American Anthropologist 90: 936-944.

Goodman, A. and J. Rose 1990 Assessment of Systematic Physiological Perturbations from Dental Enamel Hypoplasias and Associated Histological Structures. Yearbook of Physical Anthropology 33: 59-110.

Goodman, A., D. Martin, G. Armelagos, and G. Clark 1984 Indicators of Stress from Bone and Teeth. In Paleopathology at the Origins Of Agriculture, edited by M. Cohen and G. Armelagos, pp13-50. Academic Press, New York.

Gordon, C. and J. Buikstra 1981 Soil pH, Bone Preservation, and Sampling Bias at Mortuary Sites. American Antiquity 46: 566-571.

Graunt, J. 1662 Natural and Political Observations, Mentioned in a Following Index, and Made upon the Bills of Mortality. Tho. Roycroft, London.

Grobbee, D. and A. Hoes 2009 Clinical Epidemiology: Principles, Methods, and Applications for Clinical Research. Jones and Bartlett, Massachusetts.

Grupe, G., T.D. Price, P. Schorter, F. Sollner, C. Johnson, and B. Beard

333 1997 Mobility of Bell Beaker People Revealed by Stable Strontium Isotope Ratios of Teeth and Bones: A Study of Southern Bavarian Skeletal Remains. Applied Geochemistry 12:517-525.

Gurven, M. and H. Kaplan 2007 Longevity Among Hunter-Gatherers: A Cross-Cultural Examination. Population and Development Review 33: 321-365.

Harpending, H. and L. Wandsnider 1982 Population Structures of Ghanzi and Ngamiland !Kung. In Current Developments in Anthropological Genetics, Vol. 2, edited by M. Crawford, and J. Mielke, pp. 29-50. Plenum Press, New York.

Harpending, P.D. 1976 Social and Economic Constraints on Child Life among the !Kung. In Kalahari Hunter-Gatherers, edited by R. Lee and I. DeVore, pp. 199-217. Harvard University Press, Cambridge.

Hartwell, R. 1961 The Rising Standard of Living in England 1800-1850. Economic History Review 13: 397-416.

Haviland, W. 1970 Tikal, Guatemala and Mesoamerican Urbanism. World Archaeology 2: 186-198.

Hayward, M. 1986 A Demographic Study of Cholula, Mexico, from the Late Postclassic and the Colonial Period of 1642-1738. Unpublished Ph.D. Dissertation, Department of Anthropology, The Pennsylvania State University, Pennsylvania.

Healan, D., ed. 1989 Tula of the Toltecs: Excavations and Survey. University of Iowa Press, Iowa.

Hernández, C. 1970 Restos arquitectónicos del horizonte postclásico en Cholula. In Proyecto Cholula, edited by I. Marquina, pp. 89-92. INAH, Mexico.

Hernández, G., N. Quintana, G. Uruñuela, and P. Plunket 1998 Informe sondeo UA98B: área para instalación de un sistema riego. Report submitted to the Consejo Nacional de Arqueología, Instituto Nacional de Antropología e Historia, Archivo Técnico del INAH, Mexico City.

Hernández, P. 2006 Entre flores y chinampas: la salud de los antiguos habitantes de Xochimilco. In Salud y sociedad en el México prehispánico y colonial, edited by L. Márquez and P. Hernández, pp. 327-366. CONACULTA-INAH, Mexico.

334

Hess, J., M. Bender, and J. Schilling 1986 Evolution of the Ratio of Strontium-87 to Strontium-86 in Seawater from the Cretaceous to Present. Science 231: 979-984.

Hicks, F. 1982 Tetzcoco in the Early 16th Century: The State, the City, and the Calpolli. American Ethnologist 9: 230-249.

1986 Prehispanic Background of Colonial Political and Economic Organization in Central Mexico. In Supplement to the Handbook of Middle American Indians, Vol. 4 Ethnohistory, edited by R. Spores, pp. 35-54. University of Texas Press, Austin.

1991 Gift and Tribute: Relations of Dependency in Aztec Mexico. In Early State Economics, edited by H. Claessen and P. Van de Velde, pp. 199-213. Transaction Publishers, New Jersey.

Higgins, R., M. Haines, L. Walsh, J. Sirianni 2002 The Poor in the Mid-Nineteenth-Century Northeastern United States: Evidence from the Monroe County Almshouse, Rochester, New York. In The Backbone of History, edited by R. Steckel and J. Rose, pp. 162-184. Cambridge University Press, Cambridge.

Hinde, A. 1998 Demographic Methods. Arnold, Great Britain.

Hirth, K. 2003 The Altepetl and Urban Structure in Prehispanic Mesoamerica. In Urbanism in Mesoamerica, Vol. 1, edited by W. Sanders, A. Mastache, and R. Cobean, pp. 57- 84. Instituto Nacional de Antropología e Historia and The Pennsylvania State University, Mexico and University Park, PA.

Hitchcock, R. 1989 Settlement, Seasonality, and Subsistence Stress among the Tyua of Northern Botswana. In Coping with Seasonal Constraints, edited by R. Huss-Ashmore, J. Curry, and R. Hitchcock, pp. 64-85. Applied Science Center for Archaeology, Philadelphia.

Hodell, D., R. Quinn, M. Brenner, and G. Kamenov 2004 Spatial variation of strontium isotopes (87Sr/86Sr) in the Maya region : a tool for tracking ancient human migration. Journal of Archaeological Science 31: 585- 601.

335 Holland, T. and M. O’Brien 1997 Parasites, Porotic Hyperostosis, and the Implications of Changing Perspectives. American Antiquity 62: 183-193.

Holman, Darryl 2002 mle: A programming language for building likelihood models. Version 2.1. Unpublished, Seattle.

Holman, D., J. Wood, and K. O’Connor 2002 Estimating Age-at-Death Distributions from Skeletal Samples: A Multivariate Latent Trait Approach. In Paleodemography: Age Distributions from Skeletal Samples, edited by R. Hoppa and J. Vaupel, pp. 193-201. Cambridge University Press, Cambridge.

Hoppa, R. and J. Vaupel 2002 The Rostock Manifesto for Paleodemography: The Way from Stage to Age. In Paleodemography: Age Distributions from Skeletal Samples, edited by R. Hoppa and J. Vaupel, pp. 1-8. Cambridge University Press, Cambridge.

Hoppe, K., P. Koch, and T. Furutani 2003 Assessing the Preservation of Biogenic Strontium in Fossil Bones and Tooth Enamel. International Journal of Osteoarchaeology 13:20-28.

Houston, S. 1993 Hieroglyphs and History at Dos Pilas. University of Texas Press, Austin.

Howell, N. 1976a Population of the Dobe area !Kung. In Kalahari Hunter-Gatherers, edited by R. Lee and I. DeVore, pp. 137-151. Harvard University Press, Cambridge.

1976b Toward a Uniformitarian Theory of Human Paleodemography. Journal of Human Evolution 5: 25-40

1979 Demography of the Dobe !Kung. Academic Press, New York.

1982 Village Composition Implied by a Paleodemographic Life Table: The Libben Site. American Journal of Physical Anthropology 59: 263-270.

Hyman, I. 2001 Immigration and Health. Health Canada, Ottowa.

Ixtlilxochitl, F. 1975-1977 Obras históricas, Vol. I and II. UNAM, Mexico.

Johannson, S. and S. Horowitz 1986 Estimating Mortality in Skeletal Populations: Influence of the Growth Rate on

336 the Interpretation of Levels and Trends During the Transition to Agriculture. American Journal of Physical Anthropology 71: 223-250.

Jones, P. and A. Judges 1935 London Population in the Late Seventeenth Century. The Economic History Review 6: 45-63.

Keyfitz, N. 1977 Applied Mathematical Demography. John Wiley and Sons, United States.

Keyfitz, N. and W. Flieger 1971 Population: Facts and Methods of Demography. W.H. Freeman, San Francisco.

Kirchhoff, P. 1947 La Historia Tolteca Chichimeca: Un estudio historico-sociologico. In La Historia Tolteca Chichimeca: Anales de Quauhtinchan, edited by H. Berlon and S. Rendon, prologue. Antigua Librería Robredo, Mexico.

Knudson, K., D. Price, J. Buikstra, and D. Blom 2004 The use of strontium isotope analysis to investigate Tiwanaku migration and mortuary ritual in Boliva and Peru. Archaeometry 46: 5-18.

Kohn, M., M. Schoninger, and W. Barker 1999 Altered States: Effects of Diagenesis on Fossil Tooth Chemistry. Geochimica et Cosmochimica Acta 67: 2737-2747.

Konigsberg, l. and S. Frakenberg 1992 Estimation of Age Structure in Anthropological Demography. American Journal of Physical Anthropology 89: 235-256.

1994 Paleodemography: ―Not Quite Dead.‖ Evolutionary Anthropology 3: 92-105.

1997 Regress What on What? Paleodemographic Age Estimation as a Calibration Problem. In Integrating Archaeological Demography: Multidisciplinary Approaches to Prehistoric Population, edited by R. Paine, pp. 64-88. Center for Archaeological Investigations, Illinois.

Lagunas, Z. 1973 La trepanación suprainiana en cráneos de Cholula, Pue. Comunicaciones 8: 47- 48.

1994 Las practicas funerarias en Cholula prehispánica. Mirada antropológica 1: 82- 94.

Landers, J.

337 1991 London’s Mortality in the ―Long Eighteenth Century‖: A Family Reconstitution Study. In Living and Dying in London, edited by W.F. Bynum and R. Porter, pp. 1-28. Wellcome Institute for the History of Medicine, London.

1993 Death and the metropolis: Studies in the Demographic History of London 1670-1830. Cambridge University Press, Cambridge.

Latham, M. 1997 Human Nutrition in the Developing World. FAO, Rome.

Lee, R. 1972 Work Effort, Group Structure, and Land Use in Contemporary Hunter-Gatherers. In Man, Settlement, and Urbanism, edited by P. Ucko, R. Tringham, and G. Dimbleby, pp. 177-185. General Learning Press, New Jersey.

Lewis, M., C. Roberts, and K. Manchester 1994 Comparative Study of the Prevalence of Maxillary Sinusitis in Later Medieval Urban and Rural Populations in Northern England. American Journal of Physical Anthropology 98: 497-506.

Lind, M. 1994 Cholula and Mixteca Polychromes: Two Mixteca Puebla Regional Substyles. In Mixteca-Puebla: Discoveries and Research in Mesoamerican Art and Archaeology, edited by H. Nicholson and E. Quiñones Keber, pp. 79-99. Labyrinthos, California.

Lockhart, J. 1992 The Nahuas after the Conquest. Stanford University Press, Stanford.

Lomelí, L. 2001 Breve historia de Puebla. Fondo de Cultura Económica, Mexico.

Lopez, A. 1961 Problems in Stable Population Theory. Office of Population Research, Princeton.

López, A., S. Talavera, G. Uruñela, and P. Plunket 2002a Sondeo arqueológico Rancho de la Virgen 2002. Report submitted to the Consejo Nacional de Arqueología, Instituto Nacional de Antropología e Historia, Archivo Técnico del INAH, Mexico City.

2002b Sondeo arqueológico Colegio Coley B. Taylor. Report submitted to the Consejo Nacional de Arqueología, Instituto Nacional de Antropología e Historia, Archivo Técnico del INAH, Mexico City.

López Austin, A. 1988 Human Body and Ideology: Concepts of the Ancient Nahuas. University of Utah

338 Press, Salt Lake City.

López Lújan, L. 1993 The Offerings of the Templo Mayor of Tenochtitlan. University Press of Colorado, Colorado.

López, S. 1972 La cremación y los entierros en vasija de la zona arqueológica de Cholula, Puebla. Religión en Mesoamérica: XII Mesa Redonda de la Sociedad Mexicana de Antropología : 375-379.

1973 Enterramientos humanos prehispánicos de Cholula, Puebla. Comunicaciones 8: 43-45.

López, S., Z. Lagunas, and C. Serrano 1970 Sección de Antropología Física. In Proyecto Cholula, edited by I. Marquina, pp. 143-152. INAH, Mexico.

1976 Enterramientos humanos de la zona arqueológica de Cholula, Puebla. INAH, Mexico.

2002 Costumbres funerarios y sacrificio humano en Cholula prehispánica. UNAM, Mexico.

López, S. and M. Salas 1989 Los antiguos habitantes de Cholula: algunos elementos del perfil físico. Notas mesoamericanas 11: 5-18.

Lotka, A. 1907 Mode of Growth of Material Aggregates. American Journal of Science 24: 199- 216.

1922 Contributions to the Energetics of Evolution. Proceedings of the National Academy of Sciences USA 8: 147-150.

Love, B. and H. Müller 2002 A Solution to the Problem of Obtaining a Mortality Schedule for Paleodemographic Data. In Paleodemography: Age Distributions from Skeletal Samples, edited by R. Hoppa and J. Vaupel, pp. 181-192. Cambridge University Press, Cambridge.

Lovejoy, C., R. Meindl, T. Pryzbeck, T. Barton, K. Heiple, and D. Kotting 1977 Paleodemography of the Libben Site, Ottowa County, Ohio. Science 198: 291- 293.

Lovejoy, C., R. Meindl, T. Pryzbeck, and R. Mensforth

339 1985 Chronological Metamorphosis of the Auricular Surface of the Ilium: A New Method for the Determination of Adult Skeletal Age at Death. American Journal of Physical Anthropology 68: 15-28.

Luu, L.B. 2005 Immigrants and the Industries of London, 1500-1700. Ashgate Publishing, Great Britain.

Mansilla, J. 1978 Estudio de las líneas de Harris en dos poblaciones prehispánicas. Revista Mexicana de Estudios Antropológicos 24: 19-37.

Marcus, J 1983 On the Nature of the Mesoamerican City. In Prehispanic Settlement Patterns: Essays in Honor of Gordon Willey, edited by E. Vogt and R. Levanthal, pp. 195- 242. University of New Mexico Press and the Peabody Museum of Archaeology and Ethnology, Albuquerque and Cambridge.

Margerison, B. and C. Knusel 2002 Paleodemographic Comparison of a Catastrophic and an Attritional Death Assemblage. American Journal of Physical Anthropology 119: 134-143.

Márquez, M. and P. Hernández (eds.) 2006 Salud y sociedad en el México prehispánico y colonial. CONACULTA INAH, Mexico.

Márquez M., L., R. McCaa, R. Storey, and A. del Angel 2002 Health and nutrition in Pre-hispanic Mesoamerican. In The Backbone of History, edited by R. Steckel and J. Rose, pp. 307-340. Cambridge University Press, Cambridge.

Marquina, I., (ed.) 1970 Proyecto Cholula. Serie Investigaciones No. 19. INAH, Mexico.

Matos, E. 1967 Estudio de la agricultura y su relación con los patrones de asentamiento. In Cholula: Reporte Preliminar, edited by M. Messmacher, pp. 44-48. Editorial Nueva Antropología, Mexico.

1984 The Great Temple of the Aztecs. Thames and Hudson, London.

McCafferty, G. 1992 The Material Culture of Postclassic Cholula: Contextual Interpretations of the UA-1 Domestic Compunds. Unpublished Dissertation. State University of New York, Binghamton.

340 1996a Reinterpreting the Great Pyramid of Cholula, Puebla. Ancient Mesoamerica 7: 1- 17.

1996b The Ceramics and Chronology of Cholula, Mexico. Ancient Mesoamerica 7: 299- 324.

2000 Tollan Chollolan and the Legacy of Legitimacy during the Classic-Postclassic Transition. In Mesoamerica’s Classic Heritage: From Teotihuacan to the Aztecs, edited by D. Carrasco, L. Jones, and S. Sessions, pp. 341-367.

2001 Mountain of Heaven, Mountain of Earth: The Great Pyramid of Cholula as Sacred Landscape. In Landscape and Power in Ancient Mesoamerica, edited by R. Koontz, K. Reese-Taylor, and A. Headrick, pp. 279-316. Westview Press, Colorado.

2007 Altar Egos: Domestic Ritual and Social Identity in Postclassic Cholula, Mexico. In Commoner Ritual and Ideology in Ancient Mesoamerica, edited by N. Gonlin and J. Lohse, pp. 213-250. University of Colorado Press, Boulder.

McCafferty, G. and S. McCafferty 1994 The Conquered Women of Cacaxtla: Gender Identity or Gender Ideology? Ancient Mesoamerica 5: 159-172.

2006 Boys and Girls Interrupted: Mortuary Evidence of Children from Postclassic Cholula, Puebla. In The Social Experience of Childhood in Ancient Mesoamerica, edited by T. Ardren and S. Hutson, pp. 25-52. University of Colorado Press, Boulder.

McKeown, T. 1976 The Modern Rise of Population. Edward Arnold Press, London.

McKern, T. and T. Stewart 1957 Skeletal Age Changes in Young American Males, Analyzed from the Standpoint of Age Identification. Headquarters Quartermaster Research and Development Command. Technical Report EP-45. Natick, MA.

McNeill, W. 1979 Historical Patterns of Migration. Current Anthropology 20(1): 95-102.

Meindl, R., C. Lovejoy, R. Mensforth, and L. Carlos 1985a A Revised Method of Age Determination Using the Os Pubis, with a review and tests of accuracy of other current methods of pubic symphyseal aging. American Journal of Physical Anthropology 68: 29-45.

1985b Accuracy and Direction of Error in Sexing of the Skeleton: Implications for Paleodemography. American Journal of Physical Anthropology 68: 29-45.

341

Mendieta, G. de 1971 Historia eclesiástica indiana. Porrua, Mexico.

Mensforth, R., C. Lovejoy, J. Lallo, and G. Armelagos 1978 The Role of Constitutional Factors, Diet and Infectious Disease in the Etiology Of Porotic Hyperostosis and Periosteal Reactions in Prehistoric Infants and Children. Medical Anthropology 2: 1-58.

Messmacher, M. 1967 Los patrones de asentamiento y la arquitectura en Cholula. In Cholula: Reporte Preliminar, edited by M. Messmacher, pp. 6-17. Editorial Nueva Antropología, Mexico.

Mielke, J., L. Jorde, P. Trapp, D. Anderton, K. Pitkanen, and A. Eriksson 1984 Historical Epidemiology of Smallpox in Aland, Finland 1751-1890. Demography 21: 271-295.

Milner, G. 2010 Transition Analysis and Subjective Estimates of Age in Adult Skeletons. In ADBOU 1992-2009 Forskningsresultater, edited by J. Boldsen and P. Tarp. Syddansk Universitet, Denmark.

Milner, G., E. Anderson, and V. Smith 1991 Warfare in Late Prehistoric West-Central Illinois. American Antiquity 56: 581- 603.

Milner, G., D. Humpf, and H. Harpending 1989 Pattern Matching of Age-at-Death Distributions in Paleodemographic Analysis. American Journal of Physical Anthropology 80: 49-58.

Milner, G., J. Wood, and J. Boldsen 2000 Paleodemography. In Biological Anthropology of the Human Skeleton, edited by M. Katzenburg and S. Saunders, pp. 467-497. Wiley –Liss, New York.

Moholy-Nagy, H. 1999 Mexican Obsidian at Tikal, Guatemala. Latin American Antiquity 10: 300-313.

Molloy, J and W. Rathje 1974 Sexploitation among the Late Classic Maya. In Mesoamerican Archaeology: New Approaches, edited by N. Hammond, pp. 431-444. University of Texas Press, Austin.

Montgomery, J., P. Budd, and J. Evans 2000 Reconstructing the lifetime movements of ancient people: a Neolithic case study from southern England. European Journal of Archaeology 3: 370-385.

342

Monzón, M. 1990 Casas Prehispánicas en Teotihuacan. UAEM, Mexico.

Motolinía, T. 1971 Libro de las cosas de la Nueva España y de los naturales de ella. UNAM, Mexico.

Mountjoy, J. 1987 The Collapse of the Classic at Cholula as seen from Cerro Zapotecas. Notas Mesoamericanas 10: 119-151.

Mountjoy, J. and D. Peterson 1974 Man and Land in Prehispanic Cholula. Vanderbilt University Publications in Anthropology, No. 4, Nashville.

Müller, F. 1973 La extensión arqueológica de Cholula a través del tiempo. Comunicaciones 8: 19-21.

Murray, K. and T. Murray 1991 A Test of the Auricular Surface Aging Technique. Journal of Forensic Science 36: 1162-1169.

Nagao, D. 1985 Mexica Buried Offerings. British Archaeological Reports, vol. 35.

Noguera, E. 1954 La cerámica arqueológica de Cholula. Editorial Guarania, Mexico.

1956 Un edificio preclásico en Cholula. In Estudios antropológicos publicados en homenaje al Dr. Manuel Gamio, pp. 213-224. UNAM, Mexico.

Noh, S. and V. Kaspar 2003 Diversity and Immigrant Health. University of Toronto, Toronto.

Noyola, A. 1992 Unidades habitacionales prehispánicas excavadas en el estado de Puebla. Notas Mesoamericanas 14: 19-36.

Olivera, M. 1970 La importancia religiosa de Cholula. In Proyecto Cholula, edited by Ignacio Marquina, pp. 211-242. INAH, Mexico.

1979 Pillis y macehuales: Las formaciones sociales y los modos de producción de Tecali del siglo XII al XVI. Ediciones de la Casa Chata, Mexico.

343

Ortiz de Montellano, B. 1990 Medicina, salud, y nutrición aztecas. Rutgers University Press, New Brunswick.

Ortner, D. 1991 Theoretical and Methodological Issues in Paleopathology. In Human Paleopathology: Current Syntheses and Future Options, edited by D. Ortner and A. Aufderhheide, pp. 5-11. Smithsonian Institution Press, Washington, D.C.

1999 Probable Evidence of Scurvy in Subadults from Archaeological Sites in Peru. American Journal of Physical Anthropology 108: 321-331.

2002 Identification of Pathological Conditions in Human Skeletal Remains. Academic Press, San Diego.

Ortner, D. and W. Putschar 1981 Identification of Pathological Conditions in Human Skeletal Remains. Smithsonian Institution Press, Washington.

Paddock, J. 1985 Cholula en Mesoamérica. Notas Mesoamericanas 10: 21-70.

1992 Cholula de vidas varias. Notas Mesoamericanas 14: 3-18.

Paine, R. 1997 Uniformitarian models in osteological paleodemography. In Integrating Archaeological Demography: Multidisciplinary Approaches to Prehistoric Populations, edited by R. Paine, pp. 191-204. Center for Archaeological Investigations, Southern Illinois University, Carbondale.

2000 If a population crashes in prehistory, and there is no paleodemographer there to hear it, does it make a sound? American Journal of Physical Anthropology 112: 181-190.

Palka, J. 1996 Sociopolitical Implications of a New Emblem Glyph and Place Name in Classic Maya Inscriptions. Latin American Antiquity 7(3): 211-227.

Palkovich, A. 1987 Endemic Disease Patterns in Paleopathology: Porotic Hyperostosis American Journal of Physical Anthropology 74: 527-537.

Perrenoud, A. 1975 L’inegalite sociale devant la mort a Geneve au XVIIieme siecle. Population 30: 221-243.

344 1978 La mortalite a Geneve de 1625 a 1825. Annales de Demographie Historique, pp. 209-233. Societe de Demographie Historique, Paris.

Peterson, D. 1987 The Real Cholula. Notas mesoamericanas 10: 71-117.

Peterson, P., C. Chao, T. Molitor, M. Murtaugh, F. Strgar, and B. Sharp 1991 Stress and Pathogenesis of Infectious Disease. Reviews of Infectious Disease 13: 710-720.

Phenice, T. 1969 A newly developed visual method of sexing in the Os Pubis. American Journal of Physical Anthropology 30: 297-301.

Plunket, P. 1995 Cholula y su cerámic postclásica: algunas perspectivas. Arqueología 13/14: 103-108.

Plunket, P. and G. Uruñuela 1993 Proyecto de salvamento arqueológico: la residencia de monjas del convento de San Gabriel de Cholula, primera temporada. Report submitted to the Consejo Nacional de Arqueología, Instituto Nacional de Antropología e Historia, Archivo Técnico del INAH, Mexico.

1998a Preclassic Household Patterns Preserved Under Volcanic Ash at Tetimpa, Puebla. Latin American Antiquity 9: 287-309.

1998b Cholula y Teotihuacan. In Rutas de intercambio en Mesoamérica, edited by E. Rattray, pp. 101-114. UNAM, Mexico.

1998c Las unidades domésticas de Tetimpa. In Antropología e historia del occidente de México: XXIV Mesa Redonda, edited by R. Brambila, pp. 1789-1798. Sociedad Mexicana de Antropología and UNAM, Mexico.

2002 Shrines, Ancestors, and the Volcanic Landscape at Tetimpa, Puebla. In Domestic Ritual in Ancient Mesoamerica, edited by P. Plunket, pp. 31-42. The Cotsen Institute of Archaeology, California.

2003 From Episodic to Permanent Abandonment: Responses to Volcanic Hazards at Tetimpa, Puebla, Mexico. In Studies of Site Abandonment Processes in Middle America, edited by T. Inomata and R. Webb, pp. 13-27. University of Utah Press, Salt Lake City.

2005 Recent Research in Puebla Prehistory. Journal of Archaeological Research 13: 89-127.

345 Plunket, P., G. Uruñuela, and G. Hernández 1994 Proyecto de salvamento arqueológico: la residencia de monjas del convento de San Gabriel de Cholula, segunda temporada. Report submitted to the Consejo Nacional de Arqueología, Instituto Nacional de Antropología e Historia, Archivo Técnico del INAH, Mexico.

Polgar, S. 1972 Population Histories and Population Policies from an Anthropological Perspective. Current Anthropology 13 (2): 203-211.

Price, T.D., R. Bentley, J. Luning, D. Gronenborn, and J. Wahl 2001 Prehistoric human migration in the Linearbandkeramik of Central Europe. Antiquity 75: 593-603.

Price, T. D., J. Burton, and R. Bentley 2002 The characterization of biologically available strontium isotope ratios for the study of prehistoric migration. Archaeometry 44: 117-135.

Price, T. D., J. Burton, P. Fullagar, L. Wright, J. Buikstra, and V. Tiesler 2008 Strontium Isotopes and the Study of human Mobility in Ancient Mesoamerica. Latin American Antiquity 19: 167-180.

Price, T.D., J. Burton, R. Sharer, J. Buikstra, L. Wright, K. Miller In press Kings and Commoners at Copan: Isotopic Evidence of Origins and Movement in the Classic Maya Period. Journal of Anthropological Archaeology (2009).

Price, T.D., G. Grupe, and P. Schroter 1998 Migration in the Bell Beaker period of central Europe. Antiquity 72: 405-411.

Price, T.D., C. Johnson, J. Ezzo, J. Ericson, and J. Burton 1994 Residential mobility in the prehistoric southwest United States: a preliminary study using strontium isotope analysis. Journal of Archaeological Science 21: 315-330.

Price, T. D., C. Knipper, G. Grupe, and V. Smrcka. 2004 Strontium isotopes and prehistoric human migration : the Bell Beaker period in central Europe. European Journal of Archaeology 7: 9-40.

Price, T. D., L. Manzanilla, and W. Middleton 2000 Immigration and the ancient city of Teotihuacan in Mexico: a study using strontium isotope ratios in human bone and teeth. Journal of archaeological science 27: 903-913.

Quiñones Keber, E. 1995 Codex Telleriano-Remensis: Ritual, Divination, and History in a Pictorial Aztec Manuscript. University of Texas Press, Texas.

346

Rappaport, S. 1989 World within Worlds: Structure of Life in Sixteenth-Century London. Cambridge University Press, New York.

Razum, O., H. Zeeb, and S. Rohrmann 2000 The ―healthy migrant effect‖ – not merely a fallacy of inaccurate denominator figures. International Journal of Epidemiology 29: 191-192.

Reyes García, L. 1988a Cuauhtinchan del Siglo XII al XVI: Formación y desarrollo histórico de un señorío prehispánico. CIESAS, Mexico.

1988b Documentos sobre tierras y señoríos en Cuauhtinchan. CIESAS, Mexico.

Roberts, C. and K. Manchester 1997 The Archaeology of Disease. Cornell University Press, New York.

Rogers, A. and L. Castro 1984 Migration, Urbanization, and Spatial Population Dynamics. Westview Press, Boulder.

Rojas, G. de 1985 Relación de Cholula. In Relaciones Geográficas del Siglo XVI: Tlaxcala, Vol. 2, edited by Rene Acuña, pp. 123-145. UNAM, Mexico.

Romero, J. 1937 Estudio de los entierros de la Pirámide de Cholula. Anales del Museo Nacional de México II: 5-36.

Roth, E. 1985 On the Demographic Concomitants of Sedentism. American Anthropologist 87: 380-382.

Sahagún, F. 1950-1969 Florentine Codex. University of Utah Press, Salt Lake City.

1997 Primeros Memoriales. University of Oklahoma Press, Oklahoma.

Sahlins, M. 1972 Stone Age Economics. Aldine-Atherton, Chicago.

Salas Cuesta, M. 1982 La población de México-Tenochtitlán. INAH, Mexico.

Sanders, W.

347 1971 Settlement Patterns in Central Mexico. In Handbook of Middle American Indians, Vol. 10, edited by G. Elkholm and I. Bernal, pp.3-44. University of Texas Press, Austin.

1977 Ethnographic Analogy and the Teotihuacan Horizon Style. In Teotihuacan and Kaminaljuyu, edited by W. Sanders and J. Michels, pp.397-410. Penn State University Press, Pennsylvania.

1981 Ecological Adaptation in the Basin of Mexico: 23,000 B.C. to the Present. In Archaeology, Supplement to the Handbook of Middle American Indians, edited by J. Sabloff, pp. 147-197. University of Texas Press, Austin.

1989 The Epiclassic as a Stage in Mesoamerican Prehistory: An Evaluation. In Mesoamerica after the Decline of Teotihuacan, edited by R. Diehl and J. Berlo, pp. 211-218. Dumbarton Oaks, Washington, D.C.

2000 Tenochtitlan-Tlatelolco: A Pre-Hispanic Megalopolis. In Arqueologia, historia, y Antropología, edited by J. Litvak and L. Mirambell, pp. 349-371. INAH, Mexico.

Sanders, W. and J. Michels, eds. 1977 Teotihuacan and Kaminaljuyu. Penn State University Press, Pennsylvania.

Sanders, W., J. Parsons, and R. Santley 1980 The Basin of Mexico: Ecological Processes in the Evolution of a Civilization. Academic Press, New York.

Sanders, W. and R. Santley 1983 A Tale of Three Cities: Energetics and Urbanization in Pre-Hispanic Central Mexico. In Prehistoric Settlement Patterns: Essays in Honor of Gordon Willey, edited by E. Vogt and R. Levanthal, pp. 243-291. University of New Mexico Press and the Peabody Museum of Archaeology and Ethnology, Albuquerque and Cambridge.

Sanders, W. and D. Webster 1988 The Mesoamerican Urban Tradition. American Anthropologist 90: 521-546.

Santos, J. 1993 Nutrition, Infection, and Immunocompetence. Infectious Disease Clinics of North America 8: 243-267.

Sattenspiel, L. and H. Harpending 1983 Stable populations and skeletal age. American Antiquity 48: 489-498.

Schweissing, M. and G. Grupe

348 2003a Stable strontium isotopes in human teeth and bone : a key to migration events of the late Roman period in Bavaria. Journal of Archaeological Science 30: 1373- 1383.

2003b Tracing migration events in man and cattle by stable strontium isotope analysis of appositionally grown mineralized tissue. Bone chemistry 13: 96-103.

Scrimshaw, N. 1977 Effect of Infection on Nutrient Requirements. American Journal of Clinical Nutrition 30: 1536-1544.

Scrimshaw, N., C. Taylor, and J. Gordon 1968 Interactions of Nutrition and Infection. World Health Organization, Geneva.

Scrimshaw, N. and C. Tejada 1970 Pathology of Living Indians as Seen in Guatemala. In Handbook of Middle American Indians: Physical Anthropology, edited by T.D. Stewart, pp. 203-225. University of Texas, Austin.

Serrano, C. 1972 Un sitio de entierros ceremoniales en Cholula, Puebla. In Religión en Mesoamérica, XII Mesa Redonda de la Sociedad Mexicana de Antropología: 369-373.

1973 La población prehispánica de Cholula: Tipo físico y aspectos paleodemográficos. Comunicaciones 8: 53-56.

Sharlin, A. 1978 Natural Decrease in Early Modern Cities: A Reconsideration. Past and Present 79: 126-138.

1981 Natural Decrease in Early Modern Cities: A Rejoinder. Past and Present 92: 175-180.

Silberbauer, G. 1972 The Gwi Bushmen. In Hunters and Gatherers Today, edited by M. Bicchieri, pp. 271-326. Holt, Reinhart, and Winston, New York.

Singh, G. and M. Siapush 2001 All-cause and Cause-Specific Mortality Among Immigrants and Native Born in the United States. American Journal of Public Health 91: 392-99.

Sisson, E. 1973 First Annual Report of the Coxcatlan Project. Robert S. Peabody Foundation for Archaeology, Andover.

349 1974 Second Annual Report of the Coxcatlan Project. Robert S. Peabody Foundation for Archaeology, Andover.

Sjoberg, G. 1960 The Preindustrial City, Past and Present. The Free Press, Glencoe.

Skinner, M. and A. Goodman 1991 Anthropological Uses of Developmental Defects of Enamel. In Skeletal Biology of Past Peoples: Research Methods, edited by S. Saunders and M. Katzenburg, pp. 153-174. Wiley-Liss, New York.

Smith, B. 1992 Standards of Human Tooth Formation and Dental Age Assessment. In Advances in Dental Anthropology, edited by M. Kelley and C. Larsen, pp. 143-168. Wiley- Liss, New York.

Smith, M. 1984 The Aztlan Migrations of the Nahuatl Chronicles: Myth or History? Ethnohistory 31: 153-186.

1992 Archaeological Research at Aztec-period Rural Sites in Morelos, Mexico. University of Pittsburg, Pittsburg.

1996 The Aztecs. Blackwell Press, Oxford.

Solar, P. 1995 Poor Relief and English Economic Development before the Industrial Revolution. Economic History Review 48: 1-22.

Spence, M. 1996 A Comparative Analysis of Ethnic Enclaves. In Arqueología Mesoamericana: Homenaje a William T. Sanders, edited by G. Mastache, J. Parsons, R. Santley, and M. Serra, pp. 333-353. INAH, Mexico.

Steckel, R. and J. Rose, editors 2002 The Backbone of History. Cambridge University Press, Cambridge.

Stone, A. D. Reents, and R. Coffman 1985 Genealogical Documentation of the Middle Classic Dynasty, Caracol, El Cayo, Belize. In Fourth Palenque Round Table, 1980, edited by E. Benson, pp. 267- 275. Pre-Colombian Art Research Institute, San Francisco.

Storey, Rebecca 1985 An Estimate of Mortality in a Precolumbian Urban Population. American Anthropologist 87: 519-535.

350 1992 Life and Death in the Ancient City of Teotihuacan. University of Alabama Press, Tuscaloosa.

Stuart-Macadam, P. 1985 Porotic Hyperostosis: Representative of a Childhood Condition. American Journal of Physical Anthropology 66: 391-398.

1987 New Evidence to Support the Anemia Theory. American Journal of Physical Anthropology 74: 511-520.

1988 Porotic Hyperostosis: Changing Interpretations. In Human Paleopathology: Current Syntheses and Future Options, edited by D. Ortner and A. Aufderhheide, pp. 36-39. Smithsonian Institution Press, Wahington, D.C.

Suárez, S. 1985 Un entierro del Clásico Superior en Cholula, Puebla. Cuaderno de Trabajo, No. 6. INAH, Mexico.

1990 Ultimos descubrimientos de entierros postclásicos en Cholula. INAH, Puebla.

1994 El policromo laca de Cholula, Puebla. In Mixteca-Puebla: Discoveries and Research in Mesoamerican Art and Archaeology, edited by H. Nicholson and E. Quiñnes Keber, pp. 45-51. Labyrinthos, California.

Swedlund, A. and G. Armelagos 1990 Disease in Populations in Transition. Bergin and Garvey, New York.

Tanaka, J. 1976 Subsistence Ecology of the Central Kalahari San. In Kalahari Hunter-Gatherers, edited by R. Lee and I. DeVore, pp.98-119. Harvard University Press, Cambridge.

Thomas, T. 1995 Acculturative Stress in the Adjustment of Immigrant Families. Journal of Social Distress and the Homeless 4: 131-142.

Tobias, P. 1976 Bushmen Hunter-Gatherers: A Study in Human Ecology. In Human Ecology: An Environmental Approach, edited by P. Richerson and J. McEvoy, pp. 271-284. Duxbury Press, Massachusetts.

Todd, T. 1921 Age Changes in the Public Bone. American Journal of Physical Anthropology 4: 407-424.

Torquemada, J. de

351 1943 Monarquía Indiana, Vol. I. Editorial S. Chávez Hayhoe, Mexico.

1976 Monarquía Indiana, Vol. III. Instituto de Investigaciones Históricas de la UNAM, Mexico.

Truswell, A. and J. Hansen 1976 Medical Research among the !Kung. In Kalahari Hunter-Gatherers, edited by R. Lee and I. DeVore , pp. 166-194. Harvard University Press, Cambridge.

Ubelaker, D. 1989 Human Skeletal Remains: Excavation, Analysis, and Interpretation. Taraxacum Press, Washington.

Uruñuela, G. 1989 Los entierros del proyecto: rescate drenaje Cholula 85. Notas Mesoamericanas 11: 64-80.

Uruñuela, G., A. Meza, and G. Hernández 1998 Patrones funerarios en Tetimpa. In Antropología e historia del occidente de México: XXIV Mesa Redonda, edited by R. Brambila, pp. 1821-1839. Sociedad Mexicana de Antropología and UNAM, Mexico.

Uruñuela, G. and P. Plunket 1998 Áreas de actividad en unidades domésticas del Formativo Terminal en Tetimpa, Puebla. Arqueología 20: 3-19.

2001 ―De piedra ha de ser la cama…‖? Las tumbas en el Formativo Terminal de Puebla, Tlaxcala y la Cuenca de México, a partir de la evidencia de Tetimpa, Puebla. Arqueología 25: 3-22.

2002 Lineages and Ancestors: The Formative Mortuary Assemblages of Tetimpa, Puebla. In Domestic Ritual in Ancient Mesoamerica, edited by P. Plunket, pp. 20- 30. The Cotsen Institute of Archaeology, California.

2005 La trancisión del Clásico al Posclásico : reflexiones sobre el Valle de Puebla- Tlaxcala. In Reacomodos demográficos del Clásico al Postclásico en el centro de México, edited by L. Manzanilla, pp. 303-324. UNAM, Mexico.

2007 Tradition and Transformation: Village Ritual at Tetimpa as a Template for Early Teotihuacan. In Commoner Ritual and Ideology in Ancient Mesoamerica, edited by N. Gonlin and J. Lohse, pp. 33-54. University of Colorado Press, Boulder.

Usher, B. 2000 A Multistate Model of Health and Mortality for Paleodemography: Tirup Cemetery. Unpublished Dissertation, The Pennsylvania State University.

352 Usher, B., G. Milner, and J. Boldsen 2000 Handbook for Paleopathology Coding Sheet. In A Multistate Model of Health and Mortality for Paleodemography: Tirup Cemetery, by B. Usher. Unpublished Ph.D. Dissertation, The Pennsylvania State University, Pennsylvania.

Vaillant, G. 1944 La civilización azteca. Fondo de Cultura Económica, Mexico.

Van der Woude, A. 1982 Population Developments in the Northern Netherlands (1500-1800) and the Validity of the ―Urban Graveyard‖ Effect. Annales de demographie historique: 55-75.

Vaupel, J. 1991 Relative’s Risks: Frailty Models of Life History Data. Theoretical Population Biology 37: 220-234.

Vaupel, J., T. Johnson, and G. Lithgow 1979 The impact of Heterogeneity in Individual Frailty on the Dynamics of Mortality. Demography 16: 439-454.

Vaupel, J. and A. Yashin 1985 Heterogeneity’s Ruses: Some Surprising Effects of Selection on Population Dynamics. The American Statistician 39: 176-185.

Waaler, H. 1984 Height, Weight, and Mortality: The Norwegian Experience. Acta Medica Scandinavica Supplementum 679: 1-56.

Walker, P. 1995 Problems of Preservation and Sexism in Sexing: Some Lessons from Historical Collections for Paleodemographers. In Grave Reflections: Portraying the Past through Cemetery Studies, edited by S. Saunders and A. Herring, pp. 31-47. Canadian Scholars’ Press, Toronto.

Walker, P., R. Bathurst, R. Richman, T. Gjerdrum, and V. Andrushko 2009 The Causes of Porotic Hyperostosis and Cribra Orbitalia: A Reappraisal of the Iron-Deficiency-Anemia Hypothesis. American Journal of Physical Anthropology 139: 109-125.

Weber, M. 1958 The City. The Free Press, New York.

Weinstein, R.

353 1991 New Urban Demands in Early Modern London. In Living and Dying in London, edited by W.F. Bynum and R. Porter, pp. 29-40. Wellcome Institute for the History of Medicine, London.

Weiss, K. 1973 Demographic Models for Anthropology. American Antiquities Memoir 27.

1990 The Biodemography of Variation in Human Frailty. Demography 27: 185-206.

White, T. 2000 Human Osteology. Academic Press, San Diego.

White, C., F. Longstaffe, and K. Law 2001 Revisiting the Teotihuacan Connection at Altun-Ha: Oxygen Isotope Analysis of Tomb f-8/1. Ancient Mesoamerica 12: 65-72.

White, C., F. Longstaffe, M. Spence, and K. Law 2000 Teotihuacan State Representation at Kaminaljuyu: The Evidence from Oxygen Isotopes. Journal of Anthropological Research 56: 535-558.

White, C., M. Spence, F. Longstaffe, and K. Law 2004a. Demography and Ethnic Continuity in the Tlailotlacan enclave of Teotihuacan: The Evidence from Stable Oxygen Isotopes. Journal of Anthropological Archaeology 23: 385-403.

White, C. M. Spence, F. Longstaffe, H. Stuart-Williams, and K. Law 2002 Geographic identities of the sacrificial victims from the Feathered Serpent Pyramid, Teotihuacan : implications for the nature of state power. Latin American Antiquity 13: 217-236.

White, C., M. Spence, H. Stuart-Williams, and H. Schwarcz 1997 Oxygen isotopes and the identification of geographical origins: the Valley of Oaxaca versus the Valley of Mexico. Journal of Archaeological Science 25: 643-655.

White, C., R. Storey, M. Spence, and F. Longstaffe 2004b Immigration, Assimilation, and Status in the Ancient City of Teotihuacan: Isotopic Evidence from Tlajinga 33. Latin American Antiquity 15: 176-198.

Whyte, I. 2000 Migration and Society in Britain: 1550-1830. MacMillan Press, Great Britain.

Wilmsen, E. 1982 Studies in Diet, Nutrition, and Fertility among a group of Kalahari Bushmen in Botswana. Social Science Information 21(1): 95-125.

354 Wirth, L. 1938 Urbanism as a Way of Life. American Journal of Sociology 44: 1-24.

Wood, J. 1998 A Theory of Preindustrial Population Dynamics. Current Anthropology 39: 99- 135.

Wood, J., D. Holman, K. O’Connor, and R. Ferrell 2002 Mortality Models for Paleodemography. In Paleodemography: Age Distributions from Skeletal Samples, edited by R. Hoppa and J. Vaupel, pp. 129-168. Cambridge University Press, Cambridge.

Wood, J. and G. Milner 1994 Reply to ―The Osteological Paradox Reconsidered.‖ Current Anthropology 35(5): 631-637.

Wood, J., G. Milner, H. Harpending, and K. Weiss 1992 The Osteological Paradox: Problems of Inferring Health from Skeletal Samples. Current Anthropolgy 33: 343-370.

Wright, L. 2005a Identifying Immigrants to Tikal, Guatemala: Defining Local Variability in Strontium Isotope Ratios of Human Tooth Enamel. Journal of Archaeological Science 32: 555-566.

2005b In Search of Yax Nuun Ayiin I: Revisiting the Tikal Project’s Burial 10. Ancient Mesoamerica 16: 89-100.

In press The Children of Kaminaljuyu: Isotopic Insight into Diet and Long-Distance Interaction in Mesoamerica. Journal of Anthropological Archaeology (2010).

Wright, L. and L. Chew 1999 Porotic Hyperostosis and Paleoepidemiology: A Forensic Perspective on Anemia among the Ancient Maya. American Anthropologist 100: 924-939.

Wright, L. and H. Schwarcz 1998 Stable carbon and oxygen isotopes in human tooth enamel: Identifying breastfeeding and weaning in prehistory. American Journal of Physical Anthropology 106: 1-18.

Wrigley, E. 1966 Family Reconstitution. In An Introduction to English Historical Demography, edited by E.A. Wrigley, pp. 96-159. Basic Books, Inc., New York.

1967 A Simple Model of London’s Importance in Changing English Society and Economy 1650-1750. Past and Present 37: 44-70.

355

1969 Population and History. McGraw-Hill, New York.

1975 Baptism Coverage in Early Nineteenth-Century England: The Colyton Area. Population Studies 29: 299-316.

1987 People, Cities, and Wealth: The Transformation of Traditional Society. Blackwell, Oxford.

Wrigley, E.A. and R.S. Schofield 1981 The Population History of England, 1541-1871: A Reconstruction. Edward Arnold, London.

1983 English Population History from Family Reconstitution: Summary Results 1600- 1799. Population Studies 37: 157-184.

Zacarias, B. 1975 Los enterramientos. In Teotenango, el antiguo lugar de la muralla, Vol. II, pp. 365-409. Gobierno del Estado de México, Toluca.

Zorita, A. de 1994 Life and Labor in Ancient Mexico. University of Oklahoma Press, Norman.

356 APPENDIX A

THE PROVENIENCE OF SKELETONS

Table A-1: Excavation unit in which each burial was found.

SKELETON # UNIT # DATING 133 ? CHOLULTECA II 137 GRANCALA 19j-CUADRO1 CHOLULTECA II 139 ? CHOLULTECA II 76B-2 20L-3C CHOLULTECA III 88B 21L-11E/11F CHOLULTECA III 107 GRAN CALA CHOLULTECA III 108 GRAN CALA CHOLULTECA III 115 21K-35T CHOLULTECA III 180/181- 3ASSOC 21K-28G CHOLULTECA III 198-2 21K-24H CHOLULTECA III 202-2 21K-34S CHOLULTECA III 251 21K-40X CHOLULTECA III 280 21K-40D CHOLULTECA III 421 20J-39N CHOLULTECA III VASIJA9-2 19J-34B CHOLULTECA III 154 21K-38Y CHOLULTECA II 80 21L-10C CHOLULTECA III 100-2 19J-CUADRO18 CHOLULTECA III 149 19J-CUADRO1 CHOLULTECA II 104-2 GRAN CALA CHOLULTECA III 122 21L-28M? CHOLULTECA III 132 21L-32A CHOLULTECA III 261 21K-40F CHOLULTECA III 270 21K-41D CHOLULTECA III 142 19J-CUADRO4 CHOLULTECA II 174-1 21K-26G CHOLULTECA III 264 21K-40F CHOLULTECA III 282 21K-40G CHOLULTECA III 418 20J-38K CHOLULTECA III 419 20J-38K CHOLULTECA III VASIJA9-1 19J-34B CHOLULTECA III VASIJA13 CALA1-19J CHOLULTECA III 124 GRANCALA-19J CHOLULTECA II 140 19J-SEC1-CUADRO4 CHOLULTECA II 76A 20L-2C CHOLULTECA III 111A GRAN CALA-CUADRO17 CHOLULTECA III 128 GRAN CALA-CUADRO2 CHOLULTECA III

357 174-2 21K-26G CHOLULTECA III 207 21K-28G CHOLULTECA III VASIJA12 21L-28M CHOLULTECA III 129 GRANCALA-CUADRO4 CHOLULTECA II 143 19J-CUADRO5 CHOLULTECA II 203 21K-35S,SEC3 CHOLULTECA II 76B-1 21L-2C/3C CHOLULTECA III 79 21L-6Z CHOLULTECA III 83A 21L-11G CHOLULTECA III 95 21L-2H CHOLULTECA III 114 21K-35T CHOLULTECA III 116 19J-CUADRO5 CHOLULTECA III 175 21K-26G CHOLULTECA III 198 21K-24H CHOLULTECA III 247 21K-33V CHOLULTECA III 248 21K-33V CHOLULTECA III 265 21K-40F CHOLULTECA III 320 21K-43W CHOLULTECA III 141 19J-SEC1-CUADRO4 CHOLULTECA II 222 12K-27K CHOLULTECA II 68C 21L-7H CHOLULTECA III 135 21L-32A CHOLULTECA III 269 21K-40G/D CHOLULTECA III 92 ? CHOLULTECA III 97 19J-34B CHOLULTECA III 104-1 GRAN CALA CHOLULTECA III 182-2 21K-36S CHOLULTECA III 229-2 21K-26G CHOLULTECA III 372-3 20J-34X CHOLULTECA II 90A ? CHOLULTECA III 188 21K-36S CHOLULTECA III 192-1 21K-24H CHOLULTECA III 202-1 21K-24G CHOLULTECA III 225 21L-4H CHOLULTECA III 308 21L-26M CHOLULTECA III 405 20J-41Y CHOLULTECA III 90B ? CHOLULTECA III 380 20K-37B CHOLULTECA III 130 ? CHOLULTECA II 415a 20J-40O CHOLULTECA II 131 21L-32A CHOLULTECA III 173a 21K-26S CHOLULTECA III 262 21K-40F CHOLULTECA III 271 21K-10T CHOLULTECA III 319 21L-42G CHOLULTECA III 348 20L-1A CHOLULTECA III 368 18K-3A CHOLULTECA III 373-2 20K-17U CHOLULTECA III

358 403 21J-41Y CHOLULTECA III 82A-2 21L-9A CHOLULTECA III GRAN CALA 19J- 110 CUADRO15 CHOLULTECA III 230 21K-19M CHOLULTECA III 268-1 21K-14U CHOLULTECA III 283 21K-40H CHOLULTECA III 150 19J-CUADRO7 CHOLULTECA II 231 19J-31D CHOLULTECA II 313 21K-42W CHOLULTECA II 78C-2 21K-16N~ CHOLULTECA III 182-1 21K-36S CHOLULTECA III 232 21K-19M CHOLULTECA III 267 21K-14U CHOLULTECA III VASIJA10 ? CHOLULTECA III 104-3 GRAN CALA CHOLULTECA III 387 20K-44C CHOLULTECA III 83B 21L-11G CHOLULTECA III 400-1 20J-33X CHOLULTECA III 400-2 20J-33X CHOLULTECA III 422 20J-39F CHOLULTECA III 196 21K-35S CHOLULTECA III 213 21K-24G CHOLULTECA III 239 21K-7V CHOLULTECA III 266-1 21K-14U CHOLULTECA III 76 21L-6Y CHOLULTECA III 384 19J-32F CHOLULTECA III 101 7A-7Z CHOLULTECA II 134 21L-32A CHOLULTECA III 256 21K-40D CHOLULTECA III 278 21K-40G CHOLULTECA II 103 19J-CUADRO18 CHOLULTECA III 78C-1 21K-16N~ CHOLULTECA III 400-3 20J-33X CHOLULTECA III 102 7A-7Z CHOLULTECA II 94 21L-8B CHOLULTECA III 229-1 21K-26G CHOLULTECA III 268-2 21K-14U CHOLULTECA III 69A-2 21L-7H CHOLULTECA II 408 20J-40W CHOLULTECA III 326 20L-4A CHOLULTECA III 377 22K-5X CHOLULTECA III 285-2 21K-14U CHOLULTECA III 295 21K-39Y CHOLULTECA II 296 21K-40Y CHOLULTECA II 100-1 19J-CUADRO 1 CHOLULTECA III 100-3 19J-CUADRO18 CHOLULTECA III 183 21K-32T CHOLULTECA III 191 21K-24H CHOLULTECA III

359 214 21K-24H CHOLULTECA III 216 21L-4W CHOLULTECA III 341 21K-12M CHOLULTECA III 404-2 20J-41Y CHOLULTECA III 411-2b ? CHOLULTECA II 266-2 21K-14U CHOLULTECA III 69C 21L-14G CHOLULTECA II 328 21K-32Z CHOLULTECA III 433 20J-39O CHOLULTECA III 321 21K-42X CHOLULTECA III 173c ? CHOLULTECA III 263 21K-40F CHOLULTECA III 199 21K-24H CHOLULTECA III 404-1 20J-41Y CHOLULTECA III 151 19J-CUADRO4 CHOLULTECA II 127 21L-27M CHOLULTECA II 69A-1 21L-14G CHOLULTECA II 397-1 21K-1N~ CHOLULTECA II 173b ? CHOLULTECA III 299 21K-39Y CHOLULTECA II 347-1 20L-1A CHOLULTECA III 88A 21L-11E CHOLULTECA III 259 21K-41D CHOLULTECA III 72A 21K-21P,21L-6H CHOLULTECA II 345 20L-1A CHOLULTECA III 240 21K-8V CHOLULTECA III 210-1 21K-36S CHOLULTECA III 86A 21L-11E CHOLULTECA III 184 21K-32T CHOLULTECA III 304 19K-3M CHOLULTECA III 302 21K-39Z CHOLULTECA II 228 19J-30D CHOLULTECA II 417 20'J-42K CHOLULTECA II 301 21K-40Z CHOLULTECA II 260 21K-41D CHOLULTECA III 155 21L-9L CHOLULTECA III 77A 21L-12E,21L-12F CHOLULTECA II 432 20J-39P CHOLULTECA III 69B 21L-14G CHOLULTECA II 208 21K-23R CHOLULTECA III 186 21K-24H CHOLULTECA III 176 21K-36S CHOLULTECA III 370b ? CHOLULTECA III 427 20J-38O CHOLULTECA III 245 21K-29F CHOLULTECA III 211 21K-36S CHOLULTECA III 371 21K-1N CHOLULTECA II 318 21L-9A CHOLULTECA III

360 178 ? CHOLULTECA II 347-2 21L-1A CHOLULTECA III 370a ? CHOLULTECA III 147 21K-37P CHOLULTECA III 331 21K-43Z CHOLULTECA III 126 19J-CUADRO2 CHOLULTECA III 382 22K-11X CHOLULTECA III 385a 20K-43C CHOLULTECA III 68B 21L-7H/7I CHOLULTECA III 69E 21L-14G,21L-12E CHOLULTECA II 193 21K-24H CHOLULTECA III 388 20K-42C CHOLULTECA III 424 20J-39F CHOLULTECA III 426-1 20J-39P CHOLULTECA III 434 20J-39P CHOLULTECA III 396 21K-1N~ CHOLULTECA II 276 21K-40F CHOLULTECA III 281 21K-40F CHOLULTECA III 185-1 21K-32T CHOLULTECA III 68A 21L-7H/7I CHOLULTECA III 279 21K-23N CHOLULTECA II 257 21K-41D CHOLULTECA III 87A ? CHOLULTECA III 187 21K-32T CHOLULTECA III 244 21K-12P CHOLULTECA III 416 20J-42K CHOLULTECA III 385b ? CHOLULTECA III 190 21K-36S CHOLULTECA III 298 21K-40Y CHOLULTECA II 290 21K-12L CHOLULTECA III 414-1&2a 20J-40P CHOLULTECA II 83C 21L-11G CHOLULTECA III 317 21L-9A CHOLULTECA III 426-2 20J-39P CHOLULTECA III 372-1 20J-34X CHOLULTECA II 374 20K-17U CHOLULTECA III 413-1 20J-40P CHOLULTECA II 81 20L-2C/2D CHOLULTECA III 423 20J-39F CHOLULTECA III 376 22K-5X CHOLULTECA III 221 21L-4R CHOLULTECA III 292 21K-40Y CHOLULTECA II 294 21K-40Y CHOLULTECA II 398 21K-1O CHOLULTECA II 88C 21L-11E CHOLULTECA III 227 19J-30D CHOLULTECA II 78B 21K-16N~ CHOLULTECA III 327 21K-43Z CHOLULTECA III

361 413-2b 20J-40P CHOLULTECA II 375 20K-16U CHOLULTECA III 404-3 20J-40Y CHOLULTECA III 246-1 21K-8V CHOLULTECA III 153 21K-38V? CHOLULTECA III 383 19J-33P CHOLULTECA III 386 20K-42C CHOLULTECA III 395 21K-1O CHOLULTECA II 415c ? CHOLULTECA II 373-1 20K-17U CHOLULTECA III 394-2 19J-32K CHOLULTECA II GRAN CALA19J- 112 CUADRO2 CHOLULTECA III 204 21K-35S CHOLULTECA III 394-3 19J-32K CHOLULTECA II 156 ? CHOLULTECA III 212 21K-24H CHOLULTECA III 347-3 ? CHOLULTECA III 164 21K-36S CHOLULTECA III 291 21K-14Y CHOLULTECA II 372-2 20J-34X CHOLULTECA II 297-1 21K-39Z CHOLULTECA II 250 21K-40D CHOLULTECA III 300 21K-39Y/39Z CHOLULTECA II 78A 21K-16N~ CHOLULTECA III 415b ? CHOLULTECA II 246-3 21K-8V CHOLULTECA III 200 21K-35S CHOLULTECA III 201 21K-34S CHOLULTECA III 157 ? CHOLULTECA III 394-1 19J-32K CHOLULTECA II 407 20J-41Y CHOLULTECA III 430 20J-39O CHOLULTECA III 189b ? CHOLULTECA II 343 20L-1A CHOLULTECA III 255 21K-41D CHOLULTECA III 206 21K-33Q CHOLULTECA III 226 19J-30D CHOLULTECA II 420 20J-38L CHOLULTECA III 82B 21L-9A CHOLULTECA III 406 20J-41Y CHOLULTECA III 246-2 21K-8V CHOLULTECA III 189-1 21K-33S CHOLULTECA II 428 20J-39P CHOLULTECA III 413-2a ? CHOLULTECA II 293-1 21K-40Y CHOLULTECA II 284 21K-40G CHOLULTECA III 152 21K-36C CHOLULTECA III 409 20J-40W CHOLULTECA II

362 411-1 20J-40Q CHOLULTECA II 414-1&2b ? CHOLULTECA II

363 APPENDIX B

AGE AND SEX DATA

Table B-1: Age and sex for each skeleton SKELETON # AGE SEX 133 0.000 X 137 0.000 X 139 0.000 X 76B-2 0.000 X 88B 0.000 X 107 0.000 X 108 0.000 X 115 0.000 X 180/181- 3ASSOC 0.000 X 198-2 0.000 X 202-2 0.000 X 251 0.000 X 280 0.000 X 421 0.000 X VASIJA9-2 0.000 X 154 0.083 X 80 0.250 X 100-2 0.250 X 149 0.330 X 104-2 0.500 X 122 0.500 X 132 0.500 X 261 0.500 X 270 0.500 X 142 0.750 X 174-1 0.750 X 264 0.750 X 282 0.750 X 418 0.750 X 419 0.750 X VASIJA9-1 0.750 X VASIJA13 0.750 X 124 1.000 X 140 1.000 X 76A 1.000 X 111A 1.000 X 128 1.000 X 174-2 1.000 X

364 207 1.000 X VASIJA12 1.000 X 129 1.500 X 143 1.500 X 203 1.500 X 76B-1 1.500 X 79 1.500 X 83A 1.500 X 95 1.500 X 114 1.500 X 116 1.500 X 175 1.500 X 198 1.500 X 247 1.500 X 248 1.500 X 265 1.500 X 320 1.500 X 141 2.000 X 222 2.000 X 68C 2.000 X 135 2.000 X 269 2.000 X 92 2.500 X 97 2.500 X 104-1 2.500 X 182-2 2.500 X 229-2 2.500 X 372-3 3.000 X 90A 3.000 X 188 3.000 X 192-1 3.000 X 202-1 3.000 X 225 3.000 X 308 3.000 X 405 3.000 X 90B 3.500 X 380 3.500 X 130 4.000 X 415a 4.000 X 131 4.000 X 173a 4.000 X 262 4.000 X 271 4.000 X 319 4.000 X 348 4.000 X 368 4.000 X 373-2 4.000 X 403 4.000 X

365 82A-2 4.500 X 110 4.500 X 230 4.500 X 268-1 4.500 X 283 4.500 X 150 5.000 X 231 5.000 X 313 5.000 X 78C-2 5.000 X 182-1 5.000 X 232 5.000 X 267 5.000 X VASIJA10 5.500 X 104-3 6.000 X 387 6.000 X 83B 6.500 X 400-1 6.500 X 400-2 6.500 X 422 6.500 X 196 7.000 X 213 7.000 X 239 7.000 X 266-1 7.000 X 76 7.500 X 384 8.000 X 101 8.500 X 134 8.500 X 256 8.500 X 278 9.000 X 103 9.000 X 78C-1 10.000 X 400-3 10.000 X 102 10.500 X 94 10.500 X 229-1 11.000 X 268-2 11.500 X 69A-2 12.000 X 408 12.000 X 326 13.500 X 377 13.500 X 285-2 14.500 X 295 15.000 X 296 15.000 F 100-1 15.000 F 100-3 15.000 F 183 15.000 M? 191 15.000 X 214 15.000 F?

366 216 15.000 F? 341 15.000 X 404-2 15.000 F 411-2b 15.500 X 266-2 15.500 X 69C 16.060 M 328 17.500 F 433 17.500 F? 321 17.810 F 173c 18.000 M 263 18.000 X 199 18.500 M? 404-1 19.240 F 151 19.630 F 127 21.710 M 69A-1 21.800 M 397-1 22.070 F 173b 22.500 M 299 23.090 M 347-1 23.260 M 88A 24.250 F 259 25.290 M 72A 26.760 F 345 28.080 M 240 28.520 M 210-1 28.580 M 86A 29.040 F 184 29.220 F 304 31.560 F 302 31.910 M 228 32.190 M 417 34.500 M 301 35.860 M 260 36.010 M 155 37.000 F 77A 37.400 F 432 39.830 M 69B 40.610 M 208 40.720 M 186 40.850 F 176 41.930 F 370b 42.060 M 427 42.110 F 245 42.260 F 211 44.730 F 371 46.150 M 318 47.070 F 178 48.130 F

367 347-2 48.230 M 370a 49.370 M 147 49.500 F 331 50.030 M 126 50.120 F 382 50.420 F 385a 50.540 M 68B 50.710 F 69E 50.770 M 193 50.770 F 388 50.770 M? 424 50.770 M 426-1 50.770 M 434 50.770 M 396 51.280 F 276 51.580 M 281 51.830 F 185-1 51.910 M 68A 51.970 F 279 53.000 F 257 53.090 M 87A 53.460 M 187 53.560 F 244 53.630 M 416 53.730 F 385b 53.990 M 190 54.050 F 298 54.430 M 290 54.490 M 414-1&2a 54.590 F 83C 54.590 M 317 54.790 M 426-2 54.830 F 372-1 54.970 F 374 55.440 M 413-1 55.490 F 81 55.970 F 423 56.070 F 376 56.190 M 221 56.390 M 292 56.650 M 294 56.990 F 398 57.230 F 88C 57.260 M 227 57.540 M 78B 57.660 M 327 57.920 M 413-2b 57.990 F

368 375 58.390 F 404-3 58.400 F 246-1 58.630 F 153 59.010 M 383 59.420 F 386 59.680 F 395 59.760 M 415c 59.990 M 373-1 60.070 F 394-2 60.230 F 112 60.880 M 204 61.100 F 394-3 61.210 F 156 61.330 M 212 61.610 F 347-3 61.710 M 164 62.060 M 291 62.570 F 372-2 62.600 F 297-1 63.450 M 250 64.320 M 300 64.360 M 78A 64.510 F 415b 64.740 F 246-3 64.850 F 200 64.920 F 201 65.320 F 157 65.640 M 394-1 65.930 F 407 66.090 F 430 66.360 M 189b 66.930 M 343 67.770 F 255 67.970 M 206 68.010 M 226 68.250 M 420 68.470 M 82B 68.840 F 406 69.150 F 246-2 69.270 M 189-1 71.110 F 428 71.150 M 413-2a 71.460 F 293-1 72.270 F 284 75.060 F 152 85.360 F 409 110.000 F 411-1 110.000 M

369 414-1&2b 110.000 M

370

APPENDIX C

SEX DETERMINATION

Sex determination was based primarily on features of the skull and pelvis, although a subjective assessment of general robusticity was considered as well. The following five features of the skull were scored on a scale of 1 to 5, according to the scoring system presented in

Buikstra and Ubelaker (1994): the supraorbital ridge, the mastoid process, the occipital protuberance, the gonial angle, and the mental eminence (see Figure C-1 below for the identification of these features and the scoring procedures used). The following five features of the pelvis were also scored on a scale of 1 to 5: the ventral arc, the subpubic concavity, the ischiopubic ramus, the greater sciatic notch, and the preauricular sulcus (see Figures C-2 and C-3 below for identification of these features). Attributes of the pelvis were scored as follows:

X = unobservable 1 = male 2 = probable male 3 = ambiguous 4 = probable female 5 = female

Dorsal pitting, scarring of the dorsal aspect of the pubic symphysis due to childbirth, was also noted and taken into consideration when determining sex. Characteristics of the pelvis were given more weight than those of the skull whenever a discrepancy between the two arose.

371

Figure C-1: Features of the skull and scoring system used to determine sex. (Buikstra and Ubelaker 1994).

372

Figure C-2 : Features of the pelvis used to determine sex (Buikstra and Ubelaker 1994: 18-19).

373

Figure C-3: Features of the pubic symphysis used to determine sex. (Buikstra and Ubelaker 1994)

374 APPENDIX D

JUVENILE AGE ESTIMATION

Juvenile age determination was primarily based upon dental development and epiphyseal closure. In cases in which neither of these methods was possible, diaphyseal length was used.

Dental development and the stage of epiphyseal closure were preferred over diaphyseal length since the rate of development of these features, particularly dental development, is more genetically stable across populations than is diaphyseal length. In all cases, age was assigned using the criteria established by Ubelaker (1989) for the Arikara population. As these data pertain to a North American indigenous population, they were considered to be more appropriate than data from European or American children to reconstruct juvenile age for the Cholula collection.

375

Figure D-1: Chart used to assign age to juveniles based on dental development. The stage of development and eruption of available teeth are compared to the diagram to determine the age of the individual (Buikstra and Ubelaker 1994: 51).

376 Table D-1: Chart used to assign ages to juveniles based on epiphyseal closure. (Ubelaker 1989)

Table D-2: Chart used to assign ages to juveniles based upon diaphyseal length (Ubelaker 1989).

377 APPENDIX E

SCORING PROCEDURE FOR TRANSITION ANALYSIS

The following table shows the scoring procedure for transition analysis. A detailed description of the criteria for determining scores can be found in Boldsen et al. (2002). Screen shots of the ADBOU age estimation program are also shown.

378

Table E-1: The scoring procedures used for transition analysis. (from Boldsen et al. 2002)

SKELETAL FEATURES SCORES SKELETAL FEATURES SCORES

Pubic Symphysis Cranial Sutures: (continued): coronal pterica 1. open Dorsal Symphyseal Margin 1. serrated sagittal obelica 2. juxtaposed 2. flattening incomplete lambdoidal asterica 3. partially obliterated 3. flattening complete zygomaticomaxillary 4. punctuated 4. rim interpalatine 5. obliterated 5. breakdown

Pubic Symphysis: Auricular Surface:

Superior and Inferior Symphyseal Relief 1. sharp billows Demiface 1. undulating 2. soft, deep billows Topography 2. median elevation 3. soft, shallow billows 3. flat to irregular 4. residual billows 5. flat Superior, apical, and 1. billows over >2/3 of the surface 2. billows over 1/3 to 2/3 of the 6. irregular inferior surface morphology surface 3. billows over <1/3 of the surface Symphyseal Texture 1. smooth 4. flat 2. coarse 5. bumps 3. microporosity 4. macroporosity Inferior surface texture 1. smooth 2. microporosity Superior Apex 1. no protuberance 3. macroporosity 2. early protuberance Superior and inferior 3. late protuberance posterior 1. smooth 4. integrated iliac exostoses 2. rounded exostoses 3. pointed exostoses Ventral Symphyseal Margin 1. serrated 4. jagged exostoses 2. beveled 5. touching exostoses 3. rampart formation 6. fused 4. rampart completion I 5. rampart completion II Posterior Iliac Exostoses 1. smooth 6. rim 2. rounded exostoses 7. breakdown 3. pointed spicules

379

Figure E-1: Screen shot from the ADBOU age estimation program showing the data entry screen.

380

Figure E-2: Screen shot from the ADBOU age estimation program showing the age calculation for a particular individual.

381 APPENDIX F

TEST OF PROFICIENCY

In order to determine my proficiency in applying transition analysis, I performed a blind test on a small sample of casts of known-age pubic symphyses supplied by France Casting as part of the Suchey-Brooks system prior to beginning my study of the Cholula material. Only four features of the pubic symphysis were used: Symphyseal texture was not scored because casting obscures this feature. The results of this test appear in the table below. The average difference between the estimated and the actual age is 5.56 years. Twenty-seven out of 33 age estimates fall within one standard deviation, and 32 out of 33 fall within 2 standard deviations.

The one individual who was not aged within the 95% confidence interval was a 22 year old aged between 15 and 20.83.

382

Table F-1: Results of a test of accuracy in applying transition analysis.

Specimen Estimated Age Standard error 95% Confidence Interval Actual Age Difference 6 58.04 22.46 31.43-98 58 -0.04 7 15 -1 15-20.83 16 1 8 31.49 4.59 24.4-43.02 25 -6.49 9 37.29 7.01 27.01-59.4 48 10.71 10 78.43 21.1 37.27--1 63 -15.43 11 31.07 4.44 24-42.18 34 2.93 12 22.12 2.56 17.48-27.81 22 -0.12 13 42.79 11.54 28.84-80.57 38 -4.79 14 15 -1 15-20.83 18 3 15 35.87 6.28 26.4-54.22 34 -1.87 16 80.09 19.16 41.1--1 68 -12.09 17 52.57 26.53 30.52--1 68 15.43 1 AB 28.64 4.07 22.17-39.56 29 0.36 1 CD 78.95 20.55 40.18--1 75 -3.95 2 AB 23.77 2.43 19.64-29.1 21 -2.77 2 CD 18.79 2.47 -1-23.59 23 4.21 3 AB 30.96 6.59 21.93-52.2 27 -3.96 3 CD 24.8 2.64 20.44-30.98 25 0.2 4 AB 37.57 5.96 28.32-55.34 45 7.43 4 CD 56.34 19.65 28.59-96.87 48 -8.34 5 AB 31.31 3.95 24.85-40.7 26 -5.31 6 CD 89.81 17.68 56.23--1 83 -6.81 7 CD 26.63 2.86 21.81-33.49 29 2.37 8 CD 45.83 15.05 26.94-84.95 42 -3.83 9 CD 23.43 2.47 19.22-29.07 25 1.57 10 CD 89.69 19.95 53.46--1 78 -11.69 11 CD 45.83 15.05 26.94-84-95 48 2.17 12 CD 110 -1 72.63-110 79 -31 13 CD 48.66 13.81 29.48-80.95 42 -6.66 14 C 19.9 2.32 -1-25.08 20 0.1 15 D 15 -1 15-20.83 22 7 16 C 25.85 4.06 19.65-37.24 26 0.15

383 APPENDIX G

TRADITIONAL AGING METHODS

In addition to transition analysis, Todd (1921), McKern and Stewart (1957), Gilbert and

McKern (1973), Brooks and Suchey (1990), and Lovejoy et al. (1985) were used to age adult skeletons in the Cholula collection. The illustrations below provide an overview of these methods. For detailed descriptions of each phase, refer to the original source.

Figure G-1 : Todd’s ten stage method of aging using the pubic symphysis (Todd 1921).

384

Figure G-2: McKern and Stewart’s three component method of aging using the pubic symphysis (McKern and Stewart 1957). McKern and Stewart’s method is applicable only to males. A complimentary method of aging female pubic symphyses was developed by Gilbert and McKern (1973) and is very similar to McKern and Stewart’s technique.

385

Figure G-3: Brooks and Suchey’s six phase method of aging the pubic symphysis (Brooks and Suchey 1990).

386

Figure G-4: Lovejoy et al.’s method of aging the auricular surface of the ilium (White 2000: 358).

387 APPENDIX H

CRIBRA ORBITALIA AND POROTIC HYPEROSTOSIS

X = unobservable 1 = absent 2 = present

Table H-1: Cribra orbitalia and porotic hyperostosis

SKELETON # AGE Frontal Parietal Occipital Orbits 133 0.000 1 1 1 1 137 0.000 1 1 1 1 139 0.000 1 1 1 1 76B-2 0.000 X X X X 88B 0.000 1 X 1 1 107 0.000 1 1 X 1 108 0.000 X X X X 115 0.000 X 1 X X 180/181- 3ASSOC 0.000 1 1 1 1 198-2 0.000 X X X X 202-2 0.000 X X X X 251 0.000 X X X X 280 0.000 1 1 1 1 421 0.000 1 1 X 1 VASIJA9-2 0.000 X 1 X X 154 0.083 1 1 1 1 80 0.250 X X 1 X 100-2 0.250 X X X X 149 0.330 2 2 1 1 104-2 0.500 X 1 X X 122 0.500 1 1 1 X 132 0.500 X 1 X 1 261 0.500 X X X 1 270 0.500 1 1 1 1 142 0.750 1 1 1 1 174-1 0.750 1 2 1 2 264 0.750 X X X 1 282 0.750 1 X X 1 418 0.750 1 1 1 2 419 0.750 1 2 X 2 VASIJA9-1 0.750 1 1 1 1 VASIJA13 0.750 X X X X

388 124 1.000 X X X X 140 1.000 1 1 1 1 76A 1.000 X X X X 111A 1.000 X X X 1 128 1.000 1 1 1 X 174-2 1.000 X X X X 207 1.000 1 2 2 X VASIJA12 1.000 X X X X 129 1.500 X 1 1 X 143 1.500 1 1 1 1 203 1.500 1 1 1 1 76B-1 1.500 1 1 X X 79 1.500 1 X 1 1 83A 1.500 1 1 X 1 95 1.500 1 1 1 X 114 1.500 1 1 X 1 116 1.500 1 1 1 1 175 1.500 1 1 1 1 198 1.500 1 1 X X 247 1.500 1 1 2 1 248 1.500 1 1 1 1 265 1.500 X 1 X 1 320 1.500 2 2 2 2 141 2.000 1 2 X 2 222 2.000 X 1 1 X 68C 2.000 X X X X 135 2.000 1 1 X X 269 2.000 1 1 1 1 92 2.500 1 1 X X 97 2.500 1 1 X X 104-1 2.500 1 1 X 1 182-2 2.500 X 1 X X 229-2 2.500 1 X X 1 372-3 3.000 1 1 1 1 90A 3.000 X X X 1 188 3.000 1 X 1 X 192-1 3.000 X X X X 202-1 3.000 X X X X 225 3.000 X 1 1 X 308 3.000 X 1 X X 405 3.000 1 1 1 1 90B 3.500 X X X X 380 3.500 1 1 1 2 130 4.000 1 2 2 X 415a 4.000 X X X X 131 4.000 1 1 X 1 173a 4.000 X X X X 262 4.000 1 1 1 1

389 271 4.000 1 1 1 1 319 4.000 1 1 1 X 348 4.000 1 1 1 X 368 4.000 1 1 1 1 373-2 4.000 X 1 1 X 403 4.000 X X X 1 82A-2 4.500 1 1 X 1 110 4.500 1 1 1 2 230 4.500 X X X X 268-1 4.500 X 1 X X 283 4.500 1 1 1 1 150 5.000 1 1 1 1 231 5.000 X X X X 313 5.000 X X X 1 78C-2 5.000 X X X X 182-1 5.000 X 1 1 X 232 5.000 X 1 X X 267 5.000 1 1 1 1 VASIJA10 5.500 1 1 1 1 104-3 6.000 X 1 X X 387 6.000 1 1 1 X 83B 6.500 1 1 1 1 400-1 6.500 1 X X 1 400-2 6.500 1 1 1 1 422 6.500 1 1 1 1 196 7.000 X X X X 213 7.000 X X X X 239 7.000 1 1 X X 266-1 7.000 X 1 1 1 76 7.500 1 1 1 1 384 8.000 X 1 1 X 101 8.500 X X X X 134 8.500 1 1 1 X 256 8.500 1 2 X 2 278 9.000 1 1 1 1 103 9.000 1 2 2 2 78C-1 10.000 2 X X 2 400-3 10.000 1 1 1 X 102 10.500 X X X X 94 10.500 1 X 1 1 229-1 11.000 1 1 1 X 268-2 11.500 X X X X 69A-2 12.000 1 X X X 408 12.000 1 2 2 1 326 13.500 1 1 1 X 377 13.500 X X X X 285-2 14.500 X X X 1 295 15.000 1 1 1 1

390 296 15.000 1 1 1 1 100-1 15.000 X X X X 100-3 15.000 1 1 1 X 183 15.000 X X X X 191 15.000 X X X X 214 15.000 X X X X 216 15.000 1 1 1 X 341 15.000 1 1 1 1 404-2 15.000 1 2 2 2 411-2b 15.500 X X X X 266-2 15.500 X X X X 69C 16.060 X X X X 328 17.500 1 1 1 2 433 17.500 1 X X 1 321 17.810 X X X X 173c 18.000 X X X X 263 18.000 X X X X 199 18.500 1 1 1 X 404-1 19.240 X 1 1 1 151 19.630 1 1 1 X 127 21.710 1 1 X X 69A-1 21.800 1 1 X X 397-1 22.070 X 1 1 X 173b 22.500 X X X X 299 23.090 X X X X 347-1 23.260 1 1 1 1 88A 24.250 1 1 X 1 259 25.290 1 1 1 1 72A 26.760 X 1 1 1 345 28.080 1 1 1 1 240 28.520 1 2 2 1 210-1 28.580 1 1 1 1 86A 29.040 1 1 1 1 184 29.220 X X X X 304 31.560 1 1 1 1 302 31.910 1 1 1 X 228 32.190 1 1 1 X 417 34.500 X X X X 301 35.860 X 1 1 X 260 36.010 X X X X 155 37.000 1 1 1 1 77A 37.400 1 1 1 1 432 39.830 X 1 1 X 69B 40.610 X X X X 208 40.720 X X X X 186 40.850 1 1 1 1 176 41.930 1 1 1 X 370b 42.060 1 1 1 X

391 427 42.110 1 1 1 1 245 42.260 1 2 1 1 211 44.730 X X X X 371 46.150 1 1 1 X 318 47.070 1 1 1 X 178 48.130 1 1 1 1 347-2 48.230 1 1 X 1 370a 49.370 1 1 X 1 147 49.500 X X X X 331 50.030 1 1 1 1 126 50.120 X 1 1 X 382 50.420 X X X X 385a 50.540 1 1 1 1 68B 50.710 1 1 1 X 69E 50.770 X 1 X X 193 50.770 X X X X 388 50.770 X X X X 424 50.770 1 1 X X 426-1 50.770 X 1 X X 434 50.770 X X X X 396 51.280 X X X X 276 51.580 1 1 1 1 281 51.830 1 1 1 1 185-1 51.910 X 1 1 X 68A 51.970 1 1 1 X 279 53.000 1 1 1 1 257 53.090 1 2 2 1 87A 53.460 1 1 1 X 187 53.560 1 2 1 X 244 53.630 1 1 1 1 416 53.730 X X X X 385b 53.990 1 1 1 1 190 54.050 X 1 1 X 298 54.430 1 1 1 X 290 54.490 1 1 1 1 414-1&2a 54.590 1 X X 1 83C 54.590 1 1 1 X 317 54.790 1 1 1 1 426-2 54.830 X X X X 372-1 54.970 1 1 1 1 374 55.440 1 1 1 1 413-1 55.490 1 1 1 X 81 55.970 X X X X 423 56.070 1 1 1 X 376 56.190 X X X X 221 56.390 X X 1 X 292 56.650 1 X X 1 294 56.990 1 1 1 X

392 398 57.230 X X X X 88C 57.260 1 1 1 1 227 57.540 X 1 X X 78B 57.660 X X X X 327 57.920 X X X X 413-2b 57.990 X X X X 375 58.390 1 1 1 1 404-3 58.400 X 1 1 X 246-1 58.630 1 1 1 X 153 59.010 X X 1 X 383 59.420 1 1 1 1 386 59.680 X 1 X X 395 59.760 1 1 1 1 415c 59.990 X X X X 373-1 60.070 1 1 1 1 394-2 60.230 X X X X 112 60.880 1 1 1 1 204 61.100 X 1 X 1 394-3 61.210 X X X X 156 61.330 1 2 2 X 212 61.610 1 X X X 347-3 61.710 1 1 1 X 164 62.060 1 1 1 X 291 62.570 1 2 2 1 372-2 62.600 1 1 1 1 297-1 63.450 1 1 1 1 250 64.320 1 1 1 1 300 64.360 X X X X 78A 64.510 1 1 1 1 415b 64.740 1 1 1 1 246-3 64.850 1 1 1 X 200 64.920 1 1 1 1 201 65.320 X 1 1 X 157 65.640 X X X X 394-1 65.930 X X X X 407 66.090 X 1 1 X 430 66.360 X X X 1 189b 66.930 X X X X 343 67.770 1 1 1 1 255 67.970 1 1 1 1 206 68.010 X X X X 226 68.250 X X X X 420 68.470 1 X X X 82B 68.840 1 1 1 X 406 69.150 X X X X 246-2 69.270 1 1 1 1 189-1 71.110 X X X X 428 71.150 X X X X

393 413-2a 71.460 1 X X X 293-1 72.270 1 1 1 1 284 75.060 1 1 1 1 152 85.360 1 1 1 1 409 110.000 X 1 1 X 411-1 110.000 X X X X 414-1&2b 110.000 X X X X

394 APPENDIX I

ENAMEL HYPOPLASIAS

X = unobservable 1 = absent 2 = present

Table I-1: Enamel hypoplasias

SKELETON # AGE I C M1 M2 M3 133 0.000 X X X X X 137 0.000 X X X X X 139 0.000 X X X X X 76B-2 0.000 X X X X X 88B 0.000 X X X X X 107 0.000 X X X X X 108 0.000 X X X X X 115 0.000 X X X X X 180/181- 3ASSOC 0.000 X X X X X 198-2 0.000 X X X X X 202-2 0.000 X X X X X 251 0.000 X X X X X 280 0.000 X X X X X 421 0.000 X X X X X VASIJA9-2 0.000 X X X X X 154 0.083 X X X X X 80 0.250 X X X X X 100-2 0.250 X X X X X 149 0.330 X X X X X 104-2 0.500 X X X X X 122 0.500 X X X X X 132 0.500 X X X X X 261 0.500 X X X X X 270 0.500 X X X X X 142 0.750 X X X X X 174-1 0.750 X X X X X 264 0.750 X X X X X 282 0.750 X X X X X 418 0.750 X X X X X

395 419 0.750 X X X X X VASIJA9-1 0.750 X X X X X VASIJA13 0.750 X X X X X 124 1.000 X X X X X 140 1.000 X X X X X 76A 1.000 X X X X X 111A 1.000 X X X X X 128 1.000 X X X X X 174-2 1.000 X X X X X 207 1.000 X X X X X VASIJA12 1.000 X X X X X 129 1.500 X X X X X 143 1.500 X X X X X 203 1.500 X X X X X 76B-1 1.500 X X X X X 79 1.500 X X X X X 83A 1.500 X X X X X 95 1.500 X X X X X 114 1.500 X X X X X 116 1.500 X X X X X 175 1.500 X X X X X 198 1.500 X X X X X 247 1.500 X X X X X 248 1.500 X X X X X 265 1.500 X X X X X 320 1.500 X X X X X 141 2.000 X X X X X 222 2.000 X X X X X 68C 2.000 X X X X X 135 2.000 X X X X X 269 2.000 X X X X X 92 2.500 X X X X X 97 2.500 X X X X X 104-1 2.500 X X X X X 182-2 2.500 X X X X X 229-2 2.500 X X X X X 372-3 3.000 X X X X X 90A 3.000 X X X X X 188 3.000 X X 1 1 X 192-1 3.000 X X X X X 202-1 3.000 X X 2 X X 225 3.000 X X 1 X X 308 3.000 X X 1 X X 405 3.000 X X 2 X X 90B 3.500 1 X 1 X X 380 3.500 X X X X X 130 4.000 2 X X X X 415a 4.000 X X X X X

396 131 4.000 X X X X X 173a 4.000 1 X 1 X X 262 4.000 X X 2 X X 271 4.000 X X X X X 319 4.000 2 X X X X 348 4.000 X X 1 X X 368 4.000 X X X X X 373-2 4.000 X X X X X 403 4.000 X X X X X 82A-2 4.500 X X X X X 110 4.500 X X X X X 230 4.500 X X X X X 268-1 4.500 X X X X X 283 4.500 X X X X X 150 5.000 X X X X X 231 5.000 2 2 X X X 313 5.000 1 X 2 X X 78C-2 5.000 X 2 X X X 182-1 5.000 2 X 2 X X 232 5.000 X X 1 X X 267 5.000 X X X X X VASIJA10 5.500 X X 2 X X 104-3 6.000 2 X X X X 387 6.000 X X X X X 83B 6.500 X X 1 X X 400-1 6.500 X X 2 X X 400-2 6.500 2 X 1 X X 422 6.500 X X X X X 196 7.000 2 X 1 1 X 213 7.000 1 X 1 1 X 239 7.000 2 X 1 X X 266-1 7.000 2 2 1 1 X 76 7.500 X X 2 X X 384 8.000 1 2 1 1 X 101 8.500 X X X X X 134 8.500 2 2 2 X X 256 8.500 1 X 1 2 X 278 9.000 2 X 2 X X 103 9.000 2 X 2 X X 78C-1 10.000 X X 2 2 X 400-3 10.000 2 2 2 2 X 102 10.500 X X X X X 94 10.500 1 1 1 X X 229-1 11.000 1 1 1 1 X 268-2 11.500 X 2 2 2 X 69A-2 12.000 2 2 1 1 X 408 12.000 1 2 1 2 X 326 13.500 2 2 2 1 X

397 377 13.500 1 2 1 2 2 285-2 14.500 2 2 2 1 2 295 15.000 2 X 2 2 X 296 15.000 2 X 1 1 X 100-1 15.000 1 2 2 2 X 100-3 15.000 X X X X X 183 15.000 1 1 X 1 X 191 15.000 X X X X X 214 15.000 2 2 2 2 X 216 15.000 1 X X X X 341 15.000 2 2 2 2 X 404-2 15.000 1 2 1 2 2 411-2b 15.500 1 X 2 2 X 266-2 15.500 X X X X X 69C 16.060 2 2 1 1 X 328 17.500 2 2 X 2 X 433 17.500 2 2 1 2 2 321 17.810 X X X X X 173c 18.000 1 2 1 2 2 263 18.000 2 X 2 1 1 199 18.500 X 2 1 1 1 404-1 19.240 1 X X X X 151 19.630 1 2 2 2 X 127 21.710 2 X 2 2 1 69A-1 21.800 1 2 2 2 X 397-1 22.070 X 2 1 1 1 173b 22.500 1 2 2 1 X 299 23.090 X X X X X 347-1 23.260 X X 2 1 2 88A 24.250 2 2 1 1 1 259 25.290 2 2 2 2 2 72A 26.760 2 2 2 2 2 345 28.080 2 2 2 2 2 240 28.520 1 2 2 2 X 210-1 28.580 2 2 2 2 2 86A 29.040 1 2 1 X 1 184 29.220 X 2 1 X X 304 31.560 1 2 2 1 1 302 31.910 2 2 2 2 2 228 32.190 1 X X X X 417 34.500 X X 1 1 1 301 35.860 2 X 1 X X 260 36.010 X X X X X 155 37.000 2 X 1 1 1 77A 37.400 2 2 X 2 X 432 39.830 X 2 1 2 1 69B 40.610 X X X X X 208 40.720 X X X X X

398 186 40.850 2 2 1 1 1 176 41.930 2 2 2 1 1 370b 42.060 X X X X X 427 42.110 2 X X 1 1 245 42.260 2 2 2 2 2 211 44.730 2 2 1 2 1 371 46.150 1 1 1 1 X 318 47.070 1 1 1 1 1 178 48.130 X 1 X X 2 347-2 48.230 X X X X X 370a 49.370 X X X X X 147 49.500 X X X X X 331 50.030 X 2 X 1 X 126 50.120 1 2 2 2 2 382 50.420 2 2 1 1 1 385a 50.540 X X X X X 68B 50.710 2 2 1 1 1 69E 50.770 1 2 X X X 193 50.770 X X X 2 X 388 50.770 2 X X X X 424 50.770 1 2 X 2 X 426-1 50.770 1 2 X X 1 434 50.770 2 2 X 1 2 396 51.280 1 2 2 2 2 276 51.580 1 2 1 1 1 281 51.830 2 2 1 1 1 185-1 51.910 X 2 1 1 1 68A 51.970 2 2 X 2 X 279 53.000 X 2 1 1 2 257 53.090 1 2 1 1 2 87A 53.460 1 1 1 1 1 187 53.560 1 2 1 1 1 244 53.630 2 2 1 1 X 416 53.730 1 X X 1 1 385b 53.990 X 2 1 1 1 190 54.050 2 2 X X X 298 54.430 1 2 2 1 1 290 54.490 1 2 2 1 2 414-1&2a 54.590 X 2 2 2 1 83C 54.590 1 X 1 1 2 317 54.790 X 2 1 X 1 426-2 54.830 X X X X X 372-1 54.970 2 2 1 2 X 374 55.440 2 2 2 1 1 413-1 55.490 X X X X X 81 55.970 1 1 1 2 2 423 56.070 2 X X X X 376 56.190 X X X X X

399 221 56.390 X X X X X 292 56.650 2 2 1 1 1 294 56.990 X X X X X 398 57.230 X 2 2 2 X 88C 57.260 2 2 1 1 X 227 57.540 X X X X X 78B 57.660 X X X X X 327 57.920 X X X X X 413-2b 57.990 X 2 2 1 1 375 58.390 X X 1 1 1 404-3 58.400 X X X X X 246-1 58.630 X X X X X 153 59.010 2 2 1 1 1 383 59.420 2 2 1 2 1 386 59.680 1 2 1 1 X 395 59.760 X 2 X 1 X 415c 59.990 X X X X X 373-1 60.070 X 2 X 2 X 394-2 60.230 X 1 X 2 X 112 60.880 1 1 1 1 1 204 61.100 X 2 X X X 394-3 61.210 X X X X X 156 61.330 X 2 1 X 1 212 61.610 2 2 1 1 X 347-3 61.710 2 2 1 1 1 164 62.060 X 2 X X X 291 62.570 X 2 X X X 372-2 62.600 X 2 X X X 297-1 63.450 2 2 1 1 X 250 64.320 X 1 1 1 1 300 64.360 X 2 1 X X 78A 64.510 1 2 1 1 1 415b 64.740 X X 1 1 X 246-3 64.850 X X X X X 200 64.920 X X X X X 201 65.320 X X X X X 157 65.640 X X X X X 394-1 65.930 X 2 X 1 X 407 66.090 X 2 X X X 430 66.360 1 1 1 1 1 189b 66.930 X X X X X 343 67.770 X 2 1 1 1 255 67.970 1 2 1 1 1 206 68.010 X 2 1 1 X 226 68.250 X X X X X 420 68.470 X X X 1 X 82B 68.840 1 1 1 1 X 406 69.150 X 2 1 X 1

400 246-2 69.270 1 2 1 1 X 189-1 71.110 X X X X X 428 71.150 X 2 X 1 X 413-2a 71.460 X X X X X 293-1 72.270 X 2 X X X 284 75.060 X X X X X 152 85.360 1 1 X X X 409 110.000 X X X 1 X 411-1 110.000 X X X X X 414-1&2b 110.000 1 1 1 X X

401 APPENDIX J

PROLIFERATIVE LESIONS

X = unobservable 1 = absent 2 = present

Table J-1: Proliferative lesions

SKELETON # AGE Humerus Radius Ulna Femur Tibia Fibula 133 0.000 1 X X X 1 1 137 0.000 1 1 1 1 1 1 139 0.000 1 1 1 1 1 1 76B-2 0.000 X X X X X X 88B 0.000 1 X X X X X 107 0.000 1 1 1 1 1 1 108 0.000 X 1 1 1 1 X 115 0.000 1 1 1 1 1 1 180/181- 3ASSOC 0.000 1 1 1 1 1 1 198-2 0.000 X X X X X X 202-2 0.000 X X X 1 1 X 251 0.000 1 1 1 1 1 1 280 0.000 1 1 1 1 1 X 421 0.000 1 X 1 1 1 X VASIJA9-2 0.000 X X X X X X 154 0.083 2 2 1 X X X 80 0.250 1 X 1 1 1 1 100-2 0.250 1 X X X X X 149 0.330 1 X 1 1 1 1 104-2 0.500 X X X X X X 122 0.500 1 1 1 1 1 1 132 0.500 X X X X X X 261 0.500 X X X X 2 X 270 0.500 1 1 1 1 2 1 142 0.750 2 2 2 2 2 1 174-1 0.750 1 X 1 X X X 264 0.750 X X X X X X 282 0.750 1 X 1 1 1 1 418 0.750 X 1 1 1 X X 419 0.750 1 1 1 X 1 X VASIJA9-1 0.750 X X X X X X

402 VASIJA13 0.750 X X X X X X 124 1.000 X X X X X X 140 1.000 1 1 X 1 1 1 76A 1.000 X X X X X X 111A 1.000 X X X X X X 128 1.000 1 1 1 1 1 1 174-2 1.000 X X X X X 1 207 1.000 1 1 1 1 1 1 VASIJA12 1.000 X X 1 X X X 129 1.500 X 1 1 X X X 143 1.500 1 1 1 1 1 1 203 1.500 X X X X X 1 76B-1 1.500 1 1 1 X 1 1 79 1.500 X X X 1 X X 83A 1.500 X X X X X X 95 1.500 X X X X 2 X 114 1.500 X X 1 X X X 116 1.500 1 1 1 1 1 1 175 1.500 2 1 2 2 2 1 198 1.500 X X X X X 1 247 1.500 1 1 1 1 1 1 248 1.500 X X 1 1 1 1 265 1.500 X X 1 X X X 320 1.500 1 1 1 1 2 1 141 2.000 1 1 1 1 1 1 222 2.000 X X 1 X 1 1 68C 2.000 X 1 1 1 1 1 135 2.000 X X X X X X 269 2.000 1 1 1 1 1 1 92 2.500 X X X X X X 97 2.500 2 2 2 2 2 2 104-1 2.500 X X X X X X 182-2 2.500 X 1 X X X 1 229-2 2.500 1 X X X X X 372-3 3.000 X X 1 X X X 90A 3.000 X X X X X X 188 3.000 1 1 X 2 X X 192-1 3.000 X X X X X X 202-1 3.000 X X X X 1 1 225 3.000 X X X X 2 X 308 3.000 X X X X X X 405 3.000 1 1 1 X X X 90B 3.500 X X X X X X 380 3.500 X X X X X X 130 4.000 X X X X X X 415a 4.000 X X X X X X 131 4.000 X X X X X X 173a 4.000 X X X X X X

403 262 4.000 X X 1 X X X 271 4.000 X X X 1 1 1 319 4.000 1 1 X 1 2 1 348 4.000 1 X X 1 1 1 368 4.000 1 1 1 1 2 1 373-2 4.000 X X X X X X 403 4.000 X X X X 2 1 82A-2 4.500 2 X 2 2 1 1 110 4.500 1 1 1 1 2 2 230 4.500 1 1 1 X X X 268-1 4.500 X X X X X X 283 4.500 1 1 1 1 2 1 150 5.000 1 1 1 1 1 1 231 5.000 X X X X X X 313 5.000 1 1 1 1 2 1 78C-2 5.000 X X X X X X 182-1 5.000 X X X 1 X 1 232 5.000 1 X X X X X 267 5.000 1 1 1 1 1 1 VASIJA10 5.500 1 1 X 1 1 1 104-3 6.000 X X X X X X 387 6.000 X X X 1 X X 83B 6.500 1 1 1 1 1 1 400-1 6.500 1 1 X X 2 2 400-2 6.500 1 X 1 1 2 1 422 6.500 X X X X 1 X 196 7.000 1 1 1 X 1 1 213 7.000 1 1 1 X 1 1 239 7.000 X X X X X X 266-1 7.000 1 X X X 1 X 76 7.500 X X X X X X 384 8.000 1 X X 2 2 X 101 8.500 X 1 1 1 1 1 134 8.500 1 1 1 1 2 2 256 8.500 1 X 1 2 2 2 278 9.000 1 1 1 2 1 X 103 9.000 1 1 1 1 2 1 78C-1 10.000 X X X 1 1 X 400-3 10.000 X X X X X X 102 10.500 X X X 2 1 1 94 10.500 1 1 1 X X X 229-1 11.000 1 1 X 1 2 X 268-2 11.500 X X X X X X 69A-2 12.000 1 1 X 1 1 X 408 12.000 1 1 X 1 1 1 326 13.500 1 1 1 1 1 1 377 13.500 X X X X X X 285-2 14.500 1 X 1 2 1 X

404 295 15.000 2 X X 1 2 1 296 15.000 1 1 1 1 1 1 100-1 15.000 1 1 1 1 2 1 100-3 15.000 1 1 1 1 1 1 183 15.000 1 X X X X X 191 15.000 X X 2 X X X 214 15.000 X 1 X 1 X X 216 15.000 1 1 X X X 1 341 15.000 1 1 1 1 2 2 404-2 15.000 1 1 1 1 1 1 411-2b 15.500 X X X X X X 266-2 15.500 1 1 1 X 2 X 69C 16.060 X X X 2 X 2 328 17.500 X 1 1 2 2 2 433 17.500 X X X X X 1 321 17.810 1 2 2 X 2 2 173c 18.000 X X X X X X 263 18.000 X X X X X X 199 18.500 1 1 1 2 2 2 404-1 19.240 1 1 X 1 2 1 151 19.630 1 1 2 1 2 2 127 21.710 1 1 1 1 1 1 69A-1 21.800 1 1 1 X X X 397-1 22.070 X X X X X X 173b 22.500 1 X X X X 2 299 23.090 1 X X X X X 347-1 23.260 1 1 1 1 X 1 88A 24.250 1 1 1 1 1 1 259 25.290 1 1 2 1 2 2 72A 26.760 1 X 2 2 2 2 345 28.080 1 1 1 1 2 2 240 28.520 1 1 1 1 2 1 210-1 28.580 1 1 1 2 2 2 86A 29.040 1 1 1 X 1 X 184 29.220 X 1 1 X 2 1 304 31.560 1 1 2 2 2 2 302 31.910 1 X X X 1 1 228 32.190 2 1 2 2 2 2 417 34.500 1 X 1 X X X 301 35.860 X X 1 2 2 1 260 36.010 1 1 1 1 1 1 155 37.000 X 1 X 1 2 2 77A 37.400 1 1 1 1 1 1 432 39.830 X X X X X X 69B 40.610 1 1 1 X X 1 208 40.720 X X X X 2 1 186 40.850 1 1 1 1 2 1 176 41.930 1 1 1 1 1 1

405 370b 42.060 X X X X X X 427 42.110 1 1 1 1 2 2 245 42.260 1 1 2 1 2 2 211 44.730 1 1 X X 2 2 371 46.150 1 1 1 1 1 1 318 47.070 1 2 2 2 2 2 178 48.130 1 1 1 1 1 1 347-2 48.230 X 1 1 1 1 2 370a 49.370 X X X X X X 147 49.500 X X X X X X 331 50.030 1 1 1 1 1 1 126 50.120 1 1 1 1 2 1 382 50.420 1 1 1 1 2 1 385a 50.540 1 1 1 1 1 1 68B 50.710 1 1 1 1 1 1 69E 50.770 X X X X X X 193 50.770 1 1 X 1 1 X 388 50.770 X 1 1 1 2 X 424 50.770 1 1 X 1 2 X 426-1 50.770 X X 1 1 1 2 434 50.770 1 X 1 1 1 X 396 51.280 X X X X X X 276 51.580 1 1 1 1 2 2 281 51.830 1 1 1 1 1 1 185-1 51.910 1 X X X 1 1 68A 51.970 1 X 1 1 X 1 279 53.000 1 1 1 X X 1 257 53.090 1 1 1 2 2 2 87A 53.460 1 X X X 1 X 187 53.560 1 1 2 2 2 2 244 53.630 1 1 1 1 2 1 416 53.730 1 1 X X X 2 385b 53.990 1 X X 1 1 1 190 54.050 1 1 1 1 1 1 298 54.430 1 1 X X 2 X 290 54.490 1 1 1 1 1 2 414-1&2a 54.590 X X X X X X 83C 54.590 1 1 2 1 1 1 317 54.790 1 X X X 1 1 426-2 54.830 X X X X 1 X 372-1 54.970 1 1 1 1 2 2 374 55.440 1 1 1 1 1 1 413-1 55.490 X X X 1 1 1 81 55.970 1 X X 1 1 1 423 56.070 X X X X X 1 376 56.190 X X X X X X 221 56.390 X X X X X X 292 56.650 1 1 1 1 1 1

406 294 56.990 1 1 1 1 2 X 398 57.230 1 1 X X X X 88C 57.260 X X 2 1 1 X 227 57.540 X X X X X 2 78B 57.660 1 X X 1 1 1 327 57.920 X 1 1 1 1 X 413-2b 57.990 X X X X X X 375 58.390 1 1 1 1 1 1 404-3 58.400 1 X 1 1 2 X 246-1 58.630 1 1 1 1 1 X 153 59.010 1 X 1 1 1 1 383 59.420 X 1 1 1 2 2 386 59.680 X X X X 2 2 395 59.760 1 1 2 1 2 2 415c 59.990 X X X X X X 373-1 60.070 1 X 1 1 1 1 394-2 60.230 X X X X X X 112 60.880 1 1 2 2 2 2 204 61.100 X X X X 2 2 394-3 61.210 X X X X X X 156 61.330 1 1 X X X X 212 61.610 1 1 1 1 1 1 347-3 61.710 1 1 1 1 1 1 164 62.060 X 1 1 1 1 1 291 62.570 X 2 X 2 2 2 372-2 62.600 1 1 1 1 2 1 297-1 63.450 1 1 1 1 2 1 250 64.320 1 1 2 2 2 2 300 64.360 1 2 2 2 2 2 78A 64.510 1 1 1 X X X 415b 64.740 X X X X X X 246-3 64.850 1 1 1 1 2 1 200 64.920 1 1 1 1 1 2 201 65.320 X X 1 X 1 2 157 65.640 X X X 1 2 2 394-1 65.930 X X X X X X 407 66.090 X X X X X 2 430 66.360 X X 1 2 2 2 189b 66.930 X X X X X X 343 67.770 1 1 1 X X X 255 67.970 1 1 1 2 2 2 206 68.010 X 1 1 X X X 226 68.250 X 1 2 1 2 2 420 68.470 X X X X 2 1 82B 68.840 1 1 1 1 1 X 406 69.150 X X X 1 X X 246-2 69.270 1 1 1 1 1 1 189-1 71.110 X X 1 X 1 1

407 428 71.150 X X X X X X 413-2a 71.460 X X X 1 X X 293-1 72.270 1 1 2 2 2 2 284 75.060 1 1 1 2 2 2 152 85.360 1 1 1 1 1 1 409 110.000 X X X X X X 411-1 110.000 X X X 1 1 X 414-1&2b 110.000 X X X X X X

408 APPENDIX K

STRONTIUM AND OXYGEN ISOTOPE RESULTS

Table K-1: Oxygen and strontium isotope values. Abbreviations: II = Cholulteca II, III = Cholulteca III; L = left, R = right, Mx = Maxilla, Ma = Mandible, I = Incisor, C = Canine, P3 = Premolar 3, P4 = Premolar 4, M1 = Molar 1, M2 = Molar 2, M3 = Molar 3. A strontium value could not be assessed for individual 373-1. Time ID Skeleton Tooth Sex Period ∂18O 87Sr/86Sr

F5537 69A-1 RMaM2 M II -6.52 0.70606

F5538 81 RMxM1 F III -5.78 0.70576

F5539 127 RMxM2 M II -5.16 0.70493

F5540 134 RMxP3 ? III -5.34 0.70597

F5541 151 LmaM1 F II -6.79 0.70438

F5542 178 LMxM1 F II -6.05 0.70680

F5543 182-1 RMxM1 ? III -5.95 0.70600

F5544 186 LmaP4 F III -6.51 0.70450

F5545 196 RMxM1 ? III -5.81 0.70602

F5546 204 LmaP4 F III -5.53 0.70589

F5547 206 RmaM1 M III -6.45 0.70620

F5548 208 LmaM2orM3 M III -6.64 0.70527

F5549 210-1 R?Mx?C M III -5.38 0.70602

F5550 213 RMxP3 ? III -5.91 0.70654

F5551 214 L?MaM1 F III -4.57 0.70560

F5552 216 LMxM2 F III -5.45 0.70600

F5553 232 RmaM1 ? III -4.70 0.70582

F5554 240 RMxM2 M III -6.15 0.70603

F5555 245 LmaM2 F III -4.79 0.70681

F5556 260 LmaM1 M III -6.42 0.70572

F5557 263 LMxM1 ? III -6.00 0.70583

409 F5558 276 RMxP3 M III -5.40 0.70584

F5559 291 RmaM3 F II -5.50 0.70516

F5560 292 RmaM1 M II -2.13 0.70722

F5561 294 LmaC F II -6.42 0.70461

F5562 295 RMxM3 ? II -5.56 0.70528

F5563 299 LmaM2 M II -5.63 0.70534

F5564 301 R?Mx?M1 M II -6.50 0.70475

F5565 302 LMxM1 M II -6.36 0.70505

F5566 328 LmaM1 F III -5.97 0.70465

F5567 343 LMxP3 F III -4.44 0.70593

F5568 347-1 RMxM1 M III -5.89 0.70561

F5569 370a RMxM1or2 M III -5.17 0.70576

F5570 370b LMxM1 M III -5.18 0.70570

F5571 372-2 LMxM1 F II -5.33 0.70571

F5572 373-1 LMxM1 F III -3.16 *

F5573 385 RMxM1 M III -5.10 0.70556

F5574 395 LMxM2 M II -5.39 0.70548

F5575 397-1 LmaM1 F II -5.32 0.70538

F5576 398 LMxM2 F II -6.97 0.70554

F5577 400-3 LmaP4 ? III -5.68 0.70615

F5578 414 LmaM1 F II -6.57 0.70563

F5579 417 LmaM1 M II -5.19 0.70538

F5580 420 LmaP3orP4 M III -6.77 0.70619

F5581 424 LmaP3 M III -5.92 0.70523

F5582 426-1 RmaM1 M III -4.98 0.70581

F5583 428 RmxI2 M III -4.81 0.70528

F5584 432 LmaM1 M III -5.86 0.70553

F5585 433 LMxM3 F III -5.30 0.70548

F5586 Vasija 10 LmaM2 ? III -5.67 0.70595

Mean -5.60 0.70566

410 St Dev 0.8742646 0.0005766

Median -5.65 0.70571

Mode * 0.70576

Table K-2: Strontium values arranged in ascending order.

Time Skeleton Age Sex Period 87Sr/86Sr

151 19.63 F II 0.70438

186 40.85 F III 0.70450

294 56.99 F II 0.70461

328 17.5 F III 0.70465

301 35.86 M II 0.70475

127 21.71 M II 0.70493

302 31.91 M II 0.70505

291 62.57 F II 0.70516

424 50.77 M III 0.70523

208 40.72 M III 0.70527

295 15 ? II 0.70528

428 71.15 M III 0.70528

299 23.09 M II 0.70534

397-1 22.07 F II 0.70538

417 34.5 M II 0.70538

395 59.76 M II 0.70548

433 17.5 F III 0.70548

432 39.83 M III 0.70553

398 57.23 F II 0.70554

385 53.99 M III 0.70556

214 15 F III 0.70560

347-1 23.26 M III 0.70561

411 414 54.59 F II 0.70563

370b 42.06 M III 0.70570

372-2 62.6 F II 0.70571

260 36.01 M III 0.70572

81 55.97 F III 0.70576

370a 49.37 M III 0.70576

426-1 50.77 M III 0.70581

232 5 ? III 0.70582

263 18 ? III 0.70583

276 51.58 M III 0.70584

204 61.1 F III 0.70589

343 67.77 F III 0.70593

Vasija 10 5.5 ? III 0.70595

134 8.5 ? III 0.70597

182-1 5 ? III 0.70600

216 15 F III 0.70600

196 7 ? III 0.70602

210-1 28.58 M III 0.70602

240 28.52 M III 0.70603

69A-1 21.8 M II 0.70606

400-3 10 ? III 0.70615

420 68.47 M III 0.70619

206 68.01 M III 0.70620

213 7 ? III 0.70654

178 48.13 F II 0.70680

245 42.26 F III 0.70681

292 56.65 M II 0.70722

373-1 60.07 F III *

412 Table K-3: Oxygen isotope values arranged in ascending order.

Time Skeleton Age Sex Period ∂18O

398 57.23 F II -6.97

151 19.63 F II -6.79

420 68.47 M III -6.77

208 40.72 M III -6.64

414 54.59 F II -6.57

69A-1 21.8 M II -6.52

186 40.85 F III -6.51

301 35.86 M II -6.50

206 68.01 M III -6.45

294 56.99 F II -6.42

260 36.01 M III -6.42

302 31.91 M II -6.36

240 28.52 M III -6.15

178 48.13 F II -6.05

263 18 ? III -6.00

328 17.5 F III -5.97

182-1 5 ? III -5.95

424 50.77 M III -5.92

213 7 ? III -5.91

347-1 23.26 M III -5.89

432 39.83 M III -5.86

196 7 ? III -5.81

81 55.97 F III -5.78

400-3 10 ? III -5.68

Vasija 10 5.5 ? III -5.67

299 23.09 M II -5.63

295 15 ? II -5.56

413 204 61.1 F III -5.53

291 62.57 F II -5.50

216 15 F III -5.45

276 51.58 M III -5.40

395 59.76 M II -5.39

210-1 28.58 M III -5.38

134 8.5 ? III -5.34

372-2 62.6 F II -5.33

397-1 22.07 F II -5.32

433 17.5 F III -5.30

417 34.5 M II -5.19

370b 42.06 M III -5.18

370a 49.37 M III -5.17

127 21.71 M II -5.16

385 53.99 M III -5.10

426-1 50.77 M III -4.98

428 71.15 M III -4.81

245 42.26 F III -4.79

232 5 ? III -4.70

214 15 F III -4.57

343 67.77 F III -4.44

373-1 60.07 F III -3.16

292 56.65 M II -2.13

Table K-4: Adjusted oxygen values arranged in ascending order.

Time Adjusted Skeleton Tooth Age Sex Period ∂18O O 87Sr/86Sr 69A-1 LMaM1 21.8 M II -6.52 -7.22 0.70606 186 RMxM1 40.85 F III -6.51 -7.21 0.70450 301 LMaM1 35.86 M II -6.50 -7.20 0.70475 420 LMaP3orP4 68.47 M III -6.77 -7.12 0.70619

414 302 RMxM1 31.91 M II -6.36 -7.06 0.70505 208 LMaP4 40.72 M III -6.64 -6.99 0.70527 398 LMaM2 57.23 F II -6.97 -6.97 0.70554 206 LMaP3 68.01 M III -6.45 -6.80 0.70620 151 RMaM2 19.63 F II -6.79 -6.79 0.70438 178 RMxI2 48.13 F II -6.05 -6.75 0.70680 182-1 RMxM1 5 ? III -5.95 -6.65 0.70600 424 RMxM1 50.77 M III -5.92 -6.62 0.70523 213 RMaM1 7 ? III -5.91 -6.61 0.70654 414 RMaM3 54.59 F II -6.57 -6.57 0.70563 196 LMaM1 7 ? III -5.81 -6.51 0.70602 240 LMaP4 28.52 M III -6.15 -6.50 0.70603 294 RMxM2 56.99 F II -6.42 -6.42 0.70461 260 RMxM3 36.01 M III -6.42 -6.42 0.70572 400-3 LMaM1 10 ? III -5.68 -6.38 0.70615 Vasija 10 LMxM1 5.5 ? III -5.67 -6.37 0.70595 263 LMxP3 18 ? III -6.00 -6.35 0.70583 328 RMxP3 17.5 F III -5.97 -6.32 0.70465 295 RMaM1 15 ? II -5.56 -6.26 0.70528 347-1 RMxP3 23.26 M III -5.89 -6.24 0.70561 276 RMxM1 51.58 M III -5.40 -6.10 0.70584 134 LMxM1 8.5 ? III -5.34 -6.04 0.70597 433 RMaM1 17.5 F III -5.30 -6.00 0.70548 299 R?Mx?C 23.09 M II -5.63 -5.98 0.70534 417 L?MaM1 34.5 M II -5.19 -5.89 0.70538 370b RMaM1 42.06 M III -5.18 -5.88 0.70570 370a R?Mx?M1 49.37 M III -5.17 -5.87 0.70576 127 LMxM1 21.71 M II -5.16 -5.86 0.70493 432 RMxM2 39.83 M III -5.86 -5.86 0.70553 291 LMaP4 62.57 F II -5.50 -5.85 0.70516 385 LMaM1 53.99 M III -5.10 -5.80 0.70556 81 LMaM2 55.97 F III -5.78 -5.78 0.70576 372-2 LMaC 62.6 F II -5.33 -5.68 0.70571 426-1 LMxM1 50.77 M III -4.98 -5.68 0.70581 397-1 RMxP3 22.07 F II -5.32 -5.67 0.70538 204 RMxM1or2 61.1 F III -5.53 -5.53 0.70589 245 LMaM1 42.26 F III -4.79 -5.49 0.70681 216 LMxM2 15 F III -5.45 -5.45 0.70600 232 LMaM1 5 ? III -4.70 -5.40 0.70582

415 395 LMxM2 59.76 M II -5.39 -5.39 0.70548 210-1 LMxM2 28.58 M III -5.38 -5.38 0.70602 214 LMxM1 15 F III -4.57 -5.27 0.70560 343 LMxM1 67.77 F III -4.44 -5.14 0.70593 428 LMaM2orM3 71.15 M III -4.81 -4.81 0.70528 373-1 LMaM2 60.07 F III -3.16 -3.16 * 292 LMxM3 56.65 M II -2.13 -2.13 0.70722 Mean -5.60 -6.03 0.705660 St Dev 0.8743 0.92169 0.00058 Median -5.65 -6.07 0.70571

416 VITA

MEGGAN M. BULLOCK KREGER

Education:

2010 Ph.D., Anthropology. The Pennsylvania State University, State College, PA, USA. 2001 M.A., Anthropology. The Pennsylvania State University, State College, PA, USA. 1999 B.S., Anthropology, Summa Cum Laude. Tulane University, New Orleans, LA, USA. 1999 B.A., International Relations, Summa Cum Laude. Tulane University, New Orleans, LA, USA.

Research Interests:

Mesoamerica; human osteology and paleopathology; gender in anthropology; health and the evolution of complex societies.

Research Grants:

2009 Travel Grant, PSU ($600) 2007 Wenner-Gren Foundation ($10,050) 2007 Travel Grant, PSU ($300) 2005 Pennsylvania State University RGSO ($2000) 2004 Sanders Grant, PSU ($1000) 2004 Hill Grant, PSU ($1730) 2004 FAMSI Grant ($9400) 2003 Hill Grant ($2000)

Professional Experience:

2005-2010 Research on the paleodemography and paleopathology of skeletons excavated during the Proyecto Cholula; Strontium and oxygen isotope analyses of Cholula skeletons; Aging Preceramic Mexican skeletons using transition analysis; Pathologies in Preceramic Mexican skeletons

2000-2003 Teaching/Research Assistant, The Pennsylvania State University. State College, PA. Introduction to Biological Anthropology, Anthropology of Gender, Introduction to Archaeology, Environmental Archaeology Lab

Selected Presentations and Publications:

2009 Morbidity and Mortality in a Preindustrial New World City: The Effects of Urbanism on the Postclassic Population of Cholula. Paper presented at XV Coloquio Internacional de Antropología Física Juan Comas. Merida, Mexico.

2008 Reconstructing age at death in the Postclassic Population of Cholula, Puebla: A comparison of transition analysis and traditional methods of aging adult skeletal remains. 77th AAPA Annual Meeting. Columbus, OH.

2007 Reconstructing the age-at-death distribution of the Postclassic population of Cholula: A comparison of transition analysis and traditional aging methods. XIV Coloquio Internacional de Antropología Física Juan Comas. Chiapas, Mexico.